While policymakers are understandably focused on short-term growth prospects in the current environment, Asia continues to face longer-term challenges to which attention will need to return over time. One of the key challenges is rapid population aging, which could have a significant macroeconomic impact on the region in coming decades. In particular, diverse demographic trends across Asia could influence external positions and, thereby, capital flows over the long term. Moreover, aging could affect returns on asset classes and change the structure of financial markets. Preemptive policy responses at both the country and regional levels would help to accommodate these aging-related changes.
Asia is facing a demographic shift that will see its population age significantly over the next half century. There are a variety of ways in which aging could affect the region’s economies over the long term:
-
Their growth potential could decline as their labor forces shrink.
-
Their fiscal positions could come under pressure from rising pension and health care spending needs and shrinking income tax bases.
-
Both private savings and investment could be adversely affected. As populations age, aggregate savings are expected to decline as the older cohorts draw down their savings to smooth lifetime consumption. At the same time, investments could also fall as the capital stock shrinks in tandem with the labor force. The interplay between savings and investment will have significant impacts on the dynamics of current accounts and, thereby, both inter- and intraregional capital flows. In turn, this could also have important implications for regional economic and financial integration.
-
The changes in savings associated with aging may have an impact on financial markets. Asset returns and relative asset prices could be affected as risk appetite changes with age, leading to a redistribution of savings from higher- to lower-risk instruments. In addition, financial product structures may also change in response to age-related demands.
The first two of these issues have been covered extensively in the economics literature and have been at the center of the public debate for years. However, much less work has been done in regard to Asia on the latter two issues, and these are the main focus of this chapter.
Demographic Trends and Patterns
Like other regions of the world, Asia19 has experienced a demographic shift over the last half century, reflecting a drop in fertility rates and an increase in life expectancies due to higher incomes, healthier lifestyles, and medical advances (Figure 3.1). Most industrial countries experienced a steep increase in births immediately after World War II, the emergence of the so-called baby boomer generation. But fertility rates subsequently dropped off and, together with rising life expectancies, this led to a higher median age. These trends were less pronounced and took hold somewhat later in less-developed countries (LDCs), including in Asia, where the median age started to rise only after the mid-1970s and remains well below advanced-country levels. Within Asia, aging trends have varied considerably. Japan has been at the forefront throughout, and aging is now approaching a more-advanced stage in the NIEs. Australia, China, New Zealand, and Thailand are in the middle phase, while the rest of the ASEAN-5 and India remain relatively young.
Looking ahead, populations across the world are set to age markedly, as baby boomers retire and life expectancies continue to improve.20 Among industrial economies outside of Asia, Europe will experience the most dramatic aging, with the old-age dependency ratio (the ratio of the elderly to the working-age population) expected to more than double over the next 40 years. Less-developed countries, on the other hand, are faced with a much more gradual aging of their populations, as their fertility rates are expected to remain relatively high. Aging in Asia will be rapid over the next four decades, with the old-age dependency ratio tripling to approach the U.S. level by 2050 (Figure 3.2).
Asia’s Impending Demographic Transformation
(Population in millions)
Source: United Nations Population Database.Aging trends within Asia will become even more diverse (Figure 3.3). Higher-income countries in the region will generally age much faster, with Japan and the NIEs seeing their old-age dependency ratios rise by around 3 to 4 times, and are also expected to age much faster than their current trading partners. Countries such as Australia, China, and Thailand will also age markedly, but the pace will be broadly in line with their trading partners. At the other end of the spectrum, India, Indonesia, Malaysia, the Philippines, and Vietnam are expected to age at a significantly slower pace.
How Demographics Affect Capital Flows and Financial Markets
Aggregate saving and investment are the channels through which population aging affects the current account, capital flows, and financial markets. Saving behavior, according to the life cycle hypothesis, is age-dependent, with people saving during their working years and running down their savings during retirement. Hence, aggregate saving should fall with population aging, as the relative number of pensioners—or prime consumers—increases.21 Investment depends positively on the return on capital, which, in turn, depends on capital’s relative availability. As the working-age population shrinks, capital becomes relatively more abundant, and, other things equal, returns on capital should fall. Hence, investment should fall with population aging.
This chapter finds that the current account deteriorates as populations age. Since both domestic savings and investment fall as the share of pensioners increases, the effect on the current account is theoretically indeterminate. However, econometric evidence presented in this chapter suggests that the effect of population aging on saving outweighs the effect on investment, leading to a deterioration in the current account as the population ages.
Age Structure
(In percent of total population)
Source: United Nations Population Database.1 Excludes Taiwan POC.By the same token, population aging should affect international capital flows. The effect follows from the discussion above, since capital flows reflect the financing of current account deficits or the foreign investment of current account surpluses. The channel through which population aging generates capital flows is interest rate arbitrage between countries at different stages of the demographic transition.22 A country that is relatively advanced in the demographic transition should experience a fall in savings that—according to the empirical findings of this chapter—outweighs the fall in investment. As a result, real interest rates would tend to move upward and attract foreign capital. A country that is at an earlier stage of the demographic transition will see a boost in investment to accommodate its rising labor force, but the saving generated by the larger labor force will be more than sufficient to cover the additional investment. This will lead, other things equal, to downward pressures on the interest rate and capital outflows. In equilibrium, the real interest rate—which is identical to the return on capital under perfect competition—should be equalized across countries.
