A Financial-Fiscal Trilemma

Financial crises and sovereign debt crises are, of course, not a new phenomenon. But the strong connection between fiscal crises and financial crises is relatively recent, primarily developing since the Great Depression and especially since the 1980s. In a new and ambitious NBER working paper, Michael Bordo and Chris Meissner survey the literature on financial and fiscal crises and their interconnections, providing both a history of thought and a catalog of open questions.

The key to the growing link between fiscal and financial crises, they explain, is the increased use of government guarantees of financial institutions. This means that banking crises are often followed by a rise in the debt-to-GDP ratio that can be partially attributed to costs of reconstructing the financial sector. Based on a synthesis of the research in this area and some preliminary empirical analysis, Bordo and Meissner posit that countries face a “financial/fiscal trilemma.” As they explain:

This financial/fiscal trilemma suggests 43 that countries have two of the following three choices: a large financial sector, a large bailout package, and a strong discretionary reaction to the downturn associated with financial crises. The logic is as follows by way of an example. Assume a country with a large financial sector faces a banking crisis. If so, then the government can provide a bailout package of a size that is commensurate with the size of the financial sector. If so it uses up its fiscal space. Otherwise it could lower the size of the bailout and devote its fiscal space to discretionary fiscal policy. With a smaller financial sector, and the same amount of fiscal space, since the size of the bailout would by definition be smaller, the size of the rise in debt due to expansionary policy could rise (p. 42-43). 

They use data from Laeven and Valencia (2012) on 19 systematic banking crises to estimate the equation:

Fiscal costs refer to the fiscal costs of bailouts in the three years following a crisis. Discretion is the change in debt not due to the fiscal costs of bailouts, also in the three years following a crisis. The estimation results, with standard errors in parentheses, are:

Notice that the estimated coefficients on the fiscal cost and discretion to GDP ratios sum to approximately 1, suggestive of a tradeoff. If the financial sector is smaller, or if the bailout package is smaller, then the change in fiscal costs to GDP ratio is likely to be smaller, which could allow a larger change in the discretion to GDP ratio, hence the "trilemma." The trilemma is illustrated by Figure 5, below. The discretion to GDP ratio is on the y-axis and the fiscal costs of bailout to GDP ratio is on the x-axis. For a given change in the debt to GDP ratio, the regression estimates imply an "iso-line" showing the fiscal costs of bailouts and discretion to GDP ratios that are possible.

Source: Bordo and Meissner (2015)

As further evidence of the trilemma, they present Figure 6, which illustrates that countries with a larger financial sector, as measured by the domestic credit to GDP ratio, tend to have a larger rise in the share of the debt to GDP ratio explained by bailouts.

Source: Bordo and Meissner (2015)

This evidence of a new "trilemma" certainly merits more rigorous empirical evaluation. As the authors note, however, empirical studies of financial and fiscal crises face the challenge of inconsistent classification and measurement. Alternative crisis chronologies lead to contradictory results. Bordo and Meissner thus propose the following:

If economists and policy makers truly believed that crises were an important phenomenon to understand and possibly avoid then it might be the case that an independent crisis dating committee could help set the standard in much the same way the NBER business cycle dating committee works. The advantage of following this model is that the NBER is a respected non-governmental, non-partisan organization. Other organizations such as the IMF are not sufficiently politically independent. If crises are becoming increasingly global and crisis fighting is a global public good, then the importance of such a reform should be obvious. 

Financial Networks and Contagion

"Financial Networks and Contagion," a recent paper by Matthew Elliott, Benjamin Golub, and Matthew Jackson, uses network theory to study how financial interdependencies among governments, central banks, investment banks, and other institutions can lead to cascading defaults and failures.

