Professor Hendrik Bessembinder urged investors to be suspicious of stock market analysis that relies on averages, including mean-variance optimisation.
Professor Hendrik Bessembinder, who is currently Professor of Finance at the WP Carey School of Business at Arizona State University, caused a stir in 2017 when he published a research paper that showed US stocks on average do not perform much better than US Treasury bills, a form of short-term government bond, in the long run.
The finding was remarkable because investment theory tells us stocks are far more risky than Treasury bills and should attract a premium for investing in them.
Bessembinder then caused further uproar in 2019 when he extended his research to global stocks and showed that despite the fact a sample of 64,000 stocks generated US$75.7 trillion in wealth between 1990 and 2020, the majority of these stocks underperformed one-month US Treasury bills.
All of the wealth generated during that period could be attributed to the performance of a mere 2.4 per cent of stocks in the sample. Remarkably, a whopping 57.7 per cent of stocks in the sample actually caused wealth destruction.
The research has been used to support all kinds of arguments, most notably by proponents of both active and passive investing.
But in an address to an investment forum organised by Baillie Gifford, MFS Investment Management and Orbis Investments in Sydney earlier this week, Bessembinder urged people to show restraint in any interpretation of the data.
Mean-variance analysis was motivated by [the idea that] stocks are more or less normally distributed. That is just nowhere close to the truth when we look at compound, long-run outcomes
The findings of the reports hold little predictive power and merely show what happened in the past, he said.
If anything, the research indicates that when it comes to stock market investing, investors should be suspicious about averages and, in extension of this, mean-variance optimisation.
“I just want to throw out there [that] there is a ton of work on mean-variance optimisation and we focus a lot on things like alphas and Sharpe ratios, whose foundation is a mean-variance analysis. Alphas are an arithmetic mean,” Bessembinder said.
“I just think we’ve really been missing something. Mean-variance analysis was motivated by [the idea that] stocks are more or less normally distributed. That is just nowhere close to the truth when we look at compound, long-run outcomes. And we should be looking at compound, long-run outcomes,” he said.
The finding that all of the wealth generated in global stock markets over the past 30 years is due to a small number of companies doesn’t mean it is impossible to pick the winners or at least some of them, he said.
Although for the vast majority of people passive investments would make sense, there are indications active management works for skilled asset managers. And not only that, Bessembinder also believes certain institutional investors have the skill to identify those managers who can consistently outperform.
“What does this mean for active versus passive? I think the idea of comparative skill, what are you good at, is really crucial,” he said.
“I firmly believe some asset managers have the right comparative advantage and among those who place their funds with asset managers, some have comparative advantage in identifying the right asset managers. But that doesn’t mean everybody should do it,” he said.
Predictive Powers
Bessembinder was quick to point out his research doesn’t identify any drivers for returns that have predictive power.
“Of course what everybody wants to know is: Can we predict ahead of time which of these are going to [drive performance]? That is where the hard work comes in; it is not going to be easy. It is not going to be run [by] a simple computer program,” he said.
But he does point out that there are a number of fundamental measures that explain which companies have done better in the past.
I haven't considered everything that could possibly be considered here. But among the things we can readily measure, it's literally the bottom line that is most important: gross growth in income
He identifies a number of metrics, including income growth, asset growth, sales growth and average income-to-asset ratios.
Of these metrics, income growth is by far the most influential.
“I haven’t considered everything that could possibly be considered here. I haven’t considered more intangible things. But among the things we can readily measure, it’s literally the bottom line that is most important: gross growth in income,” Bessembinder said.
Australia
Bessembinder also analysed the data for the Australian stock market and found it is no exception to the patterns found elsewhere in the world. In fact, the trend is more pronounced in Australia.
“Australia essentially [shows] the same pattern and, if anything, the pattern is a little stronger in Australia than in the rest of the sample. The mean returns for almost 3000 Australian stocks, where we could get data, was almost 600 per cent, but the median [was] lower than in the other countries,” Bessembinder said.
“So just let that sink in for just a second: a randomly picked stock, traded publicly in Australia in the 30-year period, had a median negative 43 per cent return.”
The Impact of the CEO
Judging from the questions from the audience during the forum, they were most interested in Bessembinder’s research into the role of the Chief Executive Officer in the success of a company.
But again, he did not give in to the temptation to draw hasty conclusions from his research.
“The data only shows what happened during the time a particular CEO was there. It didn’t tell you whether any value created could be attributed to the CEO’s efforts or to any external factors,” he said.
But it did throw up some interesting findings.
For example, the most wealth created in a company under any CEO was during Tim Cook’s tenure at Apple: US$3.1 trillion. More so than during Steve Jobs’ time, when $355 million in wealth was created.
Let that sink in for just a second: a randomly picked stock, traded publicly in Australia in the 30-year period, had a median negative 43 per cent return
But it should be noted that when Jobs was Apple CEO, the company was a lot smaller and in relative terms he added more than Cook did.
During Jobs’ tenure, Apple produced 34 per cent returns per annum on average, while during Cook’s tenure the company produced 29 per cent returns.
Bessembinder was quick to point out the influence of a CEO or management in general is hard to measure. “Things like fire in your belly is really important, but it is hard to measure,” he said.
Own Experience
As is often inevitable with stock market research, Bessembinder was asked about his personal approach to investing, a question which he graciously answered.
Although he largely sticks to long-run, broadly diversified buy-and-hold strategies, he admitted to occasionally straying from this philosophy.
“A substantial part of my investments is in long-term, buy-and-hold [strategies], but I have deviated from that,” he said.
“I have purchased some individual stock positions and I have occasionally made investments that amount to market timing. And I’ve done well, but it can be hard to tell the difference between luck and skill, and it is a small sample.”
__________
[i3] Insights is the official educational bulletin of the Investment Innovation Institute [i3]. It covers major trends and innovations in institutional investing, providing independent and thought-provoking content about pension funds, insurance companies and sovereign wealth funds across the globe.