III. Corporate Savings and Rebalancing in Asia

International Monetary Fund. Asia and Pacific Dept
Published Date:
April 2009
  • ShareShare
Show Summary Details

In recent years, despite relatively flat investment in most of the region, corporate savings have risen in Asia, more than in other regions. Since Asian households have not reduced their savings to the same extent, the increase in corporate savings has led to higher private savings in the region. Decisions over retained earnings are related to the strength of corporate governance and the degree of financial market development, particularly in Asia. Hence, reforms in these areas must be further deepened to reduce private savings, contribute to higher private consumption, and help regional rebalancing in the coming years.

In recent years the corporate sector in much of Asia has begun to save an increasing share of national income. This has been true in the regional powerhouses of Japan, Korea, China, and India, but also in the ASEAN region and among Asia’s newly industrialized economies (NIEs) (Figure 3.1).

Figure 3.1.Change in Savings since 2000

(In percent of GDP)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

Although high profits and rising corporate savings have been a global trend through much of this decade (Cardarelli and Ueda, 2006), the increase in Asian corporate savings has outstripped that in all other regions of the world (Figure 3.2). Further, in other regions, this large increase in corporate savings has been largely offset by lower household savings (Figure 3.3). This has not been true of emerging Asia (Figure 3.4). As a result, higher corporate savings in the region have driven up national savings. Corporate savings have come to account for not only an increasing share of the region’s GDP, but also a rising fraction of global savings as a whole (Figure 3.5).

Figure 3.2.Difference in Mean Corporate Savings, Post-2000 versus Pre-2000

(In percent of GDP; regression coefficients and confidence intervals)

Source: IMF staff estimates.

Figure 3.3.Non-Asian Emerging Markets: Savings

(In percent of GDP, relative to 1997 = zero)

Sources: United Nations; and IMF staff calculations.

Figure 3.4.Emerging Asia (excluding China): Savings

(In percent of GDP, relative to 1997 = zero)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

Figure 3.5.Emerging Asia Corporate Savings

(Share in world savings, in percent)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

This rapid rise in Asian corporate savings has been particularly surprising since it occurred during a period of relatively stagnant levels of investment in the region. The decline in investment following the Asian crisis is well known and has been extensively studied (September 2005 World Economic Outlook; May 2006 Asia and Pacific Regional Economic Outlook) and resulted from the unwinding of the excessive levels of investment during the 1990s fueled by rapidly rising asset prices, loose credit, and weak corporate governance. Since then, outside of China and India, investment in Asia has risen broadly in line with output. Despite this flattening out of investment, in the same set of countries, corporate savings had increased by more than 5 percent of GDP by 2005, and this can be explained only partially by the region-wide deleveraging that followed the Asian crisis (Figure 3.6). These three phenomena—rising corporate savings, stagnant corporate investment, and a lack of offsetting behavior by households—have led to a widening of the region’s current account surplus, directly contributing to the global imbalances of the past decade. This chapter investigates these interlinking themes and examines whether a policy response is warranted to limit the incentive for Asian firms to save so much.

Figure 3.6.Leverage in Emerging Asia

(Debt-to-assets ratio, in percent)

Sources: Worldscope; and IMF, Corporate Vulnerabililty Utility database.

The reasons for the rise in corporate savings in Asia, as well as elsewhere, have been extensively studied (Cardarelli and Ueda, 2006) and Asian Development Bank, 2009). For example, in G-7 countries recent corporate savings have been driven by a desire to finance share buybacks at home and net purchases of equity overseas and to acquire liquid financial assets (Cardarelli and Ueda, 2006). In Asia, the rise in gross corporate savings has been variously linked in country-specific cases to energy and land subsidies, cheap credit, low dividend payout ratios, and robust economic growth (Asian Development Bank, 2009). Given this previous work, this chapter focuses on two other questions that directly relate to the policy discussion: (i) Do households in Asia offset corporate savings decisions? and (ii) What impact do institutional and policy factors have on corporate savings?

The structure of this chapter is as follows. The second section examines how the household reaction to corporate savings has differed between Asia and elsewhere. The analysis suggests that, unlike elsewhere, higher corporate savings in Asia are not offset by an increase in household consumption (see Box 3.1 for a detailed analysis of Japan’s experience). As a result, corporate behavior has had a potent effect on the region’s savings. Consequently, for a given level of investment, a decline in corporate savings should reduce regional imbalances and boost consumption.

The third section traces the role played by corporate governance and financial sector institutions in the accumulation of corporate savings. The evidence indicates that decisions on the level of retained earnings are intimately related to the strength of corporate governance and to the degree of financial market development and that the size of these effects is particularly large in Asia. In countries where market-financing options are plentiful and firms are able to raise capital readily from banks or capital markets, the motivation for retaining high levels of corporate earnings is lessened. Similarly, if corporate managers’ incentives are more fully aligned with those of shareholders, then this should lead corporations to pay out as dividends any profits that are not required to finance investment projects.

The fourth section draws out the policy implications of these results, finding that since the Asian crisis, improvements in corporate governance and financial development have mitigated the need for corporations to amass a large war chest of savings. However, these effects have been more than offset by the rising tide of corporate savings. Nevertheless, further improvements in corporate governance and financial reform will still need to be a critical component of the region’s rebalancing strategy in order to address the underlying structural factors that create incentives for the region’s corporations to save. In particular, the analysis finds that, for example, if Asia can converge to the average level of financial development and corporate governance that is currently seen in the G-7, then corporate savings could be lowered by as much as 7 percent of GDP.

The chapter’s final section provides a brief conclusion.

Does Higher Corporate Saving Affect Household Behavior?

