On Friday 5 June, in the face of growing Coronavirus cases and a worsening domestic (and global) economy, all three major US indexes made record moves to the upside (with the Nasdaq recording an all-time record high) in response to ‘better than expected’ unemployment rates in the US. The percentage of US citizens unemployed in April was 14.7%, this number coming down to 13.3% in the month of May -a welcome reduction, and perhaps a positive sign for the US economy.
Although a promising turn for the US economy, it was not such a great time for the many analysts and economists who tried to forecast the number… and got it terribly wrong.
The consensus estimate (or ‘expected’) unemployment rate for May, set by the brightest and most ‘in the know’ on Wall Street, predicted an increase on April’s numbers to 19.5% unemployment in May.
An increase of over five points on April’s figures but a whopping six percentage point miss on the actual of 13.5% for May. And, in fact, “the greatest miss in forecasting history”. Not only was the consensus estimate several deviations away from the actual, it was also in the completely wrong direction.
Not one of the 90 economists and analysts surveyed predicted an improvement in the numbers from April to May. It begs the question – how could they have been so wrong?
A number of those surveyed have scrambled to defend their position by claiming flaws in the data-gathering processes used at the Bureau of Labor statistics (BLS), who research and report on the actual number and have done for decades. Some others have simply blamed the unprecedented nature of the times we find ourselves in.
I can’t comment on the processes of the BLS, but I do agree that these unprecedented times have made it very difficult to make decisions and forecasts on the future. Partly because we have no previous similar experiences from which we can build assumptions, but more importantly because of the mindset and behaviour it encourages. A fearful and gloomy one.
Could the emotional tail be wagging the forecasting dog?
Is it possible that these learned and objective beings have allowed their emotional self to steer their predictions on the future? Personal bias and mood can subconsciously frame our thinking in such a way that we ourselves may not even see it. We talk about market momentum, but this is driven by mood momentum and the sentiment of the market, which at present is incredibly bearish. Where these economists have perhaps had to ‘fill in the gaps’ due to missing or inconsistent data, could they be doing so with a bias to the downside?
Is this perhaps a major flaw in the validity of forecasts given just how influential on market mood, but susceptible to personal bias (and wrong), they can be?
The evidence on forecasting
Whether it’s the weather or lottery numbers, we (as humans) prove time and time again that we are unable to predict the future with any real consistency over time. This phenomenon boasts a rich body of academic study, some of which I would be remiss not to share.
A leading and relevant example of the folly of forecasting can be seen in a study published by Professors John Graham and Campbell Harvey in the Quarterly Journal of Economics. They asked 500 CFOs (Chief Financial Officers) from some of the largest public companies in the US to predict the behaviour of the S&P500 in the following year and how certain they were of their predictions. They did this every quarter for 10 years, generating some 13,000 predictions. Just over 70% of those predictions were wrong by some margin but boasted an 80% confidence rate.
So, they were very wrong, but very sure. AND very influential!
This was more a test of over-confidence and hubris than predictive ability, demonstrating not just how wrong we can be but how certain we are about the unknown. They found those CFOs with a greater confidence in their incorrect forecasts tended to over-lever their businesses and make poor operational judgements and so this behaviour comes with risks.
Some other more playful examples are the predictions of Decca records who thought “guitar music is on the way out” in 1965 and turned down the opportunity to sign The Beatles. Variety Magazine went so far as to say rock and roll will be “gone by June” in 1966.
The point being, we are not always quite the prophets we like to think we are, and even the brightest in any given industry are prone to making mistakes about the future. This is just as much a debate on the distorting effect of one’s biases and emotional make up as it is their predictive capacity, one heavily influencing the other no matter how Spock-like and rational we try to be.