journal proposal / call for papers - ESI Publications

JOURNAL PROPOSAL / CALL FOR PAPERS

JOURNAL OF ASIAN ECONOMICS, ACCOUNTING AND FINANCEEditior-in-Chief: Prof. V. K. Gupta, Indian Institute of Management, Indore

Publisher: Academic Research Foundations, New Delhi (India)

(Accounting)• Auditing and assurance services• Audit quality, audit fees, auditor ’s tenure and auditor ’s

independence• Audit report lag• Auditors’ reporting decisions for accounting estimates• Audit evidence• Pricing initial audit engagements• Audit committees• Effectiveness of of international standards for auditing (ISA)• Internal audit effectiveness and information technology• Effectiveness of external auditor reports• Auditdata analytics• Artificial intelligence and its impact on accounting• Role of auditors in risk assessment• Big data analytics and auditing and accounting• Continuous auditing and role of Big Data• Auditing and accounting in the age of digitalization• E-commerce and auditing• ERP audits• Impact of Blockchain accounting on accountants and auditors‘

skills and practices• Impact of Crypto currencies on accounting and auditing• Information systems and computer auditing• Issues and challenges created by Fintech in auditing procedures• Financial Accounting and reporting• Impact of technology in financial reporting• Financial reporting users and their needs• The future of reporting and its relevance, including digital

reporting• International Accounting• Impact of convergence of International Financial Reporting

Standards(IFRS)• IFRS and cost capital• Capital market research• Earnings management in family versus non family firms

AIMS OF THE JOURNALJournal of Asian Economics, Accounting and Finance (JAEAF)is double blind reviewed international journal that publishes research,intensive articles, and scientific manuscripts focusing on all aspectsof Economics, Accounting and Finance topics.Publication Frequency: QuarterlyTopics of Research Interests(Economics and Finance)• Macroeconomcis• International Economics• Econometrics• Business Economics• Growth and Development• Regional Economics• Tourism Economics• International Trade• Finance• International Finance• Macroeconomic Aspects of Finance• General Financial Markets• Financial Institutions• Behavioral Finance• Public Finance• Asset Pricing• Financial Management• Options and Futures• Taxation, Subsidies and Revenue• Corporate Finance and Governance• Money and Banking• Markets and Institutions of Emerging Markets• Public Economics and Public Policy• Financial Economics• Applied Financial Econometrics• Financial Risk Analysis• Risk Management• Portfolio Management• Financial Econometrics

• Earnings management and corporate governance• Fraud, ethics and corruption• Compliance and value approaches for accounting ethics• Accounting and human rights• Accounting, information technology, and corporate governance• Impact of Fintech on corporate financing decisions• Accounting communication • Public Sector Accounting• Corporate social and environmental accounting and reporting• Integrated and sustainability development and reporting• Carbon accounting and climate change• Ethical issues in accounting and financial reporting• Measurement and reporting of Risks• Sustainability and corporate governance• Measurement and valuation of intellectual capital• Role of professional bodies in the development of accounting

standards• Political issues, political linked companies and accounting practices• Islamic accounting• Accounting and Financial Education• Cost and Managerial accounting and control practices• Strategic managerial accounting• Behavioural accounting• Budgeting practices and their behavioural implications• Contemporary performance measurement and management (PMM)• Environmental cost management and reporting• Accounting information systems• Taxation and tax avoidanceReview Process and Acceptance of manuscriptsJAEAF follows a double blind review process of all the manuscripts.The review process time may take between 8 to 12 weeks. Thejournal follows a review form for the reviewers. Reviewers may alsoadd their comments in the second section of the Review Form.Having reviewed the paper, the reviewers will be requested to makeany of the following decisions:• Accept as it is• Accept with minor revisions• Accept with major revisions• Send me the revised paper• RejectJAEAF publishes original and unpublished manuscripts. All themanuscript submitted to the journal should be the original work ofthe authors and the manuscript should not be under review of anyother journal.

GUIDELINES TO THE AUTHORS

JAEAF: publishes research articles only in English language. It followsHarvard style of citation in the text (e.g. Joshi, 2018). For preparingthe manuscript, authors should use the following guidelines:• When single author is used, the author’s name (without initials)

and the year of publication• When two authors are used: both authors’ names and the year of

publication, and• When three or more authors are used: first author’s name followed

by et al. and the year of publication.• Authors are advised to see that every reference cited in the text

should be presented in the reference list (and vice versa).Reference to a journal articleAll references should be listed in alphabetical orders. The style of

reference at the end should be in the following way:Joshi, P L., (2001) Diffusion of new management accounting practices:

the case of India, Journal of Asian Economics, Accounting andFinance, 10 (1), pp. 85-109.

More than one reference from the same author(s) in the same yearmust be identified by the letters ‘a’, ‘b’, ‘c’, etc., placed after theyear of publication.

Reference to a BookStrunk, Jr., W., White, E.B.,( 2000) The Elements of style, fourth ed.

New York: Longan.ProofsIt is the responsibility of the first author or corresponding author tocorrect theproofs of the accepted article which will be sent electronically.The same should be returned within three weeks of the receipt.Corrections should be restricted to typesetting errors only; Anyadditional changes will be charged to the authors. No late or last-minutecorrections will be entertained.Reprints: A copy of the the published paper in PDF file will be sent tothe authors which will be reprint copy of the published article.Copyright: Once the manuscript is accepted, it will be the responsibilityof the corresponding author to send the copyright form, signed byeach author and co-authors.Prepation of manuscript• The size of the manuscripts submitted to JAEAF should be between

3,000 to 6, 000 words.• The title page should include title of the manuscript, all authors

names, institutional affiliation, full address, email addresses.• Title of the manuscript should be appealing and concise. Do not

include any mathematical sign in it.• An Abstract of not more than 200-300 words should be prepared.• All pages in the manuscripts should be properly paged.• If any footnotes are used, they should appear at the bottom of the

text page where they are quoted.• All tables and figures should be included at the end, just after the

references. All tables and figures should be numbered consecutively.• All acknowledgements should be included just before the

refrencences.

Electronic Submissions should be sent in MS Word format to:[email protected]

CONTENTS

1. Expanding the A Priori Procedure (APP) to Address .................... 119-136Proportions

David Trafimow, Hui Li, Tonghui Wang,Liqun Hu, Cong Wang and Abigail Rodriguez

2. Socio-Economic and Demographic Analysis of International ..... 137-161Migration from Rural Punjab: A Case Study of Patiala District

Gurinder Kaur, Gian Singh, Dharampal,Rashmi, Rupinder Kaur, Sukhvir Kaur and Jyoti

3. Effect of Lending Rate on the Performance of Nigerian ............. 163-180Deposit Money Bank

Owolabi, A

4. Effect of Macro-economic Variables on Profitability of ............... 181-198Selected Commercial Banks in Rwanda

Daniel Twesige and Faustin Gasheja

5. On The Determinants of Unemployment Rate in Nigeria: .......... 199-225Evidence from Fully Modified OLS andError Correction Model

Adenomon, M. O.; Okoro-Ugochukwu,N. A. and Adenomon, C. A.

AFRO-ASIAN JOURNAL OFECONOMICS AND FINANCE

Volume 1 • Number 2 • 2020

EXPANDING THE A PRIORI PROCEDURE (APP) TOADDRESS PROPORTIONS

David Trafimow, Hui Li, Tonghui Wang, Liqun Hu,Cong Wang and Abigail Rodriguez

New Mexico State University, Las Cruces, NM 88003-8001Correspondence author E-mail: [email protected]

Received: 4 July 2020; Revised: 14 August 2020; Accepted: 20 August 2020; Online: 29 December 2020

Abstract: The a priori procedure (APP) helps researchers determine the necessary sample size tosimultaneously reach goals pertaining to precision (the closeness of sample statistics to thepopulation parameters they are used to estimate) and confidence (the probability of being withinthe precision specification). However, an important APP limitation is the lack of equationspertaining to proportions. Researchers sometimes have binary data, where the crucial issue is theproportion of participants that fall into a category, such as the proportion of students who graduatefrom an academic program. The present research addresses the limitation, with research implications.

Keywords: a priori procedure, proportions, precision, confidence level

Wordcount: 5257

Based on a growing dissatisfaction with traditional null hypothesissignificance testing, traditional pvalues, and traditional confidenceintervals, researchers continue to search for alternatives.1 One suchalternative is the a priori procedure (APP) (Trafimow, 2017; Trafimow &MacDonald, 2017; Trafimow, Wang, & Wang, 2019; in press). The APP isdifferent from traditional frequentist procedures because the maininferential work is performed prior to data collection. There is no necessityto calculate postdata pvalues or confidence intervals, though there isnothing in the APP to prevent the researcher from performing thosecalculations.

Conceptually, the a priori procedure hinges on the researcher asking adifferent question than normal. Typical inferential questions might be"What hypotheses can I reject or accept?" or "What kind of interval can Iconstruct that likely will contain the population parameter of interest?" Asthe many articles in the recent special issue of The American Statisticianshow, it is highly debatable whether inferential statistical procedures canprovide valid answers to questions such as these. There is no need to engage

Journal of Asian Economics, Accounting and FinanceVol. 1, No. 2, 2020, 119-136© ESI Publications. All Right ReservedURL : www.esijournals.com

120 Journal of Asian Economics, Accounting and Finance © 2020 ESI

that debate here;it can be sidestepped by asking different sorts of inferentialquestions.

To commence with the questions, it is easy to imagine someone askinga statistician about the necessity to obtain large samples, or at least samplesas large as is consistent with feasibility. The statistician doubtless wouldanswer that the larger the sample, assuming random and independentsampling, the more the sample resembles the population. Thus, the largerthe sample, the more confident the researcher can be that the sample statisticof concern is close to its corresponding population parameter. In turn,however, this brings up two more questions.

• Precision: "How close is close?

• Confidence: "How confident is confident?

The researcher who uses the APP decides, prior to data collection, thedesired level of precision and the desired level of confidence. Once thosedecisions have been made, APP equations can be used to find the samplesize needed to meet specifications.

For illustration, consider the simplest case where a researcher collectsa sample of participants from a normally distributed population and isinterested in the sample mean; but with the desideratum that the samplemean to be obtained will be reasonably likely to be reasonably close tothe population mean. Trafimow (2017) provided an accessible proof ofEquation 1:

2

(1 ) / 2;

Czn

f(1)

where f is the precision in terms of the fraction of a standard deviationwithin which the researcher wishes the sample mean to be of the populationmean, z

(1–c)/2 is the zscore that corresponds to the degree of confidence the

researcher wishes to have that the sample mean will be within f of thepopulation mean, and n is the necessary sample size to meet the precisionand confidence specifications.

For example, suppose the researcher wishes the sample mean to bewithin twotenths of the population mean, and wishes to be 95% confidentof being that close. To find the necessary sample size, it is merely necessaryto instantiate 0.2 for f and to find the zscore that corresponds to 95%confidence (it is 1.96). Thus, the necessary sample size to meet specifications

is:

21.96

96.04 970.2

(it is customary to round upwards). If the researcher

Expanding the a Priori Procedure (APP) to Address Proportions 121

goes on to collect 97 participants, she can be assured that her sample meanhas a 95% chance of being within twotenths of a standard deviation of thepopulation mean. Of course, researchers are usually interested in multiplemeans rather than simply a single mean, or in differences in means, or insamples obtained from skewed distributions, or in distribution shapes; butthese have been addressed by subsequent research.2

The focus on obtaining the necessary sample size may seem to implythat the APP is the same as power analysis. But this is not true for bothgeneral and specific reasons (see Trafimow & Myüz, 2019 for a morecomplete discussion). The general reason is that the typical goal of poweranalysis is to find the sample size needed to have a good chance of obtaininga pvalue below a threshold level; whereas the APP goal is to find the samplesize needed to meet specifications for precision and confidence, with noexpectation of computing a pvalue after data collection. But there are twovery specific differences too. First, power analysis is strongly influencedby the expected effect size; but the APP is not. Second, the APP is stronglyinfluenced by the desired precision; but power analysis is not. As a dramaticexample (Trafimow & Myüz, 2019), suppose that the expected effect size is0.8. According to a power analysis, only 13 participants would be neededto exceed 80% power; whereas according to the APP, the precision wouldbe an atrocious value of 0.54.

APP EQUATIONS FOR PROPORTIONS AND RELATED ISSUES

Previous APP work has assumed continuous data but sometimesresearchers are interested in binary data. For example, a marketer mightbe interested in the proportion of people exposed to an ad who buy theproduct. A problem the marketer might have is in deciding how manyparticipants to collect to have confidence that the obtained proportion willbe close to the population proportion. Or an economist might be interestedin the unemployment rate. Again, there is the issue of the necessary samplesize to render theeconomist confident that the sample proportion ofunemployed people is close to the population proportion.

We propose two Methods for performing the requisite calculations.Both methods are based on the normal approximation to the binomialdistribution. Method 1 has the advantage of being easy to use; but is lessexact. Method 2 is more difficult to use; but is more exact. Both Methodsuse the following variables:

• p is the sample proportion of successes, not to be confused with

the population proportion of successes p;


• q is the sample proportion of failures, ˆ ˆ1 ;q p

• E is an arbitrary closeness designation, so that the best unbiased

estimator p is within E of p, with a predetermined probability;

• z�/2 is the zscore such that the probability of standard normal

Z > z�/2 is �/2. For example, if ��= 0.05, then z�/2

= 1.96;

• n is the necessary sample size to meet specifications for precisionand confidence.

With the foregoing in mind, let us proceed to Equation 2 (Method 1)and Equation 3 (Method 2).

Derivations of Equations 2 and 3 can be found in Appendix A andAppendix B, respectively. Let us commence with Method 1, exemplifiedby Equation 2:

2/ 2

2

ˆ ˆ.

pqzn

E(2)

As an example, suppose the researcher does not have a strong reasonfor making a prediction about the proportion of successes, and wishes tobe conservative in her calculations. Equation 2 will render the largest value

for n when p is set at 0.5. Suppose that we set E = 0.03, and confidence at

the traditional 95% level so that z�/2 = 1.96. In that case, according to Equation

2, the necessary sample size is as follows: 2

2

0.25 1.961067.11 1068.

0.03 This

may seem like an excessive sample size; but remember it is based on themost conservative scenario possible. For comparison, a less conservativescenario might be that the percentage of successes will be 0.20, in whichcase the necessary sample size would reduce to 682.95 � 683. Another wayto reduce the necessary sample size is to insist on less precision, such asletting E = 0.04 instead of E = 0.03. We will discuss this issue further in thesubsequent subsection.

To use Method 2, it is necessary to use Equation 3 rather thanEquation 2:

2/ 2 2 2 2 22

ˆ ˆ ˆ ˆ ˆ ˆ[ 2 (1 4 )].2

zn pq E p q E pq

E(3)

To see that Method 2 can result in different, though comparable, valuesrelative to Method 1, let us consider the same example where E = 0.03 and


ˆ ˆ 0.25pq . The resulting value is 1063.27 (rounding upwards renders 1064),

as opposed to 1068 using Method 1.

THE EFFECTS OF DESIRED PRECISION AND PROPORTION OFSUCCESSES ON SAMPLE SIZE

Because Method 2 is more precise than Method 1, we used Method 2 toexplore the effects of desired precision and proportion of successes on thesample size necessary to meet specifications. Figure 1 illustrates these

effects. In Figure 1, p varies from 0.01 to 0.50 along the horizontal axis;

and E was set at 0.01, 0.02, 0.03, and 0.04. The figure shows that sample

sizes increase as p decreases (deviates from the middle value of 0.5) and

as E decreases (the researcher demands increased precision). There also isan interactive effect whereby the importance of demanding increasedprecision is dramatically augmented when p ? increases.

ADJUSTING E TO ACCOUNT FOR EXTREME VALUES FOR P^

Imagine that the sample proportion of successes is 0.5 versus the extremevalue of 0.01. Let us also imagine remaining with E = 0.03 as the precisionlevel. In the context where the sample proportion of successes is 0.5, where

the product of ˆ ˆ 0.25pq , using E = 0.03 may be sensible. But in the context

of the 0.01 value, where the product of ˆ ˆpq = 0.0099, using E = 0.03 would be

ridiculous. In those cases where the researcher has reason to suspect shemay obtain an extreme value for the proportion of successes, it might beworthwhile to adjust E accordingly.

One way to make the adjustment is to consider E as a proportion of

ˆ ˆpq , as in Equation 4 below:

Precision as a proportion of ˆ ˆ .ˆ ˆE

Epq P

pq (4)

To use Equation 4, consider the maximal case where ˆ 0.5,p so that p

ˆ ˆ 0.25,pq and 0.03

0.12.0.25

EP More generally, staying with the maximal

case, Equation 4 can be simplified: PE = 4E. The idea, then, is to imagine the

precision that would be desired in the maximal case, say, E = 0.03, and thenobtain an adjusted value for E to account for an extreme value for p . Too


continue, let us imagine an extreme case where we expect p to be near

0.01, so that ˆ ˆpq � 0.0099. To adjust E for this extreme value, keeping PE

constant, we can multiply ˆ ˆpq by PE. That is, the new value for E equals

0.0099�0.04 = 0.001188. More generally, the adjusted value for E can beobtained using Equation 5 below:

ˆ ˆ.EAdjAdjusted E E P pq (5)

Figure 2 shows how EAdj

varies with E and p . As p increases and as E

increases, so does EAdj

.

SIMILARITIES AND DIFFERENCES IN THE TWO METHODS

Let us consider the difference between the sample sizes indicated by thetwo methods. Figure 3 shows this difference. The vertical axis displays thesample size indicated by Method 1 minus the sample size indicated byMethod 2, setting E at 0.01, 0.02, or 0.03. When the curves are in positiveterritory, it means that the sample size indicated by Method 2 is less thanthe sample size indicated by Method 1; whereas when the curves dip intonegative territory, the reverse is true. Figure 3 shows that (a) Method 2renders smaller samples sizes than does Method 1 when the expectedproportion of successes is not an extremely low value and (b) the differencebetween the two methods is minimal when the expected proportion ofsuccesses is not an extremely low value. Therefore, we recommend thatMethod 2 be used more generally than Method 1; but Method 1 can beused, with reasonable accuracy, provided that the expected proportion ofsuccesses is not an extremely low value.

There is another way to compare Method 1 and Method 2. Considerthe condition necessary for Method 2 to render a smaller value than Method1, for the necessary sample size to meet specifications. This condition ismet whenever Inequation 6 is true:

2 ˆ ˆ1 8.

4

pqE (6)

ESTIMATING PRECISION, A POSTERIORI

Figure 1 shows that it takes a larger than typical sample size, under manycombinations of E and p , to meet specifications for precision andconfidence. Researchers often might not be able to collect such a largesample size. Is there a way to estimate precision based on the sample size


that the researcher finds feasible to obtain? There is a way to estimateprecision E . It is merely a matter of algebraically rearranging Equation 2to obtain Equation 7:

2/ 2 / 22

/ 2

ˆ ˆ ˆ .4

z zE npq

n z(7)

According to Equation 7, once the researcher obtains the sample success

rate p , she can estimate the precision level at the sample size used.

Figure 4 shows E , along the vertical axis; as a function of p , along the

horizontal axis; and at different sample sizes (n = 100, n = 200, n = 300, and

n = 400). As n and p increase, so does E . A caveat, however, harks back to

the subsection on E as a proportion of the ˆ ˆpq product. We saw earlier that

as this product decreases, E also must decrease too, to remain sensible inthe context of the product. Thus, the most important message illustratedby Figure 4 pertains to the sample size. It takes a large sample size to haveimpressive precision. In Figure 4, a sample size of 400 is insufficient to

bring E even to the 0.04 level, when ˆ ˆpq is maximized at 0.25. And being

95% confident that the p is higher or lower than p by 0.04 is not impressive.

DIFFERENCE IN PROPORTIONS ACROSS INDEPENDENT SAMPLES

Sometimes researchers wish to compare proportions across independentsamples. For example, an economist may wish to determine the proportionof males versus females in executive positions. The question to be answeredis, "how many males and females need to be collected to reach a particulardegree of precision for the difference in proportions between the twogroups?"

The answer to the question can be obtained using Equation 8 below,

derived in Appendix C; where 1p is the proportion of successes in the first

sample, 1q is the proportion of failures in the first sample, 2p is the

proportion of successes in the second sample, and 2q is the proportion offailures in the second sample:

2/ 2 1 1 2 2

2

ˆ ˆ ˆ ˆ( )z p q p qn

E(8)


For the sake of illustration, it is useful to consider 1 1 2 2ˆ ˆ ˆ ˆ( )p q p q as a single

term. That is, it is the sum of products or SOP. Thus, Equation 8 can be rewritten as Equation 9, below, that is convenient for constructing figures:

2/ 2

2 .z SOP

nE

(9)

The reason Equation 9 is convenient for figures is that SOP is at its

maximum value when 1 2ˆ ˆ 0.50.p p In that case the SOP = 0.50.

Figure 5 illustrates the minimum sample size needed for one of thegroups, along the vertical axis; as a function of SOP along the horizontalaxis; and with E set at 0.01, 0.02, 0.03, or 0.04. The more precision that isdemanded, particularly as SOP approaches 0.50, the larger the minimumsample size requirement. Because extremely small values for E result inminimum sample sizes that may appear prohibitive and render the curvesfor more reasonable values for E difficult to see, Figure 6 only providescurves for when E is set at 0.03 or 0.04.

Let us consider the example, but with specific values. Suppose againthat a researcher wishes to test the proportion of males versus femaleswho like a new social media presentation style. The researcher has no priorinformation and consequently has no idea what the resulting proportionsare going to be. Thus, the worstcase scenario is to assume the proportionsare 0.5 for both groups so SOP is maximized at 0.50. In addition, supposethe researcher wishes to use E = 0.03. In that case, the minimum samplesize for one of the groups would be

2 2/ 2 1 1 2 2

2 2

ˆ ˆ ˆ ˆ( ) 1.96 0.52134.222 2135.

.03

z p q p qn

E

In contrast, suppose that prior research had been performed, so theresearcher had reason to suspect that the proportion of males would beonly 0.10 and the proportion of females would be only 0.2. In that case,SOP = (0.1�0.9) + (0.2�0.8) = 0.25. Thus,

2 2/ 2 1 1 2 2

2 2

ˆ ˆ ˆ ˆ( ) 1.96 0.251067.111 1068.

.03

z p q p qn

E

The presence of prior research that removes the necessity to use theworstcase scenario dramatically reduces the minimum necessary samplesize.


DISCUSSION

The point was made earlier that researchers sometimes have binary dataand are interested in the proportion of successes, as opposed to beinginterested in a sample mean. Until now, an important APP limitation hasbeen the lack of equations suitable for application to proportions. Thepresent equations address this limitation. But in addition to addressingthe limitation, there are other potential gains too, particularly with respectto reproducibility.

Typically, researchers define a successful replication as occurring whena statistically significant pvalue has been obtained in an original andreplication experiment. Trafimow (2018) provided a comprehensivediscussion of the problems inherent in this conceptualization and proposedan alternative. Trafimow suggested that because it is impossible to replicateall systematic aspects of a study (e.g., the time, participants, and so on willnecessarily differ), it is worth distinguishing replication in the real universefrom replication in an idealized universe that cannot be realized but canbe imagined. In the idealized universe, all systematic aspects of the originalstudy are duplicated in the replication study, with the only differencesbetween the two studies being those due to randomness. Interestingly, APPequations can be made to yield the probability of replication in the idealizeduniverse. At first glance, this may not seem to mean that much, asresearchers tend to be interested in replication in the real universe. ButTrafimow showed that the probability of replication in the idealizeduniverse sets an upper limit on the probability of replication in the realuniverse. Consequently, if the probability of replication in the idealizeduniverse is low, the probability of replication in the real universe is evenlower. The pessimistic implications of APP equations for the requisitesample sizes needed for satisfactory precision are even more pessimisticin the context of the replication issue.

