What are the reasons behind the wealth and poverty of nations? Since this question has exercised the minds of thinkers from Adam Smith to David Landes, Jared Diamond and Richard Lynn, I decided to take a look at it myself. I came to the conclusion that while geography, macroeconomic policies, resource windfalls and the microeconomic environment do play important roles, by far the most important factor is the state of a country’s human capital – things like literacy rates, school life expectancy and performance on international student assessments.
This is not a new idea. A Goldman Sachs report, Dreaming with BRICs, noted that:
Many cross-country studies have found positive and statistically significant correlations between schooling and growth rates of per capita GDP—on the order of 0.3% faster annual growth over a 30-year period from an additional one year of schooling.
However, I think education is much more central to this. The problem with using years of schooling as a yardstick is that in many middle-income countries, like Argentina, Turkey or Brazil, the amount of schooling is converging to that of the developed world, but the quality isn’t. This is attested to by their performance on international student assessments like PISA. For instance, in the 2006 PISA Science assessment, only 15.2% of Brazilians were at Level 3 or higher (the threshold for moving beyond purely linear problem-solving), compared with 47.6% of Russian, 51.3% of American and 66.9% of Australian students. Is it really then surprising to discover that from 1997 to 2007 purchasing power GDP per capita in Brazil and Russia, both medium-income countries, has grown at 1.3% and 6.0%, respectively, i.e., that Russia is playing the game of economic catch-up much more successfully?
I collected educational statistics on 65 countries and used a formula to work out a Human Capital Index (HCI), relying on three main stats – the literacy rate, PISA/TIMSS/PIRLS performance and tertiary attainment. I then compared this with their purchasing power GDP per capita and its average growth rate for 1997-2007. The results are in the table below.
Source: CIA World Factbook for literacy rates, GDP per capita, 2007 GDP per capita growth; PISA 2006 executive summary for Maths, Science, Reading stats (note: China, India are guessed); eighth-grade Maths, Science performance from Highlights from TIMSS 2003; fourth-grade Reading from PIRLS 2006; tertiary enrolment, 1997-2006 GDP per capita growth from World Bank. M1 and M2 refer to the mean of a country’s scores across PISA and TIMSS/PIRLS, respectively, divided by average across all participating countries; if there’s a figure for both M1 and M2, then M3 = (M1+M2) / 2; if not, M3 = M1 or M2, as appropriate. HCI = literacy * M3 * tertiary enrolment ^ (1/3). Figures in italics are those for which I’ve had to use other sources.
The chart below shows how closely educational capital and wealth correlate in 2007. We see an inverse square relationship or possibly a kind of S-curve with two inflection points.
Let us note a few things:
1. Notice that out of the 30 countries with an HCI below 0.80, with the marginal exception of Saudi Arabia, not a single one had a GDP per capita exceeding 20,000 $. In fact, this chart understates the pattern, because the detailed educational stats produced by programs like PISA and TIMSS typically don’t include low-income countries, where human capital is going to be typically very low.
2. Practically all outliers can be explained by one of two things – resource windfalls and socialist legacies. Among all countries with HCI’s of less than 0.80, the top outliers are all big oil or minerals exporters. This artificially inflates their GDP’s, varying in extent from Iran, South Africa and Mexico (where oil and minerals production co-exists with a burgeoning manufacturing base) to Saudi Arabia and Botswana (which are dominated by hydrocarbons and diamonds, respectively). The latter are green and the former are cyan/green. The reason Norway is the world’s most affluent country also comes down to the oil boost (its human capital is unremarkable by average OECD standards).
3. Similarly, the vast majority of low outliers come from the former Communist bloc. East European former satellites are red, post-Soviet countries are dark red and Russia is black. Note how far off the vast majority of them are from where their HCI seems to indicate they should be (the exceptions being Azerbaijan, the Czech Republic, Slovakia and to a lesser extent Bulgaria and Romania). Poland, Hungary and the post-Soviet world are currently well below their potential. The explanation is that these countries have spent much of this century languishing under central planning with all its inefficiencies and contradictions and thereafter being subjected to a decade-long period of brutal restructuring, before normal economic growth could again resume towards the mid to late 1990’s. Meanwhile, the Communist emphasis on education payed off bigtime, as human capital is comparable to that of much richer countries. It is worth noting also that the gap between potential and actual is greatest in the former Soviet bloc, presumably because the socialist legacy was strongest there (no folk memory of pre-WW2 capitalism, no Visegrad-like experiments with creeping capitalism, etc). The big exception is Azerbaijan, which has an oil windfall; Russia, the other country in a similar position, doesn’t replicate this because its potential (based on human capital) is much higher – its socialist legacy outweighs its resource windfall.
