III. How Vulnerable Is Corporate Asia?

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

As Asia has plunged into recession, anxieties about the region’s corporate sector have grown. And not without reason. The collapse of global demand has decimated corporate revenues, forcing many companies to scramble to find financing to tide them over until their earnings revive. But financing has proved hard to find. For eight months from mid-2008 not a single emerging Asian corporate was able to issue an international bond, even as obligations on previous issues continued to mature. Finally, in March 2009, international bond markets opened again, but to only a handful of companies, the highest-rated and best-established ones. Others found they still could not borrow internationally, and many even encountered difficulty borrowing domestically, as banks became reluctant to lend in the face of deteriorating economic prospects.

Consequently, many firms are now in a race against time. The longer the current situation persists, the greater the risk that a wave of corporate bankruptcies could sweep over the region, potentially pulling down banks in their tow. Asia could then find itself trapped. The combination of corporate and bank failures would quickly transform the recession into a full-blown crisis, making the problem much more intractable to handle and causing serious social dislocation. In other words, the region could find itself back in the same painful situation that it experienced a decade ago during the Asian crisis.

The objective of this chapter is to assess the likelihood of such a scenario. It attempts to answer two broad questions. First, how high is the risk of corporate sector defaults? Second, how large are the expected losses from defaults, and how badly will they affect the banking sector? To answer these questions, the chapter relies heavily on the approach known as Contingent Claims Analysis (CCA). This methodology combines balance sheet information with prices prevailing in financial markets in order to obtain forward-looking measures of the risk of defaults and the potential losses they might cause. Put another way, the CCA attempts to uncover the market’s view of what is likely to happen to the corporate sector, by teasing out the scenario implicit in current market prices. Then, alternative scenarios can be constructed, including what would happen if things turn out worse than the market expects.

The main conclusions of the chapter are as follows:

  • The risk of corporate defaults is unusually high, but still much smaller than that which prevailed during the Asian crisis.

  • Accordingly, the impact on the corporate and banking sectors is likely to be significant but manageable. Losses to creditors (excluding shareholders) from defaults in Asia as a whole could amount to about 2 percent of GDP, while bank losses could amount to about 1⅓ percent of their assets.

  • The main reason the risks are manageable is that the corporate sector entered the crisis in robust health, with low leverage ratios and high profitability.

  • These findings, however, are based on a market-based scenario in which Asia’s economy stabilizes and then gradually recovers. Although this view is consistent with the outlook presented in Chapter 1, the downside risks are sizable and the costs of getting trapped in a corporate-banking sector bankruptcy loop could be immense. So, prudence would suggest taking preemptive measures, especially to shore up the banking system and prepare for a possible surge in corporate bankruptcies, if global demand plunges anew.

How Badly Has the Corporate Sector Been Hit?

In some respects, the situation now confronting Asia’s corporate sector is without precedent in the post-war era. Never have the declines in exports been so large, neither during the Asian crisis, nor during the collapse of the IT bubble in 2001. In most countries, exports have plunged by more than 20 percent year-on-year; in some cases, the decline has reached an astonishing 50 percent year-on-year (see Box 3.1 for an analysis of the impact on China’s corporate sector).

Box 3.1.How Is the Economic Downturn Affecting China’s Corporate Sector?

In China, the downturn has affected corporate profits and increased financial stress, particularly in the real estate and export-related sectors and in sectors where overcapacity has built up in recent years. Growth in corporate profits has slowed since late 2007 across virtually all sectors, and turned sharply negative (year-on-year) in early 2009. Signs of distress have emerged in particular sectors (e.g., real estate, steel, heavy machinery, textiles, toys, automobiles) where firms have been adjusting by running down inventories, cutting prices, and laying off workers. The layoffs mainly involve migrant workers, and are thus not fully reflected in the measured unemployment rate, which covers only registered urban workers. There have been anecdotal reports of pay cuts and wage arrears in some areas and of distress among smaller firms.

China: Year-to-Date Profits and Sales

(Year-on-year percent change)

Source: CEIC Data Company Ltd.

The corporate sector entered the downturn with relatively strong balance sheets in the aggregate, which provided a cushion. Corporate deposit growth in the banking system, while decelerating during 2008 as profits slowed, has remained relatively strong. Compared with other countries, firms in China rely relatively more on retained earnings than on bank credit, and very little on foreign funds and capital market financing. As a result, declines in these latter components have not had much effect on corporate balance sheets and investment.

The corporate sector outlook for 2009 is difficult, with GDP growth in China set to slow to 6.5 percent and output expected to decline in key partner countries. The real estate sector is likely to remain weak as the overhang of vacancies, fall-off in mortgage growth, and the gap between rental and mortgage rates could presage a continued softness in the sector in 2009. Manufacturing investment could weaken if inventories continue to be run down and falling profits drive down retained earnings. Corporate weakness could intensify further if significant overcapacity emerges in certain industries or global demand falters further or protectionist barriers emerge. Defaults on informal lending, through interenterprise credit and loans to customers, could be an additional source of risk. However, the fiscal stimulus and other policy measures should mitigate the negative impacts, particularly if they come alongside a step-up in reforms of key social services that would strengthen the basis for private consumption and strong medium-term growth.

China: Electronics Industry – Probability Distribution of Firm Profits1

Source: IMF staff estimates.

