Raising private consumption in emerging Asia is critical both for ensuring sustainable growth in the region and for contributing to an orderly resolution of global imbalances. Consumption as a share of GDP in emerging Asia is low by international standards and has been declining for some time. Demographics has played an important role in this regard, and the prospective aging of populations across much of the region will likely raise consumption significantly. But government policies can also play a supporting role. In particular, adequate provision of health care, education and social safety nets can reduce recourse to precautionary savings, while financial sector development can improve the ability of households to smooth consumption and to share in rising corporate profits.29
Introduction
As current account imbalances around the world have widened, attention has focused on the role of Asia as one of the sources of these imbalances. Some have argued that the emergence of global imbalances is linked to developments in emerging Asia, where saving rose but investment declined in the wake of the Asian crisis (See Box 5.1 for one take on the decline in investment). According to this view, the swing in Asia’s saving-investment balance is a major contributor to the global saving “glut” (Bernanke, 2005), which has allowed the United States to run record current account deficits. Within emerging Asia, attention has focused largely on China, where national saving has risen from 35 percent of GDP in 1980 to 50 percent in 2004, due in large part to an increase in private saving. While, in part, this rise in private saving is seen as integral to China’s economic development, needed to finance domestic capital formation, the rapid rise in private saving in recent years has raised concern among policymakers that private consumption in China needs to be strengthened, not only to address global imbalances but also to rebalance the economy towards more sustainable sources of growth (see the May 2006 REO). Meanwhile, there has been relatively little focus on private consumption trends in the rest of emerging Asia.
Credit, Investment and Growth in Post-Crisis Asia: A Tale of Two Sectors1
The protracted investment decline in post-crisis Asia, in contrast to a more rapid recovery of GDP, has remained a puzzle.2 This Box argues that an important source of emerging Asia’s post-crisis investment decline was that producers of nontradable (N) goods were starved of both domestic and foreign financing. As a result, Asia’s rapid recovery has been driven by the growth of the tradable (T) sector, which has been much less constrained by the credit shortfall. However, sustainable economic growth, as well as a successful unwinding of global imbalances, favors more balanced development, suggesting a need for policies to overcome impediments to investment in the N sector.
Gross Fixed Investment
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.Gross Fixed Investment
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.Gross Fixed Investment
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.In the aftermath of the Asian crisis, domestic bank credit and investment fell more sharply than GDP and has not yet fully recovered. Average investment in Korea, Indonesia, Malaysia and Thailand has declined—albeit from very high levels—by an average of 10.5 percentage points of GDP from 1990-1997 to 1998-2005. Over the same period, bank credit to corporations in the four countries decreased from 58.2 percent of GDP to 42.4 percent of GDP.
Total Lending by Banks to Corporations
(In percent of GDP)
Sources: CEIC Data Company Ltd; and IMF, APD country desks.Total Lending by Banks to Corporations
(In percent of GDP)
Sources: CEIC Data Company Ltd; and IMF, APD country desks.Total Lending by Banks to Corporations
(In percent of GDP)
Sources: CEIC Data Company Ltd; and IMF, APD country desks.The post-crisis decline in international bank flows, largely channeled to the local economy through domestic banks, further reduced available bank funds. The breakdown of private capital flows for the four crisis countries (Korea, Indonesia, Malaysia and Thailand) shows that the precipitous decline of net private capital flows reflected a sharp turn around in bank lending flows,3 while portfolio and FDI flows to the region remained relatively stable.
Private Capital Flows to Selected Asian Countries
(In billions of U.S. dollars)
Source: IMF, WEO database.Private Capital Flows to Selected Asian Countries
(In billions of U.S. dollars)
Source: IMF, WEO database.Private Capital Flows to Selected Asian Countries
(In billions of U.S. dollars)
Source: IMF, WEO database.The prolonged post-crisis credit slump and the slowing of international bank flows affected the N sector more seriously than the T sector. There are asymmetric financing opportunities for T and N firms. The T sector, typically large and able to pledge export receivables as collateral, has greater access to international capital markets, and so is not limited to domestic bank lending. The N sector, heavily populated by small and medium sized enterprises, relies predominantly on domestic and external bank credit for operations.4
What’s more, in the aftermath of the Asia crisis, with the implementation of stricter prudential regulations and stepped up supervision, banks became more cautious about lending to N firms, which often have relatively weak balance sheets.5 These N firms were hit especially hard during the crisis and benefited little from subsequent exchange rate depreciation due to their domestic focus. In addition, small and medium N firms have made slower progress in restructuring than larger export-oriented corporates, which worsened their profitability. As a result, loans to the N sector declined sharply as a percentage of total loans.
