Chapter 1: The Shortest Investment Joke: My Forecast Has a Decimal Point
On 7 December 2018, the Financial Times reported, under the headline “Market predictions: 2019 to bring new level of uncertainty”, that investment strategists expected market uncertainty to prevail in 2019. Furthermore, the FT reported that the “median forecast of strategists” indicated that the US economy would grow by 2.6% in real terms and the S&P 500 would end the year at 3,090 points.
At the time the article was published, the S&P 500 stood at 2,633 points, so the average expectation of strategists was for a 17.5% rally in little more than a year. Most strategists expected the S&P 500 to rise in 2019, but the most pessimistic forecast called for a 9% decline.
It has become somewhat of a rite of passage for strategists to be interviewed by newspapers at the end of each year about their expectations for the next 12 months. The requests for forecasts come in by the dozen from newspapers, TV stations and magazines alike. Some outlets even run horse races comparing the accuracy of forecasts at the end of the year to crown the best forecaster, who will then forever be known as the person who predicted the current bull/bear market or the last recession/recovery (choose your adjectives depending on the actual market behaviour). In extreme cases, economists or strategists who call for a financial crisis or a severe bear market in advance can become global superstars.
The temptation is high for economists and strategists not to provide forecasts that are qualitative in nature, but to instead make precise forecasts for the return of the stock market or the growth of the economy with several digits after the decimal point. This reminds me of a quip by novelist William Gilmore Simms, that “economists put decimal points into their forecasts to show they have a sense of humour”.
The excerpts from the newspaper article above already demonstrates why this approach is problematic. The most pessimistic forecast called for a significant decline in the stock market from the already depressed levels at the time (remember that the S&P 500 dropped more than 10% in the fourth quarter of 2018), while the median forecast called for a massive rally. The fact that the median was 17.5% means that half of the surveyed strategists called for stock market returns in excess of that. Thus, according to a bunch of experts on financial markets, the return of the S&P 500 in 2019 could be anything from −10% to more than 20%.
What use is a single forecast from this collection, if the uncertainty around it is so large?
Some would argue that these return forecasts are marketing exercises that no investor would ever take seriously. Yet I know of several strategists and economists who produce such forecasts on a regular basis in their jobs, even though they know they are not worth the paper they are printed on. Instead, they say that their clients, be they professional or retail investors, want these numbers so they can decide how to allocate their portfolios.
I have been in several meetings where clients have greeted me with the question: “What’s your number?” In one job I had, after I provided an outlook for the new year, the first question asked was about the consensus forecasts for the stock market. Investors take these forecasts not just seriously, but literally.
Nevertheless, I don’t think strategists make these precise forecasts simply to please their clients or journalists. After all, they keep making extremely precise forecasts even in circumstances when hardly anyone is looking. Here is an excerpt from a research report on Geely Automotive, the Chinese car manufacturer and owner of Volvo and other brands:
“In view of the slightly lower sales volume expectation due to market weakness, more than offset by better-than-expected price resilience, we raise our fiscal year 2019–2020 revenue forecast by 2.5–3.6%. However, with more prudent margin assumptions, given increasing costs, such as emission standard upgrade and research and development, we only increase our earnings forecast by 0.8%.”
The analyst then goes on to predict that Geely would sell 1,618,939 cars in 2019, 1,753,079 cars in 2020 and 1,858,030 cars in 2021. Profit margins are expected to be 19.7% in 2019 and 19.9% in 2020, before falling back to 19.7% in 2021. And these forecasts are presented alongside 15 pages of extremely detailed tables modelling every aspect of the company’s balance sheet, cash flow statements and income statements. All this is used to recommend the share as an attractive investment opportunity for investors.
I am not sure if the recommendation would change if Geely sold 1,618,940 or even 1,618,938 cars in 2019, or if a profit margin of 19.8% instead of 19.7%, would make a difference to the analysis. But if it didn’t make a difference, then I would argue that this analyst is using excessive precision which is irrelevant for the investment decision. Just because Excel can show seven digits of a number, does not mean you need to.
This chapter will focus on the impact this tendency towards forecasting developments that are essentially unknowable, or that have such high uncertainty around them that a point forecast is not reasonable, can have on investor portfolios. While precise forecasts can create a sense of certainty about the future, such certainty is entirely misplaced. First, I will show that forecasts like those above are highly unreliable and can create massive losses in a portfolio if they are only a little bit off the mark. Second, investment portfolios can be made more robust in changing market environments if the uncertainty of financial markets is taken seriously and incorporated into the portfolios of investors. How this can be done in practice will be the focus of the second part of this chapter.
The future is uncertain – deal with it
I would not have a problem with precise forecasts if they were in any way reliable. However, as Figure 1.1 shows, the forecasts of strategists have been far off the mark in the past.
The figure shows the difference between the return of the S&P 500 as predicted by strategists at the beginning of each year and the actual return of the index in that year. For example, at the beginning of 1999, strategists predicted the S&P 500 to rise by 8.8%. It then rallied 19.5%, so the analysts were too pessimistic about the market by 10.8% (leftmost bar in Figure 1.1). In 2018, on the other hand, strategists expected the S&P 500 to rally 10.3% at the beginning of the year, but it declined by 6.2%. This meant that their forecasts were too optimistic by 16.6% (rightmost bar in Figure 1.1).
Figure 1.1: Forecast error of analysts
Source: Bloomberg.
In 13 of the last 20 years, strategists have been off the mark by more than 10%, while the median error of forecasts has been 4.6%. Compare this to the average annual return of 3.6% of the S&P 500 during the last 20 years and it becomes clear that these forecasts have an uncertainty associated with them that is roughly of the same order of magnitude as recent market returns.
Maybe I am being too hard on these poor strategists. Maybe we should pay less attention to the actual number and more to the direction they predict. Will the stock market go up or down next year? As it turns out, they got the direction of the stock market right in only nine out of the last 20 years (Figure 1.2). In other words, the ability of strategists to forecast the direction of the stock market is no better than the toss of a coin.
Figure 1.2: How often do analysts get the direction of the stock market right?
Source: Bloomberg.
These results are no accident. Professor Markus Spiwoks of the University of Applied Sciences in Wolfsburg, Germany, has investigated the forecasting ability of economists and strategists for years. As a former research analyst and portfolio manager, he and his colleague Oliver Hein looked at the forecasts of more than 400 experts surveyed by the ZEW (Zentrum für Europäische Wirtschaftsforschung) in Germany each month between 1995 and 2004. T...