1 Disributions are based on shocks to exports sales of zero percent, 10 percent, and 30 percent, respectively.

A simulation exercise suggests that additional external and real estate shocks,1 were they to occur, could significantly hurt the corporate sector. But the magnitude of shocks required to induce widespread distress and defaults is very large. Weaker exports would have a large impact on the electronics and textile industries, and weaker domestic demand would have a large impact on ferrous metal and nonmetal minerals sectors. Losses would surge, especially in the electronics sector because the distribution of profits in this sector is skewed significantly to the left, with many firms having thin profit margins. The number of bankruptcies in most sectors would increase but remain manageable, because many firms have large equity buffers against losses, and firms’ financial costs are very low. A deeper and more protracted slowdown would dramatically increase bankruptcies in the electronics, textiles, and ferrous metal industries.

Note: The main authors of this box are Vivek Arora, Tarhan Feyzioğlu, and Xu Wei.1 The simulation assumes a 10 percent or a 30 percent decline in external demand or domestic demand. The chart presents the results for the external demand shock.

Two aspects of the situation are striking and very different from the Asian crisis experience:

  • Firms in more advanced economies have been hit much harder than those in developing markets. That is because many specialize in cyclically sensitive sectors that have been particularly affected, such as electronics or automobiles.

  • Companies in the tradable sector have been affected more than domestically oriented firms, because the shock originated from abroad. The major exception is real estate/construction firms, which are typically highly leveraged and have been wounded by sharp falls in real estate prices.

But the problem is not confined solely to the export and real estate sectors. There has been an unprecedented collapse of the manufacturing sector as a whole in industrial Asia and the NIEs, with industrial production falling by about 12– 15 percent year-on-year (Figure 3.1). Production has even declined in the ASEAN-4, though by far less than a decade ago. As a result, corporate profits have been severely undermined. In the December quarter, robust profit growth in China, India, and Australia suddenly ground to a halt, while in Japan manufacturing profits suddenly and completely vanished.

Figure 3.1.Selected Asia: Decline in Industrial Production

(Year-on-year decline in 1998Q3 and 2008Q4, in percent)

Source: CEIC Data Company Ltd.

With cash flows diminishing, the number of financially vulnerable firms has soared. One key measure of corporate health is the Interest Coverage Ratio (ICR), the degree to which cash flows are sufficient to cover the interest on debt. Firms where earnings before interest and taxes are less than interest payments due, that is, with ICRs of less than one, are sometimes referred to as “technically bankrupt.” Many of these firms can survive for a time by selling assets to meet their debt obligations. But if their ICRs remain below one for a prolonged period, eventually they will run out of assets and actual bankruptcy will ensue.

How have Asian firms fared on this measure? At end-2007, only about 15 percent of listed firms had ICRs less than one. Moreover, most of these firms were small, together accounting for only 7 percent of corporate sector debt (Figure 3.2). But in 2008, the share of firms with insufficient profits is estimated to have risen to about 17 percent of all firms, accounting for about 10 percent of total corporate debt.21 The estimated increase is stark in ASEAN-4, because many of the firms in these countries were close to the threshold before the crisis, and so quickly crossed over when earnings suffered at the end of last year.

Figure 3.2.Asia: Share of Debt of Firms with Interest Cover Ratio Less than One


Sources: Worldscope; and IMF staff estimates.

1 Assuming a 15 percent decline in profits.

How Large Are the Default Risks?

So far, there is little sign that these stresses have been translating into an increase in corporate bankruptcies. In Taiwan Province of China, the number of companies dissolved has actually been falling in recent months (Figure 3.3); in Korea, there was initially an uptick in bankruptcies after September 2008, but the rate has subsided again. Only in Japan (among economies for which there is ready data) does the crisis seem to have caused bankruptcy rates to jump.

Figure 3.3.Selected Asia: Corporate Bankruptcies1

(Units per month, from January 2007 to March 2009)

Source: CEIC Data Company Ltd.

1 For Taiwan Province of China, data refer to companies dissolved or licenses revoked, up to February 2009.

However, financial markets have reacted to growing vulnerability with trepidation. By February 2009, share prices in emerging Asia had plunged by about 60 percent year-on-year, a drop as precipitous as the one that occurred during the Asian crisis (Figure 3.4). Particularly punished have been the cyclically sensitive sectors, such as financials, industrials, and consumer discretionary (autos, consumer services, and durables) (Figure 3.5). Mainly, this is because the shock has hit these sectors hardest. But in some cases it is also because the sectors had preexisting vulnerabilities. For example, estimates suggest that at end-2008, about 20 percent of the debt of electronics firms in ASEAN-4 was owed by financially vulnerable firms with ICRs less than one.

Figure 3.4.Emerging Asia: Stock Market Performance

(Year-on-year percent change)

Source: Morgan Stanley.

Figure 3.5.Asia Excluding Japan: Equity Performance by Sector

(Percent change over September 2008 to-April 2009)

Source: Morgan Stanley.

Financial markets have also drawn a sharp distinction between large firms and small and medium-sized enterprises (SMEs). When the crisis hit, large well-established firms quickly turned to domestic banks for funding, in some cases to repay foreign debts, in other cases as a precautionary measure. But SMEs were unable to do the same (Figure 3.6). To the contrary, they found that their access to bank credit was curtailed quickly and sharply, a development that prompted governments to step in with loan guarantees.

Figure 3.6.Korea: Lending by Size of Companies

(Year-on-year percent change)

Source: CEIC Data Company Ltd.