Because the present equations were not available to Trafimow (2018),he was unable to explore the probability of replication with respect tostudies concerned with proportions. It is now possible to address that

limitation. For illustrative purposes, suppose we set p = 0.5 and E = 0.03.In that case, after algebraic rearrangement, Equation 1 reduces to the

following: / 2 .16.67

nz From there, it is merely necessary to find the

corresponding area under the standard normal distribution to obtain theprobability. Once that probability is obtained, it can be squared to obtainthe probability of replication in an idealized universe. For example, suppose


Figure 1: The sample size necessary to meet specification for precision and confidenceis expressed along the vertical axis, as a function of the expected value for p

expressed along the horizontal axis (commencing at 0.01), at four levels ofprecision (E = 0.01, 0.02, 0.03, or 0.04), and with the confidence

level set at the traditional 95% level throughout

Figure 2: The adjusted E expressed along the vertical axis, as a function of theexpected value for p expressed along the horizontal axis (commencing at 0.001),at three levels of precision (E = 0.01, 0.02, or 0.03), and with the confidence level

set at the traditional 95% level throughout


Figure 3: The sample size difference engendered by Method 1 and Method 2 (TheMethod 1 value minus the Method 2 value) is expressed along the vertical axis, as a

function of the expected value for p expressed along the horizontal axis(commencing at 0.01), at three levels of precision (E = 0.01, 0.02, or 0.03), and with

the confidence level set at the traditional 95% level throughout

Figure 4: The estimated value for E is expressed along the vertical axis, as a functionof the sample value for p expressed along the horizontal axis, at four sample

sizes (n = 100, 200, 300, or 400), and with the confidence level set atthe traditional 95% level throughout


Figure 5: The sample size necessary to meet specification for precision and confidenceis expressed along the vertical axis, as a function of the sum of the products (SOP)

expressed along the horizontal axis (commencing at 0.01), at four levels of precision(E = 0.01, 0.02, 0.03, or 0.04), and with the confidence level set at the

traditional 95% level throughout

Figure 6: The sample size necessary to meet specification for precision and confidenceis expressed along the vertical axis, as a function of the sum of the products (SOP)

expressed along the horizontal axis (commencing at 0.01), at two levels of precision(E = 0.03, or 0.04), and with the confidence level set at the

traditional 95% level throughout


Figure 7: Probability of replication in an idealized universe is expressed along thevertical axis as a function of the sample size expressed along the horizontal axis

(commencing at 10). The proportion of successes is set at 0.50 and theprecision (E) is set at 0.03

we set n at 200, 400, 600, 800, or 1000. The resulting probabilities ofreplication are 0.36, 0.59, 0.74, 0.83, and 0.89, respectively. Figure 7 illustratesthis effect, letting the sample size range from 10 to 1000 along the horizontalaxis. As Figure 5 shows, a sample size of 728 is necessary merely to get theprobability of replication past 80%, and even a sample size of 1000 isinsufficient to get the probability of replication past 90%. As we pointedout earlier, because replication probabilities in the idealized universe setan upper bound on replication probabilities in the real universe, theforegoing conclusionspessimistic as they aretend towards optimismrelative to replication probabilities in the real universe.

CONCLUSION

The present equations address an APP limitation pertaining to proportions,by explicating how researchers can find the sample size needed to meetspecifications for precision and confidence. In addition, we suggested away to adjust the desired precision to apply sensibly to atypically lowproportions. Furthermore, when there is a feasibility issue limiting thesample size that can be obtained, the APP nevertheless can be used toestimate the precision that corresponds with the feasible sample size.


Finally, the present advancement of the APP can be applied to Trafimowʹs(2018) notion of replication in an idealized universe. Thus, the presentequations, by extending APP capabilities to research concerned withproportions, constitute an advance.

Notes

1. Hubbard (2016) and Ziliak and McCloskey (2016) provided reviews. Also see a recentdiscussion in Basic and Applied Social Psychology about how significance testing inflateseffect sizes (Grice, 2017; Hyman, 2017; Kline, 2017; Locascio 2017a; 2017b; Marks, 2017).

2. There are many recent APP advances (e.g., Trafimow, 2017, Trafimow & MacDonald,2017; Trafimow, 2018; Trafimow, 2019a; 2019b; Trafimow & Myüz, in press; Trafimowet al., 2019a; 2019b; Wang, Wang, Trafimow, & Myüz, 2019)

3. There is no reason to illustrate what happens when p exceeds 0.5. The designation ofp and q is arbitrary, and we are using p to designate the proportion � 0.50.

References

Grice, J. W. (2017). Comment on Locascioʹs results blind manuscript evaluation proposal.Basic and Applied Social Psychology, 39(5), 254255.

Hubbard, R. (2016). Corrupt research: The case for reconceptualizing empirical managementand social science. Los Angeles, California: Sage Publications.

Hyman, M. (2017). Can ʹresults blind manuscript evaluationʹ assuage ʹpublication biasʹ?Basic and Applied Social Psychology, 39(5), 247251.

Kline, R. (2017). Comment on Locascio, results blind science publishing. Basic and AppliedSocial Psychology, 39(5), 256257.

Locascio, J. (2017a). Results blind publishing. Basic and Applied Social Psychology. 39(5),239246.

Locascio, J. (2017b). Rejoinder to responses to ʺresults blind publishing.ʺ Basic and AppliedSocial Psychology. 39(5), 258261.

Marks, M. J. (2017). Commentary on Locascio 2017. Basic and Applied Social Psychology.39(5), 252253.

Trafimow, D. (2017). Using the coefficient of confidence to make the philosophical switchfrom a posteriori to a priori inferential statistics. Educational and PsychologicalMeasurement, 77(5), 831854.

Trafimow, D. (2018). An a priori solution to the replication crisis. Philosophical Psychology,31(8), 11881214.

Trafimow, D. (2019a). Five nonobvious changes in editorial practice for editors and reviewersto consider when evaluating submissions in a post p < .05 universe. The AmericanStatistician, 73(sup1), 340345.

Trafimow, D. (2019b). My ban on null hypothesis significance testing and confidenceintervals. In V. Kreinovich and S.Sriboonchitta (Eds.), Structural changes and theireconomic modeling (pp. 3548). Cham, Switzerland: Springer.

Trafimow, D. (2019c). What to do instead of null hypothesis significance testing or confidenceintervals. In V. Kreinovich, N. N. Thach, N. D. Trung, and D. Van Thanh (Eds.), Beyondtraditional probabilistic methods in econometrics (pp. 113128). Cham, Switzerland:Springer.


Trafimow, D., & MacDonald, J. A. (2017). Performing inferential statistics prior to datacollection. Educational and Psychological Measurement, 77(2), 204219.

Trafimow, D., & Myüz, H. A. (2019). The sampling precision of research in five major areasof psychology. Behavior Research Methods, 51(5), 20392058.

Trafimow, D., Wang, T., & Wang, C. (2018). Means and standard deviations, or locationsand scales? That is the question!New Ideas in Psychology, 50, 3437.

Trafimow, D., Wang, T., & Wang, C. (2019). From a sampling precision perspective, skewnessis a friend and not an enemy! Educational and Psychological Measurement, 79(1), 129150.

Trafimow, D., Wang, C., & Wang, T. (2020). Making the a priori procedure (APP) work fordifferences between means. Educational and Psychological Measurement. 80(1), 186198.

Wang, C., Wang, T., Trafimow, D., & Myüz, H. A. (2019). Desired sample size for estimatingthe skewness under skew normal settings. In V.Kreinovich and S.Sriboonchitta (Eds.),Structural changes and their economic modeling (pp. 152162). Cham, Switzerland:Springer.

Ziliak, S. T., & McCloskey, D. N. (2016). The cult of statistical significance: How the standarderror costs us jobs, justice, and lives. Ann Arbor, Michigan: The University of MichiganPress.

To cite this article:

David Trafimow, Hui Li, Tonghui Wang, Liqun Hu, Cong Wang and Abigail Rodriguez.Expanding the a Priori Procedure (APP) to Address Proportions. Journal of AsianEconomics, Accounting and Finance, Vol. 1, No. 2, 2020, pp. 119136


Appendix A: Method 1

Let X be a binomial random variable with parameters n and p, denoted as X ~ B(n, p),where X is the number of success in nindependent and identical trials with success rate p.The mean and variance of X are

E[X] = np and Var[X] = npq,

where q = 1 – p.

A point estimator of p is P = X/n. By the Central Limit Theorem, for n sufficiently large, Pis approximately normally distributed with mean p, and variance pq/n. Therefore, we canassert that

/ 2

ˆ1 .#

//2

P pP z Z z with Z

pq n (A1)

When n is large, very little error is introduced by substituting the point estimate ˆ /p x n forp on the radical sign so that

/2 / 2

ˆ ˆ ˆ ˆˆ ˆ 1 .pq pq

P P z p P zn n

Let E be the error in estimating the p by P , then

/2

ˆ ˆ.#

pqE z

n(A2)

Solving for n in Equation A2, we obtain

2/ 2

2

ˆ ˆ,

pq zn

E

which is Equation 2.

Appendix B: Method 2

On the other hand, by solving for p in the quadratic inequality given in Equation A1, weobtain

22 2 2/ 2 / 2 /2 2

0 2/ 2

ˆ ˆ ˆ2 2 4 1

.#

2 1

a z zP P P

n n np

z

n

(A3)

In practice, we only use ʺ+ʺ sign of the roots p0, which will be given as a remark at the end

of this appendix. Therefore,

2 2/2 / 2

2 2/ 2 / 2

ˆ ˆ2 2 1 .

1 1

z zP P

n nP E p Ez z

n n


where

2/ 2 / 22 2/ 2

ˆˆ.

41

z zPQE

nz n

n

#(A4)

If we have previous sampled data, we should use sample proportion p to replace P and

ˆ ˆ1q p to replace Q in Equation A4. Then solving for n in Equation A4, we obtain

2 2 2 2 2/ 2

2

ˆ ˆ ˆ ˆ ˆ ˆ2 (1 4 ),

2

z pq E p q E pqn

E

which is Equation 3.

Remark. The following is the proof that we only use the ʺ+ʺ sign in Equation A3.

The sample size n must be a positive integer, and it depends on the top part (without 2/2z

of the fraction above.

For the ʺʺ case, let

2 2 2 2ˆ ˆ ˆ ˆ ˆ ˆ2 (1 4 ) 0,pq E p q E pq

by moving the square root part to another side and taking square at both sides, we have

4E4 – E2 > 0,

factoring the lefthand side, we get

E2 (2E + 1) (2E – 1) > 0.

Then, we can see this inequality holds only when E > 0.5 = 50%. In real applications, theerror for the proportion estimation is 5% or less.

Appendix C: Estimation of difference of proportions p1 – p

2

Let X1, X

2 beindependent random variables such that X

i ~ Binomial (n

i, p

i), i = 1,2.

We know that the best unbiased estimators of p1 and p

2 are, respectively,

1 21 2

1 2

ˆ ˆx x

p and pn n

with means

1 21 2ˆ ˆp pµ p and µ p

and variances

1 2

2 21 1 2 2ˆ ˆ

1 2

.p p

p q p qand

n n

Therefore the best unbiased estimator of 1 2p p is 1 2ˆ ˆp p with variance

1 1 2 2

1 2

.p q p q

n n


By Central Limit Theorem, the limiting distribution of 1 2ˆ ˆp p is normal so that

1 1 2 2 1 1 2 21 2 / 2 1 2 1 2 / 2

1 2 1 2

ˆ ˆ ˆ ˆ 1 .p q p q p q p q

P p p z p p p p zn n n n #(1)

Let n = min{n1, n

2} and

1 1 2 2/ 2 .

p q p qE z

n n Then we obtain from Equation (1),

1 2 1 2 1 2ˆ ˆ ˆ ˆ( ) 1 .P p p E p p p p E (2)

Now for given confidence level 1 – � and the error, E, associated with the estimator, we cansolve for n, which is given by

2/ 2 1 1 2 2

2

ˆ ˆ ˆ ˆ( ).

Z p q p qn

E#(3)

Note that the 1p and 2p in Equation (3) can be approximated by previous data. If no

approximation of 1p and 2p is known, then 1p = 2p = 0.5 is used in Equation (3).

SOCIO-ECONOMIC AND DEMOGRAPHIC ANALYSISOF INTERNATIONAL MIGRATION FROM RURALPUNJAB: A CASE STUDY OF PATIALA DISTRICT

Gurinder Kaur1, Gian Singh2, Dharampal3, Rashmi4,Rupinder Kaur5, Sukhvir Kaur6 and Jyoti7

1Professor, Department of Geography, Punjabi University, Patiala, E-mail: [email protected] Professor, Department of Economics, Punjabi University, Patiala, E-mail: [email protected] Professor, Department of Economics, GGDSD College, KheriGurna, BanurE-mail: [email protected] Faculty, Department of Geography, Punjabi University, Patiala, E-mail: [email protected] Professor, Department of Economics, Punjabi University, PatialaE-mail: [email protected] Professor, Department of Economics, Dashmesh Khalsa College, ZirakpurE-mail:[email protected] Professor, Department of Economics, GGDSD College, KheriGurna, BanurE-mail: [email protected]

Received: 5 September 2020; Revised: 6 October 2020; Accepted: 10 Dec. 2020; Online: 29 December 2020

Abstract: Based on the primary data collected from 296 international migrants belonging to 207households of rural Punjab, the study highlightsthe problems of ‘Brain Drain’, ‘Capital Drain’,and ‘Loss of Demographic Dividend’. As many as 96.62 per cent of the people migrated from theage group of 15 to 45 years. A large majority of the migrants (92.57 per cent) belong to theGeneral category. The Jat caste dominates among all the castes, with a proportion of 91.56 percent. Most of the youngsters are migrating just after completing secondary level education.Unemployment, desire to earn more, better living conditions and good administration at destination,and peer pressure are the main reasons for migration of the people form rural Punjab.The processof migration saw a spurt after 2014. The most widely chosen destination countries are Canada,Australia, New Zealand, the U.S.A., and Italy. On an average, the migrants spend Rs.1190572.67for their migration. This amount works outto Rs. 1529084.97 in case of the student categorywhile it turns to be Rs. 828388.16 in the case of non-student category. Two-thirds of the migrants(65.88 per cent) have sent no remittancesto their families. Because of the high cost of migrationand low remittances, two-thirds of the households (66.67 per cent) are under debt.

Keywords: International migration, brain drain, capital drain, demographic dividend,unemployment, remittances, and debt.

JEL Codes: F22, F24, J11, J61, O15

International migration is the dimension that impacts the economicrelations between the developed and the developing countries. It is alsowell recognized that migrant workers make huge contributions to economicand social development in both their host and home countries (Kumar



and Hussain, 2008). Since the 1880s there has been a regular stream ofadventurous young men who have left their villages in Punjab to traveloverseas. The initial destinations for the migrants were countries closer tohome like Singapore and Hong Kong. Later, these became the steppingstones for journeys to more distant lands like Australia, Canada, and theU.S.A. (UNODC, 2009).

After World War II, the Great Britain needed labour for reconstruction;Canada initiated an economic expansion programme and the U.S.A. alsoopened its doors to Indians. Punjabis made use of all these developmentsand migrated in large numbers to these countries. Moreover, internationalmigration started with recruitment to the British Army that opened avenuesfor migration to several other colonies of the British Empire, where Punjabiswere posted to maintain law and order. Soon, voluntary migrations started.Among the destinations of early migrants were British colonies in the FarEast, New Zealand, Australia, the U.K., Canada, the U.S.A., and Africa,especially East Africa. In 1970s, countries of the Middle East appeared assignificant region of destination for migrant workers from India and Punjabtoo contributed to the pool of migrants (Kapuria, 2018). More recently,migration to countries of continental Europe has come into focus, especiallycountries of South Europe that have undergone a transformation fromcountries of origin of migrants to countries of destination (Jacobsen andMyrvold, 2011).

Until recently, Punjab was one of the best performing states in thecountry in terms of per capita income (highest among all major states in199293). The state is rural in nature, with 62.51 per cent of the totalpopulation still residing in rural areas and the remaining 37.49 per cent inurban areas.Total workforce of the state was 9897362, out of which 3522966were dependent on agriculture and allied activities which accounted for35.60 per cent (Census, 2011). With the advent of ‘Green Revolution’, Punjabhas emerged as the most advanced state in agricultural development.Overtime, though agricultural sector experienced a decline in theimportance in terms of its share in Gross State Domestic Product (GSDP)and work force, yet it remains the single most important sector of the stateeconomy (Grover et al. 2017). State’s agricultural sector grew at 5.7 percent per annum during 197172 to 198586, while corresponding figure forIndia was 2.31 per cent (Gulatiet al., 2017). Soon, agricultural sector startedto lose its sheen. Its growth rate fell to 3 per cent during 198687 to 200405and further to 1.61 per cent during 200506 to 201415 (Gulati et al., 2017).The ‘Green Revolution’ has brought the negative impact on soil and waterlevel, increased the cost of cultivation, accumulated debt level of the farmersand agricultural labourers and force them to commit suicides. Agricultural

Socio-Economic and Demographic Analysis of International Migration... 139

sector’s contribution to GSDP, which was about 44 per cent during the 1970s,declined to 39 per cent during the 1990s, further declined to 31 per cent in200405 and to 23 per cent in 201011. In Per Capita Income terms, the stateslipped to fifth position in 200405 and eleventh in 201314 at 200405 prices(Kapuria, 2018).

Punjab, the leading agricultural state of India, is also home to a vastshare of Indians living abroad. The state of Punjab ranks second in termsof sending international migrants after Kerala (The Tribune, 2019).Emigration from Punjab has been consistently growing over last manydecades and it ranks among the top states in India from where large scaleemigration has happened. As per the Annual Report (201819) of Ministryof External Affairs, Punjab had a share of 6 per cent emigrants in the year2018 only. Punjab has emerged as a testimony to the “Culture of Migration”as the proportion of households who have sent international migrants hasimproved from 3 per cent in the year 19921993 (NFHS1) to about 11 percent in 201011(Nanda and Veron, 2015). At state level, the socioeconomicdifferences exacerbates the incidence of emigration as 13 per centhouseholds in rural areas and only 6 per cent households in urban areasshowed likelihood to send migrants to international destinations (Nandaand Veron, 2015).In the present study, an endeavour has been made toanalyse the socioeconomic and demographic aspects of the internationalmigration from rural Punjab.

METHODOLOGY

Since information from the secondary sources is either limited or sketchy,the present study is based on the primary data. For the purpose of thestudy, five villages from Patiala district have been randomly selected. Fromthese villages, all the households from where persons migrated to othercountries during the period from 1951 to 2019 have been surveyed. Thestudy pertained to 296 migrants from 207 households of these villages.The required primary data have been collected from the family membersof migrants through the well prepared questionnairecumschedule by theinterview method. Standard statistical tools such as mean values andproportions have been used while carrying out tabular analysis; and theresults have been shown through piechart, vertical and horizontal bardiagrams and line graph.

RESULTS AND DISCUSSION

International migration is not about the migration of a single person of afamily rather more than one or the entire family. Table 1 reveals that from


the total 207 surveyed households, the number of migrated persons is 296which clearly reflects the fact that there are multiple migrants from somehouseholds. More than twothirds of households (67.63 per cent) have singlemigrants and the remaining 32.37 per cent have multiple migrants. Thepercentage of households with two migrants is 23.67 per cent while 6.76per cent have three migrants. Only 1.93 per cent households have fourmigrants. As many as 12 families have entirely migrated to other countriesfrom these five villages.

Table 1: Number of migrants and households surveyed

Number of migrants 296

Number of households surveyed 207

Households with single migrant 140(67.63)

Households with two migrants 49(23.67)

Households with three migrants 14(6.76)

Households with four migrants 4(1.93)

Number of families which entirely migrated 12

Source: Field Survey, 2020Note: The figures given in parentheses indicate percentages.

Caste and migration has no direct linkage but are indirectly linkedwith each other. The influence of caste on migration stems in the economiccondition of the household. As per Census 2011, the state of Punjab has31.94 per cent Scheduled Caste population, i.e., the highest proportion ofScheduled Caste population reported by any other state or union territoryof India.The data showing the castewise distribution of the migrants arepresented in Table 2 and Figure 1. The table highlights that a large majorityof the migrants, i.e., 92.57 per cent belong to the General category, 4.73 percent belong to the Scheduled Caste category and 2.70 per cent are fromBackward Class category. The Jat caste dominates among all the castes,with the proportion of 91.56 per cent.

Indian society is a closely knit family where there was a culture ofjoint families. This cultural characteristic faded away in urban areas longtime ago but its roots remained intact in the rural areas of India. Punjabprospered widely due to its agriculture that gained its strength from jointfamily system present in rural areas. However, with the change in socioeconomic requirements, the land divisions gained importance and nuclearfamily system stated becoming an innate feature of rural areas as well.Table3 reveals that a majority of the migrants, i.e., 56.42 per cent were living in


Table 2: Caste-wise distribution of migrants

Category Number of migrants Percentage

General

Jat 271 91.56

Brahmin 3 1.01

Subtotal (A) 274 92.57

Scheduled Caste

Gadriya 8 2.70

Ramdasia 5 1.69

Bazigar 1 0.34

Subtotal (B) 14 4.73

Backward Class

Mehre/Jhewar 3 1.01

Ghumiar/Parjapat 2 0.68

Nai 2 0.67

Lohar 1 0.34

Subtotal (C) 8 2.70

Total (A+B+C) 296 100.00

Source: Field Survey, 2020

Figure 1: Caste-wise distribution of migrants

Source: Based on Table 2


the nuclear families while the remaining 43.58 per cent had the joint familysystem. Almost a similar pattern is observed in the case of General category.Out of the total, 54.02 per cent of the migrants had nuclear family systemand 45.98 per cent were living in the joint families. However, in the case ofScheduled Caste and Backward Class categories, a large majority of themigrants, i.e., 92.86 and 75.00 per cent were living in the nuclear familiesrespectively. Only 7.14 and 25.00 per cent of the migrants belonging toScheduled Caste and Backward Class categories had joint family systemrespectively.

Table 3: Distribution of migrants on the basis of type of family

Type of family Cast category

General Scheduled Caste Backward Class Aggregate

Joint 126(45.98) 1(7.14) 2(25.00) 129(43.58)

Nuclear 148(54.02) 13(92.86) 6(75.00) 167(56.42)

Total 274(100.00) 14(100.00) 8(100.00) 296(100.00)


Note: The figures given in parentheses indicate percentages.

Education and migration share a strong relationship with each otheras the better educated person has bright chances of assimilating in thesocioeconomic environment at the place of destination (Browne, 2017).The migration of highly educated migrants is often referred as ‘BrainDrain’ because the migrant moves to international destinations afteracquiring best of the education and gives their services to the place ofdestination. The data showing the educational status of the migrantsbefore going abroad aregiven in Table 5. The table reveals that out of thetotal 296 migrants, 43.58 per cent migrated after acquiring secondary leveleducation. As many as 29.73 per cent migrated after completing theirgraduation while 10.14 per cent were postgraduates at the time of theirmigration. A small proportion of the migrants (2.37 per cent) had thequalification of nursing/GNM/BDS at the time of their migration. Twopeople (0.67 per cent) were M.Phil./Ph.D. holders when they migrated tothe other countries. The analysis clearly highlights that even the peopleof the state with higher educational level consider a better option tosettle in the foreign countries. It is noteworthy that the migrants withlow level of education were either the elderly population who migratedalong with their families or were female spouses of previously settledmale migrants.


Table 4: Educational status of migrants at the time of migration

Educational Status Number of migrants Percentage

Illiterate 3 1.01

Primary 5 1.69

Middle 5 1.69

Matric 27 9.12

Secondary 129 43.58

Graduation 88 29.73

Postgraduation 30 10.14

Nursing/GNM/BDS 7 2.37

M.Phil./Ph.D. 2 0.67

Total 296 100.00


Figure 2 clearly shows that most of the persons migrated to the othercountries after completing their secondary level education followed by thegraduates.

Figure 2: Educational status of migrants at the time of migration


Age and migration has a strong linkage as the decision to migrate isundertaken mainly by population belonging to young and mature agegroup. It is primarily because migration is a risky affair and has severalchallenges and opportunities. Those who belong to active or working age


group as the capacity to endure the challenges and encash the opportunitiesprovided by the process of migration.Table 6 gives the data regarding ageof the migrants at the time of their migration to the foreign countries. Thetable shows that a very large majority of the migrants (96.62 per cent) werein the age group of 15 to 45 years at the time of their migration. Out ofthese migrants, slightly less than threefourths of the total migrants (74.66per cent) were in the age of 20 to 30 years when they migrated. This resultgoes well with the notion that migration in general and internationalmigration in particular is a youth centric phenomenon. This finding of thestudy matches with another research study conducted by SamitaBehl(2017)which reveals that at the time of migration approximately 78 per cent ofthe migrants belonged to the age up to 30 years or below. This shows thatthe migration of youth at their early age somehow signifies that preferencesto work for longer durations in the international job markets. Thus,international migration is supported by easy job opportunity in theinternational unskilled job market and this employment security somehowhelps the young migrants to get easy and certain matrimonial alliances.This age group is considered as the most energetic and talented havingnew ideas of income generation and growth for the country. The studyhighlights the fact that migration of the young generation translates into agreat loss of human resources, with the direct benefit accruing to therecipient countries who have not forked out the cost of educating them.The youth of any country are the most expensive resource because of theirtraining in terms of material cost and time, and most importantly, becauseof lost opportunity (Dodani and Laporte, 2005).

Table 5: Age-wise distribution of migrants at the time of migration

Age (In years) Number of migrants Percentage

Less than 15 6 2.03

1520 31 10.47

2025 133 44.93

2530 88 29.73

3035 22 7.43

3540 6 2.03

4045 6 2.03

45 and above 4 1.35

Total 296 100.00


It is clear from Figure 3 that maximum number of the persons migratedto the foreign countries in the age group of 20 to 25 years followed by thepersons in age group of 25 to 30 years.


Besides being youth centric, the migration is also a male centricphenomenon. The predominance of males in the migration process isfuelled by high demand of males in the work scenario throughout the world.The male dominated society gives more precedence to male workers insteadof their female counterparts and both the skilled as well as unskilled jobsopen a major share of their opportunities for the male population. Thedata given in Table 4 exhibits the sexwise distribution of the migrantsfrom rural Punjab. The table shows that out 296 migrants, 173 are malesand the remaining 123 are females. The males account for 58.45 per centand the females are 41.55 per cent. A similar pattern is observed in the caseof General category. But in the case of Scheduled Caste category, all themigrants are male while in the case of Backward Class category, 87.50 percent are male and the remaining 12.50 per cent are female.