Selected Asian Countries: Output Ratio between the N Sector and T Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Indonesia, the T sector includes agriculture, mining and quarrying, and manufacturing industries. The N sector includes construction, trade, hotel and restaurant, transport and communication, banking, leasing and business, and private services. For Malaysia, the T sector is proxied by export-oriented industries. The N sector is proxied by domestic oriented industries. For Thailand, the T sector includes agriculture, mining and quarrying, and manufacturing. The N sector includes construction, transportation and communication, financial intermediation, real estate, renting and business activities, sales trade, health and social work, other community, and social and personal service activities.Selected Asian Countries: Output Ratio between the N Sector and T Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Indonesia, the T sector includes agriculture, mining and quarrying, and manufacturing industries. The N sector includes construction, trade, hotel and restaurant, transport and communication, banking, leasing and business, and private services. For Malaysia, the T sector is proxied by export-oriented industries. The N sector is proxied by domestic oriented industries. For Thailand, the T sector includes agriculture, mining and quarrying, and manufacturing. The N sector includes construction, transportation and communication, financial intermediation, real estate, renting and business activities, sales trade, health and social work, other community, and social and personal service activities.Selected Asian Countries: Output Ratio between the N Sector and T Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Indonesia, the T sector includes agriculture, mining and quarrying, and manufacturing industries. The N sector includes construction, trade, hotel and restaurant, transport and communication, banking, leasing and business, and private services. For Malaysia, the T sector is proxied by export-oriented industries. The N sector is proxied by domestic oriented industries. For Thailand, the T sector includes agriculture, mining and quarrying, and manufacturing. The N sector includes construction, transportation and communication, financial intermediation, real estate, renting and business activities, sales trade, health and social work, other community, and social and personal service activities.Lack of financing appears to have constrained N sector investment and output.6 In the aftermath of the crisis, N sector output also dropped more sharply than that of the T sector, experienced a sluggish recovery, and still lags the tradable sector’s development. In fact, post-crisis growth has been driven mainly by exports. In the wake of the crisis, the tradable sector experienced an acceleration of growth after a mild recession, thanks to real depreciation and the tradable sector’s use of cheaper non-tradable sector resources. The growth path of exports has closely tracked the growth of overall GDP in the region.
Selected Asian Countries: Share of Total Bank Loans to the N Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Thailand, the N sector includes construction, wholesale and retail trade, and services. For Indonesia, the N sector includes service rendering. For Malaysia, the N sector includes wholesale, retail, restaurants and hotel, and construction.Selected Asian Countries: Share of Total Bank Loans to the N Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Thailand, the N sector includes construction, wholesale and retail trade, and services. For Indonesia, the N sector includes service rendering. For Malaysia, the N sector includes wholesale, retail, restaurants and hotel, and construction.Selected Asian Countries: Share of Total Bank Loans to the N Sector1
Sources: CEIC Data Company Ltd; and IMF staff calculations.1 For Thailand, the N sector includes construction, wholesale and retail trade, and services. For Indonesia, the N sector includes service rendering. For Malaysia, the N sector includes wholesale, retail, restaurants and hotel, and construction.A recovery in investment in Asia would be helped by steps to ensure adequate financing for small N firms. Such firms are often important sources of employment. Moreover, because the development of the T sector depends on N sector inputs, the observed decline in N sector output may generate bottlenecks that eventually hinder T sector growth. The weak recovery in the N sector has also contributed to global imbalances. For Asian countries playing their role in adjusting the global imbalance, N sector investment, will likely need to increase. While the right approach will depend on country-specific circumstances, several types of policy may help:
Net Exports for Selected Asian Countries
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.Net Exports for Selected Asian Countries
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.Net Exports for Selected Asian Countries
(In percent of GDP)
Sources: IMF, WEO database; and IMF staff calculations.Expanding the potential sources of N firms’ financing. Asia needs to develop and deepen capital markets, especially bond markets, so that N firms can have well-diversified funding channels to avoid excessive reliance on bank credit.
Bank lending to small N firms can also be encouraged by eliminating roadblocks to such lending. For example, the establishment of credit bureaus or other forms of information exchange would allow banks to lend to N firms based on individualized credit analysis, rather than simply avoiding such lending.7
This chapter examines the determinants of private consumption in emerging Asia. The severity of the Asian crisis has given rise to the view that precautionary saving may have risen in recent years, given a perception of greater uncertainty. The effects of greater economic uncertainty, according to this view, are compounded by structural factors, such as limited social safety nets, underdeveloped financial markets that prevent households from smoothing the effects of adverse shocks on consumption, and population aging. This chapter attempts to explain why consumption has grown less rapidly than GDP in Asia. It presents a regression model that links private consumption in countries around the world to economic fundamentals, including economic growth, youth- and old-age dependency ratios, measures of financial development, and public consumption. It then runs two experiments. First, it compares actual consumption in emerging Asia to its fitted value, in order to assess whether private consumption in the region is consistent with fundamentals, based on evidence from a broader set of countries. Second, it estimates the model for a sample period that ends in 1996 and compares actual consumption in recent years to the out-of-sample forecast from this regression, to assess whether the link between private consumption and fundamentals has changed in the wake of the Asian crisis, an indication that precautionary saving has risen in recent years.
Private Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.Private Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.Private Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.
Change in Private Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.Change in Private Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.Change in Private Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.
Public Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.Public Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.Public Consumption
(In percent of GDP)
Sources: WDI; IMF, WEO database; and staff estimates.
