The Relationship between Population and Economic Growth in Asian Economies

The Relationship between Population and Economic Growth in Asian EconomiesAuthor(s): Wong Hock Tsen and Fumitaka Furuoka

Source: ASEAN Economic Bulletin, Vol. 22, No. 3 (December 2005), pp. 314-330Published by: ISEAS - Yosuf Ishak Institute

Stable URL: http://www.jstor.org/stable/25773869Accessed: 11-10-2016 19:41 UTC JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusteddigital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information aboutJSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available athttp://about.jstor.org/terms

ISEAS - Yosuf Ishak Institute is collaborating with JSTOR to digitize, preserve and extend access toASEAN Economic BulletinThis content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

ASEAN Economic Bulletin Vol. 22, No. 3 (2005), pp. 314-30 ISSN 0217-4472 The Relationship between Population and Economic Growth in Asian Economies Wong HockTsen and Fumitaka Furuoka The main aim of the study is to investigate the relationship between population and economic growth in Asian economies. Generally, the results of the Johansen (1988) and Gregory and Hansen (1996),cointegration methods show that there is no long-run relationship between population and economic growth. Nonetheless, the study finds that there is bidirectional Granger causality between population and economic growth for Japan, Korea, and Thailand. For China, Singapore, and the Philippines, population is found to Granger cause economic growth and not vice versa. For and Malaysia, economic growth is found to Granger cause population and not vice versa. For Taiwan and Indonesia, there is no evidence of Granger causality between population and economic growth. On the whole, the relationship between population and economic growth is not straightforward. Population growth could be beneficial or detrimental to economic growth and economic growth could have an impact on population growth. I. Introduction The issue of population and economic growth is as product (GDP) per capita were able to increase (Meier 1995, p. 276).1 The debate between force and, therefore, increases economic growth. population increases at a geometrical rate while A large population also provides a large domestic production increases at an arithmetic rate. Thus, market for the economy. Moreover, population the unfettered population growth in a country growth encourages competition, which induces could plunge it into acute poverty. However, the technological advancements and innovations. pessimist view has proven unfounded for Nevertheless, a large population growth is not rate to surpass the production growth rate because old as economics itself. Malthus (1798) claimed positive and negative sides of population growth is ongoing. Population growth enlarges labour that there is a tendency for the population growth developed economies in that they managed to only associated with food problem but also achieve a high level of economic growth and thus, imposes constraints on the development of both population and the real gross domestic savings, foreign exchange and human resources ASEAN Economic Bulletin 314 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

(Meier 1995). Generally, there is no consensus whether population growth is beneficial or detrimental to economic growth in developing economies. Moreover, empirical evidence on the matter for developing economies is relatively limited. The issue of population and economic growth is also closely related to the issue of minimum wage. Population growth enlarges labour force and, therefore, will push wage down. The standard economic labour demand model predicts that low wage will raise the demand for labour. As a result, Therefore, it is important to examine the population and economic growth nexus. The main aim of the study is to investigate the and other natural resources. Conversely, a large population could encourage greater specialization, and a large market increases returns to human capital and knowledge. Thus, the net relationship between greater population and economic growth depends on whether the inducements to human capital and expansion of knowledge are stronger than diminishing returns to natural resources. relationship between population and economic growth in Asian economies, namely, China, Japan, the welfare of the economy is likely to increase. Moreover, low wage would encourage industries that are labour intensive. Low wage is said to be Asian NIEs (Korea, , Taiwan, and Singapore) and the ASEAN-4 countries an important factor that has contributed to the industrialization of Asian newly industrialized economies (NIEs), namely Korea, , Taiwan, and Singapore. Moreover, it is also argued to be an important factor that contributes to Thailand), generally for the period 1950-2000. (Indonesia, Malaysia, the Philippines, and Those countries vary in population size, economic growth, stage of economic development, and openness to international trade (see Table 1). The economic growth in China. Conversely, the that the introduction or rising of minimum wage study employs the Johansen (1988) (J) standard economic labour demand model predicts will break the mechanism, i.e., there would be no cointegration method to examine the long-run relationship between population and economic growth. Moreover, the possibility of a structural break in the long-run relationship between the two link between population and economic growth. Nonetheless, a range of monopsony, efficiency wage, and search models shows that in some circumstances minimum wage could indeed raise employment. The empirical evidence on the matter variables is examined using the Gregory and Hansen (1996) (GH) cointegration method. The is mixed, with some studies showing negative effects and others showing positive or zero effects occurrence of a break. In other words, it determines the break point endogenously from the advantage of the method is that it does not require information regarding the timing of or indeed the of minimum wage. Thus, there is no clear relationship between population and economic growth. Nevertheless, the studies regarding minimum wage and employment are conducted mainly for developed economies (Stewart 2004, p. C110; Rama 2001; Warr 2004). economic growth is complex and the historical evidence is ambiguous, particularly concerning data rather than on the basis of a priori can be avoided. Furthermore, the Granger causality between population and economic growth is addressed. information, which the problem of data mining The relationship between population and population and economic growth in Asian some evidence of the relationship between population and growth in those countries. Moreover, the empirical studies on the relationship The empirical study on the relationship between economies is limited. Thus, the study provides the causes and impacts (Thirlwall 1994, p. 143)?2 Becker, Glaeser, and Murphy (1999, p. 149) demonstrated in a theoretical model that a large population growth could have both negative between population and economic growth in the literature are mainly conducted using cross-section and positive impacts on productivity. A large population may reduce productivity because of diminishing returns to more intensive use of land data (Thornton 2001, p. 4). Nevertheless, some studies are conducted using time series data (Dawson and Tiffin 1998; Thornton 2001). ASEAN Economic Bulletin 315 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