Economic theory suggests that, absent tax distortions, and assuming that households have relatively easy access to well-working financial markets, the level of national savings should be relatively unrelated to the savings of corporations. As households ultimately own the corporations, they should be indifferent between holding their savings directly or indirectly via the savings of the firms that they own. However, to be indifferent, households must be able to smooth the impact of changes in corporate savings (the timing of corporate dividend payments), either by borrowing using their financial assets as collateral or by liquidating some part of those assets.

But there are a number of reasons why this neutrality result may not hold. For example, if dividends are retained in order to finance low-rate-of-return projects, then households may not realize any increase in their wealth from higher corporate savings and therefore may not adjust their consumption and savings decisions. Similarly, if corporations are government owned and pay dividends to the state, but these revenues are then deployed inefficiently with limited direct benefit to households, neutrality will again break down. Therefore, key to this neutrality proposition is the existence of an active corporate ownership culture and strong market discipline on corporate managers so that their incentives are aligned with those of shareholders.

This section examines empirically how households react to corporate savings by looking at corporate and household behavior across a range of emerging markets, using a methodology developed by Poterba (1987). In particular, a panel of emerging market economies from 1979 to 2007 is used to test whether a change in the level of corporate savings has any impact on private (that is, household plus corporate) savings. If households are indifferent to the timing of dividends and the level of retained earnings, it should be the case that corporate savings will have no explanatory power in the regression. Various factors such as age dependency (to capture life cycle savings motives), the level of economic volatility (to capture precautionary motives), the real interest rate (to capture the financial incentive to save), the level of disposable income, and lagged private savings (to proxy for accumulated wealth) are also controlled for. The results indicate that, in emerging markets as a whole, private savings are not significantly affected by the level of corporate savings (Figures 3.7 and 3.8). However, this result does not hold in an Asian subsample. Indeed, Asian households offset only about 20 percent of the increase in corporate savings. In other words, for every additional US$100 saved by Asian firms, households in Asia reduce their savings only by US$20 (Figure 3.9). It is notable that these results are not driven by the strong trends for rising savings that have been seen in China. Indeed, dropping China from the sample leads to the result that only 7 percent of higher corporate savings are offset by lower household savings. These results are very robust to changes in sample period and the use of alternate measures of private savings (see Appendix 3.1 for detailed results).

Figure 3.7.Private and Corporate Savings

(In percent of GDP)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

Figure 3.8.Household and Corporate Savings

(In percent of GDP)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

Figure 3.9.Corporate Savings Offset by Households

(In percent of corporate savings)

Source: IMF staff estimate.

Box 3.1.Corporate Savings, Household Income, and Private Spending in Japan

Despite a robust recovery of the Japanese economy and the associated pickup in corporate saving during 2003-07, household spending remained muted. Private consumption increased by only 5 percent over the period, compared with a roughly 10 percent increase in GDP. This box reviews Japanese household income and balance sheets with the aim of identifying the main drags on private consumption.

Low Investment Income

One reason for the sluggish spending may have been weak household investment income despite record corporate profits.1

  • Household holdings of risky assets, including equities, have historically been smaller in Japan than in other advanced economies. Despite financial deregulation, demand for risky assets has remained low, possibly due to relatively high transactions costs (especially trust fund fees) and stronger risk aversion.2

  • In addition, Japanese corporations have continued to pay out only a small share of their net profits, despite strong earnings (especially in the manufacturing sector) and record low interest costs. This may have reflected factors such as shareholder differences and the preference of Japanese corporations to reduce their debt to more manageable levels.

Financial Wealth Held by Households1

(In percent of total financial wealth)

Source: Bank of Japan.

1 As of end March 2009.

That said, investors and households have made some progress in piercing the corporate veil, with the dividend-to-GDP ratio jumping from 1 percent to 1½ percent during the 1980s and 1990s to about 3½ percent in 2007. However, dividend payments still remain well below German levels (about 14 percent in 2007) despite a similar level of operating profits. Japanese corporations have also not engaged much in equity buyouts, which have become a popular means of distributing profits to shareholders among Anglo-Saxon corporations.

Nonfinancial Corporations: Dividend Payout

(In percent of net profit)

Source: OECD, Annual National Accounts.

Falling Share of Labor Income

In addition, household spending has been held down by sluggish wage growth. Low wage growth can be attributed to a variety of factors, many of which also apply to other advanced economies: these include effects of foreign competition and technological progress (manufacturers have become more sensitive to cross-country wage differentials and restricted wage growth in their home countries), deregulation (liberalization measures in the second half of the 1990s expanded the list of industries that can hire low-wage “nonregular” workers),3 and population aging. Low wage growth has also partly reflected the unbalanced nature of Japanese growth—low productivity growth in Japan’s nonfinancial services sectors, accounting for about two-thirds of total employment, has put a drag on total wages (Sommer, 2009).

Stagnating Net Worth and Financial Constraints

In contrast with many other G-7 economies, household asset values and net worth have stagnated in Japan, as falling house prices offset equity price increases. Unlike in some other major economies, Japanese households could not therefore engage in a debt-financed spending supported by rising equity and housing wealth.

In this regard, downside risks to Japanese household spending from the ongoing adjustment in global asset prices seem limited. However, it remains to be seen whether Japanese consumption will respond meaningfully to rising asset values once the global recovery is firmly in place given the low marginal propensity to consume out of wealth (IMF, 2008) and limited availability of home equity loans (including reverse mortgages).4 Furthermore, despite recent improvements, consumer lending remains constrained, in part due to limited sharing of credit information.

Household Wealth

(In percent of disposable income)

Source: IMF, World Economic Outlook, April 2009.

How to Revive Consumer Spending?

