Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
SAVING TRANSITIONS* Dani Rodrik Harvard University John F. Kennedy School of Government 79 Kennedy Street Cambridge, MA 02138 (617) 495-9454 Fax: (617) 496-5747 E-mail: [email protected] Draft July 1998 * Paper prepared as part of the World Bank research project on saving. I am grateful to Luis Serven for detailed suggestions, Chad Steinberg for excellent research assistance, and Joanna Veltri for editorial and research help. I. Introduction Accumulation of physical capital is the proximate source of economic growth. While the short-run relationship between investment and growth tends to be weak at best, in the long-run the investment rate turns out to be he most robust correlate of growth.1 As a matter of accounting necessity, investment has to be financed by saving, from either domestic or foreign sources. In practice, there have been few cases of high-investment countries, perhaps none at all, where foreign saving has accounted for more than 20 percent of investment over long stretches of time. In an economy investing, say, 30 percent of its GDP, relying on foreign saving beyond this limit would imply running a persistent current account deficit in excess of 6 percent of GDP, which would be courting with disaster. Hence the critical importance of domestic saving in economic growth follows immediately from a few straightforward facts of economic life. Indeed, differences in saving rates are a clear distinguishing factor between thriving economies and stagnant ones. Over the 1984-94 period, there were 31 countries in which percapita GDP grew at an annual average rate of 2.5 percent or higher. In this set of successful countries, the median saving rate was 24 percent.2 By contrast, the median saving rate in the 59 countries in which per-capita income grew at less than 1 percent per year stood at 16 percent. Assuming all domestic saving translates into domestic investment, and that the long-run incremental capital-output ratio is around 5, virtually all of the gap in growth between these two groups of countries can be “accounted” for by the difference in their saving performance. 1 2 See Easterly (1997) on the short-run, and Levine and Renelt (1992) on the long-run. Unless otherwise mentioned, all saving rates in this paper refer to the ratio of gross national saving to gross national disposable income, as defined in the World Bank saving database. 2 But of course such calculations tell us precious little about the underlying economics of high growth and the policies that enable it. High-growth countries share many other characteristics besides high saving and investment: they tend to have lower inflation rates, smaller budget deficits, better human resources, lower current account deficits, and higher shares of trade in GDP (see Table 1, taken from IMF 1997). Which if these factors drive growth, if any? Do we need to have them all together, or are some simply the consequence of growth? And even if we were to accept the causal role played by investment, we can ask whether increases in saving are sufficient as well as necessary for investment and growth. How likely are saving transitions to result in higher growth? Finally, to the extent that saving is responsible for investment and growth, what policies and institutional arrangements generate increased saving? The empirical literature that deals with saving has three strands. One line of research has focused on the cross-national determinants of saving, applying econometric techniques to large cross-national or panel data sets (Giovannini 1985, Edwards 1996, Harrigan 1996, Loayza, Schmidt-Hebbel, and Serven 1998). This research has emphasized quantifiables, in particular the roles of demographic conditions, fiscal policy, financial depth, and economic growth itself. The initial focus in this literature was on the role of deposit interest rates in mobilizing saving. Partly because of negative findings, attention has turned to a broader set of structural and institutional determinants. A second strand has focused on the question of causality between saving and growth (Carroll and Weil 1993, Attanasio, Picci, and Scorcu 1997). There are strong hints in this research that it is growth that drives saving rather than vice versa, especially over short horizons. This result has led some analysts to suggest that saving should not receive high priority in 3 designing growth strategies: once the obstacles to growth are removed, the response of saving could be nearly automatic (see for example Gavin, Hausmann, and Talvi 1996). Finally, there are a number of analytical case studies that focus on high-saving countries or countries that have undergone transitions to high saving, such as Japan, Korea, Taiwan, and Chile, to uncover the determinants of saving and growth transitions in specific settings (Marfan and Bosworth 1994, Hayashi 1986, Rodrik 1995). This strand reinforces some of the findings of the cross-national regressions—the importance of demography, for example—as well as pointing out idiosyncratic conditions—such as investment subsidies (Korea and Taiwan) or pension-system reforms (Chile). The present paper relates to all three of these strands. My central pre-occupation is with countries that have undergone sustained saving transitions, to be defined more precisely below. The objective is to understand the causes and consequences of saving transitions. So the paper has a natural link to the case study literature. However, I take a systematic cross-national approach to identifying saving transitions, and this throws up a lot of cases that have scarcely received attention in the past. This allows me to check the importance of the standard determinants of saving for the sub-sample of transition countries. It also allows me to re-examine the question of causality between growth and saving, using a different approach and over varying time horizons. The outline of the rest of the paper is as follows. In section II, I discuss my definition of a “saving transition” and list the 20 cases that qualify. Section III provides an overview of the contours of a “typical” saving transition. Section IV focuses on the inter-relationships among saving, growth, and investment during saving transitions. Section V analyzes saving behavior in countries that undergo growth transitions. Short-run leads and lags among saving, growth, and 4 investment in our sample of countries are examined in section VI. Section VII provides econometric evidence on the determinants of saving in our sample. Section VIII discusses at greater length the experiences of the following countries: South Korea, Taiwan, Singapore, Mauritius, and Chile. Finally, section IX provides some concluding remarks. II. Defining a saving transition The central problem in the theory of economic development, wrote W. Arthur Lewis (1954, 155), “is to understand the process by which a community which was previously saving and investing 4 or 5 percent of its national income converts itself into an economy where voluntary saving is running about 12 to 15 percent of the national income or more.”3 Hence, Lewis viewed saving transitions as key to economic development. Countries with the most successful growth records in the post-war period have indeed gone through spectacular saving transitions. Figure 1 shows the examples of South Korea and Botswana. In Korea, the saving rate stood barely above 10 percent in the early 1960s. By the mid-1970s, it had risen to above 20 percent, and by the late 1980s it was above 30 percent. In Botswana, the saving rate has been more erratic, but has gone from 11 percent in 1971 (the earliest year for which the World Bank’s saving data base provides a figure), to above 30 percent in the mid-1980s, and has reached all the way to 53 percent in 1989 before declining thereafter. Lewis would have been astonished to see saving rates so high, but not surprised to learn that these two countries were placed at the very top of the economic growth league in the last three decades. 