Why were banks so quick to clamp down on SME credit? They seem to have responded to the SMEs’ greater vulnerability. In the previous decade, many SMEs had borrowed heavily to expand their activities, notably as suppliers to larger manufacturing enterprises. But these expansions proved insufficiently profitable. The financial health of listed SMEs was relatively weak—and listed firms were most likely in better shape than their unlisted brethren. In 2007, the share of small firms with ICRs less than one was about 25–30 percent (Table 3.1). In contrast, the share of debt of financially vulnerable large firms was only in the range of 1–10 percent.

Table 3.1.Selected Asia: Share of Debt of Firms with Interest Cover Ratio Less than One, by Size, 2007(Ratio)
Industrial Asia
Sources: Worldscope; and IMF staff calculations.
Sources: Worldscope; and IMF staff calculations.

Worries about default, however, extend far beyond the SMEs. There are no observable prices that directly measure the expected probability of defaults. But, such probabilities can be inferred in two ways. One is by looking at the spreads on credit default swaps (CDS) on the dollar bonds issued by Asian corporates abroad. These indeed indicate that perceived risk has soared, even for large companies that were able to borrow internationally.

Spreads on investment-grade firms, which had been running about 150 basis points (bps) before September 2008, almost immediately afterward jumped to about 300 bps, while those on high-yield companies soared to about 1,300 bps (Figure 3.7). In other words, in March 2009, bondholders seeking protection on non-investment-grade companies were willing to pay on average 13 percent of the face value of their bonds per year as an insurance premium against default, an astonishingly high amount.

Figure 3.7.Asia: Credit Default Swap Spreads1,2

(In basis points)

Source: Bloomberg LP.

1 Excluding Japan.

2 Uses Traxx indices, which are a set of credit default swap indices covering regions or sectors and contain the most liquid names in that market, based on a dealer poll.

This suggests that financial markets perceive default risks for large Asian corporates to be exceptionally high. This would be worrisome, for if significant numbers of large companies go under, a wave of defaults could cascade down Asia’s tightly integrated supply chain. But it is also possible that the figures are misleading, because Asian CDS markets are thin, distorted by the financial difficulties of the major global banks, and in any case involve only a handful of companies, possibly unrepresentative of corporate Asia as a whole.

Indeed, the second method of inferring default probabilities based on the CCA methodology, gives a very different view. This approach has the advantage of using share prices, so the analysis can be extended to a much larger universe, namely all the listed firms in Asia. Under this approach, the risk of default is related to the probability that the value of a firm’s assets will fall below the value of its liabilities. This, in turn, depends on two main factors: firm leverage (debt relative to the market value of its equity) and uncertainty about the value of firm assets, which reflects the expected value of future profits. Both of these factors are related to share prices, because when stock prices fall, this diminishes the market equity base; and when price volatility increases, this implies growing uncertainty over asset values. Both increase the probability that a firm will default. With this and other information, expected default probabilities one year ahead can be calculated using the CCA framework.

What do such calculations show? They indicate that expected default probabilities have increased markedly. For example, back in July 2007, only a small fraction of firms—those with less than 1 percent of corporate sector assets—had a default risk one year ahead that exceeded 5 percent. But by March 2009, this proportion had increased to about 14 percent (Figure 3.8). A key reason is the collapse of share prices and the soaring volatility. In addition, tighter financial conditions have raised debt service costs and rollover risks, because short-term debt represents more than 60 percent of corporates’ total debt in the sampled countries.

Figure 3.8.Asia: One-Year-Ahead Default Probability of Nonfinancial Corporates

Sources: Moody’s KMV; and IMF staff estimates.

The same calculations can be looked at a slightly different way. One could examine how default probabilities have changed for firms at different levels of risk. For example, the default probability for the median firm, an “average” firm in the sense that half the firms in Asia have a higher risk of default, has risen by 1¾ percentage points from September 2008 to March 2009, an amount equivalent to a 10-standard-deviation increase relative to the post-2004 average. For firms at the 75th percentile, the risk of default reached 8 percent in March, implying that nearly one-quarter of listed Asian corporates have even higher default risk22 (Figure 3.9). But, significantly, these levels of risk remain far below the levels reached during the Asian crisis, which for firms at the 75th percentile peaked at about 19 percent. Current levels are also smaller than the previous peak, during the IT bubble collapse in 2000–01.

Figure 3.9.Selected Asia: Historical Expected Default Frequency1

(In percent)

Source: Moody’s KMV.

1 Includes NIEs, ASEAN-4, and China.

These market-based default probabilities are based on a bad but not disastrous scenario for the corporate sector. Unlike in advanced countries, there are no consensus forecasts of corporate earnings. However, econometric techniques can be employed to tease the scenario out. Accordingly, a vector autoregression (VAR) model was estimated to establish the relationship between the expected default probabilities and the actual rate of industrial production, an observable and timely proxy for corporate revenue.23 This estimation suggests that changes in default probabilities are indeed good predictors of industrial production one year ahead, especially when financial conditions are also taken into account. Specifically, the 1¾ percentage point increase that had occurred in the default probability during May–October 2008 for the median firm predicts declines in industrial production by mid-2009 that are slightly larger than the declines that have already taken place. For example, industrial production would fall by about 30 percent in ASEAN-4, 45 percent in the NIEs, and 60 percent in industrial Asia (Figure 3.10), taking into account the change in financial conditions. Put another way, the default probabilities seem to be based on a scenario similar to the one presented in Chapter 1, in which things do not get much worse but will take some time to get better.