Household spending accounts for about 55 percent of GDP in Japan, compared with almost 60 percent among the other G-7 economies, suggesting that the scope for boosting private consumption could be significant. Policy measures could help in several areas:

  • Labor and product market reforms. Adopting reforms to increase productivity of services sectors would help reduce Japan’s economic dualities and boost per capita incomes. After the recovery takes hold, greater labor flexibility could encourage firms to pay higher wages, as the implicit insurance premium incorporated in the contracts of permanent workers would diminish.

  • Diversifying household portfolios. Further development of the capital markets would help reduce transaction costs of investing, while measures to promote understanding of investments among the general population would make stockholding more prevalent.

  • Financial deepening. Financial reforms to further enhance access of households and corporations to credit could facilitate spending out of wealth and future income, and aid the emergence of new high-productivity enterprises. Encouraging the use of reverse mortgages would ease liquidity constraints faced by the elderly.

  • Removing uncertainty. Strengthening the public pension system could boost consumption by reducing precautionary savings (IMF, 2009b).

Note: The main authors of this box are Martin Sommer, Suchanan Tambunlertchai, and Kiichi Tokuoka.1 Gross operating surplus of Japanese nonfinancial corporations increased from about 21 percent of GDP in the early 2000s to 23 percent of GDP in 2007—this surplus level is higher than in the United States and about the same as in Germany. For comparison, property income accounted for 3.8 percent of disposable income in Japan, 19.2 percent in the United States and 23.7 percent in Germany in 2007.2 The structural increase in unemployment and weakening of the lifetime employment model (one-third of the labor force now has flexible contracts) may have discouraged some households from investing in risky assets (Nakagawa and Shimizu, 2000).3 Nonregular workers are much easier to dismiss, have limited social insurance, and typically earn about 40 percent less than regular workers.4 Tight collateral requirements combined with high real estate prices may have encouraged Japanese households to accumulate large holdings of cash and deposits, further reducing investment income. See, for example, Iwaisako (2003) and Iwaisako, Mitchell, and Piggott (2004).

To examine the neutrality hypothesis from another angle, the question is asked, following Auerbach and Hassett (1991): to what extent is private consumption dependent on the level of dividend income paid to households? If households are not credit constrained, consumption decisions should be independent of the flow of dividend income they receive. However, numerous studies have found that, even in advanced economies, liquidity constraints are generally binding (for example, Flavin, 1981; Carroll, 2001). Auerbach and Hassett (1991), however, argue that although it is true that many households may be liquidity constrained, it is unlikely that households with significant levels of dividend income will be subject to such constraints. If these households are not credit constrained, then their consumption decisions should be independent of the level of dividends they receive and their consumption should depend only on wealth, not on their current income (Figure 3.10).

Figure 3.10.Households: Consumption and Disposable Property Income

(Change in percent)

Sources: United Nations; CEIC Data Company Ltd.; and IMF staff calculations.

This same test is undertaken for a sample of emerging markets, to examine whether a broad measure of property income—which encompasses interest, rent, and dividends—has a significant effect on household behavior. The results indicate that households, in general, are not indifferent to their levels of property income. However, the data also suggest that the marginal propensity to consume from property income is significantly higher in emerging Asia than elsewhere, even though liquidity constraints seem equally binding. In particular, Asian households appear to consume about US$7 more for every additional US$100 in property income they receive (compared with only US$1 for non-Asian emerging markets) (Figure 3.11). In sum, Asian households do not appear to offset the behavior of firms in making their consumption decisions. Therefore, the high levels of corporate savings in Asia are an important factor in regard to the size of national savings and have a clear role in determining the level of current account imbalances in Asia. This result is specific to Asia and also appears not to be driven solely by the savings trends in China or India (see Box 3.2).

Figure 3.11.Marginal Propensities to Consume

(Share of current real disposable income)

Source: IMF staff estimates.

Box 3.2.Savings in China

China’s saving rate has risen rapidly in recent years, and it is among the highest in the world. National saving is currently more than 55 percent of GDP and has increased more than 12 percentage points over the past five years. The high savings rate reflects high levels of savings by firms (including public companies), households, and to some extent the government. By 2005, both corporate and household saving rates were greater than 20 percent of GDP and while corporate savings slipped slightly lower in 2006 and 2007, both corporate and household savings remain very high.

China: Domestic Savings

(In percent of GDP)

Sources: CEIC Data Company Ltd.; and IMF staff estimates.

The drivers of rising household savings are relatively well known. Chinese households face considerable expenditure risks. In addition, their consumption opportunities have been stymied by the falling share of household disposable income in GDP (Aziz and Li, 2007). Following the reform of state-owned enterprises (SOEs) in the 1990s, the industry-based social safety net disappeared, limiting the support available to many individuals for their health care, their pensions, and education for their children. Consequently, households had to increase savings significantly to cover their retirement and the risks associated with sickness (Barnett and Brooks, 2009). Household income has also lagged broader economic growth, particularly reflecting (at times) modest wage growth and low investment income, the latter held down by financial repression. Recent government initiatives, however, have sought to strengthen safety nets (reducing the precautionary motives for savings) and boost household disposable income. In particular, the government is expanding health care provision and insurance coverage, increasing pensions, targeted consumption subsidies, and improved funding for rural livelihoods.

China: Industrial Enterprise Profits1

(In percent of GDP)

Sources: CEIC Data Company Ltd.; and IMF staff estimates.

1 Staff estimates for 2008 are based on data up to November 2008.

The rise in corporate savings reflects a combination of rapid growth, limited competition (which has buoyed profits), financial underdevelopment, and low input costs that increase the incentives to invest. Both listed and nonlisted companies have been highly profitable and pay little in dividends. Industrial enterprises (both listed and unlisted) account for more than 40 percent of domestic value added and almost as much in aggregate profits. Until 2008, most of the profits of these enterprises accrued to the state-controlled sector, although the share of profits made by non-state-controlled firms has increased considerably over time.

