Understanding the World Economy Master in Economics and Business Long-term economic growth Growth and factors of production Lecture 2 Nicolas Coeurdacier nicolas.coeurdacier@sciencespo.fr
Lecture 2 : Long-term economic growth Growth and factors of production 1. The neoclassical growth model 2. Do countries catch with economic leaders? 3. The Asian growth miracle?
Output per capita (constant USD) 25,000 What makes countries rich? A Tale of Two Countries 20,000 The average income on a Japanese person 15,000 10,000 5,000 The average income on a Ghanaian person 0 1950 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001
By 1996, Japan had 19 times the income of Ghana Three-quarters of Ghanaians have no access to health care; all Japanese do. Forty percent of all Ghanaians do not have clean drinking water; all Japanese do. Half of all Ghanaian women cannot read; all Japanese women can. Ghanaian mothers are 91 times more likely to die in childbirth than Japanese mothers.
Growth experiences are heterogeneous across countries
What are the engines of economic growth? The full picture is complicated and involves many different aspects. Next two lectures build up a complete picture. Focus first on just one factor - capital accumulation. Increases in the stock of physical capital (buildings and machinery). Stress many other things important not least how efficiently this capital stock is used. Capital is just a starting point.
Production function Output produced Buildings and machinery Labour input Technical knowledge and efficiency This lecture focuses on first input capital accumulation
How does capital accumulation work? How does increasing the capital stock lead to higher output? The marginal product of capital (MPK) is the increase in output that comes from increasing the capital stock leaving everything else unchanged The MPK can be : decreasing (Solow neoclassical model) each new machine adds less than the last constant (Endogenous Growth I AK) each new machines adds the same as the last increasing (Endogenous Growth II Poverty Trap) each new machine adds more than the last
THE RETURN ON CAPITAL Diminishing Marginal Product of Capital Return on Capital (MPK) 0 < < 1 = () () = = A(/) Capital
THE RETURN ON CAPITAL Constant Marginal Product of Capital Return on Capital (MPK) = 1 = K = = A Capital
Return on Capital (MPK) THE RETURN ON CAPITAL Diminishing Marginal Product of Capital > 1 = () () = = A() () Capital
The neoclassical MPK Assumptions about MPK purely technological no economics involved. Different assumptions may be needed for different technologies. We stick on decreasing returns - neoclassical production function with 0 < < 1 and k = = capital per capita: = = A() Remark: Different assumptions lead to dramatically different implications. What would cross country growth patterns look like if MPK were initially increasing and then became decreasing?
The neoclassical production function Output per worker y { k } y 2 = { k } y1 y > y 1 2 Capital per worker
Will countries which invest more grow faster? Ultimately - No Short Term - Yes In long-run, investment affects only level not growth of output. To show this we need to add some economics to our technology assumption need to introduce investment.
The Solow growth model Production: = ( ) ( ) Capital accumulation: = (1 ") +$ "= depreciation rate; $ = investment $ = Savings = constant fraction s of income: $ = s = & ( ) ( ) Thus, capital obeys to: = (1 ") +& ( ) ( )
The Solow growth model: steady state = (1 ") +& ( ) ( ) Assuming for now constant and. In per capita terms: = 1 " +& ( ) Steady-state is defined as: = = " = & ( ) Steady-state capital and output per worker: = ( () * )/() ; = ( )
Steady-state capital stock per worker Steady state investment per worker + = " + = & Capital per worker
Convergence to the steady state When capital stock is low: Each new machine leads to a big increase in gross output The amount of output needed to replace machines that have worn out is low As a result net output increases with the capital stock When capital stock is (too) high: Each new machine leads to a small increase in gross output Every period a substantial part of output is needed to replace machines that have worn out Eventually net output decreases with capital stock Key to these results is decreasing marginal return on capital.
Higher savings increase the steady-state capital stock Steady state investment per worker + = " Higher & + = & Capital per worker
Investment Rate and Level of Per Capita Real GDP Log(GDP per capita) in 2008 12 11 10 9 8 7 6 5 0 5 10 15 20 25 30 35 40 45 50 Average Investment Rate 1950-2008 Source: PWT
Savings or Investment? We are using investment and savings interchangeably If a country cannot borrow from overseas then its savings has to equal investment. Most countries do not borrow much from overseas (as a % of GDP) so savings and investment closely linked. When country borrows I > S but then it has to repay funds so S>I. Therefore on average I = S When country can borrow abroad (I > S, convergence towards steady-state can happen faster.
