Intelligence 38 (2010) 93–100

Contents lists available at ScienceDirect

Intelligence

In Italy, north–south differences in IQ predict differences in income,

education, infant mortality, stature, and literacy
Richard Lynn
University of Ulster, Coleraine, Northern Ireland, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Regional differences in IQ are presented for 12 regions of Italy showing that IQs are highest in
Received 13 January 2009 the north and lowest in the south. Regional IQs obtained in 2006 are highly correlated with
Received in revised form 27 July 2009 average incomes at r = 0.937, and with stature, infant mortality, literacy and education. The
Accepted 27 July 2009
lower IQ in southern Italy may be attributable to genetic admixture with populations from the
Available online 19 August 2009
Near East and North Africa.
© 2009 Elsevier Inc. All rights reserved.
Keywords:
IQ
Income
Infant mortality
Stature
Education
Italy

1. Introduction into the twenty first century (Tabellini, 2009). Numerous

theories have been advanced to explain what has become
Regional differences in per capita income are exception- known as “Italian economic dualism”. An Italian economist
ally large in Italy. The north is as prosperous as central and has written that “there is a huge literature dealing with the
northern Europe, but the south is much poorer. The American so-called “questione meridionale” — the social, cultural and
sociologist Robert Putnam (1993, p. 4) has written that “To economic backwardness of southern Italy” (Felici, 2007, p. 1).
travel from the north to the south in the 1970s was to return Another Italian economist has written that “works dedicated
centuries into the past… many lived in one- and two-room to the southern question would fill an entire library but many
hovels; farmers still threshed grain by hand… transportation of the economists' questions as to the size and causes of
was provided by donkeys that shared their rocky shelters, Italian economic dualism remain unanswered” (Toniolo,
alongside a few scrawny chickens and cats”. 1990, p. 159).
Statistics showing the differences in living standards Despite the attention given to this question, no consensus
between the rich north and the poor south in Italy became has been reached on the answer. Some have raised the
available in the mid-nineteenth century and these differences possibility that “socio-cultural factors” might be involved.
persist to the present day, as shown in Table 1. It is estimated Putnam believes that “the historical record, both distant and
that in 1861 per capita incomes were about 15–20% higher in recent, leads us (like others) to suspect that socio-cultural
the north than in the south (Peracchi, 2008). By 1911 “the factors are an important part of the explanation” (1993, p.
north–south gap had widened appreciably: northern incomes 159). But what are these socio-cultural factors? Putnam
were about 50 per cent higher” (Putnam, 1993, p. 158) and favors the theory of low “civic trust” in the south as a crucial
this difference persisted into the early 1960s (Lutz, 1962) and factor, but concedes that other socio-cultural factors are likely
present. More recently, the Italian economist Guido Tabellini
(2009) has proposed that “culture measured by indicators of
E-mail address: lynnr540@aol.com. individual values and beliefs, such as trust and respect for

0160-2896/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.intell.2009.07.004
94 R. Lynn / Intelligence 38 (2010) 93–100

others, and confidence in individual self-determination” is a

Latitude
determinant of regional differences in economic develop-

46.0
46.0
45.5
43.5
45.0
45.0
44.5
44.5
43.0
41.5
41.0
40.5
40.0
40.0
39.0
37.0
ment in Italy and western Europe.
Years

2001
educ
1.1. IQ and income: individuals
9.0
8.9
8.8
8.6

8.6
9.0
8.7
8.7
9.4
8.5
8.2
8.0
8.2
8.0
8.0
–
In this paper it is proposed that regional differences in
Years

1971
educ

5.7
5.7
5.3
5.2

5.5
5.9
5.2
4.9
5.8
4.6
4.7
4.5
4.6
4.5
4.5
intelligence are the major factor responsible for the regional

–
differences in Italy in per capita income and in the related
Years

1951
educ

variables of stature, infant mortality, and education. This

5.2
5.1
4.6
4.4
5.2
5.1
5.1
4.6
4.1
4.8
3.8
3.6
3.4
3.4
3.5
3.5
hypothesis is derived from the extensive research showing
that intelligence is positively related to income. This rela-
Literacy

tionship holds at three levels of generality: among indivi-

1880

45.9
45.9

38.1
63.0
67.8
55.5
52.2
28.6
41.8
18.1
24.6
20.0
19.1
14.6
19.1
duals, across regions within nations, and across nations. At
–

the level of individuals, the classical study of the contribu-

mortality
1999–02

tion of intelligence to differences in income is Jencks' (1972)

Infant

2.50
3.47
3.17
3.24
3.61
3.86
4.05
3.73
3.76

4.56
5.21
5.88
4.10
5.54
6.62
Inequality. In this he estimated a correlation of 0.31 (corrected
–

for attenuation to 0.35) between IQ and income for a white

male non-farm American sample aged 25–65. He concluded
mortality
1955–7

that this is a causal relationship such that IQ differences make

Infant

38.3
44.9
36.7
35.2
45.4

40.8
36.2
39.8

68.1
62.2
70.4
53.6

57.0
117.5

a significant contribution to differences in incomes. He also

–

concluded that IQ has a heritability of about 50%, and there-

Per cap
income

20,750
23,079
20,338
19,666
22,639
20,519
20,000
22,439
17,070
20,207
15,480
11,862
12,030
13,722
11,595
12,488

fore that genetic factors contribute to income differences.