Figure 3.10.Asia: Change in Industrial Production—Actual vs. Predicted1

(In percent)

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

1 Predicted values are the model prediction for 10 months ahead, based on changes in expected default frequencies from May to September 2008.

Another way to assess the market scenario is by comparing the current implicit forecasts with those that would have prevailed at the time of the Asian crisis. To do this, a model was estimated relating changes in default probabilities for individual firms to the amount of their investment over the subsequent year. It proved to have strong predictive power (Appendix 3.2). Then, the predicted change in investment based on the annual change of default probability in 2008 was compared to the prediction the model would have made in 1997–98.24 It turns out that the currently expected decline in corporate investment in Asia as a whole is indeed smaller than what was expected at the time of the Asian crisis (Figure 3.11), across all country groupings, especially in the ASEAN-4 countries.

Figure 3.11.Expected Implication of Higher Default Risks on Firm Investment1

(Decline in capex/total assets ratio)

Source: IMF staff estimates.

1 Estimated by β(1)*ΔDefault Risk, market capitalization weighted average within country, purchasing power parity weighted average across countries.

Finally, it is important to consider whether the market expects corporate distress to spill over into the banks. Estimates from a VAR model confirms that they do. The model shows that the cumulative impact on banks’ default risk from a 1 percentage point shock to corporate sector default probabilities exceeds one in most country groupings. The results are particularly strong for industrial Asia, where a 1 percentage point increase in corporate default risk leads to a nearly 2 percentage point rise in banks’ default risks (Figure 3.12). The larger increase in bank default risk reflects the fact that banks’ assets, which are in part composed of corporate loans (loans to corporates account for 45–60 percent of bank loans), are leveraged. This means it will be important to examine the impact of expected corporate losses on the banking system, which will be done below.

Figure 3.12.Selected Asia: Cumulative Impact on Banks’ Default Probabilities from Shock to Corporate Default Probabilities, after 10 Months

(In percent)

Source: IMF staff estimates.

Summing up, the scenario that markets seem to be expecting is similar to the one outlined in Chapter 1. Markets expect a substantial drop in industrial production and corporate investment, and a significant rise in corporate defaults. But they do not seem to be expecting defaults on anywhere near the scale that occurred during the Asian crisis, even for country groupings such as industrial Asia that emerged relatively unscathed a decade ago. This immediately raises a question: how large are the default losses expected to be?

How Large Are the Likely Default Losses?

Although mathematically complex, the basic strategy for deriving expected losses using the CCA is relatively straightforward.25 However some definitions are needed. The term “expected losses” refers to the present value of expected losses due to default, estimated using information on corporates’ equity, market value of assets, debt, and the volatility (or the risk) associated with the assets.26 Importantly, these figures are not comprehensive estimates of corporate losses, or even losses at defaulting corporates, but rather the expected losses that bondholders and banks would be forced to absorb after the equity of defaulting companies’ shareholders has been entirely wiped out (IMF, 2008a). In other words, they are partial figures, focusing on losses to creditors only. So, they are most useful as a comparative guide to indicate which regions will suffer the most. Also, because they focus on losses to creditors, they can provide a good base for estimating the potential impact on the banking system.

The model suggests that default losses could be significant. For Asia as a whole, expected losses would amount to about 2 percent of GDP (using equity price data as of mid April 2009)—a relatively high figure considering its partial nature and the fact that it is derived from a scenario in which the economic situation does not deteriorate much further. Interestingly, the expected losses correlate well with the degree to which country groupings have been affected so far, with losses in ASEAN-4 estimated to amount to about 1½ percent of GDP while those in the NIEs are predicted to reach a sizable 4 percent of GDP (Figure 3.13). Estimates for industrial Asia are close to the lower ASEAN-4 level, even though this region has been hit particularly hard, largely because the big companies there are well established and thus less likely to default.

Figure 3.13.Asia: Nonfinancial Corporate Sector—Annual Average Expected Losses One Year Ahead

(In percent of GDP)

Sources: Moody’s KMV - Credit Edge; and IMF staff estimates.

1 Dots represent point estimates. Confidence intervals assume 80 percent and 40 percent recovery rate in case of default.

There is, naturally, a range of uncertainty around these estimates of creditors’ losses. Perhaps the most important unknown is the degree to which creditors would be able to recover on their collateral when the defaults occur. In the baseline case, recovery rates on corporate assets are assumed to be the same as the historical average recovery rate of the industry that each defaulting corporate operates in. As alternatives, the recovery rate is assumed to be as much as 80 percent or as low as 40 percent.

As Figure 3.13 shows, the point estimates are generally close to the top of the range, implying that in most countries historical recovery rates are closer to 40 percent. Not surprisingly, in industrial Asia the situation is much better, as the point estimate is in the middle of the range, implying a recovery rate of about 60 percent. Should Asian countries be able to improve recovery rates even further to 80 percent, the sensitivity analysis shows that expected losses could be cut in half, to about 1 percent.

In sum, the calculated expected losses seem significant, but manageable. But how can this be? How can the extent of default losses be manageable when the collapse in industrial production is even greater than during the Asian crisis?

Why Is Asia’s Corporate Sector Expected to Remain So Resilient?

