Genetic Distance and Cognitive Human Capital: A Cross-National Investigation

J Bioecon (2016) 18:33–51DOI 10.1007/s10818-015-9210-7

Genetic distance and cognitive human capital:a cross-national investigation

Oasis Kodila-Tedika1 · Simplice A. Asongu2

Published online: 25 November 2015© Springer Science+Business Media New York 2015

Abstract This paper explores the correlates of the intelligence quotient and cognitiveability by focusing on genetic distance to frontier nations. The results based on cross-sectional data from 167 countries suggest that genetic distance to global frontiers hasa negative relationship with the employed human capital variables. Countries thatare genetically far from leading nations tend to have lower levels of human capitalwith the negative correlation to the USA frontier averagely higher relative to the UKfrontier. The sign is consistent and survives the control ofmacroeconomic, geographic,institutional and other covariates. Policy implications are discussed.

Keywords Intelligence · Human capital · Genetic distance

JEL Classification G15 · O50 · O16 · F15 · N7

1 Introduction

Why does human capital or intelligence differ across countries? The question isimportant because a substantial body of literature has established a significant rela-tionship between human capital and development (e.g. Lynn and Vanhanen 2012;

B Oasis [email protected]

Simplice A. [email protected]

1 Department of Economics, University of Kinshasa, Kinshasa,Democratic Republic of the Congo

2 African Governance and Development Institute, Yaoundé, Cameroon

123

http://crossmark.crossref.org/dialog/?doi=10.1007/s10818-015-9210-7&domain=pdf

34 O. Kodila-Tedika, S. A. Asongu

Kalonda-Kanyama and Kodila-Tedika 2012; Kodila-Tedika 2014; Rindermann et al.2015).1

There is a stream of studies that has addressed the question with a number ofvariables (Wicherts and Wilhelm 2007; Hunt 2012). Many explanations have beenadvanced to elucidate the Flynn effect.2 Some explanations to the cross-country differ-ences from Barber (2005) include: nutrition, health, mass media diffusion, education;which have been documented to promote gains in intelligence quotient (IQ). Accord-ing to Wicherts and Wilhelm (2007) and Wicherts et al. (2010), IQ differences acrosscountries are correlated with, inter alia: the number of personal computers per 1000people (0.66), urbanisation (0.67), fertility rates (−0.86), pupil-teacher ratio (−0.72)and secondary school enrolment ratio (0.78).

A genetical angle has also been engaged. Accordingly, studies have established alinkage between cognitive test results and the genotype of individuals (Davies et al.2011; Hunt 2012; Johnson et al. 2011). Notable authors that fall along this stream.include: Lynn andVanhanen (2002, 2006). Rushton and Jensen (2005), Schwekendiek(2009), Lai (2006) andPak (2010). Somehave put forward an argument of evolutionarynecessity (Lynn and Vanhanen 2006; Kanazaw 2008). However, Wicherts et al. (2010)have expressed doubts about the underlying arguments and conclusions. Rushton andJensen (2005) and Nisbett (2009) have for the most part pointed to the race anddifferences that exist in the ethnic composition.3 Hunt (2012) has expressed scepticismin the face of these arguments because they are unsustainable. On the importanceof genetics, the author states “Until then. the question “Is there a genetic basis forinternational differences in intelligence?” has a simple answer: We do not know”(Hunt2012, p. 295).

The present line of inquiry attempts to address the underlying issue by respondingto the question put forward by Hunt. We employ data on genetic distance compiledby Spolaore and Wacziarg (2009) who have provided a summary indicator of howpopulations are related to genealogy over time. According to the authors, geneticdistance between populations is highly linked to per capita income differences acrosscountries. They have disputed that genetic distance among people captures a widerange of characteristic and trait differences. In line with the narrative, the effects ofthis distance represent barriers to the diffusion of economic development from globaltechnological frontier, since variations in these characteristics stifle the flow of goods,technologies and ideas across population, which ultimately hinder development.

Ashraf and Galor (2013) along the same line have recently shown that ethnic diver-sity is linked to human development in a non-linear or Kuznets manner, with thecountries of medium diversity having the highest levels of economic development.We do not employ this measure of genetic diversity because it has been severely crit-icised (Guedes et al. 2013). Whereas genetic distance refers to genetic differencesacross populations, genetic diversity is defined in terms of heterogeneity across pop-

1 Also see: Kodila-Tedika and Mutascu (2014) and Kodila-Tedika and Bolito-Losembe (2014).2 The Flynn effect represents a phenomenon where-by on average the Intelligence Quotient (IQ) scoreshave been increasing worldwide over time, with younger generations performing relatively better than theirolder counterparts. The average IQ score increases by about 10 points per generation.3 We do not believe in race superiority, whatsoever.

123

Genetic distance and cognitive human capital... 35

ulations (Spolaore and Wacziarg 2009). Compared to genetic diversity, the geneticdistance measurement is a more interesting indicator because it has been subject toless criticism in the academic literature (Guedes et al. 2013;Asongu andKodila-Tedika2015).

The argument of this study is very intuitive and is summarised in the following ques-tion. Are genetic differences the basis for variations in human capital across countries?Human capital is appreciated with the IQ and cognitive ability. These two indicatorsspecifically tackle the issue raised by Hunt because they are linked to the concept ofintelligence. Themeasurement of Spolaore andWacziarg (2009) is interesting because,it enables us to assess if genetics explain development differences across countries.Moreover, the indicator also permits us to evaluate the role of cultural transmissions.Distant genetic populations tend to differ in many characteristics that are transmittedbetween generations, notably: appearance, language, norms, habits, beliefs, customsand values. Authors of this index suppose that their measurement contains all sortsof intergenerational traits. On the basis that this indicator combines the highlighteddimensions, it is reasonable to think that deviations between countries can be due todifferences in the perception of human capital between generations on the one handand between geographic regions on the other hand. Hence, we could logically imaginethat a generation that begun sending their children early to school should benefit froman educational advantage over that which did not. In this respect, the psychology ofeducation plays a major role in the development of cognitive capacities (Becker etal.. 2013). More insights into the IQ and education debate are available in Deary andJohnson (2010).

