Labor Productivity Divergence: Role of Age Dependency

Transcription

1 Labor Productivity Divergence: Role of Age Dependency Misbah Tanveer Choudhry University of Groningen, The Netherlands October 2009 Abstract This study finds strong empirical evidence in favor of the hypothesis that age composition of population matters for labor productivity growth. We applied the fixed effects panel model using data of large number of countries over the period Our results suggest that higher age dependency impacts the labor productivity negatively not only directly but also modifies the impact of other determinants of labor productivity. Child dependency has more adverse effect on labor productivity as compared to old age dependency. We specifically find that marginal effect of gross capital formation, labor market reforms and information and communication technology investment on labor productivity is high and significant at lower level of age dependency. However, the marginal effect of stock market capitalization on labor productivity increases at high level of age dependency in developing economies. Diversity in size and nature of age dependency across regions and different income groups helps to explain labor productivity differential across them. JEL Classifications: C22, C23, O47 Keywords: labour productivity growth, Age dependency, interaction term, panel fixed effects Correspondence : Misbah Tanveer Choudhry,University of Groningen, Department of Economics and Econometrics, Faculty of Economics & Business, P.O. Box 800, 9700 AV, Groningen, The Netherlands, phone: , m.t.choudhry.rug@gmail.com. 1

2 1. Introduction This paper explores the relationship between age composition of population and labor productivity and evaluates its direct as well as indirect role in explaining the cross country disparity in labor productivity. Total Factor Productivity (TFP) plays an important role in explaining the differences in output per capita. This emphasis on TFP comes from the literature which shows that an understanding of differences of output per worker is needed as only workers can contribute to production (see e.g. Prescott 1998, King and Levine 1994, Hall and Jones, 1999, Islam 2008 and Kogel 2005). Easterly and Levine (2001) find that TFP explains the cross country income differences rather than factor accumulation. Solow (2001) also suggests that comparative growth studies should focus on analyzing the various sources of divergence of total factor productivity. Moreover, almost every country in the world is passing through demographic changes. While developed nations are characterized by a rapidly increasing aging population share, in developing and emerging economies, youth dependency is still quite high, despite showing a declining trend. An understanding of how these changes in age composition will impact the labor productivity over the coming decades is very important for a meaningful analysis of economic growth. In this study, we use a panel of 110 countries 1 to examine the determinants of labor productivity (output per worker) during Along with other socio- economic determinants of labor productivity, our focus is mainly on the role played by the demographic structure of a country. The effect of participation and age composition of labor force on labor productivity has been analyzed in literature (see e.g. Feyrer 2007, Choudhry and van Ark, 2009, Kogel 2005). Our hypothesis is that age structure of total population not only directly affects the cross country labor productivity differential across the world but also modifies the effectiveness of other labor productivity determinants. To the best of our knowledge no study before has investigated this indirect impact. Moreover, except for some studies (see e.g. Islam, 2008) the analysis of productivity determinants is mainly focused on high-income OECD countries. In this paper, in addition to focusing on mixed sample, we also check the differences in the effect of certain determinants of productivity in high-income OECD 2 and other countries separately. 1 Sample of countries varies from 71 to 110 countries depending on data availability for explanatory variables. 2 For the sake of brevity, we will refer to high-income OECD countries as developed economies in the rest of the chapter and the other economies will be referred as developing economies. 2

3 Our results suggest that higher age dependency impacts the labor productivity negatively not only directly but also modifies the impact of other social, economic and infrastructural determinants of labor productivity. Child dependency has more adverse effect on labor productivity as compared to old age dependency. We specifically find that marginal effect of gross capital formation, labor market reforms and information and communication technology investment on labor productivity is high and significant at lower level of age dependency. This finding holds for both developed and developing economies. Marginal effect of stock market capitalization on labor productivity increases at high level of age dependency in developing economies. Moreover, in high income economies the slope of the marginal effect line is little bit flatter as compared to other countries. However, in developed economies, marginal impact of stock market capitalization is lower at high age dependency. The structure of the rest of the chapter is as follows. Section 2 discusses the labor productivity growth trends across countries and regions during and presents a brief literature review of labor productivity determinants which provide rationale for the selection of various explanatory variables. Section 3 develop our empirical model and describe our data used in empirical estimation. Section 4 presents the empirical results and also offers a sensitivity analysis. Section 5 concludes the chapter. 2. Labor productivity trends across regions In this section, we present the labor productivity growth trends across different regions for Productivity growth not only varies across different regions, but even within the regions, there is a reasonable diversity with respect to productivity performance. To examine the performance of labor productivity across different regions and countries we constructed a Labor Productivity Index (LPI) and it is presented in Figure 1 below. In South America, Chile showed an impressive and consistent increase in labor productivity during period. Peru also performed well but Ecuador s performance remained poor and its productivity in 2005 is lower as compared to 1980s. In Southeast and East Asian region, all economies registered an increase in productivity during 1990s. In East Asia, all economies in the region were more or less on an equal growth path until 1993, but China s productivity growth has increased more rapidly since then. In Southeast Asia, there has been a moderate upward trend in productivity since 1990, slower than other Asian sub regions but faster than non-asian developing economies. 3

