Friday, December 22, 2017

Why is the business cycle elongating?

First, let me wish all my Econoblog readers a Merry Christmas and a Happy New Year!!

Economists have traditionally put the business cycle at between 3 to 8 years long.  But of the last 4 business cycles 2 out of the last 4 have lasted longer than the 8 year limit that economists typically look at.  In fact counting this business cycle, which officially hits 10 years (or 120 months) in December 2017, 3 out of the last 5 business cycles have had periodicity longer than the 8 year economist "consensus" upper limit.

For those geeky enough to be interested in US business cycles see the table below which is lifted from the NBER website.  The longest expansion we have seen in the US economy was through the 1990s through until the tech stock bubble in 2001, which lasted a full 128 months, 8 months than where we are right now.
Source: NBER website

But this then begs another question.  What are the specific reasons as to why the current cycle would be elongated, and how persistent will these effects be in preventing us from entering the contractionary phase of the business cycle?  Well there are several reasons why I believe economists and economic commentators think that we might have an elongated cycle this time.  I will run through each of these reasons below, but to summarize these reason up front:

1.  We started the expansionary phase from a lower base, as the "great recession" of 2007-09 was more severe than all previous downturns with the exception of the "great depression";  

2.  Quantitative Easing (QE) has provided an additional stimulus which combined with the usual countercyclical fiscal policy, allowed the economy to achieve escape velocity, but QE is only now being unwound;

3.  The tax reform bill just passed by the Trump administration, plus the Infrastructure spending bill that the Trump administration has promised in the first half of 2018, will continue the fiscal stimulus for the economy through at least 2018, and possibly to the end of 2019; and

4.  That the "great moderation" which started in the 1980s, has seen a dramatically lowering in volatility for short term cycles in growth, but due to a couple of reasons, this cyclical volatility has transferred to longer cycles in growth which, for the moment implies that the usual business cycle frequency of 3 to 8 years quoted by economists is now incorrect.

So let's start with the first reason.  The main insight here comes from the Great depression, and the fact that when a macroeconomy experiences a really deep recession, where the financial sector is involved, the recovery will be slow and arduous.  The accompanying chart from an IMF publication shows this quite clearly.  The horizontal axis shows the number of quarters into the recession and the reaction of various economic variables (averaged across countries and across time). So for example, residential investment starts to recover after around 4 quarters for non-financial recessions, but for financial recessions that recovery starts 11 quarters after the beginning of the recession.  As the great recession was caused by both the housing market and the financial markets, the recovery pattern has clearly been slower than for other recent recessions in the US. This is also clearly seen in GDP growth itself, which is shown below.  The most recent recession is the first recession since the Great depression where the economy was recovering from a financial recession and it is clear from the rate of growth coming out of the recent recession that the economy has had difficulty growing above roughly a 2% rate.

Source: BEA and authors calcs
While the logic of this argument appears sound, there are a couple of things to notice about the nature of the reasoning here.  First, the fact that the great recession was a financial recession would tend to suggest that the business cycle elongation will be only relevant for this current recovery rather than for business cycles in general. If this reason is correct, then the macroeconomy should return to its usual 3-8 year cycle after the next recession.  Second, it also flies in the face of business cycle dating that we referenced above - the business cycle has definitely been getting longer, and has not remained within the usual 3-8 year periodicity that economists so often cite.

The second reason as to why the current business cycle will be elongated is monetary policy. The amount of stimulus provided by central banks has continued to rise albeit at a slower pace.  As the graph below shows, the US is now reversing it's massive QE program, and that is one of the reasons why rates are rising in the US relative to rates elsewhere.  If we look at the chart below, we can see that indeed Global QE is still rising, mostly thanks to the ECB, who have still not started to taper.  That in itself is a massive boost to financial companies around the world as central banks have brought large amounts of financial assets off the commercial bank balance sheets, thus freeing up capital to be lent elsewhere, and stabilizing balance sheets.  If we look at this in terms of the rate of change of QE globally, we come up with a different impression, which is given by the chart below ( - please ignore the forecasts of a research group which were made in 2015).  These liquidity injections clearly have largely dissipated for most central banks, but net injections are still reversals have not yet been substantial enough to make an impact on the total, and some central banks are still continuing their QE buying programs.

