Saturday, January 29, 2011

Foreign aid and development economics

The aim of main research agenda of development economics in the last century was to provide an evolving approach to curing the persistence of poverty and underdevelopment in world's least developed and developing countries. High economic growth in developing countries in the last decades has changed many developing nations into middle-income countries. For instance, real economic growth rate in China and India from 1960 onwards averaged 6.67 percent and 3.49 percent, respectively. In 2010, China and India were already classified as lower middle-income countries, belonging to the same income group as El Salvador, Armenia and Philippines. In the recent year, China's GDP per capita was higher than GDP per capita of many high-growing developing nations such as Ukraine, Nambia, Armenia and Bosnia and Herzegovina, and roughly at the same level as Algeria. Over the last decade, the economic growth in developing countries accelerated, driven by an increase in global commodity prices, robust investment rates, expansionary monetary policy and a growing domestic consumption. The economic growth in a majority of African states stagnated, consequently leading to a decrease in the overall standard of living. Between 1960 and 2009, average real GDP growth was negative in countries such as Congo, Democratic Republic (-2.26 percent), Liberia (-1.51 percent), Niger (-1.02 percent), Zambia (-0.52 percent) and Zimbabwe (-0.02 percent) with many other African countries with little or no growth in the second half of the 20th century. The stagnation of income per capita in countries such as Sierra Leone is largely the result of civil war and severe political instability, creating domestic violence and the persistence of poverty, malnutrition and AIDS/HIV prevalence. From the second half of the 20th century onwards, international aid donors have contributed significant amounts of foreign developmental assistance in various forms such as medical care and vaccination against polio, AIDS/HIV, measles and malaria, direct cash transfers and physical infrastructure. Despite significant official and unofficial developmental assitance from international aid donors, dispersion of real income per capita, measuring the level of cross-country convergence or divergence of income per capita, the gap in economic development widened in the course of the last century. In 2010, the percentage of countries with the level of real GDP per capita $1,500 or below equaled almost 20 percent (link).

The rise of development economics in the 20th century was a natural response to growing disparities in income per capita between rich and poor countries. In the framework of neoclassical theory, development economics emerged from a neoclassical growth theory, pioneered by the famous Solow-Swan model. In the simplest possible form, the growth of output per capita depends on the capital per worker and the initial level of output under stable rate of national saving and capital depreciation. Assuming diminishing returns to scale and constant rate of population growth, the increase in capital per worker would increase the output per worker that would, hence, approach its steady-state equilibrium. Theoretical notions of the Solow-Swan model were tested against the empirical data on economic growth. The key assumption of the neoclassical growth model is that poor countries would tend to catch-up rich countries, assuming higher output growth in poor countries. The convergence of income per capita would imply a neg relationship between the initial level of output per capita and output growth over time. Thus, countries with lower levels of output per capita in the initial period would experience faster rates of output growth. Consequently, the output per capita and the standard of living would approach to the level in rich countries. The empirical tests of the Solow-Swan model failed to confirm the theoretical hypothesis since economic growth rates in 20th century in developed countries were higher compared to developing countries. The divergence of income per capita led to the subsequent modifications of the Solow-Swan model. In fact, the main criticism of the model points out that the model itself failed to capture the role of technological progress in determining the level of output per worker. The mysterious growth episode in Japan and other East Asian nations posed a difficult question. How can a country with low initial level of output per worker at the end of the WW2, exceed the productivity level in rich countries? The obvious answer is that Solow-Swan growth model failed to capture the role of technology shocks which violate the assumption of diminishing capital returns, what could explain why initially poor countries subsequently converged to the level of productivity in rich countries and then exceeded the level. The phenomenon, known as growth residual, has subsequently reduced the predictive power of the Solow-Swan model since a considerable share of economic growth was not ascribed to capital and labor inputs but rather to the persistent role of technological change.