4 The South Asian region has seen improvements in terms of productivity growth since The productivity levels vary considerably between countries. India s output per person employed increased by more than 75 percent during , which was faster than Pakistan, Sri Lanka and Bangladesh. Pakistan and Sri Lanka started off well in the early 1990s but entered into a period of productivity decline in the late 1990s. Pakistan revived again in 2001 but Sri Lanka s productivity has stagnated since Bangladesh showed a slow and gradual increase in labor productivity during the same time period. In Europe, all the economies showed upward trend in productivity. United Kingdom s productivity growth remained better as compared to other European countries. In contrast, Spain s labor productivity is stagnant since 1995 and there is a decline in LPI for Italy after A closer focus on the transition economies from Eastern and Central European countries reveals that they made rapid advances in productivity growth during 1990s. Poland today stands out as a success story among transition economies, coming from relatively low levels of productivity and having undergone rapid structural transformation from a large and low productive agricultural sector. But Poland still has a relatively high unemployment rate and labor productivity is still much below than the EU average. In most transition economies, high labor productivity growth is the result of a rapid shake-out of employment from non-competitive plans leading to a decline in employment during the transition process. In Middle East and North Africa, labor productivity in selected economies showed mixed trends. There has been a steady upward trend in Egypt and Tunisia, a declining trend in Syria since 1998 and a volatile pattern in Morocco. Africa is the only region that had seen a decrease in labor productivity since Productivity levels in 2005 are lower than in Only Tanzania, Nigeria and South Africa showed slight increases in productivity level after

5 Figure 1: Growth in output per person employed in different regions (selected Economies, Index 1990=100) 140 Growth in output per person employed in Africa (selected economies,index 1990=100) 200 Growth in output per person employed in South Asia (selected economies,index 1990=100) Index(1990=100) Index(1990=100) Kenya Madagascar Nigeria South Africa Tanzania Zambia Zimbabwe Bangladesh India Pakistan Sri Lanka Growth in output per person employed in Europe (selected economies,index 1990=100) Growth in output per person employed in East and South East Asia (selected economies,index 1990=100) Index(1990=100) Albania Bulgaria Hungary Poland France Romania Italy Belgium Netherlands Spain United Kingdom Index(1990=100) China Indonesia Japan Malaysia Philippines South Korea Thailand Index(1990=100) Growth in output per person employed in Middle East and North Africa (selected economies,index 1990=100) Index(1990=100) Growth in output per person employed in South America (selected economies,index 1990=100) Egypt Morocco Syria Tunisia Argentina Brazil Chile Colombia Ecuador Peru

6 2.1 Determinants of labor productivity The literature focusing on labor productivity has mainly used two methods for analysis. First one is a short term business cycle approach. In this approach, productivity growth has been often related to the growth of employment (see, e.g. Beaudry et al., 2002, Becker and Gordon, 2008 and Choudhry and van Ark, 2009). Second one can be described as the long term growth approach. In this approach, growth models starting from neoclassical growth model have focused on the main determinants of labor productivity growth both in the steady state as well in the transition state. In a conditional convergence version of Solow model (Barro,1991), some other exogenous variable models are also introduced in the growth literature. The most discussed variable in recent studies as a potential determinant of labor productivity is ICT investment (see e.g. Ark et al., 2003, Belorgey et al., 2006, and Jorgenson, 2007). These studies show that ICT investment and diffusion explain the part of difference between Europe and US. So we incorporate the variable of per capita ICT expenditure as a percentage of GDP as an explanatory variable and software investment for estimating the differences in productivity levels in Positive impact of new technologies for developing economies depends on other conditions, e.g. availability of skilled labor. It is also argued that many technologies adopted by the low developing countries are developed in the OECD economies and are designed to make an optimal use of skills of these richer countries workforces. Low income developing economies lack required skilled workers for an efficient utilization of these technologies. Technology-skill mismatch could account for a large part of observed output per worker differences around the world (Acemoglu and Zilibotti, 2001). The role of various socio economic indicators can not be denied in evaluating the economic and labor productivity growth (Barro 1995, Bourles and Cette 2007 and Marelli et al., 2008). We include inflation rate as a measure of price volatility and also include the gross capital formation, urbanization, financial depth and foreign direct investment in our empirical analysis to evaluate their impact on labor productivity growth. There is, however, not much research on the effects of the age structure on the aggregate productivity in economies. These studies focus on the impact of age structure on economic growth (see, e.g. Bloom et al and Choudhry and Elhorst 2009) or on firm (plant) productivity (see, e.g. Skirbekk 2003, IImakunnas and Maliranta 2005). Moreover there are some studies examining the effect of age structure of labor force on labor productivity (see e.g. Feyrer 2007 and Choudhry and van Ark 2009). Kogel (2005) finds that the relationship between youth 6