The empirical approach consists of regressing the index of human capital fromMeisenberg and Lynn (2011) on genetic distance. This human capital index is inter-esting in the perspective that it also combines education and IQ. Hence, it takes boththe input and output dimensions of human capital into account, which is not the casewith traditional indicators (Lutz 2009). Our econometric results show a solid statisticallinkage between genetic distance and employed human capital indicators.The theoretical intuition is to reconcile genetic and environmental dimensions. Thegenetic distance measurement enables the affirmation that the environment or cultureof a person influences his/her IQ. Accordingly, an educated parent or a parent with ahigh IQ is more likely to send his/her children to school. Moreover, the parents wouldwant their children to evolve in an environment that is conducive for the developmentof their IQs

The rest of the study is organised as follows. Section 2 discusses the data andempirical strategies. The results are presented and discussed in Sect. 3. Section 4provides robustness checks while Sect. 5 concludes.

2 Data and methodology

2.1 Data

Weassess a cross-sectional sample of 167 countrieswhich consists of 1996–2009 aver-ages. The human capital dependent variables are the IQ and cognitive ability, while

123


the independent variable of interest is the genetic distance. The dependent and inde-pendent variables are in accordance with Meisenberg and Lynn (2011) and SpolaoreandWacziarg (2009) respectively. The control variables include: latitude, temperature,longitude, institutions, malaria and GDP per capita (in logarithm). The latitude whichis measured in absolute value is calculated from La Porta et al. (1999) with someadditions from the Central Intelligence Agency (CIA) World Factbook (CIA 2010).Temperature represents average monthly temperature of nations in degrees Celsiusover the period 1961–1990, computed utilising the geographical mean of data frommonthly temperature reported by the Geographically based Economic data (G-Econ)project (Nordhaus 2006) at a resolution of one degree which is based on similar datathat is spatially disaggregated at a ten minute resolution (New et al. 2002). Therefore,the indicator is a spatial average of the intertemporal monthly average temperaturewithin a nation across grid cells.4 The Longitude indicator is consistent with Easterly(2011).

On the measurement of institutions, we consider a composite indicator consistingof six governance dynamics from Kaufmann et al. (2010) averaged over 1996–2009,notably: government effectiveness, regulation quality, rule of law, corruption-control,voice and accountability and political stability/no violence. TheMalaria measurementis the malaria ecology index which appreciates the suitability of a nation’s climateto the breeding of mosquito as well as the prevalence of species of mosquito thatessentially depends on human for survival (Kiszewski et al. 2004). GDP per capita(log) measurement is in natural logarithm of the Gross Domestic Product per capitaaveraged between 1996 and 2009 from Penn World Tables 7.0.

The summary statistics of the variables is presented in Appendix 1. From thedescriptive statistics, it can be noticed that variables are quite comparable. Moreover,from the standard deviations, we can be confident that reasonable estimated linkageswould emerge owing to the substantial degree of variations. The correlation matrix isalso presented in Appendix 2 to provide us with a feeling of expected signs.

2.2 Empirical specification

Consistent with Kodila-Tedika and Asongu (2015) who have investigated the effectof intelligence on financial development, we employ baseline Ordinary Least Squares(OLS) in order to assess the linkages motivating this study. Hence, Eq. (1) belowexamines the relationship between genetic distance and human capital across 167countries.

HCi = α1 + α2G Di + α3Xi + εi (1)

where HCi (G Di ) represents a human capital variable (genetic distance) for countryiα1 is a constant,X is the vector of control variables, and εi the error term. X entails:Latitude, Temperature, Longitude, Institutions, Malaria and GDP per capita (log).Equation (1) is estimated by OLS using heteroscedasticity consistent standard errors.

4 We invite the interested reader to refer to the G-Econ project for more information.

123


3 Empirical results

Figures 1 and 2 below show the visual relationship between the IQ and genetic distancefrom the United States (US). Figure 1(2) lays emphasis on countries (continents). Wenotice a negative relationship, implying that countries/continentswhich are geneticallyfar from leading nations (like the US) tend to have lower levels of human capital.Accordingly, the presence of most African countries at the bottom-right in Fig. 1 isconfirmed with continent’s representations in Fig. 2. This is not the case of Europethat is closer to the US and is abundantly represented at the top-left with respect tocountry-specific (Fig. 1) and continental (Fig. 2) representations.Table 1 below presents the OLS findings based on Eq. (1). Column 2 or Specification1 shows univariate regression estimates where the IQ is regressed on the geneticdistance (GD) to the United States. It is evident from the findings that these twovariables demonstrate a substantial negative correlation. The GD coefficient of the firstregressions is−28.158 and significant at the 1% level. This estimate shows that nations

AFG

AGO

ALB

ARE

ARGARM

AUSAUT

AZE

BDI

BEL

BEN

BFA

BGD

BGR

BHRBHS

BLR

BLZ

BOLBRA

BRB

BRN

BTN

BWA

CAF

CANCHE

CHL

CHN

CIV CMR

COG

COL

COM

CPV

CRICUB

CYP

CZE

DJIDMA

DNK

DOMDZA

ECU

EGY

ERI

ESP

EST

ETH

FIN

FJI

FRA

GAB

GBR

GEO

GER

GHA

GIN

GMB

GNB

GNQ

GRC

GTM

GUY

HKG

HND

HRV

HTI

HUN

IDNIND

IRL

IRNIRQ

ISL

ISRITA

JAM

JOR

JPN

KAZ

KENKGZ

KHM

KWT

LAO

LBN

LBR

LBY

LKA

LSO

LTULUX LVA

MAR

MDA

MDG

MEX

MGD

MKD

MLI

MLT

MMR

MNG

MOZ

MRT

MUS

MWI

MYS

NAM

NER

NGA

NIC

NLD

NOR

NPL

NZL

OMNPAK

PAN

PER PHL

POLPRT

PRYQAT

ROM

RUS

RWA

SAUSDN

SEN

SGP

SLE

SLV

SOM

SPG

SUR

SVKSVNSWE

SWZSYCSYR

TCDTGO

THA

TJK TKM

TONTTOTUN

TUR

TWN

TZAUGA

UKR

URY

USA

UZB

VEN

VNM

VUT

ZAF

ZARZMBZWE

6070

8090

001011

0 .2 .4 .6 .8 1fst_gendist_to_US

hc Fitted values

Fig. 1 Human capital (IQ) and genetic distance (with emphasis on countries)

