Running head: The Effect of the Internet on Economic Growth The Effect of the Internet on Economic Growth: Evidence from Cross-Country Panel Data Changkyu Choi, Myung Hoon Yi Department of Economics, Myongji University, 50-3 Namgajwadong, Seodaemungu, Seoul 120-728, Korea September 20, 2005 (revised) Abstract Using cross-country panel data, we found evidence that the Internet plays a positive and significant role in economic growth after investment ratio, government consumption ratio, and inflation were used as control variables in the growth equation. Keywords: Internet; Growth; Panel data JEL Classification: C23; O40 Corresponding author. Tel.: +82-2-300-1587; fax: +82-2-300-1912. E-mail address: yimh@mju.ac.kr (M.H. Yi).
1. Motivation The Internet has influenced the economy in every respect. The history of the Internet, however, is not that long. Also, there is little research on the Internet and economy. The effect of computers, as opposed to the Internet, on an economy has been studied. For example, Krueger (1993) analyzed the effect of computer use on wage structure using Current Population Surveys (CPS) data and found that workers who use computers earn higher wages. Sichel (1999) found that computer hardware contributes to economic growth. Oliner and Sichel (2000, 2003) said that productivity growth after 1995 in the US has been driven in large part by greater use of information capital goods. According to Freund and Weinhold (2000) and Choi (2003), the Internet has had a positive effect on bilateral trade and foreign direct investment, respectively. With broad use of the Internet starting in the 1990s, we used crosscountry panel data to gather enough observations to analyze the Internet growth nexus. In section 2, we derived a simple growth equation incorporating the Internet variable. In section 3, we perform several estimations for the growth equation. Section 4 concludes the paper. 2
2. Model Romer s (1986, 1990) endogenous growth model explains that balanced growth is positively influenced by knowledge spillover. We hypothesize that the Internet plays a great role in spreading knowledge in an economy. Therefore, economic growth is positively related with the use of the Internet. From Barro s (1997) growth equation, we choose the ratio of investment to GDP, the ratio of government consumption to GDP, and inflation as explanatory variables along with our Internet variable. Therefore the real per-capita GDP growth rate is determined by the Internet, investment, government consumption, and inflation. We set the following growth equation for estimation, Growth Inflation + u, it = β0+ β1internetit + β2investmentit + β3governmentit + β4 it it (1) where uit = η + ν + ε, η i is an individual (country) effect, and ν t is a time i t it effect, and ε it is independently and identically distributed among countries and years. Growth it is the real per-capita GDP growth rate of country i at year t; Internet is the ratio of the Internet users to total population; Investment is the ratio of gross domestic investment to GDP; Government is the ratio of government expenditure to GDP. The coefficient of Internet is 3
expected to be positive as it contributes to the knowledge spillover. The coefficient of Investment is expected to have a positive sign (Levine and Renelt 1992; Makiw et al. 1992; DeLong and Summers, 1991). The coefficient of Government is expected to be negative as the government distorts the private decisions (Barro, 1997). As high inflation is known to be associated with low economic growth in general, the coefficient of Inflation expected to be negative (Barro 1995; Fernández Valdovinos, 2003). 3. Data and Empirical Results Data for 207 countries from 1991 to 2000 were taken from the World Development Indicators 2002 CD-ROM of World Bank (2002). Internet users, the number of people with access to the worldwide network, is divided by total population to get the internet users ratio. Annual percentage growth rate of GDP per capita (gross domestic product divided by midyear population) is based on a constant local currency. Gross domestic investment consists of outlays on additions to the fixed assets of the economy plus net changes in the level of inventories. General government final consumption expenditures includes all government expenditures for purchases of goods and services. Inflation is measured by the consumer price. 4
Table 1 lists the regression results. We estimated the growth equation (1) by various estimation methods: (a) pooled ordinary least squares (OLS), (b) individual random effects, (c) individual fixed effects, (d) time fixed effects, (e) individual random effects and time fixed effects, and (e) generalized method of moments (GMM) estimation. According to the benchmark pooled OLS regression (column (a) in Table 1), the estimated coefficient of Internet is 5.710 and significant at the 1% level as expected. This means that when the Internet-user ratio increases by 1% point, the growth rate increased by 0.057% point. The estimated coefficient of Investment is 0.167 and significant at the 1% level. This means that when the investment ratio increases, growth rate increases, too. The estimated coefficient of government consumption is insignificant. The estimated coefficient of Inflation is -0.003 and significant at the 10% level. When the inflation rate increases by 1% point, growth rate decreased by 0.003% point. As we used panel data in our regressions, we re-estimated growth equation (1) by panel data regression methods such as individual random effects (column (b) in Table 1), individual fixed effects (c), time fixed effects (d), and individual random effects and time fixed effects (e). The estimated coefficients of Internet range from 4.931 to 5.886 and are significant at the 1% level in (b) and (d) and at the 5% level in (c) and (e). 5