2003

Jencks' conclusions have been confirmed by a number of

subsequent studies in the United States (e.g. Brown & Reynolds,
Per cap
income

1975; Jencks & Phillips, 1998; Crouse, 1979; Bishop, 1989;

8985

9223

9517

7815

6814
6481
6313
8054
6128
6525
10,930

10,022
11,693
10,964

10,058

10,317
1970

Herrnstein & Murray (1994); Neal & Johnson, 1996; Murray,

1997, 1998; Zax & Rees, 2002; Zagorosky, 2007), and also in
Sweden (Zetterberg, 2004). An analysis of eight studies of the
Stature

178.0
177.1
177.0
175.8
175.2
175.3
175.1
175.4
175.8
175.5
174.0
173.1
173.3
171.6
172.4
172.7
1980

relation between intelligence and income concluded that the

correlation is 0.27 (Ng, Eby, Sorensen & Feldman, 2005). In a
recent meta-analysis of 85 data sets drawn from the United
Stature

171.7
169.2
169.0
169.8
168.1
168.9
169.7
169.2
167.2
167.6
164.2
164.9
164.3
162.1
163.3
164.7
1927

States, the United Kingdom, Norway, Australia, New Zealand,

Estonia, Netherlands and Sweden, Strenze (2007) concluded
that in all studies the correlation between intelligence and
Stature

167.4

167.1
166.2
166.2
167.0
167.7
166.6
164.6
164.6
162.8
162.9
163.1
160.6
162.2
163.3
1910

income is 0.20, in the best studies the correlation is 0.23, and in

35–78 year olds the correlation weighted by sample size is 0.25.
This meta-analysis did not include a recent study of a national
Stature

165.3

164.8
163.7
162.5
162.1
163.2
163.4
161.9
161.8
159.6
160.2
159.7
158.5
158.3
160.2
1855

sample in Britain in which a correlation of 0.37 between IQ

obtained at the age of 8 years and income at the age 43 years
was found for men, and for women the correlation was 0.32
97

92
90
91
90

89
103
101
101

100
100

100
IQ

–
–

(Irwing & Lynn, 2006).

Descriptive statistics for IQs and related variables for Italian regions.

The positive correlation between IQs in childhood and

education

income in middle age suggests that IQ is causal to subsequent

Mean

income. This has been confirmed by studies of sibling pairs

522
512
515

492
502
481
500

447
439
441
438

427
–

–
–

that have shown that siblings with higher IQs have higher
earnings than their lower IQ brothers and sisters (Bound,
Science

Grilliches, & Hall, 1986; Rowe, Vesterdal, & Rodgers, 1999;

534
521
524

499
508
488
510

451
442
447
449

433
–

–
–

Murray, 2002). The use of sibling pairs controls for possible

family and neighborhood effects that might affect both IQ and
Math

513
508
510

487
492
473
494

443
436
435
429

423

income. The likely explanation for the positive correlation

–

–
–

between IQ and income is that those with higher IQs work

Reading

more efficiently (Schmidt & Hunter, 1998) and can supply

519
508
511

491
506
483
496

446
438
440
435

424

goods and services with greater value than those with lower
–

–
–

IQs, and consequently can command higher incomes.

Abruzzi Basilicata
Emilia Romana

1.2. IQ and income: groups

Friuli-Venezia

Puglia Arulia
Lombardy

Campania
Piedmont
Trentino

Sardinia
Tuscany

Calabria
Umbria

At a higher level of generality, positive correlations

Veneto
Region

Liguria
Table 1

Sicily
Lazio

between IQ and per capita income have been reported for

populations in geographical regions within countries. Studies
 R. Lynn / Intelligence 38 (2010) 93–100 95

