Wednesday 17 June 2015

Innocence & Experience & 'The Wealth Effect'

One of the popular and current rationales for why the US economic expansion has never quite developed the normal panoply of cyclical accelerators is that the ‘wealth effect’ of a surging stockmarket has been far less than previous calculations suggested.

The observation is important, because as of March 2015, the US household sector was keeping 31.1%  of its total financial assets in equities or mutual funds, compared with only 19.5% in bank deposits & credit market instruments. In fact, the proportion kept in equities and mutual funds is the highest since end-2000.

Now, the S&P500 has attained a CAGR of 14.7% over the last six years, and as it has done so, the amount of wealth households have tied up in the S&P has risen from $8.2tr in 1Q2009 to $21.56tr in 1Q2015.

How much is that? Well, in the national accounts, compensation of employees currently runs at $9.477tr pa, and it is growing at about $380bn pa.  Given the tally of of equities and mutual funds held by households, it would require a rise in equity values of just 1.75% a year to generate that $380bn pa pay rise.   Or put it another way: the $13.36tr rise in the value of households’ equity and mutual fund investments is equivalent to just over one and a half year’s worth of average employees’ compensation during the same period.

So how households’ do with that new wealth plainly matters, a lot, to the economic cycle.

Looking through the literature on why the wealth effect is failing, I’m struck first by the certainty which allowed economists and econometricians to assume that ‘the wealth effect’ would be stable over time.  I’m also struck by the absence in the literature of any reference to the permanent income hypothesis  (the idea that a person will adjust his spending according to the his life-time income expectations  - or alternatively that his spending pattern will itself reveal that life-time income expectation).

It seems to me very likely that the proportion of the ‘stockmarket wealth’ a person is prepared to spend will not be stable, but rather will be informed by (and reveal) his underlying expectations about the likely volatility of those holdings.  If a person’s long-developed experience is that the stockmarket only goes up, he is likely to spend a higher proportion of the gains than if his experience has led him to believe that a significant portion of it is likely to be given back in the near future.  In other words, people will only ‘believe’ a portion of the increase in their wealth. How much they ‘believe’ it, will depend on their recent experience.

Now these ‘beliefs’ are likely to be developed as a result of long experience, and are likely to change over time as those experiences change. Here’s a  very simple (primitive, even) model: the ‘belief’ in the secure return of stockmarket investments is formed over a 10yr period, with more recent years having greater weight than in previous years. In the chart below, I’ve averaged a straight-line declining balance of years, so the change in the most recent year has 100% weight, the year before than 90%, the year before 80% etc.

What this chart suggests is that very sharp stockmarket falls may have a strong and long-lasting negative impact on people’s beliefs about what proportion of a stockmarket’s subsequent gains can be ‘believed’. And in turn, those beliefs may be incorporated into permanent income calculations, and hence spending/saving decisions.  Thus in 1999, the experience of the previous 10yrs would have led to an expectation of ‘safe’ gains of 11.1%, during a year in which the S&P rose 19.5%.  By contrast, in 2014, the previous 10 years would have taught that only 5.7% gains were ‘safe’, and any excess gains were unlikely to be factored into spending/saving decisions. Adjusted for the changing base, one would thus expect the impact on consumer spending of a rise in the stockmarket in 1999 to be approximately double what it was in 2014. 

If expectations are formed over a long period of time, no-one should expect ‘wealth effects’ from the stockmarket to be stable. But they might, given more investigation along these lines, be not entirely mysterious. 

Meanwhile, the other insight into how households actually think about their ‘wealth, or financial situation, is revealed in the household saving rate.  One would expect that when lengthy experience has led households to believe more strongly in the security of their stockmarket gains, their saving rate would decline; when they become more sceptical, one would expect the savings rate to rise. And, in general, this does appear to be the case. 

What can one conclude? That stockmarket investments are not money in the bank, and repeated warnings that prices can go down as well as up are understood, albeit that belief is moderated by experience over an unknown length of time. If that time is lengthy, as one would expect, then the impact of a sharp fall in equity values, or a series of sharp falls, will compromise the ‘wealth effect’ of subsequent rises for years to come.  In 1999 the S&P rose 19.5%, and households perhaps ‘believed’ that 11.1% of the rise in wealth would stick, and could thus be spent. In 2014 the S&P rose 11.4%, but households perhaps believed that only 5.7% of it would stick.  And one other thing. . . given a fair wind and no catastrophes (how I wish!) the wealth effect of a rising stockmarket is likely to gradually recover, but only as the disasters of 2008 recede from memory.