Table 6: Sex-wise distribution of migrants

Sex Caste category


Male 152(55.47) 14(100.00) 7(87.50) 173(58.45)

Female 122(44.53) 0(0.00) 1(12.50) 123(41.55)

Total 274(100.00) 14(100.00) 8(100.00) 296(100.00)


Figure 3: Age-wise distribution of migrants at the time of migration



High proportion of males in the emigration strengthens the fact thatin the contemporary world, where almost equal level of education isprovided to both the genders, the male counterpart is considered to be theprimary bread winner and is given the responsibility to uplift the standardof living of his family. It is worthwhile to note that females also have asignificant proportion (41.55 per cent) in the total surveyed migrantpopulation. The female international migration from Punjab is mainlymarriage centric where the girls move to international destinations as aspouse. However, in the present Punjab, the scenario has shifted completelyas those girls who are educated and have scored high band scores areconsidered to be a ladder for international migration. Families which canafford money but do not have educated and able boys tend to allure familiesof these girls and after marrying them off, they sponsor their daughterinlaw’s international travel, expenses and education. These girls later callupon their husbands living at the native place on spouse visa and with thepassage of time, they work together to secure citizenship of the country ofdestination and then become a ticket to immigration for their respectivefamilies.

Table 7 shows the information regarding the destination countries ofthe migrants from rural Punjab. The table depicts that although the list ofthe destination countries is long yet the most widely chosen destinationcountries are Canada, Australia, New Zealand, the United States, and Italy where 129, 66, 25, 17 and 16 persons from rural Punjab migrated. Thisfinding of the study matches with another research study (Kapuria andBirwal, 2017) which shows that migration from Punjab has mostly beentowards the developed countries of the West.

Table 7: Country-wise distribution of migrants

Destination General category Scheduled Cates Backward ClassCountry category category

Jat Brah Gadr Ramd Bazi Mehre/ Ghumiar/ Nai Lohar Totalmin iya asia gar Jhewar Parjapat

Canada 125 1 2 1 129

Australia 64 1 1 66

New Zealand 24 1 25

U.S.A. 17 17

Italy 15 1 16

Cyprus 1 4 5

Dubai 1 1 1 2 5

U.K. 5 5

contd. table 7


Germany 3 1 4

Jordan 2 2 4

Bahrain 3 3

South Africa 2 1 3

Malaysia 1 1 2

Saudi Arabia 1 1 2

Singapore 2 2

Spain 2 2

Sweden 2 2

Kuwait 1 1

Philippines 1 1

Portugal 1 1

Thailand 1 1

Total 271 3 8 5 1 3 2 2 1 296


With the Canadian government being the most liberal among thedeveloped countries in the grant of Permanent Residency, and opening upas many as 200 colleges to international students, Punjabi youth are makingfull use of the opportunity. As many as 1.25 lakh students from the statechose Canada this year for education while only 25,000 picked Australia,New Zealand, the U.S.A., and the U.K., where laws and policies havebecome very stringent (The Tribune, 2018).The table further shows that inthe case of General category, the most preferable destination countries areCanada, Australia, New Zealand, the United States, and Italy. However, inthe case of Scheduled Caste category, Cyprus and Jordan are the mostpreferable destination countries where four persons each from this categorymigrated to Cyprus, and Jordan. Two persons from Backward Classcategory migrated to Dubai, and one person each from this categorymigrated to Canada, Malaysia, Australia, New Zealand, South Africa, andSaudi Arabia.

It is clearly evident from Figure 4 that the most preferable destinationcountryfor the migrants of rural Punjab is Canada, followed by Australia,New Zealand, the U.S.A., and Italy.

The details of the farmsizecategories of the migrants’ families are givenin Table 8. The table reveals that out of the total families of migrants fromrural Punjab, the maximum proportion of the families, i.e., 24.64 per cent

Destination General category Scheduled Cates Backward ClassCountry category category

Jat Brah Gadr Ramd Bazi Mehre/ Ghumiar/ Nai Lohar Totalmin iya asia gar Jhewar Parjapat


have land holdings in the range of 2.51 to 5.00 acres.Slightly more than 20per cent of the families have land holdings of more than 10 acres.This isfollowed by 17.39, 14.01 and 13.04 per cent of the migrants’ families whohave land holdings in the range of 7.51 to 10.00 acres, less than 2.51 acres,and 5.01 to 7.50 acres respectively. Out of total, 10.63 per centof the migrants’families have no farm land.

Table 8: Distribution of migrants on the basis of farm-size categories

Farm size Caste category(In acres)


Landless 4(2.15) 12(92.31) 6(75.00) 22(10.63)

less than 2.50 26(13.98) 1(7.69) 2(25.00) 29(14.01)

2.515.00 51(27.42) 0(0.00) 0(0.00) 51(24.64)

5.017.50 27(14.52) 0(0.00) 0(0.00) 27(13.04)

7.5110.00 36(19.35) 0(0.00) 0(0.00) 36(17.39)

10.01 and above 42(22.58) 0(0.00) 0(0.00) 42(20.29)

Total 186(100.00) 13(100.00) 8(100.00) 207(100.00)



The castewise analysis of the table shows that majority of the familiesbelonging to the Scheduled Caste and Backward Class categories are

Figure 4: Country-wise distribution of migrants



landless.The data in the table have an important implication thatinternational migration is too expensive task. Mostly the Jatcasteis capableto afford it as they have land holdings which can be mortgaged for raisingthe funds.

A look at Table 9 and Figure 5 reveals that out of the total 296 migrantsfrom rural Punjab, 22.30 per cent of the people migrated in 2019, 15.88 percent in 2018, 12.16 per cent in 2017, 7.43 per cent in 2016, 7.09 per cent 2015,5.41 per cent in 2010 and less than 5 per cent in the other years. Thesefigures clearly highlight the fact that the process of migration saw a spurtafter 2014.Travel agents engaged in facilitating the Canadian visa forstudents saidthat the trend has seen a spurt since 2016, when around 75,000students from Punjab had gone to the country(The Tribune, 2018).

Table 9: Distribution of migrants according to year of migration

Year Number of migrants Percentage

2019 66 22.30

2018 47 15.88

2017 36 12.16

2016 22 7.43

2015 21 7.09

2014 14 4.73

2013 5 1.69

2012 3 1.01

2011 12 4.05

2010 16 5.41

2009 11 3.72

2008 11 3.72

2007 5 1.69

2006 6 2.03

2005 5 1.69

2004 3 1.01

2003 0 0.00

2002 0 0.00

2001 1 0.34

2000 and before 12 4.05

Total 296 100.00


The channel of migration refers to the way adopted by migrants duringthe migration process. Throughout the world, migration is attemptedthrough both legal and illegal channels. Where legal channel is timeconsuming and involves less risk, the illegal migration is an attempt to


enter other country by bypassing the administrative and legal set up of thecountry, thereby making it a highly risky affair. One of the positive aspectsof the study is that in 92.91 per cent cases, the channel of migration is legalwhile only in 7.09 per cent cases, it is illegal. The table further reveals thatout of the illegal migrants, 4.05 per cent of the people migrated to the othercountries on work visa while 3.04 per cent on visitor visa (Table 10). AsUNODC (2009) report shows a large number of Punjabi youth under theimpression of finding greener pastures in foreign lands, is trapped intothe agentsmuggler nexus. Illegal status in destination countries only makesthem vulnerable to exploitation.

Table 10: Distribution of migrants on the basis of channels of migration

Channels of migration Number of migrants Percentage

Legal 275 92.91

Illegal 21 7.09

Total 296 100.00

Types of visa for illegal migration

1) Work 12 4.05

2) Visitor 9 3.04

Total 21 7.09



Figure 5: Distribution of migrants according to year of migration



Table 11exhibits that more than half of the persons (51.69 per cent)migrated on study visa while slightly more than onefourth, i.e., 25.34 percenton work visa. As many as 17.23 per cent of the persons migrated onspouse visa whereas a small proportion of the persons (3.72 per cent)migrated on visitor visa. Only 2.03 per cent of the persons have familybased or blood relation type of visa to migrate abroad. The table furthershows that among the Scheduled Caste and Backward Class categories,maximum proportion of the persons migrating to the foreign countries forwork purpose as they have no job in the country. However, in the case ofGeneral category, 54.01 per cent of the persons migrated for study purpose.Only 21.53 and 18.25 per cent persons from General category have workvisa and spouse visa respectively to migrate abroad.

Table 11: Type of visa for migration

Particulars Caste category

General caste Scheduled Caste Backward Class Aggregate

1) Student 148(54.01) 3(21.43) 2(25.00) 153(51.69)

2) Work 59(21.53) 11(78.57) 5(62.50) 75(25.34)

3) Spouse 50(18.25) 0(0.00) 1(12.50) 51(17.23)

4) Visitor 11(4.01) 0(0.00) 0(0.00) 11(3.72)

5) Family based/blood relation 6(2.19) 0(0.00) 0(0.00) 6(2.03)

Total 274(100.00) 14(100.00) 8(100.00) 296(100.00)


Note: The figures given in parentheses indicate percentages

If anyone wants to migrate to the foreign countries for acquiring highereducation, working, professional recognition or for the permanentresidency, then he/she is required to present evidence of English languageproficiency. There are some basic tests of English language such as IELTS/TOEFL/GRE/SAT/PTE which are accepted by the foreign countries whereEnglish is the main mode of communication. The details regarding thebasic test of English language passed by the migrants to go abroad arepresented in Table 12. The table shows that out of total 296 migrants, morethan twothirds of the migrants (68.24 per cent) have passed the test whileremaining 31.76 per cent have not passed any such type of test.

The reasons or the factors that affect migration are also called driversof migration. These are usually multitude in number as the decision tomigrate is never influenced by a single factor; rather it is a combination ofseveral contributing factors that play a decisive role in inducing or inhibiting


migration. The decision to migrate to international destinations is alwaysa tricky one as this type of migration contains high risks as well as highgains. The factors that affect migration are usually categorised as pushand pull factors. These factors vary for both the place of origin and place ofdestination. The push factors includeincreasing unemployment,agricultural distress, population growth, and social issues like drug menaceetc. The international destinations also generate several pull factors suchas economic development, better educational and administration facilities,family unification, better living condition, greener pastures of employment,high remuneration from work etc.When the migrants or their familymembers are interviewed regarding the reasons for the migration thenthey cited the following push and pull factors (Table 13):

Table 13: Driversof migration*

S. No. Particulars Number of migrants Percentage

A. Push factors

1. Unemployment 154 52.03

2. Peer pressure 73 24.66

3. Drug menace 27 9.12

B. Pull factors

1. Desire to earn more 137 46.28

2. Better living conditions and good 126 42.57administration at destination

3. Family migration 13 4.39

4. Acquiring higher education 6 2.03

Source: Field Survey, 2020*Multiple reasons

• In the case of 52.03 per cent migrants, their family membersconfessed that unemployment is the main reason for migrationof their children.

• As many as 46.28 per cent of the family members of migrantsadmitted that their children’s desire to earn more motived themfor migration.

Table 12: Basic tests of English language passed by migrants to go abroad

Particulars Number of migrants Percentage

Migrants having English language 202 68.24test, i.e., IELTS/TOEFL/GRE/ SAT/ PTE

Migrants not having any english language test 94 31.76

Total 296 100.00



• Better living conditions and good administration at destinationattracted 42.57 per cent of the migrants.

• Slightly less than onethird of the people (24.66 per cent) migratedto the foreign countries because they feel peer pressure, as theirfamily members stated.

• As many as 9.12 per cent of the family members of migrants citedthat their children migrated abroad because of the drug menacein Punjab.

• A small proportion of the people (4.39 per cent) migrated becausetheir families are settled in the foreign countries.

• Only 2.03 per cent of the people migrated to the foreign countriesfor acquiring higher education, as admitted by their familymembers.

The above analysis clearly reflects that unemployment, desire to earnmore, better living conditions and administration at destination, and peerpressure are the main reasons for migration form rural Punjab.

The study shows that 153 persons migrated to theforeign countriesforstudy purpose. Therefore, the information regarding the courses opted bythe students in the foreign countries is given in Table 14. The table showsthat more than onefourth of the students (26.14 per cent) opted degree/diploma courses in the stream of commerce and management. This isfollowed by 10.46, 9.80, 7.84, 4.58 and 2.62 per cent of the students whoopted degree/diploma courses in the stream of nonmedical, medical, hotelmanagement, IT, and arts respectively. Interestingly, the parents of 38.56per cent of the students don’t know about the course opted by their wards.

Table 14: Courses opted by the student migrants in foreigncountries

Particulars Number of migrants Percentage

Degree/diploma in commerce and management 40 26.14

Degree/diploma in medical 15 9.80

Degree/diploma in nonmedical 16 10.46

Degree/diploma in hotel management 12 7.84

Degree/diploma in IT 7 4.58

Degree/diploma in arts 4 2.62

Don’t know 59 38.56

Total 153 100.00


The data pertaining to occupation of the migrants prior to theirmigration and after the migration from rural Punjab is given in Table 15.


The table highlights that more than half of the migrants (52.03 per cent)were the students prior to their migration. Slightly less than onefifth ofthe migrants (18.92 per cent) were farmers. As many as 11.48 per cent ofthe migrants reported that they were unemployed and the same proportionof the migrants, i.e., 11.48 per cent were doing their household work.

Table 15: Occupation of the migrants prior to and after migration

Particulars Prior to migration After migration

1) Agriculture 56 5(18.92) (1.69)

2) Household work 34 19(11.48) (6.42)

3) Private job 10 96(3.38) (32.43)

4) Study 154 111(52.03) (37.50)

5) Unemployed 34 0(11.48) (0.00)

6) Labour 4 5(1.35) (1.69)

7) Government job 1 0(0.34) (0.00)

8) Own business/selfemployment 2 12(0.68) (4.05)

9) Driving 0 46(0.00) (15.54)

10) Others 1 2(0.34) (0.68)

Total 296 296(100.00) (100.00)



As far as the occupations of the migrants after the migration areconcerned, the table reveals that the maximum proportion of the migrants,i.e., 37.50 per cent is pursuing their study in the foreign countries. Lessthan onethird of the migrants, i.e., 32.43 per cent are in private jobs while15.54 per cent are in the occupation of driving in the foreign countries. It isevident from the analysis that Punjabis have migrated to various countriesof the world to meet the demand of less skilled workers. Aresearch studyconducted by Kapuria and Birwal (2017) also produces the same results.

The data regarding the expenditure incurred by the migrants formigration to the foreign countriesare presented in Table 16. The table


exhibits that the maximum proportion of the migrants (27.02 per cent)spends in the range from Rs. 15 to 20 lakh formigration. Slightly less than24 per cent of the migrants spend less than Rs. 5 lakh for migration. This isfollowed by 15.53, 14.86, 11.83, and 4.40 per cent of the migrants who spendin the ranges from Rs. 10 to 15; 5 to 10; and 20 to 25 lakh respectively. Only2.36 per cent of the migrants spend Rs. 30 lakh and above for this purpose.

Table 16: Expenditure-wise distribution of migrants

Expenditure Types of visa(In Rs., lakhs)

Student Work Spouse Visitor Family based/ Aggregateblood relation

Less than 5 6 33 19 7 6 71(3.92) (44.00) (37.25) (63.64) (100.00) (23.99)

510 14 21 8 1 0 44(9.15) (28.00) (15.69) (9.09) (0.00) (14.86)

1015 29 13 3 1 0 46(18.95) (17.33) (5.88) (9.09) (0.00) (15.53)

1520 68 5 6 1 0 80(44.44) (6.67) (11.77) (9.09) (0.00) (27.02)

2025 28 2 5 0 0 35(18.31) (2.67) (9.80) (0.00) (0.00) (11.83)

2530 6 0 6 1 0 13(3.92) (0.00) (11.77) (9.09) (0.00) (4.40)

30 and above 2 1 4 0 0 7(1.31) (1.33) (7.84) (0.00) (0.00) (2.36)

Total 153 75 51 11 6 296(100.00) (100.00) (100.00) (100.00) (100.00) (100.00)



In the case of student visa, the maximum proportion of the migrants(44.44 per cent) spends in the range from Rs. 15 to 20 lakh for the migration.A report posted in The Tribune (2018) also explains that the emigrationprocess comes at a cost Rs. 15 to 22 lakh for the first year of study, dependingon the institute, course and country. Multiplied by the number of studentsflying out, that amounts to approximately Rs. 27,000 crore going out ofPunjab each year on account of student education. In the case of nonstudenttype visa, the maximum proportion of the migrants spends less than Rs. 5lakh for migration.

The data showing the average amount of funds arranged by themigrants from different sources given in Table 17 highlight that the migrantsspend, on an average, Rs.1190572.67 for their migration. This amount works


out to Rs. 1529084.97 in case of the student category. On the other hand,the migrants belonging to nonstudent category spend Rs. 828388.16 forthis purpose.

Table 17: Average amount of funds arranged by migrants from different sources

S. No. Sources Student Others AggregateCategory

1. Family savings 505555.56 290185.36 401508.47(33.06) (35.03) (33.72)

2. Loan from banks 452941.18 141958.05 302702.71(29.62) (17.14) (25.42)

3. Selling of land/plot/ornaments/vehicle/ 215359.48 193986.01 205033.79animals agricultural machinery (14.08) (23.42) (17.22)

4. Moneylenders 141503.27 118671.33 130472.97(9.25) (14.33) (10.96)

5. Loan/help from relatives 116339.87 58342.66 88320.95(7.61) (7.04) (7.42)

6. Funding by would be inlaws 68627.45 13286.71 41891.89(4.49) (1.60) (3.52)

7. Cooperative societies/banks 28104.58 10209.79 19459.46(1.84) (1.23) (1.64)

8. Employers/institutions 653.58 1748.25 1182.43(0.05) (0.21) (0.10)

Total 1529084.97 828388.16 1190572.67(100.00) (100.00) (100.00)


As far as the aggregate sources of funding are concerned,more thanonethird proportion (33.72 per cent) of the total expenditure, i.e., Rs.401508.47 comes from the savings of the families of the migrants. It hasbeen observed thatthe migrants arrange 25.42 per cent of the totalexpenditure from banks which amounts to Rs. 302702.71. In some of thecases, the migrants have to sell their land/plot/ornaments/vehicle/animals/agricultural machinery for arranging the funds. In the present study, thissource of funding has contributed Rs. 205033.79 which forms 17.22 percent of total cost of migration. The study further reveals that the migrantsalso take loans from moneylenders which provide 10.96 per cent of thetotal fundswhich amounts to Rs. 130472.97.It is worth mentioning herethat relatives of the migrants and even would be inlaws have also to bearthe cost of their migration. The data shows that the contribution of relativesin total cost of migration is 7.42 per cent (Rs. 88320.95). As far as the cost ofmigration borne by would be inlaws isconcerned, it is Rs. 41891.89 whichis 3.52 per cent of the total migration cost. The migrants also received Rs.


88320.95 from their relatives in the form of loan/help. The relativescontribute 7.42 per cent of the total funds. The migrants received only Rs.19459.46 and 1182.43 from cooperative societies/banks and employers/institutions respectively which form 1.64 and 0.10 per cent of the total fundsused for migration respectively

As far as thesources of funding among the student and nonstudentcategories are concerned, a slight different pattern has been observed.Thedata shows that migrants of student category are taking more loans frombanks (29.62 per cent) than the other migrants (17.14 per cent). On thecontrary, the other migrants are raising more funds through selling of theirland/plot/ornaments/vehicle/animals/agricultural machinery (23.42 percent) and through moneylenders (14.33 per cent) whilethe student migrantsare raisingless funds through these sources with the proportions of 14.08and 9.25 per cent respectively.It is pertinent to note here thattheproportionof funding by would be inlawsis more in student categorymigrants than other migrants. While the proportion of funding throughthe other remaining sources is almost same amongthe student and the othermigrants.

Migration becomes a favourable experience for the families whosemigrant members start sending remittances after joining the workforce atthe place of destination. The term ‘remittance’ refers to the monetary

Table 18: Remittances inflow to the migrants’families

a) Has the family received any remittances? Number of migrants Percentages

Yes 101 34.12

No 195 65.88

Total 296 100.00

b) Amountwise distribution of migrants

Amount (In Rs.)

Less than 50000 22 7.43

50000100000 24 8.11

100000150000 18 6.08

150000200000 10 3.38

200000250000 5 1.69

250000300000 7 2.36

300000350000 7 2.36

350000400000 3 1.01

400000 and above 5 1.69

Total 101 34.12



Table 19: Debt position ofmigrants’ families

Amount Caste category(In Rs., lakhs)

General Scheduled Backward AggregateCaste Class

Zero 67 1 1 69(36.02) (7.14) (14.29) (33.33)

less than 5 32 7 6 46(17.20) (50.00) (85.71) (22.22)

510 44 5 0 48(23.66) (35.72) (0.00) (23.19)

1015 23 0 0 23(12.37) (0.00) (0.00) (11.11)

1520 12 0 0 12(6.45) (0.00) (0.00) (5.80)

20 and above 8 1 0 9(4.30) (0.14) (0.00) (4.35)

Total 186 14 7 207(100.00) (100.00) (100.00) (100.00)



transfers made by migrants to their country of origin. These can also includeinvestments, deposits and donations made by migrants in the country oforigin (OSCE, IOM, ILO 2006). However, studying abroad is an expensiveaffair with a variety of costs involved like tuitionfees, living expenses andtravel expenses, etc. Many students opt for parttime jobs abroadwhilestudying to meet of their extra expenses. To attract international students,manycountries allow students to engage in some kind of parttime jobs tohelp them funding theirexpenses. While some countries ask students toapply for separate work permits to workduring studies, others let thestudents to work on their student visa itself, although, only forrestrictedhours usually spanning 1020 hours a week (Kaur, 2019). Table 18 showsthat only slightly more than onethird of the migrants (34.12 per cent) havesent remittances to their families while twothirds of the migrants (65.88per cent) have sent no remittancesto their families.

The table further shows that out of 34.12 per cent migrants who havesent remittances, 74 migrants (25 per cent) have sent the remittances ofless than Rs. 200000. Only 1.69 per cent of the migrants have sent the totalremittances of Rs. 400000 and above to their families.

Because of the high cost of migration and low remittances, twothirdsof the families (66.67 per cent) reported that they are under debt (Table 19).


From indebted households, 45.55 per cent households are under debt ofup to Rs. 10 lakh. The households which are under debt of Rs. 1015 lakhand Rs. 1520 lakh are 11.11 and 5.80 per cent respectively. Even 4.35 percent are such households which are under debt of Rs. 20 lakh and above.

The castewise analysis of this issue highlights that the maximumproportion of indebted households belong to Scheduled Caste category(92.86 per cent) followed by Backward Class category (85.71 per cent) andGeneral category (63.98 per cent).

CONCLUSIONS AND POLICY IMPLICATIONS

The results of the study and field survey conducted in the rural areas ofPunjab have the following important implications:

• The study reveals that 96.62 per cent of the people migratedin theage group of 15 to 45 years. This age group is considered as themost energetic and talented having new ideas of incomegeneration and growth for the country. Most of the youngstersare migrating for study purpose just after completing secondarylevel education. Even the people of the state with highereducational levels consider a better option to settle in the foreigncountries. More than half of the family members of migrants (52.03per cent) confessed that unemployment is the main reason formigration of their children. Therefore, the Central and Stategovernments must take strong initiatives for creating sufficientemployment opportunities by implementing propeople approachso that Punjabi people have not to send their family membersabroad.

• As many as 9.12 per cent of the family members of migrants citedthat their children migrated abroad because of the drug menacein Punjab. Therefore, the Central and State governments musttake strong steps to control the drug menace in the state.

• On an average, the migrants have spent Rs.1190572.67 for theirmigration. This amount is Rs. 1529084.97 in case of the studentcategory while it turns to be Rs. 828388.16 in the case of nonstudent category. Further, more than onethird proportion of thetotal expenditure comes from the savings of the families of themigrants. The study also shows that twothirds of the migrants(65.88 per cent) have sent no remittancesto their families. Only 74migrants (25 per cent) have sent the total remittances of less thanRs. 200000. Because of the high cost of migration and lowremittances, twothirds of the families (66.67 per cent) are under


debt. It is again a serious problem that compels the farmers andagricultural labourers to commit suicides . Nowadays,international migration has become an expensive task.International migration from rural Punjab doesn’t only results in‘Capital Drain’, but it also involves ‘Brain Drain’ and ‘Loss ofDemographic Dividend’. Therefore, the Central and Stategovernments should create sufficient employment opportunitiesalongwith other measures to overcome these problems.

• The people are so eager to go abroad that they don’t hesitate togo illegally as the study shows 7.09 per cent such cases. The illegalstatus in the destination countries only makes them vulnerableto be exploited by the foreigners. The problem of illegal migrationshould be solved by creating awareness among the youth againstillegal channels of migration, and to improve educationaloutcomes of the youth to increase their employability in thecountry as well as in other countries. Besides, the governmentalso needs to act strictly, through legislation and implementation,to break the agentsmuggler nexus.