reporting this have been published for the British Isles, duals, and populations with higher IQs can supply goods and
France, and the United States. The first of these studies was services with greater value than those with lower IQs, and
concerned with IQ differences in 13 regions of the British Isles hence command higher incomes.
in the mid-twentieth century (Lynn, 1979). It was found that In this paper we examine the possibility that the north–
the highest IQ (102.1) was in London, and the lowest IQs in south difference in per capita income in Italy may be due to
Scotland (97.3), Northern Ireland (96.7), and the Republic of differences in intelligence. There is some existing evidence
Ireland (96.0). These regional IQs were positively correlated suggesting this may be the case. In northern Italy, Prunetti
with per capita income at 0.73. They were also positively (1985) has reported a standardization of the Colored
correlated with intellectual achievement indexed by fellow- Progressive Matrices on 500 6–11 year olds in Pisa and the
ship of the Royal Society (r = 0.94), and negatively with surrounding countryside, and Tesi and Young (1962) have
infant mortality (r = −0.78) (Lynn, 1979). It has been shown reported a standardization of the Standard Progressive
subsequently that these regional differences in IQ are strongly Matrices on 2462 11–16 year olds in Florence and the sur-
associated negatively with differences in stature (Boldsen & rounding countryside. Both studies found that the mean IQ in
Mascie-Taylor, 1985). northern Italy is approximately the same as in Britain and
Similar results have been found in France, where regional other countries of northern and central Europe. It has not
differences in intelligence were reported for the mid-1950s proved possible to find normative data for IQs in the south of
by Montmollin (1958). IQs were obtained from 257,000 Italy. However, Peluffo (1962, 1964, 1967) has reported that
18 year old male conscripts into the armed forces, and mean the cognitive development of children in southern Italy and
IQs were given for the 90 French departments. The highest Sardinia (one of the poorest regions and part of the south)
IQs were obtained by conscripts from the Paris region and the lags behind that of children in Genoa in northern Italy and in
lowest by conscripts from Corsica. As in the British Isles, it Switzerland in the performance of Piagetian tasks of the
was shown that these departmental IQs were moderately understanding of conservation and causality. For example,
well positively correlated with average earnings (r = 0.61), 65% of 9 year olds in Genoa succeeded in the conservation of
with intellectual achievement indexed by membership of volume task, compared with only 35% of 9 year olds in the
the Institut de France (r = 0.26), and negatively with infant south. Piagetian tasks can be regarded as tests of intelligence.
mortality (r = 0.30) (Lynn, 1980). A correlation between the two of 0.49 is reported by Jensen
An association between regional IQ and per capita income (1980, p. 674) as the average of 14 studies.
has also been reported in the United States. It has long been The present paper examines three hypotheses. First, that
known that in the United States the populations of the northern IQs in Italy are higher in the north than in the south. Second,
states have higher average IQs than those of the south east that these IQ differences explain most of the per capita
(Kaufman, McClean, & Reynolds 1988). This has been con- income differences. Third, that regional IQ differences in Italy
firmed by McDaniel (2006a) who has calculated the IQs of the are also manifest in variables that can be regarded as
populations of the American states and found that these are correlates or effects of IQs, including stature, infant mortality,
highest in the north eastern states of Massachusetts (104.3), literacy, and years of education.
New Hampshire (104.2) and Vermont (103.8), and lowest in
the southern states of Mississippi (94.2) and Alabama (95.7), 2. Method
and in California (95.5). The McDaniel (2006a) average state
IQs are positively correlated with gross state product per capita Data have been assembled for 12 Italian regions for mean IQ,
(a measure of per capita income) at Pearson's r = 0.28. These average per capita income in euros for 1970 and 2003 given by
state differences in average IQ are partly determined by the the Italian Statistical Office (2008), percentages of the popula-
proportions of blacks and Hispanics, who have lower average tions that were literate in 1880, taken from Tabellini (2007,
IQs than Europeans at approximately 85, 89, and 100, 2009), statures of military conscripts born in 1855, 1910, 1927
respectively (Lynn, 2006). McDaniel (2006a) calculated that and 1980, taken from A'Hearn, Peracchi, and Vecchi (2009) and
state IQs are correlated at −0.51 with the percentage of blacks Arcaleni (2006), infant mortality 1955–57 and 1999–2002,
and −0.34 with the percentage of Hispanics. Similar state taken from Felici (2007), years of education in 1951, 1971 and
differences in IQ using a different methodology have been 2001, taken from Felici (2007), and latitude taken as the
reported by Kanazawa (2006). The different methodologies are approximate geographical mid-point of the regions.
discussed by McDaniel (2006b). The regional IQs have been calculated from the 2006 PISA
At a third level of generality, positive correlations between (Program for International Student Assessment) study of
IQ and per capita income have been reported across nations at reading comprehension, mathematical ability, and science
a magnitude of approximately 0.7 (Lynn & Vanhanen, 2002, understanding administered to 15 year olds in 52 countries
2006). This finding has been confirmed in studies that have (OECD, 2007). Scores on these tests are used as a proxy for
re-examined the data using alternative measures of per capita IQs, adopting the procedure that Rindermann (2007, 2008)
income (Barber, 2005; Dickerson, 2006; Whetzell & McDa- has used for nations. The PISA reading test is a measure of
niel, 2006; Templer & Arikawa, 2006; Hunt & Wittmann, verbal comprehension and the mathematics test is a measure
2008; Gelade, 2008), and by studies that have used national of “quantitative reasoning”, and both of these are major
scores in math, science, and literacy as proxies for intelligence components of general intelligence (e.g. Carroll, 1993, p. 597;
(Rindermann, 2007, 2008; Hunt & Wittmann, 2008). The McGrew & Flanagan, 1998, p. 14–15), while science under-
positive correlation between IQ and per capita income across standing is highly correlated with general intelligence (e.g. at
populations is to be expected from the correlation among 0.68 in the study by Deary, Strand, Smith, & Fernandes, 2007).
individuals, because populations are aggregates of indivi- More generally, numerous studies have shown that tests of
96 R. Lynn / Intelligence 38 (2010) 93–100

educational attainment are highly correlated with intelli-

Infant mortality Infant mortality Literacy Years educ Years educ Years educ
gence at around 0.5 to 0.7, reviewed in Lynn and Mikk (2007),

0.726
2001
and sometimes more highly, e.g. at 0.81 in a recent study of
70,000+ English children whose IQs were measured at the
age of 11 years and educational achievement was measured
at the age of 16 years (Deary et al., 2007). It has been shown

0.908
0.802
1971
that there is a strong genetic correlation between cognitive
ability measured by tests of intelligence and educational, i.e.
the same genes determine ability measured in both kinds of

0.965
0.862
0.888
test (Bartels, Rietveld, van Baal, & Boomsma, 2002; Petrill &

1951
Wilkerson, 2000). These are designated “generalist genes” by
Kovas, Harlaar, Petrill and Plomin (2005) because they

0.924
0.863
0.689
0.842
1880
determine many expressions of cognitive ability including
IQs, math, reading, science, etc. The terms “intelligence” and
“IQ” are used in this paper in the sense of the sum of all
cognitive abilities or global IQ, as measured by intelligence

1999–02
tests such as the Wechslers and the Binets. The PISA tests

− 0.642
− 0.765
− 0.750
− 0.868
− 0.897
measure some mix of g (Spearman's g, the general factor
present in all cognitive abilities), gf (fluid intelligence or
reasoning ability) and gc (comprehension/knowledge) but it
is not considered possible to quantify the contributions of

1955–57
these three factors to the PISA scores.

0.670
0.876 − 0.661
0.936 − 0.631
0.877 − 0.642
0.850 − 0.716
0.899 − 0.676
To calculate IQs for the Italian regions the scores on
reading comprehension, mathematic ability, and science

− 0.841 − 0.775 − 0.671 − 0.716 − 0.661 − 0.684 − 0.718

− 0.861 − 0.774 − 0.666 − 0.758 − 0.807 − 0.745 − 0.826
understanding have been averaged and these averages have

Income
2003
been expressed in standard deviation unit deviations from
the British PISA mean (502, SD = 99). This gives scores for the

0.934

0.902
0.889
0.864
0.774
0.798
Income
Italian regions expressed in standard deviation units in 1970

relation to the British mean. These figures are then converted

to conventional IQs by multiplying them by 15. Thus, the

0.604
0.821

0.666
0.830
0.788
0.796
0.875
Stature

regional Italian IQs are expressed in relation to British mean

1980

IQ of 100, SD 15. Data are missing for Tuscany, Umbria, Lazio

and Calabria.

