Some light end of summer reading.
Cecchetti, Kohler & Upper / Bank for International Settlements:
Financial Crises and Economic Activity [36 page .pdf]-- Abstract:
This paper studies the length, depth and output costs of a sample of 40 systemic banking crises in 35 countries since 1980 to assess the likely real impact of the current crisis. Most, but not all, systemic banking crises in our sample coincide with a sharp contraction in output from which it takes several years to recover. The current financial crisis is unlike any others in terms of initial conditions, industrial and institutional structures, levels of development, degrees of openness, policy frameworks and external conditions. Simply averaging outcomes of past crises to get a reading on the current one is therefore likely to be misleading regardless of the sample or subsample. With this in mind we go on to study the determinants of the output losses from past crises.
Our findings suggest that the costs are higher when the banking crisis is accompanied by a currency crisis or when growth is low immediately before the onset of the crisis. Furthermore, when it is accompanied by a sovereign debt default, a systemic banking crisis is less costly. The final part of the paper takes a longer-term view and study the impact of crises on potential output several years down the road. We find that many systemic banking crises have had lasting negative effects on the level of GDP. And even in those cases in which trend growth was higher after the crisis than it had been before, making up for the output loss resulting from the crisis itself took years.
August 2009 marks the second anniversary of the start of the first global financial crisis of the 21st century. World output has experienced its sharpest drop since the Great Depression of the 1930s, with most economies contracting in late 2008 and early 2009. The severity of the crisis has surprised nearly everyone....
Our objective here is not to explain the causes of the current crisis. Instead we study the consequences...
Cluster analysis allows us to assign sets of observations to subsets that are similar, given a set of characteristics. Apart from the choice of the characteristics, however, the technique allows the investigator to remain agnostic. It groups observations into clusters by minimising differences within clusters and maximising differences across clusters. Cluster analysis is widely used in quantitative social research to analyse datasets with a large number of variables. For example, it allows firms to group clients that may be receptive to particular ways of marketing, and it helps authorities sort through immigration files in their hunt for terrorists.
The results of the cluster analysis are represented in a dendrogramme (Graph 7)...
One of the most striking conclusions we draw from this way of looking at the data is that
current events are unique. While some of the countries suffering from the current crisis cluster fairly close together – the United Kingdom and the United States appear close – they are very dissimilar from all other episodes.19 In fact, the cluster analysis joins the current crises only after almost all previous crises have joined. The implication is that the current crisis is less similar to all of the crises in our database than, say, the Japanese financial crisis of the 1990s is to the crisis experienced by Ecuador in 1998 or than it is to the crisis that occurred in Bulgaria during the transition!
The uniqueness of the current crisis is an important, if discouraging, result...
Can we predict the length and depth of the current crisis given historical experiences? The analysis in the previous sections suggests that crises are very dissimilar and that the current financial crisis is especially different from those that have come before it. As a result, it is difficult to use (unconditional) average past experience to draw conclusions about how deep and long the current contraction of the real economy is likely to be...
Saturday, September 5, 2009
Some light end of summer reading.