Change in Public Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.Change in Public Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.Change in Public Consumption
(In percent of GDP)
Source: IMF, World Economic Outlook.Private Consumption Trends in Emerging Asia
The consumption-to-GDP ratio in emerging Asia has fallen since the 1980s, in contrast with other emerging market regions. This decline is driven in large part by private consumption, though public consumption has also fallen. In part, this long-term trend reflects the rapid pace of economic development in the region, as countries mobilized domestic saving in order to raise capital formation. What has surprised policymakers and analysts is that the consumption-to-GDP ratio has continued to fall in recent years, even as standards of living in some countries have approached those in industrialized countries.
Within emerging Asia, the recent decline in consumption relative to GDP is driven by China. Since the 1980s, the decline in private consumption relative to GDP has gathered pace in China, while it has slowed in Hong Kong SAR, Korea, Malaysia, Singapore and Thailand, perhaps reflecting the fact that economic development in some of these countries took off earlier. In India, the private consumption-to-GDP ratio has fallen steadily, though the pace of this decline has slowed, while in Indonesia, the Philippines and Taiwan Province of China private consumption has increased relative to GDP. Meanwhile, there is less of a regional pattern in public consumption, and China stands out in the region as the only country with a continuous decline in the public consumption-to-GDP ratio.
Though private consumption has fallen relative to GDP, it should be emphasized that real consumption growth is high by international standards. Since 1970, real private consumption growth in emerging Asia excluding China (EAXC) has averaged 5.7 percent, compared to 4.8 percent across industrial and emerging markets. Private consumption growth in EAXC has however not kept pace with real GDP growth, which has averaged 6.1 percent compared with 3.9 percent across developed and emerging markets.
Determinants of Private Consumption in Emerging Asia
In the wake of the Asian crisis, popular attention has focused on precautionary saving as a possible explanation for the declining importance of private consumption relative to GDP. However, in explaining the evolution of private consumption over time, a number of factors need to be taken into account:
Economic development: part of the longer term decline in consumption relative to GDP in emerging Asia could reflect a catch up with the industrialized world, as countries mobilized resources to boost capital formation. Indeed, the fact that the fall in private consumption has slowed in some wealthier countries (Hong Kong SAR, Korea, Malaysia, Singapore and Thailand), while it has accelerated in China, suggests that economic development plays at least some role in explaining private consumption patterns in emerging Asia.
Disposable income: where household survey data is available (China, the Philippines, and Thailand), there is evidence that a decline in household disposable income relative to GDP explains much of the decline in the private consumption-to-GDP ratio. For example, in Thailand private consumption declined from 66 percent of GDP in 1980 to 56 percent of GDP in 1990. Over the same period, disposable income fell from 79 percent of GDP to 69 percent of GDP. Since 1990, however, both ratios have been quite stable. Similar trends are apparent in China. In the Philippines, by contrast, disposable income has risen relative to GDP, likely reflecting growing remittances. The fall in disposable income relative to GDP in much of the region reflects primarily falling wages (as a share of GDP), driven by a slowdown in job creation as the capital intensity of production has risen. In addition, household ownership rates of equities are low by international standards, which has tended to limit positive wealth spillovers into consumption from rising corporate profits.
Social safety nets: despite the fall in the ratio of disposable income to GDP, this ratio measured 67 percent in emerging Asia in 2004, 5 percentage points of GDP above that in industrial countries, and therefore does not fully explain lower private consumption-to-GDP ratios. One widely held view is that the relative lack of social safety nets has interacted with the perception of greater economic uncertainty in the wake of the Asian crisis, such as the loss of lifetime job security, to push up precautionary saving. In the absence of cross-country data on social safety nets, we use total public consumption and government spending on health and education as proxies to examine this hypothesis. Public consumption averaged 11 percent of GDP in emerging Asia in 2004, compared with 18 percent in industrial countries, while public spending on health and education was only 6 percent on average in emerging Asia, while it was 15 percent in industrial countries. Factoring in these low levels of expenditure, the difference in “adjusted” private consumption-to-GDP ratios between industrial countries and emerging Asia is even more striking, as households in industrial countries consume publicly provided healthcare and education not captured in national accounts-based measures of personal consumption. Moreover, the differences in spending point to a relative lack of social safety nets in emerging Asia and are large enough that a rise in precautionary saving could have resulted from perceived greater uncertainty in the wake of the Asian crisis.
Selected Countries: Consumption, 2004
(In percent of GDP, unless otherwise indicated)
2003 figures.
2001 figures.
Personal consumption and government consumption on health and education.