It should be kept in mind that there are other determinants of capital flows that work in the opposite direction of demographic trends. According to the development literature, capital should flow from slow-growing industrial countries to fast-growing emerging markets. This is because emerging markets exhibit higher productivity growth than industrial countries. They should, therefore, offer higher returns on capital, which would be arbitraged away by capital inflows. In reality, capital flows to emerging markets are much smaller than theory would suggest, which has been blamed on capital market imperfections (Lucas, 1990).
Population aging can also have an impact on asset prices. The movement of saving and investment in response to population aging is mirrored by, respectively, the demand and supply of financial assets. As the relative number of pensioners rises, aggregate saving and the demand for financial assets fall. Since the working-age population shrinks at the same time, capital investment and the supply of financial assets also fall. However, the empirical findings in this chapter suggest that this effect is insufficient to prevent a fall in asset prices. Of course, such a scenario requires imperfect capital mobility: if relatively young countries can make up for the falling asset demand in relatively old countries, steep drops in asset prices should be avoided.
Relative asset returns and the structure of financial markets could also respond to population dynamics. Young and working people should prefer stocks over bonds since the returns on stocks have historically been higher and the long investment horizon of young workers allows them to accommodate the relatively higher return volatility. In contrast, without wage income after retirement, people tend to prefer the relatively safer investment in bonds. As a result, relative asset prices and returns should be affected by population aging. Moreover, the structure of financial markets could change as the supply of different types of assets responds to these changing preferences. For example, bond markets could expand relative to equity markets, and companies could rely more heavily on debt versus equity financing as the cohort of elderly people approaches retirement.
Empirical Findings23
This section provides empirical evidence of the impact of aging on net external flows—and, thereby, capital flows—and financial markets in Asia.
Demographics and the External Position
The empirical analysis suggests that the current account deteriorates as the population ages. A multi-country panel regression is used to estimate determinants of current account balances with a focus on the impact from aging. The exercise expands on much of the existing literature in two important aspects. First, relevant variables, including demographic variables, are expressed relative to trading partners, reflecting the fact that countries need to be at different stages of the demographic transition in order for it to have an impact on their external positions.24 Second, the estimation includes both the ratios of prime savers (aged 40–65) and prime consumers (aged 65+) to the working-age population. The latter corresponds to the “old-age dependency ratio.” This distinction has been made to better reflect the dynamics of aging in determining the current account. As shown in Table 3.1, the coefficients on both demographic variables (prime saver share and old-age dependency) are highly significant, showing that the current account improves with the share of prime savers and deteriorates with the share of prime consumers.
Aging and Current Account Positions—Global Panel1
Time fixed effects included. Standard errors asymptotically robust to both heteroscedasticity and serial correlation. Robust t-statistics in parentheses.
Aging and Current Account Positions—Global Panel1
| OLS | GMM | |||
|---|---|---|---|---|
| Prime saver share | 0.31 | *** | 0.42 | *** |
| (8.25) | (6.32) | |||
| Old-age dependency | -0.22 | *** | -0.16 | *** |
| (4.83) | (2.89) | |||
| Income level (lagged) | 0.01 | ** | -0.01 | |
| (2.34) | (1.63) | |||
| Income growth (lagged) | -0.19 | *** | -0.95 | *** |
| (4.19) | (3.43) | |||
| Openness (lagged) | 0.03 | *** | 0.04 | *** |
| (10.36) | (8.34) | |||
| Financial liberalization | -0.07 | *** | -0.03 | * |
| (6.43) | (1.91) | |||
| Fiscal balance | 0.15 | *** | 0.27 | *** |
| (5.20) | (3.00) | |||
| Oil balance | 0.35 | *** | 0.29 | *** |
| (6.77) | (3.81) | |||
| Net foreign assets (lagged) | … | 0.00 | *** | |
| (3.03) | ||||
| Asian crisis | … | 0.01 | *** | |
| (4.09) | ||||
| Constant | 0.04 | *** | -0.01 | |
| (5.48) | (0.77) | |||
| R -squared | 0.39 | 0.32 | ||
| Countries | 55 | 55 | ||
| Observations | 932 | 561 |
Time fixed effects included. Standard errors asymptotically robust to both heteroscedasticity and serial correlation. Robust t-statistics in parentheses.
Aging and Current Account Positions—Global Panel1
| OLS | GMM | |||
|---|---|---|---|---|
| Prime saver share | 0.31 | *** | 0.42 | *** |
| (8.25) | (6.32) | |||
| Old-age dependency | -0.22 | *** | -0.16 | *** |
| (4.83) | (2.89) | |||
| Income level (lagged) | 0.01 | ** | -0.01 | |
| (2.34) | (1.63) | |||
| Income growth (lagged) | -0.19 | *** | -0.95 | *** |
| (4.19) | (3.43) | |||
| Openness (lagged) | 0.03 | *** | 0.04 | *** |
| (10.36) | (8.34) | |||
| Financial liberalization | -0.07 | *** | -0.03 | * |
| (6.43) | (1.91) | |||
| Fiscal balance | 0.15 | *** | 0.27 | *** |
| (5.20) | (3.00) | |||
| Oil balance | 0.35 | *** | 0.29 | *** |
| (6.77) | (3.81) | |||
| Net foreign assets (lagged) | … | 0.00 | *** | |
| (3.03) | ||||
| Asian crisis | … | 0.01 | *** | |
| (4.09) | ||||
| Constant | 0.04 | *** | -0.01 | |
| (5.48) | (0.77) | |||
| R -squared | 0.39 | 0.32 | ||
| Countries | 55 | 55 | ||
| Observations | 932 | 561 |
Time fixed effects included. Standard errors asymptotically robust to both heteroscedasticity and serial correlation. Robust t-statistics in parentheses.