TABLE 1 Asian Economies ? Some Facts 1985 1990 1995 2000 2001 2002 Population (million) 1,058.5 1,143.3 1,211.2 1,267.4 1,276.3 1,284.5 Population annual change (%) 2.1 1.4 1.1 0.8 0.7 0.6 GDP annual change (%) 1.4 7.0 6.6 9.7 6.8 6.3 Agriculture of GDP (%) 28.4 27.0 20.5 16.4 15.2 14.5 China Exports of GDP (%) 8.2 13.3 18.3 23.1 23.0 26.3 Imports of GDP (%) 12.5 10.9 15.7 19.9 20.0 22.8 Services of GDP (%) 28.5 31.4 30.7 33.4 33.6 33.8 Industry of GDP (%) 43.1 41.6 48.8 50.2 51.1 51.7 Population (million) 121.0 123.6 125.6 126.9 127.3 127.4 Population annual change (%) 0.6 0.3 0.2 0.2 0.3 0.1 GDP annual change (%) 3.9 5.1 1.0 2.2 ?1.1 -1.6 Agriculture of GDP (%) 2.4 1.9 1.5 1.1 1.1 1.1 Industry of GDP (%) 46.9 43.2 38.5 37.1 35.4 34.9 Exports of GDP (%) 14.2 10.4 9.1 10.8 10.4 11.2 Imports of GDP (%) 10.8 9.4 7.7 9.4 9.8 9.9 Korea Japan Services of GDP (%). 50.7 .9 60.0 61.8 63.5 .0 Population annual change (%) 1.0 1.0 1.0 0.8 0.7 0.6 GDP annual change (%) 6.5 9.0 8.9 9.3 3.1 6.3 Agriculture of GDP (%) 12.6 8.5 6.2 4.7 4.3 4.0 Industry of GDP (%) 41.0 43.1 43.2 42.4 42.0 40.9 Services of GDP (%) 46.4 48.4 50.6 52.9 53.7 55.1 Exports of GDP (%) 28.5 25.2 25.5 38.1 35.4* 34.1 Imports of GDP (%) 28.5 26.2 26.4 34.5 32.2 31.1 Population (million) 40.8 42.9 45.1 47.0 47.3 47.6 Population (million) 5.5 5.7 6.2 6.7 6.7 6.8 Population annual change (%) 1.0 0.3 2.0 0.9 0.8 1.0 GDP annual change (%) -0.8 3.7 3.9 10.2 0.6 2.3 Agriculture of GDP (%) 0.5 0.2 0.1 0.1 0.1 0.1 Industry of GDP (%) 30.0 25.4 16.2 14.2 13.4 11.3 Exports of GDP (%) 86.2 108.9 122.6 122.6 116.4 122.8 Imports of GDP (%) 84.8 109.3 136.0 127.5 121.5 126.0 Population (million) 19.1 20.2 21.4 22.3 22.4 22.5 Population annual change (%) 1.4 1.1 0.8 0.9 0.6 0.5 GDP annual change (%) 5.0 5.4 6.4 5.9 -2.2 3.5 ASEAN Economic Bulletin 316 Vol. 22, No. 3, December 2005 Services of GDP (%) 69.5 74.4 83.7 85.7 86.5 88.6 TaiwanThis content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

TABLE 1 (continued) Agriculture of GDP (%) 5.8 4.2 3.5 2.1 1.9 1.9 Industry of GDP (%) 46.3 41.2 36.4 32.4 31.1 31.0 _1985 1990 1995 2000 2001 2002 Exports of GDP (%) 49.4 41.9 42.0 47.7 43.4 46.1 Imports of GDP (%) 31.4 32.7 37.0 43.1 36.2 37.4 Population annual change (%) 0.4 4.1 3.2 1.7 2.8 0.8 GDP annual change (%) 14.6 9.0 8.0 9.4 -2.4 2.2 Services of GDP (%) 47.9 .6 60.1 65.5 67.0 67.1 Population (million) 2.7 3.1 3.5 4.0 4.1 4.2 Singapore Agriculture of GDP (%) 1.0 0.4 0.2 0.1 0.1 0.1 Industry of GDP (%) 30.2 33.0 33.3 34.4 32.2 33.3 Exports of GDP (%) 131.1 148.2 141.1 152.9 146.6 147.5 Imports of GDP (%) 139.6 152.6 140.0 139.5 129.2 126.2 Population annual change (%) 0.7 2.0 1.7 1.2 1.2 1.2 GDP annual change (%) 2.5 9.0 8.2 4.9 3.4 3.7 Services of GDP (%) 68.8 66.6 66.5 65.5 67.7 66.6 Indonesia Population (million) 162.7 179.4 194.8 206.3 208.6 211.1 Agriculture of GDP (%) 23.3 19.4 17.1 17.2 17.0 17.4 Industry of GDP (%) 35.8 39.1 41.8 46.1 45.5 44.5 Services of GDP (%) 40.9 41.5 41.1 36.7 37.5 38.1 Exports of GDP (%) 21.2 23.4 23.5 43.5 40.6 33.5 Imports of GDP (%) 14.6 18.8 20.2 26.9 24.5 20.1 Malaysia GDP annual change (%) -1.1 9.0 9.8 8.3 0.4 4.2 Agriculture of GDP (%) 15.4 15.2 12.9 8.7 8.5 9.1 Industry of GDP (%) 44.1 42.2 41.4 46.5 48.1 48.3 Services of GDP (%) 40.5 42.6 45.7 44.8 43.4 42.6 Population annual change (%) 2.7 2.5 2.9 2.4 2.2 2.1 Population (million) 15.7 18.1 20.7 23.5 24.0 24.5 Exports of GDP (%) 48.5 65.0 80.7 109.3 99.9 98.1 Imports of GDP (%) 37.0 59.1 80.7 86.2 79.0 79.1 Philippines Agriculture of GDP (%) 24.5 21.9 21.6 15.8 14.9 14.7 Industry of GDP (%) 35.1 34.5 32.1 32.2 32.5 32.5 Services of GDP (%) 40.4 43.6 46.4 52.0 52.6 52.8 continued on next page GDP annual change (%) -7.3 3.0 4.7 4.4 3.0 4.4 Population (million) .7 62.0 70.3 78.4 80.1 81.8 Population annual change (%) 2.5 3.2 2.4 2.1 2.1 2.1 ASEAN Economic Bulletin 317 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