Policy implications from Solow-Swan model imply that the essential requirement to boost economic growth in a country with low initial level of output per capita is to increase the amount of capital per worker, namely by boosting public and private investment in infrastructure. From 1950s onwards, World Bank had repeatedly boosted the growth of infrastructure by facilitating developmental assistance into world's least developed countries. According to the neoclassical growth theory, higher capital-labor ratio would provide additional investment stimulus, thereby increasing the employment-to-population ratio. Proponents of the foreign aid provided the rationale for higher foreign aid spending by the analogy of post-WW2 Europe when Marshall Plan provided $13 billion, or roughly $100 billion in today's prices, to Western European economies to recover the physical infrastructure which had been destroyed during WW2. Marshall Plan intervention was rather short, quick and finite. The efficacy of foreign aid in Africa is questionable since little or no growth occured in many African states such as Burundi, Benin, Zimbabwe and Congo. Official forecasts from the United Nations from 1950s onwards, based on the famous Harrod-Domar growth model (link), often assumed a rapid increase in the level of GDP per capita in response to the increase in investment rates. The forecasts, based on the theoretical assumption of diminishing capital returns, predicted a persistent convergence of GDP per capita to the level sustained in richer countries. The fact that the launch of extensive investment in infrastructure resulted in further economic stagnation of many African states, has questioned both the validity and quality of prescriptions laid by the mainstream development economics.

The philosophy of the mainstream development economics was sharply criticized in the light of the fact that foreign aid failed to alleviate poverty and made the growth of African economies slower. The efforts by the World Bank have been diverted from correct diagnosis of the developmental issues in African states to repeated initiatives such as the commitment of the international community to increase the share of foreign aid to least-developed countries to at least 1 percent of the GDP. The criticism of the mainstream development economics was already formulated in 1958 when Mont Pelerin Society organized the 9th meeting and development economics seminar where professor Herbert Frankel of the Nuttfield College put forth the criticism of foreign aid and the failure of development economics:

"The lesson that flows from it is that it does pay to go to these remote areas and find out what the problem is, instead of assuming that one knows the problem before one begins. Until recent years, people have simply assumed in many of these territories in Africa, that there were no real, positive signs of enterprise among the indigenous population, which was supposed to be so uninstructed or inert that it was not able to fend for itself, experiment for itself, or improve itself. It was not realised that a reason why there was this apparent lack of initiative in the population was that there were serious customary or legal obstacles to the exercise of ordinary enterprise, even on a small scale."
Given the lack of the comprehensive diagnosis of the causes of underdevelopment in African countries, the mainstream development economics failed to capture the appropriate assumptions in the theoretical models of economic development, upon which developmental assistance was justified. A more reasonable theoretical solution to the economic stagnation and social conflict in Africa has been put forth by the human capital theory. In its broadest and most general form, the theory stated that the economic stagnation of African countries is a consequence of the lack of skills and investment in education that could provide the necessary input to increase the economic growth and, subsequently, alleviate the issues of AIDS/HIV, malaria, child malnutrition and domestic violence. There is no doubt that the growth of education initiatives in Africa has sent many children to school. In addition, many universities in Western Europe and the United States have expanded the initiative and offered students from African states preferential admission criteria in various forms such as graduate fellowships, student grants and lower required standardized test scores, to boost admission rates of African nationals at U.S. universities. The efforts of developed countries to bring educational initiatives to Africa encouraged school participation as well as international opportunities of African citizens to study abroad, even at world's most prestigious and highly-ranked universities. Notwithstanding the importance of education in creating the stock of human capital for the wealth of nations, educational initiatives should address the essential obstacles that creates the failure of African expatriates to return to home countries, hence, bring skills, knowledge and various other forms of human capital, which are essential to the process of long-run growth, the issues of labor market distortions in African countries. These distortions crucially impede the ability of young African graduates to matching jobs in regional labor market.

What the mainstream development economics failed to take into account is the institutional paralysis which prevails in a majority of African countries, plagued by the destructive tribal institutions based on widespread corruption, bribes and domestic violence as means of achieving political power. The prevalence of hybrid institutions, marred by the complete absence of the rule of law and judicial institutions that could facilitate efficient contract enforcement and the protection of private property rights, is not only a severe obstacle to higher economic growth but also the apparent mechanism that captures the set of explanatory features that could possibly account for what caused the misdiagnosis of the African development dilemma. Back in 2002, African Union estimated that each year, corruption costs African economies more than $148 billion or 25 percent of Africa's GDP. The significance of corruption in state structures in Africa manifests itself in poor quality and provision of public services, the absence of judicial independence from political regimes, cumbersome contract enforcement and unprotected private property rights. Such distortions impede the level of trust and provide evolving incentives to subvert the institutional independence into political cronyism, in which corruption substitutes the tax system through bribes and extortion as methods of lowering transaction costs in overcoming the malfunctioning of the judicial system. In 1978, Erwin Blumenthal of the central bank of the Federal Republic of Germany, warned the international community that "Zaire's political system is so corrupt that there's no prospect for Zaire's creditors to get their money back." (link)