3 Lewis’ answer was based on the classical model emphasizing the functional distribution of income: as the profit share of national income rose, the rate of aggregate saving would rise alongside it. 5 When one mentions “saving transition” one immediately thinks of a handful of East Asian countries and a few others elsewhere (such as Botswana, and Chile since the second half of the 1980s). The approach in this paper is to look at saving transitions more systematically, by applying a common definition to the cross-national data. This has the advantage that we can look beyond the usual suspects, and avoid the optical illusions produced by a focus on a narrow set of countries. My definition of a saving transition is inspired by Lewis. I define a transition as a sustained increase in the saving rate by more than 5 percentage points of national income. To make this definition operational, I apply the following filter to the time-series of saving rates for each country. A country is said to undergo an investment transition at year T if the three-year moving average of its investment rate over a nine-year period starting at T exceeds by more than 5 percentage points the five-year average of its investment rate prior to T. I excluded from the analysis countries that have been the recipient of large resource windfalls, such as the major oil exporting countries. I also excluded cases where the post-transition saving rate remained below 10 percent. ~ More precisely, define ST as the three-year moving average of the saving rate with year T the first year of the average, and S$T as the five-year moving average with year T the terminal ~ year. For example, S1965 corresponds to the average for the years 1965, 1966, and 1967, while S$1965 is the average for the years 1961 through 1965. Applying our filter amounts to searching through the data for occurrences of any T such that the following are true: ~ 1. S T + i > S$T −1 + x for all i=0,1,..,n. 6 ~ 2. S T + i > 0.10 for all i=0,1,..,n. The parameter x stands for the threshold increase in saving rate (set to 0.05) while n captures the length of the horizon over which the transition is expected to be sustained. With a nine-year horizon starting at year 0, n=6. The first of these conditions checks that the (moving-average of the) saving rate after year T exceeds the prevailing average prior to T by more than 5 percentage points. The second condition ensures that the average saving rate subsequent to the candidate transition year exceeds 10 percent. If for any country there is more than a single year for which these conditions are satisfied, we check whether ten years or more separate the dates. If not, we assume there is a single transition and designate the earliest year in the sequence as the transition year.4 The World Bank saving database covers the years 1960-1995.5 However, given the leads and lags involved in the definition of a transition, the earliest possible year for a transition is 1965, and the latest year is 1987. The resulting list of transitions is shown in the first column of Table 2. A total of twenty transitions is listed, with two cases in the 1960s (Portugal 1965, Panama 1968), thirteen in the 1970s (China 1970, Egypt 1974, Jordan 1972, Lesotho 1977, Malta 1975, Pakistan 1968, Paraguay 1972, Philippines 1972, Singapore 1971, Sri Lanka 1976, Suriname 1972, Syria 1973, and Taiwan 1970), and five in the 1980s (Belize 1985, Chile 1985, Costa Rica 1983, Korea 1984, and Mauritius 1984). The list includes many well-known cases, such as Korea, Singapore, 4 Otherwise, countries would be listed with multiple transition years. However, with the values of n and x selected for our central case, there were no cases of multiple transitions. 5 Of course, not all countries have coverage over the entire period, and most countries lack data for 1995. Choosing a definition of saving that is theoretically less appropriate, such as gross domestic saving as a ratio of GDP, would have increased the available number of observations. Preliminary work indicates, however, that the qualitative conclusions would remain unaffected. 7 Taiwan, China, Chile and Mauritius, but also many others that are surprises (to me at least). As mentioned before, one of the advantages to letting the data “reveal” the transitions is that this allows us to focus on a broader range of countries that goes beyond the familiar cases.6 The transition years generally accord with conventional wisdom regarding the better known cases. Note however that Korea’s transition date is a rather late 1984. This is because prior to that date Korea’s saving rate was increasing at a steady but slow pace, and a transition is not picked up with a filter that requires a jump of 5 percentage points. When the threshold is lowered to 4 percentage points, Korea is listed as having two transitions, one in 1965 and another in 1975 (see below). In general, however, the selection and dating of transitions is not terribly sensitive to the particular thresholds used in operationalizing my definition. The remaining columns in Table 2 show alternative lists of transitions when the values of x and n are altered. Column (2) reduces by two years the horizon over which the increase in saving has to be maintained. This change results in a second transition for Mauritius (1971) and Suriname (1972), as well as twelve additional transitions in previously not listed countries. Among the additions, the case of Uganda (1988) is particularly intriguing, in view of the significant reforms that this country has undertaken since 1987. In column (3), the required increase in the saving rate is raised to 7.5 percentage points (while keeping the horizon same as in column (2)). This reduces the total number of transitions to 16, but also knocks out some clear-cut cases like as Taiwan and Pakistan. Finally, column (4) restores the original horizon but lowers the threshold to 4 percentage points. This throws up a 6 Botswana, discussed previously, is not included in the table because its diamond discoveries classify it as a resource-boom country. 8 few new countries such as Hong Kong (1971), Malaysia (1973), and Turkey (1984), and also makes a difference to the dating of Korea’s transitions as noted before. A certain arbitrariness in the definition of a saving transition is unavoidable. Each version of the filter I have used reveals its own anomalies. If the threshold value for the rise in saving is set high, we overlook cases of more gradual, but sustained increases in saving; if it is set low, we include too many instances of simple volatility. If the horizon is kept long, we lose countries with transitions in the 1980s; if it is too short, we pick up many increases in saving which are temporary. In view of these tradeoffs, the list of transitions in the first column of Table 2 strikes me as a good starting point III. The contours of saving transitions The typical pattern of a saving transition is represented in Figure 2, which shows the time paths of the mean and median of the saving rates in our 20-country sample. Rather than displaying the data chronologically, I have shown them in transition time. Country data are stacked so that year 0 corresponds to the transition year in each country, year 1 is the first year thereafter, and so on. The figure reveals that the typical jump in the saving rate around year 0 is significantly larger than 5 percentage points. The median saving rate in our sample goes from 14 percent in the five years before the transition to 23 percent in the next five years and 25 percent in the five years thereafter. The data for each country, averaged in five-year intervals, are shown in Table 3. The most spectacular cases are those of Belize (an increase from 12 percent to 24 percent), Lesotho (from 9 to 22 percent), and Suriname (from 20 to 41 percent). In one of these cases (Suriname), however, there is an equally spectacular reversal eventually. We observe similar reversals in four 9 other countries: Egypt, Philippines, Portugal, and Syria. In each one of these cases, the saving rates in years [T+10, T+14] is back to pre-transition levels. Since the reversal takes place after a long lag, however, it made sense to keep these countries in the sample. In the remaining countries, saving rates are substantially higher 10 to 15 years down the line—and in some instances by more than 20 percentage points (e.g., Lesotho and Singapore). IV. Growth, saving and investment I will discuss the specifics of some of the countries later in this paper. In this section, I analyze the three-way interrelationship among saving, investment and growth for the whole sample of transition countries. Figures 3 and 4 plot the median investment and growth rates alongside saving rates for these countries, with country data stacked according to transition years as before. The investment rate refers to gross investment as a ratio of gross national disposable income, while the growth rate is the growth of GNP. Note that the investment and growth rates are shown relative to the world average for the relevant year. The purpose of benchmarking the data in this fashion is to take out the effect of cycles common to most countries. As the figures show, saving transitions are associated with noticeable increases both in investment and growth rates. The correlation with investment is particularly strong. The median investment rate in our sample is 1.6 percentage points below the world average prior to the saving transition. In the first five years of the transition, it increases to 2.5 percentage points above the world average, and in the next five years it rises further to 4.8 percentage points above (see Table 3). In other words, the median investment rate rises by about 6.4 percentage points relative to the world average in countries undergoing saving transitions. 