The answer to this question is quite straightforward: Asia entered the global financial crisis in a relatively healthy state. The comparison with a decade ago is instructive. In 1997, corporate Asia was in an extremely vulnerable position. Its leverage was high and its profitability low; it had large unhedged foreign currency and short-term debts. So, when exchange rates fell and interest rates rose, the sector was quickly devastated.

In contrast, thanks to the postcrisis restructuring and a long global boom, by 2007 corporate positions were exceptionally strong on every standard measure.

  • Leverage. The Asian corporate sector had deleveraged significantly, with the debt-to-equity ratio (market capitalization weighted average) for the region as a whole falling by half from its peak to just 75 percent in 2007 (Figure 3.14). In addition, its debt structure had also improved, with short-term debt falling as a share of total debt.

  • Debt service. Similarly, corporate financing costs, as measured by the average interest paid on debt outstanding, had declined as interest rates declined across the globe and spreads collapsed.

  • Profitability. With demand booming and companies paying much closer attention to the bottom line, profitability increased sharply, with the average rate of return on assets reaching about 10 percent for Asia as a whole and 14.5 percent in ASEAN-4 in 2007 (Figure 3.15).

  • Liquidity. During the boom period, firms also built up their liquidity, a precaution that proved extremely helpful when the recession hit. The average quick ratio (the current assets of a firm net of inventories divided by current liabilities) had increased above one in every country grouping (Figure 3.16), indicating that firms were able to service a year’s worth of obligations simply by using their liquid assets (cash plus marketable securities plus accounts receivable).

Figure 3.14.Asia: Leverage (Debt-to-Equity Ratio)

(In percent)

Source: IMF, Corporate Vulnerability Utility.

Figure 3.15.Asia: Profitability (Return on Assets)

(In percent)

Source: IMF, Corporate Vulnerability Utility.

Figure 3.16.Asia: Liquidity (Quick Ratio)

(In percent)

Source: IMF, Corporate Vulnerability Utility.

These strong initial conditions have made Asia resilient to what has been an exceptionally large shock. In addition to the shield provided by strong balance sheets, firms are likely to attempt to slash costs by delaying investment plans and shedding labor to avoid entering into bankruptcy. Even so, there will be a significant amount of defaults, which means there will be a sizable spillover to the banking sector. To quantify exactly how large this spillover will be, we turn back to the CCA.

How Badly Will the Banks Be Affected?

Translating the corporate sector’s expected losses into banking sector losses requires several steps. Essentially, one needs to apportion the losses among the various creditors, according to the relative importance of these sources and the seniority structure of the debt. Much of the information needed to do this is unavailable, but some approximations can be made (Appendix 3.3). When this is done, the estimates suggest that new bank writedowns could range from 1¾ percent of total 2008 loans in industrial Asia to 2½ percent of loans in ASEAN-4 (Figure 3.17). Such writedowns would bring banks’ cumulative losses—that is, existing provisions plus expected new writedowns—to as much as 6 percent of banks’ total loans in ASEAN-4 economies.

Figure 3.17.Asia: Banking Sector—Expected Losses from Corporate Sector Distress One-Year-Ahead1

(In percent of total banking sector loans)

Sources: Moody’s KMV - Credit Edge; and IMF staff estimates.

1 Dots represent point estimates. Confidence intervals assume 80 percent and 40 percent recovery rate in case of default.

Note an apparent paradox. Expected losses from corporate defaults are highest in the NIEs by far, but the impact on the banks is expected to be the greatest in ASEAN-4. Although the result seems strange, the explanation is simple. The banking sector in ASEAN-4 is much smaller compared to other country groupings, so corporate losses that may seem modest as a share of GDP will have a disproportionately large impact on ASEAN-4 bank capital.

Even so, banks in all regions should be in a position to absorb the projected writedowns, because they too entered the crisis in strong capital positions, relative to the minimum Basel requirements.27 Under relatively conservative assumptions regarding banks’ operating profits this year, the toll on total and Tier 1 capital is expected to range between ½ percent and 1¼ percent of bank assets across country groupings. After these reductions, bank capital will still remain sufficient, with regulatory capital asset ratios around 10 percent for Asia as a whole and Tier 1 capital exceeding 5.4 percent in all country groupings (Table 3.2).

Table 3.2.Asia: Potential Impact on Bank Capital from Corporate Sector Distress1(In percent)
CapitalReturn on Assets4
Regulatory2Tier 13
Before writedownsAfter writedownsBefore writedownsAfter writedownsBefore writedownsAfter writedowns
Industrial Asia12.
Sources: Moody’s KMV - Credit Edge; and IMF staff estimates.

Based on estimated losses using Moody’s corporates recovery rates.

Refers to Basel Capital Adequacy Ratios, with the minimum ratio typically around 8 percent.

Minimum Tier 1 ratio is typically 4 percent.

Profits are assumed to decline by 50 percent relative to their 2008 level.

Sources: Moody’s KMV - Credit Edge; and IMF staff estimates.

Based on estimated losses using Moody’s corporates recovery rates.

Refers to Basel Capital Adequacy Ratios, with the minimum ratio typically around 8 percent.

Minimum Tier 1 ratio is typically 4 percent.

Profits are assumed to decline by 50 percent relative to their 2008 level.