Selected Equity Markets: Dividend Yields

(In percent per annum; 12-month moving average)

Sources: CEIC Data Company Ltd.; and IMF staff calculations.

Despite high profits, Chinese firms—even those with H shares listed in Hong Kong SAR, which have international investors as owners—pay very low dividends in comparison with firms in both developed and emerging markets.

Misaligned relative prices have produced an extraordinarily high level of investment, which has led to excess capacity in some sectors, and provided a strong motive for firms to save. The incentives include the low costs of energy, water, and land, an underpricing of environmental externalities, a relatively high price for capital-intensive tradables, and a regulated, relatively cheap supply of capital for large formal-sector companies with access to financial markets. Investment growth has been especially rapid in nonstate controlled companies, both private and nonlisted shareholding companies. Non-state-controlled companies, including SMEs, are likely the most credit constrained, at least in the formal financing market, and have to rely heavily on retained earnings (corporate savings) to finance their investment. In addition, the large state-owned enterprises also rely on retained earnings to a large extent.

China: Domestic Enterprise Investment Spending

(In percent of GDP)

Sources: CEIC Data Company Ltd.; and IMF staff calculations.

1 Excludes listed corporates.

The government has introduced an initiative to require state-owned companies to pay dividends that will help to reduce corporate savings. But more action might be needed. In late 2007 the government introduced a pilot program under which state-owned enterprises disburse limited dividends (at rates of 5–10 percent of profits) to the government. However, as these funds are paid into a special capital budget that is used to finance state enterprise investment, they do little to lower corporate savings or limit the availability of cheap funds for investment as they are reinvested into the corporate sector. In addition, this arrangement also prevents these funds from being reallocated to other purposes with higher social value (through the budget) or higher economic value (through financial markets).

China: Fixed-Asset Investment Financing

(In percent of GDP)

Sources: CEIC Data Company Ltd.; and IMF staff calculations.

Expanding the access of all firms to formal capital market financing would seem a key issue, as it would limit the need for retained earnings to finance investment, particularly among private firms and SMEs. Such access would include greater equity and fixed-income financing for SMEs, both of which the government has been working to address. A new growth enterprise exchange for small firms was launched in May 2009 and will become operational in October of this year, and commercial paper and medium-term notes markets were established in 2005 and 2008 respectively and are in active use. Finally, better corporate governance, including through stronger shareholder activism and the development of strong institutional investors, would likely lessen the incentives to retain earnings.

Note: The author of this box is Nathan Porter.

What Influences Corporate Savings in Asia?

This section supplements country-level analysis with a rich, firm-level data set (see Appendix 3.2) to examine what factors are relevant for individual-firm savings decisions and how these factors vary across countries. The firm-level data allow fundamental firm-level drivers of corporate savings such as profitability, Tobin’s Q, the share of intangible assets, firm size, sales volatility, leverage, and capital expenditure to be controlled for. These data are used to examine to what extent financial market development or changes in corporate governance have an impact on corporate saving behavior, once these fundamental factors are controlled for. For financial sector reform, an index constructed by Abiad, Detragiache, and Tressel (2008) is used that incorporates various dimensions of financial development, including banking supervision, privatization, barriers to entry, level of directed credit, credit ceilings, interest rate controls, and securities market reform (Figure 3.12). A corporate governance data set (outlined in De Nicolo, Laeven, and Ueda, 2006) that looks at earnings opacity, the comovement of stock prices, and the number of accounting regulations complied with in a particular country is also drawn upon (Figure 3.13).

Figure 3.12.Financial Reform, 2005


Figure 3.13.Corporate Governance Quality, 2008


Overall effects of financial reform and improvements in corporate governance on corporate savings may be ambiguous. In a liberalized financial market, firms might have easier access to finance and so need to save less. But at the same time, once close ties between firms and banks are severed, firms may no longer have access to relatively automatic credit lines, which may induce them to save more. Similarly, improvements in corporate governance, specifically transparency, may improve access to external funding and reduce the need to rely on internally generated funds. But they could also mean that firms can no longer rely on implicit intragroup cross-subsidies, and so corporate savings increase. Moreover, anecdotal evidence suggests that immediately after the Asian crisis, for example, corporate governance reforms raised fears of hostile takeovers, inducing firms to hoard cash to buy back shares. Therefore, as with financial sector reform, the net effect of corporate governance reform on corporate savings could be either positive or negative.

The firm-level data suggest that financial reform has had a significant impact by tempering the rise in corporate savings across all emerging markets and that this effect has been considerably larger in Asia (Figure 3.14). Similarly, improvements in the transparency of corporate operations also appear to have lessened the motivation for higher corporate savings (Figure 3.15).

Figure 3.14.The Effect of Financial Liberalization on Corporate Savings

(Regression coefficients)

Source: IMF staff estimates.

Figure 3.15.The Effect of Corporate Governance Reform on Corporate Savings

(Regression coefficients)

Source: IMF staff estimates.

An attempt to confirm these results is made using the World Bank’s Doing Business database, drawing on measures of shareholder protection (which capture the strength of protection for minority shareholders). Using these data, the analysis confirms the finding that improvements in corporate governance (through stronger shareholder rights) lead to a better alignment of the incentives of shareholders and managers, thereby moderating the tendency for corporations to save.

The Policy Implications

Following the 1997–98 financial crisis, Asia was a fertile ground for financial market and corporate governance reform. The corporate governance measures and financial reform indicators used in the previous section all have improved significantly (Figures 3.16 and 3.17). As the financial system in the region has become more market driven, the options for firms to finance investment projects (including equity issuance, corporate bonds, and overseas financing) have grown. This, in turn, has limited firms’ incentives to retain earnings as they become more secure that market financing will be available for high-return projects. Consequently, firms have needed to rely less on self-financing. Similarly, improvements in corporate governance over the past decade have increased the discipline on corporations and lessened their tendency to hoard cash. This progress has had a significant effect in tempering the rise in regional corporate savings.