Selected Countries: Consumption, 2004
(In percent of GDP, unless otherwise indicated)
| Personal disposable income | Taxes on Personal income1 | personal consumption/disposable income | Personal consumption | Public Consumption | Labor Income | Government consumption on health and education2 | Adjusted consumption3 | |
|---|---|---|---|---|---|---|---|---|
| United States | 74 | 9 | 95 | 70 | 16 | 57 | 22 | 92 |
| United Kingdom | 66 | 10 | 98 | 65 | 21 | 56 | 13 | 78 |
| Australia | 58 | 12 | 103 | 60 | 18 | 49 | 15 | 75 |
| Canada | 58 | 12 | 96 | 56 | 19 | 50 | 16 | 72 |
| Ireland | 49 | … | 91 | 44 | 14 | 40 | 12 | 56 |
| France | 62 | 8 | 90 | 56 | 24 | 52 | 16 | 72 |
| Germany | 66 | 9 | 88 | 57 | 19 | 51 | 16 | 74 |
| Italy | 67 | … | 90 | 60 | 20 | 42 | 14 | 74 |
| Japan | 59 | 8 | 96 | 57 | 18 | 51 | 12 | 69 |
| China | 60 | 1 | 69 | 41 | 10 | 56 | 3 | 44 |
| India | 84 | 2 | 76 | 67 | 11 | … | … | … |
| Korea | 54 | 3 | 95 | 51 | 13 | 44 | 14 | 65 |
| Singapore | 52 | 2 | 82 | 43 | 11 | … | … | … |
| Taiwan POC | 73 | 3 | 77 | 60 | 14 | 47 | … | … |
| Thailand | 65 | 2 | 87 | 57 | 10 | 30 | 6 | 63 |
| Philippines | 78 | 3 | 89 | 69 | 10 | 34 | 5 | 74 |
| Hong Kong SAR | … | … | … | … | 10 | … | … | … |
| Malaysia | … | … | … | … | 13 | … | 10 | … |
| Indonesia | … | … | … | … | 8 | … | 2 | … |
2003 figures.
2001 figures.
Personal consumption and government consumption on health and education.
Selected Countries: Consumption, 2004
(In percent of GDP, unless otherwise indicated)
| Personal disposable income | Taxes on Personal income1 | personal consumption/disposable income | Personal consumption | Public Consumption | Labor Income | Government consumption on health and education2 | Adjusted consumption3 | |
|---|---|---|---|---|---|---|---|---|
| United States | 74 | 9 | 95 | 70 | 16 | 57 | 22 | 92 |
| United Kingdom | 66 | 10 | 98 | 65 | 21 | 56 | 13 | 78 |
| Australia | 58 | 12 | 103 | 60 | 18 | 49 | 15 | 75 |
| Canada | 58 | 12 | 96 | 56 | 19 | 50 | 16 | 72 |
| Ireland | 49 | … | 91 | 44 | 14 | 40 | 12 | 56 |
| France | 62 | 8 | 90 | 56 | 24 | 52 | 16 | 72 |
| Germany | 66 | 9 | 88 | 57 | 19 | 51 | 16 | 74 |
| Italy | 67 | … | 90 | 60 | 20 | 42 | 14 | 74 |
| Japan | 59 | 8 | 96 | 57 | 18 | 51 | 12 | 69 |
| China | 60 | 1 | 69 | 41 | 10 | 56 | 3 | 44 |
| India | 84 | 2 | 76 | 67 | 11 | … | … | … |
| Korea | 54 | 3 | 95 | 51 | 13 | 44 | 14 | 65 |
| Singapore | 52 | 2 | 82 | 43 | 11 | … | … | … |
| Taiwan POC | 73 | 3 | 77 | 60 | 14 | 47 | … | … |
| Thailand | 65 | 2 | 87 | 57 | 10 | 30 | 6 | 63 |
| Philippines | 78 | 3 | 89 | 69 | 10 | 34 | 5 | 74 |
| Hong Kong SAR | … | … | … | … | 10 | … | … | … |
| Malaysia | … | … | … | … | 13 | … | 10 | … |
| Indonesia | … | … | … | … | 8 | … | 2 | … |
2003 figures.
2001 figures.
Personal consumption and government consumption on health and education.
Old Age Dependency in 2005 and Projected Aging in Emerging Asia
(In percent)
Source: United Nations.Old Age Dependency in 2005 and Projected Aging in Emerging Asia
(In percent)
Source: United Nations.Old Age Dependency in 2005 and Projected Aging in Emerging Asia
(In percent)
Source: United Nations.Demographics: emerging Asia as a whole is relatively young, but is projected to age more rapidly than other emerging market regions. Some of the decline in consumption-to-GDP ratios may, therefore, be related to life cycle factors. The prospect of rapid population aging in the decades ahead may have held back consumption growth, as households increased saving rates in preparation for retirement. If this factor turns out to be important, it can be expected to boost consumption-to-GDP ratios in the decades ahead, as more and more workers retire and consume out of savings.
Old Age Dependency in 2005 and Projected Changes to 2050
(In percent)
Source: United Nations.Old Age Dependency in 2005 and Projected Changes to 2050
(In percent)
Source: United Nations.Old Age Dependency in 2005 and Projected Changes to 2050
(In percent)
Source: United Nations.