Source: Elliott et al. 2013

While the model is quite technical, the main theoretical findings are fairly intuitive. They define two key concepts, integration and diversification. Integration refers to the level of exposure of institutions to each other through cross-holdings. Diversification refers to how spread-out the cross-holdings are; in other words, whether a typical organization is held by many others or just a few. The key finding is that at very low or very high levels of integration and diversification there is lower risk of far-reaching cascades of financial failures. The risk of a far-reaching cascade is highest at intermediate levels of integration and diversification. The authors explain:

"If there is no integration then clearly there cannot be any contagion. As integration increases, the exposure of organizations to each other increases and so contagions become possible. Thus, on a basic level increasing integration leads to increased exposure which tends to increase the probability and extent of contagions. The countervailing effect here is that an organization's dependence on its own primitive assets decreases as it becomes integrated. Thus, although integration can increase the likelihood of a cascade once an initial failure occurs, it can also decrease the likelihood of that first failure... 

With low levels of diversification, organizations can be very sensitive to particular others, but the network of interdependencies is disconnected and overall cascades are limited in extent. As diversification increases, a "sweet spot" is hit where organizations have enough of their cross-holdings concentrated in particular other organizations so that a cascade can occur, and yet the network of cross-holdings is connected enough for the contagion to be far-reaching. Finally, as diversification is further increased, organizations' portfolios are sufficiently diversified so that they become insensitive to any particular organization's failure."

Near the end of the paper, they illustrate the model using cross-holdings of debt among six European countries. The figure above is their representation of financial interdependencies in Europe. They conduct something akin to stress tests, simulating cascades of failures under various scenarios that very roughly approximate conditions in 2008. The simulations find that, following a first failure in Greece, Portugal is fails from contagion. After Portugal fails, Spain fails due to its large exposure to Portugal. The high exposure of France and Germany to Spain causes them to fail next in most simulations. Italy is always last to fail due to its low exposure to others' debt. They emphasize that this is intended only as an illustrative exercise at this stage, but could eventually be refined and incorporated into analysis of failure and contagion risk.

*Edited to fix my mistake pointed out by Phil.

Cyprus Levy: Historical Precedents

In 1991, in response to considerable debt management concerns among European nations, Barry Eichengreen wrote a paper called "The Capital Levy in Theory and Practice." A capital levy on wealth-holders with the goal of retiring public debt was perhaps the most controversial proposed solution to the European public debt problem. Eichengreen provides a theoretical framework for considering the effects of such a levy, a list of challenges to successful implementation, a "catalog of failed levies," and an example of an exceptionally successful levy. Eichengreen's framework and lessons from history are useful for considering the levy on savings in Cyprus that is planned as part of Cyprus' 10 billion euro bailout.

Eichengreen's model uses insights from the literature on time consistency and government reputation. If governments could commit to only issuing capital levies under certain circumstances (in technical terms, if capital taxation is a state-contingent claim with full commitment mechanism), then governments could optimally issue levies when government obligations are unusually high, social returns to spending are unusually high, and/or conventional revenues are unusually low. "If the contingencies in response to which the levy is imposed are fully anticipated, independently verifiable, and not under government control," he writes, "then saving and investment should not fall following the imposition of the levy, nor should the government find it more difficult to raise revenues subsequently."

Of course, there are major practical impediments  First, full government commitment technology does not exist. If a government were to precommit to a plan for how its capital taxes would respond to future contingencies, it would end up having an incentive to renege on its commitment. Knowing this, savers would shift their capital to tax havens. Reputational concerns can partially get around this problem. The model describes how a "reputational equilibrium" can result if savers refuse to repatriate their capital for a certain amount of time following a government's decision to renege on its commitment. Another major impediment is the following:

"In a democracy, there is no independent authority to verify to the satisfaction of savers that the realization of the state of the world in fact justifies a capital levy. Even if savers recognize that capital taxation is a contingent claim, they retain the incentive to dispute that the relevant contingency has arisen. Neither is there a mechanism to prevent the government from pursuing policies that strengthen the case for capital taxation -- for example, increasing ordinary expenditures as levy receipts roll in. Savers will accuse the government of succumbing to moral hazard and resist the levy on those grounds... If the levy is imposed at all, typically this will occur only at the end of a protracted and divisive political debate...If there is an extended delay between proposal and implementation of the levy, capital flight is likely to render the measure ineffectual."