• Since 42.57 per cent of the family members have cited that theirchildren migrated to the foreign countries because of the betterliving conditions and good administration at destination.TheCentral and State governments should considerably improve theliving conditions and administration in the state and country.

Note

The authors are thankful to Bebe GurnamKaur Memorial Educational Centre, Isru,Ludhianafor sponsoring this study.

References

Behl, S. (2017). International migration and extent of remittances: A case study of ruralPunjab. International Journal of Economic and Business Review, 5(3), 136146.

Browne E. (2017). Evidence on education as a driver for migration, Retrieved from https://a s s e t s . p u b l i s h i n g . s e r v i c e . g o v . u k / m e d i a / 5 9 8 0 8 6 a 0 e d 9 1 5 d 0 2 2 b 0 0 0 0 3 c /K4D_HDR__Migration_and_Education.pdf

Dodani. S. and Laporte, R.E. (2005). Brain drain from developing countries: how can braindrain be converted into wisdom gain? Journal of the Royal Society of Medicine, 98(11),487491.

Government of India. (2011). Census of India.Director of Census Operations, Punjab.Retrieved from http://censusindia.gov.in/.

Grover, D.K., Singh, J.K., Kaur, A. and Kumar, S. (2017). State Agricultural ProfilePunjab.AgroEconomic Research Centre, Department of Economics and Sociology, PunjabAgricultural University, Ludhiana.


Gulati, A., Roy, R. and Hussain, S. (2017). Getting Punjab Agriculture Back on High GrowthPath: Sources, Drivers and Policy Lessons. ICRIER, New Delhi.

Jacobsen, K.A. and Myrvold, K. (eds.) (2011). Sikhs in Europe: Migration, Identities andRepresentations. Ashgate, U.K.,

Kapuria, S. (2018). International migration from Punjab and challenges for governance.Panjab University: Research Journal (Arts), XLV (1), 119

Kapuria, S. and Birwal, D. (2017). International migration from Punjab: Trends andchallenges. Researchpaedia, 4(1), 2736.

Kaur, G. (2019). Overseas migration of students from Punjab.International Journal of Researchand Analytical Reviews, 6(1), 10531059.

Kumar, S. and Hussain, Z. (2008). Managing International Labour Migration from India: Policiesand Perspectives. ILO, Sub Regional Office for South Asia, New Delhi.

Ministry of External Affairs (2019). Annual report 20182019. Retrieved from http://www.mea.gov.in/Uploads/PublicationDocs/31719_MEA_AR18_19.pdf

Nanda, A.K & Veron J. (2015). Dynamics of International Outmigration from PUNJAB, Reporton PIMS. INED, Paris and CRRID, Chandigarh, 1,38.

OSCE, IOM, ILO (2006). Handbook on Establishing Effective Labour Migration Policies inCountries of Origin and Destination. Organization for Security and Cooperation inEurope (OSCE), International Organization for Migration (IOM), International LabourOffice (ILO), Geneva

The Tribune (2018). Cost of foreign dreams for patients in Punjab. July 29. Retrieved fromhttps://www.tribuneindia.com/news/punjab/costofforeigndreamsforparentsinpunjabrs27000cr/628584.html

The Tribune (2019). Unsettling migration’s underbelly, March 03.Retrieved from https://www.tribuneindia.com/news/archive/features/unsettlingmigrationsunderbelly737217

UNODC (2009). Smuggling of Migrants from India to Europe and in particular to U.K.: A Studyon Punjab and Haryana. United Nations Office on Drugs and Crime, New Delhi.


Gurinder Kaur, Gian Singh, Dharampal, Rashmi, Rupinder Kaur, Sukhvir Kaur andJyoti. SocioEconomic and Demographic Analysis of International Migration fromRural Punjab: A Case Study of Patiala District. Journal of Asian Economics, Accountingand Finance, Vol. 1, No. 2, 2020, pp. 137161

EFFECT OF LENDING RATE ON THE PERFORMANCEOF NIGERIAN DEPOSIT MONEY BANK

Owolabi, A

Department of Banking and Finance, The Federal Polytechnic, Ado – Ekiti, Ekiti State, NigeriaE-mail: [email protected]

Received: 24 July 2020; Revised: 21 August 2020; Accepted: 26 August 2020; Online: 29 December 2020

Abstract: The paper examined the effect of lending rate on the performance of Nigerian DepositMoney Bank. The paper made use of secondary data which were sourced from Central Bank ofNigerian Statistical Bulletin of various years. The data was analysis with the use of multipleregressions technique of Ordinary Least Square (OLS). The study revealed that positive relationshipexists between lending rate and the profit after tax of Nigerian banks. The study recommendedthat Nigerian banks should review down ward the lending rate to enhance more patronage of theloan facilities by the public.

Key Words: Profit After Tax, Bank Performance, Lending Rate, Economic Growth

I. INTRODUCTION

Commercial Banks are custodians of depositor’s funds and operate byreceiving cash deposits from the general public and loaning them out tothe needy at statutorily allowed interest rates (Ngure, 2014). In Nigeria thefinancial sector is dominated by commercial banks, therefore any failurein the sector has a grave consequence on the economic growth anddevelopment of the country. This is due to the fact that any bankruptcythat could happen in the sector has a contagion effect that can lead to bankruns, crises and bring overall financial crisis and economic tribulations(IMF, 2001). Banks play a major role in the economy through their economicfunction of financial intermediation that performs both a brokerage and arisk transformation function (Hara, 1983).

Commercial Banks as financial intermediaries perform financialintermediation function of mobilization and allocation of funds from theeconomic surplus (lenders) to the economic deficit unit (borrowers). Thisfunction is directly linked with banks profitability which encourageseconomic growth. According to Wainaina (2013), profitability of banks hasrelationships with growth and development of the economy. Deposit moneybanks are the most important savings and mobilization of financialresources and allocating them to productive investment and in return



promote their performance (Victor 2013). Interest rate however plays avital role in how a bank makes money (Haye, 2013). Hualan (1992) foundthat interest rate is one of the most important factors that affect the bankfinancial performance. Interest rates are the reward paid by a borrower(debtor) to a lender (creditor) for the use of money for a period and theyare expressed in a percentage, per annum (pa) to make them comparable.Interest rates are also quite often referred to as the price of money. Corb(2012) described interest rate as an economic tool used by the Central Bankto control inflation and boost economic development. Therefore poordecisions on an interest rate regime could spell doom for the financialsystem and the economy as a whole.

The Central Bank of Nigeria uses the interest rate is as a monetarypolicy tool to adjust the lending rates of banks and other financialinstitutions in Nigeria. Giovanni (2006) argued that high interest rate setby the Central Bank means that the other financial institution will have tocharge high because they are all profit oriented. In Nigeria, since theinception of interest rates deregulation in 1986, the government has pursueda marketdetermined interest rate regime, which does not permit a directstate intervention in the general direction of the economy (Adebiyi andBabatope, 2004). Rasheed (2010) states that the Nigerian economy sawdifferent interest rates for different sectors in 1970’s through the mid 1980(regulated Regime). Preferential interest rates were therefore applied toencourage priority sectors such as agriculture and manufacturing. However,deposit money banks decisions to lend out loans are influenced by a lot offactors such as the prevailing interest rate, the volume of deposits, the levelof their domestic and foreign investment, banks liquidity ratio, prestigeand public recognition to mention just but a few. Lending practices in theworld could be traced to the period of industrial revolution which increasethe pace of commercial and production activities thereby bringing aboutthe need for large capital outlays for projects.

Many captains of industry at this period were unable to meet up withthe sudden upturn in the financial requirements and therefore turn to thebanks for assistance (Ezirim, 2005). However, the emergence of banks inNigeria in 1872 with the establishment of the African Banks Corporation(ABC) and later appearance of other banks in the scene during the colonialera witnessed the beginning of banks’ lending practice in Nigeria. Though,the lending practices of the then colonial banks were biased anddiscriminatory and could not be said to be a good lending practice as onlythe expatriates were given loans and advances. This among other reasonsled to the establishment of indigenous banks in Nigeria. Prior to the adventof Structural Adjustment Programme (SAP) in the country in 1986, the

Effect of Lending Rate on the Performance of Nigerian Deposit Money Bank 165

lending practices of banks were strictly regulated under the closesurveillance of the bank’s supervisory bodies. The SAP period broughtabout some relaxation of the stringent rules guiding banking practices.The Bank and Other Financial Act Amendment (BOFIA) 1998, requiresbanks to report large borrowing to the CBN. The CBN also require thattheir total value of a loan credit facility or any other liability in respect of aborrower, at any time, should not exceed 20% of the shareholders’ fundsunimpaired by losses in the case of commercial banks (Felicia, 2011).

Statement of the Problem

Commercial banks in Nigeria are predominant in the banking industry.Their deposit and credits form a major portion of the total credit to alleconomy sector. However, they still face major challenges with regards togovernment regulations, institutional difficulties and other similarchallenges. This study therefore intends to identify the effect of the interestrate on the performance of Nigeria deposit money bank. This would be ofgreat assistance to the regulators in forming a favorable interest rate thatwould meet the macro economic objectives in Nigeria.

Many researchers have work on this study “effect of interest rate onperformance of Nigerian deposit money bank “ using many interest ratevariable to proxy the interest rate but no researchers have use depositinterest rate to measure the relationship between the variables. So this studywill include the deposit interest rate as one of the variable in the model.

Research Questions

The seminar paper is guided with the following research question;

i) Does lending interest rate have significant relationship with theperformance of Nigerian deposit money bank?

ii) Does deposit interest rate significantly influence the performanceof Nigerian deposit money bank?

iii) Does monetary policy rate significantly influenced theperformance of Nigerian deposit money bank?

Objectives of the Study

This study examined the effect of the lending rate on the performance ofNigeria deposit money bank.

While, the specific objective are;

i) Examine lending interest rate relationship with the performanceof Nigerian deposit money bank.


ii) Examine deposit interest rate relationship with the performanceof Nigerian deposit money bank.

iii) Examine monetary policy rate relationship with the performanceof Nigerian deposit money bank

Hypotheses of the Study

The following hypothesis are relevant for this study:

H01: Lending interest rate has no significant relationship with the

performance of Nigerian deposit money bank.

H02: Deposit interest rate has no significant influence on the

performance of Nigerian deposit money bank.

H03: Monetary policy rate has no significant influence on the

performance of Nigerian deposit money bank

Scope of the Study

The research study on lending rate and the performance of Nigeria depositmoney bank covered the period of 20072017. This period was chosenbecause of the researcher felt that it would be better to use a period ofsteady democratic dispensation in Nigeria. This study will be limited tolending rate, inflation rate and it effect on the performance of Nigeriadeposit bank. Other relevant variable could have been studied but due totime and other resource.

Significance of the Study

The study will helps us understand the impact of an effective lending rateon the performance of the Nigeria deposit money Banks. It would aid theregulators to carefully plan and forecast the effects of its policies to meetits objectives of economic growth and full employment. To bankers, it wouldexpose the relationship existing between our relevant variables, which willbe of interest to them in their respective banks. This would also benefit theacademic community which would avail them the opportunity ofconducting further research in the topic of similar areas.

II. LITERATURE REVIEW

Conceptual Review

Interest Rate

Gilchris, (2013) states that although it is difficult to determine the directionof the relationship between lending rate and profitability, studies confirm


that lending rate instability affects Nigeria deposit money bankperformance while other studies give contradictory findings. The Centralbanks also lends Commercial Banks funds. Money borrowed from theCentral Bank is to be repaid at a particular interest rate (Monetary PolicyRate). This makes interest rate (lending rate) a powerful governmentregulatory tool for determining other interest rates in the banking industry.Hualan (1992) stated that interest rate is one of the most important factorsthat affect the bank financial performance. Corb (2012) argued that interestrate is an economic tool used by the Central Bank to control inflation andto boost economic development. Ngugi (2004) explained that low interestrates and small spread promote economic growth in big ways henceencouraged.

Ngure (2014) defined interest rates as the price a borrower pays forthe use of money they borrow from a lender (financial institution) or feepaid on borrowed assets. Sayedi (2013) expressed interest rate as thepercentage rate over a period of one year. Karl et al., (2009) posits thatinterest rates are derived from macroeconomic factors which agree withIrungu (2013) that interest rates are major economic factors that influencethe economic growth in an economy. Inflation and inflationary expectationscan press interest rate upward which affects lending rates resulting toreduce credit demand and lending ability of Commercial Banks (Keynes,2006). Irungu (2013) states that interest rate is the price of money. Interestrates can either be nominal or real. Nominal interest rate can be measuredin naira terms, not in terms of goods. The nominal interest rate measuresthe yield in naira per year, per naira invested while the real interest rate iscorrected for inflation and is calculated as the nominal interest rate minusthe rate of inflation (Pandey, 1999).

Bank Profitability and Financial Performance

The profitability of a bank is determined by interior and exteriordeterminants which agrees with (Ongore, 2013; AlTamini et al., 2010). Theinterior determinants are called micro or bank specific determinants ofprofitability because they are initiated from bank accounts like balancesheet or profit and loss account. While on the other hand, the exteriordeterminants are the variables which are not in the control of banks’management such as monetary policy interest rates. Chenn (2011) explainedthat these macroeconomic factors are significant in explaining firmperformance (profitability) and subsequent returns to investment. Gilchris,(2013) agrees that the financial performance is commonly measured byratios such as Return on Equity, Return on Assets.


There are many different mathematical measures to evaluate how wella company is using its resources to make profit (Irungu, 2013). Financialperformance can be measured using the following techniques; operatingincome, earning before interest and taxes, net asset value (Gilchris, 2013).Irungu (2013) described financial performance analysis as the process ofidentifying the financial strengths and weakness of the firm by properlyestablishing the relationship between the items of the balance sheet and profitand loss account. It’s the process of identifying the relationship between thecomponent parts of financial statements to ascertain an organization position,performance and prospects. Financial performance analysis can beundertaken by management, owners, creditors, investors (Chenn, 2011)

The performance of banks gives direction to shareholders in theirdecision making (Panayiotis et al., 2006). Wainaina, (2013) says the effectof macroeconomic factors in other sectors of the economy will always affectthe banking sector and what goes on in the banking sector will affect theother sectors of the economy. Chen et al., (1986) maintains that these macroeconomic factors are significant in explaining firm performance(profitability) and subsequent returns to investors. Gilchris (2013) agreesthat financial performance is commonly measured by ratios such as returnon equity, return on assets, return on capital, return on sales and operatingmargin. A firm has several objectives but profit maximization is said to beparamount among these (Damilola, 2007; KPMG, 2005; Raheman and Nasr,2007). Profit is a tool for efficient resources allocation because it is the mostappropriate measure of corporate performance under competitive marketconditions (Pandey, 2005).

Conceptually profit connotes the excess of revenue generated by a firmover its associated costs for an accounting period. Operationally the termprofit is imprecise, as many variants exist. The term profit could refer toprofit before tax, profit after tax, gross profit, net profit, profit per share,return on assets, among other variants (Damilola, 2007; Pandey, 2005)

Performance of Nigeria Deposit Money Bank

The financial system of most developing nations has come under stress asa result of the economic shocks of the 1980s. The economic shocks largelymanifested through indiscriminate distortions of financial prices whichincludes interest rates, has tended to reduce the real rate of growth andthe real size of the financial system relative to financial magnitude. In otherwords, banks do grant loans and advances to individuals, businessorganizations as well as government in order to enable them embark oninvestment and development activities as a means of aiding their growthin particular or contributing towards the economic development of a


country in general. Deposit money banks are the most important savings,mobilization and financial resource allocation institutions. Consequently,these roles make them an important phenomenon in economic growth anddevelopment. Therefore, no matter the sources of the generation of incomeor the economic policies of the country deposit money banks would beinterested in giving out loans and advances to their numerous customersbearing in mind, the three principles guiding their operations which areprofitability, liquidity and solvency (Ajayi, 2008). This study becomesimperative because deposit money banks in Nigeria need to understandhow to manage these huge assets in terms of their loans and advances. Forthe banks to balance their main objectives of liquidity, profitability andsolvency, lending must be handled effectively and the banks must behavein a way that their potential customers are attracted and retained. Agene(2001) argued that the effects of an increase in interest rate, other thingsbeing equal, will lead to a decline in aggregate demand partly becausethese will encourage savings to earn higher returns. On the other hand,Adam (2001) added that in a situation where the interest payments form asignificant portion of product costs, increased interest rates could result inreduced capital spending, investment, output and employment.

Theoretical Review

The Theoretical framework is guided by the work of Bekaert (1998) whichtries to analyze the influence of lending rate on performance on Nigeriadeposit money bank.This section considers theories such as loan Pricingtheory, banks lending rate, firm characteristic theory, theory of multiplelending, the signaling approach, credit market theory, classical theory ofinterest etc.

Loan pricing Theory

Banks cannot always set high interest rates. Banks should consider adverseselection and moral hazard because it is difficult to determine the borrowertype at the start of the banking relationship (Stiglitz and Weiss, 1981). Ifinterest rates are too high, it might cause adverse selection problems becauseonly high risk borrowers are willing to borrow. Once they receive the loansthey may develop moral hazard behavior since they are likely to take highlyrisky projects (Chodecai, 2004)

Loanable Funds Theory

This theory synthesizes both the monetary and non monetary impact ofthe problem (saving and investment process) (Wensheng, et al., 2002). It


assumes that interest rates are determined by supply of loanable fundsand demand for credit. It recognizes that money can play a disturbing rolein the saving and investment processes and thereby causes variations inthe level of income. The theory suggests that interest rates equate thedemand and supply of loanable funds. Loanable funds are the sum ofmoney supplied and demanded at any time in the money market. Loanablefunds theory has implications on banks savers and borrowers and eachside is well compensated at equilibrium, Interest rate should be structuredin a way every party feel comfortable (Emmanuelle, 2013)

Classical Theory of Interest

According to Keynes, the classical theory of interest is the savingsinvestment theory. It states that on the general equilibrium theory, the rateof interest is determined by the intersection of demand for and supply ofcapital which agreed with Caplan (2000). Fredman (1991) explains that thesaving and investment are the two real factors determining the rate ofinterest.

Rational expectations Theory of Interest Rates

This is based on the idea that people formulate expectations based on allthe information that is available in the market. It holds that the bestestimation for future interest rates is the current spot rate and that changesin interest rates are primary due to unexpected information or changes ineconomic factors. The limiting factors of rational expectation theory aremostly related to the difficulty in gathering information and understandinghow the public uses its information to form its expectation (Caplan, 2000).If interest rate rise will avoid borrowing, this in turn will affect bankperformance and vice versa (Bekaert,1998)

Credit Market Theory

It states that the term of the credit clears the market. If collateral and otherrestrictions remain constant, interest rate is the only price mechanism. Ifthere is an increasing demand for credit and the supply remains constant,the interest rate rises and vice versa. Ewert (2000) suggest that the higherthe failure risk of the borrower, the higher the interest premium.

Bank’s Lending Rate

By far the most visible and obvious power of many modern central banksis to influence market interest rates; contrary to popular belief they rarely“set” rates to a fixed number although the mechanism differs from country


to country/ most use a similar mechanism based on a central bank’s abilityto create as much fiat money as required. The mechanism to move themarket towards a “target rate7ʹ (which specific rate is used) is generally tolend money or borrow money in theoretically unlimited quantities untilthe targeted market rate is sufficiently close to the target (Adam, 2001).Central banks may do so to by lending money to and borrowing moneyfrom a limited number of qualified banks. For example, the Bank of CanadaSets a target overnight rate, and a band of plus or minus 0.25%. Qualifiedbanks borrow from each other within this band, but never above or below,because the Central bank will always lend to them at the top of the bandand take deposits at the bottom and lend at the extremes of the band areunlimited. This mechanism also implies to the Central Bank of Nigeria

Multiple Lending Theory

The theory posits that banks should be less inclined to share lending (loansyndication) when the equity markets are well developed. Mergers,acquisition and outside equity increase banks’ lending capacity and reducesthe need for greater diversification and monitoring (Carletti, 2006; Ongeneand Smith, 2000; Karceski, 2004; Degryse, 2004).

Empirical Review

Adofu and Audu (2010) used ordinary least square method to ascertainthe assessment of the effects of interest rate deregulation in enhancingagricultural productivity in Nigeria. The study found out that interest rateplay a significant role in enhancing economic activities and as such,monetary authorities should ensure appropriate determination of interestrate level that will break the double edge effect of interest rate on saversand local investors.

Rasheed (2010) used error correction model (ECM) to investigateinterest rates determination in Nigeria. The study found out that as theNigerian financial sector integrates more with global markets, returns onforeign assets will play a significant role in the determination of domesticinterest rates.

Newman (2012) used regression and collection methods to examinedthe relationship between interest rate and the performance of Nigeriandeposit money bank. In this study he analysis published audited accountof twenty banks from 19802009 from the central bank of Nigeria statisticalbulletin. He found out that the interest rate policies have not improved theperformance of the banks significantly and also have contributedmarginally to the growth of the economy for sustainable development.


Owni and Ajaude (2013) used multiple regression in his study toexamined the influence of lending rate on the performance of Nigeriandeposit money bank, and twenty four 24 active deposit money banks inNigeria are used to formulate the population of the study. Data are extractedfrom the central bank of Nigeria statistical bulletin from 19862012. Findingshow that there is significant relationship between lending rate and othervariable with the money deposit bank in Nigeria.

Okoye and Eze (2013) used econometrics data in a regression methodto investigate the effect of bank lending rate on the performance of Nigeriandeposit money bank between the year 20002010, data were collected fromthe central bank of Nigeria statistical bulletin and others relevant journalsand the findings show that lending rate and monetary policy rate hassignificant and positive effects on the performance of Nigerian depositmoney bank

Irungu(2013), used ordinary least square method to analysis the variableon effect of monetary policy rates on the profitability of the Nigeria banks,data were collected from the central bank of Nigeria statistical bulletin andothers relevant journals and the findings show that lending rate and monetarypolicy rate has significant. The concluded that government should use afavorable monetary policy rate to have a healthy economy growth.

Ngure (2014) used ordinary least square method to ascertain theassessment of the effects of interest rate deregulation in controlling theNigeria economy. The study found out that interest rate play a significantrole in enhancing economic activities and as such, monetary authoritiesshould ensure appropriate determination of interest rate level that willbreak the double edge effect of interest rate on savers and local investors.

Enyioko (2015) used regression and collection methods to examinedthe relationship between monetary and the profitability of Nigerian depositmoney bank. In this study he analysis published audited financial accountof 15 banks from 19802009 and the central bank of Nigeria statisticalbulletin. He found out that the monetary policies rate have no significanton the performance of the banks and also have contributed to the growthof the economy.

Adeosun and Habeeb(2015) used multiple regression in his study toexamined the impact of lending rate on the economic growth. Data areextracted from the central bank of Nigeria statistical bulletin from 19802014. Finding show that there is significant relationship between lendingrate and the Nigerian economy the monetary authority should apply afavorable rate that will improve the economic activities.


III. RESEARCH METHODOLOGY

Research Design

The research design employed Descriptive and Expost facto ResearchDesign. Descriptive research design method helps in gathering informationabout the existing status of the phenomena in order to describe what existsin respect to variables.

Model Specification

The theoretical foundation is based on the study of Okoye and Eze (2013)and Udeh (2015), who used Pearson Product moment correlation techniqueto analyze the data collected

Model Estimation

This is expressed functionally as;

ROE= f ( LIR,DIR,MPR) (1)

The linear regression equation for the model are

ROE = b 0 + b1LIR + b2 DIR+b3MPR+ µ t (2)

Where

ROE = Dependent Variable (Y t)

LR = Lending interest Rate (X t)

DIR = Deposit interest rate (X t).

MPR= Monetary policy rate

B0= Intercept

B1b3= co efficient of the independent variables

t = Time series (Annual)

µt = Error or disturbance term.

Source of Data

The data required for the study was obtained from secondary sources thatwere used to investigate the relationship between dependent andindependent variables. The study used secondary data sources to gatherinformation relevant to the research objectives. The study covered data forinterest rate and Nigeria deposit bank within 20072017 ,which wascollected from the Central Bank of Nigeria; website and statistical bulletin,annual reports and the internet.


IV. DATA PRESENTATION AND ANALYSIS

Data Presentation

See appendix i

Data Analysis

Interpretation of Result and Discussion of Findings

Table 4.1: OLS Regression Results

Dependent Variable: ROE

Method: Least Squares

Date: 06/24/19 Time: 13:11

Sample: 2007 2017

Included observations: 11

Variable Coefficient Std. Error tStatistic Prob.

C 103.3616 169.9801 0.608080 0.5623

LIR 12.85291 9.318188 1.379335 0.2102

DIR 24.89479 8.310319 2.995648 0.0201

MPR 6.807494 3.444906 1.976105 0.0887

Rsquared 0.802427 Mean dependent var 33.06909

Adjusted Rsquared 0.717752 S.D. dependent var 44.91613

S.E. of regression 23.86258 Akaike info criterion 9.457788

Sum squared resid 3985.959 Schwarz criterion 9.602477

Log likelihood 48.01783 HannanQuinn criter. 9.366582

Fstatistic 9.476632 DurbinWatson stat 2.825184

Prob(Fstatistic) 0.007341

Source: Author’s computation 2019.