7 dependency and total factor productivity is negative. Moreover, some studies find that age structure of entire population affects output (Persson, 2002 and Sarel, 1995). This paper differs from above mentioned literature as it is focusing on the age composition of entire population for labor productivity and introduces an interactive model of age composition with other determinants of labor productivity. We argue in this paper that these determinants have not only a direct impact on labor productivity but also an indirect impact, conditional on age composition of total population. To the best of our knowledge, no study before has investigated this indirect impact. Our results suggest that higher age dependency impacts the labor productivity negatively not only directly but also lowers the impact of other social, economic and infrastructural determinants of labor productivity. We find that the performance of other explanatory variables is conditional on different levels of age dependency. At higher level of age dependency, we show that the impact of certain productivity determinants becomes insignificant. 3. Model specification Our dependent variable is labor productivity growth. We take the output per worker as a measure of labor productivity. Alternatively, as a part of the sensitivity analysis, we also take output per hour as a dependent variable. In our baseline model, we assume that performance of labor productivity is dependent on economic indicators, labor market institutions, age composition of population, and financial development and information and communication technology development in an economy. We argue in this paper that these determinants have not only a direct impact on labor productivity but also an indirect impact conditional on age composition of total population. So our baseline model is ΔLP + α i β β β β i t = + ΔERit + ipeyit + Inf it + ADit + Ψ (1), it β ε it Where i represent country and t is time period. LP denotes the labor productivity (output per worker) growth; ΔER is change in participation rate measured as total employment divided by total population, ipey is an initial level of labor productivity in an economy, Inf is inflation rate and AD is the ratio of dependent population to working age population. Ψ is a vector of our 7

8 variables of interest 3 which includes gross capital formation, ICT expenditure, labor market regulations and stock market capitalization. it represents country specific fixed effects. These fixed effects allow for different labor market institutions and cultural and social norms across countries. To check whether performance of other productivity determinants is conditional on the age structure of the population, we introduced interaction term of age dependency with other control variables of our interest. Consequently, our model for estimation with interaction term becomes ΔLP i t = α + β ΔERit + β ipeyit + β Inf it + β ADit + β Ψit + β ψ i it * ADit + + it ε (2), it Where Ψ *AD is interaction term of our variable of interest with age dependency in equation 2. As mentioned above that we are mainly interested in evaluating the effect of four explanatory variables on different level of age dependency. So we will estimate the four separate models with separate interaction terms 4. For example to evaluate the effect of ICT expenditure on labor productivity at different levels of age dependency we estimate the following model α i β β β ΔLP + β i t = + ΔERit + ipeyit + Inf it + ADit + ICTit + ICTit * ADit + (3), it β β ε it After estimation of these models we are interested in marginal effect of our variables of interest on our dependent variable at different level of age dependency. For this purpose we will differentiate the equation (2) with respect to that particular variable. For calculation of marginal effect of ICT expenditure on labor productivity conditional on age dependency, we need to take derivative of equation (3) with respect to ICT expenditure, so we have β β 6 LP AD ICT = + (4) 5 Equation (4) will give the marginal effect of ICT on labor productivity at different levels of age dependency. Marginal effect of other variables of interest can be calculated in a similar way. 3 For these variables we want to check their effect on labor productivity at different level of age dependency of entire population. 4 1) Age dependency*gross capital formation; 2) age dependency* ICT expenditure; 3) age dependency* labor market reforms and 4) age dependency * stock market capitalization. 8