TABLE 1 (continued) 1985 1990 1995 2000 2001 2002 Exports of GDP (%) 15.1 18.5 23.5 49.1 43.4 44.1 Imports of GDP (%) 16.6 27.5 35.6 44.1 44.4 43.6 Thailand Population annual change (%) 1.9 1.1 1.2 1.0 0.8 0.8 GDP annual change (%) 4.6 11.2 9.2 4.6 1.9 5.2 Population (million) 51.6 55.8 59.4 62.4 62.9 63.4 Agriculture of GDP (%) 15.8 12.5 11.0 10.4 10.4 9.0 Industry of GDP (%) 31.8 37.2 39.3 40.5 40.7 42.5 Exports of GDP (%) 18.2 26.8 33.2 55.4 .8 52.9 Imports of GDP (%) 21.6 34.6 37.7 45.8 47.3 45.2 Services of GDP (%) 52.4 50.3 49.7 49.1 48.9 48.5 Sources: Asian Development Bank, United Nations, and Statistics Bureau, Ministry of Internal Affairs a Communications, Japan. However, these studies do not consider the population in the long-run equilibrium (Becker, possibility of a structural break in their long-run Glaeser, and Murphy 1999, p. 145). analysis, whereas the issue is considered in the Generally, population growth is associated with study. food problem, i.e., malnutrition and hunger. The study is organized as follows. Section II Nonetheless, the food problem is more a problem gives a discussion of population and economic of poverty and inadequate income than a matter of growth. Section m explains data and methodology inadequate global food supplies. The population used in the study. Section IV provides the empirical results and discussions. Section V gives some concluding remarks. and food problem is solved when income is enough to buy adequate food as prices provide adequate incentives to produce. Developing n. Population and Economic Growth The debate on the relationship between population economies are capable of producing surpluses of food for exports. On the other hand, developing economies would have to export more, receive and economic growth could be traced back to 1798 when Thomas Malthus published the book foreign aid or borrow overseas to meet their An Essay on the Principle of Population. According to the Malthusian model, the causation went in both directions. Higher economic growth increased population by stimulating earlier marriages and higher birth rates, and by cutting down mortality from malnutrition and other factors. On the other hand, higher population also increased demand for food by increased imports (Meier 1995, p. 277). Population growth is much more than a food problem. A high rate of population growth not only has an adverse impact on improvement in food supplies, but also intensifies the constraints on development of savings, foreign exchange, and human resources. Rapid population growth tends to depress savings per capita and retards growth of depressed economic growth through diminishing returns. This dynamic interaction between physical capital per worker. The need for social population and economic growth is the centre of the Malthusian model, which implies a stationary infrastructure is also broadened and public need for a larger population rather than in expenditures must be absorbed in providing the ASEAN Economic Bulletin 318 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

providing directly productive assets (Meier 1995, pp. 276-77). increase the cost of having large families compared with investing more in each of children. Population pressure is likely to intensify the foreign exchange constraints by placing more pressure on the balance of payment. The need to import food will require the development of new industries for export expansion and/or import Therefore, the demographic transition towards smaller families in economies with initially high fertility and low income per capita may be stimulated by an initial growth in population. Thus, an increase in population may both reduce fertility and raise the accumulation of human substitution. The rapid increase in school-age population and the expanding number of labour force entrants puts ever-greater pressure on educational and training facilities and retards improvement in the quality of education, which is capital. A larger population may help overcome possibly diminishing returns to this generation's human capital in the production of the next a problem in developing economies as about 33 per cent of the children of primary-school age are generation's human capital because greater not enrolled in school and of those who enter school, 60 per cent will not complete more than three years of primary school (Meier 1995, p. 285). Also, too denser a population aggravates population growth induces more specialization and a larger market that raise returns to human capital and knowledge. If human capital per capita were sufficiently large, the economy would move to steady-state growth, whereby in the steady-state the problem of improving the health of the population. In most developing economies, the working age population had roughly doubled in the past twenty-five years. At expected growth rates, it will double again in the next twenty-five growth path, consumption per capita would increase at a slower rate than human capital if the population is growing and if the production of consumer goods has diminishing returns to still be increasing, despite these diminishing population. However, consumption per capita can years. This growth clearly intensifies pressure on employment and the amount of investment returns, if the positive impact of the growth in human capital on productivity in the consumption p. 277). available per labour market entrant (Meier 1995, demonstrated in a theoretical model that Becker, Glaeser, and Murphy (1999, p. 147) sector more than offsets the negative impact of population growth. Thus, zero population growth is not necessary for sustainable growth in per capita consumption, even with diminishing returns to population in the production of consumer goods population growth will increase parental utility if it has a sufficiently positive impact on human capital accumulation or if the impact on current (Becker, Glaeser, and Murphy 1999, p. 148). production is not too negative. Since human capital is more important at higher levels of On the other hand, an increase in population growth may lower the productivity of farming in development, greater population is likely to raise per capita welfare in more developed economies. relationship between population and economic growth. A majority of them uses cross-section regression to analyse the relationship between the There are few empirical studies on the two variables (Easterlin 1967; Thirlwall 1972; Simon 1992; Kelley and Schmidt 1996; Ahlburg poorer agricultural economies, so that output per capita there would be lower initially. However, 1996). Some of them found no statistically even in these economies, greater population significant relationship between population and economic growth while other studies were not able human capital by raising rates of return on investments in schooling and other human capital. growth would tend to raise the accumulation of to come to a conclusive opinion as the results tended to be contradictory. Dawson and Tiffin (1998) used annual time Moreover, families would lower their fertilities if series data over the period 1950-93 to analyse the population growth raises rates of return on investments in children because that would long-run relationship between population and economic growth in India. The study employed ASEAN Economic Bulletin 319 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

