Lecture 1 Economic Growth and Income Differences: A Look at the Data Rahul Giri Contact Address: Centro de Investigacion Economica, Instituto Tecnologico Autonomo de Mexico (ITAM). E-mail: rahul.giri@itam.mx
Economic growth has been the centerpiece of the evolution of macroeconomics over the last 60 years. The main issues on economic growth that macroeconomics has focused on are: Cross-Country Income Differences: There is a great inequality in income levels (income per capita and income per worker) across countries. Countries at the top of the world income distribution are are more than 30 times as rich as those at the bottom. For example, in 2000, gross domestic product (GDP) per capita in the United States was more than $34, 000 whereas in Nigeria it was about $1, 000. In Mexico it was $8, 000, in China about $4, 000 and in India just over $2, 500. These numbers are expressed in 2000 US dollars and are adjusted for purchasing power parity (PPP). Furthermore, this inequality has somewhat increased over time. 2
Figure 1: Estimates of Distribution of countries according to PPP-adjusted GDP per capita 3
Figure 2: Estimates of Distribution of countries according to PPP-adjusted Log GDP per worker 4
Figure 3: Histogram for GDP per capita in 1960 Source: Economic Growth Robert J. Barro and Xavier Sala-i-Martin, 2nd ed., The MIT Press. 5
Figure 4: Histogram for GDP per capita in 2000 Source: Economic Growth Robert J. Barro and Xavier Sala-i-Martin, 2nd ed., The MIT Press. 6
So, in 1960 (the data is for 113 countries) the mean per capita GDP was $3, 390, measured in PPP adjusted 1996 US dollars. Switzerland was the richest country with a per capita GDP of $14, 980, which was 39 times the lowest per capita GDP value of $381 for Tanzania. The US was second with a per capita GDP of $12, 270. Overall, the OECD countries along with Argentina and Venezuela were the richest countries. Most of the Latin American countries were in the middle of the distribution. The poorer countries were mix of African and Asian countries, with the exception of some Asian countries who were in the middle of the distribution. In 2000 (the data is for 150 countries) the mean per capita GDP was $8, 490, which is 2.5 times the mean of 1960. The highest value of $43, 990 for Luxembourg was 91 times the lowest value of $482 for Tanzania. The OECD countries still dominated the top group, joined by some East Asian countries. Most other Asian countries were in the middle range of per capita GDP, as were most Latin American countries. The poorest countries belonged to Sub-Saharan Africa. Income and Welfare: So, what if there are such large income differences? Should we care? Yes, we should. Income levels correlate very strongly with standards of living, and health. The next two graphs show the association between GDP per capita and consumption per capita, and that between GDP per capita and life expectancy at birth in 2000. We see that the richest countries are not only producing 30 times as much as the poorest countries, but are also consuming 30 times as much. Similarly, in the richest countries life expectancy at birth is as high as 80 years while in the poorest countries it is only between 40 to 50 years. Thus income levels show a strong positive correlation with welfare indicators. However, this does not mean that every individual in a rich country is a winner. Very often during the process of growth some individuals or groups of them are left behind. 7
Figure 5: Association between GDP per capita and consumption per capita in 2000 8
Figure 6: Association between GDP per capita and life expectancy at birth in 2000 Source Data: World Bank Development Indicators. 9
Economic Growth and Income Differences: The reason for such large cross-country income differences is the difference in growth rates. To illustrate this point, consider two countries with the same initial level of income. Suppose one country experiences 0% growth in per capita income, whereas the other country grows at 2% per capita. Then in 200 years the country with 2% growth rate will be more than 52 times richer than than the other country. Let us look at how different countries and regions of the world have grown over time. The first graph that follows shows the estimated distribution of growth rate of GDP per worker (PPP adjusted) for 1960 (geometric average over 1950 and 1969), 1980 (geometric average over 1970 and 1989) and 2000 (geometric average over 1990 and 2000). There is a significant variability in the growth rates in every period. Furthermore, the variability has increased over time. The next graph shows the detail of countries in the distribution of growth rate of per capita GDP from 1960-2000. It shows that sub-saharan Africa grew at the slowest rate, and given that these countries started with the lowest GDP per capita levels they ended at the lowest per capita GDP levels in 2000. Asia, on the other hand, started slightly higher than Africa and in may cases grew more rapidly and ended up mostly in the middle of the distribution. Latin America started in the mid to high range of per capita GDP but grew at slower rate than Asia and therefore ended up in the middle with Asia. Lastly, OECD countries started the highest and grew in the middle range of growth rates and therefore ended up at the top of income levels in 2000. The last graph shows the evolution of log per capita GDP from 1960 to 2000 for some specific countries to illustrate how some countries experienced rapid growth in per capita GDP while others fell behind even though they started with higher income levels. 10
Figure 7: Estimates of distribution of countries according to growth rate of GDP per worker 11
Figure 8: Histogram of growth rate of per capita GDP from 1960 to 2000 Source: Economic Growth Robert J. Barro and Xavier Sala-i-Martin, 2nd ed., The MIT Press. 12
Figure 9: Evolution of lof GDP per capita from 1960 to 2000 13