9 3.1 Data description and analysis We use data on labor productivity (Output per worker) from Groningen Growth and Development Centre (GGDC). Labor productivity varies significantly in sample countries. In high income economies average labor productivity level was 11 times higher as compared to low income economies in The divergence in labor productivity increased further and reached to 13 times higher in high income economies in comparison with low income economies in We used employment to population ratio as a measure of labor participation rate and also it has been calculated from GGDC. As discussed in section 3 that potential determinants of labor productivity are macroeconomic indicators, labor market institution, ICT development and financial depth. We include three macroeconomic variables: (i) adjusted inflation 5 as a proxy for the changes in the price level in the country; (ii) gross capital formation growth to capture macroeconomic developments; and (iii) Initial level of GDP per employed. Inflation and gross capital formation data is taken from World Development Indicators (WDI). To evaluate the role of labor market institutions and regulations, we incorporate Labor Market Regulations (LMR) index as an explanatory variable. LMR is an index based on further six measures of labor markets ( Minimum wage, hiring and firing regulations, centralized collective bargaining, Mandated cost of hiring, Mandated cost of worker dismissal and Conscription). LMR is un-weighted average of these six measures and its value varies from It has been taken from Economic Freedom Index of Fraser Institute. Additionally, we use ICT expenditure (as a percentage of GDP) as a proxy for ICT development and its data is taken from WDI. To check the role of financial development for labor productivity, we use Stock Market Capitalization to GDP ratio as an explanatory variable. Data is taken from the World Bank Financial Structure database. P /100 5 To adjust for extreme movements, we modify the inflation rate (P) as. 1 + ( P /100) 9

10 Figure 2: Labor productivity and age dependency by income groups Labor Productivity and age Dependency by Income Group ( ) LP AD(sec axis) US $ percent HYE UMYE LMYE LYE Age dependency is measured as a ratio of dependent population to working age population. Dependent population is defined as a sum of young population (0-14 years) and aged population (65+ years). Working age population is considered as population of age years. Data on age dependency is also taken from WDI. To evaluate the relationship between labor productivity and age dependency, a scatter plot is presented in Figure A1 in appendix which suggests a negative relationship between labor productivity and age dependency. Age dependency in low income economies is nearly double as compared to high income economies and labor productivity is also quite low in comparison with high income and upper middle income economies see Figure 2. The precise definitions and data sources of all variables used are given in Table A1 in the appendix. Table A2 in the appendix shows the correlation matrix of our variables. The low correlation of the explanatory variables suggests that multicollinearity is not a problem in our estimations. 10

11 4. Empirical results We estimate equation (3) using a fixed effects panel model for period for a panel of countries. Fixed effects model has been selected on the basis of Hausman test. The Hausman test statistic shows that a fixed effects model should be used instead of a random effects model (see Table 1). The results of empirical estimation are given in Table 1. First we do a regression analysis with age dependency and other control variables without any interactive term (see columns 1-4 of Table 1). The F-statistics indicate overall significance of the models at 1 percent level of significance. We find that initial labor productivity variable is significant and negative which reflects the conditional convergence in per capita productivity growth in our sample countries. We find that the impact of age dependency is negative and significant which implies that higher dependency burden in population lead to decline in labor productivity growth. Other control variables also have expected signs. In line with our expectations, the results suggest that higher inflation impact the productivity negatively, while higher gross capital formation, ICT development and stock market capitalization promote labor productivity. However, the impact of LMR does not appear significant in our empirical analysis (see column 1-4 in table 1). To check whether performance of other productivity determinants is conditional on the age structure of the population, we introduce an interaction term of age dependency with other control variables and estimation results are presented in column 5-8 of Table 1. Before we turn to the results regarding the impact of various indicators conditional on age dependency, it is important to note that inference cannot be based on simple t-statistics because model parameters do not provide substantial information in case of models with multiplicative terms (Brambor et al., 2006 and Shehzad and de Haan 2009). As Aiken and West (1991) point out, in interactive models we need to take the derivative of the model with respect to the variable of interest and evaluate its effect on the means of other constituent terms of the derivative. 6 Sample of countries varies from 71 to 110 countries depending on data availability for explanatory variables. 11