0.919
0.782
0.842

0.807
0.894
0.858
0.761
0.870
Stature
1927

3. Results
0.965
0.868
0.843
0.841

0.875
0.901
0.831
0.685
0.874
Stature
1910

Descriptive statistics of the data are given in Table 1. The

first column of Table 1 lists the regions in descending order of
0.929
0.964
0.939
0.691
0.815

0.748
0.820
0.782
0.721
0.863
Stature

mean IQs. The region designated Lazio is sometimes referred

1855

to as Latium and consists of Rome and the surrounding

countryside. Columns 2 through 4 give the mean scores of
0.918
0.902
0.925
0.933
0.736
0.937

0.861
0.929
0.871
0.886
15 year olds on reading comprehension, mathematics and 0.963
science understanding for the Italian regions obtained in the
Mean educ IQ

2006 PISA (Program for International Student Assessment)

study. Column 5 gives the average of these three scores.
0.993
0.929
0.894
0.919
0.944
0.694
0.913
− 0.834 − 0.860 − 0.847
− 0.861 − 0.883 − 0.873
0.838
0.905
0.866
0.882
0.964

Column 6 expresses these averages as conventional IQs in

relation to a British mean of 100 (SD 15). Columns 7 through
10 give the statures of cohorts of military conscripts born in
0.998
0.990
0.927
0.877
0.911
0.936
0.678
0.908

0.820
0.893
0.860
0.878
0.956
Science

1855, 1910, 1927 and 1980. Columns 11 and 12 give per

Correlation matrix for variables in Table 1.

capita income for 1970 and 2003 expressed in euros. Columns

0.994
0.998
0.991
0.938
0.897
0.918
0.953
0.677
0.914

0.829
0.899
0.855
0.880
0.961

13 and 14 give rates of infant mortality (deaths of infants in

Reading Math

the first year of life) per 1000 births for 1955–1957 and 1999–
2002 (Felici, 2007). Column 15 gives the percentage of the
0.993
0.993
0.997
0.993
0.918
0.906
0.926
0.936
0.729
0.914
− 0.844
Inf Mrt 1999–02 − 0.867
0.863
0.922
0.883
0.886
0.970

population that was literate in 1880 (Tabellini, 2007).

Columns 16 through 18 give the average years of education
in 1951, 1971 and 2001 (Felici, 2007). Column 19 gives the
Years educ 1951
Years educ 1971
Years educ 2001
Inf Mrt 1955–7

approximate latitude of the regions.

Literacy 1880
Income 1970
Income 2003
Height 1855
Height 1910
Height 1927
Height 1980
Mean educ

The inter-correlations between the variables are given in

Measure

Latitude
Science

Table 2. Correlations higher than 0.38 are statistically signifi-

Table 2

Math

cant at the 0.05 level (2-tailed), and correlations higher than

IQ

0.48 are statistically significant at the 0.01 level (2-tailed).

 R. Lynn / Intelligence 38 (2010) 93–100 97

4. Discussion by Savage (1946). These results corroborate studies showing

that IQ is negatively related to mortality over the life span
There are ten points of interest in the results. First, the IQ (e.g. Batty, Deary, & Gottfredson, 2007; Batty, Deary,&
in the northern regions of Italy measured by the PISA data is Macintyre, 2007; Batty, Shipley, Mortensen, & Deary, 2008;
approximately 100 and therefore about the same as in Britain Gottfredson, 2004).
and other countries of northern and central Europe given in Four, per capita incomes are also highly negatively
Lynn and Vanhanen (2002, 2006). This confirms the results of correlated with rates of infant mortality in 1954–57 (r =
the standardization of the Colored Progressive Matrices in −0.652), and 1999–2002 (r = −0.823). It is proposed that
northern Italy reported by Prunetti (1985) and shows that IQs the principal explanation for these correlations is that
measured by the PISA data and by the Colored Progressive populations with higher per capita incomes have higher IQs
Matrices data are consistent. Regional IQs in Italy decline and this is the main reason why they have lower infant
steadily through the central regions and into the south and mortality, as argued by Gottfredson (2004).
reach a low of 89 in the most southerly region of Sicily. The Five, the ability of populations with high IQs to give their
first hypothesis of this study that there may be a north–south children better nutrition makes them healthier, more resistant
gradient of IQs in Italy is supported and quantified by the to disease and reduces the risk of mortality, and also improves
correlation of 0.963 between regional IQs and latitude. their children's stature. This brings about the high correlations
Two, the second hypothesis of this study is that the between regional IQs and the statures of cohorts of military
north–south gradient of IQs in Italy may explain much of the conscripts born in 1855 (r =0.918), 1910 (r =0.902), 1927
difference in economic development between the north and (r = 0.925), and 1980 (r =0.933). Positive correlations between
south of Italy. This hypothesis is confirmed by the correlation IQ and stature of around 0.25 at the level of individuals have
of 0.937 between regional IQs obtained in 2006 and per been reported in a number of studies (Paterson, 1930; Stoddard,
