Monday, April 20, 2015

The similarities between 1990 and 2007

Benjamin Cole recently unearthed a 1992 piece from Milton Friedman in the Wall Street Journal, complaining that Fed policy was too tight.  I do think there are an interesting number of similarities between the recoveries after 1990 and 2007.

Source
Between 1989 and 1996, Price/Rent ratios declined by about 15% (and more like 23% in the Case-Shiller 10 city index).

The delinquency rate on real estate loans reached more than 7% in 1991.  The delinquency rate on real estate loans in the recent crisis didn't outpace that mark until the second quarter of 2009.  (By that time, home Price/Rent values had plummeted by 25% or more, depending on the price index you use.  It has become fashionable, in hindsight, to blame irresponsibility of households and banks for the recent crisis.  But, this measure suggests that, compared to 1990, both households and banks were more stable.  Home values had to fall much farther and more steeply in the recent crisis than they did after the 1990 recession before delinquencies and bank failures reached similar levels.  Higher inflation meant that the delinquencies in 1991 came without any notable drop in nominal home prices, while nominal home prices by early 2009 had fallen by more than 20% by the time delinquencies topped the 1991 level.  If the crisis was an inevitable result of high delinquencies, then why didn't home prices, and the wider economy, collapse after 1990?)

Nominal growth of mortgages was low until 1998.

Growth in Loans and Leases in Bank Credit at the commercial banks appears to have suffered a permanent drop from the long term trend, and continued along the new trend path without reverting to the old trend.

After a brief increase during the recession, inflation dropped to a descending trend after the recession, so that inflation rates after 1992 have remained much lower than the pre 1992 levels.  (This trend has generally been in place since 1980.)

So, why was the 1990 recession less severe than the 2007 recession?

Some things which were different:

Inflation was over 4%, and rose to over 5% during the 1990 recession.  Even though inflation was in the midst of a sharp secular decline, 4% left a lot of room for price flexibility and real wage adjustments.  It also meant that the zero lower bound was not relevant, that bond yields could reach natural equilibrium prices, and that holding money was associated with opportunity costs.

In addition, risk premiums were low and declining in the 1980s and 1990s.  This meant that, even without the inflation premiums, real interest rates tended to be high.  (Low real long term interest rates tend to be associated with high equity risk premiums.)

In the comments on previous posts, Travis V has pointed out that the persistently high equity risk premiums we have seen over the past decade could be related to the current low levels of inflation and interest rates, which cause demand shocks to be sharper than usual.  I agree.  (The previous period of high equity risk premiums might have been related to the opposite problem - instability because of excess inflation in the 1970s.)

I also wonder if the target inflation rate needs to be a little higher than it had been in the past because the large components of education and health care, which have tended to be inflationary and tend to capture an increasing portion of expenses, mean that total compensation might need to rise at a rate quite a bit higher than zero before growth in cash wages can become positive.

And I also wonder how much the current high equity risk premium reflects the current state of the corporate competitive environment, where, especially in the tech sector, firms capture high profits and valuations which are based on intangible values and network effects, and can quickly vanish.  Possibly, these high risk premiums simply reflect a large influence of the volatile state of global technological advancement on equity markets in general - creating a bifurcation of (1) capital at risk with unavoidably high variances in potential outcomes and (2) low risk capital which must balance this problem in the aggregate market portfolio.

Finally, employment was affected by unprecedented pro-cyclical policies during this cycle, which were not in place in 1990 to nearly the same scale.  The last graph here reflects an estimate of the unusual level of very long duration unemployment after the 2007 recession, which I attribute to the unusually generous extension of unemployment benefits.  In all previous time periods, very long duration unemployment could be pretty accurately predicted by using shorter duration (less than 26 weeks) unemployment levels.  After the 2007 recession, this longstanding relationship broke down.  This graph shows the estimated effect this had on unemployment levels.  Without that effect, unemployment in 1990 and 2007 look similar, except for the brief spike above 8% in 2009, which quickly fell back to about 7% in 2010.  Five years after peak unemployment, in 1997 and 2014, the unemployment rate was just falling to below 5% (after removing the remaining very long duration unemployed).  But today, there still remain more than 0.5% of additional reported unemployed workers with very long unemployment durations (currently averaging more than 100 weeks).

If inflation had been higher, equity premiums lower, and labor policy less pro-cyclical, maybe the aftermath of 2007 would have looked much more like 1990.  Maybe the disastrous FOMC decisions, especially in 2008, look especially damning because we have seen the results, but maybe they were simply following the general tenor of a policy stance that had been working pretty well for a couple of decades, and by 2008 the slow march of these trends had created new downsides to that policy stance that we hadn't had to deal with before.

How much different would things have turned out if homeowners in 2008 were sitting on 10% or more of additional inflationary gains in their home prices and rent values, just as a cumulative effect of general inflation?  What if the natural neutral rate of interest had been a couple percent higher?  It's not hard to imagine a much different scenario.  Maybe the Fed hadn't changed so much in 20 years, but the nominal economic landscape they had shaped simply had become less forgiving to tight monetary policy in ways we didn't appreciate until after the crisis happened.

As usual, I suspect that our determination to blame this all on greedy financiers and lemming-like homebuyers will cause us to miss the important lessons here.

3 comments:

  1. Really thoughtful blogging. The explosion of international trade and the Internet may have combined to make business enterprises more risky as well.
    I think the main point is clear: the Fed needs to be much more aggressive in fomenting growth.
    As Milton Friedman advised in 1992, even when inflation was 3% and real growth was 4%.
    Which raises a troubling question: when did America's right-wing monetarists become such a band of defeatists? They seem to think robust growth is impossible, or if possible would cause inflation to rise above 2% and thus is impossible. And the liberals are worse.

    ReplyDelete
    Replies
    1. I think you're right. I wonder if we should expect higher equity premiums since such a high proportion of corporate revenues and production are foreign now. Total expected returns to corporate assets appear to still be in the same range they have been for a long time, although it looks to me like those expected returns may be boosted slightly by higher expected growth in foreign revenues. GDP growth expectations are low in the US, and this doesn't seem to be reflected in corporate growth expectations. I haven't completely settled on a theory yet of what, if anything, should happen to corporate returns over the long run. It could be that demographics and emerging market savings are the largest contributors to the current risk premium profile. But, just with Apple, there is a lot of intangible capital and a lot of cash, with little or no corporate debt. The tech. and internet factor must have some influence.

      Delete
  2. TravisV here.

    Thanks for mentioning me! I've been busy lately but LOVE the 1992 Friedman WSJ op-ed!

    ReplyDelete