12 Table 1. Role of demographics in determination of labor productivity Direct Model Interactive model Age Dependency *** *** *** *** *** ** *** Change in Participation *** *** *** *** *** *** *** *** Initial level of labor productivity *** *** *** *** *** *** *** *** Inflation *** *** *** *** *** *** *** *** Change in Gross Capital Formation 0.001*** 0.002*** 0 0 Gross capital Formation*Age Dependency *** 0 ICT Expenditure 0.010*** 0.028*** ICT expenditure*age Dependency * Labor Market Reforms ** Labor Market Reforms*Age Dependency ** Stock Market Capitalization 0.020*** Stock Market Capitalization*Age Dependency Constant 0.256*** 1.647*** 0.602*** 0.606*** 0.258*** 1.641*** 0.525*** 0.607*** Significance of Fixed Effects P-value Hausman Test Joint significance of Interaction and Constituent Term Number of observations Number of Countries R F Statistic 68.64*** 13.07*** 7.34*** 25.58*** 59.16*** 13.09*** 7.71*** 21.39*** Robust standard errors are reported below the coefficients. *** indicates significance at 1 percent level, ** indicates significance at 5 percent and * indicates a significance at 1 percent level 12

13 Our key hypotheses relate to the significance of the marginal effect of our variables of interest represented by Ψ on our dependent variable. By taking a derivative of our dependent variable with respect to variable of interest as mentioned in equation (4), we will get our hypothesis to test H 0 : β 5 + β 6 *AD = 0 H 1 : β 5 + β 6 *AD 0 Rejection of the null hypothesis implies that impact of our variable of interest on labor productivity is significant and conditional on the size of age dependency. In order to assess the significance of the variables of interest, we need to draw confidence intervals, for which standard errors can be calculated following the methodology of Aiken and West (1991). On the basis of results presented in column 5-8 in table 1, we calculated the marginal impact of the multiplicative term. Figure 3 presents the marginal effect of various explanatory variables on labor productivity, conditional at different levels of age dependency. In Figure 3, the solid line indicates marginal impact of our variable of interest on labor productivity at different level of age dependency. 95% confidence intervals around the line allow us to determine the conditions under which our variable of interest has a statistically significant effect on labor productivity - they have a statistically significant positive (or negative) effect whenever the upper and lower bounds of the confidence interval are both above (or below) the zero line. The upper left panel in Figure 3 shows the marginal effect of growth in gross capital formation on labor productivity at different levels of age dependency. This figure corresponds to main result in column 5 of table 1. Downward slopping, statistically significant marginal effects of GCF reflects that as the age dependency level in an economy increases, it lowers the positive impact of investment on labor productivity growth. A possible explanation for this significant downward sloping marginal effect can be that the higher share of working age population can utilize the gross capital formation more efficiently and it promotes high growth in labor productivity. 13

14 Marginal Effect of ICT Marginal Effect of GCF Figure 3: Marginal Impact of Gross Capital Formation, ICT expenditure, labor Market Reforms and Stock Market Capitalization on Labor Productivity at different level of Age DependencyComplete Sample Age Dependency Marginal Effect of SMC Age Dependency Marginal Effect of LMR Age Dependency Age Dependency This figure examines the impact of gross capital formation, ICT expenditure, Labor market reforms and stock market capitalization on labor productivity growth at different level of age dependency. These figures correspond to our main results as given in column 5-8 in Table 1. The upper panels shows the marginal effect of Gross Capital Formation (GCF) and ICT expenditure (ICT) and the lower panels shows the marginal effect of Labor Market Reforms (LMR) and Stock Market Capitalization (SMC) at different level of age dependency. The marginal effect of ICT on labor productivity at different levels of age dependency is presented in upper right panel of figure 3 (corresponding to main results in column 6 of table 1). Downward slopping marginal impact line suggests that ICT helps more to promote labor productivity growth at lower level of dependent population. As dependency burden in an economy increases, it depresses the positive role of ICT for productivity growth. To absorb and utilize the ICT development properly, skilled active labor force is a basic requirement which is difficult to be fulfilled in the presence of high dependent population. 14