Projecting current account balances for Asia based on the estimated demographic coefficients yields the following results:
-
Hong Kong SAR, Japan, Korea, and Singapore, which are aging faster than their current trading partners, could face significant downward pressures on their savings-investment balances over the next 50 years, up to around 6 percent of GDP (Figure 3.4). These pressures would be driven by a decline in private savings as retirees draw down assets and a deterioration in public savings due to aging-related fiscal pressures. On the other hand, Korea and, to a lesser extent, Hong Kong SAR will still tend to build up savings over the next 15 to 30 years as the share of prime savers increases.
-
Countries expected to age more slowly than their trading partners, including India, Malaysia, the Philippines, and Vietnam, could experience a positive impact on their current account over the medium to long term, ranging from 6 to 7 percent of GDP. These countries will have both a relatively higher share of prime savers and a lower share of dissavers (Figure 3.5). The former will boost the current account mostly toward the end of the forecast horizon.
-
Australia, China, and New Zealand are likely to experience a less significant impact on their current account positions from demographics, since their aging trends are projected to mirror more broadly those of their trading partners.
Asia: Impact of Aging on Current Account Balances
(Change in percent of GDP)
Sources: United Nations Population Database; World Bank; and IMF staff estimates.Current Account Impact from Relative Changes in Age Cohorts
(From changes in share of savers and dissavers; in percent of GDP)
Sources: United Nations Population Database; World Bank; and IMF staff estimates.However, China’s current account impact is still positive, as its share of prime savers rises relative to those of its trading partners. This makes it all the more important for China to undertake public investments in health care, education, and pension systems to lessen precautionary savings and to liberalize the financial system to increase household disposable income and encourage consumption.
These estimated changes in current account balances are meant only to give an idea of the underlying impact of aging-related pressures and should not be interpreted as predictions of final outcomes. To begin with, the framework is not globally consistent, meaning that projected current account balances do not sum to zero. Moreover, policymakers should not be expected to stand by idly as demographic pressures become more intense. For example, in the fast-aging countries, the degree of intergenerational transfer through pay-as-you-go pension systems is likely to be reduced, as the transfer from prime savers (workers) to prime consumers (pensioners) exacerbates the effect of population aging on saving. Moreover, factors other than demographics will affect external positions over the medium term, notably relative productivity growth. The negative growth coefficient in the current account equation is evidence of this alleviating mechanism. Nevertheless, the results suggest that demographics will be an important driver of external balances and capital flows over the longer term.
Demographically induced current account surpluses and deficits broadly offset each other at the regional level. Looking at Asia as a whole (scaled by GDP), the adverse demographic pressures on Japan’s current account over the next 30 years are entirely matched by current account improvements elsewhere in the region, including in China, India, and the Philippines (Figure 3.6). Looking at the financing side, this implies that demographics would tend to drive capital flows from less-developed countries to relatively advanced countries, even if the actual capital flows could be intermediated outside the region.
Aggregate Impact of Aging on Current Account
(Contribution to change in percent of GDP)
Sources: United Nations Population Database; World Bank; and IMF staff estimates.1 Australia, New Zealand, Hong Kong SAR, Singapore, India, Malaysia, Philippines, Thailand, and Vietnam.The projected current account developments in Asia mirror broader global patterns. Fast-aging countries in Europe and the United States could face downward pressure on their current account positions from demographics. Emerging economies in Africa, Europe, Latin America, and the Middle East are broadly expected to see an improvement in their external balances, driven by their relatively favorable demographic positions (Figure 3.7). Compared to these countries, Asian countries fall at each end of the spectrum. The global perspective suggests that demographics will be of little help in correcting today’s global imbalances, which are characterized by a large U.S. current account deficit and offsetting current account surpluses in Asia and oil-exporting countries.
Impact of Aging on Current Account Balances Outside Asia
(Change in percent of GDP)
Sources: United Nations Population Database; World Bank; and IMF staff estimates.Demographics and Financial Markets
The “asset meltdown” hypothesis gained prominence in connection with the 1990s bull market in equities. It proposes that asset prices could decline sharply with population aging as the elderly shed financial assets. A cursory look at the relationship between the S&P 500 and the ratio of prime savers to the rest of the population (the population aged 40–64 over those aged 0–39 and 65+) in the United States seems to support this hypothesis (Figure 3.8). The weakness in stock prices in the 1970s and early 1980s coincided with a low share of prime savers. However, once baby boomers moved into their prime saving years, stock prices took off and (until recently) remained at unprecedented levels despite a steep correction after the dot.com bubble. While causality is debatable, some have argued that the correlation between demographics and stock indices implies that stock market performance could weaken further over the next 20 years in the United States. By extension, rapidly aging countries in Asia could also see downward pressures on stock prices, potentially threatening the adequacy of retirement savings.
Demographics and Asset Prices in the United States
Sources: Shiller (2008); and UN (2007).1 U.S. population aged 40-64 over population aged 0-39 and 65 and over.To investigate the effects of demographic changes on financial markets in the region, an Asia-specific regression was conducted. Panel data techniques with fixed effects were employed, with additional controls such as inflation, per capita income, urbanization, and openness, following authors like Davis (2006). Further, in considering the effect of demographics on asset prices, the exercise moves beyond most previous work by considering more rigorous controls incorporating nondemographic influences.25 Following the literature, the demographic variables used in the analysis are the share of the total population aged 40–64 (prime savers) and 65 and above (prime consumers).26 An openness measure is also included to control for any international effects.