10 The growth rate also displays a significant spike around the time of the transition year (Figure 4), with its median value going from 1.5 percent (relative to the world average during the five years preceding the transition) to 3.9 percent (during the five years thereafter). So saving transitions are clearly associated with sharp increases in growth rates. But the striking message that comes out of Figure 4 is that the increase in growth tends to be temporary. Following the initial spike, the growth rate starts to decline, and ten years or so into the transition, it is back at the levels that prevailed in pre-transition years. The median growth rate in years [T+10, T+14] is 1.5 percent, the same as that in years [T-5, T-1] (see Table 3). The conclusion is that on average saving transitions do not seem to produce lasting increases in growth, even when the rise in saving itself is permanent. Excluding the five countries that eventually experience a reduction in saving rates—Egypt, Philippines, Portugal, Suriname, and Syria—does not affect this conclusion. The case of Pakistan, shown in Figure 5, represents perhaps the paradigmatic case. Pakistan’s saving rate went through a sustained increased after 1976, going from around 10 percent to above 20 percent. Till about 1982, this was accompanied by a significant increase in Pakistan’s relative growth performance—a rise of more than 6 percentage points in the growth rate (relative to the world average). Throughout the 1980s and early 1990s, however, Pakistan’s relative performance steadily slipped, to the point where the gap in its favor was entirely eliminated by 1994. The actual growth rates (without the benchmarking), show a similar, if less marked, cycle. 11 What explains this general pattern of temporary growth spikes?7 Note that our figures say nothing about the direction of causality between saving and growth, and the language I have used (“accompanied by,” “correlated with,” etc.) reflects that fact. One hypothesis is that causality runs from growth to saving, and that what we observe around our transition dates is an increase in saving resulting from an increase in growth, which in turn results from determinants other than saving. If there is some hysteresis in saving behavior, arising from habit formation or irreversible changes in the financial system, a subsequent reduction in growth—arising again from other determinants—might then leave saving relatively unaffected. In other words, the pattern that we observe could be the joint product of hysteresis and of saving being driven by growth. V. Growth transitions and saving behavior One way of getting more insight into this issue is to reverse the exercise and look for saving patterns in countries that undergo sustained growth transitions. If saving remains high in such countries well into the transition, our suspicions that growth is the driving force behind saving would be strengthened. For this purpose, I have defined a growth transition in a manner analogous to a saving transition. A growth transition is taken to be a sustained increase in the growth rate of real GNP by more than 2.5 percentage points. In particular, a country is said to undergo a growth transition at year T if the three-year moving average of its growth rate over a nine-year period starting at T 7 One explanation, not considered any further, is that we are observing the implications of the Solow growth model in operation. According to this model, a permanent rise in saving would increase the steady-state capital stock, but increase the economy’s growth rate only temporarily until a new balanced-growth path (at the previous rate of growth) is reached. However, the speed of adjustment in the Solow model tends to be too slow to explain the relatively rapid growth declines we observe in our data. For example, under a typical calibration (carried out in Romer 1996, 22), the half-life of convergence is around 18 years. 12 exceeds by more than 2.5 percentage points the five-year average of its growth rate prior to T. Countries where the post-transition growth rate averages less than 4 per cent were excluded from the analysis. Also excluded, as before, were resource-boom countries.8 The resulting list of countries and transition dates, along with the relevant data, is shown in Table 4. There are 18 countries in all, many of which have also had saving transitions, although the dates for the two kinds of transitions do not always coincide. There are also many new countries, however, such as Bangladesh (1974), Brazil (1966), Ghana (1984), and Thailand (1986). The median growth rate of income in these 18 countries rises from 1.1 percent prior to the transition to 7.0 percent and 7.2 percent in the first and second five years of the transition, respectively (these are actual growth rates, not relative to the world). The growth rate falls off somewhat eventually (ten years or more after the transition date), even though it remains higher than the pre-transition rate on average. Our main interest lies in what happens to the saving rate in these countries experiencing increased growth. The results in Table 4 are interesting in that respect. Benchmarked against the world norm, our 18 countries are significant under-performers on the saving front before the transition begins: their median saving rate is 7.5 percentage points below the world average. After the growth transition, however, this performance steadily improves. The median saving rate rises to -3.7, 1.2, and 2.5 percentage points above the world average in the three five-year periods following the transition year. Notice that saving performance continues to improve even in years [T+10, T+14], when, as mentioned previously, growth slows down. The cumulative improvement in median saving (relative to the world) amounts to a striking 10 percent of national 8 I also excluded a few very small, island economies: Cape Verde, Dominica, Solomon Islands, St. Vincent and the Grenadines. 13 income. The results without benchmarking the saving rate are virtually identical. The conclusion is clear: growth transitions tend to be followed by significant, and sustained, improvements in saving performance.9 What the data show, therefore, is an interesting asymmetry between saving and growth transitions. Significant increases in either saving or growth are generally accompanied by contemporaneous increases in the other variable. But while growth transitions lead to sustained increases in saving rates, saving transitions tend to result only in temporary increases in growth. These findings are in line with the hypothesis that it is mainly growth that drives the time-series relationship between the two variables. VI. Granger tests The preceding analysis allowed us to visualize the dynamic relationship between growth and saving over varying time horizons. The standard procedure for examining time precedence issues is to use some variant of Granger “causality” tests. Such tests are better at picking up short-run leads and lags than long-term relationships. Nonetheless, it is instructive to check whether the conclusions reached above are borne out by the evidence from more formal tests of this type. Table 5 presents the results of Granger “causality” tests on saving, investment, and growth (in real GNP) using our sample of countries with saving transitions. Each variable is regressed on the lagged values of itself and of the other two variables. Results with annual data as well as with 9 Note that the median (and mean) values reported in Table 4 are calculated using the whole sample, including countries for which observations are not available in certain time periods. However, excluding such countries and using a consistent set of countries throughout makes no difference to the conclusions in this paragraph and the previous one. 14 five-year averages are presented. In the latter case, the 1960-94 period is divided into seven subperiods of five years each. Regressions were run on each country individually (using annual data only), as well as on the entire panel. Bivariate regressions (using pairs among the three variables at a time) did not yield results that were qualitatively different, so they are not shown here. The results for specific countries do not reveal clear regularities. The pattern of signs on the estimated coefficients and their significance suggest that the stories for the individual countries have their own peculiarities. However, when the countries are pooled together, the outcome is quite revealing. There is strong evidence in the pooled data that in the very short run (i.e., using lags of a single year) growth precedes saving. The estimated coefficient on growth rate in the saving regression is positive and highly significant when annual data are used. The result continues to hold when we add country and period dummies to the regression. It also holds when we exclude lagged investment as a regressor. When we use five-year averages, the evidence that growth Grangercauses saving is weaker: the estimated coefficient on lagged growth is significant only when we do not include country and period dummies. As for the reverse relationship, the results indicate, if anything, a negative effect from saving to subsequent growth. This negative effect is found in regressions using five-year averages as well as annual data, as long as fixed effects for countries are included in the regression. These results (as well as those