6070

8090

100

110

latip aC

namu

Hev iti ngo

C

0 .2 .4 .6 .8 1Genetic Distance

Africa Europe Asia Oceania Americas

Fig. 2 Human capital (IQ) and genetic distance (with emphasis on continents)

123


Table 1 OLS estimates of the impact of genetic distance

Dependent variable: human capital(1) (2) (3) (4) (5) (6)

GD_US(fst) −28.158*** −19.990*** −11.067*** −10.203*** −10.493*** −8.980***

(1.842) (2.787) (3.130) (3.270) (3.292) (3.276)

Absolute valueof latitude

−4.670 −11.889* −11.237* −11.140* −6.037

(6.562) (6.857) (6.395) (6.260) (5.815)

Meantemperature

−0.495*** −0.445*** −0.379** −0.362** −0.345**

(0.164) (0.165) (0.165) (0.165) (0.142)

Longitude 0.045*** 0.042*** 0.041*** 0.040*** 0.116***

(0.014) (0.009) (0.009) (0.009) (0.027)

GDP percapita (log)

6.084*** 5.676*** 4.716*** 4.062***

(0.833) (0.827) (0.955) (0.836)

Malaria −0.172** −0.198** −0.001

(0.078) (0.079) (0.093)

Institutions 0.058 0.067**

(0.035) (0.031)

Africa 11.538**

(4.530)

Americas 24.009***

(6.466)

Asia 11.897***

(3.087)

Europa 16.708***

(4.058)

Oceania (dropped)

Constant 97.340*** 102.270*** 47.390*** 49.724*** 55.350*** 41.442***

(0.916) (5.063) (10.782) (10.619) (10.366) (9.993)

Observations 169 98 83 83 83 83

R2 0.407 0.444 0.743 0.754 0.760 0.815

GD_US(fst): FST measure of genetic distance to the USA*** .01; ** .05; * .1

that are genetically far from the United States tend to have considerably lower levels ofhuman capital. The corresponding coefficient of determination (R2) of 0.407, indicatesthat GD explains about 40 % of the total variation in human development acrosscountries. The negative relationship between the IQ and ‘GD to the US’ is consistentacross specifications, with varying conditioning information set or covariates.

The significant control variables have the expected signs. First, the negative effectthe Latitude variable may be due to its high correlation with Temperature (see cor-

123


relation matrix in Appendix 2). Second, temperatures are negatively correlated withhuman capital, because some findings have shown that higher IQ is a compensation forcold weather (Jacob 2010). Third, longitudes have been documented to be positivelylinked with the dependent variable (Kanazaw 2008) and in the context of our study;its comparatively low magnitude may be traceable to the apparent criticism followingthe study of Kanazawa (Denny 2009). Fourth, good institutions and per capita eco-nomic prosperity have been established to be positively associated with human capital(Meisenberg and Lynn 2011, p. 434). Fifth, in the same vein, Meisenberg and Lynn(2011) have shown that infant mortality (life expectancy) is negatively (positively)linked to the development of human capital. Hence it is natural to infer that Malariaexerts a negative influence on the dependent variable, since like many diseases it hasbeen shown to significantly reduce life expectancy, by increasing infantmortality, interalia (Weil 2010; Asongu 2014a, b). Sixth, the relatively high (low) human capital lev-els of European (African) countries discussed in Figs. 1 and 2 are consistent with thecontinental magnitudes.

4 Robustness checks

In this section, we perform several robustness checks on the baseline specification inColumn VI of Table 1. These checks use alternative GD indicators to global frontiers.Specifically, we employ: (i) the FST measure of GD to the US; (ii) the FST measureof GD to the UK and (iii) the Nei measure of GD to the US. These measures areconsistent with Ang and Kumar (2014, p. 53). According to the authors, whereas theFST measure of GD is the most widely employed in the literature, it is also necessaryto complement the measurement with the alternative measure of GD which is basedon the Nei approach, with data from Spolaore and Wacziarg (2009).

The GD approach by Nei considers that genetic differences that are driven bygenetic mutations and drifts. More technical insights into the Nei formula are foundin Cavalli-Sforza et al. (1996, p. 27). The authors have argued that the FST measureis more precise relative to the Nei indicator, if only a few mutations arise duringthe investigated periodicity. This is the case in the contemporary era of human beings.Hence, consistent with Ang and Kumar (2014), whereas, we employ the FST indicatoras our default measurement for human GD, the Nei indicator is also employed for thepurpose of more robustness.

4.1 Alternate measures of genetic distance

InTable 2,we employ an alternativemeasure ofGD, notably: theFSTmeasure ofGD totheUK andNei indicator of GD to theUS. The following findings are established. Firstthe negative relation between GD to frontier nations established earlier is confirmed.Second, the negative correlation to the USA frontier is averagely higher relative to theUK frontier. Third, the significant control variables have the expected signs and theirdiscourses are consistent with those engaged for Table 1.