Below is a table with three indicators for each country – their actual GDP, potential GDP based on average macroeconomic/microeconomic policies and potential GDP based on optimal policies.
|Actual||Mean Potential||Max Potential|
4. There are five other low outliers – Slovenia, Taiwan, New Zealand, Finland and South Korea, of which the latter two are particularly big. The explanations aren’t as clear-cut here, but I’ll throw a few around. Finland is a northern country covered in permafrost that inflates construction, energy and transport costs, while New Zealand has a small population (i.e. a small market) far removed from the arteries of world trade. Slovenia has a socialist legacy. Taiwan and Korea are very densely populated, which impacts negatively on the productivity of the retail and construction sectors. Singapore is a top outlier, much richer than warranted by its human capital – I suppose that’s because of its status as a major trade hub. Not as convincing? I kind of agree. The above explanations do not have the all-encompassing unity and simplicity of the socialist legacy or the resource windfall. Which is why it’s time for us to talk about economic growth rates.
Speaking of which – see below.
Countries are marked by GDP / capita growth rates from 1997 to 2007. The colors go as follows: white (1.0-1.9%); yellow (2.0-2.9%); orange (3.0-3.9%); red (4.0-5.9%); dark red (6.0%-7.9%) and black (8.0%-14.9%). Also, GDP per capita figures (on the y-axis) are for 1997 – this is because what we are interested in is the influence of education levels on future growth, which we know for the period from 1997 up until today. Unfortunately, educational stats for 1997 will be much less comprehensive (PISA and TIMMS embraced much fewer countries then), plus it would take a lot of time digging them up – hence I made a rough assumption that they were the same as for 2007. Actually, the HCI is based on a collation of different stats from the 2000-2005 period).
One thing that immediately stands out is how countries that are below their potential tend to have much higher growth rates than those on or above their potential. In other words, excluding chaotic and cyclical trends, economies tend to a steady state depending on the level of their human capital. Thus, the blue and cyan groups above tend to have equal growth in GDP per capita (although since the population grows in most cyan countries, absolute GDP growth will be larger), meaning that the cyan countries aren’t converging and should not converge economically, no matter their degree of openness or transparency.
The most glaring exceptions are typically due to oil booms and the like, which not only increase GDP in of themselves but also fuel consumption splurges. The purple group is the most interesting, which mainly encompasses relatively well-educated post-Communist countries. Unshackled from the chains of socialism, they are now growing very quickly due to the huge ‘potential gap’ that exists between their human capital and development level. (Picture this as a question of heat diffusion – the greater the difference, the greater the pressure to close it). The green countries must massively increase their investment into education if they want to join the development bandwagon.
Now for some questions and answers:
What does this mean for development strategies?
Policy-makers must realize that education is the elixir of economic growth. Individual incomes grow due to the introduction of new technologies, which increase total factor productivity. (Granted, it is possible to increase labor participation rates and increase savings – but only up to a point. There are only so many people in any workforce, while investment is subject to diminishing returns. The only long-term development model is continuous technological adaptation, which is recognized by the exogenous growth model). However, a country’s ability to take advantage of technology diffusion is governed by its educational levels (e.g., the illiterate will have little need for a computer).
It is not enough, however, to enroll every annual cohort, give them ten years of public schooling and consider the task done, as is the pattern in much of the developing world today. School life expectancy in much of Latin America, the Middle East, South Asia and China might be drawing close to Western standards – the same cannot be said, however, for its quality, as these international student assessments reveal. Secondly, tertiary enrollment there (15-30%) remains far below Western and post-Communist standards (50-80%). To bridge the gap into society-wide participation in the ongoing technological revolutions, developing countries must make efforts to remedy the two above problems. Massive labor and capital infusions will only take you so far. Once a country achieves a GDP per capita of around 5,000 – 10,000$, growth becomes a matter of increasing productivity in the services and higher-tech manufacturing sectors. This requires annual cohorts of well-educated workers.
I am not denying that there are many other conditions that have to be fulfilled for economic convergence to happen. Goldman Sachs, for instance, has compiled a Growth Environment Index that takes into account thirteen factors: inflation, government deficits, external debt, investment, openness, years of schooling, life expectancy, political stability, rule of law, corruption and Internet, PC and telephone penetration levels. I think this approach misses the central point of development, however. An honest and well-run state simply reduces barriers to an economy reaching its maximum potential level of development; if the human resources are lacking, it will not converge to Western levels.
To illustrate this, let’s take a few middle-income countries, say, Chile and Estonia – both have solid macro-economics and perform respectably in rankings such as economic freedom, ease of doing business and corruption perceptions. Nonetheless, Chile’s per capita growth rate for the past ten years has been a sluggish 2.6%, compared to Estonia’s tigerish 7.7%. Why? I suspect it has something to do with Chile scoring 0.69 and Estonia 1.00 in my Human Capital Index. Russia, rarely cited as a paragon of economic freedom but with a good HCI of 0.94, outperformed Chile with growth of 6.0%. I suspect that the 1-2% difference from Estonia is due to the greater barriers to technology/productivity diffusion in Russia. (Incidentally, the reason the late USSR grew slowly was because its human capital was immensely burdened by the planned economy).