reported in the previous paragraph) continue to hold when the pooled regressions are run in first differences. Growth Granger-causes saving, while saving (negatively) Granger-causes growth. Taken at face value, the results suggest that an increase in saving is a deterrent to growth over the short- to medium-run. 15 With regard to investment, the regressions reveal two-way feedback effects between saving and investment in the short-run. But we find no statistically significant relationship between these two when five-year averages are used. Recapitulation. The Granger-type regressions reinforce the conclusions reached previously. Putting these results together with the findings discussed earlier, the story that emerges is one that emphasizes economic growth as the driving force behind the saving transitions we observe. In the short run, economic growth tends to have a clear positive effect on the saving rate. Increases in saving per se do not seem to produce a rise in growth, either in the short run or in the long run. The typical pattern for countries that undergo saving transitions is that their growth rates eventually return to pre-transition levels. VII. Determinants of saving What are the other determinants of saving besides (past) growth? As mentioned in the introduction, a number of studies have analyzed this question using panel techniques. Here, I take a fixed-effects approach, restricting the sample to the group of transition countries. The purpose is mainly to see whether the variables identified as determinants in the literature show up strongly in this sample as well. I use as my dependent variable private saving as well as total saving. The results are shown in Table 6. The main findings can be summarized as follows: 1. Even after other possible determinants are controlled for, lagged income growth exerts a statistically significant positive effect on the saving rate. This effect operates entirely on private saving. Lagged growth has no significant effect on public saving (not shown in the table). The estimated coefficients indicate that an increase in the growth rate of 1 percent 16 raises the saving rate the following year by between 0.2 and 0.3 percentage points of national income. 2. Public saving exerts a strong positive effect on aggregate saving. An increase in public saving of 1 percentage point of national income raises total national saving by 0.40-0.74 points. While public saving does crowd out private saving, the regressions reveal that the crowding out is far from complete. Indeed, Figure 6 shows that there is a noticeable increase in public saving in our sample around the time that the transitions begin. All in all, mobilizing public saving seems to be one of the most potent ways of raising national saving. 3. There is a positive and significant relationship between increases in the external terms of trade and saving. It is primarily private saving that is affected by the terms of trade. 4. Urbanization exerts a positive and statistically significant on national saving. This relationship is, once again, entirely due to the effect on private saving. 5. Aid flows from abroad appear to increase national saving. About 50 cents of a dollar of aid ends up as increased saving (see also Figure 7). The estimated coefficient is highly significant. 6. There is no statistically significant relationship in our sample between the dependency ratio in the population and the saving rate. 7. As regards financial variables, the stock of private credit appears to exert no effect on saving, while the estimated coefficient on the real interest rate on deposits in the banking system is actually negative and statistically significant. Lagging the real deposit rate does not change this result. As Figure 8 shows, saving transitions are associated with significant reductions in the real deposit rate. 17 8. There is a negative and significant relationship between the exports-GNP ratio and private saving. The reason for this is unclear.10 These results largely confirm those in Loayza, Schmidt-Hebbel, and Serven (1998), who use a global sample and try a variety of estimation techniques. In particular, the findings here on the positive role of income growth, public saving, and terms of trade growth mirror those in Loayza et al. Moreover, the negative coefficient on the real interest rate is reported in that paper as well. So there is little reason to think that our group of 20 countries with saving transitions behaves qualitatively differently from the larger sample. However, one respect in which our sample seems distinctive is the significant role played by workers’ remittances in many of the countries. Seven countries in our sample have received remittances in excess of one percent of GNP over sustained periods of time: Belize, Egypt, Jordan, Malta, Pakistan, Portugal, and Sri Lanka. Remittances were particularly large in Jordan, Egypt, Pakistan, and Portugal. These countries have supplied significant amounts of labor to booming economies nearby--the oil-rich Gulf states in the case of the first three countries, and West Germany in the case of Portugal. Since the early 1970s, if BOP and national-income data are to be believed, remittances have averaged 18 percent of GNP in Jordan, 8 percent in Egypt, 5 percent in Pakistan, and 8 percent in Portugal (Figure 9).11 In most of these countries, the saving transitions and the initial spurt in remittances were closely synchronized. This is certainly true of Egypt, Jordan, Pakistan, and Sri Lanka, all of which 10 Including the level of real national income as a regressor did not change any of these results. Real income itself enters with a statistically insignificant coefficient. Remember, however, that the regressions contain country fixed effects. 11 The source for these data is the World Bank’s World Development Indicators 1998 CD-ROM. Data are unavailable for Chile, Costa Rica, Lesotho, Mauritius, Singapore, Syria, Taiwan. Some of these countries are known to have been recipients of significant amounts of remittances as well. 18 benefited from the boom in the oil-producing states during the 1970s. The case of Portugal, which benefited from the German boom of the 1960s, is analogous. Hence, remittances appear to have been an important determinant of saving transitions in a sub-sample of our countries.12 VIII. Country stories As emphasized before, not all saving transitions lead to high growth in the long run. Only a small number of our countries have managed to combine increased saving with increased investment and growth on a sustained basis. How do these virtuous saving-investment-growth cycles get started? What are the respective roles of external factors, government policies, and institutional determinants? To answer these questions, I focus in greater detail on a few cases where such cycles seem to have taken hold. South Korea As Figure 10 shows, the saving rate in Korea has increased steadily since the early 1960s, from around 10 percent in 1960 to more than 35 percent by the late 1980s. But as the figure makes clear, saving has lagged behind investment until the second half of the 1980s. Our filter does not pick up a saving transition in Korea until 1984, which is quite late in view of the sharp pick-up in growth during the 1960s. Applying the same filter to the investment rate, we get an investment-transition date of 1965—two decades prior to the saving transition. Korea is a prime 12 Introducing the remittances/GNP ratio in the regressions yields a positive coefficient on this term. However, the estimated coefficient is statistically insignificant, perhaps because the sample size shrinks considerably due to the lack of data for many of the countries. 19 example of a country where high saving has been largely the product of high growth—itself the result of an investment boom starting in the early 1960s. Where did the investment boom come from? In Rodrik (1995), I have argued that the boom was largely the result of government policies that substantially increased the private profitability of investment from the early 1960s on. With the inauguration of President Park, who took power in a military coup in 1961, the investment climate in Korea improved sharply. In addition to eliminating obstacles to investment, government policy highly subsidized investment. The chief form of investment subsidy was the extension of credit to large business groups at negative real interest rates. Korean banks were nationalized after the military coup of 1961, and consequently the government obtained exclusive control over the allocation of funds in the economy. Investment was also subsidized through the socialization of investment risk in selected sectors. This came about because the government—most notably President Park himself— provided an implicit guarantee that the state would bail out those entrepreneurs investing in “desirable” activities if circumstances later threatened the profitability of these investments. The government played a direct, hands-on role by organizing private entrepreneurs into investments that they may not have otherwise made. In