123


Table 2 Other measures of human capital diffusion

Dependent variable: human capital(1) (2) (3) (4)

GD_UK(fst) −8.893*** −7.234**

(2.806) (2.747)

GD_US(nei) −10.707*** −9.362***

(3.137) (2.987)

Absolute value of latitude −10.978* −5.769 −11.589* −6.572

(6.219) (5.748) (6.403) (5.897)

Mean temperature −0.367** −0.346** −0.364** −0.354**

(0.167) (0.145) (0.163) (0.141)

Longitude 0.038*** 0.111*** 0.039*** 0.117***

(0.009) (0.026) (0.009) (0.026)

GDP per capita (log) 4.711*** 4.070*** 4.790*** 4.161***

(0.956) (0.839) (0.919) (0.809)

Malaria −0.184** 0.007 −0.201** −0.006

(0.081) (0.094) (0.080) (0.091)

Institutions 0.062* 0.071** 0.058* 0.065**

(0.035) (0.032) (0.034) (0.030)

Africa 10.965** 11.643**

(4.426) (4.410)

americas 23.185*** 24.097***

(6.389) (6.306)

Asia 11.622*** 11.602***

(3.067) (2.949)

Europa 16.444*** 16.661***

(4.030) (3.931)

Oceania (dropped) (dropped)

Constant 53.795*** 40.285*** 54.547*** 40.801***

(10.131) (10.003) (9.966) (9.832)

Observations 83 83 83 83

R2 0.760 0.814 0.767 0.820

() Robust standard errors in brackets that are corrected for heteroscedasticity via White. GD_UK(fst): FSTmeasure of genetic distance to the UK. GD_US(nei): Nei measure of genetic distance to the USA*** .01; ** .05; * .1

4.2 Robustness with respect to influential observations

In order to further improve the estimations, our empirical approach follows the M-estimators of Huber (1973) by using: (i) iteratively reweighted least squares (IRWLS),(ii) MM-estimator proposed by Yohai (1987) and (iii) S-estimator proposed bySalibian-Barrera and Yohai (2006) and Rousseeuw and Yohai (1984). As Midi andTalib (2008) have noted, compared to the OLS approach, the advantage of these robust

123


Table 3 Robustness checks

Dependent variable: human capital(1) (2) (3)

GD_US(fst) −8.015**

(3.399)

GD_UK(fst) −6.503**

(2.857)

GD_US(nei) −8.666***

(3.087)

Absolute value of latitude −4.686 −4.469 −5.418

(6.227) (6.220) (6.109)

Mean temperature −0.303** −0.304** −0.330***

(0.124) (0.125) (0.123)

Longitude 0.090*** 0.083*** 0.093***

(0.027) (0.027) (0.026)

GDP per capita (log) 3.887*** 3.882*** 3.878***

(0.937) (0.937) (0.919)

Malaria −0.056 −0.052 −0.049

(0.100) (0.100) (0.098)

Institutions 0.077* 0.082** 0.075*

(0.041) (0.041) (0.040)

Africa 8.110 7.410 8.421

(5.359) (5.303) (5.180)

Americas 18.274** 17.191** 18.973***

(7.177) (7.104) (6.998)

Asia 9.528** 9.121** 9.601**

(3.917) (3.900) (3.817)

Europa 13.281* 12.827* 13.446*

(6.954) (6.941) (6.825)

Oceania (Dropped) (Dropped) (Dropped)

Constant 45.253*** 44.634*** 45.799***

(9.704) (9.632) (9.485)

Observations 83 83 83

R2 0.807 0.807 0.814

GD_US(fst): FST measure of genetic distance to the USA. GD_UK(fst): FST measure of genetic distanceto the UK. GD_US(nei): Nei measure of genetic distance to the USA*** .01; ** .05; * .1

estimators is that theyfix simultaneously any issue arising from the existence of outliersand/or heteroskedasticity (non-constant error variances). Based on the the finding, thesigns and significance of the variables across specifications are consistent with thoseof the preceding tables (Table 3).

123


4.3 More control for outliers using Quantile regressions

In this section, we employ the quantile regression (QR) approach which enables us toassess the role of GDon human capital (IQ) throughout the distributions of human cap-ital. The QR approach is robust to the control of influential outliers and contrary to theOLS technique it does not assume that error terms are normally distributed. The resultsare presented in three panels. The first, second and third correspond respectively to theFSTmeasure ofGD to theUS,FSTmeasure ofGD to theUKandNei indicator ofGD totheUS. The findings are consistent with those established in preceding tables (Table 4).

4.4 Genetic Distance and Cognitive Ability

Consistent with the motivation of the inquiry, we employ an alternative human capitalmeasure as dependent variable. Hence, the cognitive ability (CA) is regressed on thethreeGDmeasures contingent on other covariates. It is interesting to note that cognitivehuman capital has been appreciated in various ways in the human capital literature.Rindermann (2007) has proposed an indicator for CA. The author has defined CA as

Table 4 Quantile regressions

q.25 q.50 q.75

Dependent variable: human capital

GD_US(fst) −13.283** −11.485*** −9.033**

(5.175) (4.025) (4.471)

Constant 47.113*** 58.140*** 63.352***

(14.175) (12.117) (18.676)


Pseudo-R2 0.5400 0.5615 0.5404

GD_UK(fst) −11.683*** −9.177*** −7.978**

(3.086) (2.653) (3.363)

Constant 46.408*** 52.905*** 62.966***

(9.557) (9.017) (16.090)


Pseudo-R2 0.5380 0.5623 0.5438

GD_US(nei) −12.348** −9.934** −9.074**

(5.446) (4.330) (4.047)

Constant 47.899*** 57.530*** 66.458***

(15.062) (13.637) (13.910)


Pseudo-R2 0.5354 0.5627 0.5492

() Robust standard errors in brackets that are corrected for heteroscedasticity via White. GD_US(fst):FST measure of genetic distance to the USA. GD_UK(fst): FST measure of genetic distance to the UK.GD_US(nei): Nei measure of genetic distance to the USA*** .01; ** .05; * .1

123


equal to student intelligence test and achievement examination, measured primarilyfor common macro-social level CA. This ability embodies: (i) the capacity to think(intelligence) and (ii) the rate of true and relevant know-how, the ability of acquireand use know-how (knowledge) (Table 5).