The advancement of country's economic prospects requires not only transparent, sound and efficient regulations but, more importantly, highly efficient civil service. In 2010, Transparency International published Corruption Perception Index (link) by measuring the persistence of corruption in public sectors across the world. The findings showed that the vast majority of poor African countries were plagued by extensive and extortionate corruption and ranked in the bottom 20 percent of the distribution. Comparing the level bureaucracy against GDP per capita reveals the amplified evidence of the negative correlation between the efficiency of civil service and the GDP per capita. The ease of doing business in Africa in countries such as Botswana, Ghana, Mauritius and South Africa is remarkably easier with predictable, stable and efficient regulation, compared to countries such as Burundi, Burkina Faso, Côte d'Ivoire etc. where highly burdensome administrative procedures in doing business hamper capital formation and restrain productive investment in health-care, education and private-sector infrastructure that could provide the impetus to economic growth.

The relationship between the amount of foreign aid, received by the least-developed countries, and the scope of corruption as a rough approximation of the institutional quality in the least-developed states, could provide the answer to the question whether international donors consider the scope and significance of corruption in allocating the amount of foreign aid. The experience from the last century of development policy, suggest that international donors actually allocated more foreign aid to the countries, suffering from severe state failure, widespread corruption, government failure and the complete absence of judicial independence that could provide a system of checks and balances and the necessary restraint on the violiation of private property rights, extortion and violence by the political elites. In 1999, Alberto Alesina and Beatrice Weder (see "Do Corrupt Governments Receive Less Foreign Aid," American Economic Review, 92(4), pp. 1126-1137) found that, contrary to arguments of aid supporters, foreign aid is not used to reward good governments since more corrupt governments received more foreign aid and official development assistance from international donors. The most striking evidence, presented by Alesina and Weder, suggests that U.S donors seem to neglect the persistence of corruption in allocating foreign aid to poor countries while, on the other hand, Scandinavian donors deem the persistence of corruption as highly important, hence, rewarding governments with lower extent of corruption.

In the following graph, I estimated the impact of corruption on official development assitance in the sample of 41 least-developed countries in 2008. In the model, I set the official development assitance to be determined by the scope of corruption in least-developed countries. The official development assitance is expressed as a share of representative country's gross national income (GNI) for it provides a better measure of aid dependence than foreign aid per capita since the size of population is controlled by the main assumptions of the model. The data on official development assistance were download from World Bank's World Development Indicators (link). The data on the extent of corruption in least-developed countries were provided by Transparency International's 2008 Corruption Perception Index (link). The extent of corruption varies from 1 to 10, where lower values indicate more persistent corruption. I estimated whether countries with more corrupt governments receive a higher share of foreign aid from international donors. On the basis of 41 least-developed countries, sample estimates suggest that a 1 point improvment in corruption perception index tends to decrease, on average, the share of foreign aid in gross national income, on average, by 2.37 percentage points. Sample estimate of the slope coefficient is statistically significant at 5 percent level. Even though, the variation in corruption perception index accounts for 5.51 percent of the variation in official development assitance, the influence of the extent of corruption on the share of foreign aid in gross national income is not spurious but systematic and persistent.

Corruption and official development assistance
Source: World Bank, World Development Indicators, 2010. Transparency International, Corruption Perception Index, 2008.

The estimate suggests that international donors indeed reward more corrupt governments by increasing the share of official development assitance. In 2002, African Union estimated that corruption was costing the African continent $150 billion per year. The estimates of the total cost of corruption provide an ample evidence that, over the last century, international donors consistently allocated foreign aid to more corrupt governments, creating aid-dependent economies, prone to bloated bureaucracies and extractive institutions which subsequently led to the stagnation of income per capita in the last decades. An ample criticism of foreign aid initiative was put forth by Dambisa Moyo (link) in the WSJ two years ago: "The most obvious criticism of aid is its links to rampant corruption. Aid flows destined to help the average African end up supporting bloated bureaucracies in the form of the poor-country governments and donor-funded non-governmental organizations."