the words of Amsden (1989, 80-81), "[t]he initiative to enter new manufacturing branches has come primarily from the public sphere. Ignoring the 1950s... every major shift in industrial diversification in the decades of the 1960s and 1970s was instigated by the state...." Finally, public enterprises played a very important role in enhancing the profitability of private investment. They did so by ensuring that key inputs were available locally for private producers downstream. The government established many new public enterprises in the 1960s 20 and 1970s, particularly in basic industries characterized by high degree of linkages and scale economies (Jones and Sakong 1980). Not only did public enterprises account for a large share of manufacturing output and investment in Korea, their importance actually increased during the critical take-off years of the 1960s. It is true that many of the same policies were applied in other countries, with much less favorable results. The differences in Korea were that: (a) there was much greater consistency, predictability, and coherence in the application of the investment incentives; (b) the government bureaucracy exhibited less corruption and more competence; and (c) the educational-attainment level of the labor force was quite high for a country at Korea’s level of development. These factors reduced the administrative and rent-seeking costs of the interventions while enhancing their efficiency consequences. Taiwan and Singapore In Taiwan and Singapore, the saving transitions are dated earlier than in Korea (1970 and 1971, respectively), but otherwise the story is quite similar. In both cases, investment booms that took place in the 1960s led growth. Government policies that encouraged and subsidized private investment played a critical role in these booms. In both countries, the saving rate eventually overtook the investment rate, but not until the 1980s. In Taiwan, an important turning point was the Nineteen-Point Reform Program instituted in 1960. This contained a wide range of subsidies for investment, and signaled a major shift in government attitudes towards investment. The most important direct subsidies in Taiwan came in the form of tax incentives. The Statute for Encouragement of Investment (enacted in 1960 in conjunction with the Nineteen-Point Reform Program) entailed a significant expansion of the prevailing tax credit system for investment. These incentives were further expanded in 1965, at 21 which time specified manufacturing sectors (in basic metals, electrical machinery and electronics, machinery, transportation equipment, chemical fertilizers, petrochemicals, and natural gas pipe) were given complete exemption from import duties on plant equipment. As in Korea, it was common for the state to establish new plants in upstream industries. These were then either handed over to selected private entrepreneurs (as in the case of glass, plastics, steel, and cement) or run as public enterprises (Wade 1990, 78). In Singapore, investment was heavily subsidized as well. According to Young (1992, 21), there was a dramatic expansion in 1968 of the Singaporean government’s involvement in investment activities, with the Development Bank of Singapore (DBS) increasing its financial commitments eight-fold over a two-and-a-half year period. A substantial share of the economy came to be owned, directly or indirectly, by the Singaporean government. The year 1968 also marks the beginning of an investment boom, ahead of the saving transition in 1971 (Figure 11). These large investments were funded in part by surpluses on the current account of the government budget and by borrowing from the Central Provident Fund (Young 1992). But foreign saving also played an important role in closing the investment-saving gap until the early 1980s. Unlike Korea and Taiwan, Singapore focused its incentives heavily on foreign investors. The year 1968 marked a turning point with regard to foreign investment as well (Young 1992). Labor legislation passed that year significantly strengthened management’s bargaining power over issues of pay, benefits, and other working conditions. A wide range of tax incentives for investors were phased in or expanded after 1967, with exemptions from profits taxes taking the lead. While these incentives in principle did not discriminate between domestic and foreign investors, “because 22 they are usually linked to sizable investments involving advanced technologies in new (targeted) industries, the overwhelming majority of participants are foreign” (Young 1992, 23). Young emphasizes that Singapore’s Pioneer Industries Ordinance, the source of the most significant tax holidays provided to foreign investors, dates from 1959. He notes that the Ordinance failed to attract much foreign investment “until after 1968, when the Singaporean government began to expand its own financial participation in manufacturing and other sectors” (1992, 24). He suggests that after that date the government subsidized foreign investment beyond the tax incentives themselves, at quite exorbitant rates. Another distinguishing factor in Singapore is the role played by the Central Provident Fund (CPF). Established in 1955 as a compulsory, individualized social security account, the CPF has played an important role in mobilizing saving in Singapore. The contribution rates to the CPF were initially set at 5 percent of salary (by employee and employer alike). In 1968, the rates were raised to 6.5 percent, and since then they have been steadily raised until the mid-1980s, reaching 25 percent in 1984 (Lim et al. 1993, 118). The assets of the CPF are invested predominantly in government securities. The CPF must be considered an important factor behind the rise in the national saving rate in Singapore. The accumulation of CPF balances has accounted for more than 20 percent of gross domestic saving since 1971 (Lim et al. 1993, Table 3-4). The CPF has enabled an investment rate that would reach almost 50 percent by the early 1980s—higher than in any other East Asian country. At the same time, the root cause of the increase in investment must be sought not in the CPF itself, but in other government policies which, as noted above, resulted in higher public investment levels as well as higher private returns to investment in Singapore. Mauritius 23 Mauritius has experienced two spurts in saving, one in 1971 and another in 1984. The first of these, arising from a sugar boom, was short-lived and collapsed in 1974 along with sugar prices. The second one (which is the only one picked up by our filter) has survived so far, with the saving rate hovering around 26 percent (Figure 12). Even though the first saving boom was temporary, it played a crucial role in Mauritius’ development because it set the stage for a significant jump in investment, a jump which was to prove more durable than the saving boom itself. The increase in the island’s saving rate during the early 1970s resulted primarily from an improvement in the island’s terms of trade. World sugar prices began to rise in 1971, and the prices received by sugar producers more than tripled between 1972 and 1975 (Wellisz and Saw 1993, 235). The government had established an export-processing zone (EPZ) in 1970, shortly before the sugar boom got started. Enterprises operating under the EPZ—which entailed no particular geographic designation—were provided tariff-free access to imports of machinery and inputs, free repatriation of profits, a ten-year tax holiday (for foreign investors), and an implicit guarantee of wage moderation. The EPZ enabled saving to be channeled into productive, export-oriented investments, in turn setting the stage for an export boom in garments to European markets, where Mauritians could export quota-free. Without the EPZ, there is a good chance that these investments would not have taken place, or else would have been wasted in high-cost, inward-looking projects. In 1971, there were a mere nine enterprises in the EPZ, employing 644 people. Five years later, there were 85 EPZ enterprises, employing 17,171 workers, and making up 13 percent of the island’s exports (Wellisz and Saw 1993, p. 241). The economy’s investment rate rose from below 15 percent in the late 1960s to 30 percent a decade later. By 1976, the trend in world sugar prices was reversed, and domestic saving began to fall. 24 However, domestic investment remained high and the government maintained expansionary fiscal policies. This resulted in increasing resort to foreign borrowing and a deterioration of the balance of payments. A tripling of the country’s external debt between 1976 and 1979 brought the country to the verge of bankruptcy and forced the government to turn to the IMF and the World Bank for assistance (World Bank 1989, 4). Hit by cyclones and floods which caused extensive damage to crops and housing, the economy suffered a reduction in real income of more than 10 percent in 1980 (Wellisz and Saw, 245). The saving rate fell to its lowest level in two decades. During the early 1980s, Mauritius followed a classic adjustment program that