cointegration and Granger causality methods and the data for Taiwan are over the period 1951-98; reported that there is no long-run relationship the data for and Indonesia are over the between the two variables. Moreover, population period 1960-2000, respectively; the data for growth neither Granger causes economic growth Singapore are 1960-96, and the data for Malaysia nor is caused by it. Thornton (2001) conducted a are 1955-2000. All the data were transformed into similar research on the long-run relationship logarithms. The plots of logarithms of population between population and economic growth in seven and the real GDP per capita are given in Figure 1. Latin American countries, namely, Argentina, Generally, the figure shows no relationship Brazil, Chile, Colombia, Mexico, Peru, and between population and the real GDP per capita in Venezuela. The study used annual time series data all economies examined. The trends of population generally over the period 1900-94 and employed tend to be stationary while the trends of the real the same methods of analysis as Dawson and GDP per capita show some fluctuations. Tiffin (1998). The study concluded that there is no The empirical estimation in the study begins with the unit root tests. The aim of unit root tests long-run relationship between the two variables in any of the seven countries. Furthermore, population growth neither Granger causes eco nomic growth nor is caused by it. is to examine whether a series is stationary or non stationary, which is important to avoid spurious regression. In the study, the Dickey and Fuller (1979) (DF) and Phillips and Perron (1988) (PP) III. Data and Methodology The population and economic growth data are based on annual data. In the study, the economic growth is expressed by the real GDP per capita. The population and economic growth data were obtained from Heston, Summers, and Aten (2002). unit root test statistics are employed. The DF unit root test statistic uses the parametric approach, i.e., to change the estimation regression to solve term. In contrast with the DF unit root test statistic, the PP unit root test statistic uses the non the heterogeneity and serial correlation in an error parametric approach, i.e., to modify statistic to obtain estimator and statistic. The DF unit root test The data for Japan, the Philippines, and Thailand are over the period 1950-2000, respectively; the statistic is a low power test under the null hypothesis of a unit root that posits root close to the unit circle or tends to stationary. On the other data for China are over the period 1952-2000; the data for Korea are over the period 1953-2000; FIGURE 1 The Plots of Logarithms of Population and Real GDP per capita against Time ASEAN Economic Bulletin 320 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

Korea FIGURE 1 (continued) 1953 1963 19J3 1983 1993 1960 1965 1970 1975 1980 1985 1990 1995 Years Years Taiwan Singapore 1951 19561961 19661971 19761981 19861991 1996 1998 1960 1965 1970 1975 1980 1985 1990 1995 Years Years Indonesia Malaysia 1960 1965 1970 1975 1980 1985 1990 1995 1955 1960 1965 1970 1975 1980 1985 1990 1995 Years The Philippines 11 Thailand ?: Notes: The light colour (P) line indicates logarithm of population. The dark colour (Y) line indicates logarithm of the real GDP per capita. ASEAN Economic Bulletin 321 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

hand, the PP unit root test statistic is known to be more robust in an error term process, i.e., an error term is allowed to be weakly heterogeneous. to consider the rank of n, that is to seek the number of linearly independent columns in n. According to Engle and Granger (1987), series that are integrated of the same order may cointegrate together. The cointegrated series may drift apart from each other in the short run but the distance between them tends to be constant or in a The J cointegration method can be use to compute two likelihood ratio tests for testing the number of cointegrating vectors in the system, namely the maximum eigenvalue (A.Max) and trace (Kr^) statistics, which are respectively computed as: a vector of series (n x 1), y, is said to be cointegrated if each of the series is integrated of the same order, an existing non-zero cointegrating stationary process in the long run. More formally, W=-rin(l-^,) (2) vector (n x 1), a' such that the linear combination of these series, a'y, are stationary or is said to be where T is the sample size; In is the integrated of zero and denoted by 1(0). In the study, the J cointegration method is used xTtac. = -rzri^lin(i- Xi is the eigenvalue. The A^ax test stat null hypothesis (H0) of r cointegrat to test the number of cointegrating vectors. against the alternative hypothesis (H Suppose all variables are to be integrated of order are (r + 1) cointegrating vectors. Th one, 1(1). A vector autoregressive (VAR) process k statistic tests the H0 that has could be re-parameterized in the error correction cointegrating vectors in the system form as: number of cointegrating vectors is equal to r (Johansen 1988). The like test statistics can be sensitive to the c az, = n + rx azf_i + ... + Tw az,_*+l + n z,_k + ut (i) lag length in equation (1). Thus, the where A is the first difference operator; z, is a (p x 1) vector of variables; |i is a (p x 1) vector of constants; II is a (pxp) matrix of coefficients; ut is lag length in equation (1) is dete Schwarz Bayesian criterion (SBC). Moreover, the GH cointegration which allows for the possibility of r Gaussian; T, = -(I - n, - ... - n^); n = -(I - Yl{ - ... - n*) and I is an identity matrix. Equation (1) combines information on both the short-run and a (p x 1) vector of error terms, assumed to be is employed to examine the co relationship between population a growth. They consider three different long-run adjustment to changes in z, via the estimates of T, and II, respectively. When the rank made concerning the nature of the cointegrating regression, namely the level shift (C), level shift and a time of n is r, where 1 < r < (p - 1) or it has reduced independent. When the rank of II is zero, the meaning is that all the elements in the matrix are zero. Therefore, there is no long-run relationship rank, meaning its rows are not linearly and regime shift (C/S), respectively as In y, = Hu + H12 All use subject to http://about.jstor.org/terms