The results suggest that demographic changes have influenced financial market valuations in Asia (Table 3.2). Looking first at asset price data, demographic variables do not appear to affect price-earnings ratios in Asia, while an increase in prime consumers tends to reduce stock price growth, consistent with falling demand for equities as populations age (columns 1–2). Stronger evidence of demographic effects emerges when asset return data are used: a rise in the share of prime savers increases stock returns and decreases bond yields, consistent with an upward effect on prices associated with high demand for financial assets during this stage of the life cycle (columns 3–5).27 In column 5, most of the additional variables have the expected signs, although lower growth appears to boost bond yields.
Demography and Asset Prices and Returns1
Time fixed effects included. t-statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Demography and Asset Prices and Returns1
| Stock Prices | Returns | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| P/E Ratios | Growth | Stocks | Bonds | |||||||
| (1) | (2) | (3) | (4) | (5) | ||||||
| Prime savers (40-64) | 3.313 | -0.010 | 2.148 | * | -0.269 | ** | -0.602 | ** | ||
| (1.40) | (0.74) | (1.74) | (2.35) | (3.06) | ||||||
| Prime consumers (65+) | 2.086 | -0.022 | ** | 2.931 | -0.216 | -0.053 | ||||
| (0.70) | (2.55) | (1.31) | (1.02) | (0.30) | ||||||
| Real per capita GDP | -0.017 | -0.932 | -0.045 | |||||||
| (0.01) | (1.08) | (0.24) | ||||||||
| Real GDP growth | -1.552 | 2.179 | * | -0.115 | ||||||
| (0.45) | (1.81) | (1.27) | ||||||||
| Inflation | -0.857 | -0.883 | * | -0.506 | ** | |||||
| (0.76) | (1.86) | (4.38) | ||||||||
| Urbanization rate | 1.847 | 0.417 | -0.300 | ** | ||||||
| (1.02) | (0.91) | (3.85) | ||||||||
| Openness | -0.059 | -0.017 | 0.007 | |||||||
| (0.37) | (0.19) | (0.28) | ||||||||
| Trend real GDP growth | 0.159 | * | -1.333 | ** | ||||||
| (1.79) | (4.27) | |||||||||
| Cyclical GDP growth | 0.055 | ** | -0.092 | ** | ||||||
| (8.21) | (3.33) | |||||||||
| Real long-term interest rate | -0.012 | |||||||||
| (1.23) | ||||||||||
| Lagged inflation | -0.594 | ** | ||||||||
| (5.23) | ||||||||||
| Change in inflation | -0.790 | ** | ||||||||
| (10.77) | ||||||||||
| R-squared | 0.109 | 0.529 | 0.426 | 0.681 | 0.854 | |||||
| Countries | 12 | 7 | 14 | 10 | 11 | |||||
| Observations | 206 | 140 | 284 | 271 | 270 | |||||
Time fixed effects included. t-statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Demography and Asset Prices and Returns1
| Stock Prices | Returns | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| P/E Ratios | Growth | Stocks | Bonds | |||||||
| (1) | (2) | (3) | (4) | (5) | ||||||
| Prime savers (40-64) | 3.313 | -0.010 | 2.148 | * | -0.269 | ** | -0.602 | ** | ||
| (1.40) | (0.74) | (1.74) | (2.35) | (3.06) | ||||||
| Prime consumers (65+) | 2.086 | -0.022 | ** | 2.931 | -0.216 | -0.053 | ||||
| (0.70) | (2.55) | (1.31) | (1.02) | (0.30) | ||||||
| Real per capita GDP | -0.017 | -0.932 | -0.045 | |||||||
| (0.01) | (1.08) | (0.24) | ||||||||
| Real GDP growth | -1.552 | 2.179 | * | -0.115 | ||||||
| (0.45) | (1.81) | (1.27) | ||||||||
| Inflation | -0.857 | -0.883 | * | -0.506 | ** | |||||
| (0.76) | (1.86) | (4.38) | ||||||||
| Urbanization rate | 1.847 | 0.417 | -0.300 | ** | ||||||
| (1.02) | (0.91) | (3.85) | ||||||||
| Openness | -0.059 | -0.017 | 0.007 | |||||||
| (0.37) | (0.19) | (0.28) | ||||||||
| Trend real GDP growth | 0.159 | * | -1.333 | ** | ||||||
| (1.79) | (4.27) | |||||||||
| Cyclical GDP growth | 0.055 | ** | -0.092 | ** | ||||||
| (8.21) | (3.33) | |||||||||
| Real long-term interest rate | -0.012 | |||||||||
| (1.23) | ||||||||||
| Lagged inflation | -0.594 | ** | ||||||||
| (5.23) | ||||||||||
| Change in inflation | -0.790 | ** | ||||||||
| (10.77) | ||||||||||
| R-squared | 0.109 | 0.529 | 0.426 | 0.681 | 0.854 | |||||
| Countries | 12 | 7 | 14 | 10 | 11 | |||||
| Observations | 206 | 140 | 284 | 271 | 270 | |||||
Time fixed effects included. t-statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Fears of rapid and pronounced declines in prices as suggested by the asset meltdown hypothesis appear largely unfounded. Indeed, the results imply a relatively modest and gradual impact of aging on bond yields and stock returns over long horizons. For the more-advanced economies in Asia, the fall in the share of prime savers could lower demand for financial assets and result in an increase in bond yields of between 100 and 200 basis points over the next half century (Figure 3.9). These trends may be exacerbated as older cohorts divest their assets as they retire and age-related pressures on fiscal positions require larger public bond issuance. At the other end of the spectrum, a rise in the share of prime savers in younger countries, such as India and the ASEAN-5, will tend to push yields down. These interest rate responses will support the potential redirection of capital flows suggested by the findings in the previous section. Demographic pressures on stock returns are similarly varied, ranging from a 10 percent decline in Japan to a 30 percent increase in Vietnam over the same extended period (Figure 3.10).