The same goes for the arguments of geographic determinism. Yes, being landlocked and frozen, or suffering the scourges of endemic debilitating diseases in tropical climes, tends to negatively affect development. I don’t see, however, how these disadvantages are different in quality from factors like macroeconomic incompetence or failing institutions.
How will this affect the world’s future?
In my previous Core Article Towards a New Russian Century?, I identified economic convergence, doubly exponential growth in IT and climate change as key drivers of world geopolitics in the decades ahead. On the topic of the former, I wrote:
Of course, it’s not sufficient merely to be behind to catch up. One must also have the human capital and physical infrastructure in place. One of the best proxies for human capital is education. Now as we can see from the info above, Russia’s (and eastern Europe’s) educational profile is of a First World character. Hence it is likely that the region’s impressive post-millennial growth will be sustained, resulting in convergence with west European countries by a 2020-30 time frame.
I stand by this prediction. According to the data, Latin America, the Middle East and South Asia all have educational systems that leave much to be desired, and little sign of fundamental change can be observed (there is no discernable improvement in Mexico’s and Brazil’s PISA scores from 2000 to 2006; tertiary enrolments in the above regions are increasing at a glacial pace).
India is still plagued by illiteracy and as late as 2005 a tenth of the youth cohort didn’t receive primary education, 43% didn’t receive secondary and only 11% received a higher-level education (an unimpressive over 6% in 1991), according to the World Bank. India will be a Great Power, but its few economic/technological centres will remain islands of prosperity amidst a sea of backwardness.
China has a decent school life expectancy, literacy rate and enrollment rates (universal primary and 74% secondary), but its tertiary enrollment ratio is low at 20% in 2005. (Nonetheless, it has increased very rapidly, from 3% in 1991, and if the experience of other countries is anything to go by, a concerted effort could see their rates rise to developed-country standards within the next twenty years). I have no idea how average Chinese students would score on the PISA tests (comparing them to Hong Kong or Macau is a pointless exercise, because of the vast disparity in development), so I guessed 420. If so, then China’s current 10% growth rates should soon moderate to around 5% – as my second graph shows, it is a) fast approaching its potential and b) it’s hampered by bureaucracy and corruption. Today most Chinese growth, unlike in east-central Europe, comes from infusions of labor and capital rather than productivity improvements – growth which could experience a severe and protracted slowdown once the surplus labor pool in the countryside is expended and investment rates are hit by a financial crisis, as happened with the other east Asian tigers after 1997.
In conclusion, China seems set to do considerably better than other developing regions of the world (Latin America, Middle East, South Asia and Africa), but will still be very far from converging to advanced industrial levels in 2025. Meanwhile, eastern Europe will converge with western Europe, while Finland, Korea and Estonia may become some of the richest countries in the world.
How did you work out the Human Capital Index?
Explained beneath the big table. Basically, literacy rate * international student assessment scores mean average * tertiary enrollment ^ (1/3). The reasoning is that a) you must be literate, at a minimum, to participate in a modern economy, b) international student assessments give a clue as to the quality of those who are educated (better than school life expectancy) and c) tertiary education improves human capital further, though not to the same absolute extent as elementary schooling – hence we take the cube root of that figure.
I am aware that in an ideal situation, we would make a sample representing everyone in the country take a skills test to gauge human capital; since we don’t have that luxury, we must rely on available statistics, two of which (test scores and tertiary enrolment) apply mostly to the newest cohorts entering those countries’ labor forces.
What has this got to do with Russia?
When I say things like “Russia will have a GDP per capita of 30,000$ by 2020”, I work by a set of assumptions and beliefs that may not be entirely clear to the casual reader of this blog, particularly the Russophobe variety which believes Russia’s economy is an oil bubble about to pop like a balloon. Which is understandable, given that the Western MSM’s discourse on Russia’s prospects is mostly negative.
Nonetheless, some facts must be acknowledged. The government since 1998 has handled the economy well, balancing inflation and ruble depreciation by maintaining fiscal discipline. Since 2006, they have embarked on large-scale basic (agriculture, housing, health and education) and strategic (nanotechnology, venture capital) investment programs. Yes, the bureaucracy is unwieldy, corruption is a problem and life is hard for small businesses, but as the last few years have showed, these problems are not fatal – at worst, they have shaved off 1-2% of annual GDP growth, and in any case the Baltics are the exception rather than the rule in the post-Soviet space. Most importantly, Russia’s human capital is of First World standards – and as we’ve argued here, this is the key component of development.
If anything, it would be exceptional if Russia didn’t converge to west European levels of development by the 2020’s.