Higher TFP increases the steady-state capital stock Steady state investment per worker $ = " Higher $ = & Capital per worker
Optimal consumption and welfare A high savings rate leads to high income but low consumption A low savings rate leads to low income but high consumption Consumption per worker in the steady-state:, = (1 &) ( ) = (1 &)( & " )/() - + The Golden rule - level of savings that maximizes consumption in the steady state. Simple calculations suggest that optimal rate around 30-35% of GDP
Optimal consumption and welfare, = (1 &) ( ) = (1 &)( & " )/(), 0% 100% &
60 50 40 30 20 10 0 Savings rate (% of GDP) Selected countries, average 1998-2008. Source: PWT Cote d'ivoire Senegal Madagascar Rwanda Argentina United Kingdom United States Canada South Africa Mexico Australia France Italy Poland Morocco Germany India Czech Republic Indonesia Switzerland Japan Malaysia Korea, Rep. of China Singapore
The Solow growth model: dynamics Steady-state: = ( () * )/() Dynamics: = 1 " +& ( ) = -( ) with = -( ) Convergence: converges to. As approaches, the growth rate of (and output ) decreases. The further away from a country is, the faster it grows.
Dynamics of the capital stock per worker = = -( ) 45.
30 20 10 0-10 -20-30 -40-50 German GDP and capital growth 1936-1955 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 GDP Growth Change in Capital Stock
Implications of the neoclassical growth model 1. Countries always eventually reach their steady state 2. In the steady state/long run growth only comes from TFP. Countries at their steady state no longer grow from capital accumulation. Think of the OECD economies as those at their steady state. 3. Richer countries should grow slower than poorer ones: Higher returns on investment in poorer countries. Emerging markets are moving towards a steady state and growing fast. According to this story we would expect countries to catch up with economically most advanced nations
Catch up amongst Europe s big 4 Log output per capita (US $ 1990) 11.0 10.5 10.0 9.5 9.0 France Germany Italy UK 8.5 8.0 7.5 7.0 6.5 6.0 1861 1866 1871 1876 1881 1886 1891 1896 1901 1906 1911 1916 1921 1926 1931 1936 1941 1946 1951 1956 1961 1966 1971 1976 1981 1986 1991 1996 2001 2006 Source: Maddison, GGDC and DataStream
Lecture 2 : Long-term economic growth Growth and factors of production 1. The neoclassical growth model 2. Do countries catch up with economic leaders? 3. The Asian growth miracle?
Output per worker growth (/ 0 ) dynamics / 0 Everything else equal, countries starting further away from their steady-state capital stock are growing at at a faster pace. Zero growth from capital accumlation.
Do countries converge? According to the Solow neoclassical growth model, poorer countries (in terms of capital stock per worker should) grow faster. Does it hold in the data? Investigate link between future growth and initial income per capita.
Do countries converge? The Evidence - Round I: The World 12 Average GDP growth rates 1970-2002 10 8 6 4 2 0-2 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 GDP per Capita 1970, USD If economies converge then expect negative correlation : countries with high GDP in 1970 should grow more slowly as capital earns a low return. Source: PWT
Do economies converge? The evidence - Round II: OECD Evidence for convergence is much stronger
Do economies converge? The evidence Round III: The States
Do economies converge? The evidence - Round IV: Japanese Prefectures
Do economies converge? The evidence - Round V: European Regions
Do economies converge? The evidence - Round VI: canadian provinces
The Iron Law of Convergence On average 2% of the gap between poor and rich countries that are similar vanishes in a year. Very slow - takes 35 years to eliminate half of the initial output gap.
Reconciling the evidence When we look at all economies no evidence of convergence. When we examine very similar countries strong evidence of convergence. How can we explain this mixed evidence concerning convergence? - take into account that steady states can differ across countries!
Conditional convergence Neoclassical Solow model does not predict convergence unconditionally. Everything else equal, countries with lower capital stock should grow faster. Countries with same steady-state should converge = conditional convergence In particular, countries with lower TFP (or lower savings) should not catch-up. Importance for growth is distance from steady-state which can be different across countries.