Age-Related Pressures on Real Bond Yields
(Change in percentage points)
Sources: United Nations Population Database; and IMF staff estimates.Age-Related Pressures on Real Stock Returns
(Change in percentage points)
Sources: United Nations Population Database; and IMF staff estimates.Demographics also affect the structure of financial markets in Asia (Table 3.3). Using data drawn from the World Bank’s Financial Structure Database (Beck, Demirgüç-Kunt, and Levine, 2007) (Figure 3.11), equations relating the GDP shares of bank loans, equity market capitalization, bond market capitalization, and the overall size of the financial sector in Asian economies to demographic variables were estimated. The specifications include a number of controls, including lagged bank credit to GDP as a proxy for financial development. The results point to substitution between asset classes of the kind suggested by the life cycle hypothesis. A rise in the elderly share tends to decrease the role of the banking sector in the economy, reflecting lower demand for consumer borrowing during old age (column 1 in the table), and reduces the size of the equity market while increasing bond market capitalization, consistent with a substitution toward safer assets in old age (columns 2 and 3). However, demographics do not affect the overall size of the financial sector (column 4), but only the structure. In this way, demographics would tend to support changes already under way in the region, by promoting securitization in less-developed economies, while spurring bond market development in more-advanced economies (Figure 3.12).
Demography and Financial Structure1
Time fixed effects included. t -statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Demography and Financial Structure1
| Loans/GDP (1) |
Equities/GDP (2) |
Bonds/GDP (3) |
Total size/GDP (4) |
|
|---|---|---|---|---|
| Prime savers (40-64) | 0.020 ** (2.21) |
-0.018 (0.87) |
-0.002 (0.17) |
0.048 (0.79) |
| Prime consumers (65+) | -0.037 ** (3.82) |
-0.129 ** (3.36) |
0.153 ** (7.70) |
-0.059 (0.98) |
| Real per capita GDP | 0.006 (1.10) |
0.145 ** (3.07) |
-0.014 ** (2.24) |
0.184 ** (2.52) |
| Inflation | 0.000 (1.04) |
-0.004 (1.12) |
0.000 (0.11) |
-0.010 (1.46) |
| Urbanization rate | 0.005 (1.42) |
-0.020 (1.52) |
0.017 ** (2.78) |
0.057 ** (2.16) |
| Openness | -0.001 ** (2.66) |
0.010 ** (4.68) |
0.000 (0.06) |
0.004 (1.21) |
| Lagged bank credit to GDP | 0.843 ** (15.14) |
-0.133 (0.90) |
-0.085 (0.90) |
0.377 ** (2.20) |
| R-squared | 0.920 | 0.596 | 0.206 | 0.366 |
| Countries | 21 | 16 | 10 | 10 |
| Observations | 578 | 300 | 136 | 136 |
Time fixed effects included. t -statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Demography and Financial Structure1
| Loans/GDP (1) |
Equities/GDP (2) |
Bonds/GDP (3) |
Total size/GDP (4) |
|
|---|---|---|---|---|
| Prime savers (40-64) | 0.020 ** (2.21) |
-0.018 (0.87) |
-0.002 (0.17) |
0.048 (0.79) |
| Prime consumers (65+) | -0.037 ** (3.82) |
-0.129 ** (3.36) |
0.153 ** (7.70) |
-0.059 (0.98) |
| Real per capita GDP | 0.006 (1.10) |
0.145 ** (3.07) |
-0.014 ** (2.24) |
0.184 ** (2.52) |
| Inflation | 0.000 (1.04) |
-0.004 (1.12) |
0.000 (0.11) |
-0.010 (1.46) |
| Urbanization rate | 0.005 (1.42) |
-0.020 (1.52) |
0.017 ** (2.78) |
0.057 ** (2.16) |
| Openness | -0.001 ** (2.66) |
0.010 ** (4.68) |
0.000 (0.06) |
0.004 (1.21) |
| Lagged bank credit to GDP | 0.843 ** (15.14) |
-0.133 (0.90) |
-0.085 (0.90) |
0.377 ** (2.20) |
| R-squared | 0.920 | 0.596 | 0.206 | 0.366 |
| Countries | 21 | 16 | 10 | 10 |
| Observations | 578 | 300 | 136 | 136 |
Time fixed effects included. t -statistics in parentheses, asymptotically robust to both heteroscedasticity and serial correlation. ** and * indicate that coefficients are significant at the 5 and 10 percent levels, respectively.
Financial Market Structures in Asia, 2005
(In percent of GDP)
Source: Beck, Demirgüç-Kunt, and Levine (2007).Age-Related Pressures on Representative Market Structures
(As percent of total financial sector)
Sources: Beck, Demirgüç-Kunt, and Levine (2007).Overall, our results point to plausible effects of demographic variables on asset prices and composition in Asia, but the actual impacts may turn out to be more limited. For instance, more fully integrated global capital markets could change the picture in coming decades: in particular, capital flows from younger economies in the region could help support asset demand and prices in rapidly aging countries. Equally, firms in fast-aging countries could diversify their operations toward faster-growing and younger economies to offset declines in their valuation. Moreover, other determinants of asset values, such as productivity gains and fiscal and monetary policy, may offset demographic influences.