15 The effect of labor market regulations on labor productivity also changes at different level of age dependency in an economy. Contrary to the results in model without interaction term (column 3 of table 1), we find that labor market regulations promote labor productivity and their impact is higher in the presence of high working age population (see lower right panel of Figure 3). In countries with high dependency rates, the role of labor market regulations is relatively small in promoting labor productivity. Financial development promotes labor productivity (see e.g. Nourzad, 2002 and Arizala et al, 2009). We use Stock Market Capitalization (SMC) to GDP ratio as a proxy for financial development. We find that its marginal effect is positive at high level of age dependency. It reflects that an increase in SMC to GDP ratio increases the labor productivity at different levels of age dependency but it is significant only at lower age dependency. In financially integrated world, cross country age structure differences motivate to invest in younger economies, characterized by higher capital demand and higher expected asset returns (Taylor and Williamson, 1994, Higgins, 1998, IMF, 2004 and Santis and Luhrmann, 2009). It implies that as more investment is made through stock markets in firms, it increases the labor productivity. A possible explanation can be the shareholders monitoring of firm performance. Moreover, to maintain investor s interest in the firms shares management opts for more competitive production strategies. The finding of high stock market capitalization at high dependency level is consistent with the findings of De Santis and Luhrmann (2009) that youth and old age dependency ratios affect countries net equity flows. 4.1 Robustness and extension This section presents a number of robustness checks. We first examine whether our results hold when we restrict our sample to only developed or developing economies. The results are presented in appendix in Tables A3 and A4 respectively and the marginal effects corresponding to these results are presented in Figures 4 and 5 respectively. Our findings almost remain the same for both groups of economies with the exception of marginal effect of stock market capitalization on labor productivity at different levels of age dependency. 15

16 Figure 4: Marginal Impact of Gross Capital Formation, ICT expenditure, labor Market Reforms and Stock Market Capitalization on Labor Productivity at different level of Age Dependency- OECD Countries Marginal Effect of GCF Marginal Effect of ICT Age Dependency Age Dependency Marginal Effect of LMR Marginal Effect of SMC Age Dependency Age Dependency This figure examines the impact of gross capital formation, ICT expenditure, Labor market reforms and stock market capitalization on labor productivity growth at different level of age dependency. These figures correspond to our main results as given in column 5-8 in Table A3. The upper panels shows the marginal effect of Gross Capital Formation (GCF) and ICT expenditure (ICT) and the lower panels shows the marginal effect of Labor Market Reforms (LMR) and Stock Market Capitalization (SMC) at different level of age dependency. We find that marginal effect of gross capital formation, ICT expenditure and labor market reforms declines at high level of age dependency. Separate analysis for devloped and developing economies also provide very similar kind of marginal impact of GCF on labor productivity conditional on age dependency. However, in case of developed Countries, the slope is a little bit flatter indicating a slower decline in labor productivity as compared to developing countries (see upper left graphs of Figure 4 and 5 respectively). Same is true for ICT expenditure and LMR (see upper right and lower left graphs in Figure 4 and 5 respectively). Difference in marginal impact of SMC is mainly because of different nature of age dependency in developed and developing countries. Developed countries are characterized by high old age dependency contrary to developing countries where main high child dependency represents the age dependency. 16

17 Moreover, we have also checked the impact of child dependency and old age dependency separately. Marginal effects of various determinants of labor productivity at different levels of child dependency and old age dependency are presented in Figures A2-A2.1 and A3-A3.1 in appendix 7. We find that child dependency affects the labor productivity adversely and it is also true for sub samples of developed and developing countries. However, in case of low levels of old age dependency, their impact is not negative in case of mixed sample and developing countries. In case of developed countries, marginal effect of GCF is slightly lower at different level of old age dependency. Stock market capitalization impact on labor productivity declines at high old age dependency and it is true for complete as well as for sub-samples. We have also checked the marginal effects of various determinants of labor productivity at different level of working age population share (results available on request). Findings reconfirm that at higher levels of age dependency the marginal impacts of some socio-economic determinants of labor productivity decline or become insignificant. Finally, we change our dependent variable from output per worker to output per hour. There is a significant decline in the number of observations, as data is not available for a large number of countries. Findings are very similar to the results obtained for developed countries as data on output per worker is only available mostly for developed economies. Our results still hold for ICT expenditure and Labor market reforms. The slope of SMC becomes flatter and significant only at low level of age dependency. Again, it appears that most results are very similar to our previous findings. 5. Conclusion We examine the effect of gross capital formation, labor market institution, ICT expenditure and financial development on labor productivity (measured by output per worker) conditional at different levels of age dependency for about 110 countries in the period We find that age dependency of population matters for labor productivity growth. This result is in line with the findings of Kogel (2005). Our results suggest that higher age dependency impacts the labor productivity negatively not only directly but also lowers the impact of other social, economic and infrastructural determinants of labor productivity. We find that the performance of other explanatory variables is conditional on different levels of age dependency. At higher level of age dependency, we show that the impact of certain productivity determinants becomes insignificant. 7 The empirical results tables are not presented here but are available on request. 17

