Thursday, February 11, 2016

Note To Self: Bear of The Dark

“Going negative is daring but appropriate monetary policy. But it is a sign of a terrible policy failure by fiscal policymakers.” - this succinct quote from former Fed official Narayana Kocherlakota captures the very essence of the limitation of the central banks today.

I started turning bearish - like most market participants - in early January. I turned neutral here, and bearishness followed. See here and here and here. That was mostly based on reading the market information and assessing them, subjectively. Which is NOT particularly a great way to develop an outlook. Here we attempt a more formal process.

A bad time to assess if the markets are turning a bearish corner is on days like today: the markets sell off a quite a bit. First thing one has to fight even before starting to reason are a whole bunch of cognitive biases!

Let's assume we have already done that. The next thing is to take an objective view, and establish based on the current information what are the chances of a bearish market setting its foot hold firmly.

First, the base rates: to start with, only the price information and no context whatsoever. First we notice we are in some kind of cusp. The 3 month, 6 month and 12 month performance of, say, S&P 500 are more or less similar within some tolerance. That means, we are not in a steady trend, yet, in either direction - bullish or bearish. In a steady trend, it is much easier (and almost useless) to establish if we are in a bullish or bearish phase. It is very difficult to predict a trend at the onset of it. Many investors spend considerably amount of time and effort trying to achieve this, and more often than not, they fail.

But looking at data, we can make some probabilistic observation. Here we look at the historical probability distribution of S&P 500 returns - over next 3 months and 6 months - when it behaved similar to today. That is: it sold of significantly (we look at 5%, 7.5%, 10%, 12.5% and 15% prior sell-off) and that happens in a turning phase (i.e. 3, 6 and 12 months performance are equivalent within some tolerance). Here is how it looks, since the post-world ear.

For next 3 months return distribution - on the milder sell-off side, the distribution tends to the usual - hardly useful to prognosticate. However, as the sell-off amount increases it tends to widen in a bi-modal distribution - tending to either settle down in to a bearish phase or come-back sharply. And only beyond 15% the upside skew tends to dominate. We are at around 12%. Looking at the next 6 months returns - the historical distributions are more extreme, and chance of an onset of bearish trends increases with increasing sell-off, till the upside skew tilts it in bull's favor beyond 15%.

The obvious conclusion here is we are not going to settle at this stage, and likely to move to one of the peaks. And we can also say, although with less conviction, the sell-off is perhaps not large enough to be perceived over-sold and is slightly tilted in favor of the bears.

Continuing with base rate case: Equity market is related but distinct from the underlying economy. One way to approach is to ask what are the chances of a sell-off given a recession, and also what are the chances the markets sell off anyways even without any recession in the economy.

Here is a list of historical recessions in the US since the great depression, and the corresponding market sell-off

So excluding minor recession introduced by sudden fiscal tightening or monetary tightening or the very distinctive dot com crash - the average sell-off has been around 12 percentage point per percent point of GDP lost. And roughly 30% on an average.

It can be noted that we had two occasions of a market sell-off (greater than 10% peak-to-trough) without any recessions - one during 1965-66, presumably under a credit crunch and followed by a recession. The other time was 1976-78, which was actually straddled by the 1973-75 recession and early 1980s recession. So the probabilities of a sustained bear markets without a recession is much less. But of course we are looking at a more globalized world now. Hence we can expect equity underperformance in the US even without a technical recession, if there is a large scale disruption else where in the world.

Overall conclusion - we have a considerable chance of a bear market setting in. Note, a 50-50 chance of a bear market is actually a much stronger conviction than it sounds. On a rough measures, since post war era till before the current phase, we have spend roughly a 30% of months in a bearish market. So unconditional probability of a bear market will be around 30%. And given the above observations, it should be higher than 30%. May be closer to, but quite less than 50%. A good number to anchor is mid-way at 40%. And this is a conditional estimate, without considering the context or forward outlook. It depends on the probability of a economic recessions, global or in a major economy. Which comes up in the next post.

As a tail-piece, here is how the classic equity risk factors are performing lately. This is in the US, but should be similar across the board.

The value is beaten down, and it is the momentum that is the last man standing.

No comments:

Post a Comment