eventually produced a return to high growth rates after 1984. The currency was devalued and there was a significant reduction in the fiscal deficit. The improvement in public saving around the year of the saving transition (1984) amounted to 6.6 percent of national income. There was also a range of structural reforms: imports were liberalized, price controls on most commodities were removed, and the tax system was reformed. At the same time, growing protectionism in the advanced industrial countries in textiles and clothing led major exporters (notably Hong Kong entrepreneurs) to look for production sites not yet subject to quantitative restrictions (Wellisz and Saw 1993, 249). These developments resulted in a second wave of investments in the EPZ and an export-led economic boom. The rise in saving after 1984 appears to be a clear case of growth-led saving. Following 1987, a number of financial-sector reforms enhanced the operation of the financial sector. First, interest rates were liberalized in July 1988 by abolishing the minimum deposit rate and the maximum loan rate guideline. Second, in 1987 two Bank of Mauritius saving bonds were issued to non-financial institutions. Third, Mauritius Housing Corporation tax-free saving bonds were introduced. The latter allows individuals to save for a down payment for a 25 house. Finally, the development of the stock market has made significant progress (World Bank 1989, 70-74). It is possible that these reforms have contributed to keeping the saving rate up. Chile13 National saving behavior in Chile has been extremely volatile (Figure 13). The saving rate during the late 1960s stood at around 15 percent. This rate then declined to 5.6 percent in 1973 and immediately peaked in 1974 at 24.3 percent. During the latter part of the 1970s the saving rate eventually leveled out at around 17 percent while investment steadily rose to levels exceeding 25 percent. At the beginning of the 1980s, the saving rate once again plummeted to low single digits. Then in 1985 Chile experienced a sustained saving transition. The average rate over the period from 1989 through 1995 has been above 25 percent As in the other cases, income growth has played a leading role in raising the level of saving in Chile. The saving transition in 1985 is associated with a turnaround in real income growth from -1.4 percent per annum to 7.8 percent (these figures are averages over five years prior to and following 1984). The correlation coefficient between growth and the saving rate over the entire period is 0.55. Chile’s case however, also provides some evidence that structural economic reforms may have had a positive if lagged effect on increasing the saving rate. These include financial sector liberalization, social security reform, and the stabilization policy. During the 1950s and 1960s Chile followed isolationist policies which created a highly distorted economy. The Allende government that took office in 1970 followed a socialist program that included policies to nationalize industry, banking and mining. This initially created an economic boom which however “quickly degenerated into an explosion of inflation, shortages, black markets, and huge losses in the state enterprises.” (Bosworth et al 1994, 5). What resulted 26 was a military coup in 1973, led by General Pinochet. The new government implemented a series of radical economic reforms that would extend into the next decade. The budget deficit was reduced from 25 percent of GDP in 1973 to 1 percent in 1975 (Bosworth et al 1994, 5). This was combined with restrictive monetary policy, trade and exchange rate liberalization and a comprehensive privatization plan. The reforms included extensive financial sector liberalization without adequate provision for prudential regulation of financial markets. The austerity program in combination with several external shocks resulted in the 1975 recession. To control inflation the currency was anchored to the U.S. dollar after 1978. A strong recovery soon followed from 1976 to 1982. A domestic investment boom was financed mainly by foreign capital inflows, which began to flow freely into the domestic economy with the removal of capital controls. However, saving remained depressed throughout the late 1970s for a number of reasons: the volatility of the economy in the 1970s created uncertainty about future earnings; consumer saving behavior did not radically change in response to the deregulation of the financial system and interest rates; and the large inflow of foreign capital between 1979 and 1981 pushed up asset prices generating a consumption boom through the wealth effect (Marfan and Bosworth 1994, 181-187). The increasing overvaluation of the currency and the growing current account deficit were the weak points of the system. In 1982, foreign commercial banks cut off credit to Chile and the economy spiraled into a deep recession. In the words of Bosworth, Dornbusch and Laban (1994, 8): The major economic crisis was partly a result of several external shocks—the drying up of voluntary external financing; the deterioration of the terms of trade; and the major increase in foreign interest rates. But the effect of external developments was exacerbated by the 13 This account draws extensively on a note prepared by Chad Steinberg. 27 mishandling of several domestic policies: the fixed exchange rate policy coupled with mandatory indexation of wages at 100 percent of past inflation; the sweeping opening of the capital account at the time of the boom; the radical liberalization of the domestic financial markets without the provision of proper regulations and controls; and the belief in the “automatic adjustment” mechanism, by which the market was expected to produce a quick adjustment to the new recessionary conditions without interference by the authorities. The bankruptcy of the Chilean financial sector required a bail-out that eventually cost the public sector more than 30 percent of GDP. The crisis of 1982 caused the saving rate to fully collapse. According to Marfan and Bosworth (1994, 192-1933) this “was due exclusively to reduced saving in the public sector, which was faced with the costs of rescuing the financial sector, a decline in the price of copper, the effects of the recession on tax collections, and the costs of the pension reform.” In the decade that followed, the Chilean economy made a remarkable recovery. Government policies in this period included a large currency devaluation, tight fiscal policy, a second round of privatization, and social security reform. In addition, the Central Bank was given a greater supervisory role in rebuilding the financial system. While the rise in saving after 1985 has all the hallmarks of a growth-led recovery, the magnitude and strength of the recovery are arguably attributable to structural changes in the economy. Some of the key changes are briefly listed below. Private sector incentives. Marfan and Bosworth (1994) emphasize the importance of the government’s efforts to increase private saving by reducing the private debt burden and reforming the tax system in 1984. The over-indebtedness of the private sector was dealt with by writing-off or rescheduling debts of troubled companies. These programs were successful in inducing firms to contribute additional resources to debt reduction. The efforts made by productive firms to reduce their debt may have been a significant factor in the rise of private saving during this period. (Marfan and Bosworth 1994, 193.) 28 Second, the tax reform in 1984 lowered both income and corporate tax rates. This may have increased incentives for corporations to retain earnings. Social security reforms. The Chilean effort in the early 1980s to privatize the social security system has received worldwide attention. The new system is based on a defined contribution plan with a mandatory contribution equal to 10 percent of wages. Workers have individual retirement accounts managed by private pension funds that are subject to government regulation and oversight. Nearly a quarter of private saving is in pension funds. The Copper Stabilization Fund. The CSF was designed to help avoid Dutch-disease type crises resulting from cycles in copper prices. The law setting up the CSF forces the public sector to save part of its income from copper sales when the market price of copper exceeds some reference price. With the dramatic rise of copper prices in the late 1980s, this law has had a positive effect on public saving. Stability and confidence. There were no major changes in financial conditions immediately before and after the saving transition in 1985. However, the stability of the Chilean macroeconomy has considerably improved, in comparison with the high inflation rates of the 1970s and the crisis years of the early 1980s. Confidence in the economy and in the government’s economic management has increased greatly. Transition to democracy has gone smoothly, with few changes in the underlying rules of the game in economic policy. These factors may have helped sustain the new higher levels of saving. Hence, the strength of