Figure 3.16.Financial Reform since the Asian Crisis


Figure 3.17.Measure of Investor Protection in Asia


Source: World Bank, Doing Business database.

Nevertheless, despite the impressive and steady progress in reforming the financial system and improving corporate governance, emerging Asia still lags behind competitors along a range of indicators. Certainly, the difference has narrowed over the past decade, but there is still much to do. Further corporate governance and financial reform not only would improve the allocation of capital and raise productivity but would also have a significant impact on corporate savings. Such reform could be an instrumental factor in reducing current account surpluses across the region, given the stagnant investment levels in most of Asia over the last decade. For example, if Asia were able to reach the average level of financial liberalization in advanced economies, it would be able to lessen corporate savings by as much as 5 percent of GDP. Similarly, improvement in corporate governance to bring Asia to the average levels prevailing in advanced economies could reduce corporate savings by 2.4 percent of GDP. Taken together, these results suggest that reforming corporate incentives and financial market institutions should be an integral part of regional rebalancing in the coming years.


Savings by Asian firms seem to have been a major contributor to the increase in the gap between savings and investment seen in Asia during recent years. Although high corporate profits and savings have been seen across the globe, the rise in Asia has been particularly striking, especially given anemic investment demand. Moreover, it has had a marked effect on national savings because, unlike those elsewhere, Asian households have not offset the rise in corporate savings with higher consumption. It is nevertheless true that the rise in savings would have been considerably larger if the region had not made the important advances it did over the past decade in both financial market development and corporate governance. It is important that these reforms be further deepened, and doing so will help contribute to declining savings, rising consumption, and regional rebalancing in the coming years. With the marginal propensity from property income higher in Asia, a rise in property income that follows reform is likely to lead to a larger boost in private consumption and to assist in rebalancing.

Appendix 3.1: Corporate and Household Savings

Two tests of the interaction of households and corporate savings are considered:

  • In the first, suggested by Poterba (1987), a savings function is estimated based on some fundamental drivers of savings (S) to test whether changes in corporate savings (CS) are offset by households. In particular, the following equation is estimated:


    where lagged savings proxies wealth. Generally private savings should be driven by life cycle factors (such as aging) and precautionary motives. In light of this, the age dependency ratio, household disposable income, the (ex post) real interest rate, and a measure of aggregate volatility are used as fundamentals (X); disposable income is also included. Although the relationship may be biased by the holdings of equities outside the domestic market, this is unlikely to be a problem, given the extent of home bias seen in equity holdings (well over 90 percent held domestically in emerging markets). To ensure robustness, the relationship is estimated using measures of both private savings and household savings. If the corporate veil is pierced, then χ should not be significantly different from 0 when a private savings equation is estimated and not significantly different from –1 when a household savings relationship is estimated. The relationship is estimated with generalized method of moments (GMM), using as instruments twice-lagged savings, a United Nations (UN) measure of the change in household financial assets, and the lagged value of the fundamentals, to deal with possible endogeneity issues.

  • The second set of tests, proposed by Auerbach and Hassett (1991), is based on determining whether consumption deviates from that implied by optimizing consumer theory, even after liquidity-constrained consumers are controlled for (by controlling for current disposable income). Specifically the relationship between real consumption growth, interest rates, and both current disposable property and nonproperty income is estimated. Deviations based on consumption from current nonproperty income indicate the extent of liquidity constraints faced by households. However, because individuals who receive dividends (or other property income) are arguably not credit constrained, then if consumption decisions depend significantly on dividend income, these households are arguably unable to realize the same wealth out of retained earnings (corporate savings) as out of disbursed earnings (dividends). In other words, instead of saving the expected part of dividend payments (adding it to their wealth to increase lifetime consumption), they consume a significant share of it today. The following equation is therefore estimated:


    where YNPDI is nonproperty disposable income, and YPID disposable property income. If consumers are credit constrained, then γY > 0, and if γPI > 0 then households savings behavior is unlikely to offset corporate savings behavior. We estimate this equation using GMM and instrument for twice- and thrice-lagged consumption growth, real interest rates, and income variables.

Empirical Results

Table 3A.1 shows the relationship between private savings (constructed as the sum of corporate and household savings using UN data), fundamentals, and corporate savings.1 This is the analogous relationship to the baseline relationship estimated for the United States in Poterba (1987). In the pooled sample, corporate savings have no significant impact on total private savings—consistent with households offsetting changes in corporate savings—thereby offsetting the impact of higher corporate savings with lower households savings. However, if the relationship for corporate savings in emerging Asia and elsewhere is estimated separately, it is found that private savings in Asia rise significantly with Asian corporate savings. Specifically, in emerging Asia, households offset (through lower savings) an additional dollar in corporate savings by only less than 20 cents, whereas elsewhere they offset almost three-quarters of the rise in corporate savings. The results are not an artifact of China’s high savings and are robust even when Asian savings are broken down between China and the rest of emerging Asia. These results are also robust across measures of savings. Table 3A.2 shows the results of the test on corporate savings when various measures of private and household savings are used. The only substantial difference is that when the test is constructed using household savings directly, it seems non-Asian households also fail to breach the corporate veil.

Table 3A.1.Private and Corporate Savings(Regression results)
First lag of private savings/GDP0.44**0.47***0.48***
Age dependency-0.19-0.09-0.09
Household disposable income/GDP0.310.340.33*
Real interest rate-0.03-0.05-0.05
Corporate savings/GDP
Asia (excluding China)0.93+
Adjusted R-squared0.950.940.94
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. + indicates significance at 10.6 percent level.

Standard deviation of Hodrick-Prescott-filtered GDP cycles.

Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. + indicates significance at 10.6 percent level.

Standard deviation of Hodrick-Prescott-filtered GDP cycles.

Table 3A.2.Impact of Corporate Savings on Private Savings by Region: Sensitivity to Measures of Private Savings(Regression results)
Corporate savings/GDPPrivate savings1Private savings2Household

Asia (excluding China)0.68**0.96***0.89*
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively.

Private savings as measured in IMF WEO database.

Private savings derived using national savings less government savings. Data obtained from IMF WEO database and the United Nations.

Household savings as measured by the United Nations. The null hypothesis (pierce the corporate veil) is that the corporate savings coefficient is -1. The significance test is adjusted accordingly.

Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively.

Private savings as measured in IMF WEO database.

Private savings derived using national savings less government savings. Data obtained from IMF WEO database and the United Nations.

Household savings as measured by the United Nations. The null hypothesis (pierce the corporate veil) is that the corporate savings coefficient is -1. The significance test is adjusted accordingly.

Following Auerbach and Hassett (1991), Equation (2) is estimated using real disposable nonproperty disposable income and real disposable property income to test whether current consumption depends on property income (see Table 3A.3). The results suggest that households in all regions do not fully offset corporate savings. However, although all households have a similar propensity to consume out of nonproperty income (suggesting equally binding liquidity constraints), Asian households have a significantly stronger propensity to consume directly out of property income than households in other regions. This suggests that Asian households are much less likely to offset corporate savings decisions than those elsewhere.

Table 3A.3.Private Consumption and Household Property and Nonproperty Income(Regression results)
Real interest rate (T-bill)-0.103***-0.067***
Disposable income growth
Nonproperty (Pooled)0.989***
Property (Pooled)0.033**
Nonproperty (Non-Asia)0.729***
Property (Non-Asia)0.013**
Nonproperty (Asia)0.740***
Property (Asia)0.068***
Equality of disposable income coefficients (Asia=Non-Asia) 1
Nonproperty income growth0.018
Property income growth24.372***
R -squared0.2450.376
Adjusted R -squared0.2340.359
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively.

Wald test for the null hypothesis that the disposable income coefficients in Asia and that outside Asia are equal.

Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively.

Wald test for the null hypothesis that the disposable income coefficients in Asia and that outside Asia are equal.

Appendix 3.2: Regressions on the Determinants of Corporate Savings

This chapter relies on a series of regressions using macro-level and firm-level data to explain the impact of changes in the policy environment on corporate savings.

Macro-Level Regressions

At the aggregate level, the chapter uses a crosscountry panel data set compiled from various sources: the United Nations Statistical Database, CEIC Data Company Limited, World Development Indicators, the Lane and Milesi-Ferretti (2006) database on financial openness, and two new data sets that construct indices for financial development (Abiad, Detragiache, and Tressel, 2008) and corporate governance quality (De Nicolo, Laeven, and Ueda, 2006). The covariates considered are manufacturing value-added share of GDP, growth rate of GDP, volatility of GDP growth, the trade/GDP ratio, terms of trade volatility, financial openness, and domestic credit as a share of GDP. Fixed-effects regressions are used to control for time-invariant country-specific factors.

Changes in the mean corporate savings/GDP ratio since 2000 relative to prior years are tested for. The results, presented in Table 3A.4, indicate that corporate savings have increased in general across the world, with the increases ranging from 3.06 percent of GDP in Asia to 1.17 percent of GDP in non-Asian emerging economies (the change is not, however, statistically significant for this latter group). The point estimates indicate that the magnitude of the increase is appreciably larger in Asia than it is outside Asia and that this result is not being driven entirely by China and India.

Table 3A.4.Difference in Mean Corporate Savings, Post-2000 versus Pre-2000(Regression results)


China and



R -squared0.
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. Robust standard errors are in parentheses.
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. Robust standard errors are in parentheses.

To study the policy levers that influence corporate savings at the macro level, the earnings-smoothing indicator of corporate governance constructed by De Nicolo, Laeven, and Ueda (2006) is used. This indicator tracks the degree to which managers are able to hide the true performance of firms by using accruals to smooth fluctuations of annual profits. It is calculated as the rank correlation between cash flows (before any accounting adjustments) and profits (after accounting adjustments) in a cross-section of firms within a country at each point in time.

Table 3A.5 reports the results from a country fixed-effects regression (which controls for time-invariant country-specific factors that influence corporate savings). As the coefficient on the interaction term for Asia post-2000 and corporate governance quality indicates, the association between corporate savings and corporate governance is significantly more negative than in preceding years and when compared with that for the rest of the world. This is consistent with improvements in corporate governance in Asia having an offsetting negative effect on corporate savings (against the backdrop of an overall rising trend during this period).

Table 3A.5.Corporate Savings and Corporate Governance(Fixed-effect regression results)

Asia post-20004.24

Corporate governance quality0.81

Post-2000 x corporate governance quality-0.24

Asia x corporate governance quality-1.52

Asia post-2000 x corporate governance quality-1.70


Number of countries36
R -squared0.23
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. Standard errors are in parentheses.
Source: IMF staff estimates.Note: ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. Standard errors are in parentheses.

Firm-Level Regressions

The firm-level perspective controls for firm and industry characteristics in regard to overall corporate liquidity. The data are from the Worldscope database and are based on corporate balance sheets and income statements. The sample includes approximately 20,000 firms from 60 countries. It excludes financial firms, as they operate differently and hold cash and other liquid assets to meet statutory capital requirements.

The chapter analyzes the impact of financial liberalization on corporate savings, conditional on a number of firm-level determinants of corporate savings. The measure of financial liberalization used is an index of financial reform constructed by Abiad, Detragiache, and Tressel (2008), which looks at financial market development along seven dimensions (banking supervision, privatization, entry barriers, directed credit, credit ceilings, interest rate controls, and securities market reform) and then takes an average across the seven indicators to compile the index.

