Soviet Lessons on Ethnic Disparities in STEM

The main tenet of DEI is attribution of all significant group disparities to deep-seated, multi-faceted, systemic oppression. This claim is hard to test thoroughly, even in the restricted sphere of STEM, absent socialist revolution. Fortunately, the Soviet Union provided the world’s longest-running, best-documented experiment in promoting STEM participation and countering group disparities on a thoroughly anti-capitalist foundation.

Results were mixed. The main achievements were widespread STEM education on a generally strong analytic foundation, with special gains for women and previously mistreated minorities. The main shortcomings were the semi-feudal treatment of farm workers and significant under-representation of central Asians. The huge over-representation of Jews boosted Soviet STEM until Jews emigrated to Israel and the US. Key lessons include the usefulness of (1) broad access to good-quality instruction, (2) recognition that bottom-up social influences often defy top-down efforts at equalization, and (3) veneration of outstanding performance regardless of group identifiers.

Unlike woke leftists today, Marx, Engels and Lenin were wholehearted supporters of STEM. They clearly distinguished the “forces of production”, which embrace technological know-how and the hardware that embeds it, from the social “relations of production” transferring the fruits of human labor to others. They prized natural science’s contributions to technological progress, refused to deprecate scientists for their privileges, and framed their own research as scientific analysis of society. They viewed STEM as an invaluable ally, as in Lenin’s aphorism “Communism is Soviet power plus electrification of the whole country”.

Accordingly, once the Bolsheviks secured power, they set a high priority on massive extension of STEM training. This was less liberating than their fans hoped, due to the tension between laws of nature that willful humans can master and laws of nature that override human will. Soviet authorities became increasingly uncomfortable with the latter, especially after their wrong-headed agricultural policies ran aground in the 1920s and the massive industrialization program in the 1930s caused mammoth waste. Biology and economics were stunted thereafter, including the notorious Lysenkoist campaign to cancel Mendelian genetics. But math and the hard sciences generally flourished, with attempts to cancel quantum physics rebuffed so as not to jeopardize nuclear bomb development.

The number of higher education enrollments more than tripled between 1928 and 1940, with STEM students accounting for 60% of enrollments. (The corresponding increases in the US were barely one-third, though from a much larger base, and with less focus on STEM.) While innovation slowed after World War II due to a mix of chronic shortages of critical components and accumulating inertia in huge state firms, core training remained strong, as evidenced by both Nobel Prizes and the achievements of Soviet émigré scientists. Soviet STEM instruction in primary and secondary schools was generally both more advanced and delivered more broadly and effectively than its US counterparts, facilitated by Soviet practices holding parents accountable for students’ misbehavior. Soviet youth dominated International Math and Physics Olympiads in the 1980s.

The best-documented evidence of high Soviet STEM quality involves the emigration of over 1,000 Soviet mathematicians (nearly 10% of the previous total) after the collapse of the Soviet Union. Comprehensive research found that émigrés tended to dominate their fields. They shrank publication opportunities for domestic rivals.

The Soviet Union also pioneered in encouraging STEM education for women. Girls were not segregated in classes or assigned a less demanding curriculum. By the 1980s, about half of Soviet engineers were women. However, women were concentrated at middle or lower levels of STEM. In 1989, women reportedly comprised only 2% of members of the elite USSR Academy of Sciences. I do not have the data to analyze the disproportion.

The Soviet Union also narrowed differences by nationality. “Nationality” in Soviet parlance was what Americans call “ethnic group” but it was defined far more clearly and tracked more thoroughly. Nationality was inscribed at birth to match the father’s or mother’s, with limited rights to change at age 16. No mixed classifications or strategic reclassifications were allowed. No nationalities were subdivided by color. No nationalities were fused to make a larger unit, although all nationalities were encouraged to view themselves as parts of a greater Russian-speaking Soviet whole. In comparison, US racial/ethnic/color classification is an inconsistent, overly amalgamated, politically manipulated mess.

Soviet tracking by nationality was facilitated by the concentration of each nationality in a home region. The Soviet Union was comprised of 15 largely national republics, with Russia and a few other republics containing some autonomous regions for smaller nationalities. Soviet authorities sought to promote variants on home rule without encouraging fissile tendencies, and while they ultimately lost control, they did manage to promote relatively uniform standards in STEM education with one huge exception.

The exception involved the distinctly second-class treatment of agricultural workers. Mistrust of “peasants”, which peaked in the brutal collectivization in eastern Ukraine and subsequent Holodomor, always preponderated. They were largely tied to their collective or state farms, with limited opportunities to travel outside, much less move permanently. This was enforced through a system of mandatory internal passports introduced in 1933 and which revived many of the semi-feudal aspects of the tsarist passport system that the Bolsheviks had proudly abolished in 1917. Without passport authorization and associated residency permits, urban immigrants were illegal aliens and generally treated more harshly than their Western counterparts. This limited rural incentives to gain STEM skills and even opportunities to gain them, since few skilled teachers wanted to live in the countryside.

With that background, let’s skip forward to the last years of the Soviet Union, and examine some statistics that I gathered from Soviet yearbooks and were published here in Appendix II-4 of Volume 1. Table 24 lists numbers of scientists of a given nationality at end-1987 per 10,000 nationals in 1989. Per Soviet definitions, scientists include small numbers of social scientists and humanities scholars.

In 1989, scientists comprised 0.53% of the Soviet population or 1.00% of the civilian labor force. Scientist shares ranged from 0.12% for Tajiks to 0.71% for Russians. I have charted the shares below as percent of the Russian benchmark.