Themainfindings in terms of significance andmagnitude of significance are broadlyin accordance with those established in previous tables. Moreover, the significantcontrol variables also display expected signs and their discourse is consistent with thatengaged for Table 1.

4.5 Genetic distance, human capital and Africa

In this section, we briefly consider some criticisms that have been levelled on AfricanIQ scores. In essence,Wichert et al. have made substantial claims that IQs fromAfricabased on the Lynn/Vanhanen database are relatively low. According to the authors, thelowers scores should be raised to 76 (Wicherts et al. 2010a) and 80 (Wicherts et al.2010b).

Tables 6, 7 and 8 below respectively provide findings corresponding to baselineOLS estimations, control for outliers and use of cognitive ability as alternative humancapital variable. The findings are consistent with those established in preceding tableswith a slim exception, notably: GD to the US based on the Nei measure slightly lowerin terms of negativemagnitude compared theGD to theUKbased on the FST indicator.However, it should be noted that the magnitude corresponding to ‘GD to the US’ basedon the FST measure is substantially higher compared to the UK estimate.

5 Concluding implications

The objective of this study has been to explore correlates of intelligence by focus-ing on genetic distance to frontier nations. The results based on cross-sectional datafrom 167 countries suggest that genetic distance to global frontiers has a nega-tive relationship with human capital. Countries that are genetically far from leadingnations tend to have lower levels of human capital with the negative correlation tothe USA frontier averagely higher relative to the UK frontier. The sign is consis-tent and survives the control of macroeconomic, geographic, institutional and othercovariates.

While the significant correlations confirm the findings of Spolaore and Wacziarg(2009) on the role of genetic distance to theUSAas a barrier to economic development,they are however not consistent with a stream of studies that have concluded that therole of genetic distance disappears after controlling for additional variables, notably: (i)Angeles (2012)who has extended the empirical work of Spolaore andWacziarg (2009)and found that the role of genetic distance dissipates after controlling for dynamicsin population decent and (ii) Campbell and Pyun (2015), who have examined whysocieties are less developed to establish that, contrary to the mainstream narrative, thenexus between GDP per capita and ‘genetic distance from the US’ disappears aftercontrolling for the distance from the equator, sub-Saharan Africa and geography. The

123


Table 5 Other measures of cognitive capital (from Reinderman)

Dependent variable: cognitive ability(1) (2) (3)

GD_US(fst) −12.163***

(4.569)

GD_UK(fst) −10.282***

(3.804)

GD_US(nei) −13.289***

(4.120)

Absolute value of latitude 4.920 5.016 3.853

(8.047) (8.052) (8.131)

Mean temperature −0.239 −0.243 −0.254*

(0.148) (0.150) (0.146)

Longitude 0.140*** 0.135*** 0.143***

(0.029) (0.029) (0.028)

GDP per capita 4.300*** 4.298*** 4.425***

(1.354) (1.359) (1.291)

Malaria 0.009 0.019 0.001

(0.117) (0.117) (0.112)

Institutions 0.061 0.067 0.058

(0.049) (0.050) (0.048)

Africa 7.954 7.870 8.721

(5.825) (5.574) (5.480)

Americas 26.009*** 25.551*** 26.737***

(7.549) (7.364) (7.141)

Asia 11.409** 11.448*** 11.328***

(4.416) (4.262) (4.100)

Europa 14.261*** 14.392*** 14.600***

(5.183) (5.047) (4.933)

Oceania (dropped) (dropped) (dropped)

Constant 34.933*** 33.202*** 34.064***

(11.478) (11.206) (10.920)


R2 0.831 0.831 0.839


implication is that contemporary domestic human capital values may be traceable tothe genetic distance to frontier or developed countries.

The negative role of genetic distance found in this study implies that the patternmaybe more likely U-shaped than hump-shaped. This is essentially because if interactiveregressions were involved in the specifications, corresponding estimated parameters

123


Table 6 OLS estimates of the impact of genetic distance


GD_US(fst) −9.413**

(3.542)

GD_UK(fst) −7.988**

(3.011)

GD_US(nei) −7.925**

(3.217)


(6.888) (6.898) (6.644)

Mean temperature 0.109 0.120 0.089

(0.298) (0.289) (0.305)

Longitude 0.048* 0.047* 0.048

(0.027) (0.027) (0.029)

GDP per capita (log) 2.574** 2.500** 2.770**

(1.116) (1.128) (1.082)

Malaria −0.254** −0.250** −0.246*

(0.121) (0.121) (0.125)


(0.040) (0.040) (0.041)

Constant 58.576*** 57.510*** 55.886***

(13.607) (13.162) (13.267)


R2 0.661 0.663 0.653


will be interpreted as marginal impacts. Therefore, evidence of diminishing marginaleffects of genetic distance on the dependent variable suggests a hump-shaped nexus.It follows that the exploration of non-linear patterns in the underlying nexus couldprovide more region specific insights into how the distance to the USA affects humancapital development in other regions of the world. This challenging research task isnot within the scope of the present inquiry and is shelved for future research.

We have also found that genetic distance to the USA is averagely more importantthan the distance to the UK in the diffusion to human capital. This is logical becausethe USA is a relatively higher frontier nation in terms of human capital. According tothe 2014 Shanghai Academic Rankings of World Universities (ARWU), in terms ofcontribution to knowledge, the USA has 146 universities in the Top 500 whereas theUK has 38 (ARWU 2014).