The consequence of rootedness of corruption and extractive political institutions in African tribal cultures can be, in a considerable part, drawn upon the colonial heritage that spread throughout the African continent from 19th century onwards. The colonial experience across the African continent (link) served not only as a conquest of newly discovered areas but, moreover, also as an experiment of developing political and economic institutions on the basis of European influence. The colonial heritage in Africa was mainly derived from the European occupation of African lands. Hereto, the presence of European colonisers in Africa provided a long-lasting foundation of the institutional lessons from which the African states went forth.

Given the heterogenity of the European perspectives on institutional development, the colonial period in Africa left a long-lasting impact on the economic and political development in Africa. Africa's richest countries, namely Botswana, South Africa and Mauritius, were influenced tremendously by the colonial heritage. In Botswana and South Africa, the colonial influence of English and Dutch on further economic development was mainly derived from setting strong institutional foundations of economic development such as the rule of law, judicial independence and limited government compared to other African states. Apart from the setting of formal institutions, fostering contract enforcement and the integrity of the political institutions, English and Dutch colonizers provided the establishment of cultural setting not prone to fraud, extortion and extractive institutions. Favorable institutional conditions furthered the advertance of trust and institutional efficiency, which are deemed essential in fostering the development of financial markets. Even the German presence in Namibia from 1884 to 1915 during Deutsch-Südwestafrika (link) fostered, to a certain extent, independent judiciary, relatively sound institutions and cohesive framework of the rule of law. As a result, Nambia retained the status of one of the least corrupt countries in Africa, known for relatively high degree of economic freedom in a regional comparison with other African states.

While the influence of German, English and Dutch colonizers was largely beneficial to African countries from the perspective of economic growth and development over the last century, the presence of French, Italian and Belgian colonizers arises serious concerns over the prospects of economic development across the African continent. The myraid of violence, in countries such as Congo Dem. Rep. and Somalia, which ultimately led to civil wars and the settlement of extractive institutions, largely reflects the innate ability of the colonial policies to provide the necessary conditions for the institutional integrity, the rule of law and stringent property rights that could underline the basis of economic development by restraining the power and domination of political elites and their ability to expropriate private property rights in pursuit of extractive monopoly rents from natural resources. That easily explains why countries such as Congo, Zambia, Nigeria and Zimbabwe, in spite of vast reserves of natural resources, were seized by the state capture of political elites. The colonial presence largely determined the size and scope of aid dependency in African states. The most plausible and persuasive explanation of the impact of European colonial policies in African countries was presented by Daron Acemoglu, Simon Johnson and David Robinson (see "Disease and Development" Journal of European Economic Association, 1(2/3), pp. 397-405):
"European colonists were much more likely to develop institutions of private property, encouraging economic and social development, in places where they settled. In contrast, in places where they did not settle, they were more likely to opt for extractive institutions, designed to extract resources without investing in institutional development. In these places, institutions were highly centralized, with political power concentrated in the hands of small elites and with almost no checks on this elite. The property rights and more general rights of the majority of the population were not protected."

The political and economic circumstances of the European institutional legacy in African states imparted aid dependency on those countries where the combination of tribal institutions, hostile to free enterprise and judicial restraint of political dictatorships, and unequivocally detrimental colonial policies dominated the development of political and economic institutions, setting the rules of the game. Therefore, the inability of many African societies to establish sensible and effective institutions resulted in the political capture of the state by the elites. The monopoly power of the political elites, enforcing anti-growth public policies, led to consistently poor economic outcomes, plagued by high rates of poverty and infectious diseases such as polio, malaria and measles.

The challenge of development economics is not to design aid schemes, which inevitably lead to aid dependency, marred by persistent corruption and political fraud, but to ascertain correct diagnosis of why foreign aid repeatedly resulted in the poor economic outcomes and the consequent stagnation of income per capita in many African states in 20th century. The failure of African societies to establish a rigorous system of incentives, which could significantly improve economic outcomes, is not a response to market failures (which deemed highly of early development economics) but a result of severe government failure to establish effective institutions of the rule of law, contract enforcement and stringent property rights. These institutions are the broadest foundations of economic development and the only viable alternative to political nepotism and the power of elites which, as poor development outcomes in Africa show, ultimately impose extractive institutions, causing the persistence of poverty and underdevelopment.