capita incomes in 2003. This correlation indicates that IQ 1943, p. 200; Laycock & Caylor, 1965). An extensive review of the
differences explain 88% of the variance in per capita incomes evidence for the positive effect of the quality of nutrition on IQs is
across Italian regions. The 14 IQ point difference between the given in Lynn (1990) and confirmed by Benton (2001) and Arija,
most northerly region of Friuli-Venezia (IQ = 103) and the Esparo, Fernandez-Ballart, Murphy, Biarnes (2006). This is a
most southerly region of Sicily (IQ = 89) can be compared with corroboration of the positive feedback loop in which the pop-
the 10.1 IQ point difference in the United States between the ulation IQ is a determinant of the quality of nutrition received by
highest and lowest states (Massachusetts IQ = 104.3; Mis- children, and the quality of nutrition received by children im-
sissippi IQ= 94.2) (McDaniel, 2006), and the 8.1 IQ point proves the children's IQs.
difference in the British Isles between the highest (London and Six, regional IQs in 2006 are highly correlated with the years
the south east IQ = 102.1) and the Republic of Ireland of education of adults in 1951 (r = 0.929), 1971 (r = 0.871) and
(IQ = 96.0) (Lynn, 1979). The large regional IQ differences in 2001 (r = 0.886). The likely explanation for these high correla-
Italy help to explain the large regional differences in per capita tions is that the populations with higher IQs keep their children
income. It is proposed that these correlations between at school longer. This improves the IQs of the children. The
population IQ and per capita income arise through a positive positive effect of years of education on IQ has been shown
feedback loop in which the population IQ is a determinant of in numerous studies recently reviewed by Cliffordson and
per capita income, and per capita income is a determinant of the Gustafsson (2008). However, the regional differences in years
population IQ. Thus, the population's IQ is both a cause and a of schooling amount to only about one year and it is estimated
result of its per capita income. The population's IQ is a cause of that this raises IQ by about 2.5 IQ points (Cliffordson &
its per capita income because individuals and populations with Gustafsson, 2008). Hence regional differences in years of
high IQs are able to work more efficiently than those with low schooling do not account for much of the 14 IQ point difference
IQs and consequently command higher incomes. The popula- between the highest and lowest IQ regions. The positive
tion's IQ is a result of its per capita income because populations relationship between regional IQs and years of education is
with high IQs provide a better environment (better nutrition, best envisioned as another positive feedback loop in which the
health care and education) for the development of the intel- population IQ is a determinant of the amount of education
ligence of their children. This positive feedback loop is known received by children, and the amount of education received by
in behavior genetics as genotype–environment correlation children is a determinant of their IQs.
(Plomin, DeFries & McClearn, 1990). Seven, there is a high correlation between contemporary
Three, the third hypothesis we set out to examine is that regional IQs and the percentage of the population that was
regional IQ differences in Italy are also manifest in variables literate in 1880 (r = 0.861). The likely explanation for this
that can be regarded as correlates or effects of IQs, including high correlation is that the percentages of the population that
stature, infant mortality, literacy, and years of education. This were literate in 1880 was a function of IQs and therefore that
hypothesis is substantiated for all of these phenomena. the regional differences in IQs were present in 1880 and have
Regional IQs are highly correlated negatively with rates of been stable over the period 1880 to 2006. This inference is
infant mortality in 1954–57 (r = −0.847), and 1999–2002 confirmed by the high correlation between contemporary IQs
(r = − 0.873). It is proposed that the explanation for these and the statures of cohorts of military conscripts born in 1855
correlations is that populations with high IQs are more com- (r = 0.918). Stature is a function of nutrition, which is itself a
petent in looking after their babies, e.g. by avoiding accidents, function of IQ.
and are able to give them better nutrition, which makes them Eight, it is an interesting question whether the differences
healthier and more resistant to disease. An association be- in Italian regional IQs were present in earlier historical pe-
tween infant mortality and low IQ mothers has been reported riods. Some useful data bearing on this question have been
98 R. Lynn / Intelligence 38 (2010) 93–100

