The single most important force leading to long-run increases in living standards is economic growth. The benefits of economic growth include increases in income levels (or 'welfare'), greater room for redistributive policies and favourable changes in lifestyle. The costs include forgone consumption today and environmental damage due to increased production. In this analysis I am going to assume that growth is a 'good' thing and it is the desire to know why growth has been achieved by some countries and has not been achieved by others that motivates this inquiry. The investigation is also being carried out to empirically test whether Keynesian demand management or Classical supply-side policies should be used to stimulate economic growth. Finally, this phenomena is being explored because it is not understood. As yet there are no firm theoretical conclusions about the causes of growth or about the reasons why growth rates differ. My investigation is therefore aiming to shed some light on a currently uncertain area.
My dependent variable in this investigation is the average annual rate of economic growth between 1970 and 1985. This is measured as the average rate of change in Real GDP per capita (RGDPCH). I chose to average growth over this time-span rather than investigate it in any one year because there could be too many unique factors influencing it in any one year that would defer it from its normal pattern. Rates are used instead of absolute growth levels because they show how growth is related to the country's initial level of income, they are easier to use in a cross-sectional analysis and they are the most commonly used measure of growth in economic analysis.
Economic growth is a measure of the change in national income over time. In my analysis the change in RGDPCH is used as the most preferred measure for national income because it eliminates the influence of inflation or deflation on changes in GDP by taking a real level, and the per capita figure is used because it is a relevant measure of welfare change.
In order to understand why economic growth occurs independent X-variables must be introduced into the investigation. Economic theory has proposed numerous variables which may determine the rate of economic growth. These range from population growth to consumption expenditure to confidence levels to political stability. It must also be emphasised that different factors may influence the country's growth rate depending on the country's level of development. Thus, it can be seen that by limiting my exploration to three independent variables my analysis is going to be imperfect due to the absence of many other possible determining factors.
X1 : My first X-variable is the average annual rate of investment expenditure (as a % of RGDP) between 1970 and 1985 (I). This variable was chosen because it is often assumed by theorists and policy makers to have a large influence on growth. The promotion of investment to increase RGDPCH is a central demand management policy that is supported by Keynesians. Thus, the relationship between I and economic growth is being investigated to test the Keynesian viewpoint. The relationship is also being explored to decipher whether it is worth our while taking money, which could be used for consumption today, to put into investment so that consumption could be higher in the future. That is, will economic growth (a benefit of investment) occur to outweigh forgone consumption today (a cost of investment).
X2 : My second X-variable is the average rate of government expenditure (as a % of RGDP) between 1970 and 1985 (G). Like investment, this variable was chosen because it is often cited as having a large influence on growth. Specifically, Keynesians argue that government expenditure is needed to stabilise an economy and that it will cause RGDPCH to increase. This variable is also being used in my analysis because I want to see if the growing levels of public sector expenditure (PSE) worldwide have been the cause of economic growth. If this is the case then the hypothesis that increased G will lead to a better future would provide a justification for large levels of PSE.
X3 : My third, and final X-variable is the average number of school years in 1970 (HK). This variable was chosen as a 'proxy' measure of human capital in a country. I chose to use the level of human capital in one year rather than over the 15 year period because it is the educational level of the workforce (assumed to be determined by the length of time they spent in school before entering the labour market) over the 15 years which influences the rate of change of income and not the educational level of the students in school over the period. It is generally hypothesised that an economy will grow if its workforce is well - educated. The purpose of the investigation into the relationship between growth and HK is to test this classical supply-side argument.
In this investigation I am using the estimation technique of ordinary least squares. This allows me to use regression analysis to get a line of 'best fit' for my data. The model takes the form of
Y = 0 + 1X1 + 2X2 + 3X3 + e
for the multiple regression case. The purpose of this exploration, then, is to estimate the size and sign of the unknown parameters (0 , 1 , 2 , 3 ).
In my exploration of economic growth I have 40 observations. I left myself with this many countries because I feel that the growth rate is too diverse on a global scale to narrow my analysis of it down to a particular region of the world. Instead, I wanted to get a general overview of world economic growth rates. This is done by breaking my observations down into four subsections (10 countries in each) with each section representing a national income category as defined by the World Development Report (1993).
The data for RGDPCH (from which the growth rates were derived), I and G were all taken from the Penn World Tables (Mark 5). These data are consistent because the compilers have taken figures from the benchmark studies of the UN International Comparison Program for 138 countries and have adjusted them all for different purchasing power parities and they have all been indexed linked to 1985 international prices. It should be noted, however, that the data are only as reliable as the UN figures, which in most cases, are only as reliable as the individual country's National Accounts. Thus, for some countries, most notably the Centrally Planned Economies, the RGDPCH figures and thus, the growth rates may be overestimated.
