What Are the Fiscal Causes of Crisis?
A review of the theoretical literature suggests three main ways in which fiscal policy can cause a financial crisis. These are through an overly expansionary fiscal stance, leading to a lending and/or consumption boom; through concerns about sustainability, which could be triggered by news about contingent liabilities or by a shift in expectations about the government’s commitment to fiscal adjustment; and through the maturity and currency structure of public debt, which can be critical to perceptions of government liquidity, and hence increase vulnerability to self-fulfilling crises.
Conclusive evidence on the causes of crises is hard to find. In particular, most existing empirical studies of crises are more relevant to the issue of correlation than causation. That said, the evidence on fiscal variables is mixed. Some studies find the deficit and debt variables to be significantly correlated with crises, others do not. But very few studies conclude that fiscal variables provide the main “explanation” of crises. The empirical analysis of a large sample of emerging market economies finds robust evidence that a few fiscal variables are correlated with crises and with pressure in the foreign exchange market. There is also evidence suggesting that fiscal effects operate indirectly, through variables such as the exchange rate or the current account.
In contrast, the 11 case studies of financial crises in the 1990s suggest a direct and important role for fiscal policy in causing crises. In six countries—Bulgaria (1996–97), Pakistan (1998–2000), Russia (1998), Ukraine (1998–99), Brazil (1998–99), and Ecuador (1999)—fiscal problems were clearly central to, and arguably the single most important cause of, the crises. In three countries—Mexico (1994–95), Argentina (1995), and the Czech Republic (1997)—fiscal problems contributed to the crisis, but were probably not the most important factor. In the other two countries—Thailand and Korea—the key vulnerabilities were rooted in the private sector, but fiscal policy still played a role in the Thai crisis.
Several of the case studies confirm the importance of the structure of public debt as a source of vulnerability to crisis. The most obvious example is the Mexican government issuing large amounts of short-term dollar-denominated debt in 1994, but a similar reliance on short-term domestic and external financing—and low policy credibility—led to the rollover problems that precipitated the crises in Russia and Ukraine. When external public debt is high relative to tax revenue and exports, the public finances are vulnerable to a fall in the exchange rate, which increases debt-service costs without having a large positive effect on growth.
The case studies also show how fiscal policy can contribute to financial sector vulnerabilities. On the asset side, this could come through banks holding significant amounts of government paper that is restructured (Russia, Ecuador), or by contributing to nonperforming loans as a result of directed lending (Bulgaria), financing the deficits of provincial governments (Argentina), or through implicit government guarantees and tax breaks contributing to balance sheet vulnerabilities in the corporate and financial sectors (Thailand). On the liability side, banks in Ecuador suffered deposit runs as a result of tax changes prompting large-scale withdrawals, and due to a widening fiscal deficit undermining the credibility of deposit guarantees.
Which Fiscal Vulnerability Indicators Help to Predict Crises?
A number of fiscal variables are potentially useful as univariate leading indicators for signaling crises on an annual basis. The best indicators—short-term debt, foreign currency debt, and various deficit measures—perform as well as the best (annual) leading indicators in other signals EWS studies. While even the best performing indicators fail to signal around two-thirds of crises, these indicators send very few false alarms, and therefore warnings about an impending crisis should be heeded. These findings apply as much to banking crises as currency and debt crises, suggesting that fiscal variables may indeed have a part to play in predicting financial sector crises. This is somewhat surprising, given the absence of a role for fiscal variables in most other EWS models of banking crises.
In a multivariate context, fiscal variables appear to add relatively little to the IMF’s main EWS model. Bearing in mind the relatively low frequency of most fiscal data, and the premium on high-frequency data for EWS models, this suggests that fiscal variables are of limited use in parsimonious probit EWS models. But there is evidence from both the signals and probit EWS models that fiscal variables help more to identify tranquil or noncrisis periods than to signal crisis episodes.
Regarding particular types of fiscal indicators, the main findings are as follows:
Deficit and financing variables can certainly send clear signals of impending crisis. This was the case in the crises in Bulgaria, Pakistan, Russia, Ukraine, Brazil, and Ecuador in the late 1990s. But in other crises, as mentioned above, standard measures of the deficit did not send a clear signal in advance. The evidence from the event studies is that deficits are significantly higher, on average, in the two years prior to currency and debt crises than in noncrisis periods.
Total debt does not seem to add much to crisis prediction. In the event studies, most debt variables are no different, on average, prior to a crisis than in noncrisis periods; they rank poorly among fiscal variables in the signals EWS approach; and they are insignificant or have perverse signs when included in probit EWS models.
Some variables relating to the composition of debt emerge as useful predictors of crises. The event studies show that short-term debt is significantly higher in the run-up to crises; indeed, this is the only variable examined that is systematically “abnormal” prior to banking crises. Short-term debt also comes out as one of the best predictors of currency, debt, and banking crises according to the signals EWS approach. However, the fact that available cross-country data on short-term debt do not distinguish public from private sector debt limits the usefulness of the result.
There is little evidence that expenditure and revenue variables help to predict crises. None of these variables adds to the predictive power of the IMF’s main EWS model, and the event studies fail to find any expenditure or revenue variables, which differ systematically in the run-up to crises compared with noncrisis periods.
Can Fiscal Variables Explain the Severity of Currency Crises?
Using the standard approach in the literature, some deficit and financing variables are found to be weakly correlated with the severity of currency crises. However, this correlation is not robust, being dominated by other variables. Using a panel data approach, the picture changes somewhat. Deficit, debt, and interest expenditure variables are important for explaining changes in the foreign exchange market pressure index, in both crisis and noncrisis periods. This provides more support for the conclusion noted above, namely that fiscal variables help explain relatively tranquil periods better than crisis periods. It also suggests that there is an important indirect impact of fiscal policy on currency crises, operating through the real exchange rate appreciation.
What Are the Fiscal Consequences of Crises?
The event studies show that the deficit declines on average in the two years following currency and debt crises. Given that crises will typically result in a slowdown in growth, an improvement in the conventional deficit implies discretionary fiscal tightening, on average, following currency and debt crises. This will often be inevitable when a crisis hits. However, there is some evidence that the situation is different following banking crises, with some deficit variables pointing to fiscal loosening following a crisis. The appropriateness of a fiscal loosening will depend not only on the type of crises, but also its underlying cause, the precrisis fiscal position, and the response of the domestic economy and international investors to the crises.
On average, the debt-to-GDP ratio increases significantly following all types of crises. While there is some evidence of a reduction in short-term debt following currency and debt crises, foreign debt remains fairly stable in the two years following crises. The case studies show how these averages can conceal wide variations; indeed in 4 of the 11 cases, debt fell following the crisis, due to a combination of debt restructuring, fiscal adjustment, and real exchange rate appreciation.
The event studies suggest that total expenditure changes little, on average, following currency and debt crises, but remains significantly above the tranquil period average in the case of banking crises. Consistent with the increase in total debt stocks following crises, interest expenditure rises in the years following currency and debt crises, offset by lower social expenditure. The event studies imply that the big falloff in revenue experienced by some Asian crisis countries during 1997–98 is not typical of most emerging market economies following financial crises.
Finally, the case studies demonstrate that crises can catalyze significant structural fiscal reforms. In particular, they allow political barriers to reform to be overcome. Clearly, if crises clear the way to undertake much-needed structural reform, full advantage should be taken of this opportunity.