(p,tt = 0 if t < [Tx] and 1 if t > [Tx]; x e (0,1) is the Aln Yt Pio + ^w PiMAlnP0Pl4 + J?M pl2l AlnyM + vu unknown parameter which denotes the timing of the change point and [ ] denotes the integer part; (7) t is a time trend and eu (i = 1, 2, 3) is an error term, respectively. In the model with level shift (Q, \\in represents the intercept before the shift, Aln POP, Pao + ^i-i p2i, AlnPOPM + I^P^AlnrM + v^ (8) and flu represents the change in the intercept at where vu (i = 1, 2) is a zero-mean and serially the time of the shift. This implies that the uncorrelated error term. For example, if the joint test, i.e., the F-statistic of lagged variables of fashion. In the model with level shift and a time Aln POP, in equation (7) is significandy different trend (C/T), in addition to the feature in the model from zero, it implies that population Granger with level shift, a time trend is included. In the causes economic growth. If the joint test of lagged cointegrating relationship has shifted in a parallel model with regime shift (C/S), in addition to the feature in the model with level shift, a7^ denotes the change in the slope coefficient is considered. variables of Aln Yt is significandy different from zero in equation (8), then it implies that economic growth Granger causes population. The minimum Thus, this model allows a level and the slope final prediction error (FPE) criterion proposed by vector in the cointegrating relationship to shift. Akaike (1971) is used to determine the optimal The procedure is similar to the Engle and lags of the model. When series are cointegrated, Granger (1987) cointegration method that includes the simple Granger causality test becomes a dummy variable in the cointegrating regression to consider a shift in the long-run relationship. The estimation of the'above models using the ordinary inappropriate and the testing of Granger causality goes in the error correction model, i.e., an error correction term, which is obtained from the least squares estimator yields the estimated error cointegrating regression to be included in the terms, following which the unit root tests (ADF*? estimation in equations (7) and (8), respectively. Z*t) are applied to them. The unit root tests, ADF*, The testing procedure is the same as mentioned above. and Z*t are designed to test the null hypothesis of no cointegration against the alternative hypothesis of cointegration in the presence of a possible regime shift. If there is one unknown point in the sample, the standard tests for cointegration are not appropriate, since they presume that the cointegrating vector is time-invariant under the alternative hypothesis. The advantage of the GH cointegration method is that it does not require IV. Empirical Results and Discussions The results of the DF and PP unit root test statistics are reported in Table 2. The lag length used to compute the DF test statistic is based on Akaike (1973) information criterion (AIC). For the PP unit root test statistic, the results that are occurrence of a break. In other words, it information regarding the timing of or indeed the data rather than on the basis of a priori determines the break point endogenously from the are used to compute the test statistic after considering truncation lags one to three in computing the test statistic. The results of the DF reported are based on three truncation lags, which information, which the problem of data mining and PP unit root test statistics show that can be avoided. population is integrated of order one for Hong In the Granger (1969) sense of a variable X Kong, the Philippines, and Thailand. For China causes another variable Y if the current value of Y can better be predicted by using past values of X. and Taiwan, the unit root test statistics show that it is integrated of order two. For Japan, Korea, and Singapore, the DF unit root test statistic shows that it is integrated of order two while the PP unit root On the other hand, Y is said Granger causes X if the past values of Y can be used to improve the prediction of X. When series are not cointegrated, the simple Granger causality test is used: test statistic shows that it is integrated of order one. For Indonesia and Malaysia, the DF unit root ASEAN Economic Bulletin 323 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

TABLE 2 The Results of the Dickey and Fuller (1979) and Phillips and Perron (1988) Unit Root Test Statistics ty - trend In POP, -3.6993* (1) 0.8816 (3) A In POP, -2.7944 (4) -2.7936 (3) Indonesia Z(ty) - trend lnY, -2.5632 (2) -2.4985 (3) AlnY, -5.3667**(0) -4.5823**(3) Malaysia A2 In POP, -4.1781**(0) -4.10**(3) China In POP, A In POP, A2 In POP, lnK, A In Y, -.3863 (4) -0.4401 (3) In POP, -5.6534**(1) -2.2161 (3) -2.9668 (3) -2.8339 (3) A In POP, -2.2737 (1) -2.0103 (3) -4.2433**(1) -4.9677**(2) -.9582 (3) -0.5696 (3) -6.32**(3) -4.5944**(3) A2 In POP, -5.5742**(0) -5.3185**(3) lnY, -3.1462 (1) -3.02 (3) Aln7, -4.0735* (0) -4.5038**(3) Philippines In POP, A In POP, A2 In POP, Japan -1.3385 (4) 0.9588 (3) In POP, .5856 (0) 0.4609 (3) -1.9475 (4) -4.7490**(3) A In POP, -5.4782**(0) -5.7771**(3) -1.0358 '(1) -3.9416* (0) -3.9593* (3) AlnY, -4.7335**(0) -5.2682**(3) -3.78* (3) -14.2995 **(3) Alny, Korea In POP, lnr, A2 In POP, -5.3839**(2) -13.4502**(3) -0.17 (3) lnY, -1.8205 (1) -2.1852 (3) Thailand A In POP, A2 In POP, Alny, In POP, A In POP, A2 In POP, lny, lny, -5.4112**(0) -3:3939 (2) -0.2519 (3) In POP, 1.5952 (1) 3.1459 (3) -1. (1) -5.5285 **(3) A In POP, -3.9717* (0) -3.5372* (3) -4.1651* (1) -6.7566**(3) A2 In POP, -5.3733**(2) -11.9376**(3) -3.7575* (1) -2.5844 (3) In Y, -4.2863**(3) -4.2849**(3) -5.9335**(3) A In Y, -4.6561 **(3) -5.41 **(3) t-statistic. Z(fy) is the PP t-statistic. Values in parentheses are the lag length used in the estimation of ^.9878 **(0) -4.8242**(3) -1.4991 (0) -2.9850 (3) Notes: A2 is the second difference operator, ty is the Dickey-Fuller (DF) or Augmented Dickey-Fuller (ADF) -6.1947**(1) -11.1149**(3) Alny, -6.1614**(1) -5.4879**(3) -.2519 (2) -0.5049 (3) the unit root test statistics. Critical values for ty (Z(rT)) with a drift and a time trend (trend) at 1 per cent and 5 per cent for sample size 45 are -4.18 and -3.51, respectively (MacKinnon 1996). ** denotes significance at 1 per cent level. * denotes significance at 5 per cent Taiwan In POP, A In POP, A2 In POP, -1.3868 (1) -1.8783 (3) -2.00 (0) -2.2366 (3) -3.3557 (1) -2.1603 (3) -4.6196**(0) ^.**(3) -6.9120**(0) -5.6085**(3) level. Alny Singapore In POP, lny, A In POP, A2 In POP, Alny, lny, -11.7902**(0) -.91469 (3) -2.6731 (3) -6.6314**(3) -6.2605**(0) -1.9114 (1) -1.2686 (3) -3.3793 (0) -3.5283* (3) -o\\7570**(3) while the PP unit root test statistic shows that it is test statistic shows that it is integrated of zero, integrated of order two. On the other hand, the results of the DF and PP unit root test statistics of order one, except Korea and Thailand. For Korea, the DF test statistic shows that it is generally show that economic growth is integrated ASEAN Economic Bulletin 324 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