Looking at the above regression output, the DurbinWatson Statisticsof 2.83 showed that there was no presence of positive serial correlationwhich could render the estimated model result biased. Thus, the resultswere reliable and meaningful economic and standard inference could bemade. Hence, from the multiple linear regression results on table 4.1, theregression equation predicting the relationship between the Return onequity of banks in Nigeria (ROE) and Lending interest rate (LIR), Depositinterest rate (DIR) and Monetary policy rate(MPR) can be stated as:

ROE = 10.3987 – 0.1097LIR + 0.2321DIR – 0.0019MPR (4.1)

From the equation of best fit estimated above, it could be deduced thatwhile Deposit interest rate (DIR) and monetary policy rate (MPR)maintained positive relationship with the Banks performance (ROE),Lending interest rate (LIR) maintained negative relationship. Due to the


negative relationship exhibited by LIR, 1% increase in LIR was associatedwith negative impact of reducing the average mean value of ROE by about1200% and vice versa. Furthermore, since DIR maintained a directrelationship with ROE, it followed that 1% increase or decrease in Depositinterest rate would culminate in about 2400% increase or decrease in theaverage mean value of Banks performance in Nigeria. Similarly, Monetarypolicy rate (MPR) had positive relationship with Banks performance suchthat 1% increase in MPR would resort to about 600% increase in Banksperformance and vice versa. Lending interest rate showed the expectednegative relationship because it was expected that increase in lendinginterest rate by the deposit money banks should discourage lending fromthe investing public and this would eventually lead to loss of interest incomeon the part of the banks. In case of deposit interest rate it also did not meetexpected relationship because increase in the interest payable on depositby the banks would reduce the available revenue and profitability level ofthe banks as more part of the interest earned is used to pay interest ondeposit to the customers. However, increase in deposit interest rate mayencourage more deposit from the customers, increase in deposit will leadto increase in loanable and lending capacity of banks and increase in lendingwould generate more interest incomes to the bank.

He multiple correlation coefficient (R) of 0.89 indicated a strong linearrelationship between the dependent variable which was the Return onequity of banks in Nigeria (ROE) and the interest rate (LIR, DIR and MPR))since the value was close to 1. Also, the coefficient of determination (R2) of0.80 indicated that about 80% of the variation in the performance of banks(ROE) could be accounted for by the variations in the independent variables(LIR, DIR and MPR) while the remaining 20% was accounted for by otherextraneous variables not captured in the model. Furthermore, the standarderror of the model which was 23.86 was considered to be moderately highagainst expectation. The R2 adjusted for the number of parameter (nk) was0.71 which was significant.

Durbin Watson Statistic of 2.82 was higher than the R2 value of 0.8 andmore than the benchmark value of 2 which freed the model variables fromautocorrelation complicity. Tratios measured how large the coefficients ofthe parameters will vary if carried out on repeated sampling of theobservations. Thus, DIR has highest tratio of 12.99, it thus means that DIRwould have very little variation in repeated sampling than MPR and LIRwhich has lower tratios. Moreover, looking at the significance of each ofthe coefficients of the predictors, only DIR was statistically significant tothe specified model judging from their pvalues. However, the fstat of themodel which was 9.47 significant and indicated that the proportion of


variation in the banks performance accounted for by the interest rate wasnot due to chance or error.

4.3. Test of Hypotheses

The decision rule for testing hypothesis is that Null Hypothesis (H0) shouldbe rejected and Alternate Hypothesis (H1) accepted if Pvalue is less than0.05 threshold and vice versa.

1. H01: Lending interest rate does not impact significantly on banks

performance in Nigeria;

H11: Lending interest rate impacts significantly on banks performance

in Nigeria in Nigeria.

From table 4.1, since Pvalue of LIR which is 0.2102 was more than thecritical value of 0.05, H0

1 was accepted and H1

1 rejected. This meant that

Lending interest rate has no significant on Nigerian banks performance inNigeria. This might be a pointer to high interest rate that is predominant inthe economy which keeps discouraging customers from approaching banksfor credit. It also meant that interest rate during the period under study didnot determine the profit level of banks in Nigeria in a significant manner..

2. H02: Deposit interest rate has no significant impact on the Nigerian

banks performance.

H12: Deposit interest rate has significant impact on the Nigerian banks

performance.

Also, since Pvalue of 0.02 was less than the critical value of 0.05, therewas no enough reason to accept the H0

2; it thus meant that deposit interest

rate has significant impact on Nigerian banks performance.Although DIRrelationship contradicted the apriori expectation that high deposit interestrate should reduce the profit earned by the banks, nevertheless, the increasein deposit level which might associate with increase in deposit interest canpromote banks performance in terms of return on equity and profitabilityin Nigeria. Thus, the deposit interest of the money deposit banks over theperiod this study has translated to positive growth in the profitability levelof Nigerian banks.

3. H03: Monetary policy rate has no significant impact on the Nigerian

banks performance.

H13: Monetary policy rate has significant impact on the Nigerian banks

performance in Nigeria

On the contrary, MPR has Pvalue of 0.08 which was greater than criticalvalue of 0.05, there was no enough reason to reject the H0

3; this translated

that Monetary policy rate has no significant impact on banks performance


in Nigeria. The forgoing contradicted the expected outcome that increasein the rate at which BN lend to deposit money banks should discourageborrowing by the public and reduce banks performance in terms of profitsignificantly.

V. CONCLUSION

Nigerian deposit money banks remain dominant in the banking system interms of their shares of total assets and deposit liabilities. Their interestrate policy, a major component of total credits on the increase in spite ofthe major constraints posted by the government regulations, institutionalconstraints and other macro economic factors. I concludes that, bothgovernment and deposit money banks should be mindful of the facts thatinterest rate in which they operate in, for the bank performance. Wherethe interest is conducive and supportive, performance of banks will beeffective and efficient and increase the profitability of the bank. But wherethe interest rate are hash the Deposit money banks should note that theyneed to do a lot in order to ensure good lending behavior even where agood measure ofmacroeconomic stability is achieved. This is because ofthe positive and significant relationship found between bank interest rateand bank performance in both short and long run.

Recommendations

Based on the findings in this study, the following suggestions arerecommended:

1. Having seen that there exists a long run and short run relationshipbetween return on equity and explanatory variables (LIR,DIR, andMPR) through the use of multiple regression model, governmentshould adopt policies that will help Nigerian deposit money banksto improve on their performance.

2. There is need to strengthened bank lending rate policythrough effective and efficient regulation and supervisoryframework.

3. Banks should try as much as possible to strike a balance in theirloan pricing decisions. This will help them to be able to covercost associated with lending and at the same time, maintain goodbanking relationship with their borrowers.

References

Abayomi, T. O. and Adebayo, M. S. (2010). Determinants of Interest Rates in Nigerian. Journalof Economics and International Finance, 2(12): 261271.


Adofu, M. I. and Audu, S. I. (2010). An Assessment of the Effect of Interest Rate Deregulation inenhancing Agricultural production in Nigeria. Current Research Journal of EconomicTheory, 2(2): 8286.

Agene, C. E. (2001). The Principles of Modern Banking; Lagos: Gene Publishers.

Ajayi, I. S. (2008). “The Transmission of Mechanism of Monetary Policy”:Paper Presented atthe 3rd Seminar for the Press Conducted by Central Bank of Nigeria at Port Harcourt,(20th 22nd, August).

AlTamini, H. and Hassan, A. (2010). Factors Influencing Performance of the UAE Islamic andConventional National Banks.

Bekaert G. (1998). Regime Switches in Interest Rates. Cambridge:Mass National Bureau ofEconomic Research

Caplan, B. (2000). Rational Expectation. George Mason University, Department of Economic,D. A. (2007).

Carletti, E., Cerasi, V. and Daltung, S. (2006), “MultipleBank Lending: Diversification andFree riding in Monitoring, Working Paper, Department of Statistics: Universita degliStudi di Milano Bicocca.

Chen M. (2011). Interest Rate Spread. Washington D.C. International Monetary Fund:Monetary and Capital Market Dept.

Cooper, R., and Schinder, S., (2008). Business Research Methods. 10th Ed. New York: McGraw Hill.

Corb(2012). Interest Rate Swaps and other Derivatives. New York. Columbia Business School.

Degryse H. (2004). SMEs and Bank Lending Relationships: the Impact of Mergers. NationalBank of Belgium Working Paper, No. 46.

Emmanuelle, J. (2003). Monetary and Fiscal policy. Kenya University of Nairobi.

Ewert, R. and Schenk, G. (1998). Determinants of Bank Lending Performance. Working Paper,Center for Financial Studies, University of Frankfurt:World Bank working paper series,500p

Ezirim, N. (2005). Impact of Interest Rate Policy and Performance of Deposit Money banks inNigeria. Global Journal of Management and Business Research. 12(21). ISSN Online:22494588, ISSN Print: 09755853

Friedman, B.M. and Kutnner, K.N. (1991). Why does the Paper Bill Spread Predict Real EconomicActivity? Cambridge, M.A., National Bureau of Economic Research.

Gilchris, M. (2013). Influence of Bank Specific and Macroeconomic Factors on the Profitability of25 Commercial Banks in Pakistan during the Time Period 20072011. American Journal ofBusiness and Finance.

Giovanni, J. and Shambaugh J.C. (2006). The Impact of Foreign Interest rates on the Economy:The Role of the Exchange Rate Regime. Cambridge, National Bureau of Economic Research.

Hara M. (1983). A dynamic Theory of the Banking Firm. The Journal of Finance.

Hayes, K.C. (2013). How Interest Rates Impact Banks’ Bottom Lines: A look at History. ValueWalk: http://www.valuewalk.com/2013/06interest rates banks earnings good/.Retrieved April15, 2016.

Hualan, C. (1992). Fiscal Policy and Interest Rates, How Sustainable is the New Economy?Washington D.C.: International Monetary Fund.

IMF International .Monetary Fund, (2001). Regional Economic Outlook: SubSaharan Africa.World Economic and Financial Surveys, IMF Working Papers, 02587440.60.


Irungu, P.N. (2013). The Effect of Interest Rate Spread on Financial Performance of CommercialBanks in Kenya. An Msc Research Project submitted to the University of Nairobi.

Karl, E., Ray C. and Shannon, M. (2009). Principles of Economics. Pearson InternationalEdition. Pretence Hall.

Keynes (2006). The General Theory of Employment, Interest and Money. New York: AtlanticPublishers and Distributors; Original Published 1936.

KPMG, (2005). Working Capital Management Survey: How European Companies manage theirWorking Capital. KPMG International. Retrieved on 15/03/2016 from http://www.ch/library/pdf/kpmg survey Working Capital Management_e.pdf

Kothari, C. R. (2008). Research Methodology: Methods and Techniques. New Delhi: New AgeInternational Publishers,ISBN: 9788122436235, 3rd Edition.

Ngugi , R. (2004). Determinants of Interest Rates Spread In Kenya, Nairobi. Kenya Institute forPublic Policy Research and Analysis.

Ngure I.M. (2014). The effect of Interest Rates on Financial Performance of Commercial Banks inKenya. An MSc Research Project Submitted to the Department of Businees Admin.,University of Nairobi.

Ongore, V. (2013). Determinants of Financial Performance of Commercial Banks in Kenya,International Journal of Economics and Financial Issues, ISSN: 21464138.

Okoye, V. and Eze, R.O. (2013). Effect of bank lending rate on the performance of Nigeriandeposit money banks, International Journal of Business and Management Review.

Ongena, .S.,and Smith, D. C. (2000). What Determines the Number of Bank Relationships: CrossCountry Evidence, Journal of Financial Intermediation, 9(5): 2656.

Pandey, I. M. (2005). Financial Management. 9ed, New Delhi, Vikas.

Raheman, A. and Nasr, M. (2007). Working Capital Management and Profitability Case ofPakistan Firms, International Review of Business Research Papers.

Rasheed O.A. (2010). Interest Rate Determinants in Nigeria, International Journal of Financeand Economics 2(3): 112.

Sayedi, S.(2013). Bank Specific, Industrial Specific and Macroeconomic Determinants of Banks’Profitability in Nigeria. Journal of Finance.

Stigliz, K. & Weiss, U. (1981). Financial asset of economic development. Washington Dc BrookingInstitution.

Uchendu, O. (1995). Monetary Policy and the Performance of Commercial Banks in Nigeria.Central Bank of Nigeria Economic Review.

Udeh S.N. (2015). Impact of Monetary Policy Instruments on Profitability of Commercial Banksin Nigeria: Zenith Bank Experience. Research Journal of Finance and Accountingwww.iiste.org ISSN 22221697. Available on www.iiste.org

Victor (2013). Effect of Bank Lending Rate on the Performance of Nigerian Deposit Bank,International Journal of Business and Management Review.

Wainaina (2013). Effect of macroeconomic factors on Commercial Banks Lending to AgriculturalSector in Kenya. MBA Project University of Nairobi

Wensheng, P., Lai, K., Leung, F. and Shu, C. (2002). Impact of Interest Rates Shocks on the ofBanking Sector. Hong Kong: HonS


Owolabi, A. Effect of Lending Rate of the Performance of Nigerian Deposit Money Bank.Journal of Asian Economics, Accounting and Finance, Vol. 1, No. 2, 2020, pp. 163180


Appendix

Data Presentation

YEAR ROE LIR DIR MPR

2007 36.83 16.94 3.55 9.50

2008 34.11 15.14 2.84 9.75

2009 64.72 18.99 2.68 6.00

2010 16 17.59 2.21 6.25

2011 0.28 16.02 1.41 12.00

2012 22.2 16.79 1.70 12.00

2013 23.21 16.72 2.17 12.00

2014 44.84 16.55 3.38 13.00

2015 56.78 16.85 3.58 11.00

2016 96.56 16.87 3.75 14.00

2017 98.23 17.78 5.16 14.00

Source: Daniel and John, 2016 and CBN Statistical bulletin 2017

EFFECT OF MACRO-ECONOMIC VARIABLES ONPROFITABILITY OF SELECTED COMMERCIALBANKS IN RWANDA

Daniel Twesige and Faustin Gasheja

Lecturer University of Rwanda, Collage of Business and Economics, School of Business,Department of Accounting, E-mail: [email protected]

Received: 26 May 2020; Revised: 3 August 2020; Accepted: 9 August 2020; Online: 29 December 2020

Abstract: The purpose of this study was to determine the effect of macroeconomic variables onthe profitability of selected commercial banks in Rwanda.

A descriptive research design was used and data was collected from four commercial banks inRwanda licensed by the National Bank of Rwanda (NBR). Secondary data covering a period ofthree years, from 2016 to 2018 were sourced from the published financial reports of the selectedcommercial banks and the National Bank of Rwanda. The study used a multiple regression analysisin examining how macroeconomic variables affect the financial performance of commercial banksin Rwanda.

Findings of this study show that there is a negative correlation between the macroeconomicvariables performance indicators and ROE and a positive correlation between the macroeconomicvariables and the ROA. Based on the findings, the researchers conclude that macroeconomicvariables factors are very fundamental to the profitability of the commercial banks thereforecommercial banks should clearly analyse the macroeconomic environment where they operate.

Keywords: GDP, ROA, ROE Inflation rate, Interest rate, unemployment rate, Macroeconomicfactors, profitability, Banking sector

1. INTRODUCTION

Financial institutions globally facilitate the movement of funds from thesurplus units in an economy to the deficit supply units (Eakins & Mishkin,2012). Commercial banks form a part of these financial institutions. InRwanda commercial banks are the major players in the financial system(NBR, 2016). Banks undertake their financial intermediation role with thegoal of maximizing their returns, also; through improved financialperformance. As such, they accumulate and deploy assets towardsachieving the desired performance.

Banks also operate in the industry and national environment. Thisenvironment is often turbulent and volatile as a result of interaction betweenand among various forces, among them macroeconomic variables.



Governments often enact legislation desired to achieve certainsocioeconomic goals; these legislative enactments and other governmentintermediations in the market influence the macroeconomic environment(Osamwonji & Chijuka, 2014).

The agency theory (Jensen & Meckling, 1976) postulates that firmmanagers (agents) should be in a position to anticipate the macroeconomicenvironment changes and take adaptive measures for them to safeguardand maximize their firms’ returns. This study sought to investigate howmacroeconomic variables affect the financial performance of commercialbanks in Rwanda.

Macroeconomic variables are the elements that typify the nationaleconomy and business environment. In an economy, these macroeconomicfactors are not within the influence of one individual firm (Brueggeman &Fisher, 2011). However the government often influences the macroeconomicvariables through enactment of legislation and or policies. These factorsinclude the inflation rate, GDP, interest rate, foreign exchange rate, moneysupply, and so on (Simiyu & Ngile, 2015). Macroeconomic variablesinfluence the complexity and volatility of the business setting (San & Heng,2013). Due to increasing globalization and technological advances, economicturbulence in other (international) economies might creep into the localbusiness environment.

The government has a precarious role in enhancing stability of themacroeconomic variables. Businesses, among them commercial banks,prefer a stable macroeconomic environment; a stable environment is morepredictable, risk is also lower under such stable conditions. Financialperformance denotes the percentage or degree of attainment of economicgoals, objectives and or targets by a firm. Financial performance is specifiedas at a stated point in time and refers to performance in a given time period(Pandey, 2009). Financial performance is measured in various ways. TheFinancial performance of commercial banks is best measured using ratiossuch as return on assets, return on equity, net interest margin, equitymultiplier, and nonperforming loans (Eakins & Mishkin, 2012). Return onassets ratio is a ratio of net income to total assets; the return on equity is aratio of net income and shareholders’ equity; net interest margin is thedifference between interest expenses paid out and interest income earnedby a bank.

Firms seek to improve continuously their financial performance forvarious reasons, among them to maximize shareholder returns(Brueggeman & Fisher, 2011). Returns to shareholders are largelydependent and linked to the financial performance registered by

Effect of Macro-Economic Variables on Profitabilty of Selected Commercial Banks... 183

commercial banks. Good financial performance is essential as resourcesavailable to a firm are scarce; better financial performance leads to surplusinflow of resources to the banks, these resources are then available to bedeployed for further growth, undertake expansion purposes, or to justsustain the commercial banks going concerns.

Macroeconomic variables are anticipated to influence the businesssetting (Brueggeman & Fisher, 2011). These variables affect the nature andintensity of the volatility of the operating environment. According to(Markowitz, 1952), the portfolio theory states that investors will makedecisions on the riskreturn tradeoff; such investors tend to prefer morereturns to less returns, they also favor less risk to higher risk. High volatilityof variables in the macroeconomic environment creates and fosters anunstable and highly volatile environment, risk, thus becomes aggravatedand in turn threatens returns. Good and healthy financial performancethen becomes uncertain.

The theory of efficient market hypothesis (Fama, 1970) postulates that,in a market security price will reflect all the available information, always.Bank managers as such therefore ought to react fast and accurately to actualand anticipated macroeconomic variable changes by adapting the saidchanges or planning for them well in advance. Such prudence assists to assurefinancial performance not only in the present, but also in future.Macroeconomic variables affect firms’ profitability (Gerlach, Peng & Shu,2005). Changes in macroeconomic variables present opportunities as well asthreats to the industry players concurrently; those prepared for the changes,shall realize gains from the opportunities that arise, thus fostering theirfinancial performance, while those who are unprepared might suffer fromthe threats and might in turn impact their financial performance negatively.

The Rwandan banking sector has faced a challenging macroeconomicenvironment such as capping of interest rate that was affected, but thesector remains resilient. Other macroeconomic challenges that affectedthe sector include; increasing levels of prices, unpredictability of interestrates and exchange rate variability. The Rwandan francs have greatlydepreciated against most traded world currencies over the last few years,in addition to widening current account deficit. These unfavorablemacroeconomic variables may result to great problems in the bankingindustry, when management deeds are faroff reflecting the recurringnature of the economy in its decisions. Mounting stress within the bankingsystem can be experienced due to extremely unexpected cyclical fluctuation.Nevertheless, the macroeconomic variables might well deliver goodindicators, but it’s not always the case. (NBR, 2016)


In Rwanda the national bank through its monetary policy committee,sets the base lending rate and as such influences the prevailing lendingrates in the economy (NBR, 2016). Banks in Rwanda have also showedgrowth in profitability over the previous few years, with the exception of afew that reported poor performance.

The Rwandan government also influences macroeconomic variablesnot only through economic and fiscal policies, but also through marketactivities such as issuance of government debt securities, capping of interestrates in the economy. Banks are left with the only option of adapting tomacroeconomic changes in order to protect and safeguard their futurefinancial performance.

A number of research studies (foreign & regional) have embarked onthis research area. Osamwonji and Chijuka (2014) investigated howmacroeconomic variables affect the profitability of commercial banks. Thestudy finds a significant positive relationship between the return on equityand GDP, a significant negative relationship between return on equity andinterest rate, and an insignificant negative relation involving inflation rate.San and Heng (2013) found macroeconomic variables like gross domesticgrowth and inflation do not have an effect on profitability.

Local study that has been conducted by Nkurikiye and Uwizeyimana(2017) in study called the effects of GDP, interest rate, and inflation onprivate investment in Rwanda, the research has revealed that gross domesticproducts affects private investment and both in the long run and shortrun, and study has revealed that there is a positive impact of inflation oninvestment.

Another local empirical study that has been conducted by Gatsimbaziand Mulyungi (2018) in study called ‘’effects of macroeconomic variableson stock market performance in Rwanda, the study has revealed that GDPgrowth rate, inflation and exchange rate are negatively significant in theRwandan stock market.

Dukundane and Rukera (2016) in their research entitled the impact ofcredit risks management on financial performance of commercial banks inRwanda, they research findings has revealed that there is a negativecorrelation between non performing loan and financial performance ofbanks.

Even though many researchers (local) worked on financial performancecases, but they didn’t deepen on the relationship between themacroeconomic variable and financial performance, and many researchershave been putting blame on internal control of commercial towards


financial performance. Yet they are other factors that are beyond the controlof internal control that can affect the profitability of commercial banks inRwanda. This study sought to fill this gap by seeking to answer the researchquestion: What are the effects of macroeconomic variables on the financialperformance of commercial banks in Rwanda?

2. THEORETICAL AND CONCEPTUAL FRAMEWORK

Various theories have been discussed presenting arguments that guidedthis study. These theories include Schumpeter economic cycle theory,Keynesian liquidity preference theory and macroeconomic theory.

2.1. Schumpeter Economic Cycle Theory

The theory was propounded by Schumpeter (1939) who indicated theprocess of economic change or evolution that consists of two distinct phases,“prosperity” and “recession”. One under which the impulse ofentrepreneurial activity, draws away from an equilibrium position, andthe second of which it draws toward another equilibrium position.Schumpeter calls those fluctuations/cyclical processes in the economic lifebusiness cycle. Schumpeter shows the intermediary role of the financialsector between those who save and invest, through a process referred to ascredit creation by bank financing that leads to economic growth anddevelopment. The effect of this process leads to profit and loss generatedby the lender and the borrower.

Certain macroeconomic variables, typically display a unique patternof boom and recession in a business cycle. A crisis is said to occur at thepeak of expansion when growth in real GDP and domestic demand declineleading to acceleration in inflation. During periods of economic expansion,firms and their respective sectors profits increases, asset prices risesaggregate sectoral demand for credit facilities, expands leading to growthin bank lending resulting to increased interest income. Banks mayunderestimate their risk exposures, relaxing credit standards and reduceprovisioning for future losses while the economy indebtedness rises.

As the downturn sets in individual’s, firms and sector profitabilitydeteriorates (Bikker and Hu, 2002). The theory assumes that recessionsand periods of economic growth are an efficient response to exogenouschanges in the real economic environment and that decline in profitabilityresult in fall of asset prices, nonperforming loans, lowers borrowers’financial capacity, fall in employment levels, and depresses the value ofcollaterals as a secondary mean of servicing debts. Banks’ risk exposureincreases, and consequently raises the need for larger loan provisions and


higher levels of capital, exactly when it is more expensive or simply notavailable. This may lead to banks reacting by reducing the amount oflending, especially if they have low capital buffers above the minimumcapital requirement, thus increasing the effects of the economic downturnas well as increasing the lending rates.

Critics of the theory state that it is a common misconception thatmacroeconomic purely based on shocks to supply, as opposed to shockson demand, and this leads to the common criticism of Schumpeter economiccycle theory by ignoring the demand side of the economy. However, in thereal business cycles situation, consumers will change their intertemporalconsumption and savings decisions based on the real interest rate availableto them, which is a shift in demand. In relation to the study, the theoryviews interest rate changes as normal economic occurrences which willaffect commercial bank’s performance. It disregards the argument thatinterest rates are determined by the liquidity in the economy, but isdetermined by the prevailing macroeconomic environment as determinedby the business cycles. Hence, according to the theory, interest rates willkeep on changing according to the prevailing macroeconomic conditions.