Infrastructure Quality Ranking
(1 = poorly developed, 7 = among best in the world)
Source: Global Competitiveness Report 2003-2004, World Economic Forum.Infrastructure Quality Ranking
(1 = poorly developed, 7 = among best in the world)
Source: Global Competitiveness Report 2003-2004, World Economic Forum.Infrastructure Quality Ranking
(1 = poorly developed, 7 = among best in the world)
Source: Global Competitiveness Report 2003-2004, World Economic Forum.Financial markets: The perception of greater economic uncertainty in the wake of recent shocks may have interacted with a limited ability of households to borrow to smooth consumption to raise precautionary saving. Kuralbayeva and N’Diaye (forthcoming) find evidence of significant liquidity constraints in Indonesia, Korea, the Philippines, Taiwan Province of China, and Singapore.30 At the same time, it is also true that indicators of financial depth in emerging Asia compare favorably with other emerging market regions. For example, credit to the private sector has averaged 64 percent of GDP in EAXC since the 1970s, relative to 59 percent across advanced and emerging markets. It is possible, however, that indicators typically used to measure financial development may not capture constraints faced by households.
Infrastructure: a likely future driver of higher private consumption in the region would be higher purchases of durable goods and services. However, purchases of durable goods and services depend on the availability of infrastructure, including the provision of roads, electricity, telecommunications, and health services. The relatively poor provision of infrastructure in some countries in emerging Asia, principally China, India and the Philippines, may have reduced the scope for higher consumption of durable goods and services.
Estimation Results
This section uses panel regressions to explore the relative importance of these various factors in explaining the evolution of private consumption in emerging Asia. It uses a dataset that spans 44 countries from 1972 through 2004, which includes the following variables: the share of private and public consumption in GDP, real per capita GDP growth, elderly and youth dependency ratios, private sector credit (in percent of GDP) as a proxy for financial development, and real interest rates, which are generally thought to be negatively related to private consumption, though the strength of this effect tends to depend on households’ net asset position.
As a result of data limitations, it is not possible to control explicitly for the evolution of disposable incomes, the social safety net and the quality of infrastructure. Though public consumption may be seen as a proxy for the generosity of the social safety net, movements in public consumption are likely dominated by fiscal consolidation effects in the wake of the Asian crisis and therefore provide only a very weak link to the generosity of social safety nets. Moreover, public consumption may be in part a substitute for private consumption, while social safety nets can be thought of as complementary. Also, though the list of controls omits a quality of infrastructure measure, the regressions include country-specific fixed effects. As it is likely that the quality of infrastructure does not change rapidly over time, it is likely that these country dummies will control for this factor at least in part. Finally, the absence of cross-country data on disposable incomes is unfortunate, as this is likely an important driver of consumption behavior. However, the regressions include time dummies, which aim to control for global variation in private consumption. To the extent that the increased capital intensity of production has caused disposable incomes to fall over time across countries, these time dummies should capture some of this effect.
The dataset captures the main characteristics of emerging Asia outlined above. Private consumption in EAXC is lower than for the overall sample, and has tended to decline over time. Public consumption in EAXC is also lower than for the overall sample. Growth has exceeded that in other regions, though it has slowed slightly since the Asian crisis. Financial markets are relatively deep, especially compared to other emerging markets.
Data description1
Simple average of 1972-2004 or available periods for each country.
Data description1
| All | Emerging Market | Emerging Asia ex. China | Emerging Asia ex. China 1980-1989 | Emerging Asia ex. China 2000-2005 | |
|---|---|---|---|---|---|
| Private consumption (percent of GDP) | 59.8 | 61.5 | 58.8 | 59.9 | 56.6 |
| Elderly dependency ratio | 8.9 | 4.7 | 4.7 | 3.9 | 6.3 |
| Youth dependency ratio | 20.0 | 36.3 | 32.9 | 35.3 | 26.7 |
| Real interest rate | 0.7 | −1.3 | 2.3 | 3.1 | 2.0 |
| Credit to private sector (percent of GDP) | 58.8 | 41.1 | 63.9 | 49.5 | 89.1 |
| Real per capita GDP growth | 2.3 | 2.5 | 3.5 | 3.6 | 3.3 |
| Terms-of-trade growth | −0.1 | 0.0 | −0.6 | −0.7 | −1.7 |
| Public consumption (percent of GDP) | 16.2 | 13.6 | 11.0 | 11.3 | 11.2 |
| Real GDP growth | 3.9 | 4.8 | 6.1 | 6.2 | 4.7 |
| Real private consumption growth | 4.8 | 6.7 | 5.7 | 5.9 | 4.3 |
Simple average of 1972-2004 or available periods for each country.