For Thailand, the DF and PP unit root test statistics show that it is integrated of order zero. integrated of order zero, while the PP unit root test statistic shows that it is integrated of order one. istics show that population and economic growth are not cointegrated. Lastly, for the Philippines, the A,M? test statistic shows that the null On the whole, population and economic growth are treated to be integrated of order one. hypotheses are not rejected at 95 per cent critical value. In contrast with the A^ax test statistic, the A/rrace test statistic shows that there is one The results of the J cointegration method are cointegrating vector. Johansen and Juselius (1990) reported in Table 3. The results of the X.Max and suggested that the XMax test might be better than Aqvacc test statistics are computed with unrestricted the A/rrace test. Thus, it is concluded that there is no intercepts and no trends. For Japan and , the results of the XMax and ^rmce test statistics cointegration for population and economic growth show that the null hypotheses, i.e., H0: r = 0 is rejected at 95 per cent critical value while H0: r <= 1 is not rejected at 95 per cent critical value, which indicate that population and economic growth are cointegrated. For China, Korea, Taiwan, Singapore, Indonesia, Malaysia, and in the Philippines. On the whole, the study finds no long-run relationship between population and economic growth. The finding of no cointegration between population and economic growth could be because of the existence of a structural break that biases the test results in favour of not rejecting the null hypothesis of no cointegration. Therefore, Thailand, the results of the and test stat the GH cointegration method is employed. TABLE 3 The Results of the Johansen (1988) Likelihood Ratio Test Statistics Xmox Test Statistic Xmce Test Statistic H0: r = 0 r <= 1 r = 0 r <= 1 Ha: r= 1 r = 2 r > 1 r > 2 China Korea Japan 23.0334* 29.2441* 15.0365* 7.4532 Taiwan Malaysia Thailand Singapore Indonesia Philippines 17.2226* 12.8616 13.2238 9.8792 5.9426 3.1023 9.6877* 1.8037 8.2509* .0824 8.9511* 3.9969 1.3855 27.0303* 38.1952* 18.1387* 19.5669* 25.4735* 12.9440 19.4141* 16.0715 8.8387 7.7463 3.9969 8.9511* 3.1023 9.6877* 1.8037 8.2509* 1.3855 9.1933 6.1903 6.8783 6.1903 6.8783 .0824 c.v. 14.8800 8.0700 17.8600 8.0700 Notes: For China and Thailand, the VAR = 3 is used in the estimation. F Japan, Korea, Taiwan, Indonesia, Malaysia and the Philippines, the VAR = 2 used in the estimation. For and Singapore, the VAR = 1 is used the estimation, c.v. denotes 95 per cent critical value. * denotes significance 95 per cent critical value. ASEAN Economic Bulletin 325 Vol. 22, No. 3, December 20This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

The results of the GH cointegration method are ADF*, and Z*f test statistics show that the null reported in Table 4. Generally, the results of hypothesis of no cointegration against the TABLE 4 The Results of the Gregory and Hansen (1996) Cointegration Test Statistics c err c/s ADF*, -2.81(0.71) -4.68(0.14) -4.42(0.53) ADF*, -3.23(0.27) -3.31(0.49) -3.52(0.27) ADF*, -3.29(0.17) -4.45(0.48) -5.16*(0.46) China Japan Z* -2.45(0.76) -4.48(0.16) -3.32(0.59) Z* -3.10(0.25) -3.35(0.49) -3.39(0.49) Korea Z*, -3.33(0.17) -3.52(0.15) -2.41(0.38) ADF*, -3.06(0.17) -3.(0.85) -3.44(0.76) Z*, ' -3.11(0.85) -3.15(0.85) -3.63(0.76) Taiwan ADF*, -3.06(0.17) -4.17(0.44) -3.95(0.) ADF*, -4.34(0.32) -5.44*(0.68) -4.93(0.43) Singapore Z* -3.12(0.17) -3.43(0.15) -2.30(0.56) Z*, -3.(0.30) -4.62(0.30) -3.75(0.32) Z* -3.38(0.15) -3.68(0.15) -3.42(0.85) ADF*, -3.94(0.15) -4.18(0.15) -3.98(0.15) Malaysia Indonesia ADF*, -3.84(0.46) -4.57(0.61) -4.13(0.35) Philippines Z* -3.32(0.43) -3.59(0.63) -3.40(0.43) ADF*, -4.48(0.73) -4.33(0.73) -4.28(0.73) Thailand Z* -4.03(0.69) -3.96(0.69) -3.81(0.69) ADF* -4.55(0.78) -4.71(0.78) -4.56(0.78) Z*, -4.59(0.78) -4.78(0.76) -4.95(0.74) Notes: The values in brackets show the breakpoint, i.e., the point in the sample where the smallest value of the test statistic is obtained. Critical values for C, C/T and CIS at 1 per cent (5 per cent) with one regressor are ?5.13(?4.61), -5.45(-4.99) and -5.47(-4.95), respectively (Gregory and Hansen 1996). * denotes significance at 5 per cent level. ASEAN Economic Bulletin 326 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