18 In our sensitivity tests, we use a number of variations in the definitions of the variables used and samples. It turns out that our results are generally very robust and hold for both sub samples of developed and developing countries. We also find that child dependency has an adverse impact on labor productivity as compared to old age dependency. These findings imply that as age dependency in low income economies is 1.8 times higher as compared to high income developed economies, it helps to explain the productivity differential across different income groups. Our findings also suggest that in future with a decline in child dependency in developing and emerging economies, there is an expected increase in labor productivity not only directly but performance of other determinants of labor productivity will also improve. In developing and emerging economies, foremost priority should be given to an increase in the education and skill training of working age population so that they can properly utilize the demographic changes. References Acemoglu, D. and F. Zilibotti (2001) Productivity differences. Quarterly Journal of Economics 116: Aiken, L. and S.West (1991) Multiple regression: Testing and interpreting interactions. Sage Publications, London. Ark, van. B, R.Inklaar, and R.H.McGuckin ( 2003) The contribution of ICT-producing and ICTusing industries to productivity growth: A comparison of Canada, Europe and the United States. Centre for the Study of Living Standards. International Productivity Monitor 6: Barro R.J. (1991) Economic growth in a cross section of countries. Quarterly Journal of Economics 106: Barro, R, and X. Sala-i-Martin (1995) Economic growth. New York: McGraw-Hill Beaudry P. and Collard F (2002) Why has the employment-productivity tradeoff among industrialized countries been so strong? NBER working paper Becker I.D. and R. J. Gordon (2008) The role of labor market changes in the slowdown of European productivity growth. CEPR discussion paper Belorgey, N, R. Lecat, and T.Maury (2006) Determinants of productivity per employee: An empirical estimation using panel data. Economics Letters 91: Bloom, D.E. and J.G.Williamson (1998) Demographic transition and economic miracles in emerging Asia. World Bank Economic Review 12: Bourles, R. and G. Cette (2007) Trends in "structural" productivity levels in the major industrialized countries. Economics Letters 95: Brambor, T, W.R.Clark and M.Golder (2006) Understanding interaction models: improving empirical analyses. Political Analysis 14:

19 Choudhry M.T. and J.P.Elhorst (2009) Economic development, fertility decline and female labor force participation. Paper presented at European Economic Association, Annual Conference held in Barcelona, Spain. Choudhry M.T. and B. van Ark (2009) Economic Growth and Structural Features of Transition, Palgrave/Macmillan, forthcoming. De Santis, R.A. and M.Lührmann (2009) On the determinants of net international portfolio flows: A global perspective. Journal of International Money and Finance 28: Easterly, W and R. Levine (2001) It s not factor accumulation: Stylized facts and growth models. The World Bank Economic Review 15: Feyrer, F. (2007) Demographics and productivity. The Review of Economics and Statistics 89: Francisco, A, E. Cavallo and A.Galindo (2009) Financial development and TFP growth: Crosscountry and industry-level evidence. RES Working Papers 4630, Inter-American Development Bank, Research Department. Hall, R and C.I. Jones (1999) Why do some countries produce so much more output per worker than others? Quarterly Journal of Economics 114: Hausman, J. (1978) Specification tests in econometrics. Econometrica 46: Higgins, M. (1998) Demography, national savings and international capital flows. International Economic Review 39: IImakunnas, P. and M. Maliranta (2005) Technology, labour characteristics and wageproductivity gaps. Oxford Bulletin of Economics and Statistics 67: International Monetary Fund (2004) How will demographic change affect the global economy? IMF World Economic Outlook: September. Islam, N. (2003) What have we learnt from the convergence debate? Journal of Economic Surveys 17: Islam, N. (2008) Determinants of productivity across countries: An exploratory analysis. The Journal of Developing Economies 42: Jorgenson, D.W. and K. Vu (2007) Information technology and the world growth resurgence. German Economic Review 8: King, R.G. and R. Levine (1994) Capital fundamentalism, economic development, and economic growth. Carnegie-Rochester Conference Series on Public Policy 40: Kogel, T. (2005) Youth dependency and total factor productivity. Journal of Development Economics 76: Marelli E. and Signorelli M. (2008) Productivity, models of growth and the role of human capital in the European countries. Paper presented at XIII annual conference of labor economics. 19