Wednesday, January 12, 2011

Artificial states

The question of the artificial states has recently been brought up by the referendum in Southern Sudan on whether the southern part of the country should declare political independence from the northern part of the country. An article in New York Times (link) succintly discussed few notable fact-checked evidence of the increasing ethnic, political and economic North-South division within the country. As a single country, Sudan performed terribly in development outcomes. According to CIA World Factbook (link), the country suffers from high rates of extreme poverty and illiteracy. For instance, the official share of population below poverty line is estimated at 40 percent - almost twice the average of North African states. In addition to poor development outcomes, the country is plagued by significant political and ethnic fragmentation into largely Muslim, Arab-speaking north and predominantly Christain, English-speaking south. The political independence of South Sudan is the only contemporary evidence of the re-establishment of land borders alongside the ethnic division.

In the recent paper entitled Artificial States, Alberto Alesina, William Easterly and Janina Matuszeski presented two formal measures of artificial states. Aside from the measure of ethnic division, the authors constructed the measure of straightness of border lines. The hypothesis suggests that squiggly geographic border lines separate the states alongside the ethnic division. On the contrary, straight border lines suggest increase the probability of the emergence of artificial states plagued by ethnic and linguistic fractioning. The authors presented the empirical evidence, suggesting that fractional land borders are highly correlated with the GDP per capita. In addition, the share of ethnically partitioned population within the country is systematically decreasing the GDP per capita in cross-country comparison. The intuitive ideas behind the empirical evidence suggest that at the end of colonial period, colonizers that set straight borderlines between the emerging countries incured significant economic cost to newly formed African countries in terms of lost GDP. The evidence from Alesina-Easterly-Matuszeski study suggest that a 1 percentage point increase in the fraction of country's population belonging to groups partitioned by the border would decrease the GDP per capita by 1.3 percent. On the other hand, countries with squiggly geographic borderlines enjoy significantly higher GDP per capita.

The post-colonial period in Sudan was characterized by two civil wars which outbroke in 1972 and 1983. In 1956, Sudan gained the political independence from Great Britain. Contemporary borderlines were predominantly determined by the colonial authorities in African states prior to the wave of independence of many African nations. The emergence of the artificial states is rather a consequence of poor colonial policies than of high bargaining cost of ethnic groups within the country in setting country borderlines. Hence, the economic effects of colonial legacy can persist over time. Consider the evidence from Cameroon. The country was originally colonized by Germany. During the World War I, Northern Cameroon was occupied by Germany while the rest of the country was colonized by the French. Between 1916 and 1960, the country was a unique experiment of how the establishment of the institutional setting of European countries affects domestic development outcomes. A recent study by Alexander Lee and Kenneth Schultz (link) suggests that in the areas formerly occupied by the British enjoy higher levels of wealth and improved access to clean water while the rest of the rural country, after having been colonized by the French, suffers from significantly hindered access to clean water and worse provision of public goods. Even though the colonial patterns do not apply to urban areas, lessons from Cameroon suggest that the impact of post-independence public policies and colonial legacy on the level of wealth is of the same importance even when linguistic and ethnic fragmentation persists over time.

In spite of considerable degree of inefficiency, the persistence of inefficient and ethnically fragmented states is continuously marked by poor economic and development outcomes, often accompanied by civil-war conflicts such as military violence and genocide by the Sudanese army in Darfur. An interesting theory has been recently put forth by Daron Acemoglu, Davide Ticchi and Andrea Vindigni (link) who suggest that rich political elites seize state capture and democratic politics by expanding the size of bureaucracy. Hence, to gain political support, the coalition of elites chooses an inefficient structure and organization of the state.

The phenomenon of artificial states is not abridged to least-developed countries and developing world. Even in the group of advanced countries, several countries emerged despite a considerable degree of linguistic, ethnic and cultural fractioning within country borders. The evidence from Switzerland suggests that a continuous transition to a peaceful and stable democracy is possible. Amid highly fragmented linguistic and cultural characteristics such as four official languages and the persistence of GDP per capita divergence between high-income German cantons and low-income French cantons, Switzerland is characterized by envious political stability and economic performance. The genuine feature of federal political system is a consistent fiscal decentralization such as jurisdictional competition between units within the federalized structure in various areas such as taxation, regulation, health care etc. Hence, a decentralized fiscal structure and division of powers require a limited federal government and a high degree of autonomy within the federation. Thus, despite a multilingual population, the Swiss model of federalism is marked by political stability, peace and prosperity.