These estimates, however, need to be treated with caution. For a start, the figures almost certainly underestimate the extent of the likely losses. That is because they include only those bank losses that stem directly from defaults on corporate sector loans. And corporate sector loans account for only around 45–60 percent of bank loan books. In particular, the estimates exclude losses from loans to households also, which may be quite large because the recession has also put them under stress, from rising unemployment and falling housing prices.

One way to cross-check the results is to calculate bank losses directly from bank share prices using the CCA. These losses are naturally higher, amounting to 2½ percent of GDP on average for Asia.28 They are also in line with IMF staff estimates of bank losses on loans and securities of about 2 percent of GDP (excluding China), presented in the April 2009 Global Financial Stability Report. Even using these more comprehensive figures, however, bank losses will still be a far cry from those experienced during the Asian crisis. In that case, banks’ recapitalization costs varied from 20 percent of GDP to 35 percent of GDP in the cases of Indonesia, Korea, the Philippines, and Thailand (Berg, 1999).

But there is one final problem with the results: they are based on the scenario that Asia’s economy will stabilize and only gradually recover. What would happen if instead things get worse?

Stress Testing the Corporate Sector

There are two possible ways to examine a more adverse scenario. The first is by employing more pessimistic assumptions in the CCA. Consider the case, for example, where worsening economic conditions lead to a fall in corporate share prices by an additional 50 percent from March 2009 levels. If that happens, default probabilities would soar by 4 percentage points for the median firm, levels not that far away from those reached in the Asian crisis. Under this scenario, expected corporate losses would amount to 2–7 percent of GDP, with the smallest losses again occurring in ASEAN-4 and the largest in the NIEs. As a result, banks would have to write down 2–4 percent of their loans, reducing their regulatory capital ratios by about 1½– 2 percentage points. This would represent a sizable blow to banks; while average capital in all country groupings would still remain a few percentage points above the minimum 8 percent regulatory norm, this would not necessarily be true for individual banks.

Another way to examine what would happen if things go wrong is to stress test corporate balance sheets. This approach involves examining how different types of shocks to the corporate sector, such as a specified fall in profits, would affect firms’ viability as measured by their ICRs (Appendix 3.4). The results provide a guide to the types of shocks that would be particularly painful for the corporate sector, and the types of firms that would be particularly vulnerable.

The stress tests suggest that corporate Asia would be particularly vulnerable to further demand shocks. For example, in one test, profits were assumed to fall another 35 percent from their estimated 2008 level, in line with the drop that occurred during the Asian crisis. In that case, the share of Asian firms with ICRs less than one would rise to around a quarter of all firms, accounting for about 17 percent of corporate debt (Figures 3.18 and 3.19). Firms in the NIEs and ASEAN-4 would be particularly affected, as the share of low-profit firms in these groupings would exceed a third of all firms.

Figure 3.18.Profit Shock vs. Interest Rate Shock: Share of Firms with Interest Cover Ratio Less than One


Sources: Worlscope; and IMF staff estimates.

Figure 3.19.Profit Shock vs. Interest Rate Shock: Share of Impaired Debt of Firms with Interest Cover Ratio Less than One


Sources: Worldscope; and IMF staff estimates.

Breaking the results down by sector shows that three types of firms would be particularly affected:

  • Small firms.29 The earnings shock would push up the share of debt of low ICR firms to 30– 50 percent (Figure 3.20).

  • Electronics sector. If profits fall sharply, 70 percent of debt of electronics firms would be at risk in the ASEAN-4 countries (Figure 3.21).

  • Construction companies. In the ASEAN-4 countries, a major portion of construction firms’ debt would become impaired.

Figure 3.20.Share of Debt of Firms with Interest Cover Ratio Less than One, by Size


Sources: Worldscope; and IMF staff estimates.

Figure 3.21.Share of Debt of Firms with Interest Cover Ratio Less than One, by Sector


Sources: Worldscope; and IMF staff estimates.

On the positive side, corporate Asia is much less sensitive to other types of shocks. Increases in interest rates matter much less than they did a decade ago, because firms have deleveraged considerably. For example, an increase in interest rates of 100 basis points would have a much smaller impact on ICRs than a further 35 percent profit decline. Moreover, the likelihood of such a shock is extremely small—so far during the current recession, domestic interest rates have actually been falling.

Sensitivity to changes in foreign interest rates and exchange rates also appears to have diminished. It is impossible to test these sensitivities directly, because most firms report only their aggregate debt levels, without breaking them down into foreign and domestic debt. But some inferences can still be drawn. For example, firms are unlikely to be much affected by soaring foreign interest rate spreads. That is because access to new borrowing is extremely limited, whereas all of the existing direct corporate borrowing would have been done at fixed rates or fixed spreads. As for exchange rates, sensitivity is likely to have fallen sharply over the past decade, because external debt ratios have come down considerably. Indeed, for many exporters, exchange rate depreciations may now be beneficial, boosting the local currency value of their revenues by more than they increase their more modest external debts.

Overall, the stress tests are a bit worrisome. Although the corporate sector may now have less cause for concern from higher interest rates and depreciating exchange rates, plausible further falls in profit levels could swell the ranks of the “technical defaulters” to exceptionally high levels, particularly among certain vulnerable subsectors.


We are now in a position to answer the questions posed at the beginning of the chapter. Is Asia likely to see a wave of corporate defaults? Most likely, yes. Already, many industries have seen demand and profits collapse. Now, they are facing a squeeze from the financing side, particularly from external creditors. It will be difficult for firms to roll over their mounting external debt obligations, since external bond markets have been shut to all but the highest-rated companies since mid-2008 and are likely to stay shut for some months to come. Meanwhile, the scope for substituting to domestic borrowing is narrowing, since domestic banks, like those in advanced countries, have become acutely concerned about credit risks. So, defaults are bound to rise.