The negative nexus with genetic diversity on human capital implies that it is morelikely for the relationship to be U-shaped than hump-shaped. While the sign is consis-

123


Table 7 Huber to control for outliers


GD_US(fst) −11.356***

(3.894)

GD_UK(fst) −9.977***

(3.183)

GD_US(nei) −9.749***

(3.388)


(8.133) (7.969) (8.105)


(0.388) (0.377) (0.391)

Longitude 0.060 0.062 0.062

(0.045) (0.044) (0.045)

GDP per capita 2.173* 1.985* 2.427**

(1.170) (1.148) (1.151)

Malaria −0.222 −0.210 −0.206

(0.146) (0.143) (0.146)


(0.061) (0.059) (0.060)

Constant 64.327*** 63.855*** 61.169***

(14.115) (13.493) (13.589)


R2 0.609 0.618 0.610

() robust standard errors in brackets that are corrected for heteroscedasticity via White. GD_US(fst):FST measure of genetic distance to the USA. GD_UK(fst): FST measure of genetic distance to the UK.GD_US(nei): Nei measure of genetic distance to the USA*** .01; ** .05; * .1

tentwithAsongu andKodila-Tedika (2015), the potentialU-shapednexus substantiallyruns counter to Sequeira et al. (2013) who have assessed the relationship betweenhuman capital and ancestral diversity to conclude on a strong hump-shaped relation-ship. Two implications boldly standout. First, like in Cook (2013) on the influence ofgenetic diversity on the effectiveness of vaccines and medicines and Ager and Brück-ner (2013) on the impact of genetic diversity on economic development in the UnitedStates, there is a significant correlation between genetic diversity and development.Second, on the widely discussed hump-shaped nexus established by Ashraf and Galor(2013), the findings may contrast with the recent stream of literature that has con-firmed the hump-shaped nexus, notably: Williams (2013) between genetic diversityand per capita income from a productivity perspective by means of a negative effect ofsocial capital and positive impact on technological productivity; Ashraf et al. (2014),using night time light intensity and Cardella et al. (2015) on financial development.The negative relationship we find in this study has been confirmed by Asongu and

123


Table 8 Cognitive capital and genetic distance

Dependent variable: cognitive ability(1) (2) (3)

GD_US(fst) −17.638***

(5.187)

GD_UK(fst) −14.982***

(4.240)

GD_US(nei) −14.833***

(4.664)

Absolute value of latitude 18.966** 18.405** 19.294**

(8.131) (8.026) (8.530)


(0.384) (0.368) (0.391)

Longitude 0.183*** 0.181*** 0.183***

(0.044) (0.043) (0.048)

GDP per capita (log) 1.862 1.722 2.230

(1.610) (1.609) (1.602)

Malaria −0.194 −0.187 −0.179

(0.165) (0.164) (0.167)


(0.083) (0.083) (0.084)

Constant 49.983*** 48.021*** 44.908**

(17.822) (16.863) (17.233)


R2 0.704 0.709 0.692

() robust standard errors in brackets that are corrected for heteroscedasticity via White. GD_US(fst):FST measure of genetic distance to the USA. GD_UK(fst): FST measure of genetic distance to the UK.GD_US(nei): Nei measure of genetic distance to the USA*** .01; ** .05; * .1

Kodila-Tedika (2015) in a historical analysis of Ashraf and Galor (2013) within anAfrican framework.

We briefly provide alternative explanations to the results. Consistent with Di Tellaand Macculloch (2009), the absence of economic development can also be traceableto historic and bio-geographic phenomena. According to the narrative, educationalquality is endogeneous to socioeconomic features of society that ultimately exert someinfluence on intelligence. The industrial revolution as well as the genetic compositionof the population in the world was heterogeneous. While a correlation between thetwo can be expected, there is no causal linkage.

We set-out in the motivation of this study to investigate the issue raised by Hunt onwhether there is a genetic basis for international disparities in intelligence. “Until then.the question “Is there a genetic basis for international differences in intelligence?”has a simple answer: We do not know” (Hunt 2012. p. 295). We have establishedthat there are linkages from genetic diversity and genetic distance which can explain

123


such differences. However, we have only partially addressed the concern becauseour findings should be treated as correlations, not causalities. Moreover, complexnexuses cannot be captured exclusively with linear correlation analysis (Jaffe et al.2013) and other analytical approaches are needed to confirmed established linkagesand advance the extant literature. The challenging task to advancing scientific inquiryon the underlying highlighted issues is left for future research. Moreover, a usefulcontribution in future inquirieswould be to reverse theAshraf andGalor (2013) results.

Acknowledgments The authors are highly indebted to the editor and referees for the constructivecomments.

Appendices

See Tables 9 and 10.

Table 9 Summary statistics

Variables Obs Mean Std. Dev. Min Max

Human capital 175 84.208 10.853 61.2 106.9

fst_gendis 181 0.474 0.247 0 1

Latitude 180 0.280 0.191 0 0.778

Temperature 114 20.726 6.919 −7.633 28.193

Longitude 174 14.774 68.005 −175.2 177.97

GDP per capita (log) 140 8.871 1.188 5.902 11.173

Malaria 114 5.422 8.115 0 32.203

Institution 114 38.194 22.257 2.977 96.298

Obs observations, Std. Dev standard deviation, Min minimum, Max maximum, Log logarithm, fst_gendisgenetic distance

Table 10 Correlation matrix

(1) (2) (3) (4) (5) (6) (7) (8)

Human capital(1) 1.000

fst_gendis(2) −0.564 1.000

Latitude(3) 0.489 −0.473 1.000

Temperature(4) −0.526 0.293 −0.735 1.000

Longitude(5) 0.227 0.135 0.122 −0.116 1.000

GDP per capita(log)(6) 0.761 −0.530 0.426 −0.366 −0.017 1.000

Malaria(7) −0.563 0.387 −0.388 0.453 −0.021 −0.495 1.000

Institution(8) 0.5853 −0.307 0.325 −0.308 0.071 0.719 −0.240 1.000

reghcfst_gendist_to_uslatit_laportameantemp longitude lgdpcap2005 malaria institutions if africa = 1. roLog logarithm, fst_gendis : genetic distance

123


References

Ager, P., & Brückner, M. (2013). Immigrants’ genes: genetic diversity and economic development in theUS. MPRA Paper No. 51906.