alternative hypothesis of cointegration in the presence of a possible regime shift are not rejected between population and economic growth in India, at 5 per cent level, except Korea and Singapore. For Korea, ADF*, test statistic that tests the null hypothesis of the model C/S is rejected at 5 per and Thornton (2001), which reported that the long-run relationship between population and economic growth in Latin American countries, namely Argentina, Brazil, Chile, Colombia, cent level, thus implying there is a long-run Mexico, Peru, and Venezuela, does not exist. In relationship between population and economic the long run, there is no relationship between population and economic growth regardless of the size of population, openness of international trade, growth. On the other hand, Z*, test statistic that tests the null hypothesis of the model C/S is not rejected at 5 per cent level. Therefore, the results are inconclusive. For Singapore, ADF*f test statistic that tests the null hypothesis of the model trading partner, state of economic development, and minimum wage. C/T is rejected at 5 per cent level, implying there reported in Table 5. Generally, there is some and Indonesia. For Taiwan and Indonesia, the The results of Granger causality test are is a long-run relationship between population and evidence that population and economic growth are economic growth. On the other hand, Z*f test Granger causality to each other, except for Taiwan statistic that tests the null hypothesis of the model C/T is not rejected at 5 per cent level. Thus, the findings are consistent with those of Dawson and results are inconclusive. ' Tiffin (1998) and Thornton (2001), which found Generally, the results of the GH cointegration population growth neither Granger causes method show the same conclusion as the J economic growth nor is caused by it. In other cointegration method,-i.e., there is no evidence of words, population growth neither stimulates economic growth nor detracts from it. On the other hand, for Japan, Korea and Thailand, there is findings of Dawson and Tiffin (1998), which bidirectional causality between population and reported that there is no long-run relationship economic growth, which contradicts with the a long-run relationship between population and economic growth. The finding is the same as the TABLE 5 The Results of Granger Causality Test A In POPt->Aln Y, A In Yt -? A In POPt China 11.72** 3.48 .07 11.84** Taiwan 1.33 4.87 Indonesia 2.94 3.65 Korea 8.47** 10.32** Japan 3.95* 41.46** Singapore 10.18** 0.22 Malaysia 1.28 11.75** Thailand 7.99* 4.40* Philippines 8.46** 3.34 Notes: The arrow denotes no Granger causality. ** denotes significance at 1 per cent level. * denotes significance at 5 per cent level. ASEAN Economic Bulletin 327 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

results of Dawson and Tiffin (1998) and Thornton found to have a positive impact on economic (2001). For China, Singapore, and the Philippines, growth. population growth is found to Granger cause economic growth. For and Malaysia, economic growth is found to Granger cause V. Concluding Remarks The main aim of the study is to investigate the relationship between population and economic growth in Asian economies, namely China, Japan, population and economic growth is not straightforward. There is no strong evidence that a population growth. Thus, the relationship between large population will contribute to economic Asian NIEs, and ASEAN-4. The results of the DF growth. Moreover, the relationship between and PP unit root test statistics show that, generally, population and economic growth is not the same among the countries that have about the same state population and economic growth are non stationary in level but become stationary after taking the first differences. In other words, those of economic development. The size of an economy and openness of international trade do not matter. series are considered to be integrated of order The implementation of minimum wage has no one. Moreover, the results of the J cointegration strong impact on the population and economic method show that generally population and growth relationship. There is no straightforward relationship between economic growth are not cointegrated. The inability of rejecting the hull hypothesis of population and economic growth. Population no cointegration between population and growth could be beneficial or detrimental to economic growth in most of the cases examined economic growth, and economic growth could have economies in Asia, which achieve a low level of an impact on population growth. Thus, some could be because of the existence of a structural break that biases the test results in favour of not rejecting the null hypothesis of no cointegration. population growth, but are affected by other factors method, which can accommodate the existence of economic growth, may not be affected by Thus, the study employed the GH cointegration such as political instability and lack of investments. a structural break in the cointegrating vector. On the other hand, some economies in Asia, which Nonetheless, the results of the GH cointegration method show the same conclusion as the results of achieve a high level of economic growth, may not have done so because of population growth, but due the J cointegration method. Thus, the results of the to other factors. Tan (1995) claimed that political stability, efficiency of public administration, cointegration methods employed in the study reaffirm each other. There is no straightforward successful implementation of export-oriented relationship between population and economic industrialization policies, quality of labour force, growth. and macroeconomic stability are among the factors that have contributed to economic growth in Asian fast growth of the East Asian economies were partly due to their early openness to international trade, NIEs. Lloyd and MacLaren (2000) argued that the growth. The results are mixed. For Japan, Korea causality between population and economic Furthermore, the study estimates the Granger and less openness of their economies to international trade will slow down their economic and Thailand, there is bidirectional Granger causality between population and economic growth. For China, Singapore and the Philippines, population is found to Granger cause economic growth and not vice versa. On the other hand, for growth rates. Wong (2003) examined foreign direct investment (FDI) and economic growth in the and Malaysia, economic growth is ASEAN-4 countries and China and reported that found to Granger cause population and not vice FDI has contributed to economic growth in these versa. For Taiwan and Indonesia, there is no countries. In addition, human capital, domestic evidence of Granger causality between population investment, and openness to international trade are and economic growth. ASEAN Economic Bulletin 328 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