How significant will these defaults be? In the baseline case, the resulting losses will be manageable, in the sense that they will not unduly deplete bank capital. But the current fragile global situation gives little ground for complacency. If global demand plunges anew, stress tests indicate that the ranks of defaulters would grow to uncomfortably high levels. In that case, the region could suddenly find itself trapped in an adverse feedback loop. Large-scale corporate defaults could severely damage banks, rendering them unable to extend credit, which in turn will put further pressure on the corporate sector.

Given the serious implications of a more adverse scenario, it would seem important to take preemptive measures to limit the potential ramifications. In particular, countries may want to reexamine their corporate bankruptcy frameworks. Research shows that bankruptcies in emerging Asia are time-consuming, costly, and likely to yield little for creditors. If bankruptcy procedures can be improved and streamlined, then recovery rates could be improved, and the impact of corporate losses on the banking sector could be better circumscribed.

At the same time, it would be helpful to increase banks’ capital further so they can better absorb potential losses. Asian banks have raised capital of US$73 billion in the five months to March 2009, helping to ensure that their capital adequacy will remain at healthy levels. But further efforts to shore up capital are still needed, given the risks that lie ahead, as well as the market-driven demands to maintain higher capital cushions and higher quality capital, such as tangible common equity.

Putting things another way: so far, so good. Asia’s corporate sector has withstood an enormous shock reasonably well. With some luck, the final toll will be readily absorbable. But the risks are sizable, in which case, a pound of preparation could be worth much more than a ton of cure.

Appendix 3.1. Using Vector Autoregressions to Analyze the Transmission of Shocks across Sectors

To analyze the transmission of shocks across sectors, a series of vector autoregressions are estimated. In these VARs, default risks of banks and corporates are incorporated in addition to standard macroeconomic variables to predict industrial production and banks’ default risks.30 The vector includes the following variables at the monthly frequency: the default probabilities of corporates and banks, industrial production, credit growth, and inflation.

The baseline VARs are augmented by a global Financial Conditions Index (FCI), which is considered exogenous. The FCI is an equal-variance weighted average of seven variables, from the banking sector, securities market, and foreign exchange market (Cardarelli, Elekdag, and Lall, forthcoming).

The model is estimated in first differences for variables that appear not to reject the null of a unit root. The lag structure is determined according to the Akaike information criterion and the Bayesian information criterion. The generalized impulse responses (Pesaran and Shin, 1998) are utilized to avoid the sensitivity of the results to the ordering of the variables in the VAR.

Appendix 3.2. Micro-Level Evidence on Real Macro-Financial Linkages

A standard firm-level investment model based on Tobin’s Q is augmented to incorporate measures of firm vulnerability (Equation 1). A firm-level panel equation of one-year ahead capital expenditure (standardized by asset size) is estimated for each country using ordinary least squares (OLS), controlling for macroeconomic and external factors through time dummies and firm-level fixed effects.

Capex(t+1, i)/ BookValueof TotalAssets(t, i) =const.+β1DefaultRisk(t,i)+β3TimeDummies+β4FirmFE+ε(t,i) (1)

Standard deviation is clustered by firms. The Tobin’s Q is proxied by annual sales growth.31 In addition to this baseline model, alternative models with additional firm-specific characteristics, including size, equity volatility, return on assets, and leverage, are estimated for robustness checks.

All firm-level (i) data, including (end-year) default probability, are annual (t) and from the IMF’s Corporate Vulnerability Utility (CVU), based on Worldscope and Datastream.32 For most of the countries, the data start in the mid-1990s and end in 2007. The analysis is focused on nonfinancial firms. The sampled countries include China, India, ASEAN-4, the NIEs, and industrial Asia.

The estimation results indicate that for most of the sampled countries firms’ default probabilities do have statistically significant predictive power for future investment. Moreover, for most countries, the coefficient on the default probability continues to remain statistically significant and of similar magnitude when other firm-specific characteristics are included.

Table 3A.1.Summary: Estimation Results
New Zealand***noyes
Hong Kong SAR***yesyes
Taiwan Province of China***noyes
Source: IMF staff estimates.

Statistical significance of default risk coefficient (β1 in eq.1). Significant at 10 percent (*), 5 percent (**), and 1 percent (***) levels.

Robustness against inclusion of other firm characteristics.

Estimated parameters in alternative specifications stay within 95 percent confidence interval of the baseline estimation.

Source: IMF staff estimates.

Statistical significance of default risk coefficient (β1 in eq.1). Significant at 10 percent (*), 5 percent (**), and 1 percent (***) levels.

Robustness against inclusion of other firm characteristics.

Estimated parameters in alternative specifications stay within 95 percent confidence interval of the baseline estimation.

Appendix 3.3. Computation of Banks’ Expected Losses from Corporate Sector Distress

Banks’ expected losses from corporate sector distress were calculated using information from Moody’s KMV implied CDS (EICDS) spreads and banks’ exposure to the corporate sector. The calculation involved the following steps:

  • First, expected losses for the corporate sector one year ahead embedded in EICDS spreads are calculated using the contingent claim analysis framework.