The Armenian and Georgian shares were close to the Russian share, and Estonia wasn’t far behind. Latvians and Lithuanians lagged some, Ukrainians and Belarusians lagged more, but none fell as much as half below the Russian norm. The rest all did, most by 75% or more. They are all Muslim peoples except the Moldovans, who were Romanians annexed by the Soviet Union and forced to switch their Latin alphabet to Cyrillic.

History warns against treating the group differences as immutable. Moldovans descended from relatively sophisticated Roman settlers, Central Asians straddled the cosmopolitan Silk Road, and Slavs were relatively isolated backwoods farmers whose capture by technologically superior Muslim raiders bequeathed us the word “slave”. Yet the differences defied decades of Soviet efforts to eliminate them. Why?

It is hard to infer causality as many measures moved in tandem. The relation I found most interesting was the 0.88 between log scientist share and log urban share for the home republic. A regression associated each 1% increase in urbanization with an average 2.2% increase in scientist share. Yet no cause should directly induce more than a 1:1 relation. I infer that industrialization, urbanization, STEM participation and other factors resonate into what we might call a culture of productivity, even though I’m not sure what that entails. The Soviets never figured out what that entails either; what they tried to remold was less plastic than they hoped.

An alternative critique is that the Soviets failed to close the gaps because they were too imperialist, too Russian, and in Central Asia too European and non-Muslim. Here history provides us a simple test, since the 15 republics have now been independent for over 30 years and have become more nationally homogeneous. Our best single measure of economic development is GDP per capita in terms of purchasing power parity (PPP). The correlation between the logarithm of this measure and log urban share in 1990 is 0.89, and each 1% increase in urbanization in 1990 was associated with an average 2.3% increase in real per capita income. Overall, the differentials have widened considerably. GDP per capita in PPP terms averages over $40,000 in the Baltic states but under $18,000 in Moldova and all Central Asian states outside of mineral-rich, highly Russified Kazakhstan. Kyrgyzstan at $7,000 and Tajikistan at $5,500 are comparable to Pakistan at $4,800.

Ukraine is the biggest outlier. Its GDP per capita in PPP terms is $16,300, 1% less than Moldova’s. The main cause is the war with Russia. In 2013, Ukraine’s GDP per capita in PPP terms was 10% higher than it is now and 60% higher than Moldova’s. However, Ukraine was lagging other east European countries even before the war due to a mix of its own poor policies and EU reluctance to take Ukraine in.

Jewish Overrepresentation in STEM

Only one nationality outdid Russians in terms of percentage working as scientists in the Soviet Union, and its edge was enormous. In 1987, 4.24% of Jews were classified as scientists, six times the Russian share. At the higher level of kandidati/doktori, roughly equivalent to US PhDs and senior scholars, the Jewish share of 2.16% was ten times the Russian share and seven times the shares for Estonians, Georgians, or Armenians.

As noted in my recent essay on representation in basketball and football, disparate impact in specialties is most easily modeled as the tails of various Gaussian bell curves on some unknown metric of aptitude, attitude, and access. The default treatment assumes equal variances. In that case, differences can be summarized as shifts in the ratio of mean to standard deviation, which I call Margin. The bar chart below displays estimated Margins for scientists, assuming that the Soviet labor force participation rate of 53% applied to every nationality (I don’t have data on differences).

The Jewish Margin was 0.8 above the Russian benchmark, while Central Asian and Moldovan Margins were 0.4-0.6 below. The maximum difference in Margins was 1.4, slightly less than the difference between Blacks and Asians in US basketball and football. Given the understandable sensitivities, let me make clear that none of these Margins connote some overall superiority. Both biological and social evolution tend to achieve excellence in one sphere by sacrificing excellence in another, creating what Thomas Sowell calls “reciprocal advantages”. I do not claim to know the sources of these advantages, nor do I claim that advantages are permanently fixed by either genes or socialization.

I do claim that the Jewish statistical Margin in STEM was non-trivial for Soviet technological progress. Soviet authorities evidently agreed, as they could have easily capped Jewish representation much lower and did so in many other areas. However, poor treatment including chronic shortages encouraged many Soviet Jews to emigrate, taking valuable STEM talents elsewhere. Between 1971 and end-1991 when the Soviet Union was dissolved, roughly 600 thousand Soviet Jews emigrated, about 30% of the 1970 base. Most went to Israel and contributed greatly to its subsequent productivity gains; Israel now has higher per capita GDP in PPP terms than any former Soviet republic. The next most important destination was the US; prominent Soviet Jewish émigrés in STEM include Sergey Brin, co-founder of Google and frequent Heterodox STEM contributor Anna Krylov.

Jews continued to emigrate from the former Soviet Union after it collapsed, as did many prominent non-Jewish scientists. Some of this emigration was prompted by post-Soviet disorder, some by Soviet-style practices that still lingered. This doubtless has contributed to a huge STEM anomaly. Russia exhibits by far the world’s sharpest contrast between the profusion of tertiary STEM education and the rate of patent applications. Russia ranks among the world’s top five in tertiary-educated manpower but accounted for just 0.3% of foreign patents awarded by the US Patent and Trade Office and less than 0.5% of international patent applications to the UN’s World Intellectual Property Organization.

Learning from Others’ Experience

Second marriages are said to represent the triumph of hope over experience. As I am married thrice, I have no right to slur the many well-intentioned hopes for narrowing group disparities in STEM. But if the hopeful want to make a truly positive impact, it behooves them to learn from Soviet experience. Broadening access to high-quality STEM education works. Setting high standards and rewarding their achievement works. Dumbing down instruction for lagging groups doesn’t. Insisting on quotas doesn’t. If you truly respect Diversity, and if you truly aim for Inclusion, you might provide roughly Equal opportunities, but you will never achieve the roughly Equal results that so-called “Equity” demands.