Conditional convergence: an illustration Which country should grow faster? Steady state investment per worker $ = " & 12 = & 345464 & 784564 Initial conditions: 784564 < 12 = 345464 784564 < 345464 < 12 784564 345464 12 Capital per worker
Conditional convergence Only when countries share the same steady state should we see convergence. Explains why only see evidence for convergence amongst similar countries. Explains why many African countries do not catch up with Europe/U.S. Suggests that wealthier economies will persistently stay wealthier, but within wealthier economies should see evidence of catch up. The big question is what determines a countries steady state?
Estimating the steady state Convergence implies that growth increases with distance form steady-state: Don t have data on Y SS but assume GDP Growth = b (log(y SS )-log(y(0))), b>0 log(y SS )= a1 x Education + a2 x Health + a3 x Investment And can estimate for a cross-section of countries: GDP growth = b x (a1 x Education) + b x (a2 x health) + b x (a3 x Investment) b Y(0) If b> 0 then have support for convergence (conditionally on SS determinants) Can estimate if a1, a2, a3 are significantly different from zero and matter for steady state Typical cross-country growth regressions (initiated by Barro)
Dependent variable: real GDP per capita growth (1960-1990), 100 countries. (in 1960) Source: Barro (1991)
I JUST RAN TWO MILLION REGRESSIONS (Salah-i-Martin (1997)) Always significant Frequently significant Often significant Sometimes significant Education Primary school Enrolment Regional dummies (Latin America, Sub Saharan Africa, Negative) Exchange Rate Overvaluation (Negative) Government Consumption (Negative) Investment Rule of Law Black market premiums (Negative) Health Life Expectancy Political Rights Primary Products (% exports Negative) Religious dummies (confucian, Muslim, Protestant) Openness Degree of Capitalism Financial Sophistication Inflation (Negative) Ethnic Diversity (Negative) Civil Liberties Revolutions, Coups, Wars (Negative) Religious dummies (Buddhism, Catholic) Public Investment
Lecture 2 : Long-term economic growth Growth and factors of production 1. The neoclassical growth model 2. Do countries catch with economic leaders? 3. The Asian growth miracle?
Asian Tigers: growth miracle? Since the post-war period we have witnessed astonishing levels of growth in the Asian Tigers. The press and political commentators: Bad for the West: Take our jobs Western leaders and CEO s should look at the Tigers and learn how to improve efficiency. Economists: No growth miracle. Just accumulation of inputs.
Average Growth GDP per capita 1966-2009 Germany Mexico US Canada France Brazil Italy Chile Japan Singapore HongKong China Korea Taiwan 0% 1% 2% 3% 4% 5% 6% 7% Source : Penn WT 6.3 and DataStream.
70 60 50 40 30 20 10 0 Investment rates 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 % of GDP Source : Penn World Table 6.3 Singapore Hong Kong China
Large amount of compulsory savings in Singapore % of Wages 60 CPF Contributions 50 40 Employe e Employe rs 30 20 10 0 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 Source : Young 1995, Barro and Sala-I-Martin 1991
TFP 40% Growth Accounting 1966-90 for Asian Dragons Employment TFP 18% -4% Employment 38% Hong Kong Capital 62% Capital 42% Employment TFP 17% 23% Employment TFP 19% 28% Singapore Korea Taiwan Capital 59% Source : Young 1995, Barro and Sala-I-Martin 1991 Capital 53%
Asian Tigers: growth miracle? Some differences in the sources of growth. BUT: Increased capital stock MOST important factor for all of them. Krugman: Perspiration rather than inspiration! Too rapid development? No gains from the learning curve? Incorrect balance between innovation and experience? Does it matter? No: They got much richer anyway Yes: Some have paid for it - those that were young early on in the process: Low income, high savings Future? Need to shift from extensive to intensive margin
Summary Countries show enormous differences in their standard of living. Further some poor countries have shown rapid growth while others have remained poor. The neoclassical (Solow) growth model focuses on explaining these differences through capital accumulation, assuming diminishing marginal product of capital. This assumption implies convergence. Countries with high investment rates have a high steady state and are rich. At the steady state all countries grow at the rate of technological progress. The neoclassical model relies on conditional convergence conditional on countries sharing the same steady state then poorer countries will grow faster than rich ones. The steady state depends on many factors. Strong roles are found for education, health and rate of investment but many other additional variables are also found important. The analysis suggests that making a country rich will involve a broad package of economic, social and political policies don t look for magic ingredient X.