20 Nourzad, F. (2002) Financial development and productive efficiency: a panel study of developed and developing countries. Journal of Economics and Finance 26: Persson, J. (2002) Demographics, human capital, and economic growth: A study of US States FIEF working paper. Prescott, E. (1998) Needed: a theory of total factor productivity. International Economic Review 39: Sarel, M. (1995) Demographic dynamics and the empirics of economic growth. IMF Staff Papers 42: Shehzad, C.T. and J.de Haan (2009) Financial liberalization and banking crisis. Paper presented at the 2009 Conference of the Royal Economic Society, University of Surrey. Skirbekk, V. (2003) Age and individual productivity: a literature survey. Max Planck Institute for Demographic Research working paper 28. Solow, R.M.(2001) Applying growth theory across countries. The World Bank Economic Review 15: Taylor, A. and J.G. Willamson (1994) Capital flows to the new world as an intergenerational transfer. Journal of Political Economy 102: The Conference Board and Groningen Growth and Development Centre, Total Economy Database, January 2008, World Bank (2007) World Bank Development Indicators CD-ROM. 20

21 Appendix: Table A1. Description of Data and its sources Variable Variable description Source LP GDP per employed person Groningen Growth and Development Centre PART Employed to total population ratio Groningen Growth and Development Centre AD (0-14 years)+ (65+ years) / years World Development Indicators GCF Gross capital formation is the sum of fixed gross capital formation, changes in inventories and acquisition less disposables of valuables World Development Indicators INF consumer prices % annual World Development Indicators ICT Information and communication technology expenditure (% of GDP) World Development Indicators LMR LMR is un-weighted average of six measures (Minimum wage, hiring and firing regulations, centralized collective bargaining, Mandated cost of hiring, Mandated cost of worker dismissal and Conscription )and its value varies from Economic Freedom Index of Fraser Institute SMC Stock Market Capitalization (SMC) as a percentage of GDP World bank financial structure database 21

22 Figure A1. Labor Productivity and Age Dependency Labor Productivity and Age Dependency in 2000 Output per worker (000 US $) Age Dependency Table A2. Correlation Matrix LP PART ILP INF GCF ICT LMR SMC AD Labor Productivity Growth (LP) 1.00 Participation Growth (PART) Initial Labor Productivity Level (ILP) Inflation (INF) Gross Capital Formation Growth (GCF) ICT Expenditure Labor Market Regulations (LMR) Stock Market Capitalization (SMC) Age Dependency (AD) Table A3:. Role of Demographics in Determination of Labor Productivity for High Income OECD Countries Direct Model Interactive model Age Dependency ** Change in Participation *** *** *** *** *** *** *** *** Initial level of labor productivity *** *** *** *** *** *** Inflation *** ** *** *** *** ** *** ***

23 Change in Gross Capital Formation 0.001*** 0.002** Gross capital Formation*Age Dependency ICT Expenditure 0.010** ICT expenditure*age Dependency Labor Market Reforms ** Labor Market Reforms*Age Dependency ** Stock Market Capitalization 0.011*** Stock Market Capitalization*Age Dependency Constant 0.285*** *** 0.683*** 0.282*** * 0.655*** Number of observations Number of Countries R F Statistic *** 8.096*** 9.259*** *** *** 9.349*** *** *** Robust standard errors are reported below the coefficients. *** indicates significance at 1 percent level, ** indicates significance at 5 percent and * indicates a significance at 1 percent level Table A4. Role of Demographics in Determination of Labor Productivity for Non-OECD Countries Direct Model Interactive model Age Dependency *** *** ** *** *** * *** Change in Participation *** *** *** *** *** *** *** *** Initial level of labor productivity *** *** *** *** *** *** *** *** Inflation *** *** *** *** *** *** *** *** Change in Gross Capital Formation 0.001*** 0.002*** 0 0 Gross capital Formation*Age Dependency *** 0 ICT Expenditure 0.011*** 0.036*** ICT expenditure*age Dependency ** Labor Market Reforms ** Labor Market Reforms*Age Dependency ** 23

24 0.017 Stock Market Capitalization 0.020*** Stock Market Capitalization*Age Dependency Constant 0.238*** 1.824*** 0.735*** 0.588*** 0.239*** 1.862*** 0.635*** 0.589*** Number of observations Number of Countries R F Statistic *** *** 7.523*** *** *** *** 7.719*** *** Robust standard errors are reported below the coefficients. *** indicates significance at 1 percent level, ** indicates significance at 5 percent and * indicates a significance at 1 percent level 24