the new financial system, the commitment of successive governments to stability, and reforms in the social security system have probably all played a role in the recovery of saving since 1985. While income growth is the most immediate stimulus behind 29 the saving increase, the structural changes made in the Chilean economy over the past two decades have likely affected the magnitude and the length of the saving transition. At the same time, it is worth pointing out that the Chilean saving rate still remains substantially below the level prevailing in East Asia. IX. Concluding remarks The central message of this paper is a negative one: focusing on saving performance does not seem to be a profitable strategy for understanding what makes for successful economic performance. Increases in saving appear to be the outcome of economic growth, not a fundamental determinant of it. The evidence indicates that countries that undergo saving transitions do not necessarily experience sustained increases in their growth rates. In fact, the typical pattern in our sample is that growth rates—benchmarked against world norms—return to their pre-transition levels within a decade. Countries that have experienced saving transitions due to rapid increases in worker remittances exemplify this point. Very few of these countries have experienced increases in their long-run growth rates. By contrast, countries that undergo growth transitions—due to improved terms of trade, increased domestic investment, and other reasons— do end up with permanently higher saving rates. The policy implication is clear: policies geared towards raising domestic saving do not deserve priority when designing economic programs. As our case studies demonstrate, the key to generating virtuous cycles of high growth-high investment-high saving is to kindle the animal spirits of entrepreneurs by increasing the expected profitability of their activities. Enhancing production and investment incentives seem preferable to working on saving incentives. 30 In the East Asian countries (South Korea, Singapore, and Taiwan), the virtuous cycles were started by employing a wide variety of investment subsidies and by undertaking public investments that increased the return to private investments. Implemented in a context of overall macroeconomic stability, good governance, and superior human resources, these interventions enhanced enterprise profits, crowded in private investment, and led to a steady increase in corporate and household saving. In Mauritius, a temporary saving boom generated by a terms-oftrade windfall was put to good use by setting up an export-processing zone unencumbered by the restrictions placed on entrepreneurship in the rest of the economy. The result was an investment boom that proved more durable than the saving boom and set the stage for export-oriented growth. In Chile, the rise in saving after 1985 was the result of economic recovery following the crisis of 1982-83 and the restoration of economic stability after a long period of macroeconomic instability. In all of these cases, policies and institutions having a direct bearing on saving performance did make a difference. It is unlikely that the saving rate in Singapore would have reached so high in the absence of the Central Provident Fund, and therefore unlikely that the investment rate would have climbed to 50 percent by the early 1980s. Similarly, it is plausible to think that financial policies and institutions in Chile boosted the saving rate beyond where it would have otherwise gone, even in the presence of a recovery of a similar magnitude. But it would be quite difficult to argue in these and other cases we looked at that it was saving per se that acted as the trigger for economic growth. 31 REFERENCES Amsden, Alice H., Asia's Next Giant: South Korea and Late Industrialization, New York, Oxford University Press, 1989. Attanasio Orazio, Lucio Picci, and Antonello Scorcu, “Saving, Growth, and Investment: A Macroeconomic Analysis Using A Panel of Countries,” December 1997. Bosworth, Barry, Rudi Dornbusch, and Raul Laban, eds, The Chilean Economy: Policy Lessons and Challenges, Washington, D.C.: Brookings Institution, 1994. Carroll, Christopher, and David Weil, “Saving and Growth: A Reinterpretation,” NBER Discussion Paper No. 4470, September 1993. Easterly, William, “The Ghost of Financing Gap,” unpublished paper, World Bank, July 1997. Edwards, Sebastian, “Why Are Latin America’s Saving Rates So Low?” Journal of Development Economics, vol. 51, 1996, 5-44. Gavin, Michael, Ricardo Hausmann, and Ernesto Talvi, “Saving Behavior in Latin America: Overview and Policy Issues,” Inter-American Development Bank, September 1996. Giovannini, Alberto, “Saving and the Rate of Interest in LDCs,” Journal of Development Economics, vol. 18, 1985, 197-217. Harrigan, Frank, “Saving Transitions in Southeast Asia,” Asian Development Bank, July 1996. Hayashi, Fumio, “Why Is Japan’s Saving Rate So Apparently High?” in S. Fischer, ed., NBER Macroeconomics Annual, Cambridge, MA, MIT Press, 1986. International Monetary Fund, World Economic Outlook, Washington, DC, May 1997. Jones, Leroy, and Il Sakong, Government, Business, and Entrepreneurship in Economic Development: The Korean Case, Cambridge, MA, Harvard University Press, 1980. Lim. Linda, Pang Eng Fong, and Ronald Findlay, “Singapore,” in Ronald Findlay and Stanislaw Wellisz, eds, Five Small Open Economies, Oxford, New York, Toronto and Melbourne: Oxford University Press, 1993. Loayza Norman, Klaus Schmidt-Hebbel, and Luis Serven, “What Drives Saving Across the World?” World Bank, January1998. Levine, Ross, and David Renelt, “A Sensitivity Analysis of Cross-Country Growth Regressions,” American Economic Review, vol. 82, 1992, 942-63. 32 Lewis, W.A., “Economic Development with Unlimited Supplies of Labor,” The Manchester School 22, May 1954. Marfan, Manuel and Barry Bosworth. “Saving, Investment, and Economic Growth” in Barry Bosworth, Rudi Dornbusch, and Raul Laban, eds., The Chilean Economy: Policy Lessons and Challenges, Washington, D.C.: Brookings Institution, 1994. Rodrik, Dani, “Getting Interventions Right: How South Korea and Taiwan Grew Rich,” Economic Policy, April 995. Romer, David, Advanced Macroeconomics, New York, McGraw-Hill, 1996. Wade, Robert, 1990, Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization, Princeton, NJ, Princeton University Press. Wellisz, Stanislaw and Philippe Lam Shin Saw, “Mauritius” in Ronald Findlay and Stanislaw Wellisz, eds, Five Small Open Economies, Oxford, New York, Toronto and Melbourne: Oxford University Press, 1993. World Bank, Mauritius: Managing Success, A World Bank Country Study, Washington, D.C.: 1989. Young, Alwyn, “A Tale of Two Cities: Factor Accumulation and Technical Change in Hong Kong and Singapore,” in O. Blanchard and S. Fischer, eds., NBER Macroeconomics Annual, Cambridge, MA, MIT Press, 1992. Table 1 Correlates of economic performance Low Growth2 1970-84 1985-95 Medium Growth 1970-84 1985-95 High Growth 1970-84 1985-95 Initial conditions GDP per capita in initial year3 Human capital4 1,697 2.2 2,185 3.3 2,266 3.2 2,188 3.8 1,776 3.5 2,734 5.4 Macro conditions Saving5 Investment5 Inflation rate per year (median) 17.8 19.0 11.0 16.5 19.4 14.1 18.5 22.1 10.9 19.2 21.1 11.1 26.0 27.4 11.3 31.4 31.9 7.8 Fiscal conditions Fiscal balances5 Government expenditure5 Government revenue5 -5.7 19.9 14.2 -5.6 25.0 19.4 -4.2 19.6 15.4 -3.3 20.0 16.7 -2.0 19.5 17.6 -2.4 18.9 16.5 International Net private capital flows6 Balance on current account5 Exports5 Imports5 20.2 -1.0 11.3 12.2 11.8 -2.6 17.2 17.7 12.9 -3.7 14.9 17.4 19.9 -1.4 17.2 18.1 66.9 -1.9 18.2 19.5 68.3 0.3 33.0 32.4 Source: IMF (1997), Table 19. Notes: 1 Excludes major oil exporting countries, Cyprus, and Malta. 2 Low growth is defined as per capita real income growth of less than one-half of 1 percent a year, which is roughly the mean growth rate minus one-half of the standard deviation of growth in the sample for the specified period. Correspondingly, high growth refers to rates above the mean plus one-half of the standard deviation (2.9 percent). 3 Group average in U.S. dollar terms, using purchasing power parity weights. 4 Average schooling years in population aged 15 and over. See Robert J. Barro and Jong-Wha Lee, "International Measures of Schooling Years and Schooling Quality," American Economic Review, Papers and Proceedings, Vol. 86 (May 1996), pp. 218-23. 