Potential Policy Implications
Although the impact of aging is more pronounced over the long term, it is a relevant policy concern now. Given that age-related pressures are still some decades away, it may be tempting to delay reform until they begin to manifest themselves more visibly. However, such a strategy would be significantly more costly and distorting. Some of the required policy measures will have an effect only with significant lags, and it may take time to build consensus on politically sensitive reforms, possibly requiring a more gradualist approach. Importantly, deferring policy action would likely require drastic reforms to ensure “catch-up,” which would be more disruptive economically and politically. It is important to note, as well, that many policy reforms that are positive in their own right—such as prudent macroeconomic policies and structural reform aimed at providing an enabling business environment—become increasingly important in the context of large demographic shifts.
For countries at an advanced stage of aging, it will be important to address downward pressures on savings. It will be key to make progress on fiscal consolidation through tax and pension reform, including by stepping up prefunding of existing pay-as-you-go pension systems. This could be complemented by steps to promote privately funded pension schemes that invest savings abroad and whose returns are largely independent of demographic factors. There is also scope to press on with structural reforms to encourage labor force participation, raise mandatory retirement ages in line with life expectancies, and boost productivity.
For fast-aging economies that are projected to remain capital exporters for some time, increased openness and regional financial integration would bring great advantages. Demographic changes suggest that Hong Kong SAR and Korea will tend to run current account surpluses over the next 15 to 30 years. Maximizing returns on the associated savings and capital outflows during this window of opportunity will be important to prepare for the fiscal and external pressures these two countries will face over the longer term. Increased regional financial integration could help during this phase by channeling these flows towards younger and more productive economies in Asia. More generally, increased openness could enable a more efficient allocation of savings and better risk diversification. In this context, raising the ceiling on pension funds’ foreign investments would also be important, but should go hand in hand with measures to reform governance and management structures of such funds (Box 3.1).
For younger economies in Asia, the demographic transition puts a premium on ongoing efforts to raise the risk-adjusted rate of return and improve financial intermediation. The analysis suggests that, other things equal, younger and less-developed countries would tend to see capital outflows as a result of demographic transition. Hence, efforts to boost productivity and raise risk-adjusted rates of return could gain further urgency. While the projected rise in domestic saving is good news for young and developing countries, it will be important to improve financial intermediation to translate this into higher domestic investment. In this vein, it will also be important to ensure that the labor markets can absorb the growing workforces.
Financial product structures will need to adapt in response to aging. Increasingly, aging societies will demand financial products that allow the drawdown of accumulated wealth, such as annuities and reverse mortgages. However, markets in such products remain underdeveloped owing, in part, to systematic risks that cannot be diversified away. The government may have a role to play in the development of such markets:
-
Inflation risks. Annuities without inflation protection are risky, since retirees live for many years after leaving work, while annuities with inflation protection are expensive. The government could take on the inflation risks by issuing inflation-indexed bonds, which private companies could use to back inflation-protected annuities. Japan and Korea have started issuing inflation-indexed bonds, but markets are still very thin, with inflation-indexed bonds amounting to only 5 percent of global pension fund assets in 2005.
-
Duration risks. Pension funds and annuity providers routinely stress the need for bonds with maturities of 20 or more years, to match the maturity of their liabilities. The private sector provides some long-dated bonds (e.g., those for capital-intensive industries, utilities, and financial institutions), but issuance is insufficient and hampered by a number of factors, including lack of publicly traded benchmarks, tax disincentives, and—more cyclically—strong corporate earnings, until recently (Groome, Blancher, and Ramlogan, 2006).
Public Pension Funds in Asia: Maximizing Returns through Investment and Governance Reform
Asia is home to some of the largest public pension funds in the world. Asia’s 10 largest public pension funds are estimated to hold nearly US$1.8 trillion in assets. In addition to Japan’s Government Pension Investment Fund, whose assets of over $1 trillion make it the global leader, schemes in Korea, Taiwan Province of China, Singapore, Malaysia, and China rank among the world’s largest public pension funds. Looking ahead, the region’s public pension fund assets could increase threefold by 2015 (Allianz Global Investors, 2007) and even more over the medium term, mainly reflecting expanded coverage, rising income levels, and favorable return assumptions.
Raising returns on the region’s considerable pension fund assets will be important for accommodating age-related pressures. Historically, most public pension schemes in the region do not appear to have maximized risk-adjusted returns. Over the last five years, for instance, real rates of return in Indonesia, Korea, and Malaysia have averaged only between 2 and 4 percent, at best barely keeping up with the growth in real per capita incomes in those countries. Over longer horizons, too, returns have generally been low (Iglesias and Palacios, 2000). By contrast, returns for a number of public pension funds in OECD countries—including Canada, Ireland, Norway, and, within Asia, New Zealand—have been two to three times higher.