This means that when the Internet user ratio increases by 1% point, the growth rate turned out to increase by between 0.049 and 0.059% point. The estimated coefficients of Investment range from 0.168 to 0.281 and are all significant at the 1% level. This means that when the investment ratio increases, the growth rate increases, too. The estimated coefficient of government consumption is negative and significant at the 1% level in (c), at the 10% level in (b) and (e), and insignificant in (d). The estimated coefficient of Inflation is -0.003 and significant at the 1% level. When the inflation rate increases by 1% point, growth rate turned out to decrease by 0.003 % point. Because explanatory variables such as the Internet, investment ratio, and government consumption ratio, can be influenced by economic growth, we performed GMM estimation to take into account any endogeneity of the explanatory variables (column (f) in Table 1). The coefficient of Internet is 5.517 and significant at the 1% level. The coefficient Investment is 0.085 and significant at the 5% level. The coefficients of government consumption and inflation proved to be insignificant. Hansen s (1982) J- statistic is 10.346 with a p-value of 0.111, suggesting that the model is well specified (Hansen and Singleton 1982). 6
To sum up, the effect of the Internet on economic growth is positive and significant across all the regressions. This means that the result is quite robust against different estimation methods. Insert Table 1. 4. Conclusion The Internet is assumed to contribute to the spillover effect of knowledge across countries. Therefore, the increase in the use of the Internet in a country is hypothesized to have a positive impact on the economic growth. Using panel data with 207 countries from 1991 to 2000, we found evidence that the Internet plays a positive and significant role in economic growth after investment ratio, government consumption ratio, and inflation were used as control variables in the growth equation. 7
References: Barro, R.J., 1995, Inflation and economic growth, NBER Working Paper No. 5326. Barro, R.J., 1997, Determinants of economic growth, The MIT Press. Choi, C., 2003, Does the Internet stimulate inward FDI? Journal of Policy Modeling 25, 319 326. DeLong J.B. and L.H. Summers, 1991, Equipment investment and economic growth, Quarterly Journal of Economics 106, 445 502. Fernández Valdovinos, C.G., 2003, Inflation and economic growth in the long run, Economics Letters 80, 167 173. Freund, C. and D. Weinhold, 2000, On the effect of the Internet on international trade, International Finance Discussion Papers, Board of Governors of the Federal Reserve System. Hansen, L.P., 1982, Large sample properties of generalized method of moments estimators, Econometrica 50, 1029 1054. Hansen, L.P. and K.J. Singleton, 1982, Generalized instrumental variables estimation of nonlinear rational expectations models, Econometrica 50, 1269 1286. Krueger A.B., 1993, How Computers have changed the wage structure: Evidence from micro data 1984 1989, Quarterly Journal of Economics 108, 33 60. 8
Levine, R. and D. Renelt, 1992, A sensitivity analysis of cross-country growth regressions, American Economic Review 82, 942 963. Mankiw, N.G., D. Romer, and D.N. Weil, 1992, A contribution to the empirics of economic growth, Quarterly Journal of Economics 107, 407 437. Newey, W.K., and K.D. West, 1987, A simple, positive semi-definite, heteroscedasticity and autocorrelation consistent covariance matrix, Econometrica 55, 703 708. Oliner, S.D. and D.E. Sichel, 2000, The resurgence of growth in the late 1990s: Is information technology the story? Journal of Economic Perspectives 14, 3 22. Oliner, S.D. and D.E. Sichel, 2003, Information technology and productivity: Where are we now and where are we going? Journal of Policy Modeling 25, 477 503. Romer, P.M., 1986, Increasing returns and long-run growth, Journal of Political Economy 94, 1002 1037. Romer, P.M., 1990, Endogenous technical change, Journal of Political Economy 98 (5) Part 2, S71 S101. Sichel, D.E., 1999, Computers and aggregate economic growth: An update. Business Economics (April), 18 24. World Bank, 2002, World Development Indicators 2002 CD-ROM, The World Bank, Washington, Washington DC. 9
Table 1 The Internet and Economic Growth a (a) b (b) (c) (d) (e) (f) b,c Pooled OLS Individual Random Individual Fixed Time fixed Individual Random & Time fixed Panel GMM Constant -1.173* -1.447* -1.186 0.389 (0.701) (0.742) (0.801) (1.070) Internet 5.710*** 5.641*** 4.931** 5.886*** 5.678** 5.517*** (1.566) (2.018) (2.194) (2.034) (2.324) (1.724) Investment 0.167*** 0.195*** 0.281*** 0.168*** 0.197*** 0.085** (0.023) (0.024) (0.038) (0.019) (0.024) (0.041) Government -0.039-0.061* -0.198*** 0.033-0.054* -0.024 (0.032) (0.031) (0.071) (0.022) (-0.032) (-0.034) Inflation -0.003* -0.003*** -0.003*** -0.003*** -0.003*** 0.001 (0.001) (0.0004) (0.0005) (0.004) (0.0004) (0.002) R 2 0.27 0.43 0.46 0.29 0.45 J-statistic 10.346 [p-value] [0.111] Sample size 1004 1004 1004 1004 1004 565 Notes: a. ***, **, and, * indicate significance at the 1%, 5%, and, 10% levels, respectively. Standard errors are in parentheses. b. Newey and West s (1987) heteroscedasticity and autocorrelation consistent covariance matrix assuming a lag length of one is used for standard errors. c. Instrumental variables include constant, (Growth) t-2, t-3, (Internet Users/Pop) t-2, t-3, (Gross Investment/GDP) t-2, t-3, (Government Expenditure/GDP) t-2, t-3, (Inflation) t-2, t-3 10