Monetary Policy In A Weak Economy

Speech given by Martin Weale, External Member of the Monetary Policy Committee, Bank of England

At the National Institute of Economic and Social Research, London
25 November 2011


Even before the recent problems associated with the euro area, most countries were experiencing a disappointing recovery from the aftermath of the recession which began in 2008. In the United Kingdom things have been particularly slow. In this speech I would to first like say rather more about our own recent experience and some of the factors which make the last few years so different from our earlier experiences of business cycles. I will focus particularly on the unusual behaviour of both productivity and consumption.

Secondly I would like to discuss what the Bank has been doing to support the economy. I will then broaden this to discuss some of the other proposals which have been suggested and reach conclusions about the evolution of aggregate demand compatible with a well-balanced and sustainable recovery.

The UK Recovery in a Historic and International Context. Perhaps I could begin by putting some flesh round the observation that we are experiencing the slowest economic recovery since the First World War. There have been six recessions since 1920 and, working with colleagues from the National Institute and Cambridge University, it has proven possible to construct monthly profiles of these. Those for the recessions between the two wars are interpolated using the monthly activity indicators collected by The Economist during the period (Mitchell, Solomou and Weale, 2009). The post-war recessions are interpolated from quarterly data using manufacturing output and retail sales as monthly indicator variables (Mitchell, Smith, Weale, Wright and Salazar, 2005).

In Charts 1 and 2 I show the time profiles of these, measuring GDP in each month (as a three-month moving average) relative to GDP at the start of the contraction. In interpreting these charts we need to remember that all of the data are uncertain and that the most recent data are at particular risk of significant revision.

We can see from the charts that there have been two relatively mild recession/recovery periods, those of the mid-1970s, associated with the oil crisis, and the early 1990s and four recessions which were similar in their depth, those that began in 1920, 1930, 1979 and 2008. Nevertheless the first pre-War recession was slightly deeper than the more recent recessions. Its trough was over 9% below the pre-recession peak. 1 The 1932 trough, like the trough in the current recession, was about 7% below peak with the recession in 1979 being milder. But we can also see that, for the three completed cycles, the time needed for output to recover to its pre-recession level was no more than four years and one month.

We might now ask how long it is likely to be before output regains its level of early 2008. The central estimate of economic growth associated with the MPC’s latest forecast, published in the Inflation Report, shows this happening in the third quarter of 2013. Of course this is no more than the centre of a range of dates implied by our fan-chart, but it suggests a cycle lasting more than twelve months longer than its predecessors. In other words, in terms of duration if not depth, our forecast suggests that this will be the worst of the cycles for which we can produce more than annual indicators.

At this point a digression is helpful. Why do we refer to the 1930s cycle as the Great Depression if it was shallower than the 1920s cycle and slightly less protracted than the cycle around 1980? The answer is because other countries experienced much sharper contractions than we did. Over the period output fell by
between a quarter and a third in Austria, Germany, Canada and the United States as compared to our figure of around 7%. The United States did not regain its output peak of 1929 until 1936 – in other words there output was depressed below its peak for about seven years as compared to our four years and one month. What can be learned if we look at the experience of the other major advanced economies? Might our weak recovery be associated with the magnitude of the recent banking crisis? In Chart 4 we can see the paths that the G7 countries have followed since the start of the crisis. Canada is the obvious winner, but then there are two groups. The United States, Germany and France, at least so far, have shown recovery paths close to the top of what might be expected following major banking crises. In contrast the United Kingdom, Italy and Japan have performed near the bottom of the range of outcomes indicated by past experience. The
crisis in the United Kingdom was probably not more severe than in the United States while Italy and Japan, for all their other problems, did not have banking crises (Barrell, 2010). And the period I am considering finishes before the problems of the euro area started to affect Italy. So, on this basis too, UK experience is not easy to explain. So, seen from the perspective of movements in GDP, the Great Depression merits a special label because of its impact on these and other countries and not because of its impact on the United Kingdom.

Does it become any easier to understand if, instead of looking at averages of earlier recessions, we try to account statistically for the lengths of those earlier recessions on the basis of the circumstances which gave rise to them? There has, of course been a wide range of studies looking at influences on the depth and duration of recessions and the speed of economic recoveries. These studies typically either examine the probability of a recession taking place or, conditional on that happening, explore some other characteristic such as its depth, its duration, or the variable of particular interest to me today, the time taken for output to recover to its pre-recession peak.2 Ideally one should examine jointly the probability that a recession occurred and the time taken for output to regain its previous peak. To the extent that other discrete events are involved, such as the presence or absence of financial crises, one should also look at the determinants
of such events rather than treating them as exogenous.

A study by the IMF suggests that the mean time taken for output to fall during a recession and then regain its previous peak, an interval which I subsequently refer to as a period of depressed output, is 6.8 quarters, rising to 11.3 quarters if the recession is associated with a financial crisis (IMF, 2009). So all of the completed periods of depressed output in the UK have been much longer than the average for developed countries. Indeed they have all been longer than the average for periods of depressed output associated with financial crises. Even if one looks at periods of depressed output associated with financial crises that are highly synchronized, the mean duration is 14.1 quarters, while for those associated with major banking crises (which might be considered more comparable to our recent experience than the wider range of crises covered by the IMF work) the mean duration is 13 quarters. Fourteen quarters have already elapsed since the start of the current period of depressed output. So four of the six phases since 1920 will have lasted longer than the worst mean case identified by the IMF, associated with highly synchronised financial crises.

The unusual length of such periods in the UK can be seen in the chart below. This looks at the distribution of the duration of periods of depressed output. The periods associated with the post-war UK recessions are all in the tail of the distribution, as are the recent periods associated with the financial crisis.

Graphs and full account may are available on Bank of England website.