2.2. Keynes’s Liquidity Preference Theory

The theory was advanced by Keynes (2006). According to the liquiditypreference theory, the interest rates are determined by the demand forand supply of money balances. The theory assumes that people’s demandfor money is not for transactions purpose, but as a precaution and forspeculative purposes. The transaction demand and precautionary demandfor money increases with income, while the speculative demand is inverselyrelated to interest rates because of the forgone interest. The supply of moneyis determined by the monetary authority (the central bank), by the lendingof commercial banks and by the public preference for holding cash (Were,Kamau, Sichei, Kiptui, 2013).

Therefore, interest rates are expected to increase as the maturity profileof securities increases. This is so because the longer the maturity, the greateris the uncertainty; and therefore the premium demanded by investors topart with cash increases as the maturity profile increases. The expectation,therefore, is that forward exchange rates should offer a premium overexpected future spot exchange rates for those who are riskaverse demanda premium for securities with longerterm maturities. A premium is offeredby way of greater forward rates in order to attract investors to longertermsecurities. Consequently, current interest rates reflect expected inflationrates, income (GDP) and expected money supply changes (Were et al., 2013).


Critics of this theory argue that the liquidity preference theory ofinterest suffers from a fallacy of mutual determination. Keynes alleges thatthe rate of interest is determined by liquidity preference. In practice,however, Keynes treats the rate of interest as determining liquiditypreference. The critics state that “The Keynesians therefore treat the

The rate of interest, not as they believe they do as determined byliquidity preference but rather as some sort of mysterious and unexplainedforce imposing itself on the other elements of the economic system (Wereet al., 2013). In relevance to the study, the theory views interest rates asbeing mainly driven by the liquidity level in the economy. The theory doesnot recognize the role of macroeconomic policies formulated by the centralbank, but interest rates are purely driven by the demand of money in theeconomy. Therefore, interest rates will go up and down according to thelevel of liquidity in the economy and preference for the liquidity by theusers of funds.

2.3. Macroeconomic Theory

The theory was proposed by Friedman, (1963). The theory has viewedinterest rates as always and everywhere a monetary phenomenon(Friedman, 1963). Further, macroeconomic theory assumes that growingthe money supply in excess of real growth causes interest rates to rise.This is also the result from the Harberger (1963) model, which assumesthat prices adjust to excess money supply in the money market. It is on thebasis of this assumption that it is possible to invert the real money demandand control interest rates.

Interest rate volatility in open economy results from differentdisequilibria in many markets specifically, the domestic money market,external/foreign markets and the labor market. Thus an increase in interestrates emanates from three main sources that include excess money supply,foreign prices and cost push factors (Were et al., 2013). The theory is relatedto keysian liquidity preference theory, but recognizes additional sourcesof interest rates not only demand for money but also foreign prices andcost push factors.

Critics of this theory base their argument on the grounds thatgovernments would in practice be unlikely to implement theoreticallyoptimal policies. According to them, the implicit assumption underlying themacroeconomic revolution was that economic policy would be made by wisemen, acting without regard to political pressures or opportunities, and guidedby disinterested economic technocrats. They argued that this was anunrealistic assumption about political, bureaucratic and electoral behavior.


In relevance to the study, macroeconomic theory views growing moneysupply in excess of real growth as the cause of interest rates to rise. Interestrate volatility is seen by the theory as emanating from three main sourcesthat include excess money supply, foreign prices and cost push factors.Interest rate volatility will also result from different disequilibria in manymarkets specifically, the domestic money market, external/foreign marketsand the labor market. Hence, controlling interest rate volatility will involvedealing with disequilibrium in the markets.

Yuqi (2008) examined the determinants of 123 United Kingdom (UK)banks profitability and its implication on risk management from 1999 to2006. The study utilized multiple regression models and panel dataestimation. The econometric results indicate that capital adequacy hassignificant positive impacts on profitability but inflation has insignificantpositive impact on profitability. Liquidity and credit risk had significantnegative impacts on profitability though; GDP and interest rate haveinsignificant negative impacts on the profitability of banks in UK.

Buyinza (2010) investigated samples of 23 commercial banksprofitability from 1999 to 2006 in Sub Sahara Africa countries. The studyutilized panel data and the regression results revealed that capital, efficientexpenses management, bank size, credit risk, diversified earning ability ofthe banks, per capital GDP, growth rate and inflation have significant andpositive impact on banks’ profitability. Ali, Akhtar, and Ahmed (2011)examined the bank specific and macroeconomic indicators of 22 publicand private sector commercial banks profitability from 2006 to 2009 inPakistan. The research made use of multiple regression models and paneldata estimation. The study found that bank size, operating efficiency, assetmanagement and GDP had positive effect on banks’ profitability. However,capital and credit risk had negative effect on banks profitability in Pakistan.

Gul, Irshad, and Zaman (2011) studied the factors affecting samples of15 commercial banks profitability from 2005 to 2009 in Pakistan. Theinvestigation utilized a regression model, panel data estimation and PooledOrdinary Least Square (POLS) method of computation with the aid of aneconometric package. The econometric result indicated both internal andexternal factors such as bank size, loan, deposit, GDP, inflation and marketcapitalization have significant positive influence on banks profitabilitymeasured by Return on Assets (ROA). Still in Pakistan, Gilchris (2013)examined the influence of bank specific and macroeconomic factors onsamples of 25 commercial banks profitability from 2007 to 2011 in Pakistan.The regression results indicated that bank size, net interest margin, andindustry production growth rate had positive and significant impact on


the profitability (ROA and ROE). Non performing loan to total advancesand inflation have negative and significant impact on ROA while GDP haspositive impact on ROA. Capital ratio has positive significant impact onROE.

Saidu and Tumin (2011) investigated the performance and financialratios on samples of four Malaysian and nine Chinese commercial banksfrom 2001 to 2007. The research made use of panel data and the regressionresults show that credit, capital and operating ratios have influence on theperformance of banks in China which is not true for Malaysia. The studyfound that liquidity and size of the banks do not influence the performanceof the banks in both countries. Khrawish, and Siam, (2011) investigatedthe determinants on samples of three Jordan Islamic banks profitabilityfrom 2005 and 2009.

The multiple linear regression results show capital, bank size, financialrisk, GDP growth rate, inflation, and the exchange rate have significantnegative relationship with profitability but credit risk has an insignificantpositive relationship with the profitability of Islamic banks in Jordan. Rachdi(2013) examined what determines the profitability of banks during andbefore the international financial crisis. The study samples 10 Tunisianbanks from 2000 to 2010.

The regression results indicate that, before the US subprime crisis,capital adequacy, liquidity, bank size and yearly real GDP growth affectpositively the performance (ROA,ROE and NIM) of the Banks. However,costincome ratio, yearly growth of deposits and inflation rate are negativelycorrelated across all measures of bank profitability. In crisis period, bankprofitability is mainly explained by operational efficiency, yearly growthof deposits, GDP growth and inflation.

Lucas and Anne (2010) examined the effect of macroeconomicdevelopments on performance, credit quality and lending behavior of banksin Kenya, by estimating a dynamic panel data model using GeneralizedMethod of Moments. The study suggested that banks needed to continuepursuing risk sensitive loan pricing policies to ease the extent of countercyclical behavior during economic upswings/downswings respectively,which in turn reduces the chances of supply driven credit crunch effects.Macharia, (2013) studied the effects of global financial crisis on the financialperformance of commercial banks offering mortgage finance in Kenya.

The study found a negative relationship between inflation, interestrates as a result of global financial crisis and financial performance ofcommercial banks offering mortgage finance in Kenya. A unit increase in


inflation and interest rates led to a 0.543 and 0.425 decrease respectively inthe scores of financial performance of commercial banks offering mortgagefinance in Kenya. The study further found that exchange rates, as a resultof global financial crisis, had a positive effect on financial performance ofcommercial banks offering mortgage finance in Kenya. A unit increase inforeign exchange rates led to a 0.652 increase in the scores of financialperformance of commercial banks offering mortgage finance in Kenya.

Otuori (2013) investigated the determinant factors of exchange ratesand their effects on the performance of commercial banks in Kenya. Thestudy found that exports and imports Interest rates, inflation and exchangerates were all highly correlated. By manipulating interest rates, centralbanks could exert influence over both inflation and exchange rates, andchanging interest rates impact inflation and currency values. Higher interestrates offered lenders in an economy a higher return relative to othercountries which attract foreign capital and cause the exchange rate to rise.

Mboka (2013) studied the relationship between macroeconomicvariables on nonperforming loans of commercial banks in Kenya. Data wasanalyzed by applying both descriptive and inferential statistics for a 10year period (2003 to 2012). The study found a strong correlation betweeninflation and gross domestic product and current account deficit. GDP alsocorrelated strongly with inflation and Money supply. A significant andpositive correlation was also found between nonperforming loans and GDPgrowth rate, exchange rate volatility, and banking sector developmentindex. Kiruri and Olkalou (2013) studied the ownership structure onsamples of 43 banks profitability from 2007 to 2011 in Kenya. The simplelinear regression shows that ownership concentration and state ownershiphad negative and significant effects on bank profitability while foreignownership and domestic ownership had positive and significant effectson bank profitability in Kenya.

3. RESEARCH METHODOLOGY

The section focuses the research design and methodology that was appliedin conducting this study. It describes the research design, population ofthe study, sample size, sample frame, data collection methods and dataanalysis and presentation of the research findings.

3.1. Research Design

This research employed a quantitative research design. The quantitativeresearch design method helped in gathering information about the existingstatus of the phenomena in order to describe what exists in respect to


variable. This method is used because it addresses the objective of the studyin investigating the relationship between the variables of the study (Kothari,2008). The design takes into consideration aspects like the size of the samplein relation to the target population, the variables under the study, theapproaches to the research, and the methods employed in data collection.

Multiregression method was used to determine the relationshipbetween interest rates and profitability of commercial banks. The studyused time series empirical data on the Variables to examine the relationshipbetween interest rate by establishing correlation coefficients between thevariables and profitability of commercial banks.

3.2. Population of the Study

The target population for this study was 4 commercial banks in Rwanda(Bank of Kigali, CoGEBANK Bank, I&M Banks and ECO BANK). Thecommercial banks were selected because the accessibility of their financialdata by the public. All the 4 banks constituted the study sample. A censusdesign was applied where all the 4 commercial banks were studied. Acensus is a collection of information from all units in the population or acomplete enumeration of the population. A census design is used wherethe population is small and manageable (Mugenda & Mugenda, 2003).

3.3. Data Collection

The data required for the study were obtained from secondary sourcesthat were used to investigate the relationship between dependent andindependent variables. In the study, 3 years data (2016 to 2018) werecollected. The collected data related to dependent variable which is thecommercial bank’s profitability as measured by return on assets, return onequity, nonperforming loan, net profit and the independent variableswhich was Interest rate. Documentation techniques were used toinvestigate, categorize and collect physical resources, mostly commonlywritten documents.

3.4. Data processing and Analysis

The secondary data that was collected from the reports were summarizedusing excel software. Data was imported in the Statistical Package for SocialSciences (SPSS) version 21 from where analysis was made. Inferentialstatistics was conducted using a multiple – regression which was also usedto determine the relationship between macroeconomic variables andprofitability of commercial banks. The variable was considered significantwhere the Pvalue is less than 5%


3.5. Operational definition of variables

The study analyzed two study variables which are macroeconomic variablesand profitability of commercial banks. The macroeconomic variables weremeasured using Gross domestic product (GDP), Inflation (IF), Interest rate(IT), and unemployment rate (UP). On the other hand, the profitability ofthe bank was measured using the return on asset (ROA) and the return onEquity (ROE).

Profitability (P) = F(GDP, IF, IT, UP) + �

P = �0 + �

1GDP + �

2IF + �

3IT + �

4UP +��

Parameters �0 ……….�

5 was estimated using a least square method

which appropriate when one has used a multiple regression method. Thecovariance of the error terms á was assumed to be zero.

4. RESULTS AND DISCUSSION

This section presents the results collected from the survey of the selectedcommercial banks and provides a discussion to the findings.

4.1. Multicollinearity test

In the following lines, the presence of linear relationship of all the predictorsused in the model and their coefficient estimates is examined.

Table 1: Variance Inflation Factor values for each predictor

Coefficientsa

Model Collinearity Statistics

Tolerance VIF

1 GDP .953 1.049

Interest rate .950 1.053

Inflation rate .908 1.101

Unemployment rate 0.902 1.113

Survey data 2019

Variance Inflation Factor (VIF) was analyzed to test for the existenceof multicollinearity. This phenomenon occurs when “two or moreindependent variables (or combination of independent variables) in amultiple linear regression are highly correlated with each other” (Kothari2000), meaning that one can be linearly predicted from the others with asubstantial degree of accuracy. This leads to problems with understandingwhich independent variable contributes to the variance explained in the


dependent variable, as well as technical issues in calculating a multipleregression model. The VIF for each predictor is quite low compared to themaximum acceptable value of 5, hence absence of colinearity among them.

4.2. Testing violation of the normality assumption of the error term inthe model

In the line that follows, the assumption on the error terms in model isexamined. These have been assumed to be normally distributed withconstant variance. Reading from the two graphs below reveals that theseare close to being normally distributed. In fact, the right hand side graphreveals that the standard deviation of the residual is small, since theirdensity tends to conglomerate around the center or the mean.

Table 2: Tests of Normality

KolmogorovSmirnova ShapiroWilk

Statistic df Sig. Statistic df Sig.

GDP .210 11 .189 .917 11 .291

Inflation .171 11 .200* .884 11 .118

Interest Rate .231 11 .104 .815 11 .152

Unemployment .180 11 .200* .854 11 .428

a. Lilliefors Significance Correction

Table 2 tests on whether the study variables are normally distributed. Anull hypothesis which state that at 5% level of significant the data are notsignificantly different from the normal distribution. Using the Kolmogorov– Smirnov and Shapiro – Wilk significant test, the results from the surveyshows that all the tested variables are not statistically significant as their Pvalue is more than 5%. The researchers therefore accepted the null hypothesis.This, therefore, means that the tested variables are normally distributed.

4.3. Effect of Macroeconomic Variable on Return on Equity

Table 2: Model summary on the effect of macroeconomic variable on ROE

Model R R Square Adjusted R Square Std. Error of the Estimate

1 .889a .790 .788 7.46074

Source: Survey data 2019

Table 2 shows the correlation coefficient and the coefficient ofdetermination. From the table the correlation coefficient is very high (0.889).


This means that macroeconomic variable and profitability of commercialbanks are highly positively correlated. This means when themacroeconomic variable improves, the profitability of the commercial bankswill improve. The coefficient of determination is 0.790 which implies that79% of the variation in ROE is explained by the macroeconomic variables.The study revealed that it is only 21% of the variation in the ROE that iscaused by other variables.

Table 3: Significance of the model to determine profitability: ANOVAa

Model Sum of Squares df Mean Square F Sig.

1 Regression 5.8 x 107 3 1.92 x 107 311.903 .000b

Residual 1.5 x106 248 6.17 x 104

Total 7.3 x 107 251


ANOVA was conducted to assess whether the data are consistent withthe model assumptions or not. This was done on the basis of the nullhypothesis stated that “there is no difference between the model withoutindependent variables and the model with independent variables”. FromTable 3, the PValue (0.000) is less than the significance level (0.05), thusthere is enough evidence for rejecting the null hypothesis. This, therefore,implies that the coefficients used in the model are not zero.

Table 4: Coefficients of determinants of profit shifting

Model Unstandardized Standardized t Sig.Coefficients Coefficients

B Std. Error Beta

1 (Constant) 2004136077.975 643387036.508 3.115 .002

GDP .995 .037 .818 26.834 .000

Interest rate 1.271 .200 .190 6.373 .000

Inflation rate .567 .062 .096 9.226 .000

Unemployment 0.455 .004 1.56 10.22 .000

Source: survey data 2019

Table 4 shows the significance of the independent variables inpredicting the relationship. This was done on the basis of the null hypothesisthat “the independent variable has no effect on the return on equity. Thetable above shows that the PValues for GDP, interest rate, inflation rateand employment rate (0.000) which is smaller than the significance level(0.05). Thus there is enough evidence to reject the null hypothesis for theseindependent variables. We can therefore conclude that macroeconomic


variables have significant effects on return equity. GDP has positiverelationship with the ROE meaning that their increase leads to the increasein the profitability. However, interest rate, inflation rate and unemploymentrate has a negative relationship with the ROE. This means that an increasein these variables leads to a decrease in the profitability. The study findingsagree with Buyinza (2010) and Osamwanji and Chijuka (2014) who indicatesa positive correlation between GDP and ROE and a negative relationshipbetween inflation, interest rate and the ROE.

ROE = 2004136077 + 0.995(GDP) 1.271(IT) – 0.567(IF) – 0.455(UP) (3)

From the regression equation (3), the following conclusion can bedrawn: A unit change in GDP increases ROE by 0.995 units keeping allother variables constant.

4.4. Effect of Macroeconomic Variables on Return on Assets (ROA)

Table 5: Model summary of effect of macroeconomic variables on ROA

Model Summary

Model R R Square Adjusted Std. Error of the EstimateR Square

1 .874a .764 .681 6.773


Table 5 shows the correlation coefficient and the coefficient ofdetermination. From the table the correlation coefficient is 0.874. This meansthat macroeconomic variables and ROA are negatively correlated. Thecoefficient of determination is 00.764 which implies that 76.4% of the variationof ROA is determined by the variations in macroeconomic variables. Thismeans that GDP, interest rate, inflation rate and unemployment rate explains76.4% of the variation in the profitability of commercial banks. This impliesthat 23.6% of the variation in the ROA is unaccounted for by the model.

Table 6: Significance of the model: ANOVAa test

Model Sum of Squares df Mean Square F Sig.

1 Regression 5.58 x 108 3 1.86 x 108 40.575 .000b

Residual 1.11 x 108 248 4.58 x 108

Total 1.69 x 108 251

ANOVA was conducted to assess whether the data are consistent withthe model assumptions or not. This was done on the basis of the nullhypothesis stated that “there is no difference between the model without


independent variables and the model with independent variables”. Table6 shows that PValue (0.000) is less than the significance level (0.05), thusthere is enough evidence for rejecting the null hypothesis. We can thereforeconclude that there is a significant statistical difference between the modelwithout independent variables and the model with independent variableshence the model fits the data.

Table 7: Coefficients of the determinants of profit shifting

Coefficientsa

Model Unstandardized StandardizedCoefficients Coefficients

B Std. Error Beta t Sig.

1 (Constant) 9.999 5.5556 1.800 .073

Unemployment 136 .032 233 4.269 .000

Inflation rate 1.866 .172 .579 10.856 .000

Interest rate .54 .0398 .007 13.568 .000

GDP 1.243 0.234 2.123 6.871 0.021

Source: Survey Data 2019

Table 7 shows the significance of the independent variables. This wasdone on the basis of the null hypothesis that “the independent variableshave no effect on ROA The table shows that the PValues for unemploymentrate, inflation rate and interest rate is (0.000) whereas GDP has a Pvalue of0.021 which is less than the significance level (0.05), thus there is enoughevidence to reject the null hypothesis for these independent variables. Wecan therefore conclude that GDP, interest rate, inflation rate andunemployment rate have significant effects on ROA. The study findingsagrees with Gilchris, (2013) report which identified a relationship betweenmacroeconomic variables and profitability of commercial banks

ROA = 9.98 + 1.243(GDP)– 1.36(UP) – 1.866(IF) 0.54(IT) (4)

From the regression equation (4), we can say that: A unit change inGDP, the ROA will increase by 1.243. More still, a unit change inunemployment, inflation rate and interest rate the ROA will decrease by1.36, 1.866 and 0.54 respectively.

5. CONCLUSION AND POLICY RECOMMENDATIONS

5.1. Conclusion

The results from the survey indicated a strong relationship between themacroeconomic variables and profitability of commercial banks. The results


revealed that GDP, Unemployment rate, inflation rate and interest ratehighly explains the variability in the profitability of commercial banks.The results revealed that there is a positive correlation between GDP andprofitability of commercial banks and a negative correlation betweeninflation, interest rate, unemployment and profitability of the commercialbanks. Improving the economic activities is very fundamental to theprofitability of commercial banks. However, care should be taken whileimproving the economic activities on controlling the level of inflation,interest rate and unemployment as these factors negatively affect theprofitability of the commercial banks.

5.2. Policy Recommendation

The research has revealed that there are some macroeconomic variablethat affect financial performance of bank negatively (inflation rate, andinterest rate), the increase in one of those variables leads into poorperformance of commercial banks and inflation contribute to an increasein nonperforming loan (customer loan default).

The research has revealed that there is a strong negative correlationbetween lending rate and financial performance, this makes sense that theincrease in the cost of borrowing it doesn’t yield return on banks insteadthe increase in lending rate it create a heavy burden to borrower whichfinally result into a total loan default. Commercial banks and central bankof Rwanda are advised to a reasonable lending rate that is suitable forborrowers/investors.

References

Bikker and Hu, (2002). Determinants of bank profitability in a developing economy:empirical evidence from the Philippines.

Brueggeman & Fisher, (2011). Real Estate Finance & Investments (The McGrawHill/IrwinSeries in Finance, Insurance, and Real Estate) 14th Edition.

Brueggeman & Fisher, (2011). Real estate finance and investments

Dukundane and Rukera, (2016). The impact of credit risk management on financialperformance of commercial banks in Rwanda.

Eakins & Mishkin, (2012). Financial Markets and Institutions 7th (Seventh) Edition byEakins,November Vol., No. 178.

Fama, (1970). Efficient Capital Markets: A Review of Theory and Empirical Work 1969(May, 1970), pp. 383417

Friedman, (1963). Development of standard rating scales for mechanical parameters oftexture and correlation between the objective and the sensory methods of textureevaluation.

Gilchris (2013). uncertainty, financial frictions, and investment dynamics


Gul, Irshad, and Zaman (2011). The impact of macroeconomic variables on the profitabilityof listed commercial banks in Pakistan

Jensen & Meckling, (1976). Theory of the firm: Managerial behavior, agency costs andownership structure

Keynes (2006). A handbook of alternative monetary economic

Khrawish, and Siam, (2011). Samples of three Jordan Islamic banks profitability InternationalJournal of Science and Research (IJSR), India, Volume 2 Issue 3, March 2013, Online ISSN:23197064;

Lucas and Anne, (2010). The effect of macroeconomic developments on performance.

Macharia, (2013). Effects of global financial crisis on the financial performance of commercialbanks.

Markowitz, (1952). Efficient Diversification of Investments

Markowitz, (1952). The early history of portfolio theory: 1600–1960, The Journal of Finance,1952 Wiley Online Library.

Nkurikiye and Uwizeyimana (2017). Effects of Gdp, interest rate, and inflation on privateinvestment in Rwanda.

Osamwonji & Chijuka, (2014). The effect of selected macroeconomic variables on financialperformance of banking sector in Kenya, Volume 17, Number 6.

Peng and Shu, (2005). Macroeconomic conditions and banking performance in Hong KongSAR: a panel data study

Rosli A.R., (2017). The Relationship Between Hong Leong Bank’s Performance with Leverageand Inflation, MPRA Paper No. 78500, posted 17 April 2017, https://mpra.ub.unimuenchen.de/78500/;

Saidu and Tumin (2011). Impacts of internal and external factors on profitability of banksin nigeria vol. 1, issue 9, 2014.

San & Heng, (2013). Factors affecting the profitability of Malaysian commercial banks.

Schumpeter (1939). Business cycles

Simiyu & Ngile, (2015). Effect of macroeconomic variables on profitability of commercialbanks listed in the Nairobi securities exchange.

Simiyu and Ngile, (2015). The bank sector performance and macroeconomic environment:empirical evidence in Togo.

Tee1 O.C., (2011). The international banking crisis: effects and some key lessons, (pp 359363.) Bank for International Settlements, vol. 54.

Were, Kamau, Sichei, Kiptui, 2013). Assessing the effectiveness of monetary policy in Kenya:Evidence from a macroeconomic model.

Young B., (2014). Financial crisis: causes, policy responses, future challenges, Outcomes ofEUfunded research, European Union.


Daniel Twesige and Faustin Gasheja. Effect of Macroeconomic Variables on Profitabilityof Selected Commercial Banks in Rwanda. Journal of Asian Economics, Accountingand Finance, Vol. 1, No. 2, 2020, pp. 181198

ON THE DETERMINANTS OF UNEMPLOYMENTRATE IN NIGERIA:Evidence from Fully Modified OLS andError Correction Model

1Adenomon, M. O.; 1Okoro-Ugochukwu, N. A. and 2Adenomon, C. A.1Department of Statistics, Nasarawa State University, Keffi, Nasarawa State.1NSUK-LISA Stat Lab, Nasarawa State University, Keffi, Nigeria.2Department of Administration, Federal Medical Centre, Bida, Niger StateE-mail: [email protected]

Received: 24 August 2020; Revised: 29 September 2020; Accepted: 9 October 2020; Online: 29 December 2020

Abstract: This study employed the Fully Modified Ordinary Least Squares (FMOLS) and the ErrorCorrection Model (ECM) to investigate the long-run and short-run determinants of unemploymentrate in Nigeria. To achieve this annual data on unemployment rate, inflation rate, interest rate, exchangerate and population growth from 1981 to 2016 was collected from Central Bank Statistical Bulletinsand the World Bank website. The ADF test revealed that the macroeconomic variables are stationaryat first difference while the Cointegration test revealed that the variables are cointegrated. Usingunemployment rate as dependent variable, the FMOLS model revealed that exchange rate andpopulation growth are positively significantly related to unemployment rate, interest rate and inflationrate were negatively related to unemployment rate but only interest rate was significant. The shortrun relationship revealed that the coefficient of the ecm(-1) is negative and statistically significant at5% level indicating that the system corrects its previous period disequilibrium at the speed of 48.93%yearly. This study concludes that high exchange rate and population growth can lead to increase inunemployment rate in Nigeria while the government should develop the industrial sector and non-oil sector in order to generate employment and boost export in Nigeria.