Note that this reason would also only imply a temporary one-time elongation of the business cycle, and so doesn't explain why the most recent business cycles appear to have been on a elongating trend.

The third reason is due to the recent US fiscal stimulus in the form of the tax reform and the possible infrastructure package that President Trump has promised next year. This will affect the US, but does come with likely additional public debt implications, which will tend to crowd out investment and in normal circumstances would drive interest rates up.  But the tax reform essentially increases the return on US investment (purchase of plant, machinery and equipment), which will tend to increase private investment, thereby offsetting the crowding out effect.  So on balance, with the individual income tax reductions, these corporate tax changes should further stimulate the US economy, bolstering the monetary policy argument above. Once again, this is a one-time effect, and does not explain the lengthening of the business cycle.

The last reason why we might be seeing an elongation of the business cycle can be explained by recent research that I have been doing with Professor Andrew Hughes Hallett of George Mason University.  The empirical argument is shown in the figure below.

This analysis is called a "multiresolution decomposition" or MRD, and the technique essentially extracts the processes embedded within the series over different frequency ranges which are represented by the series d1 to d5 which are shown in the figure. We have 2 papers, the first of which showed statistically that the longer cycles embedded in real GDP growth (shown by d5 and d6 above (which relate to cyclical activity ranging from above 8 -16 years and from 16 - 32 years respectively) have become more volatile since the early 1980s, while the higher volatility cycles in real GDP (shown by d1 to d4, corresponding to cycles from 2 quarters to 8 years), have become less volatile [see below for academic references].

The second paper, which has been published as a discussion paper by the central bank of Finland (Suomen Pankki) [again see below for academic reference], goes through a lengthy analysis of the theoretical models typically used by macroeconomists to show the factors that could potentially cause this lengthening of the business cycle.  To cut a long story short, the factors that could be shifting volatility in the process that drive economic growth from shorter cycles to longer cycles turn out to be i) an increase in inflation aversion; and ii) a reduction in output stabilization. So let us look at each of these parameters in turn.

Has there been an increase in inflation aversion moving from the pre-mid-1980s period through to the post-mid-1980s period?  I would assert that yes, there has been, and this is due to the fact that many central banks instituted inflation targeting and if he could have done so, we know that ex Fed Chairman Ben Bernanke would have done so.  So has there been a reduction in output stabilization?  That is, has there been a reluctance to fully engage fiscal policy to it's maximum effect during downturns and to offset any rapid growth in the economy?  I think the evidence, once again, is that yes, we are seeing less output stabilization in US fiscal policy for certain, and perhaps a little more emphasis on stabilization by the Fed.  The net effect though would still be for less emphasis on output stabilization.  Now why do I assert that this is the case?  I think the evidence has been on show during the last week in the US.  As we know we are entering the final stages of the business cycle, the Trump administration has effectively announced a tax stimulus package which then will cause a spurt in growth as well as a one time elongation of the business cycle.  This tax reform package is definitely not output stabilization in the classic sense of counter-cyclical fiscal policy.

So note here that this fourth explanation would help to explain a permanent elongation of the business cycle since the mid-1980s.

But what about the markets in all of this?  One of the best visualizations I have seen relating to business cycles and the stock market came in a piece of research out of Goldman Sachs in late November this year (see below).

The figure shows that we are now approaching the 9th year of a bull market, with no signs of any correction coming.  This is not quite the record run yet, but it is fast approaching the 9.1 years of the 1920s bull market.

My own feeling about the financial markets is that we are beginning to move into "borrowed time", and that as soon as these one-time stimulus factors have passed, the downturn will happen. Whether that is in late-2018, 2019 or 2020 I am unsure.  But if there is one thing I am definitely sure of it is that the next downturn is coming sooner or later.


Crowley, P. and Hughes Hallett, A. (2015), "Great moderation or “Will o’ the Wisp”? A time–frequency decomposition of GDP for the US and UK", Journal of Macroeconomics, Vol 44, pp82-97.

Crowley, P. and Hughes Hallett, A. (2014), “Volatility transfers between cycles: A theory of why the "great moderation" was more mirage than moderation”, Bank of Finland Discussion Paper 24/2014.

Featured Post

Free Trade on Trial - What are the Lessons for Economists?

This election season in the US there has been an extraordinary and disturbing trend at work: vilifying free trade as a "job kille...

Popular Posts

Search This Blog