Angeles, L. (2012). Is there a role for genetics in economic development? University of Glasgow. AccessedApril 24, 2015, from http://www.gla.ac.uk/media/media_221391_en.pdf.

Ang, J. B., &Kumar, S. (2014). Financial development and barriers to the cross-border diffusion of financialinnovation. Journal of Banking & Finance, 39, 43–56.

ARWU(2014). Academic ranking ofworld universities. ARWU2014PressRelease. Center forWorld-ClassUniversities at Shanghai Jiao Tong University.

Ashraf, Q., Galor, O., & Klemp, M. (2014). The out of Africa hypothesis of comparative developmentreflected by night time light intensity. Working Paper. Brown University. Department of Economics2014.

Ashraf, Q., & Galor, O. (2013). The “Out of Africa” hypothesis. Human genetic diversity and comparativeeconomic development. American Economic Review, 103, 1–46.

Asongu, S.A., & Kodila-Tedika, O. (2015). Is poverty in the African DNA(Gene)? African Governance andDevelopment Institute Working Paper No. 15/011, Yaoundé.

Asongu, S. A. (2014a). The impact of health worker migration on development dynamics: Evidence ofwealth effects from Africa. The European Journal of Health Economics, 15(2), 187–201.

Asongu, S. A. (2014b). Globalization and health worker crisis:What do wealth-effects tell us? InternationalJournal of Social Economics, 41(12), 1243–1264.

Barber, N. (2005). Educational and ecological correlates of IQ: A cross-investigation. Intelligence, 33,273–284.

Campbell, D. L., & Pyun, J. H. (2015). The diffusion of development: Along genetic or geographic lines?Journal of International Development. doi:10.1002/jid.3059.

Cardella, E., Kalcheva, I., & Shang, D. (2015). Financial development and genetic diversity. Texas TechUni-versity. Accessed April 13, 2015, from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2409500.

Cavalli-Sforza, L. L., Menozzi, P., & Piazza, A. (1996). The history and geography ofhuman genes. Prince-ton: Princeton University Press.

CIA. (2010). Central Intelligence Agency. The World Factbook 2010.Cook, C.J. (2013). The natural selection of infectious disease resistance and its effect on contemporary

health. Division of Health Policy and Management. Yale School of PublicHealth. Accessed April 13,2015, from https://www.aae.wisc.edu/events/papers/DevEcon/2013/cook.09.19.pdf.

Davies, G., Tenesa, A., Payton, A., Yang, J., Harris, S. E., Liewald, D., et al. (2011). Genome-wide associa-tion studies establish that human intelligence is highly heritable and polygenic. Molecular Psychiatry,16(10), 996–1005.

Deary, I., & Johnson, W. (2010). Intelligence and education: Causal perceptions drive analytic processesand therefore conclusion. International Journal of Epidemiology, 39(5), 1362–1369.

Denny, K. (2009). On a dubious theory of cross-country differences in intelligence. Journal of EvolutionaryPsychology, 7(4), 341–346.

Di Tella, R., & Macculloch, R. (2009). Why doesn’t capitalism flow to poor countries? Brookings Paperson Economic Activity, 2009(Spring), 285–332.

Easterly, W. R. (2011). Global development network growth database. The World Bank. http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EXTRESEARCH/0,,contentMDK:20701055~pagePK:64214825~piPK:64214943~theSitePK:469382,00.html. Accessed 14 June2015.

Guedes, J., Bestor, T.C., Carrasco, D., Flad, R., Fosse, E., Herzfeld, M., Lamberg-Karlovsky, C.C., Lewis,C.M., Liebmann, M., Meadow, R., Patterson, N., Price, M., Reiches, M., Richardson, S., Shattuck-Heidorn, H., Ur, J., Urton, G., & Warinner, C. (2013). Is poverty in our genes? A critique of Ashrafand Galor, “The ‘Out of Africa’ hypothesis, human genetic diversity, and comparative economicdevelopment,” American Economic Review, Current Anthropology, 54(1), 71–79.

Huber, P. J. (1973). Robust regression: Asymptotics, conjectures and Monte Carlo. The Annals of Statistics,1, 799–821.

Hunt, E. (2012). What makes nations intelligent? Perspectives on Psychological Science, 7(3), 284–306.Jacob, T. (2010). A compensation for cold weather: Higher IQs. Pacific Stan-

dard. Accessed April 6, 2015, from http://www.psmag.com/books-and-culture/a-compensation-for-cold-weather-higher-iqs-25414.

123

http://www.gla.ac.uk/media/media_221391_en.pdf

http://dx.doi.org/10.1002/jid.3059

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2409500

https://www.aae.wisc.edu/events/papers/DevEcon/2013/cook.09.19.pdf

http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EXTRESEARCH/0,,contentMDK:20701055~pagePK:64214825~piPK:64214943~theSitePK:469382,00.html



http://www.psmag.com/books-and-culture/a-compensation-for-cold-weather-higher-iqs-25414

http://www.psmag.com/books-and-culture/a-compensation-for-cold-weather-higher-iqs-25414


Jaffe, K., Ríos, A., & Flórez, A. (2013). Statistics shows that economic prosperity needs both high scientificproductivity and complex technological knowledge, but in different ways. Universidad Simón Bolívar(USB). Accessed September 3, 2015, from http://www.interciencia.org/v38_02/150.pdf.

Johnson, W., Penke, S., & Spinath, F. M. (2011). Heritability in the era of molecular genetics: Somethoughts for understanding genetic influences on behavioral traits. European Journal of Personality,25, 254–266.

Kalonda-Kanyama I., & Kodila-Tedika, O. (2012). Quality of institutions: Does intelligence matter? Work-ing Papers No. 308. Economic Research Southern Africa.

Kanazaw, S. (2008). Temperature and evolutionary novelty as forces behind the evolution of general intel-ligence. Intelligence, 36, 99–108.

