Three behavioural biases

Around this time last year I remember starting to think that investors in general, especially those in the equity market, were becoming overly bearish. Markets had already fallen a fair amount from their peak. Equity valuations did not look hugely stretched, at least not on some measures (see below). Meanwhile, the press and all the talk seemed to be very negative. For sure, some of the bears made valid points, such as analyst estimates being way too high, and the still highly inflated prices in other risk assets such as commodities and real estate. But, for equities in particular, I couldn't help thinking that a lot of this was now priced into the markets. Of course, subsequent events, which saw equity markets in developed economies fall at least another 30%, suggested that my view was clearly wrong, or at best far too early.

Of course at the time it was not clear that the financial markets were about to almost completely collapse. Even the uber-bears were not predicting this. So I think it would be unfair to say that the subsequent path of equity markets proves that my belief this time last year was wrong. That said, I am somewhat disappointed with myself for not being able to truly see the wood from the trees. That equity markets have fallen way further than I thought they would when I initially turned positive suggests that I had at least not aqequately assessed the risks involved.

So why did I fail to assess the risks properly? Having put considerable thought to this question, I have come to the conclusion that it can mostly be explained by behavioural finance biases, rather than a specific informational or interpretative deficiency. Indeed, I think the only information gap was my lack of understanding of the banks' and non-bank financial institutions' balance sheets. However, virtually all investors were in this camp given the poor levels of disclosure. Behaviourally speaking, I woulld highlight three biases that I think adversely affected my assessment of the situation:

1) Familiarity bias. This is the tendency to draw heavily on facts that one is familiar is. Being an equity investor and not a credit investor, I tended to look closely at equity market valuations, and especially those in the sector that I was a research analyst for at the time. I did not think these valuations were hugely stretched given how far many of the stocks I was interested in had already fallen. Therefore I conlcuded that it would be very unlikely for these stocks to fall a great deal more given what the market had already priced in. The mistake I made was to apply this to the whole equity market, failing to appreciate fully that a lot of other sectors such as mining (well at least I knew that was overvalued!), industrials and other financials were not yet pricing in the same sort of outcomes. I also rather naively assumed that the outcome being implied by the credit markets should not be taken too seriously given the huge liquidity dislocation within those markets.

2) Extrapolation bias (I think this one may actually have another name, but it is obvious what it means). Here I think the mistake I made was relying too much on one indicator - the earnings yield versus the bond yield - to draw a bullish conclusion. This measure which has worked reasonable well for the last 20-25 years or so suggested equities were very cheap compared with bonds, even after adjusting the earnings down 15-20%. I knew that other measures such as cyclically adjusted p/e ratios and tobin's q still suggested the market was slightly expensive on an absolute basis, but I also knew that historically these measures have not been great buy and sell indicators in the near/medium term (i.e. 1-2 years). Perhaps I should follow Jeremy Grantham and rely more on mean reversion for my investment decisions in the future.

3) Overconfidence. Here I think my mistake was not to fully appreciate how serious the "known unknowns" could be, in particular the scale of the exposure on bank balance sheets and the knock-on effects this could have. For sure, I was aware of the problem, I knew it was one of solvency and not liquidity, and I also suspected that more capital might need to be injected at some point. But I was overconfident (naive?) to think that this was already reflected in the general level of equity prices. One of the tests that behavioural finance researches routinely run is to ask people to estimate on the spot with 90% probability (90% confidence interval) a range of answers for a specific question, such as the 150 year total return of US equities. The results usually show that respondents vastly overestimate their ability (i.e. way more that 10% of respondents' range estimates are outside of the answer). I guess if you had asked me to estimate a 90% range for the market this time last year, where we ended up would have been outside of my range.