The phenomenon of artificial states in advanced countries is not confined to Switzerland itself. What distinguishes the Swiss model of federalism from centralist political systems in its neighboring countries is a high degree of political autonomy in Swiss cantons. Competition between jurisdictions within the federation nonetheless generates different economic outcomes. However, the outcomes generated by jurisdictional competition preclude adverse effects caused by either state capture of democratic politics or redistributive taxation between jurisdictions. In Switzerland, cantons with favorable public policies such as low tax rates on labor and capital, sound regulation and competitive provision of health care, have enjoyed persistently higher levels of wealth compared to cantons in the rest of the country. Of course, the coexistence of diverse ethnic and lingual groups within the single state requires common values, integrated into formal institutions.

Contrary to common perception, the artificial state may not be characterized exclusively by ethnic and linguistic fragmentation. To a large extent, Germany and Belgium could be classifed as artificial states. In Belgium, Flemmish-speaking north of the country consistently outperformed French-speaking south on various indicators and outcomes, including income per capita, international test scores and employment-to-population ratio. The political and linguistic division of Belgium into high-income Flemmish part and less developed French part reflects the essential dilemma of artificial states - should a single country with fragmented and heterogenous population be abandonded and whether ethnicity borders should represent country borderlines. In fact, linguistic fragmentation of Belgium to the extent that Flemmish and French part of the country adopted different administrative and education systems, led to persistent inability of two majorities to form a government. In 2007, The Economist opined that Belgium should cease to exist. The unification of Germany (Wiedervereinigung) integrated two parts of the country with vastly different institutional setting into a single political unity. However, Eastern and Western Germany were known for completely different political and economic system. The unification has incured many adverse effects. A significant difference in wage and price levels between East Germany and West Germany caused continuous migration of East German labor into West Germany, thus decreasing the productivity growth in East Germany. Consequently, the unification of the country led to the adoption of West Germany level of prices and wages in Eastern part of the new country. The artificial increase in price-wage level increased the unemployment rate in Eastern Germany to double-digit level, not least triggered brain-drain and capital flight. Hence, the unification of Germany as an artificial state resulted in persistent income per capita divergence between high-income West Germany and low-income East Germany. The unification of Germany into a single country should indeed never happen. In fact, adverse effects of the unification on East German productivity and wages would not lead to continuous increase in unemployment rate. If East Germany maintained a high degree of political autonomy, the transition to market economy would not be restrained by the adoption of West German price-wage level that could not be sustained by low productivity level in East Germany. To avoid the pitfalls of the artificial states, West and East Germany would be better off, had the countries never been reunified.

Recent national referendum in South Sudan on whether the southern part of the country should declare political independence from Sudan again pondered over the persistence of artificial states. The empirical evidence on poor development outcomes in several African countries suggests that borderlines, disregarding the ethnic distribution of the population within the country, are highly correlated with low income per capita. The unique solution of the artificial states is the adoption of political federalism. Under fiscal decentralization and limited government, federalism enables peaceful and prosperous existence of fragmented ethnic and linguistic structure in a single state. For instance, former Yugoslavia, known for highly fragmented ethnic, linguistic and economic disparities, ceased to exist not because federalism would not be a genuine political system but, ultimately, because of severe economic mismanagement, powerful and centralized government that disdained the principles of political autonomy and market economy, causing severe hyperinflation and the collapse of the federation that eventually resulted in a decade of civil wars and military violence. The evidence from Yugoslavia suggests that between 1950 and 1990, drastic economic divergence occured. The lesson suggests that the essential condition for the efficiency of federalism as a political system is high political autonomy and fiscal decentralization, both of which enable the competition between public policies.

Amid linguistic fragmentation, the competition between jurisidictions rewards competitive public policies by higher income per capita which ultimately boost the inception of public policies in less developed parts of the federation. Hence, without jurisdictional competition and political autonomy, ethnic and linguistic fragmention of the country may ultimately result in political instability which, in addition, generates poor economic outcomes.

Friday, January 07, 2011

Student performance and economic growth

The 2009 PISA test study (link) of students' proficiency in reading, mathematics and science is a highly successful method of evaluating student performance across countries. In fact, the creation of human capital is the main endogenous feature of the long-run economic growth since the quality of schooling and education system are essential to the creation of human capital. The difference in GDP per capita across countries is both intuitive, theoretical and empirical challenge to search for the causes of the gap between the economic performance of nations.