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The worldwide governance indicators: Methodologyand analytical issues. World Bank Policy Research Working Paper No. 5430.

Kiszewski, A., Mellinger, A., Spielman, A., Malaney, P., Sachs, S. E., & Sachs, J. (2004). A global indexrepresenting the stability of malaria transmission. The American Journal of Tropical Medicine andHygiene, 70(5), 486–498.

Kodila-Tedika, O., &Mutascu, M. (2014). Tax revenues and intelligence: Cross-sectional evidence. MPRAPaper No. 57581. University Library of Munich, Germany.

Kodila-Tedika, O. (2014). Governance and intelligence: Empirical analysis from African data. Journal ofAfrican Development, 16(1), 83–97.

Kodila-Tedika, O., & Asongu, S. A. (2015). The effect of intelligence on financial development: A cross-country comparison. Intelligence, 51, 1–9.

Kodila-Tedika, O., & Bolito-Losembe, R. (2014). Poverty and intelligence: Evidence using quantile regres-sion. Economiic Researrch Guarrdiian, 1, 25–32.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., &Vishny, R.W. (1999). The quality of government. Journalof Law, Economics, and Organization, 15(1), 222–279.

Lai, C.-Q. (2006). How much of human height is genetic and how much is due to nutrition? ScientificAmerican. Retrieved from www.scientificamerican.com/article.cfm?id=how-much-ofhuman-height

Lutz, W. (2009). Sola scholaetsanitate: Human capital as the root cause and priority for internationaldevelopment? Philosophical Transactions of the Royal Society, B364, 3031–3047.

Lynn, R., & Vanhanen, T. (2002). IQ and the wealth of nations. Westport, CT: Praeger.Lynn, R., & Vanhanen, T. (2006). IQ and global inequality. Augusta, GA: Washington Summit Publishers.Lynn, R., &Vanhanen, T. (2012). National IQs: A review of their educational. cognitive, economic, political,

demographic, sociological, epidemiological, geographic and climate correlates. Intelligence, 40(2),226–234.

Meisenberg, G., & Lynn, R. (2011). Intelligence: A measure of human capital in nations. Journal of Social,Political and Economic Studies, 36(4), 421–454.

Midi, H., & Talib, B. A. (2008). The performance of robust estimator on linear regressionmodel having bothcontinuous and categorical variables with heteroscedastic errors. Malaysian Journal of MathematicalSciences, 2(1), 25–48.

New, M., Lister, D., Hulme, M., & Makin, I. (2002). A high-resolution data set of surface climate overglobal land areas. Climate Research, 21(1), 1–25.

Nisbett, R. E. (2009). Intelligence and how to get it. New York: Norton.Nordhaus, William D. (2006). Geography and macroeconomics: New data and new findings. Pro-ceedings

of the National Academy of Sciences, 103(10), 3510–3517.Pak, S. (2010). The growth status of North Korean refugee children and adolescents from 6 to 19 years of

age. Economics & Human Biology, 8, 385–395.Rindermann, H. (2007). The g-factor of international cognitive ability comparisons: The homogeneity of

results in PISA, TIMSS, PIRLS and IQ tests across nations. European Journal of Personality, 21,667–706.

Rindermann, H., Kodila-Tedika, O., & Christainsen, G. (2015). Cognitive capital, good governance, andthe wealth of nations. Intelligence, 51, 98–108.

Rousseeuw, P. J., &Yohai, V. J. (1984). Robust regression bymeans of S-estimators. In J. Franke,W.Hardle,& R. D. Martin (Eds.), Robust and nonlinear time series analysis (pp. 256–276). New York: Springer.

Rushton, J. P., & Jensen, A. R. (2005). Thirty years of research on race differences in cognitive ability.Psychology, Public Policy, and Law, 11, 235–294.

Salibian-Barrera, M., & Yohai, V. J. (2006). A fast algorithm for S-regression estimates. Journal of Com-putational and Graphical Statistics, 15, 414–427.

123

http://www.interciencia.org/v38_02/150.pdf

www.scientificamerican.com/article.cfm?id=how-much-ofhuman-height


Schwekendiek, D. (2009). Height and weight differences between North and South Korea. Journal ofBiosocial Science, 41, 51–55.

Sequeira, T., Santos, M., & Ferreira-Lopes, A. (2013). Human capital and genetic diversity. MPRA PaperNo. 49135.

Spolaore, E., & Wacziarg, R. (2009). The diffusion of development. Quarterly Journal of Economics, 124,469–529.

Weil, D.N. (2010). The impact ofmalaria onAfrican development over the LongueDurée. BrownUniversityand NBER.

Wicherts, J. M., Borsboom, D., & Dolan, C. V. (2010). Why national IQs do not support evolutionarytheories of intelligence. Personality and Individual Differences, 48(2), 91–96.

Wicherts, J. M., Dolan, C. V., Carlson, J. S., & van der Maas, H. L. J. (2010a). Another failure to replicateLynn’s estimate of the average IQ of sub-Saharan Africans. Learning and Individual Differences, 20,155–157.

Wicherts, J. M., Dolan, C. V., Carlson, J. S., & van der Maas, H. L. J. (2010b). Raven’s test performance ofsub-Saharan Africans: Average performance, psychometric properties, and the Flynn effect. Learningand Individual Differences, 20, 135–151.

Wicherts, J. M., & Wilhelm, O. (2007). What is the national g-factor? European Journal of Personality,21(5), 763–765.

Williams, B.J. (2013). Does human genetic diversity affect net productivity?. Master’s thesis. AccessedApril 14, 2015, from http://scholarcommons.sc.edu/etd/2419.

Yohai, V. J. (1987). High breakdown-point and high efficiency robust estimates for regression. The Annalsof Statistics, 15, 642–656.

123

http://scholarcommons.sc.edu/etd/2419

Genetic Distance and Cognitive Human Capital: A Cross-National Investigation

Documents

Transcript of Genetic Distance and Cognitive Human Capital: A Cross-National Investigation