Table 3 probable that something similar has occurred in Italy,

Numbers of “significant figures” in science born in Italian regions from although most of the migration from the poor south was to
Murray (2003).
the rich north rather than to Rome which is situated in the
Region 1400–1600 1600–1800 1800–1950 center of Italy. Putnam (1993, p. 139) describes migration
from the south from the 1890s onwards as “substantial” and
North 62 76 49
Center 8 18 6 states that between 1958 and 1963 approximately 7% of the
South 3 5 9 population of the south migrated to the north. Some political
scientists have considered the possibility that outward mi-
gration from the south could have been selective. For in-
assembled by Murray (2003, pp. 303–5) who has compiled stance, Putnam (1993, p. 239) writes that “it could be argued
the numbers of “significant figures” (i.e. those who have that selective migration could account for the backwardness
made significant contributions to science, literature, music of the south” although he does not mean selective migration
and art) and their places of birth for the whole of Europe from for intelligence but for “civic mindedness” which he considers
the year 1400 to 1950. His figures for the north, center and the crucial advantage of the northern population.
south of Italy are shown in Table 3. The north lies north of the However, it was not until the 1890s that outward
42nd line of latitude, the center lies between the 41st and migration from the south became substantial (Clark, 1984),
42nd lines of latitude and includes Rome, and the south lies to but the north–south differences in per capita income were
the south of the 41st line of latitude and includes Naples. It present before then and go back centuries (Toniolo, 1990).
will be seen that as early as the years 1400–1600 by far the The north–south differences in IQ also appear to have been
greatest number of Italian “significant figures” have come present before substantial migration from the south to the
from the north (62 out of a total of 73), and the north con- north, if the percentages of the populations that were literate
tinued to produce many more “significant figures” from 1600 in 1880 are adopted as a proxy for IQ, while Murray's data on
to 1800 and from 1800 to 1950. These results suggest that the the far greater numbers of high IQ “significant figures” from
north–south IQ gradient in Italy has been present since 1400 northern Italy as early as the period 1400–1600 given in
and can explain most of the income differences that go back Table 3 suggest that the higher IQ of the north predated by
many centuries. several centuries the mass migration from south to north.
Are other explanations possible for these regional dispa- A possible explanation for the northern regions having
rities? Murray (2003, p. 357) has noted the greater numbers had higher IQs than the southern regions at least from 1880
of significant figures from the north and writes: “Italy's and possibly from 1400 to 1600 is that the populations of the
largest city during the Renaissance was Naples, and yet north and south are genetically different and these genetic
Naples, along with the rest of southern Italy, has almost no differences are related to differences in intelligence. Both
significant figures at all. Why not? A plausible explanation is intelligence and educational attainment are strongly deter-
that for practical purposes Naples and southern Italy were not mined by the same genetic factors (Bartels et al., 2002; Petrill
part of what we think of as Renaissance Italy. They were & Wilkerson, 2000; Kovas et al., 2005), and where there are
controlled by the Spanish Hapsburgs and were politically large population differences in intelligence and educational
and culturally separated from the rest of Italy”. This may not attainment there is a probability that genetic factors are
seem a wholly satisfying answer to the question of why involved. In the case of Italy, it is known that the populations
“Naples and southern Italy were not part of what we think of of the north and south differ genetically. Cavalli-Sforza,
as Renaissance Italy”. Perhaps the true explanation for the Menozzi, and Piazza (1994) are the leading authorities on
dearth of “significant figures” from southern Italy is that the the genetics of human populations, particularly those in Italy.
IQ was lower. They write of the population genetics of Italy that “northern
Nine, Putnam (1993, p. 159) and Tabellini (2007) have Italy shows similarities with countries of central Europe,
proposed that “civic trust” is a determinant of regional dif- whereas central and southern Italy are more similar to Greece
ferences in economic development in Italy and in western and other Mediterranean countries. This corresponds to the
Europe. In this connection it is interesting to note that well-known differences in physical type (especially pigmen-
Rindermann (2008) has shown that “interpersonal trust” is tation and size) between the northern and north-central
correlated at 0.49 with intelligence across 41 nations. It may Italians on the one side and southern Italians on the other”
be that IQ is a determinant of “civic/interpersonal trust” and (1994, p. 277). By “Mediterranean countries” Cavalli-Sforza,
this accounts for the relationship between civic trust and Menozzi and Piazza mean the countries that border the
economic development in western Europe found by Tabellini Mediterranean including those of North Africa and the Near
(2009). East. They note also that the Sardinians are genetically more
Ten, we consider finally the how the regional differences closely related to the Greeks, Lebanese and North African
in IQs in Italy can be explained. The regional differences in IQs Berbers than to central and northern Europeans (Cavalli-
in the British Isles and France can be reasonably explained by Sforza et al., 1994, pp. 78, 274). Subsequent studies have
a tendency of the more intelligent in the poorer regions confirmed the genetic impact of immigration from the Near
to have migrated over the course of centuries from the East and North Africa into southern Italy. For instance, the
provinces to London and Paris, for which evidence is given in Taql, p1 2f2-8-kb allele has a high frequency in the Near East
Lynn (1979, 1980). This would have raised IQs in the capital and North Africa (Morocco, 81.8; Lebanon, 43.7; Tunisia,
cities and reduced IQs in the provinces. In the United States 34.1). The allele is also present but at a lower frequency
also there has been migration from the poor south to the (26.4) in southern Italy, including Sicily. The frequency of the
more affluent north that has likely been selective. It is allele falls to 14.1% in central and northern Italy, but the allele
 R. Lynn / Intelligence 38 (2010) 93–100 99