Data description1
| All | Emerging Market | Emerging Asia ex. China | Emerging Asia ex. China 1980-1989 | Emerging Asia ex. China 2000-2005 | |
|---|---|---|---|---|---|
| Private consumption (percent of GDP) | 59.8 | 61.5 | 58.8 | 59.9 | 56.6 |
| Elderly dependency ratio | 8.9 | 4.7 | 4.7 | 3.9 | 6.3 |
| Youth dependency ratio | 20.0 | 36.3 | 32.9 | 35.3 | 26.7 |
| Real interest rate | 0.7 | −1.3 | 2.3 | 3.1 | 2.0 |
| Credit to private sector (percent of GDP) | 58.8 | 41.1 | 63.9 | 49.5 | 89.1 |
| Real per capita GDP growth | 2.3 | 2.5 | 3.5 | 3.6 | 3.3 |
| Terms-of-trade growth | −0.1 | 0.0 | −0.6 | −0.7 | −1.7 |
| Public consumption (percent of GDP) | 16.2 | 13.6 | 11.0 | 11.3 | 11.2 |
| Real GDP growth | 3.9 | 4.8 | 6.1 | 6.2 | 4.7 |
| Real private consumption growth | 4.8 | 6.7 | 5.7 | 5.9 | 4.3 |
Simple average of 1972-2004 or available periods for each country.
The empirical framework used to explore the link between private consumption and its determinants follows Arellano-Bond (1991). This approach allows country-specific fixed effects to be included together with the lagged dependent variable. As noted above, country-specific fixed effects are important because they control for omitted variables that do not vary over time but are specific to each country in the sample. Meanwhile, including lagged private consumption as a determinant of private consumption is important for two reasons. From a theoretical perspective, it reflects growing evidence that habit formation is a key element of consumption behavior, so that shocks will affect private consumption only slowly. From an empirical perspective, the lagged dependent variable permits a distinction between long and short run effects of shocks on consumption. For the purposes of this chapter, long run effects will be the main focus, and will be discussed below. The regression model used is as follows:
cit = α · ci(t − 1) + Xit β + χi + δt + Ɛi,t i = 1,…,N; t = 1,…, T
where cit is the consumption-to-GDP ratio for country i in period t; Xit is a matrix of the explanatory variables; χi is a country dummy that controls for country-specific features that are time invariant and determine the consumption-to-GDP ratio; and δt is a time dummy that controls for global trends in consumption-to-GDP ratios. As noted above, the regressions also include the lagged consumption-to-GDP ratio as an explanatory variable.
The regression results suggest that private consumption relates to fundamentals in the expected way. Indeed, most of the results are in line with the World Economic Outlook chapter on the determinants of private saving (IMF, Chapter II, 2005):
Global Private Consumption : Panel Regression (Long-term effects)1
Period of 1972-2004 or available periods for each country. Taiwan POC is excluded from Emerging Asia, due to the lack of data.
Global Private Consumption : Panel Regression (Long-term effects)1
| Private Consumption (percent of GDP) | |||||
|---|---|---|---|---|---|
| All | Emerging Market | All 1975-96 | Emerging Market 1975-1996 | Emerging Asia 1975-1996 | |
| Elderly dependency ratio | 2.98 | 1.58 | 1.50 | -0.81 | 3.22 |
| Youth dependency ratio | 0.09 | -0.73 | -0.19 | -0.82 | 0.31 |
| Real interest rate | 0.05 | 0.05 | 0.03 | 0.04 | -0.48 |
| Credit to private sector (percent of GDP) | -0.03 | -0.08 | -0.01 | -0.04 | -0.04 |
| Real GDP growth | -0.49 | -0.41 | -0.45 | -0.36 | -1.64 |
| Terms-of-trade growth | -0.19 | -0.16 | -0.15 | -0.12 | -0.03 |
| Public consumption (percent of GDP) | -1.63 | -1.54 | -1.57 | -1.38 | 0.14 |
Period of 1972-2004 or available periods for each country. Taiwan POC is excluded from Emerging Asia, due to the lack of data.
Global Private Consumption : Panel Regression (Long-term effects)1
| Private Consumption (percent of GDP) | |||||
|---|---|---|---|---|---|
| All | Emerging Market | All 1975-96 | Emerging Market 1975-1996 | Emerging Asia 1975-1996 | |
| Elderly dependency ratio | 2.98 | 1.58 | 1.50 | -0.81 | 3.22 |
| Youth dependency ratio | 0.09 | -0.73 | -0.19 | -0.82 | 0.31 |
| Real interest rate | 0.05 | 0.05 | 0.03 | 0.04 | -0.48 |
| Credit to private sector (percent of GDP) | -0.03 | -0.08 | -0.01 | -0.04 | -0.04 |
| Real GDP growth | -0.49 | -0.41 | -0.45 | -0.36 | -1.64 |
| Terms-of-trade growth | -0.19 | -0.16 | -0.15 | -0.12 | -0.03 |
| Public consumption (percent of GDP) | -1.63 | -1.54 | -1.57 | -1.38 | 0.14 |
Period of 1972-2004 or available periods for each country. Taiwan POC is excluded from Emerging Asia, due to the lack of data.
Higher output growth tends to lower the share of consumption in GDP. In other words, households tend to save some part of a growth spurt, in order to smooth consumption over time. The coefficient estimates suggest that a one percentage point increase in per capita output growth would in the long run reduce consumption by about ½ percent of GDP.31
Population aging tends to raise consumption over time quite significantly, by about 3 percentage points of GDP for each percentage point increase in the elderly dependency ratio. Meanwhile, the youth dependency ratio does not have a statistically significant effect on consumption. This suggests that population aging may have profound effects on consumption-to-GDP ratios going forward, as a rise in consumption-to-GDP ratios driven by rising old age dependency may not be offset by a decline in consumption linked to falling youth dependency ratios.