In light of these impediments to the successful use of a capital levy, Eichengreen analyzes a number of previous levies, mostly in the aftermath of World War I, and explains why they obtained varying degrees of failure or partially success. His first example, however, is not from post-WWI but rather from the ancient Greeks. They used periodic capital levies of one to four percent which, it is said, were "phenomenally successful because property owners, out of vanity, overstated the value of their assets!" Capital levies were also proposed, but not adopted, following the Napoleonic wars, the Franco-Prussian war, and other periods of major military expenditure. But the end of WWI was the true hey-day of the capital levy.

Italy imposed a capital levy in 1920. The rates ranged from 4.5 to 50 percent; however, payment could be stretched out over 20 years. Thus, it was successful but not too comparable with the Cyprus levy, which would be paid at once. The Czech levy of 1920 is more comparable with the Cyprus levy, and was more successful than levies in Austria, Hungary, and Germany at that time. In Czechoslovakia, a levy on all property was imposed, with progressive rates from 3 to 30 percent, with a separate surtax on the wartime increment up to 40 percent. There are two main reasons this levy was relatively successful. First, the levy fell mainly on a small ethnic German minority, which was unable to mount effective political resistance to delay adoption, so capital flight was minimized. Second, the government budget was structured so that levy revenue was completely separate from day-to-day government operations. The levy was explicitly devoted to extinguishing debts and meeting the special costs of establishing a newly-independent nation. This "lent credibility to claims that the levy was an extraordinary tax whose repetition was unlikely."

Other countries' levies were less successful. In Austria, political delays dragged on so long that asset holders had more than a year to prepare, so capital flight was extensive; moreover, hyperinflation liquidated levy obligations. Similarly, delays in other countries enabled capital flight. In Britain, a levy was long debated but never imposed. Keynes himself weighed in, initially in support of the levy, but later opposing it by the mid twenties, when postwar exceptional circumstances were further in the past.

The Japanese levy after World War II is Eichengreen's example of the exception that proves the rule. It was successful because the typical impediments to success were negated by the exceptional circumstances. Capital flight was limited, and the levy applied primarily to a small minority of individuals considered to have profited greatly from the war. In 1946-47, Japan's sovereignty was severely abridged by occupation forces. Thus "with important elements of democracy in suspension, the levy could be quickly and effectively implemented. One might go further and argue that only when political sovereignty is suspended can the measure be pushed through with such alacrity." Also, since the levy was essentially imposed by outsiders, it did not damage the Japanese government's reputation too much or adversely impact the ability of the Japanese government to raise revenue subsequently.

Lessons for Cyprus

A common theme of levies in the 20th century is that delays, usually politically-induced, cause capital flight and prevent the levy from raising a successful amount of revenue. The Cypriot parliament has already delayed its vote on the deposit levy to Monday. We will see how long the proceedings drag on. In Cyprus,  the levy will fall in large part on foreign depositors, particularly Russians, like the Czech levy fell mainly on Germans. It is not clear whether foreign depositors will prove as unable to mount political resistance in Cyprus as the Germans were in Czechoslovakia.

The levy in Cyprus is progressive, like it was in Czechoslovakia, although not to the same extent. The rate is 6.75% on deposits less than 100,000 euros and 9.9% on larger deposits. According to Eichengreen, a crucially important feature of the successful Japanese levy was the "sociopolitical argument," or the fact that the levy was imposed on people who were seen as having profited unjustly from the war. Engineers of the Cypriot levy may make the case that it is imposed on Russian money launderers or other wealthy depositors funneled their money there for less than admirable reasons. This case might be easier to make if the levy were more progressive-- if it hit bigger accounts harder, and were easier on the small accounts.

The Japanese levy did not adversely impact future Japanese governments' ability to borrow because it was imposed by outsiders. The Cypriot government may also be able to point to the IMF and EU as the masterminds behind this levy. The bigger issue, though, is whether other EU countries will suffer reputational penalties by virtue of association.

Finally, what worked really well for Czechoslovakia was that the levy revenue was totally separated from other tax revenue and dedicated strictly to paying off debt and to the needs of the extraordinary circumstances. It was not used to fund day-to-day operations of the government. That made it seem less likely that the government would be tempted to impose another levy in the near future. That is a feasible option for Cyprus, which they will hopefully employ.