Keywords: Determinants, Unemployment Rate, OLS, FMOLS, ECM.

1. INTRODUCTION

Unemployment rate can be seen as a measure of the occurrence or frequencyof unemployment and which is usually calculated as a percentage simplyby dividing the number of unemployed individuals by all individualscurrently in the labour force. During periods of recession, an economyusually experiences a relatively high unemployment rate. The NationalBureau of Statistics of Nigeria stated that Nigerian youths are among themost important resources the country need is to be able to achieveprosperity and progress (Maigwa and Kipesha, 2013). In addition, thepopulation of every economy constitutes of two categories, the economicallyactive and the economically inactive (Muhdin, 2016).



In general, unemployment among young people has become the mainchallenge which all the nations in the world are facing presently. Theresultant effects of unemployment are extensive crises in psychological,social and economic perspectives, some of them are: increasing crime ratesand violence in the society, reliance on family, low selfconfidence by thevictim, poor social adaptation, unhappiness and loss of confidence(Kabaklarli & Bulus, 2011). Nasir et al (2009) in the same manner showedthat unemployment affects the socioeconomic status of the family, andalso leads to poor emotional health, dependency and surges up themagnitude of corruption, prostitution, drug addiction, kidnapping, ritualkillings, suicide and other crimes in a society.

This study examined the determinants of unemployment rate inNigeria using Fully Modified Ordinary Least Squares (FMOLS) and ErrorCorrection Model (ECM).

2. EMPIRICAL LITERATURE REVIEW

First we defined the following:

Unemployment is defined by the Bureau of Labour Statistics as peoplewho do not have a job, have actively looked for work in the past four weeks,and are currently available for work. There are three major types ofunemployment – Structural Unemployment, Critical Unemployment andFrictional unemployment.

Inflation is a sustained increase in the general price level of goods andservices in an economy over a period of time. When the price level rises,each unit of currency buys fewer goods and services. Consequently,inflation reflects a reduction in the purchasing power per unit of money –a loss of real value in the medium of exchange and unit of account withinthe economy.

Exchange rate is the rate of transformation of one currency to another.Nzotta (2004) defines exchange rates as the price of one currency in termsof another.

Previous literatures are available on the subject matter, highlightingvarious causes and consequences regarding increasing rate ofunemployment using statistical methods.

Pallis (2006) focused his study on the relationship between inflationand unemployment in new European Union member states. He obtainedthe data used in the analysis the annual data that covered the period from1994 to 2005, which was taken from European commission 2004 referred to

On the Determinants of Unemployment Rate in Nigeria 201

the new ten (10) European Union (EU) member states. The three variablesused are “the price deflator of GDP at market prices, the nominalcompensation per employee and then the total employment rate (%). Inestimating the variables used in the study, Nonlinear least square methodof estimations and Eviews techniques were used. The findings provedand concluded that the application of common policies across economymay be questionable because of the different effects of these policies onunemployment and inflation.

Ozturk & Akhtar (2009) studied and analysed a comprehensiveapproach to unemployment by using VAR of “Variance Decompositionand Impulse response function analysis”. They were intereste d in studyinginterrelationship among Foreign Direct investment (FDI), Export, GrossDomestic product (GDP) and unemployment in Turkey for the period of2000 to 2007. They found only two counteracting vectors in the system,showing long run relationship. They now concluded that foreign directinvestment (FDI) did not lead to reduce unemployment in Turkey. GDP ispositively affected by variations in exports but is also insignificant. So theydid not found any evidence of export led growth in Turkey. Again,Variations in GDP was not attached with reduction of unemployment.

Eita & Johnson (2010) studied the causes of unemployment in Namibiafor the period 1971 to 2007. The results revealed that in Namibia, there exita negative relationship between unemployment and inflation. It is notedthat if wages increases, Unemployment responds positively if actual outputis below potential output. An increase in investment results to decrease inunemployment significantly. The results provide evidence that the Phillipscurve holds for Namibia and unemployment can be reduced or decreasedby increasing aggregate demand.

ElAgrody et al. (2010) examined and evaluates the economic study ofunemployment and its impact on the GDP for Egypt. Data was collectedfrom year 1994 to 2004 in Egypt. Simple and multiple linear regressionanalysis were applied. Variables used in the study were privatization,population, consumption expenditure, interest rates, exchange rates,technology, agricultural domestic product, real wage rates and agriculturalinvestment. The results and the findings showed that there is a significantpositive impact of national unemployment, national investment, exchangerate and average per capita share of GDP on the volume of GDP. The resultsin addition also highlighted privatization and increasing population as themain reasons of increasing unemployment. They in turn recommendedthat privatization policies need to be revised and to reduce interest rates inorder to lowering the agricultural unemployment.


Afzal & Awais (2012) investigates and analyzes the relationshipbetween InflationUnemployment Trade Off: Evidence taken from Pakistan.Using the method of ordinary least squares (OLS) and also equation by themethod of nonlinear least squares (NLS). The Data on consumer priceindex and unemployment were collected from Government of Pakistan(GOP), economic survey (various issues) for the period 1973 2010. Then,the empirical results obtained for the first three periods (19742010, 197482, and 197492) and then the last period 20002010 show that the Phillipscurve holds in Pakistan because the unemployment coefficient is negativeand also very significant. For other periods (19812000 and 19812010)though there is negative relationship between inflation and unemployment,the unemployment coefficients are not very significant.

Aminu & Anono (2012) studied the relationship betweenunemployment and inflation in the Nigerian economy. Data was used from1977 and 2009 and was analyzed with the application of Augmented DickeyFuller techniques in order to examine the unit root property of the series,however, Granger causality test was conducted to determine if there iscausality between unemployment and inflation, in addition cointegrationtest was conducted through the application of Johansen cointegrationtechnique to examine the longrun relationship between the twophenomenon, they later used ARCH and GARCH technique to conductand examine if there is existence of volatility in the series. The results ofthe analysis indicated that inflation impacted negatively on unemployment.The causality test however revealed that there is no causality betweenunemployment and inflation in Nigeria during the specified period of studybut it is noted that a longrun relationship exists between them as confirmedby the cointegration test. The result of ARCH and GARCH test then revealedthat the time series data for the period under review exhibits a high volatilityclustering.

Maqbool et.al. (2013) upon examining the empirical relationship amongthe unemployment, population, foreign direct investment, gross domesticproduct, inflation, and external debt in Pakistan. The use of AutoregressiveDistributed Lag (ARDL) approach was applied to test determinants ofunemployment. The results however revealed that both gross domesticproduct, population, foreign direct investment and inflation are significantdeterminants of unemployment in Pakistan in both shortrun and longrun.

Baah Boateng (2013) studied the determinants of Unemployment ratein Ghana for a period of 1991 to 2005 using binary Regression Estimate toanalyse its data. A crosssectional estimation of a probit regression model


also shows that there is a strong effect of demand factors on unemployment,and this also shows a weak employment generating impact of economicgrowth. Empirical analysis which was employed also conûrms very highvulnerability of youth and urban dwellers to unemployment wherebyeducation and gender explained unemployment in some instances.Reservation wage is also observed to have an increasing effect ofunemployment.

Kemi & Ayo (2014) investigates the issue of Unemployment andEconomic Growth in Nigeria. The study actually validated Okun’s law inNigeria. In order to examine the relationship between unemployment rateand economic growth, Error Correction Model (ECM) and Johasencointegration test were employed to determine both the shortrun and longrun relationships among the macro variables employed in the study. TheEmpirical findings however revealed that there exist both the short andthe long run relationship between unemployment rate and output inNigeria.

Cheema & Atta (2014) analyses Economic Determinants ofUnemployment in Pakistan: Cointegration Analysis. This study revealsthe determinants of unemployment by applying the Auto RegressiveDistribution Lag Model (ARDL) bound approach using the time series datafrom the period of 1973 to 2010. The outcomes indicate that unemploymenthas statistically significant and positive relationships with output gap,Productivity and Economic Uncertainty while it has statistically significantbut negative relationships with Gross Fixed Investment and Openness ofTrade.

Muhdin (2016) studied the main issues underlying unemployment inEthopia. The data used was collected from Central Statistical Agency (CSA)in 2015 and a total of 16984 were considered for the analysis. The use ofdescriptive statistic like percentages, mean value and cross tabulation inthe study shows that youth unemployment is highly related with regionallocation, sex, marital status and education. Using descriptive and crosstabulation analysis, the study shows that youth unemployment is highlyrelated with regional location, sex, marital status and education. Out ofthe total responses obtained in the survey, 53.5 percent are female. Theaverage year of the sample under review was 23.3 years. Moderately, largerproportion, 59.9 percent, of the youth were never married, however about33.9 percent of them were married, 0.9 percent of live together and theremaining 5.3 percent were noticed to be divorced, separated and widowed.Also, on the average, household size is seen to be 2.2 family members. Thesurvey shows that the literacy level is very high for Ethiopian Youths.


3. MODEL SPECIFICATION

Fully Modified OLS

The FMOLS is an optimal singleequation method based on the use of OLSwith semiparametric correction for serial correlation and endogeneity(Phillips & Loretan, 1991).

Suppose yt be an nvector I(1) process and u

t be an nvector stationary

time series. The partition of these vectors can be seen as

1 , ,1

2

1

1

2

1 ��

��

��

�

��

�� mn

u

uu

y

yy

mt

tt

mt

tt (1)

Assuming that the generating mechanism for ty is the cointegrated

system given as

1 2 1t t ty y u (2)

2 2t ty u (3)

The basic idea in this procedure is to modified the OLS estimator

� � 121

22* yYYY �� (4)

(Phillips & Loretan, 1991).

But )uE(u kk

10

2021 ��

�

�� . If 21� is consistent estimator of �21

, then wee

have a modified OLS estimator

� � )ˆ(** 21121

22 �� TyYYY ��

Then the Fully Modified OLS (FMOLS) estimator employs both theserial correction and endogeneity corrections and is given as

1

2 2 2 1ˆ( )Y Y Y y T (5)

where

11 1 21 22 2

ˆˆt t ty y y (6)

122 21

1ˆ ˆ

ˆ (7)


20 1

0

ˆ ˆwhere is a consistent estimate of and is consistent for k

k

Δ E(u u )

Fully modified test statistics that are based on �� may be constructedin the usual way. Thus, for tratios we defined as

( ) /i i i it s (8)

1211 2 2 2

ˆwhere i .ii

(s ) σ Y Y (9)

111.2 11 21 22 21

ˆˆ ˆ ˆ ˆhere (10)

And is based on components of � which is again an estimate of the

longrun covariance matrix ��(Phillips & Hansen (1990); Phillips & Loretan(1991)). Application of this method can be found in Kuhe (2016) andAdenomon et al. (2018).

Augmented DickeyFuller (ADF) Unit Root test

Engle and Granger, (1987) considered seven test statistics in a simulationstudy to test cointegration. Engle and Granger concluded that theAugmented Dickey Fuller test was recommended and can be used as arough guide in applied work

To distinguish a unit root, we can run the regression

11

ttjt

k

jjot uYtYbbY ��

��

The model in (1) may be run without t if a time trend is not necessary.This technique was applied in Ajayi and Mougoue (1996). If there is unitroot, differencing Y should result in a whitenoise series (no correlationwith Y

t1).

The Augmented DickeyFuller (ADF) test of the null hypothesis of no

unit root test is of the form Ho: 0�� (if there is trend we use Ftest)

and Ho: � = 0 (if there is no trend we use ttest). If the null hypothesis is

accepted, we assume that there is a unit root and difference the data beforerunning a regression. If the null hypothesis is rejected, the data arestationary and can be used without differencing (Salvatore & Reagle, 2002).

Johansen and Juselius Cointegration Test

The most popular test for cointegration testing is the Johansen and Juseliuscointegration test (i.e Maximum Eigenvalue test and the trace test)


(Johansen & Juselius, 1990). The maximum eigenvalue test and the tracetest are used as procedures to determine the number of cointegrationvectors.

The maximum eigenvalue statistic test the null hypothesis of rcointegrating relations against the alternative of r+1 cointegrating relationsfor r = 0, 1, 2, . . ., n–1. This test statistic is computed as

)ˆ1ln()1,( 1max �� rTrr ��where � is the computed maximum eigenvalues and T is the sample size.

The main difference between the maximum eigenvalue test and thetrace test is that the trace test is a joint test, whereas the maximum eigenvalue test conducts separate test on the individual eigenvalues.

Trace statistic examines the null hypothesis of r cointegrating relationsagainst the alternative of n cointegrating relations, where n is the numberof variable in the system for

r = 0, 1, 2, . . ., n–1.

It is computed according to the following formula

��

��n

riitrace Tr

1

)ˆ1ln()( ��

The results of trace test should be chosen where trace and maximumeigenvalue statistic may yield different results in some case (Habte, 2014).

The Error Correction Model

The cointegrating regression considers only the longrun property of agiven model, and does not deal with the shortrun dynamics explicitly.Clearly, a good time series modelling should describe both shortrundynamics and the longrun equilibrium simultaneously. For this purposeerror correction model (ECM) was developed. Although ECM has beenpopularized after Engle and Granger, it has a long tradition in time serieseconometrics dating back to Sargan (1964) as documented by Hendry (2003).To start, we define the error correction term by

�t = y

t – �x

t

Where � is a cointegration. In fact �t is the error from a regression of y

t

on xt. Then an ECM is simply defined as

tttt uxy �� 1


where ut is iid. the ECM equation simply says that ty� can be explained by

the �t–1

and �xt. We can notice that �

t–1 can be thought of as an equilibrium

error (or disequilibrium term) occurred in the previous period (Salmon,1982).

4. MATERIALS AND METHODS

The data used for this analysis is a secondary data. The data was collectedfrom Central Bank of Nigeria Statistical bulletin from 1981 to 2016 andWorld Bank web site (www.worldbank.org). the data sets was transform usingnatural log to ensure normality, stability and to reduce skewness andkurtosis.

5. DATA ANALYSIS AND DISCUSSION OF RESULTS

The statistical analysis was carried using EViews 7.2 statistical software.The data used in this analysis are presented in Table 1 presented at theappendix. While the natural logarithm transform of the economics variablesare presented in Table 2 presented at the appendix. The data used in thiswork can be graphically represented as in Fig. 2 below

Figure 2: A Graph of some Macroeconomic variables in Nigeria from 1981 to 2016


In fig. 2 above, It is observed that Exchange Rate exhibit a positivegrowth; Under Inflation rate, there is fluctuation; Observe that PopulationGrowth is stable, Interest rate and Unemployment also fluctuates.

Table 3: Descriptive statistics of the dependent and independent variables.

LNEX LNINFL LNINTRT LNPOP LNUNEMP

Mean 3.293656 2.674964 2.828446 0.947595 2.103168

Median 3.811330 2.446680 2.864765 0.948614 1.901105

Maximum 5.535320 4.289090 3.394510 0.998815 3.401200

Minimum 0.494300 1.667710 2.047690 0.911553 0.587790

Std. Dev. 1.947536 0.762144 0.293679 0.026654 0.843132

Skewness 0.735632 0.737983 0.757065 0.098451 0.065551

Kurtosis 2.202316 2.358079 3.611314 1.669811 1.707701

JarqueBera 4.201373 3.885804 3.999444 2.712260 2.530838

Probability 0.122372 0.143288 0.135373 0.257656 0.282121

Sum 118.5716 96.29872 101.8240 34.11341 75.71405

Sum Sq. Dev. 132.7513 20.33022 3.018649 0.024865 24.88051

Observations 36 36 36 36 36

Where: LNEX= Log of Exchange Rate; LNINFL= Log of Inflation rate;LNINTRT = Log of Interest Rate; LNPOP = Log of Population Growth;LNUNEMP= Log of Unemployment

In table 1 above, the mean values for log of Exchange Rate, Log ofInflation rate, Log of Interest Rate, Log of Population Growth and Log ofUnemployment are 3.293656, 2.674964, 2.828446, 0.947595 and 2.103168respectively. JarqueBera test revealed that the variables are normallydistributed (since all Pvalues= 0.1223, 0.1433, 0.1354, 0.2577, 0.2821 Ã 0.05)which means that we accept H

0 and conclude that all the variables are

Normally Distributed and the implication of this is that we can use theproposed Model that is Fully Modified Ordinary Least Square and ErrorCorrection models.

Table 4: Unit Root Test

The unit root test for all the variables can be represented in the tabular form shown below:

VARIABLES ADF Pvalues ORDER REMARK

D(LNEX) 5.020825 0.0002 I (1) Stationary

D(LNINFL) 7.240855 0,0000 I (1) Stationary

D(LNINTRT) 8.707736 0.0000 I (1) Stationary

D(LNPOP) 7.056791 0.0000 I (1) Stationary

D(LNUNEMP) 6.641959 0.0000 I (1) Stationary


The table 4 above all the variables are of order 1 that is all the variablesare stationary at first difference. Details are in table in the appendix.

Table 5: Cointegration test

Date: 05/18/18 Time: 01:58

Sample (adjusted): 1983 2016

Included observations: 34 after adjustments

Trend assumption: Linear deterministic trend

Series: LNEX LNINFL LNINTRT LNPOP LNUNEMP

Lags interval (in first differences): 1 to 1

Unrestricted Cointegration Rank Test (Trace)

Hypothesized Trace 0.05No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None * 0.851668 139.1505 69.81889 0.0000

At most 1 * 0.744829 74.26825 47.85613 0.0000

At most 2 0.409899 27.83033 29.79707 0.0829

At most 3 0.165586 9.896657 15.49471 0.2888

At most 4 0.104213 3.741798 3.841466 0.0531

Trace test indicates 2 cointegrating eqn(s) at the 0.05 level

* denotes rejection of the hypothesis at the 0.05 level

Unrestricted Cointegration Rank Test (Maximum Eigenvalue)

Hypothesized MaxEigen 0.05No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None * 0.851668 64.88226 33.87687 0.0000

At most 1 * 0.744829 46.43792 27.58434 0.0001

At most 2 0.409899 17.93368 21.13162 0.1323

At most 3 0.165586 6.154859 14.26460 0.5934

At most 4 0.104213 3.741798 3.841466 0.0531

Maxeigenvalue test indicates 2 cointegrating eqn(s) at the 0.05 level

* denotes rejection of the hypothesis at the 0.05 level

**MacKinnonHaugMichelis (1999) pvalues

Since Pvalue = 0.0829 > 0.05 which appeared under Atmost 2, wetherefore state that the Trace Test indicates 2 cointegration eqn(s) at 0.05level which implies that the variables are cointegrated that is there exist along run relationship among the macro variables.

Furthermore the maximum eigen Test also indicated 2 cointegrationeqn(s) at the 0.05 level depict the long run relationship among theUnemployment rate and all other independent variables.


Table 6: Fully Modified Ordinary Least Squares (FMOLS)

Dependent Variable: LNUNEMP

Method: Fully Modified Least Squares (FMOLS)

Date: 05/18/18 Time: 01:59

Sample (adjusted): 1982 2016


Cointegrating equation deterministics: C

Longrun covariance estimate (Bartlett kernel, NeweyWest fixed bandwidth = 4.0000)


LNEX 0.316484 0.067104 4.716341 0.0001

LNINFL 0.159174 0.147440 1.079581 0.2889

LNINTRT 1.277057 0.480040 2.660314 0.0124

LNPOP 14.49323 4.114303 3.522646 0.0014

C 8.590843 4.343328 1.977940 0.0572



S.E. of regression 0.459763 Sum squared resid 6.341468

DurbinWatson stat 1.191382 Longrun variance 0.296870

Since there exist long run relationship among the variables, then theFully modified ordinary least square model becomes appropriate to beused to obtain the long run estimates. Then the FMOLS obtained in Table 6and the estimated model is given below:

tLNPOP + ε . LNINTRT + .LNINFL - .LNEX - . + .LNUEMP = - 93241427711159203165059088

The model above reveal that Unemployment becomes negative if thereis no influence of all other variables. The model also show that Exchangerate and population growth are positively related to unemployment rate

Table 7: Cointegration Test Hansen Parameter Instability (FMOLS)

Cointegration Test Hansen Parameter Instability

Date: 05/18/18 Time: 02:00

Equation: UNTITLED

Series: LNUNEMP LNEX LNINFL LNINTRT LNPOP

Null hypothesis: Series are cointegrated

Cointegrating equation deterministics: C

Stochastic Deterministic ExcludedLc statistic Trends (m) Trends (k) Trends (p2) Prob.*

0.288646 4 0 0 > 0.2

*Hansen (1992b) Lc(m2=4, k=0) pvalues, where m2=mp2 is the number of stochastic trends inthe asymptotic distribution.


which implies that every unit increase in exchange rate and populationgrowth will result to an increase in unemployment rate. While inflationrate and interest rate are negatively related to unemployment rate whichimplies that every unit increase in Inflation rate and Interest rate will resultto a decrease on Unemployment rate. The model further revealed thatExchange rate, interest rate and population growth are all significant sinceall the Pvalues are less than 0.05 level of significance, only inflation rate isnot significant since the Pvalue is greater than 0.05 level of significance.

The cointegration test in Table 5 shows that the Pvalue = 0.2 > 0.05 weaccept the Null Hypothesis and conclude that Variables are cointegrated.

Table 8: (Multicollinearity Test) Variance Inflation Factor (VIF)

Variance Inflation FactorsDate: 05/18/18 Time: 02:01Sample: 1981 2016Included observations: 35

Coefficient Uncentered CenteredVariable Variance VIF VIF

LNEX 0.004503 7.933429 1.789701LNINFL 0.021739 19.67270 1.478417LNINTRT 0.230438 222.6443 1.856462LNPOP 16.92749 1787.737 1.263904C 18.86449 2224.059 NA

From the table 8 above, it can be seen that the Centered VIF for all thevariables are less 10 (i.e VIF < 10), this means that there is no presence ofMulticollinearity among the variables in the FMOLS model estimated.

Figure 3A: Normality Test

The fig 3A above revealed depict that the error term are normallydistributed since the Pvalue 0.647 is greater than 0.05 level of significance.


This implies that the estimated FMOLS Model is robust. The aboveGraphical representation in Fig 3A connotes a normal distribution.

Figure 3B: Graph of the Residual, Actual and Fitted Observation

The fig 3B graph above illustrates that the Actual and the Fitted arealmost close to each other, which implies that the FMOLS model is robust.

To estimate the short run model using the ECM, the first stage is toobtain the optimum lag for the model as shown in table 9A below.

Table 9A: VAR Lag Order Selection Criteria

VAR Lag Order Selection CriteriaEndogenous variables: LNEX LNINFL LNINTRT LNPOP LNUNEMP Sample: 1981 2016Included observations: 33

Lag LogL LR FPE AIC SC HQ

0 30.14700 NA 5.79e06 2.130121 2.356865 2.206413

1 103.3089 218.3823 8.26e09 4.442962 3.082500 3.985208

2 166.7150 84.54150 9.01e10 6.770605 4.276426 5.931390

3 227.3021 62.42307* 1.42e10* 8.927399* 5.299502* 7.706722*

* indicates lag order selected by the criterion

LR: sequential modified LR test statistic (each test at 5% level)

FPE: Final prediction error

AIC: Akaike information criterion

SC: Schwarz information criterion

HQ: HannanQuinn information criterion


In table 9A above, the VAR Lag selection Criteria indicated OptimalLag as 3. Since we are using ECM Model, this then implies that theappropriate Lag selection becomes 2.

Table 9B: ECM for Short Run Analysis

Dependent Variable: D(LNUNEMP)

Method: Least Squares

Date: 05/18/18 Time: 01:53

Sample (adjusted): 1984 – 2016



C 0.012732 0.065087 0.195614 0.8468

D(LNUNEMP(1)) 0.337993 0.191764 1.762543 0.0925

D(LNUNEMP(2)) 0.123905 0.160549 0.771759 0.4489

D(LNEX(1)) 0.463811 0.209843 2.210273 0.0383

D(LNEX(2)) 0.614194 0.219781 2.794575 0.0109

D(LNINFL(1)) 0.041415 0.072109 0.574340 0.5718

D(LNINFL(2)) 0.009572 0.068833 0.139064 0.8907

D(LNINTRT(1)) 0.414488 0.366296 1.131565 0.2706

D(LNINTRT(2)) 0.585830 0.310725 1.885365 0.0733

D(LNPOP(1)) 17.28117 11.88619 1.453887 0.1608

D(LNPOP(2)) 13.30747 9.500538 1.400707 0.1759

ECM(1) 0.489328 0.194418 2.516882 0.0200



S.E. of regression 0.277420 Akaike info criterion 0.548718

Sum squared resid 1.616197 Schwarz criterion 1.092903

Log likelihood 2.946151 HannanQuinn criter. 0.731820

Fstatistic 3.302137 DurbinWatson stat 1.929024

Prob(Fstatistic) 0.008996

Table 9B above, contains the ECM Coefficients and its respective shortrun coefficients as well as their tstatistic and pvalue. The ECM (1) is thecoefficient of the error correction mechanism and it has the correct signand it is significant, which implies that the speed it takes for the system tocome to its equilibrium point is about 48.9%. The ECM model furtherrevealed that there is significant short run effect of Exchange rate onunemployment rate in Nigeria.


