PsyFi Search

Monday 16 April 2018

Forecasts? I Haven't Got A Clue ...

Small Worlds

We predict things all the time, we can't help ourselves. And in the small world that constitutes our immediate bubble of experience those predictions may have some validity. But in the big world that we inhabit they often don't.  Which of those worlds do you think more accurately represents stockmarket investment?

It's difficult for us to accept that our small world experience doesn't scale up to the big world, but it's often true. To all intents and purposes we're infants teetering on the edge of the volcano: to all intents and purposes we know absolutely nothing worth knowing.


Forecasting is a tricky business. So tricky, in fact, that I long ago decided that I haven't got a clue what's going to happen next, so I try not to worry about it.

But if you took away the ability to prognosticate about the future from most financial experts they'd actually have to go and get a proper job. There's nothing they like better than pontificating about the future, which makes it odd that mostly they can't even explain the past. You might almost think that they haven't got a clue either ...


It's almost cruel to point out the failures of past forecasters, but we'll do it anyway. Economists are quite good at this - Paul Samuelson back in 1961 stated that the success of the Soviet economy showed:
"A socialist command economy can function and even thrive".
For a while anyway. Another economist, Joan Robinson, was so impressed by North Korea in 1964 that she predicted that:
"As the North continues to develop and the South to degenerate, sooner or later the curtain of lies must surely begin to tear".
Or what about Joseph Cassano, who ran the giant insurer's AIG's financial products division. In 2007 he promised that AIG wouldn't be:
"Losing one dollar in any of those (CDO) transactions".
AIG went bust the following year and had to be bailed out.


The idea that our personal experiences simply don't translate into executable knowledge in the wider world is hard for most people to accept. To a point that's understandable, because people are very similar wherever you are, their behaviors don't change that much; in any local environment we'll usually find some level of connection with other humans, whether they're Inuits in Greenland, radical Republicans in Arkansas or bushmen from the Kalahari. At the very least we all eat, sleep, fornicate and excrete.

Despite this connectivity at a local level it still remains the case that when we scale up small worlds to big ones our experience counts for nothing because the behavior that emerges from our connected actions isn't simply a result of adding up lots of small world behavior. It's like the way consciousness emerges from a bunch of neurons and dendrites: look as hard as you want at any given brain cell you won't find a miniature version of human intelligence, the sum of the parts is magnificently greater than the whole.

CFO Bias

What applies to you and me also applies to everyone else involved in markets. Take Chief Financial Officers, for instance. Not only are they responsible for managing budgets but their forecasts are used to guide analyst expectations and determine future investment plans. So you'd hope they're a bit more rational than the average Joe in the street when it comes to not extrapolating small data sets into the future. Well, guess what:
"... when CFOs or analysts observe good or bad earnings realization, they think that similar realization persists into the future and fail to correct for mean reversion"
Expectations and Investment, Nicola Gennaioli, Yueren Ma, Andrea Shleifer.

These kinds of bias affect all participants in markets - Robin Greenwood and Andrei Shleifer studied a range of factors in Expectations of Returns and Expected Returns and came up with an interesting set of fundings - although readers here might well have been able to forecast them.

Inverse Correlation is Causality

Firstly they showed that pretty much all groups of investors have unrealistic expectations - they're correlated against past returns across different groups, across different surveys and questions and against investment flows into and out of mutual funds. The results look like they're identifying something real and aren't a facet of some research glitch.

Secondly, expectations are high after periods of market growth and low after slumps. Which is exactly what you'd expect if people aren't performing any kind of rational analysis and are simply defaulting a heuristic that extrapolates the recent past into the indefinite future, regardless of the lessons of market history.

Thirdly, peoples' expectations of market returns are inversely correlated with models looking at whether markets are overvalued or undervalued. Oh, and the models are usually correct and the people are usually wrong - essentially when the models are suggesting that markets are expensive while investor expectations for future returns are strong this is an indicator that markets will fall. Or to put it another way:
"... Both expectations of returns and ER [Expected Returns] predict future stock market returns but with opposite signs. When ER is high market returns are on average high; when expectations of returns are high, market returns are on average low."
The Future Looks Like History, Not Last Year

So what are investors expecting right now? According to the 2017 Schroders Global Investment Study they're expecting 10.2% returns per annum over the next 5 years, and Millennials are predicting 11.7%. Indonesians are the most optimistic, expecting 17.1% returns, US investors are the middle of the pack, confidently anticipating 11.2%.

To put this in perspective, the MSCI World Index has seen returns of 7.2% over the past twenty years. Schroders themselves predict below average returns going forward, expecting only a 4.2% return over the next seven years. In fact, across the board, you'll be hard pressed to find a rational expectations model that will generate much more than 5% a year, based on the current valuation level of markets.

Being Human - And Wrong

Of course, there are always arguments that it's different this time. There are revolutionary changes happening in the world: the collapse of a lot of traditional businesses under the onslaught of the internet is real, and the process of creative destructionism is well underway. But ultimately even online businesses have a real cost of capital to maintain their market position.

Clearly markets can continue to go higher in the short term buoyed by sentiment and hope, but ultimately economics determines future returns, not expectations. Markets are weighing machines, and they're telling us to we need to go on a diet. But CFOs, analysts and investors will carry on their merry ways: you can't convince a human being to trust data over feelings, it's just not the way we're built.

Related Articles

No comments:

Post a Comment