PISA test scores are aimed mainly at the evaluation of student knowledge at primary and secondary level in the fields of reading, mathematics and science. The assessment of knowledge in a particular field is subdivided into six different proficiency levels, ranging from 1 to 6. For instance, students at the 1st reading proficiency level are characterized by innate recognition of simple ideas reinforced in the text while students at the 6th reading proficiency level are characterized by a full capability of making multiple inferences, comparisons and contrasts and integrating the ideas presented in the text into a coherent conceptual framework of abstract ideas, sound evaluation and reflection. While 98.6 percent of OECD students can perform reading tasks at level 1, only 1.1 percent of students across OECD countries can perform reading tasks at the highest proficiency level. In addition, 28.4 percent of students in OECD countries exceeded the 4th (mid-range) reading proficiency level.

The reading scale has been further divided in reading continuous and non-continuous texts. However, the evidence suggest no systematic difference in reading scores between the two fields. Countries with the highest performance, measured as mean score, in reading rank are Korea (89.8 percent), Finland (89.3 percent) and Canada (87.3 percent). Countries with the largest student populations such as United States, United Kingdom and France were ranked in the upper-middle range while percent, and Israel (79 percent), Luxembourg (78.6 percent) and Austria (78.3 percent) are the lowest-ranking high-income countries on the reading scale in the 2009 PISA assessment. In the field of mathematics, 8 percent of students in OECD countries perform below level 1, 31.4 percent of students can perform mathematical tasks at 4th (mid-range) proficiency level while 3.1 percent of students perform at the highest proficiency level. In the country distribution, the percentage of students in the 6th proficiency level is the highest in Korea and Switzerland (8 percent), Japan, Belgium and New Zealand (5 percent). In a regional distribution, more than 25 percent of students in Shanghai perform at the highest level of mathematical proficiency. The proportion of students in the 6th proficiency level is very high in Singapore, Chinese Taipei and Hong Kong - 15.6 percent, 11.3 percent and 10.8 percent respectively. In addition, performance disparity in mathematics varied significantly across countries. Less than 1 percent of students in Mexico, Chile, Greece and Ireland reached 6th proficiency level A brief overview of the main empirical findings suggests a rather rigorous disparities in country ranking and performance.

The assessment of student performance in science is similar to the distribution of mean scores in the field of mathematics. About 5 percent of students perform below 1st proficiency level. In addition, only 29.4 percent of the students in OECD countries is proficient at 4th (mid-range) proficiency level in science while an average 1.1 percent of students in OECD countries can perform at the highest level of scientific literacy. In addition, the percentage of students below the lowest level of scientific proficiency is highly negatively correlated with country ranking since the proportion of students below the 1st level is the lowest in Finland (8.3 percent), Korea (6.3 percent), Estonia (8.3 percent) and Canada (9.6). Higher country ranking would thus indicate a lower proportion of students below the 1st proficiency level. All of the aforementioned countries ranked in the highest 10 percent of the distribution. If Shanghai, Hong Kong, Macao and Taipei were independent countries, their respective ranking in the field of scientific literacy would be in the top 10 percent of the distribution.

The empirical data on the distribution of mean scores in reading, mathematics and science are highly relevant to the measurement of human capital since the impact of mean scores on economic growth would differ to the certain extent from other measures of human capital. Recent attempts to capture the effect of human capital on economic growth were aimed at the definition of human capital as total years of primary, secondary and tertiary schooling. For instance, Robert Barro and Jong Wha Lee have collected disaggregated data on the total years of schooling for 146 countries between 1950 and 2005 at five-year intervals (link). The empirical evidence from the country panel suggests a strong linkage between schooling and long-run economic growth and institutional country features (link). In addition, Barro and Lee estimated the implicit return from an additional year of schooling ranging from 5 percent to 12 percent.

Gary Becker (link) and Richard Posner (link) recently discussed the 2009 PISA findings and the impact of cultural, genetic and demographic disparities on mean test scores in the United States. United States ranked in the middle of the mean score distribution. The rank of the United States (17th out of 79 countries) is above average in reading and average rank in mathematics (31st out of 79 countries) and science (23rd out of 79 countries). As Becker and Posner indicate, the relative performance of the United States should be evaluated with the consideration of cultural and demographic differences since the mean score of White and Asian students is significantly higher than the mean score of African American and Hispanic students. Disparities in mean scores between different demographic groups are typical in largely heterogenous populations. In Belgium, the regional disparity in mean scores between French and Flemish communities is even more striking. While the mean score in mathematics in 2006 in Flemish community had beenabove the OECD average, the mean score in mathematics of students in French community had been 18.56 points below the OECD average while the mean score of students in Flemish community had been 30 points above the OECD average. In 2009, such a relative difference would place French community in the rank of Italy, Portugal and Spain. On the other hand, student performance in Flemish-speaking community would reach the rank of Canada, Switzerland and Japan.