Government consumption is found to reduce private consumption, suggesting that the two types of consumption tend, in the aggregate, to be substitutes. The coefficient estimates imply that a one percentage point of GDP increase in public consumption in the long run reduces private consumption by about 1½ percent of GDP (although this point estimate is high, it is not statistically different from one, suggesting a one-for-one offset). This result contrasts somewhat with the empirical literature, which finds in general that increased government spending is either not offset, or only partly offset, by lower private consumption.32
Positive terms of trade shocks tend to lower consumption, which can again be explained by households smoothing consumption in the face of transitory shocks. The point estimates suggest that a 10 percent increase in the terms of trade reduces consumption by close to 2 percentage points of GDP.
Increases in private sector credit are not associated with changes in private consumption. This finding is similar to Eichengreen (2006) and likely reflects the fact that this measure of financial deepening fails to capture borrowing constraints that households effectively face.
Finally, the coefficient for the real interest rate is positive and marginally significant, and is therefore the wrong sign. This result is in line with the World Economic Outlook chapter on the determinants of national savings rates in industrial countries, which also failed to find the expected sign. This may point to a role for wealth effects in explaining consumption.
These regression results are subject to the caveat that controls for disposable income, the social safety net, and the quality of infrastructure are not included. The omission of disposable income is perhaps most problematic, as the evolution of this variable is likely to be dominated by country- and region-specific developments, such as labor market reforms and product market regulations. In addition, the wave of fiscal consolidation in emerging Asia following the late-1990s may have resulted in a deterioration in the quality of infrastructure, which would not be adequately captured by country dummies. Similarly, some countries in emerging Asia introduced social safety nets in the wake of the Asian crisis, so that country dummies will again only be an imperfect control for this omission.
Is Private Consumption in Emerging Asia Too Low?
This section uses the regression results to test whether private consumption in emerging Asia is too low, i.e. whether the empirical estimates suggest that consumption should be higher, given the level of fundamentals. The absence of cross-country data on disposable income and the generosity of social safety nets complicates the task of assessing whether households have become more cautious in the wake of the Asian crisis. As a result, this section devises two experiments to test whether the link between private consumption and fundamentals is somehow “out of whack.”
The first experiment assesses whether private consumption is in line with fundamentals, based on evidence from a broader set of countries. It compares actual and fitted values for consumption-to-GDP ratios in emerging Asia, to assess whether consumption relative to GDP is in line with fundamentals, based on the cross-country regressions. This comparison shows that deviations of predicted values from actuals have been minor in recent years and that the long-run trend of consumption can be well explained by changes in fundamentals. That said, this experiment should be interpreted with caution, given that the regressions include country dummies and lagged consumption, so that the deviation of actual from fitted consumption is likely to be small by design.
The second experiment explores whether the link between private consumption and fundamentals has changed in the wake of the Asian crisis. Regression results using the sample until 1996 are used to produce out-of-sample forecasts for 1997-2004. The results show that in 1998 private consumption-to-GDP fell sharply below the out-of-sample forecast in Korea (6 percentage points of GDP below), Malaysia (6 percentage points below), Singapore (2 percentage points below) and Thailand (4 percentage points below). This is a possible indication that, rather than smoothing consumption, households raised precautionary saving in the face of the Asian crisis, due to uncertainty over the scale and length of the crisis. However, private consumption in these countries bounced back in subsequent years and in many cases exceeded the out-of-sample forecast, evidence that precautionary saving was unwound as uncertainty surrounding the Asian crisis receded. The one case where the private consumption-to-GDP ratio is consistently below the out-of-sample forecast is China, which is consistent with the notion that precautionary saving may have risen permanently in the wake of the Asian crisis, so that for the same level of fundamentals as in the period through 1996, the consumption-to-GDP ratios is now lower.
Though these out-of-sample results are compelling, they are subject to a number of caveats. The omission of disposable income means that any divergence between actual and predicted private consumption could reflect the compression over time in disposable income. This factor is especially important for the out-of-sample results for China, where disposable income has fallen sharply relative to GDP in recent years. Moreover, other factors besides a rise in precautionary savings may have altered the relationship between private consumption and fundamentals in China. As a result, this out-of-sample experiment should be seen as an upper bound for the possible impact from a change in precautionary behavior.
Emerging Asia: Actual vs. Predicted Private Consumption (1972-2004)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Emerging Asia: Out-of-Sample Forecasts of Private Consumption (1997-2004)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Hong Kong SAR: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Philippines: Private Consumption
(In percent of GDP)
Potential Policy Implications
This chapter finds that private consumption in emerging Asia has a large life cycle component. In the years ahead, as population aging in emerging Asia accelerates, this suggests that consumption-to-GDP ratios will rise. Indeed, in Korea, Singapore, Hong Kong SAR, and Thailand, the elderly dependency ratio is projected to rise between 3-6 percentage points over the next decade. That alone could push consumption up by 10-18 percentage points of GDP.