The relationship between population and detrimental to economic growth and economic economic growth is not straightforward. growth could have an impact on population Population growth could be beneficial or growth. NOTES The authors would like to thank the referees and co-editors of the bulletin for their comments on the early versions of the article. All remaining errors are ours. 1. Economic growth in the literature of population and economic growth is measured mainly using the real GDP per 2. Thirlwall (1994) discussed the issue of the relationship between population and economic growth mainly for developing economies. capita. REFERENCES Ahlburg, D. A. \"Population ^nd Poverty\". In The Impact of Population Growth on Well-being in Developing Countries, edited by Dennis A. Ahlburg, Allen C. Kelley, and Karen Oppenheim Mason, pp. 219-58. Berlin: Springer, 1996. Akaike, H. \"Statistical Predictor Identification\". Annals of the Institute of Statistical Mathematics 22, no. 2 (1971): 203-17. -. \"Maximum Likelihood Identification of Gaussian Autoregressive Moving Average Models\". Biometrika 60, Becker, G. S., E. L. Glaeser, and K. M. Murphy. \"Population and Economic Growth\". American Economic Review Papers and Proceedings , no. 2 (1999): 145-49. Case of India\". Journal of Development Studies 34, no. 5 (1998): 149-56. Dickey, D. A. and W. A. Fuller. \"Distribution of the Estimators for Autoregressive Time Series with a Unit Root\". Journal of the American Statistical Association 74, no. 366 (1979): 427-31. Easterlin, R. A. \"Effect of Population Growth in the Economic Development of Developing Countries\". Annals of the American Academy of Political and Social Sciences 369 (1967): 98-108. Engle, R. F. and C. W. J. Granger. \"Co-integration and Error Correction: Representation, Estimation, and Testing\". Econometrica 55, no. 2 (1987): 251-76. Granger, C. W. J. \"Investigating Causal Relations by Econometric Models and Cross Spectral Models\". Econometrica Dawson, P. J. and R. Tiffin. \"Is there a Long-run Relationship between Population Growth and Living Standards? The no. 2 (1973): 255-65. 37, no. 3 (1969): 424-38. Gregory, A. W. and B. E. Hansen. \"Residual-based Tests for Cointegration in Models with Regime Shifts\". Journal of Econometrics 70 (1996): 99-126. Heston, A., R. Summers, and B. Aten. \"Penn World Table Version 6.1\". Center for International Comparisons at the University of Pennsylvania (CICUP), October 2002. Johansen, S. \"Statistical Analysis of Cointegration Vectors\". Journal of Economic Dynamics and Control 12, no. 2-3 Johansen, S. and K. Juselius. \"Maximum Likelihood Estimation and Inference on Cointegration ? with Applications to the Demand for Money\". Oxford Bulletin of Economics and Statistics 52, no. 2 (1990): 169-210. Kelley, A. C. and R. M. Schmidt. \"Toward a Cure for the Myopia and Tunnel Vision of the Population Debate: A Dose of Historical Perspective\". In The Impact of Population Growth on Well-being in Developing Countries, edited by Dennis A. Ahlburg, Allen C. Kelley, and Karen Oppenheim Mason, pp. 11-35. Berlin: Springer, 1996. (1988): 231-. Lloyd, P. J. and D. MacLaren. \"Openness and Growth in East Asia after the Asian Crisis\". Journal of Asian MacKinnon, J. G. \"Numerical Distribution Functions for Unit Root and Cointegration Tests\". Journal of Applied Econometrics 11, no. 6 (1996): 601-18. Malthus, T. An Essay of the Principle of Population and a Summery View of the Principles of Population. London: Pickering, 1798. Economics 11 (2000): -105. ASEAN Economic Bulletin 329 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms

Meier, G. M. Leading Issues in Economic Development 6th ed. New York: Oxford University Press, 1995. Phillips, P. C. B. and P. Perron. 'Testing for a Unit Root in Time Series Regression\". Biometrika 75, no. 2 (1988): 335-46. Rama, M. \"The Consequences of Doubling the Minimum Wage: The Case of Indonesia\". Industrial and Labour Relations Review , no. 4 (2001): 8-81. Simon, J. L. Population and Developing Countries. Princeton, New Jersey: Princeton University Press, 1992. Stewart, M. B. \"The Employment Effects of the National Minimum Wage\". Economic Journal 114 (2004): Cl 10-16. Tan, G. The Newly Industrializing Countries of Asia. 2nd ed. Singapore: Times Academic Press, 1995. Thiriwall, A. P. \"A Cross-section Study of Population Growth and the Growth of Output and Per Capita Income in a Production Function Framework\". The Manchester School 40 (1972): 339-59. -. Growth and Development. Basingstoke: Macmillan Press, 1994. Thornton, J. \"Population Growth and Economic Growth: Long-run Evidence from Latin America\". Southern Economic Journal 68, no. 2 (2001): 4-68. Warr, P. \"Globalization, Growth, and Poverty Reduction in Thailand\". ASEAN Economic Bulletin 21, no. 1 (2004): 1-18. Wong, H. T. \"Foreign Direct Investment and Economic Growth: An Empirical Study on ASEAN-4 and China\". Labuan Bulletin of International Business and Finance 1, no. 1 (2003): 65-78. Wong Hock Tsen is a lecturer at the School of Business and Economics, Universiti Malaysia Sabah. Fumitaka Furuoka is a lecturer at the School of Business and Economics, Universiti Malaysia Sabah. ASEAN Economic Bulletin 330 Vol. 22, No. 3, December 2005This content downloaded from 216.57.242.172 on Tue, 11 Oct 2016 19:41:35 UTC

All use subject to http://about.jstor.org/terms