Sovereign Pension Funds in Asia
Sovereign Pension Funds in Asia
| Global Rank | Fund | Economy | Assets | |
|---|---|---|---|---|
| US$ billions | Percent of GDP | |||
| 1 | Government Pension Investment Fund | Japan | 1,072 | 24 |
| 3 | National Pension Fund | Korea | 232 | 24 |
| 4 | Postal Savings Fund | Taiwan POC | 130 | 34 |
| 7 | Central Provident Fund | Singapore | 95 | 59 |
| 8 | Employees Provident Fund | Malaysia | 95 | 51 |
| 9 | National Social Security Fund | China | 71 | 2 |
| 12 | Future Fund | Australia | 44 | 5 |
| 13 | Employees’ Provident Fund | India | 43 | 4 |
| 24 | Government Pension Fund | Thailand | 13 | 5 |
| 26 | New Zealand Superannuation | New Zealand | 10 | 8 |
Sovereign Pension Funds in Asia
| Global Rank | Fund | Economy | Assets | |
|---|---|---|---|---|
| US$ billions | Percent of GDP | |||
| 1 | Government Pension Investment Fund | Japan | 1,072 | 24 |
| 3 | National Pension Fund | Korea | 232 | 24 |
| 4 | Postal Savings Fund | Taiwan POC | 130 | 34 |
| 7 | Central Provident Fund | Singapore | 95 | 59 |
| 8 | Employees Provident Fund | Malaysia | 95 | 51 |
| 9 | National Social Security Fund | China | 71 | 2 |
| 12 | Future Fund | Australia | 44 | 5 |
| 13 | Employees’ Provident Fund | India | 43 | 4 |
| 24 | Government Pension Fund | Thailand | 13 | 5 |
| 26 | New Zealand Superannuation | New Zealand | 10 | 8 |
Public Pension Fund Performance: Real Rate of Return, 2003–2007
(In percent; deflated by average CPI inflation)
Sources: OECD; IMF, WEO database; and pension fund annual reports.Two key factors help explain this gap: investment strategies and governance structures.
Investment portfolios of most public pension funds in emerging Asia are not very well diversified, in terms either of asset class or of geographic allocation. In much of emerging Asia, domestic government bonds, deposits in state-owned banks, housing loans, or infrastructure projects historically account for the lion’s share. In 2007, public pension schemes in Korea, Malaysia, and Thailand allocated only 10 to 20 percent of their portfolios to equity investments, and less than 10 percent of their investments were made abroad. These allocation patterns—and the low associated rates of return—broadly mirror those observed in emerging markets (Hess and Impavido, 2003). While much better diversified in recent years, Japan’s portfolio also displayed similar biases prior to reforms in 2001.
By contrast, pension funds in the four aforementioned OECD economies invest a much higher proportion of their assets in equities (50 to 75 percent) and outside their home countries (35 to 75 percent). As current global events demonstrate, this search for higher return comes with greater short-term risk, but pension funds typically have long investment horizons, allowing them to ride out limited periods of financial market turbulence. However, it is critical that before investment restrictions are loosened and allocations broadened, risk management and governance structures be strengthened.
The governance and management of pension funds in emerging Asia generally lag best international practice. Over the last decade, some OECD countries—notably Canada (1997), Ireland (2000), Sweden (2000), New Zealand (2001), and Norway (2003)—have undertaken major reforms of their pension frameworks, with a view to raising rates of return in response to pressures resulting from aging. These experiences suggest several broad reform directions that, together with the guidelines for pension fund governance set out by Yermo (2008) and the ISSA (2005), help define best practices, notably (i) selecting professional board members through a process that maintains an arm’s-length relationship with the government, (ii) adopting explicit risk-adjusted return targets, (iii) allowing commercial investment policies to achieve targets and avoiding public policy considerations not justified by returns, (iv) making greater use of external fund managers, selected using objective criteria, (v) avoiding strict portfolio limits, especially on foreign investment, and (vi) ensuring high standards of transparency and disclosure. While progress is being made, the general structure of most public pension funds in the region does not adhere to these principles.
Asset Allocation of Public Pension Funds
(In percent of total assets)
Sources: OECD; CEIC Data Company Ltd; and pension fund annual reports.Across Asia, governments are taking steps to deregulate pension fund management, remove restrictions directing investment toward low-yielding domestic assets, ease pressures for policy-based investment, open up investment in equity and foreign assets, and outsource funds to external professional asset managers.1 In particular, Japan and Korea are leading the way, and their experiences could be helpful for the rest of Asia:
-
Japan. Until 2001, pension fund reserves were entrusted to the Fiscal Investment and Loan Programs (FILPs) under the Ministry of Finance. FILPs provided long-term financing for public purposes and contributed to the rapid economic growth of the 1960s and 1970s. However, as the Japanese economy matured, the effectiveness of some of the FILP operations came into question, particularly in areas where returns were low and positive spillovers had become less significant. In 2001, the Japanese government decided to discontinue its direct management of pension reserves and to set up the Government Pension Investment Fund (GPIF). In 2006, the GPIF was reorganized and made more independent of the government. The GPIF is externally audited by the Independent Administrative Corporation Evaluation Committee, consisting of financial and pension experts. Reserve funds have been progressively entrusted from the FILPs to the GPIF (2001–2008). The GPIF is mandated to conduct investments within a “basic management policy” framework, issued by the Ministry of Health, Labor, and Welfare. Within this framework, the GPIF investment committee, consisting of independent experts in finance and economics, draws up a more concrete and diversified “principal portfolio” to obtain the targeted rate of return and minimize risks. Pension funds are invested mainly in long-term, passive, and indexed financial products to avoid excessive risks and are entrusted to trust banks, insurance companies, and investment advisory firms.