Table 10: Variance Inflation Factors (VIF)

Variance Inflation Factors

Date: 05/18/18 Time: 02:05

Sample: 1981 2016

Included observations: 33

Coefficient Uncentered CenteredVariable Variance VIF VIF

C 0.004236 1.816468 NA

D(LNUNEMP(1)) 0.036774 2.134072 2.108777

D(LNUNEMP(2)) 0.025776 1.479169 1.454946

D(LNEX(1)) 0.044034 2.318361 1.763720

D(LNEX(2)) 0.048304 2.525635 1.937603

D(LNINFL(1)) 0.005200 1.342768 1.342756

D(LNINFL(2)) 0.004738 1.287106 1.285391

D(LNINTRT(1)) 0.134173 2.868059 2.854975

D(LNINTRT(2)) 0.096550 2.160367 2.139422

D(LNPOP(1)) 141.2815 4.245635 4.234142

D(LNPOP(2)) 90.26023 4.746861 4.731695

ECM(1) 0.037799 2.691902 2.691498

From the Table 10 above, it can be seen that the Centered VIF for allthe variables are less 10 (i.e VIF < 10), this means that there is no presenceof Multicollearity among the variables or in the model.

Table 11: Serial Correlation LM Test

BreuschGodfrey Serial Correlation LM Test:

Fstatistic 0.306562 Prob. F(2,19) 0.7395

Obs*Rsquared 1.031610 Prob. ChiSquare(2) 0.5970

From Table 11 above, the test of serial correlation was carried out onthe model, the Pvalue = 0.7395 > 0.05 revealed that there is no serialcorrelation in the Error of the estimated ECM model. The implication ofthis also is that the fitted unemployment rate model is good for forecasting.

Table 12: Heteroskedasticity Test

Heteroskedasticity Test: BreuschPaganGodfrey

Fstatistic 3.131981 Prob. F(11,21) 0.0118

Obs*Rsquared 20.50266 Prob. ChiSquare(11) 0.0389

Scaled explained SS 5.633895 Prob. ChiSquare(11) 0.8966

From table 12 above, the test of Heteroskedasticity test was carriedout on the model, the result revealed that the variance of the residual are


not constant, since the (Pvalue = 0.0118 < 0.05). This implies that there isevidence of Heteroskedastic in the Residual. This could be as a result ofthe transformation of the variables.

Figure 4: Plot of CUSUM Test for Stability of ECM Model

Figure 4 above, the results indicated the absence of any instability ofthe coefficients because the plot of the CUSUM statistic falls inside thecritical bands of the 5% confidence interval of parameter stability. In essence,we say the model is Stable because the blue line is strictly between the RedLine (at 5% significance). This also further tells us that this model is goodfor forecasting.

Table 13: Pairwise Granger Causality Tests

Pairwise Granger Causality TestsDate: 05/18/18 Time: 12:02Sample: 1981 2016Lags: 2

Null Hypothesis: Obs FStatistic Prob.

LNINFL does not Granger Cause LNEX 34 2.22268 0.1264

LNEX does not Granger Cause LNINFL 1.00963 0.3768

LNINTRT does not Granger Cause LNEX 34 0.08249 0.9210

LNEX does not Granger Cause LNINTRT 0.46355 0.6336

LNPOP does not Granger Cause LNEX 34 1.31113 0.2850

LNEX does not Granger Cause LNPOP 2.17474 0.1318


LNUNEMP does not Granger Cause LNEX 34 0.11264 0.8939

LNEX does not Granger Cause LNUNEMP 9.82263 0.0006

LNINTRT does not Granger Cause LNINFL 34 0.27479 0.7617

LNINFL does not Granger Cause LNINTRT 3.27355 0.0523

LNPOP does not Granger Cause LNINFL 34 2.79819 0.0774

LNINFL does not Granger Cause LNPOP 0.59109 0.5603

LNUNEMP does not Granger Cause LNINFL 34 2.96622 0.0673

LNINFL does not Granger Cause LNUNEMP 1.86380 0.1732

LNPOP does not Granger Cause LNINTRT 34 0.21096 0.8110

LNINTRT does not Granger Cause LNPOP 7.32613 0.0027

LNUNEMP does not Granger Cause LNINTRT 34 1.12379 0.3388

LNINTRT does not Granger Cause LNUNEMP 0.04886 0.9524

LNUNEMP does not Granger Cause LNPOP 34 3.33373 0.0498

LNPOP does not Granger Cause LNUNEMP 0.46011 0.6357

From Table 13 above, it revealed that in the short run, Exchange ratecauses Unemployment. This corresponds to the result obtained in the ECMmodel.

5. DISCUSSION OF FINDINGS

The descriptive statistics revealed mean values for log of Exchange Rate,Log of Inflation rate, Log of Interest Rate, Log of Population Growth andLog of Unemployment are 3.293656, 2.674964, 2.828446, 0.947595 and 2.103168 respectively. The descriptive statistics further revealed that thevariables are normally distributed, this agrees with the work of Fuhrer(2017). The implication of this is that the FMOLS and ECM models areappropriate because these models assume normal distribution.

The Augmented Dickey Fuller test revealed that Exchange Rate,Inflation rate, Interest Rate, Population Growth and Unemployment areintegrated at order 1 that is I(1), this is expected because most macroeconomic variables have trends in them. This is in line with the work ofLibanio (2005). Since the variables are integrated, one may suspect thepossibility of cointegration which means Long run relationship. TheJohansen cointegration test for the macro variables revealed that thevariables are cointegrated which is in line with economic expectation thatstate that most macroeconomic variables tends to exhibit long runrelationship (Kwon and Shin, 1999). Since the variables are integrated oforder one and also cointegrated, the FMOLS model was then applied tothe variables using Unemployment rate as the dependent variables while


others are independent variables. The estimated FMOLS revealed that theexpected unemployment figure in Nigeria will be 8.6% if there are noeffects of these and other macro variables. The estimated FMOLS modelalso revealed that there is a positive relationship between Exchange rateand Unemployment Rate in Nigeria. By this relationship, it means that ifthere is unit increase in Exchange Rate, Unemployment Rate will Increaseby 0.3 per unit increase in Exchange Rate in Nigeria. This contradicts theresult of Bakhshi and Ebrahimi (2016) for Iran economy. The implication ofour result shows that the Nigerian economy depends more on Importation.The FMOLS model also revealed a significant positive relationship betweenPopulation growth and unemployment rate in Nigeria that is if there is aunit increase in Population Growth, Unemployment Rate will increase by14.49 per unit increase in Population Growth. The implication is that anuncontrolled population growth will lead to explosion on unemploymentRate, this is in line with the work of Loku and Deda (2013).

The FMOLS model also revealed that Interest Rate and Inflation Rateare negatively related to Unemployment but only Interest Rate is Significant(Pvalue = 0.0124 > 0.05). This means that if there is unit increase in InflationRate, Unemployment Rate will decrease by 0.16 per unit increase in InflationRate. Also, if there is unit increase in Interest Rate, Unemployment Ratewill decrease by 1.28 per unit increase in Interest Rate. This agrees with(Alisa 2015)., The Coefficient of determination (Rsquared) shows that 74% variation in unemployment rate can be explained jointly by fourindependent variables such as log of Interest Rate, log of Exchange Rate,log of inflation rate and log of Population growth. The rest 26% variationin Unemployment Rate can be explained by residuals or other variablesother than the four independent variables. The FMOLS model also revealedthat the variables are cointegrated, no presence of Multicollinearity, theexpected fitted values of the model are close to the actual while the Errorfrom the model are normally distributed which fulfil the assumption ofthe OLS .

To understand the Short run Dynamics among these macro variables,the ECM model was adopted. The VAR selection criteria revealed that anoptimum Lag of 2 was selected. The fitted ECM model revealed a ShortRun impact from Exchange Rate to Unemployment Rate in Nigeria. Thisalso implies that Nigerian economy depend on Importation. This agreeswith the work of Nyahokwe & Ncwadi (2013). While the ECM coefficientis –0.489 which has the expected sign and it is also significant, this meansthat for the system to come to equilibrium, it takes a speed up of 48.9%annually. This agrees with the work of Alisa (2015).


The diagnostic Test revealed that there are no presence ofMulticollinearity, Serial correlation. Lastly, the expected ECM model issuitable for forecasting.

6. CONCLUSION AND RECOMMENDATIONS

This study investigated the Lon Run and the short Run relationship amongall the macro variables (that is Unemployment, Inflation Rate, ExchangeRate and Interest Rate). A secondary data was sourced from CBN StatisticalBulletin and World Bank website. The descriptive statistics revealed meanvalues for log of Exchange Rate, Log of Inflation rate, Log of Interest Rate,Log of Population Growth and Log of Unemployment are 3.293656,2.674964, 2.828446, 0.947595 and 2.103168 respectively. The descriptivestatistics further revealed that the variables are normally distributed. TheAugmented Dickey Fuller test revealed that Exchange Rate, Inflation rate,Interest Rate, Population Growth and Unemployment are integrated atorder 1 that is I(1), this is expected because most macroeconomic variableshave trends in them. The Johansen cointegration test for the macro variablesrevealed that the variables are cointegrated. Since the variables areintegrated of order one and also cointegrated, the FMOLS model was thenapplied to the variables using Unemployment rate as the dependentvariables while others are independent variables. The estimated FMOLSrevealed that the expected unemployment figure in Nigeria will be 8.6%if there are no effects of these and other macro variables. The estimatedFMOLS model also revealed that there is a positive relationship betweenExchange rate and Unemployment Rate in Nigeria. The FMOLS modelalso revealed a significant positive relationship between Population growthand unemployment rate in Nigeria that is if there is a unit increase inPopulation Growth, Unemployment Rate will increase by 14.49 per unitincrease in Population Growth. The FMOLS model also revealed thatInterest Rate and Inflation Rate are negatively related to Unemploymentbut only Interest Rate is Significant (Pvalue = 0.0124 Â 0.05). The Coefficientof determination (Rsquared) shows that 74 % variation in unemploymentrate can be explained jointly by four independent variables such as log ofInterest Rate, log of Exchange Rate, log of inflation rate and log of Populationgrowth. The rest 26% variation in Unemployment Rate can be explainedby residuals or other variables other than the four independent variables.

The FMOLS model also revealed that the variables are cointegrated,no presence of Multicollinearity, the expected fitted values of the modelare close to the actual while the Error from the model are normallydistributed.


To understand the Short run Dynamics among these macro variables,the ECM model was adopted. The VAR selection criteria revealed that anoptimum Lag of 2 was selected. The fitted ECM model revealed a ShortRun impact from Exchange Rate to Unemployment Rate in Nigeria. Whilethe ECM coefficient is –o.489 which has the expected sign and it is alsosignificant, this means that for the system to come to equilibrium, it takesa speed up of 48.9% annually. The diagnostic Test revealed that there areno presence of Multicollinearity, Serial correlation. The estimated ECMmodel is suitable for forecasting.

In summary, this study revealed that Exchange Rate and PopulationGrowth are significant and are positively related to Unemployment ratewhile Inflation Rate and Interest Rate are negatively related toUnemployment in Nigeria with only Interest Rate being significant.

The following are recommended due to the findings of this study:

i) Government should create other sources of employment likesports activities that will adequately engage the youths.

ii) Also youth empowerment should be encouraged in order toencourage small and medium scale enterprises through the useof some low interest finance banks like BOI (Bank of Industry)for the grassroots. This is to encourage the Small and Mediumenterprise for the youths with great insights.

iii) Government should endeavor to create new classes of loan withreduced interest rate which is targeted at financing youngentrepreneurial and based on that it means that banks and otherfinancial institutions still have a role to play.

iv) Improvement is required on other sector such as the nonoil sectorof the economy like Agriculture

v) Control of the population may be necessary by Government.

vi) Skill acquisition and trainings should be introduced in a wise wayso that it will engage the unemployed in order to reduce theunemployment situation in the country.

References

Adenomon, M. O.: Ahmed, I. & Abubakar, M. A. (2018). Fully Modified Ordinary LeastSquares Analysis of the Long Run Interrelationship Among Macroeconomic Variablesin Nigeria. A Paper presented at the 2nd FNAS National Conference, Nasarawa StateUniversity, Keffi, 27th to 30th May, 2018.

Afzal Muhammad & Awais S. (2012): InflationUnemployment Trade Off: Evidence fromPakistan, Journal of Global Economy 8(1), 2132.


Ajayi, A. and Mougoue, M. (1996). “On the Dynamic Relationship Between Stock Pricesand Exchange Rates” Journal of Financial Research 9(2): 193207.

Alisa, M.(2015): The Relationship Between Inflation and Unemployment: A theoreticalDiscussion about the Phillips Curve. Journal of International Business and Economics,3(2): 8997.

Aminu U. & Anono A. Z. (2012). An Empirical Analysis Of The Relationship BetweenUnemployment And Inflation In Nigeria From 19772009, Economics and Finance Review.1(12): 42 – 61.

BaahBoateng,W. (2013). Determinants of Unemployment in Ghana. African DevelopmentReview. 25(4): 85–399.

Bakhshi, Z. & Ebrahimi, M. (2016). The Effect of Real Exchange rate on UnemploymentMarketing and Branding Research, 3:413.

Central Bank of Nigeria (2003). Contemporary Economic Policy Issues in Nigeria. Abuja.

Cheema A. R., & Atta A. (2014): Economic Determinants Of Unemployment in Pakistan:CoIntegration Analysis, International Journal of Business and Social Science 5(3):209221

Countries: Evidences from Tanzania. Journal of Economics and Sustainable Development, 4(14):2222 2855

Eita, J. H. & Johanson M. A. (2010). Determinants of Unemployment in Namibia. InternationalJournal of Business and Management. 5(10): 92 – 104.

ElAgrody, N. M., Othman, A. Z., & Hassan, M. B.D. (2010). Economic Study ofUnemployment in Egypt and Impacts on GDP. Nature and Science, 8(10): 102111.

Engle, R. F. and Granger, C. W. J. (1987). CoIntegration and Error Correction:Representation, Estimation and Testing. Econometrica. 55(2):251276.

Fuher, J. (2017). Expectations as a source of macroeconomic persistence: Evidence from surveyexpectation in a dynamic macro model. Journal of Monetary Economics 86: 2235.

Habte Z, Market Integration for Oxen Prices Using Vector Error Correction Model (VECM)in Ethiopia. Int. J. of Tech. Enhancement and Emerging Eng. Research 2: 69(2014).

Hendry, D. F.(2003). J. Denis Sargan and the Origins of LSE Econometric Methodology.Econometric Theory, 19(3): 457480.

Johansen S & Juselius K, Maximum Likelihood Estimation and Inference on Cointegrationwith Application to the Demand for Money. Oxford Bullentin of Econs & Stat. 52: 169210 (1990).

Kabaklarli, E., & Bulus, A. (2011). Economic Determinants of Turkish Youth UnemploymentProblem: CoIntegration Analysis. International Conference on Applied Economics. 267.

Kemi F. A. & Dayo B. O. (2014). Unemployment and Economic Growth in Nigeria. Journalof Economics and Sustainable Development. 5(4): 138–144.

Kuhe, D. A.(2016): An Econometric Analysis of Crime Determinants in Nigeria. Journal ofthe Nigerian Statistical Association, 28:115.

Kwon, C.S and Shin, T.S (1999): Cointegration and causality between macroeconomicVariables and stock Market Returns, Global Finance Journal, 10(1): 7181.

Loku, A. & Deda, G (2013). Unemployment in Relation of Growth Population. InternationalJournal of Research in Social Sciences, 2(2):3138.

Maigwa R. & Kipesha E. F. (2013). Determinants of Youth Unemployment in DevelopingCountries: Evidences from Tanzania. Journal of Economics and Sustainable Development,4(14): 2222 2855.


Maqbool M. S, Sattar T., Mahmood A. & Bhalli M. N. (2013). Determinant of UnemploymentEmpirical Evidences from Pakistan Economics and Social Review. 51(2): 191207.

Muhdin, M.B. (2016). Determinants of youth unemployment of urban areas of Ethopia.International Journal of Scientific and Research Publications, 6(5): 43350.

Nasir, F., Cheema, M. A., Zafar M.I., & Batool, Z. (2009). SocioEconomic Impacts ofUnemployment in Urban Faisalabad, Pakistan. Journal of Social Science, 18(3): 183188.

Nyahokwe,O. & Ncwadi, R (2013). Import of Exchange Rate Volatility on Unemploymentin South Africa. Mediterranean Journal of Social Science, 4(3): 109120.

Okun, A. (1962). ‘Potential GNP: Its Measurement and Signiûcance’, in the 1962 Proceedingsof Business and Economic Statistics Section, American Statistical Association.

Pallis, D. (2006). The Tradeoff between inflation and unemployment in new EuropeanUnion member states. International Research Journal of Finance and Economics 1:8197.

Phillips, P. C. B. & Hansen, B. E. (1990). Statistical Inference in Instrumental Regressionwith I(1) Processes. The Review of Economic Studies, 57(1): 99125.

Phillips, P. C. B. & Loretan, M. (1991). Estimating LongRun Economic Equilibra. The Reviewof Economic Studies, 58(3): 407436.

Salmon, M.(1982). Error Correction Mechanisms. Economic Journal 92, 615629.

Salvatore, D. and Reagle, D. (2002). Schaum’s Outline of Theory and Problems of Statisticsand Econometrics (2 ed). New York: McGrawHill.

Sargan, J. D.(1964). Wages and Prices in the United Kingdom: A Study in EconometricMethodology (with discussion). In P. E. Hart, G. Mills & J. K. Whitaker (eds),Econometric Analysis for national Economic Planning, Vol. 16 of Colston Papers, Pp.2563, London: Butterworth.


Adenomon, M.O., OkoroUgochukwu, N.A. and Adenomon, C.A. On the Determinantsof Unemployment Rate in Nigeria: Evidence from Fully Modified OLS and ErrorCorrection Model. Journal of Asian Economics, Accounting and Finance, Vol. 1, No. 2,2020, pp. 199225


Appendix

Table 1: Annual Data on Exchange Rate, Inflation Rate, Interest Rate, UnemploymentRate and Population Growth from 1981 to 2016

Year Exchange rate Inflation rate Interest Rate Unemployment Rate Pop Growth

1981 0.61 20.9999 7.75 6.4 2.71506

1982 0.673 7.6 10.25 6.4 2.60268

1983 0.724 23.1999 10 6.4 2.53541

1984 0.765 39.6 12.5 6.2 2.52929

1985 0.894 5.5 9.25 6.1 2.56273

1986 2.021 5.4 10.5 5.3 2.6032

1987 4.018 10.2 17.5 7 2.62564

1988 4.537 56.1002 16.5 5.3 2.63093

1989 7.392 50.4998 26.7999 4.5 2.61241

1990 8.038 7.5 25.5 3.5 2.57904

1991 9.909 12.9 20.01 3.1 2.54561

1992 17.298 44.5998 29.8 3.4 2.52124

1993 22.051 57.1998 18.3199 2.7 2.50297

1994 21.886 56.9999 20.9999 2 2.493

1995 21.886 72.9001 20.18 1.8 2.48943

1996 21.886 30.3999 19.7401 3.4 2.48837

1997 21.886 8.2 13.54 3.2 2.48818

1998 21.886 10.3 18.2899 3.2 2.49072

1999 92.694 6.7 21.3201 3 2.49581

2000 102.105 6.9 17.98 18.1 2.5034

2001 111.943 18.9 18.2899 13.7 2.51121

2002 120.97 12.9 24.8501 12.2 2.52111

2003 129.356 14 20.7101 14.8 2.53684

2004 133.5 14.9 19.18 11.8 2.55924

2005 132.146 17.9 17.95 11.9 2.58522

2006 128.652 8.2 17.26 13.7 2.61039

2007 125.834 5.3 16.94 14.6 2.63165

2008 118.567 11.6001 16.94 14.9 2.64897

2009 148.88 13.7001 15.14 19.7 2.66122

2010 150.298 10.8 18.99 21.1 2.66875

2011 153.861 10.3 17.59 23.9 2.67475

2012 157.499 11.5 16.02 24.3 2.67766

2013 157.312 8.5 17.0899 28.5 2.67292

2014 158.553 8.05 16.28 30 2.65955

2015 192.439 8.2 16.8599 24 2.64036

2016 253.489 9.6 16.54 25 2.61903

Source: CBN Statistical Bulletin (2016) & World Bank Website (www.worldbank.org)


Table 2: Natural Log Transformation of the Annual Data on Exchange Rate, Inflation Rate,Interest Rate, Unemployment Rate and Population Growth from 1981 to 2016

year LNEX LNINFL LNINTRT LNPOP LNUNEMP

1981 0.4943 3.04452 2.04769 0.998815 1.8563

1982 0.39616 2.02815 2.32728 0.95654 1.8563

1983 0.32283 3.14415 2.30259 0.930356 1.8563

1984 0.26801 3.67883 2.52573 0.927938 1.82455

1985 0.11227 1.70475 2.22462 0.941074 1.80829

1986 0.70339 1.6864 2.35138 0.956742 1.66771

1987 1.39076 2.32239 2.8622 0.965324 1.94591

1988 1.5122 4.02714 2.80336 0.967338 1.66771

1989 2.00034 3.92197 3.2884 0.960275 1.50408

1990 2.08416 2.0149 3.23868 0.947416 1.25276

1991 2.29349 2.55723 2.99623 0.934371 1.1314

1992 2.85061 3.79773 3.39451 0.924751 1.22378

1993 3.09336 4.04655 2.90799 0.917478 0.99325

1994 3.08585 4.04305 3.04452 0.913485 0.69315

1995 3.08585 4.28909 3.00469 0.912056 0.58779

1996 3.08585 3.41444 2.98265 0.911626 1.22378

1997 3.08585 2.10413 2.60565 0.911553 1.16315

1998 3.08585 2.33214 2.90635 0.912574 1.16315

1999 4.5293 1.90211 3.05965 0.914615 1.09861

2000 4.626 1.93152 2.88926 0.917649 2.89591

2001 4.71799 2.93916 2.90635 0.920766 2.6174

2002 4.79554 2.55723 3.21286 0.924698 2.50144

2003 4.86257 2.63906 3.03062 0.930919 2.69463

2004 4.8941 2.70136 2.95387 0.93971 2.4681

2005 4.88391 2.8848 2.88759 0.949811 2.47654

2006 4.85711 2.10413 2.84839 0.9595 2.6174

2007 4.83496 1.66771 2.82968 0.967613 2.68102

2008 4.77548 2.45101 2.82968 0.97417 2.70136

2009 5.00314 2.6174 2.71734 0.978785 2.98062

2010 5.01262 2.37955 2.94391 0.981609 3.04927

2011 5.03605 2.33214 2.86733 0.983858 3.17388

2012 5.05942 2.44235 2.77384 0.984943 3.19048

2013 5.05823 2.14007 2.83849 0.983171 3.3499

2014 5.06609 2.08567 2.78994 0.978157 3.4012

2015 5.25978 2.10413 2.82494 0.970914 3.17805

2016 5.53532 2.26176 2.80578 0.962805 3.21888


Table 14: Unit Root Test for Exchange Rate

Null Hypothesis: D(LNEX) has a unit root

Exogenous: Constant

Lag Length: 0 (Automatic based on AIC, maxlag=9)

tStatistic Prob.*

Augmented DickeyFuller test statistic 5.020825 0.0002

Test critical values: 1% level 3.639407

5% level 2.951125

10% level 2.614300

*MacKinnon (1996) onesided pvalues.

Table 15: Unit Root Test for Inflation Rate

Null Hypothesis: D(LNINFL) has a unit root

Exogenous: Constant


tStatistic Prob.*



5% level 2.954021

10% level 2.615817


Table 16: Unit Root Test for Interest Rate

Null Hypothesis: D(LNINTRT) has a unit root

Exogenous: Constant


tStatistic Prob.*



5% level 2.951125

10% level 2.614300


Table 17: Unit Root Test for Population Growth Rate

Null Hypothesis: D(LNPOP) has a unit root

Exogenous: Constant


tStatistic Prob.*



5% level 2.954021

10% level 2.615817



Table 18: Unit Root Test for Unemployment Rate

Null Hypothesis: D(LNUNEMP) has a unit root

Exogenous: Constant


tStatistic Prob.*



5% level 2.951125

10% level 2.614300


journal proposal / call for papers - ESI Publications

Documents

Transcript of journal proposal / call for papers - ESI Publications