  • Second, the corporate sector’s expected losses are expressed as ratios of the corporate sector’s total liabilities. It is then assumed that all the corporate sector’s creditors will suffer the same relative losses in their books—for example, if the corporate sector’s expected losses represent 10 percent of the corporate sector’s total liabilities, then the banking sector will write down 10 percent of its current performing loans to the corporate sector. This approximation is necessary because a more precise calculation requires information on the seniority structure of the debt and on the relative importance of domestic versus foreign financing sources—data that we do not have.

  • Third, banks’ current performing loans to the corporate sector are calculated. Here in the absence of information on banks’ current provisions for losses on loans to the corporate sector, banks’ overall provisions for losses are subtracted from the current stock of their loans to the corporate sector, and the resulting amount is scaled by banks’ exposure to the corporate sector.

  • Fourth, the relative losses calculated in the second step are multiplied by the current stock of performing loans to the corporate sector calculated in the third step. The resulting amount is the expected increase in banks losses stemming from banks’ exposure to the corporate sector.

The calculations were made for individual countries and then aggregated into regional groupings using 2008 purchasing power parity (PPP) weights.

Appendix 3.4. Stress Testing Corporate Balance Sheets

The methodology focuses on analyzing how shocks affect the debt servicing capacity of corporates, following the framework utilized in Jones and Karasulu (2006), Heytens and Karacadag (2001), and Oura and Topalova (2009). As is standard in the literature, the debt-servicing capacity of a firm is measured by the interest cover ratio. Specifically, the idea is to analyze how many firms go into “technical default” (interest cover ratio less than one) when buffeted by various shocks. The results are expressed in terms of the debt of firms with low interest cover, or the “impaired” debt, as a percent of total corporate sector debt. An interest cover ratio less than one does not necessarily imply that the firm will default as it may have other liquid assets to draw upon. Nevertheless, it does point to some level of corporate distress that is not sustainable over the long term.

The stress tests use firm-level data from Worldscope for 13 countries.33 In the absence of complete balance sheets for 2008, the 2007 balance sheets are adjusted to arrive at an estimate of the end-2008 balance sheet position, assuming an estimated decline in profits of 15 percent across the board. This is considered the baseline.

The profit shock of 35 percent is based on the average decline during the Asian crisis.34 This is also in line with a one-standard-deviation shock to profits. Next, we consider a shock to the cost of financing just to give an idea of the impact of a significant rise in interest rates. So far, however, interest rates on domestic bank borrowing have actually declined while there has been almost no new international borrowing.

Note: The main authors of this chapter are Sonali Jain-Chandra, Papa N’Diaye, and Hiroko Oura. Adil Mohommad provided research assistance. The authors thank Kenichi Ueda for sharing his Matlab code and the Worldscope data used in the IMF’s Corporate Vulnerability Utility, and Petia Topalova for her Stata code.

Since 2008 balance sheets are not available in many cases, results for that year were derived by applying an estimated profit decline (15 percent) to 2007 balance sheet data.

In April, the default risk at the 75th percentile receded to 7 percent.

For details, see Appendix 3.1.

Because many firms still have not reported their 2008 results, the end-2008 default probability is computed using 2007 balance sheet and 2008 equity price data. The change in default risk during the Asian crisis is measured using the largest four quarter change in default risk between 1997Q1 and 1998Q4, because the timing of the increase in default risks varied significantly from country to country.

In the underlying Merton model (see Gray and Malone, 2008), these expected losses are equal to the value of a put option, the underlying assets of which are the assets of the firm, and the strike price is defined by the firm’s liabilities.

For discussion of the issue of capital buffers higher than the minimum Basel capital requirements, see Bank for International Settlements (2006).

These estimates exclude losses for Singapore and New Zealand banks owing to data constraints. They also exclude China for comparability with the GFSR estimates (see IMF, 2009a).

The sample is divided into three categories—small, medium, and large—on the basis of market capitalization. Small refers to firms below the median, medium to firms between the 50th and 90th percentile, and large to firms above the 90th percentile. These results are robust to alternative cutoffs; they also hold if total assets were used to rank firms rather than market capitalization.

Papers that examine the relationship between macro variables and expected default frequencies of firms include Castrén, Dées, and Zaher (2008), and Åsberg Sommar and Shahnazarian (2008).

We also estimated models with the Tobin’s Q proxied by (market value of equity + book value of debt)/(book value of assets). Because market value of equity appears in both this measure of Tobin’s Q and default risks, multicollinearity could undermine their parameter estimation. While this seems to affect the estimated parameter and standard deviation of the coefficient of Tobin’s Q, rather than that of default risks, we adopt a measure of Tobin’s Q without using market value of equity.

We resort to CVU data because the KMV data are aggregated, do not accompany firm-level balance sheet data, and cover only five years.

Worldscope data consists only of listed companies, so the analysis excludes unlisted SMEs. Worldscope data do not provide the currency composition of debt, so exchange rate shocks—which may be an important source of risk—are not considered.

In absence of aggregate profits data for all countries in the sample going back to the Asian crisis, we aggregated the firm level profits from the Worldscope database for the standard 13 economies analyzed in the REO (excluding Vietnam). These are then aggregated using PPP weights. During the Asian crisis, profits fell by 37 percent for all 13 countries. Profits in the crisis hit countries fell more dramatically, by around 200 percent, with many firms going into losses.

    Other Resources Citing This Publication