The finding is that the lack of financial development has led to an external finance premium forcing Asian firms to save more. The results suggest that financial liberalization does indeed lower the need for precautionary savings and tends to moderate the increase in corporate savings. The coefficient on the overall index is negative and significant, more so for Asia than for outside Asia.

The chapter uses the retained earnings—to—sales ratio as the measure of gross corporate savings. Retained earnings are defined as the accumulated after-tax earnings of a firm that have not been distributed as dividends to shareholders. Following the literature on corporate liquidity (Opler and others, 1999; Cardarelli and Ueda, 2006), a number of determinants are considered in the chapter, in a dynamic-panel setting (using Arellano-Bond GMM estimation and robust standard errors). The firm-level control variables include firm size (measured by the logarithm of the book value of assets), share of intangible assets in total assets, Tobin’s Q (or the market-to-book value, defined as the market value of firms assets to book value of firms’ assets), rate of return on assets, volatility of sales (measured by the rolling standard deviation of sales), capital expenditure, and leverage. In addition to the above firm-level variables, the analysis controls for country, industry, and time effects by including dummy variables. Industries are defined at the two-digit Standard Industrial Classification level (Table 3A.6).

Table 3A.6.Results from Firm-Level Regression
(11.37)* * *
Share of intangible assets0.362
(5.40)* * *
Market to book value0.018
(3.16)* * *
Return on assets0.01
(8.91)* * *
Capital expenditure-0.001
Financial reform index-1.181
(4.93)* * *
Asia x financial reform index-3.416
(9.51)* * *
Number of observations105897
Source: IMF staff estimates.Note: *** indicates significance at 1 percent level. Robust z -statistics are in parentheses.
Source: IMF staff estimates.Note: *** indicates significance at 1 percent level. Robust z -statistics are in parentheses.

The effect of corporate governance reforms on corporate savings is also analyzed using two databases: that from De Nicolo, Laeven, and Ueda (2006) (discussed above) and the World Bank’s Doing Business database. The latter database gauges the strength of minority shareholder protections against directors’ misuse of corporate assets for personal gain. The indicators distinguish three dimensions of investor protection: transparency of transactions (extent of disclosure index), the extent of director liability, and shareholders’ ability to sue officers and directors for misconduct (ease of shareholder suits index).


    AbiadAbdulEnricaDetragiache and ThierryTressel2008“A New Database of Financial Reforms,”IMF Working Paper 08/266 (Washington: International Monetary Fund).

    AlmekindersGerard and Zeno RonaldRuiz Abenoja2009“Remittances in South Asia and the Philippines: Determinants and Outlook,”forthcomingIMF Working Paper (Washington: International Monetary Fund).

    Asian Development Bank2009Asian Development Outlook 2009 (Manila, Philippines).

    AuerbachAlan J. and Kevin A.Hassett1991“Corporate Savings and Shareholder Consumption,”NBER Working Paper No. 2994 (Cambridge, Massachusetts: National Bureau of Economic Research).

    AzizJahangir and LiCui2007“Explaining China’s Low Consumption: The Neglected Role of Household Income,”IMF Working Paper 07/181 (Washington: International Monetary Fund).

    BakerCarol2009“Cambodia: Background Note for the 2009 Article IV Consultation on Garment Export Performance in 2009 and Future Prospects,” (unpublished; Washington: International Monetary Fund, mimeo).

    BalakrishnanRaviStephanDanningerSelimElekdag and IrinaTytell2009“How Linkages Fuel the Fire: The Transmission of Financial Stress from Advanced to Emerging Economies” Chapter 4 in World Economic Outlook (WashingtonInternational Monetary FundOctober).

    BarnettSteve and RayBrooks2009“China: Does Government Health and Education Spending Boost Consumption?” (unpublished; Washington: International Monetary Fund).

    BeckThorsten and AsliDemirgüç-Kunt2006“Small and Medium-Size Enterprises: Access to Finance As a Growth Constraint,”Journal of Banking and FinanceVol. 30 pp. 293143.

    BeckThorstenLeora F.Klapper and JuanCarlos Mendoza2008“The Typology of Partial Credit Guarantee Funds Around the World,”Policy Research Working Paper No. 4771 (Washington: World Bank).

    CardarelliRoberto and KenichiUeda2006“Awash with Cash: Why Are Corporate Savings So High?” in World Economic Outlook April 2006 World Economic and Financial Surveys (Washington: International Monetary Fund).

    CarrollChristopher2001“A Theory of the Consumption Function, with and without Liquidity Constraints,”Journal of Economic PerspectivesVol. 15No. 3 pp. 2345.

    CerraValerie and SwetaSaxena2008“Growth Dynamics: The Myth of Economic Recovery,”American Economic ReviewVol. 98 (December) pp. 43957.

    De NicoloGianniLucLaeven and KenichiUeda2006“Corporate Governance Quality: Trends and Real Effects,”IMF Working Paper 06/293 (Washington: International Monetary Fund).

    DunnJonathan2007The Ready-Made Garment Industryin Bangladesh: An UpdateIMF Staff Country Report No. 07/230 (Washington: International Monetary Fund).

    Espinosa-VegaMarco and AlessandroRebucci2003“Retail Bank Interest Rate Pass-Through: Is Chile Atypical?,”IMF Working Paper 03/112 (Washington: International Monetary Fund).

    EstevãoM.2009U.S. Potential Growth in the Aftermath of the Crisis IMF Staff Country Report No. 09/229 (Washington: International Monetary Fund).

    European Commission2009“Impact of the Current Economic and Financial Crisis on Potential Output” Occasional Paper 49 (Brussels).