Emerging Asia: Projected Impact of Aging on Consumption1
Excluding Taiwan POC.
Emerging Asia: Projected Impact of Aging on Consumption1
| Elderly dependency ratio 2005 | Elderly dependency ratio 2015 | Projected rise in consumption (in percent) | |
|---|---|---|---|
| China | 10.7 | 13.3 | 7.8 |
| Hong Kong SAR | 16.3 | 19.8 | 10.6 |
| India | 8.4 | 9.4 | 3.0 |
| Indonesia | 8.3 | 9.4 | 3.3 |
| Korea | 13.1 | 18.1 | 15.0 |
| Malaysia | 7.3 | 9.1 | 5.3 |
| Philippines | 6.4 | 7.5 | 3.4 |
| Singapore | 11.8 | 18.0 | 18.6 |
| Thailand | 10.2 | 13.4 | 9.4 |
Excluding Taiwan POC.
Emerging Asia: Projected Impact of Aging on Consumption1
| Elderly dependency ratio 2005 | Elderly dependency ratio 2015 | Projected rise in consumption (in percent) | |
|---|---|---|---|
| China | 10.7 | 13.3 | 7.8 |
| Hong Kong SAR | 16.3 | 19.8 | 10.6 |
| India | 8.4 | 9.4 | 3.0 |
| Indonesia | 8.3 | 9.4 | 3.3 |
| Korea | 13.1 | 18.1 | 15.0 |
| Malaysia | 7.3 | 9.1 | 5.3 |
| Philippines | 6.4 | 7.5 | 3.4 |
| Singapore | 11.8 | 18.0 | 18.6 |
| Thailand | 10.2 | 13.4 | 9.4 |
Excluding Taiwan POC.
With the exception of China, there is little evidence that private consumption in emerging Asia is “too low”, based on the historical association between consumption and fundamentals. There is some evidence that precautionary saving in some countries (Korea, Malaysia, Singapore, Thailand) rose around the time of the Asian crisis, as households scaled back consumption in the face of greater uncertainty, but this rise in precautionary behavior has been unwound since then. Only in China are there strong indications that precautionary behavior increased permanently in the wake of the Asian crisis.
Government can play a supportive role in raising consumption over the medium term. Short-term fiscal policy aimed to give a boost to consumption may have limited effectiveness, since there is such a large offset between private and public consumption. However, there are important structural reforms that countries in the region can take to help ensure a more permanent increase in consumption and help rebalance GDP growth away from external demand. While the best approach will vary from country to country, the following considerations may be relevant:
The apparent rise in precautionary saving during the Asian crisis suggests that governments in the region may need to do more to reduce uncertainty faced by households, including through education, health care, and pensions. Such reforms would likely entail higher public spending, but may not entail an offset in private consumption because the provision of an enhanced social safety net would reduce the precautionary savings motive.
Although financial markets are relatively deep, especially compared to other emerging market regions, more can be done do develop household credit instruments that would better allow intertemporal smoothing of consumption. This in turn would have the potential to reduce precautionary saving, as households would be less likely to save for a “rainy day.” Ongoing banking sector reforms in the region, including the implementation of improved risk management systems, should also help banks better cater to the borrowing needs of small- and medium-sized enterprises, again providing a channel for household savings rates to decline over time.
These reforms could be enhanced by further developing capital markets. Developing more liquid equity and bond markets with greater household participation would allow households to better diversify risk away from purely wage income-based risk, and encouraging international diversification among households would further strengthen this trend. In addition, greater financial market participation among households would allow them to share in rising profits and market valuations of companies, permitting greater wealth effects on consumption.
Finally, greater exchange rate flexibility would likely imply an appreciation of some regional currencies, which would boost households’ purchasing power and likely raise the consumption-to-GDP ratio.
References
Arellano, Manuel and Bond, Stephen, 1991, “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations.” Review of Economic Studies, Vol. 58, No. 2: 277–97 (April).
Bernanke, Ben, 2005. “The Global Saving Glut and the U.S. Current Account Deficit,” remarks at the Sandridge Lecture, Virginia Association of Economics, Richmond, Virginia.
Briotti, M. Gabriella, October 2005, “Economic Reactions to Public Finance Consolidation: a Survey of the Literature,” European Central Bank Occasional Paper No. 38.
Eichengreen, Barry, 2005, “The Blind Men and the Elephant.” Mimeo available at http://www.econ.berkeley.edu/~eichengr/policy/KyotoPaper.pdf. (November).
Kuralbayeva, Karlygash, and Papa N’Diaye, 2006, “Determinants of Private Consumption in a Selected Number of Asian Countries,” IMF Working Paper (forthcoming) (Washington: International Monetary Fund).
International Monetary Fund, 2006, “The Consumer Finance Boom: Is it a Problem?” Asia and Pacific Regional Economic Outlook, May (Washington).
International Monetary Fund, 2005, “Global Imbalances: A Saving and Investment Perspective,” World Economic Outlook, September (Washington).