is rare at 3.8% in France and 3.5% in the Netherlands (Semino, References

Passarino, Brega, Fellous, & Santachiara-Benerecetti, 1996). A
similar gradient is present for the pYα1 allele which has a A'Hearn, B., Peracchi, F., & Vecchi, G. (2009). Stature and the normal distri-
high frequency in North Africa represented by Egypt (85.0), a bution: Evidence from Italian military data. Demography (to appear).
Arcaleni, E. (2006). Secular trend and regional differences in the stature of
lower frequency in southern Italy (includes Sicily and Italians, 1854–1980. Economics and Human Biology, 44, 24–38.
Sardinia) (70.5), falling to 51.4 in north-central Italy, and Arija, V., Esparo, G., Fernandez-Ballart, J., Murphy, M. M., Biarnes, E., & Canals,
falling further to 33.0 in England (Mitchell, Earl, & Fricke, J. (2006). Nutritional status and performance in test of verbal and non-
verbal intelligence in 6 year old children. Intelligence, 36, 141–151.
1997).
Aubet, M. E. (2001). The Phoenicians and the West: Politics, colonies and
The explanation of the genetic difference between northern trade Cambridge: Cambridge University Press.
and north-central Italians and southern Italians together with Barber, N. (2005). Educational and ecological correlates of IQ: A cross-national
investigation. Intelligence, 35, 141–151.
Sardinians is that over the course of many centuries there has
Bartels, M., Rietveld, M. J. H., van Baal, G. C. M., & Boomsma, D. I. (2002).
been considerable immigration into the southern Italian Heritability of educational achievement in 12 year olds and the overlap
mainland, Sicily, and Sardinia by peoples from North Africa with cognitive abilities. Twin Research, 5, 544–553.
and the Near East. Around 750 BC the Phoenicians and later the Batty, G. D., Deary, I. J., & Gottfredson, L. S. (2007). Premorbid (early life) IQ
and later mortality risk: Systematic review. Annals of Epidemiology, 17,
Carthaginians from North Africa colonized all these southern 278–288.
Italian regions (Aubet, 2001; Marston, 2002). Later, “Arabs Batty, G. D., Deary, I. J., & Macintyre, S. (2007). Childhood IQ in relation to risk
occupied Sicily, Sardinia and Southern Italy in the seventh to factors for premature mortality in middle aged persons: The Aberdeen
children of the 1950s study. Journal of Epidemiology & Community Health,
the ninth centuries” (Cavalli-Sforza et al., 1994, p. 261). Sicily 61, 241–247.
was ruled by Arabs from North Africa until 1060 and “during Batty, G. D., Shipley, M. J., Mortensen, L. H., & Deary, I. J. (2008). IQ in late
and after the Arab conquest, large Arab immigration took place” adolescence/early adulthood, risk factors in middle aged persons and
later all-cause mortality in men: The Vietnam experience study. Journal
(M.M.C., 1960, p. 608). Arab rule ended in 1060 when Sicily was of Epidemiology & Community Health, 62, 522–531.
conquered by the Normans, but the Arab occupation left a Benton, D. (2001). Micro-nutrient supplementation and the intelligence of
genetic footprint. These results have been further confirmed by children. Neuroscience and Behavioral Reviews, 25, 297–309.
Bishop, J. H. (1989). Is the test score decline responsible for the productivity
Zalloua, Platt, El Sibai, and Khalife (2008) who report genetic
growth decline? American Economic Review, 79, 178–197.
similarities between male Phoenicians (from present day Boldsen, J. L., & Mascie-Taylor, C. G. N. (1985). Analysis of stature variation in
Lebanon) and males in the populations of Sicily, Sardinia, and a contemporary British sample. Human Biology, 57, 473–480.
Bound, J., Grilliches, Z., & Hall, B. (1986). Wages, schooling and IQ of brothers
the far south of mainland Italy.
and sisters: Do family factors differ? International Economic Review, 27,
The diffusion of genes from the Near East and North Africa 217–230.
may explain why the populations of southern Italy have IQs in Brown, W., & Reynolds, M. (1975). A model of IQ, occupation and earnings.
the range of 89–92, intermediate between those of northern American Economic Review, 65, 1001–1007.
Carroll, J. B. (1993). Human cognitive abilities Cambridge: Cambridge University
Italy and central and northern Europe (about 100) and those Press.
of the Near East and North Africa (in the range of 80–84) Cavalli-Sforza, L. L., Menozzi, P., & Piazza, A. (1994). The history and geo-
(these IQs are given in Lynn, 2006). This also explains the graphy of human genes Princeton, NJ: Princeton University Press.
Clark, M. (1984). Modern Italy, 1871–1982 New York: Longman.
north–south gradient of IQ in Italy in which the regional IQs Cliffordson, C., & Gustafsson, J. -E. (2008). Effects of schooling on intellectual
do not show a clear dichotomy between north and south but performance: Estimates obtained from analysis of continuous variation
rather a gradient in which IQs decline steadily with more in age and length of schooling. Intelligence, 36, 143–152.
Crouse, J. (1979). The effects of academic ability. In J. Crouse (Ed.), Who Gets
southerly latitude. This gradient is shown by the latitudes of Ahead? The Determinants of Economic Success in America. New York: Basic
the regions given in the right hand column of Table 1. Here we Books.
see that the two most northerly regions of Friuli-Venezia and Deary, I. J., Strand, S., Smith, P., & Fernandes, C. (2007). Intelligence and
educational achievement. Intelligence, 35, 13–21.
Trentino have the highest IQs (the same as those in central
Dickerson, R. E. (2006). Exponential correlation of IQ and the wealth of
and northern Europe), and that regional IQs decline progres- nations. Intelligence, 34, 291–295.
sively with latitude to the five most southerly regions of Felici, E. (2007). I divari regionale in Italia sulle base degli indicatori sociali
1871–2001. Revista di Politico Economico, 5, 259–405.
Abruzzi Basilicata, Puglia Arulia, Campania, Sardinia and
Gelade, G. A. (2008). IQ, cultural values, and the technological achievement
Sicily, which have the lowest IQs (89–92). This gradient is of nations. Intelligence, 36, 711–718.
quantified by the correlation of 0.963 between latitude and Gottfredson, L. S. (2004). Intelligence: Is it the epidemiologists' elusive
IQ given in Table 2, suggesting that 93% of the variance in “fundamental cause” of social class inequalities in health? Journal of
Personality and Social Psychology, 86, 174–199.
IQs across Italian regions is explained by latitude. The pro- Herrnstein, R. J., & Murray, C. (1994). The bell curve: Intelligence and class
posed explanation for this gradient is that with increasingly structure in American life New York: The Free Press.
higher latitude the proportion of the Italian population with Hunt, E., & Wittmann, W. (2008). National intelligence and national pros-
perity. Intelligence, 36, 1–9.
European ancestry and high IQs increases steadily until in the Irwing, P., & Lynn, R. (2006). The relation between childhood IQ and income
far north the IQ reaches the average (100) of the populations in middle age. Journal of Social, Political and Economic Studies, 31,
of central and northern Europe. This explanation also ac- 191–196.
Italian Statistical Office. (2008). Regio-It 1970–2003 Rome: Italian Statistical
counts for the IQs of around 90 for several countries in the Office.
Balkans shown in Lynn (2006) and confirmed for Serbia Jencks, S. (1972). Inequality. London: Penguin.
by Rushton and Čvorović (2009), whose populations are of Jencks, C., & Phillips, M. (1998). The Black-White Test Score Gap. Washington,
D.C.: Brookings Institution.
partly European and partly Near Eastern origin (Semino et al.,
Jensen, A. R. (1980). Bias in mental testing London: Methuen.
1996). All these data taken together indicate that the north– Kanazawa, S. (2006). IQ and the wealth of states. Intelligence, 34, 593–600.
south gradient of intelligence in Italy has a genetic basis going Kaufman, A. S., McClean, J. E., & Reynolds, C. R. (1988). Sex, race, region and
education differences on the 11 WAIS-R subtests. Journal of Clinical
back many centuries, and hence predicts the social and
Psychology, 44, 231–248.
economic differences documented in the nineteenth century Kovas, Y., Harlaar, N., Petrill, S. A., & Plomin, R. (2005). Generalist genes and
up to the present day. mathematics in 7 year old twins. Intelligence, 35, 473–489.
100 R. Lynn / Intelligence 38 (2010) 93–100