The data for the average number of school years in 1970 was taken from an NBER working paper by Barro and Lee. The authors got the figures from national censuses so the data are only as reliable as these censuses. The measure used should be fairly consistent as interpretations of this straight forward measure can not differ too much from country to country.
The results of the multiple regression analysis are shown in the table below. A priori, the problem of multicollinearity is recognised by the author although it shall not be dealt with in detail. The line of best fit has been estimated as :
Y = 0.58 + 0.06 X1 - 0.05 X2 - 0.05 X3 + e
where 'e' is the error term or residual of the regression. The correlation coefficient , R2, indicates that 32% of the variation in Y can be explained by the linear influence of all the Xs.
Dependent Variable: Rate of Economic Growth
Independen Parameter t - t Variable Estimate statistic H0 : i = 0 Constant 0.5786 0.93612 X1 = I 0.0583 2.75994 X2 = G -0.0450 -2.25457 X3 = HK -0.0513 -0.73746R2 = O.3246
To evaluate the regression results I am going to compare my estimations with those predicted by theory. In all, the three X-variables have, as theory claims, a relatively significant impact on economic growth (32%). This result is especially interesting when we consider that consumption expenditure, the largest influence on RGDPCH, has not been included in the analysis. Looking more closely at the results, however, we see that the outcome is not so positive. The Keynesian theory that increased G means increased RGDP and the Classical Theory that increased HK would lead to economic growth are both contradicted by my regression results (i.e. the parameter estimates have a negative sign).
In the case of human capital, no result can be taken as absolute because of the 'proxy' nature of my variable. It should be noted, however, that the number of years spent in school may not have an influence on growth. It may be the quality of the education and not the length of it that matters for increasing productivity and thus, growth. In the case of government expenditure there is probably some truth in the result that increased G will not automatically stimulate economic growth. This is particularly true if the expenditure was aimed at sectors that had no growth potential or if it was mismanaged. Thus, badly spent PSE, no matter how large it is, may not lead to an increase in RGDPCH.
The case for investment is far more hopeful. Although the parameter estimate is small (0.06), indicating a not so great influence of I on growth, the sign of the coefficient corresponds to the one proposed by economic theory. That is, the positive nature of 1 coincides with the Keynesian theory that an increase in investment will stimulate economic growth.
To make a simple statistical evaluation of my results I am going to examine the t-statistics given in the table. Considering the hypothesis that there is no relationship between X and Y (H0 : i = 0) we can look at the t-statistic which is the ratio of the estimate to the standard error. An estimate of a parameter is statistically significant if the t-statistic associated with it causes us to reject, at a particular significance level, the hypothesis that i is equal to zero. At both the 5% and 10% significance levels the estimate for 1 is statistically significant while, at the same significance levels, the estimates for 2 and 3 are statistically insignificant.
In today's world, economic strength is the key to political power. Thus countries that are striving to become powerful and to increase living standards all look for economic growth. A final means of evaluating my results, therefore, is to see if they are relevant to policy makers who are trying to achieve this primary policy objective.
My research would indicate to policy makers that the best method of getting growth is to encourage investment. It should be noted, however, that we can not assume that increased investment will automatically lead to large scale growth. Only if investment is directed to the right areas will RGDPCH be increased. Secondly, my results will, perhaps, encourage policy makers to reduce PSE. They could not justify large scale expenditure on the grounds that it will lead to an improved future using my analysis. Instead, it would be suggested that lower but better managed PSE is what is needed to stimulate economic growth. Finally, I would discourage policy makers from taking my results vis à vis human capital too seriously. They should be aware of the very slight possibility that the results are right and could use this as a guideline to improve the quality rather than the quantity of education. They must not, however, take the results so seriously that they would actively discourage people from staying in education. This would have disastrous results for society and for the economy.
It is hoped that my research provides an insight for policy makers and other economists into the workings of economic growth across the globe. The results can be taken as possible indicators of how growth may be generated but they should not be taken as being infallible. Like all econometric work, my exploration into economic growth should only become completely acceptable if and when it is verified by the research of others.
Barro, R & Lee, J (1993) International Comparisons of Educational Attainment Mankiw, G (1992) Macroeconomics Summers, R & Heston, A (1991) "The Penn World Table (Mark 5): An Expected Set of International Comparisons, 1950 - 1988" in the Quarterly Journal of EconomicsWorld Bank (1993) World Development Report