Going Back in History: So, should we conclude that the differences in the growth rates across countries in post-war era can explain the per capita income differences? The answer is no. This is because large per capita income differences already existed in 1960. The first figure that follows plots log GDP per worker of a country in 2000 versus log GDP per capita of the same country in 1960 (in both cases relative to the US value). Most observations are around the 45 degree line, which implies that the relative ranking of countries has changed little between 1960 and 2000. Thus, the ultimate origins of of very large income differences across nations are older than we think. In other words, the world income distribution has been fairly stable over a long time period, with a slight tendency to become more unequal. But, then the question to ask is when did this growth gap emerge? The answer is that much of this gap was created in the nineteenth and early twentieth century. The second figure that follows, using PPP adjusted data compiled by Angus Maddison, shows the evolution of GDP per capita for five groups of countries. The Western offshoots of Europe include Australia, Canada, New Zealand and the US. It shows that during the nineteenth century Western Europe and Western offshoots of Europe experienced rapid growth while Africa and Asia remained stagnant and Latin America showed little growth. The relatively small income gap in 1820 became much larger by 1960. Going back further, evidence suggests that income gaps were even smaller. The third figure that follows shows evolution of average of GDP per capita for the same group of countries starting from 1000 A.D.. Although the data is not very reliable, its shows the same trends as we observed in the earlier graph. Another aspect of the process of economic growth that becomes clear from this graph is that each set of the countries experiences a take-off (or industrial revolution ) in GDP per capita. The first to experience the take-off was Western Europe starting in 1800s, followed by the Western offshoots, followed by the others though not so much Africa. Macroeconomists and historians, alike, have asked why the take-off did not take place before 1800s; why it started in Western Europe and then spread to other parts of the world and what the are the drivers of such rapid and sustained growth in GDP per capita? The last figure shows that evolution of income for some specific countries since 1820. 14
Figure 10: Log GDP per worker in 200 versus log GDP per capita in 1960 15
Figure 11: Evolution of average GDP per capita, 1820-2000 Source Data: Historical Data compiled by Angus Maddison. 16
Figure 12: Evolution of average GDP per capita, 1000-2000 Source Data: Historical Data compiled by Angus Maddison. 17
Figure 13: Evolution of GDP per capita in specific countries, 1820-2000 Source Data: Historical Data compiled by Angus Maddison. 18
Conditional Convergence: The data so far suggests that large income differences across countries emerged in ninteenth and early twentieth century, and over time these differences have increased. However, our analysis has looked at the unconditional distribution of income across countries, i.e. we looked at whether income gap between countries increased or decreased regardless of countries characteristics. Barro and Sala-i-Martin argue that one should look at conditional distribution of income. In fact, they find that in the post-war era income gaps between countries with similar characteristics become smaller over time. The first graph that follows shows that for the world there is no tendency for convergence. There is no relationship between the average growth rate of GDP from 1960 to 2000 and the log GDP per worker in 1960. However, when we plot the same relationship for the OECD countries, there is a strong negative relationship, i.e. countries that had higher GDP per worker in 1960 had lower growth rate of GDP over 1960-2000. Thus, the relatively poor countries in the OECD showed a tendency to catch-up with the rich ones. The OECD countries share similar characteristics in terms of institutions, policies, and initial conditions. Thus, there might be some type of conditional convergence when we control for certain country characteristics that potentially affect economic growth. 19
Figure 14: Conditional Convergence for the World 20
Figure 15: Conditional Convergence for the OECD 21
Correlates of Economic Growth: So what factors drive the process of economic growth, i.e. what are the causal factors? Isolating the causal factors is not an easy task because of the interaction between the potential factors and that between economic growth and these factors. Therefore, for now, we will look at the correlation between economic growth and potential factors that theories of economic growth have focused on. The first figure that follows shows a strong positive correlation between growth rate of GDP per capita and the average investment rate. The second figure that follows shows a positive correlation between growth rate of GDP per capita and years of schooling. These figures suggest that rapid growth countries have invested more in physical and human capital. These are not the only factors that correlate with growth, but are the factors that theory has focused on. Another important factor that theory has focused on is technology. 22
Figure 16: Correlation between average growth rate of GDP per capita and average growth of investments to GDP ratio, 1960-2000 23
Figure 17: Correlation between average growth rate of GDP per capita and average years of schooling, 1960-2000 24
So, even if these correlations between economic growth and physical capital, human capital and technology are in fact also causal relationships and these factors are the drivers of economic growth, the question that one would like to answer is that why some of the countries failed to invest in physical capital, human capital and technology? Thus, fundamental factors that lie behind these proximate causes of growth are what we would like to understand in order to truly understand the process of economic growth. 25