Laycock, F., & Caylor, J. S. (1965). Physiques of gifted children and their less Peracchi, F. (2008). Stature and economic development in Italy. American
gifted siblings. Child Development, 35, 63–74. Economic Review, 98, 475–481.
Lutz, V. (1962). Italy: A study in economic development Oxford: Oxford Petrill, S. A., & Wilkerson, B. (2000). Intelligence and achievement: A behav-
University Press. ioural genetics perspective. Educational Psychology Review, 12, 185–199.
Lynn, R. (1979). The social ecology of intelligence in the British Isles. British Plomin, R., DeFries, J. C., & McClearn, G. E. (1990). Behavioral genetics New York:
Journal of Social and Clinical Psychology, 18, 1–12. Freeman.
Lynn, R. (1980). The social ecology of intelligence in the France. British Prunetti, C. (1985). Dati normative del test PM47 colored sur un calpioni di
Journal of Social and Clinical Psychology, 19, 325–331. bambini italiani. Bollettino di Psicologia Applicata, 176, 27–35.
Lynn, R. (1990). The role of nutrition in secular increases of intelligence. Putnam, R. D. (1993). Making democracy work Princeton, NJ: Princeton
Personality and Individual Differences, 11, 273–285. University Press.
Lynn, R. (2006). Race differences in intelligence: An evolutionary analysis Rindermann, H. (2007). The g-factor of international cognitive ability com-
Athens, GA: Washington Summit Publishers. parisons: The homogeneity of results in PISA, TIMMS, PIRLS and IQ tests
Lynn, R., & Mikk, J. (2007). National differences in intelligence and across nations. European Journal of Personality, 21, 667–706.
educational attainment. Intelligence, 35, 115–121. Rindermann, H. (2008). Relevance of education and intelligence at the national
Lynn, R., & Vanhanen, T. (2002). IQ and the wealth of nations Westport, CT: level for the economic welfare of people. Intelligence, 36, 127–142.
Praeger. Rowe, D., Vesterdal, W., & Rodgers, J. (1999). Herrnstein's syllogism: Genetic
Lynn, R., & Vanhanen, T. (2006). IQ and global inequality Athens, GA: and shared environmental influences on IQ, education and income.
Washington Summit Publishers. Intelligence, 26, 405–423.
Marston, E. (2002). The Phoenicians New York: Benchmark Books. Rushton, J. P., & Čvorović, J. (2009). Data on the Raven's Standard Progressive
M.M.C. (1960). Sicily. Encyclopedia Britannica Chicago: William Benton. Matrices from four Serbian samples. Personality and Individual Differences,
McDaniel, M. A. (2006). State preferences for the ACT versus SAT complicates 46, 483–486.
inferences about SAT-derived state IQ estimates: A comment on Kanazawa Savage, S. W. (1946). Intelligence and infant mortality in problem families.
(2006). Intelligence, 34, 601–606. British Medical Journal, 19, 86–87 Jan.
McDaniel, M. A. (2006). Estimating state IQ: Measurement challenges and Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection
preliminary correlates. Intelligence, 34, 607–619. methods in psychology: Practical and theoretical implications of
McGrew, K. S., & Flanagan, D. (1998). The intelligence test desk reference 85 years of research findings. Psychological Bulletin, 124, 262–274.
Boston: Allyn and Bacon. Semino, O., Passarino, G., Brega, A., Fellous, M., & Santachiara-Benerecetti, A. S.
Mitchell, R. J., Earl, L., & Fricke, B. (1997). Y-chromosome specific alleles and (1996). A view of the Neolithic demic diffusion in Europe through two Y
haplotypes in European and Asian populations: Linkage disequilibrium chromosome-specific markers. American Journal of Human Genetics, 59,
and geographic diversity. American Journal of Physical Anthropology, 104, 964–968.
167–176. Stoddard, G. D. (1943). The meaning of intelligence New York: Macmillan.
Montmollin, M. (1958). Le niveau intellectuel des recrues du contingent. Strenze, T. (2007). Intelligence and socioeconomic success: A meta-analytic
Population, 13, 259–268. review of longitudinal research. Intelligence, 35, 401–426.
Murray, C. (1997). IQ and economic success. Public Interest, 128, 21–35. Tabellini, G. (2007). Personal communication.
Murray, C. (1998). Income Inequality and IQ. Washington, D.C.: AEI Press. Tabellini, G. (2009). Culture and institutions: Economic development in the
Murray, C. (2002). IQ and income inequality in a sample of sibling pairs from regions of Europe. Journal of the European Economic Association, Papers
advantaged family backgrounds. American Economics Association: Papers and Proceedings (to appear).
and Proceedings, 92(2), 339–343. Templer, D. I., & Arikawa, H. (2006). Temperature, skin color, per capita
Murray, C. (2003). Human accomplishment New York: Harper Collins. income, and IQ: An international perspective. Intelligence, 34, 121–139.
Neal, D. A., & Johnson, W. R. (1996). The role of premarket factors in black- Tesi, G., & Young, B. H. (1962). A standardisation of Raven's Progressive Matrices.
white wage differences. Journal of Political Economy, 104, 98–123. Archive de Psicologia Neurologia e Psichiatrica, 5, 455–464.
Ng, T. W. H., Eby, L. T., Sorensen, K. L., & Feldman, D. C. (2005). Predictors of Toniolo, G. (1990). An economic history of liberal Italy, 1850–1918 New
objective and subjective career success: A meta-analysis. Personnel Psychology, York: Routledge.
58, 367–408. Whetzell, D., & McDaniel, M. A. (2006). Prediction of national wealth. Intelligence,
OECD (Organization for Economic Co-operation & Development). (2007). 34, 449–458.
PISA 2006 competencies for tomorrow's world Paris: OECD. Zagorosky, J. L. (2007). Do you have to be smart to be rich? The impact of IQ
Paterson, D. G. (1930). Physique and intellect New York: Century. on wealth, income and financial distress. Intelligence, 35, 489–501.
Peluffo, N. (1962). Les notions de conservation et de causalité chez les Zalloua, P. A., Platt, D. E., El Sibai, M., & Khalife, J. (2008). Identifying genetic
enfants provenant de différents milieux physiques et socioculturels. traces of historical expansions: Phoenician footprints in the Mediterra-
Archives de Psychologie, 38, 275–291. nean. American Journal of Human Genetics, 83, 633–642.
Peluffo, N. (1964). La nozione di conservazione del volume e le operazioni di Zax, J. S., & Rees, D. I. (2002). IQ, academic performance, environment and
combinazione come indice di sviluppo del pensiero operatorio in soggetti earnings. Review of Economics and Statistics, 84, 600–614.
appartenenti ad ambienti fisici e socioculturali diversi. Rivista di Psicologia Zetterberg, J., 2004. The impact of cognitive and non-cognitive ability on
Sociale, 11(99-1), 32. earnings – Swedish evidence. Unpublished.
Peluffo, N. (1967). Culture and cognitive problems. International Journal of
Psychology, 2, 187–198.