In the U.S, the demographic disparities in mean scores in reading, mathematics and science do not reflect the quality of the American education system. While the overall quality of the public and private American high school education system raised considerable concerns in previous performance of U.S. students in international mathematics and science ranking, the ranking of U.S. universities in science, mathematics and social sciences is the highest in the world. The output U.S. universities resulted in the highest number of Nobel laureates in physics, chemistry, medicine and economics as well as into cutting-edge accomplishments in R&D and technology. The emphasis on creativity and innovative thinking embodied in the American education system has enabled the United States to emerge as a world leader in technology, R&D, innovation and entrepreneurship.

The openness of the U.S. education system to international students, ideas and creative thinking could account for the remarkable achievements and academic quality of top American universities. On the other hand, poor teacher quality in American high school system is detrimental to the reading and quantitative literacy of American high school graduates, as a consequence of what Becker calls "teaching-to-the-test" syndrom where many public school teachers teach students topics not relevant to the command of knowledge but to the tests since test scores presumably determine teacher pay. Eric Hanushek of Hoover Institution recently found (link) that replacing bottom 5-8 percent of high school teachers with average teachers could near the United States on top of the international mathematics and science ranking. In addition, the measure is worth $100 trillion according to Hanushek (2010).

In the international perspective, student performance has been viewed as a significant determinant of the difference in cross-country economic performance. Recent paper by Atherton, Appleton and Bleaney (2010) found that higher mean test scores in mathematics, reading and science significantly improve per capita GDP. The authors showed that holding per capita income constant, average years of schooling is less important than mean test scores in predicting the economic growth. The 2009 PISA study highlighted the relationship between international test scores and economic growth. The evidence suggests almost non-existent correlation between reading performance and GDP per capita and cumulative education expenditure. The evidence simply suggests that socioeconomic variables matter more for reading performance than simple and often inconclusive aggregate indicators. In analysing the impact of various social and economic variables on reading performance, the results suggest a high correlation between parents' education and children's reading performance. For instance, 1 percentage point increase in the percentage of the population aged 35-44 with tertiary education returns 1.36 point increase in average reading performance where parents' education accounts for 44 percent of the variation in children's reading performance. The impact is shown in the following graph.

Parents' education and student reading performance

Source: OECD, PISA 2009.

A brief look also reveals another striking implication: cross-country reading performance is strongly affected by social and cultural status of the child's parents. A simple estimate of the relationship between reading performance and socio-economic status suggests that 1 percentage point increase in the share of students with very low social, cultural and economic status tends to decrease the average reading score by 1.13 points. Hence, social, economic and cultural status explains about 46 percent of the variation in student reading performance.

A considerable improvement of primary and secondary education system is vitally essential to the long-run economic growth. International test scores are an important method of evaluating international disparities in student performance and the subsequent impact on economic growth. Modern knowledge-driven economy requires not only intelligence, attentiveness but also comprehensive, integrated and developed social skills. Low reading, mathematics and science performance is generally attributed to student's low social, cultural and economic status. Genetic, cultural and socioeconomic variables, rather than education expenditure and GDP per capita, tend to play a major role in early childhood development as a basis of future student performance.

The economic and social future of countries requires considerable investment in children and student. Professor James Heckman of the University of Chicago brilliantly argued in Heckman equation (link) why the most gainful benefit of early childhood development is increased social productivity, greater motivation and developmental stimulation that every child needs. Our society should be not neglect an indisputable fact that early childhood development is a major determinant of student performance which sets the conditions for future advantage in school, college, career and life in general. Without these essential measures, student performance would suffer heavily from the spread of crime, teenage violence and high dropout rates.

The evidence from the 2009 PISA study suggests that higher quality of the education system is a necessary condition for higher test scores in reading, mathematics and science. As the evidence suggests, that the quality of human capital is strongly associated with higher standard of living. Without a prudent step towards improving developmental stimulation of students, considerably low student performance may seriously harm the prospects of future generations.