5 In percent of GDP. 6 In percent of total private capital flow to developing countries. Excludes Asian newly industrialized economies. Page 1 Table 2 Saving transitions using different filters Country Belize Chile China Costa Rica Egypt, Arab Rep. Jordan Korea, Republic of Lesotho Malta Mauritius Pakistan Panama Paraguay Philippines Portugal Singapore Sri Lanka Suriname Syrian Arab Rep. Taiwan, China Barbados Dominican Republic Gambia, The Kiribati Malawi Mali Mauritania Mozambique, Rep. of Seychelles St. Lucia Thailand Uganda Burkina Faso Cameroon Hong Kong Malaysia Tunisia Turkey Morocco total no. of transitions Notes: See text. transition year x=5%, n=6 x=5%, n=4 x=7.5%, n=4 (1) (2) (3) 1985 1985 1985 1985 1985 1986 1970 1970 1983 1983 1974 1974 1972 1972 1973 1984 1984 1985 1977 1977 1977 1975 1975 1984 1971, 1984 1984 1976 1976 1968 1968 1972 1972 1972 1972 1965 1965 1971 1971 1972 1976 1976 1972 1972, 1986 1972, 1986 1973 1973 1973 1970 1970 1979 1978 1973 1987 1987 1984 1984 1972 1972 1986 1987 1986 1986 1984 1985 1985 1987 1988 x=4%, n=6 (4) 1985 1985 1970 1983 1974 1972 1965, 1975 1977 1975 1984 1976 1966 1972 1972 1965 1970 1976 1972 1973 1970 1973 1965, 1986 1984 1976 1971 1973 1970 1984 1969, 1984 20 34 16 33 Table 3 Contours of saving transitions (percent) transition Country year Belize 1985 Chile 1985 China 1970 Costa Rica 1983 Egypt 1974 Jordan 1972 South Korea 1984 Lesotho 1977 Malta 1975 Mauritius 1984 Pakistan 1976 Panama 1968 Paraguay 1972 Philippines 1972 Portugal 1965 Singapore 1971 Sri Lanka 1976 Suriname 1972 Syria 1973 Taiwan 1970 Median Mean savings (GNS/GNDI) [T-5,T-1] [T,T+4] [T+5,T+9] 11.5 7.8 22.6 13.6 11.7 10.2 24.7 8.5 19.5 14.5 10.4 15.5 11.7 20.0 20.3 17.4 11.9 20.2 12.2 22.4 23.5 18.0 29.1 20.9 18.6 18.1 33.2 22.3 26.6 24.6 17.7 22.8 18.0 26.0 25.6 25.3 17.0 40.6 23.2 30.4 14.1 15.3 23.4 24.1 investment (relative to world average) growth (relative to world average) [T+10,T+14] [T-5,T-1] [T,T+4] [T+5,T+9] 24.8 25.4 32.6 22.2 17.8 24.7 35.8 22.2 27.8 26.8 22.6 22.8 21.4 26.9 28.6 33.6 19.6 29.3 22.7 31.9 25.8 34.6 23.1 12.2 17.0 35.0 32.8 25.6 24.9 21.6 21.5 18.8 20.0 20.9 42.3 18.2 11.0 14.1 32.0 -2.0 -5.3 0.7 3.2 -8.9 -7.8 7.6 -11.6 0.0 -1.4 -8.6 -1.8 -5.4 -0.3 2.4 6.0 -6.8 4.8 -6.1 2.3 1.0 2.0 6.3 4.7 3.1 -6.3 9.1 -8.6 -2.7 3.5 -8.5 6.7 -2.3 3.0 1.9 19.2 -3.4 13.7 -0.9 6.9 6.7 5.9 8.0 4.8 4.7 0.7 14.6 -3.2 0.8 7.5 -7.2 7.8 3.9 5.1 4.2 17.3 2.4 3.3 -2.8 5.8 5.8 10.0 5.9 2.6 0.2 14.5 5.1 3.6 9.8 -4.9 5.3 1.4 0.3 1.5 24.6 -0.1 -4.1 -0.1 3.3 25.1 26.0 21.6 23.8 -1.6 -2.0 2.5 2.4 4.8 4.5 3.3 4.5 Note: T is the first year of transition for each country. [T+10,T+14] [T-5,T-1] [T,T+4] 0.4 -3.5 2.6 -4.0 -0.7 4.5 4.7 4.1 2.3 4.4 [T+5,T+9] [T+10,T+14] 2.9 4.0 1.7 2.2 5.1 3.0 8.7 2.8 -2.0 -1.0 3.1 -0.6 -0.7 1.5 6.2 -1.1 5.5 4.6 4.5 6.9 1.8 8.0 4.4 3.3 0.1 2.3 1.6 0.9 4.6 1.0 1.9 3.9 4.8 4.4 -0.8 1.2 3.0 4.7 -2.0 7.3 1.3 2.0 2.9 3.1 -0.6 3.6 5.2 -2.6 3.7 6.8 1.7 2.6 0.3 5.7 0.7 1.5 0.8 1.4 4.6 -2.8 -3.9 -2.0 5.4 0.2 5.3 -3.4 5.1 1.5 1.5 3.9 3.4 2.9 2.7 1.5 1.6 Table 4 Contours of growth transitions (percent) transition Country year Bangladesh 1974 Brazil 1966 Cameroon 1976 Chile 1984 China 1977 Costa Rica 1983 Dominican Republic1969 Ghana 1984 Lesotho 1969 Mali 1985 Malta 1966 Mauritius 1983 Pakistan 1976 Paraguay 1972 Philippines 1986 Seychelles 1985 Syrian Arab Rep. 1969 Thailand Median Mean 1986 growth savings (relative to world average) investment (relative to world average) [T-5,T-1] [T,T+4] [T+5,T+9] [T+10,T+14] [T-5,T-1] [T,T+4] [T+5,T+9] [T+10,T+14] [T-5,T-1] [T,T+4] -1.9 4.2 3.6 -1.1 3.0 -1.1 2.0 -3.1 6.4 -2.6 0.8 0.9 3.4 4.6 -1.8 -0.7 1.3 5.0 4.6 8.0 7.2 6.8 8.1 5.0 12.3 5.2 12.0 5.8 8.7 6.3 7.7 6.9 5.2 5.6 7.8 10.0 4.8 9.8 9.3 8.0 10.6 5.2 4.7 4.3 11.3 1.2 9.3 6.1 6.7 11.0 3.1 5.3 11.3 7.8 3.9 6.1 -3.2 5.7 7.8 4.6 3.4 3.9 3.0 7.1 10.5 4.0 4.6 -0.6 -14.4 1.0 -12.4 -8.4 7.4 -6.0 -12.4 -14.3 -9.8 -1.5 1.2 6.9 17.1 3.5 -3.7 -9.4 -8.8 -1.2 -1.5 8.3 4.9 0.8 1.2 2.9 1.8 16.6 -9.0 -2.0 -1.8 6.8 17.7 4.4 -1.4 -7.6 4.7 -13.7 -2.2 -2.9 -4.5 5.5 3.2 -5.0 -19.4 -8.6 -1.5 0.7 -8.6 -5.4 1.6 4.4 -7.7 6.2 -15.4 -0.9 2.6 -0.3 9.4 4.7 -2.4 -12.0 -11.3 -2.9 3.3 0.9 -8.5 -2.3 -2.0 1.9 -6.8 11.5 -11.1 1.1 4.8 5.8 13.2 4.8 -1.6 -8.3 -11.8 -1.3 -1.7 7.6 -7.2 3.9 -0.1 0.0 1.1 20.4 -9.5 -1.3 1.0 5.8 14.0 5.9 -1.2 -6.8 -6.2 -11.9 4.9 -4.5 -11.0 -7.5 3.4 3.2 -8.0 6.0 -17.4 -1.1 -6.0 -5.1 13.7 2.8 -5.9 -11.3 -13.1 -6.7 7.5 4.1 -3.6 -3.8 0.8 12.8 -5.7 11.0 1.1 1.3 7.0 7.4 7.2 7.2 4.6 4.4 -7.5 -5.0 -3.7 -1.5 1.2 1.6 2.5 2.3 -2.9 -3.4 -1.4 -1.7 0.5 1.1 -1.2 0.2 Note: T is the first year of transition for each country. 5.3 5.9 8.0 2.7 1.1 2.4 [T+5,T+9] [T+10,T+14] -2.5 9.1 -4.9 1.4 -2.6 Table 4 Table 5 Granger "causality" tests: saving, investment, and growth growth preceded by: savings investment Full sample, annual data no dummies w/country dummies w/country and period dummies -*** -** Full sample, 5-year averages no dummies w/country dummies w/country and period dummies -* -* investment preceded by: savings growth +* +* +* saving preceded by: growth investment +* +* +* +** +** +*** +** +** Countries, annual data Belize Chile +*** +*** -** +** -** -* +** +*** China Costa Rica Egypt Jordan -** South Korea Lesotho Malta Mauritius -*** -** +** +* +** +* -** -** +* +*** +* +** -* Pakistan Panama -* +* -** Paraguay +** Philippines +*** -* Portugal -*** Singapore +*** Sri Lanka Suriname Syria Taiwan -*** +*** -*** +** +** +* Note: Each variable is regressed on the lagged annual values of itself and the other two variables. Asterisks denote levels of statistical significance: * 99 percent; ** 95 percent; *** 90 percent. Blanks denote estimated coefficients that are not statistically significant. Table 6 Panel (within) regressions on determinants of saving in countries with saving transitions Dependent variable gross national savings/GNDI (gnsgndi) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) income growth-1 0.20* 0.26* 0.27* 0.31* 0.27* 0.26* 0.32* 0.27* 0.20* 0.27* (0.05) (0.05) (0.05) (0.06) (0.05) (0.05) (0.05) (0.05) (0.05) (0.05) public savings/GNDI 0.70* 0.62* 0.61* 0.40* 0.59* 0.60* 0.51* 0.74* -0.29* -0.37* (0.08) (0.08) (0.08) (0.13) (0.09) (0.08) (0.11) (0.09) (0.08) (0.09) terms of trade growth 0.09* 0.08* 0.06** 0.08* 0.08* 0.02 0.09* 0.09* (0.03) (0.03) (0.03) (0.03) (0.03) (0.04) (0.03) (0.03) urbanization 0.29* 0.35* 0.31* 0.30* 0.18 0.32* (0.11) (0.10) (0.11) (0.11) (0.14) (0.11) aid/GNP 0.54* (0.15) dependency ratio 0.15 (0.22) private credit/GNP -0.03 (0.03) real deposit rate -0.10** (0.04) exports/GNP -0.11* (0.03) N R2 438 0.27 350 0.29 350 0.31 287 0.32 350 0.31 350 0.31 203 0.47 348 0.33 438 0.18 350 0.22 private savings/GNDI (prsgndi) (11) (12) (13) (14) 0.27* 0.32* 0.27* 0.27* (0.05) (0.06) (0.05) (0.05) -0.39* -0.60* -0.40* -0.40* (0.08) (0.13) (0.09) (0.09) 0.08* 0.06** 0.08* 0.08* (0.03) (0.03) (0.03) (0.03) 0.28* 0.34* 0.30* 0.29* (0.11) (0.10) (0.11) (0.11) 0.56* (0.15) 0.14 (0.22) -0.02 (0.03) (15) 0.31* (0.05) -0.49* (0.11) 0.01 (0.04) 0.18 (0.14) (16) 0.27* (0.05) -0.25* (0.09) 0.09* (0.03) 0.31* (0.11) -0.10** (0.04) -0.11* (0.03) 350 0.24 287 0.28 Country and year dummies included (coefficients not shown). Levels of statistical significance: *99%, **95%, ***90%. 350 0.24 350 0.24 203 0.46 348 0.27 Figure 1 Saving rates in South Korea and Botswana 0.6 0.5 0.4 0.3 0.2 0.1 Botswana Korea, Republic of 94 92 90 88 86 84 82 80 78 76 74 72 70 68 66 64 62 60 0 Figure 2 Chart 1 Saving transitions 28% 26% 24% 22% 20% 18% 16% 14% 12% 10% -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 T Years S/Y median S/Y mean 10 12 14 16 18 20 Figure 3 Chart 1 Saving transitions and investment 28% 8% 26% 6% 24% 4% 22% 20% 2% 18% 0% 16% -2% 14% -4% 12% 10% -6% -14 -12 -10 -8 -6 -4 -2 0 2 S Median 4 6 8 I Median 10 12 14 16 18 20 Figure 4 Chart 2 Saving transitions and growth 0.06 0.32 0.05 0.27 0.04 0.03 0.22 0.02 0.17 0.01 0 0.12 -0.01 -0.02 0.07 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 T Years S Median Growth Median Page 1 12 14 16 18 20 Figure 5 Saving and growth in Pakistan 0.3 0.07 0.06 0.25 0.05 0.04 0.15 0.02 0.01 0.1 0 -0.01 0.05 -0.02 savings rate growth (relative to world, 3-yr avg.) 19 93 19 91 19 89 19 87 19 85 19 83 19 81 19 79 19 77 19 75 19 73 -0.03 19 71 0 growth 0.03 19 69 savings rate 0.2 Figure 6 Chart 1 Saving transitions and public savings 0.32 0.06 0.05 0.27 0.04 0.22 0.03 0.17 0.02 0.12 0.01 0.07 0 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 T Years S Median PS Median Page 1 8 10 12 14 16 18 20 Figure 7 Chart 1 Saving transitions and aid 0.32 0.06 0.05 0.27 0.04 0.22 0.03 0.17 0.02 0.12 0.01 0.07 0 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 T Years S Median Aid Median Page 1 8 10 12 14 16 18 20 Figure 8 Chart 1 Saving transitions and real deposit interest rate 0.32 0.2 0.15 0.27 0.1 0.22 0.05 0 0.17 -0.05 0.12 -0.1 0.07 -0.15 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 T Years S Median DepRate Median Page 1 12 14 16 18 20 Figure 9 Wokers' remittances as a share of GNP 0.25 0.2 0.15 0.1 0.05 Belize Egypt, Arab Rep. Jordan Malta Pakistan Portugal 1996 1995 1994 1993 1992 1991 1990 1989 1988 1987 1986 1985 1984 1983 1982 1981 1980 1979 1978 1977 1976 1975 1974 1973 1972 1971 1970 0 Sri Lanka Figure 10 Korea: saving and investment 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 60 62 64 66 68 70 72 74 76 GDI/GNDI 78 80 82 84 GNS/GNDI 86 88 90 92 94 Figure 11 Singapore: saving and investment 0.6 0.5 0.4 0.3 0.2 0.1 GDI/GNDI GNS/GNDI 93 91 89 87 85 83 81 79 77 75 73 71 69 67 65 0 Figure 12 Mauritius: saving and investment 0.35 0.3 0.25 0.2 0.15 0.1 0.05 GDI/GNDI GNS/GNDI 93 91 89 87 85 83 81 79 77 75 73 71 69 67 65 0 Figure 13 Chile: saving and investment 0.35 0.3 0.25 0.2 0.15 0.1 0.05 GDI/GNDI GNS/GNDI 95 93 91 89 87 85 83 81 79 77 75 73 71 69 67 65 0