    FlavinMarjorie1981“The Adjustment of Consumption to Changing Expectations about Future Income,”Journal of Political EconomyVol. 89No. 5 pp. 9741009.

    FurceriDavide and AnabelleMourougane2009“The Effect of Financial Crises on Potential Output: New Empirical Evidence from OECD Countries,”OECD Economic Department Working Paper No. 699 (Paris: Organization for Economic Cooperation and Development).

    HarvieCharles2004“East Asian Capacity Building, Competitiveness and Market Opportunities in a Global Economy,”SMEs in a Global Economy III Conference—Economic Resilience in East Asia—Role of SMEs and Stakeholders pp. 1741University of Wollongong - Universiti Teknologi MARA MalaysiaKuala Lumpur (July6–7).

    HortonMarkManmohanKumar and PaoloMauro2009 “The State of Public Finances: A Cross-Country Fiscal Monitor,”IMF Staff Position Note SPN/09/21International Monetary FundJuly.

    International Monetary Fund2005World Economic Outlook (WashingtonSeptember).

    International Monetary Fund2006aWorld Economic Outlook (WashingtonSeptember).

    International Monetary Fund2006bRegional Economic Outlook, Asia and Pacific—Asia’s Investment Decline (WashingtonMay).

    International Monetary Fund2008The Changing Housing Cycle and the Implications for Monetary Policy, Chapter 3 in World Economic Outlook (WashingtonApril).

    International Monetary Fund2009aRegional Economic Outlook, Asia and Pacific—Global Crisis: The Asian Context (WashingtonMay).

    International Monetary Fund2009bJapan: 2009 Article IV Consultation—Staff ReportIMF Staff Country Report No. 9/210 (Washington).

    International Monetary Fund2009c“What’s the Damage? Medium-Term Output Dynamics after Financial Crises,”Chapter 4 in World Economic Outlook (WashingtonOctober).

    International Monetary Fund2009dGlobal Financial Stability ReportOctober (Washington).

    IwaisakoTokuo2003“Household Portfolios in Japan,”NBER Working Paper No. 9647 (Cambridge, Massachusetts: National Bureau of Economic Reseach).

    IwaisakoTokuoOliviaMitchell and JohnPiggott2004“Strategic Asset Allocation in Japan: An Empirical Evaluation,”Presented at the 2004 ESRI International Conference2004. Tokyo.

    KangKennethMurtazaSyed and KiichiTokuoka2009“Revisiting Japan’s Lost Decade,”in Regional Economic Outlook—Asia and Pacific (Washington: International Monetary FundMay).

    LanePhillip and Gian MariaMilesi-Ferretti2006“The External Wealth of Nations Mark II: Revised and Extended Estimates of Foreign Assets and Liabilities, 1970–2004,”IMF Working Paper 06/09 (Washington: International Monetary Fund).

    NakagawaShinobu and TomokoShimizu2000“Portfolio Selection of Financial Assets by Japan’s Households—Why Are Japan’s Households Reluctant to Invest in Risky Assets?” (Tokyo: Bank of Japan).

    N’DiayePapaPingZhang and WenlangZhang2009“Structural Reform, Intra-Regional Trade, and Medium-Term Growth Prospects of East Asia and the Pacific—Perspectives from a New Multi-Region Model,”Journal of Asian Economics (forthcoming).

    OplerT L PinkowitzR.Stulz and R.Williamson1999“The Determinants and Implications of Corporate Cash Holdings,”Journal of Financial EconomicsVol. 52.

    Organization for Economic Cooperation and Development2009“Beyond the Crisis: Medium-Term Challenges Relating to Potential Output, Unemployment and Fiscal Positions”Chapter 4 in OECD Economic Outlook 85 (Paris).

    PeltonenTuomasA. and GaborPula2009“Has Emerging Asia Decoupled? An Analysis of Production and Trade Linkages Using the Asian International Input-Output Table,”ECB Working Paper Series No. 993 (Frankfurt: European Central Bank).

    PoterbaJames1987“Tax Policy and Corporate Saving,”Brookings Papers on Economic Activity 2 pp. 455503 (Washington: Brookings Institution).

    PoterbaJames1992“Dividends, Capital Gains, and the Corporate Veil: Evidence from Britain, Canada, and the United States,”NBER Working Paper No. 2975 (Cambridge, Massachusetts: National Bureau of Economic Research).

    PrasadEswar2009“Rebalancing Growth in Asia,” (unpublished; Ithaca, New York: Cornell University).

    ReinhartCarmenM. and Kenneth S.Rogoff2009“The Aftermath of Financial Crises,”American Economic ReviewVol. 99No. 246672.

    ShimIlhyock2006“Corporate Credit Guarantees in Asia,”Quarterly Review Basel: Bank for International SettlementsDecember.

    SommerMartin2009“Why Are Japanese Wages So Sluggish?,”IMF Working Paper 09/97 (Washington: International Monetary Fund).

    TerronesMarco and RobertoCardarelli2005“Global Imbalances: A Saving and Investment Perspective,”in World Economic Outlook (Washington: International Monetary FundSeptember).

    TokuokaKiichi2009The Outlook for Financing Japan’s Public DebtIMF Country Report 09/211 (Washington: International Monetary Fund).

    UedaKenichi2008“Corporate Governance and Liquidity Management” (unpublished; Washington: International Monetary Fund).

Note: The main authors of this chapter are Sonali Jain-Chandra, Malhar Nabar, and Nathan Porter, with assistance from Souvik Gupta. The authors would like to thank Kenichi Ueda for sharing data from Worldscope, as well as the corporate governance dataset used in this chapter.

These relationships are estimated using system GMM with fixed effects and lagged values as instruments to control for potential endogeneity. Significance is determined using robust standard errors.

    Other Resources Citing This Publication