-
Korea. The Korean government is reviewing the governance and management structure of the National Pension Fund (NPF). Currently falling under the Ministry of Health and Welfare, the pension fund’s investment decisions are not insulated from public policy considerations. The authorities are exploring ways to bring the fund’s governance closer to international practice, notably ensuring greater accountability and independence of the governing board and increased reliance on private expert management. Such reforms are welcome and could be complemented by simple and transparent investment objectives, focused on maximizing risk-adjusted returns; replacing strict portfolio restrictions with a transparent structure and sound prudential requirements; and increasing the role of external specialists—including auditors, actuaries, and asset managers. Concurrently with these changes in the governance framework, the authorities are pursuing a reform to diversify the investment portfolio of the NPF further in terms of both asset class and geographic allocation.
Appendix
Where possible, data were obtained from the April 2008 vintage of the World Economic Outlook database maintained by the IMF. Where necessary, data were retrieved from other databases maintained by the IMF or the World Bank Group. In particular, net foreign asset position data were retrieved from the International Investment Position database, and population share data were extracted from the World Development Indicators database. The projected demographic data were based on United Nations population data (2006 revision, medium variant).
Demographics and the External Position
The panel data set consists of annual observations on several macroeconomic variables for 55 developed and emerging market economies over the period 1973 through 2007, building on a data set prepared by Francis Vitek (Brooks, Edison, and Vitek, forthcoming). This panel data set is unbalanced, in the sense that the number of observations varies across macroeconomic variables along both the cross-sectional and time-series dimensions. The economies considered are Algeria, Argentina, Australia, Austria, Belgium, Brazil, Canada, Chile, China, Colombia, Croatia, the Czech Republic, Denmark, Egypt, Finland, France, Germany, Greece, Hong Kong SAR, Hungary, India, Indonesia, Ireland, Israel, Italy, Japan, Korea, Luxembourg, Malaysia, Mexico, Morocco, the Netherlands, New Zealand, Norway, Pakistan, Peru, the Philippines, Poland, Portugal, Russia, Saudi Arabia, Singapore, the Slovak Republic, Slovenia, South Africa, Spain, Sweden, Switzerland, Taiwan Province of China, Thailand, Tunisia, Turkey, the United Kingdom, the United States, and Venezuela.
The empirical approach applies the so-called macroeconomic balance approach (see Lee and others, 2008). The dependent variable is the ratio of the current account balance to output, while the explanatory variables considered are the ratio of the retirement-age population (65+) to the working-age population (15–64), expressed as a deviation from an arithmetic trade-weighted average across trading partners; the ratio of the prime saving population (40–64) to the working-age population (15–64), expressed as a deviation from an arithmetic trade-weighted average across trading partners; the logarithm of income per capita expressed in terms of purchasing power, expressed as a deviation from a geometric trade-weighted average across trading partners; the growth rate of income per capita expressed in terms of purchasing power, expressed as a deviation from an arithmetic trade-weighted average across trading partners; the ratio of the oil trade balance to output; the ratio of the fiscal balance to output, expressed as a deviation from an arithmetic trade-weighted average across trading partners; the ratio of external trade (exports plus imports) to output; the lagged ratio of the net foreign asset position to output; and a financial liberalization index that combines information on interest rate controls, credit controls, competition restrictions, state ownership, quality of banking supervision and regulation, policies to encourage capital market development, and capital account openness (access to domestic stock market for foreigners). The index is from Abiad, Detragiache, and Tressel (forthcoming).
The equations do not include country fixed effects and retain the cross-sectional information. Time fixed effects are included to capture developments that affect all countries similarly in a given year. The models are estimated using unrestricted panel regressions by ordinary least squares as well as restricted panel regressions by the generalized method of moments, both with heteroscedasticity-robust standard errors.
Demographics and Financial Markets
The empirical analysis of the effects of aging on financial market valuations relied on monthly share price and returns data and annual long-term government bond yields drawn from CEIC Data Company, Ltd, and Bloomberg. The additional control variables were drawn from the WEO database, with openness defined as the ratio of the sum of imports and exports to GDP. Trend GDP growth was estimated using a Hodrick-Prescott filter, with the cyclical component defined as the deviation of actual growth from this trend. The specifications are based on the framework for asset valuation suggested by Davis and Li (2003), which relates stock prices and bond yields to economic fundamentals. The full panel data set was unbalanced, consisting of annual observations on 14 Asian economies over the period 1975 to 2006. The economies considered are Australia, Bangladesh, China, Hong Kong SAR, India, Indonesia, Japan, Korea, Malaysia, New Zealand, the Philippines, Singapore, Sri Lanka, and Thailand. In some specifications, data unavailability limited the regressions to a subsample of this data set.
The empirical analysis of the effects of aging on financial market structure relied on the World Bank’s Financial Structure Database (Beck, Demirgüç-Kunt, and Levine, 2007), which provides data on volumes of equities, bonds, and bank assets for a wide set of economies. The full panel data set was unbalanced, consisting of annual observations on 22 Asian economies over the period 1975 to 2006. The economies considered are Australia, Bangladesh, Bhutan, Cambodia, China, Hong Kong SAR, India, Indonesia, Japan, Korea, Lao PDR, Malaysia, Maldives, Mongolia, Nepal, New Zealand, Papua New Guinea, the Philippines, Singapore, Sri Lanka, Thailand, and Vietnam. In some specifications, data unavailability again limited the regressions to a subsample of this data set.
In both sets of empirical work, the equations include country fixed effects and year dummies. The models are estimated using fixed-effects regressions, with heteroscedasticity-robust standard errors.
