What happens if countries in a monetary union set ambitious goals for budget deficits and growth? They start making overly optimistic forecasts. This paper by Jeffrey Frankel and Jess Schreger from Harvard has some informative graphs. The first graph below shows how the two year ahead forecasting error differs per country. Greece, Ireland, Portugal, and Spain come out on top. The second graph shows how the average forecasting error varies over time. The forecasting error dropped just before the introduction of the euro and rose soon after.
The authors then go on to show that “euro area governments making forecasts while in violation of the EDP with an independent fiscal institution that makes independent budget forecasts have a mean bias that is smaller by 2.9% of GDP at the one-year horizon and 2.6% of GDP at the two year horizon, compared to a euro area country violating the EDP without such an independent fiscal institution.” I rest my case.
The most relevant economic studies do not necessarily coincide with those published in top economic journals. It is easy to miss a great study. This paper by the Sverigse Riksbank is a point in case. It was released in december 2011 and I only saw it yesterday. It tries to answer a crucial question I’ve discussed on this blog here and here (in Dutch) – how much bank capital is optimal from a welfare point of view. The answer: somewhere between 10 and 17 percent. This is in line with previous studies that try to tackle this nasty question, as shown in the table below.
But what is really interesting is the following graph. It shows the effect on gdp of an increase in capital ratios for the various studies, starting from an initial level of 6 percent. Clearly, when capital ratios are too low the benefits of increasing them is very large: the curve increases steeply for low capital ratios. Remarkably, however, it flattens out when capital ratios reach the optimal level. This implies that the costs of setting the capital ratio too high (i.e. at 20 percent if 15 percent is optimal) are much smaller than the costs of setting the capital ratios too low (i.e. at 10 percent instead of 15 percent).
Let me put this differently. In all studies considered here, bluntly setting the capital ratio at 20 percent results in only a very minor loss relative too choosing the optimal level, wheras it results in a significant benefit when compared to the initial level of 6 percent. So why worry about the optimal level of bank capital if you can just set it at 20 percent?
An important policy question is: how big should the financial sector in a country be. Can there be too much finance? A recent IMF working paper by Jean-Louis Arcand, Enrico Berkes, and Ugo Panizza suggests there is a threshold above which financial development no longer has a positive effect on economic growth. They find that finance starts having a negative effect on output growth when credit to the private sector reaches 100% of GDP. Below is a graph of the effect of credit to the private sector on growth that shows this inverted U-shape relation between credit to the private sector and growth.
And here are the countries that are just below or above the optimal level
Proposals for eurobonds abound: Delpla and Von Weizsacker, Hellwig and Phillipon, Brunnermeier and others , German council, Boonstra. Most aim to shield countries from sovereign debt crises. Indeed, a common observation is that Europe as a whole has a lower debt to GDP ratio than the United States or Japan. At a more abstract level, eurobonds are one answer to the following question: how can eurozone countries create some form of mutual insurance while stopping short of full-fledged political integration of federative states like Germany or the US. But what do we know about the economics of eurobonds? Eurobond proposals need to deal with four issues: moral hazard, externalities, contagion and credibility.
A central issue in designing optimal insurance schemes is how to address moral hazard. Eurobonds insure against events that raise interest rates. In particular, they insure against both liquidity risk and solvency risk. Most economists believe that solvency risk is at least to some extent driven by economic policy. Thus, insured countries can make a costly effort to reduce the risks they are insured against by reducing debt (e.g. by cutting spending), increasing growth (e.g. by labor market reforms or reducing entry barriers in product markets), or increasing resilience to shocks (e.g. by strict banking regulation or limiting private sector debt). When such effort is unobservable or noncontractible, this raises the issue of moral hazard.
In the presence of externalities, it may be difficult to agree upon the optimal amount of insurance. Suppose problems in one country have significant negative consequences for another country, but vice versa the impact is much less. In that situation, the two countries will want different levels of insurance. The optimal level of insurance will lie somewhere in between the two levels. To get the countries to agree might be difficult if transfers are not possible.
Suppose two countries have mutually insured themselves. This implies that shocks that raise one countries’ interest rate will now also raise the other countries’ interest rate as well. It is in principle possible that a shock that would pull only one country down now pulls both countries debt levels into unsustainable territory. The figure below illustrates this in a simple setting. Suppose countries without mutual insurance go bankrupt if the shock exceeds some level A. On the right hand side the region where both countries survive is smaller than on the left-hand side. These are benefits of insurance. The area where both countries fail together, however, is larger on the left-hand side. These are costs of mutual insurance.
Finally, insurance should be credible. When a country could withdraw from the insurance scheme if that were desirable for the country in question, this would undermine the effectiveness of the scheme. It may happen that while insurance is optimal for a country ex ante when it is not known who will be hit by the shock, ex post once shocks have occurred it may not be optimal for a country to pay up.
When will the big bad market come: if a country implements too little austerity, or if economic growth falters? Soon after the fall of the Dutch government (because the freedom party withdrew support for the minority government), the Dutch Central Bank published on its website a short analysis of Dutch – German bond spreads. They argued that the spreads increased as a direct consequence of the fall of the government. The supposed conclusion: the government budget should be put into order as soon as possible. Here’s the Dutch-German bond spread since 1987. It looks pretty high.
But the claim raises two questions. First, did the spread between Dutch and german interest rates really react to the news or was the larger spread just coincidence? In reality, bond prices jitter around all the time. What are the chances that the movement after the fall of the Dutch government were due to accident? The graph below shows the change in the spread between the 10-year Dutch and German treasury bonds and probability distribution associated with those changes. If we look at changes outside a 99.5% confidence interval, the day the Dutch cabinet fell, the 23rd of april, with a jump in the spread of 0.155 classifies as a significant event. The very next day, by the way, prices dropped by 0.136 percent.
The second question is what did markets actually react to? Did they react to a potential worsening of government budget or to a potential deterioration of growth prospects? In her presentation for the IMF on lessons from the crisis for fiscal policy, Christina Romer provided an interesting analysis for Spain. She made an overview of the ten largest increases in Spanish interest rates and coupled them to the information arriving in the market that caused these interest rate hikes. She concludes that news on a lack of fiscal consolidation as well as news on deteriorating growth prospects were associated in roughly equal proportions with the movements in interest rates.
In the period after 1-1-1999 there were 20 days with upward or downward jumps in the interest rate that were equally significant. Here’s the list.
The conclusion seems to be that Dutch bond prices did indeed increase after the government fell. To attribute jumps in spreads to a lack of austerity measures, however, is a bridge too far. Bonds spreads might just as well have reacted to a perceived drop the probability of reform measures being enacted. Or to a temporary increase in uncertainty. Who knows.
A recent working paper by Alfredo Martín-Oliver, Sonia Ruano and Vicente Salas-Fumás from the Spanish Central Bank nicely illustrates how equity of (Spanish) banks slowly melted away over the past two decade. In 1992 Spanish banks held on average 12% equity capital. In 2007 only 5% remained (top graph). Unsurprisingly, the solvency ratio – which involves risk-weighted assets – dropped much less, from somewhat more than 17% in 1992 to roughly 14% in 2007 (bottom graph). The ratio of risk-weighted assets to total assets has been dropping steadily over the past two decades for banks from all countries, see my earlier post on this paper.
Interestingly, the paper also contains a lot of information on the distribution of equity, showing that bank equity levels have been converging. In 1992 the gap between banks with 10% highest equity level and banks in the 10% lowest categorie was more than 15%. In 2007 this gap was reduced to roughly 6%.