22-02-2022 · Interview

‘Most people are not good investors because they are driven by greed and fear’

Exciting stocks can pose significant risks to investors. We discuss this and other topics with our guest Ilia Dichev, Professor of Accounting at Emory University’s Goizueta Business School.

    Authors

  • Lusanele Magwa - Quant Investment Specialist/CPM

    Lusanele Magwa

    Quant Investment Specialist/CPM

What is you main takeaway from your research on investment returns versus investor returns?

“The main finding is that market participants typically experience poorer performance (based on investor returns or dollar-weighted returns) compared to the securities or vehicles they invest in (based on investment returns or buy-and-hold returns).1 This is primarily due to the timing of capital flows. In the early years of the hedge fund industry, for example, a fairly small pool of capital chased attractive opportunities. This setting initially helped hedge funds to generate strong returns, leading investors to pour significant capital into these strategies.”

“However, most of the late-arrival investors have since not earned the publicized high returns. To illustrate this, if you calculate the buy-and-hold returns of hedge funds, the outcome is a generally good average which combines the early period of strong performance with the less-than-stellar returns that have since followed. But most investors have not pocketed this average return as they were typically late to the party. Therefore, the investor experience is masked by just calculating buy-and-hold returns. And all investors are subject to this effect.”

“For instance, most of us save for retirement and we increase our exposure to capital markets as we edge closer to retirement. Firstly, imagine if markets perform strongly in the early years of your retirement savings journey when the capital invested is still relatively small or moderate. Secondly, picture them delivering weak returns closer to your retirement date when the savings pot is significantly larger. This would be devasting as the average buy-and-hold returns over the period would still look good, while your experience as an investor would be worse off. Thus, investor returns depend on the timing of investments and how much skin one has in the game over time.”

Can you relate this to the growth versus value debate?

“This topic still requires some careful consideration from an academic perspective. But a key takeaway from my experience as a researcher is that most people are not good investors because they are driven by greed and fear. This means they typically enter close to the top when securities are doing well, and exit near the bottom when their investments are struggling. What this implies is that they usually experience lower investor returns than investment returns.”

“So the strong returns that are quoted in newspaper headlines can paint an optimistic picture and fuel investor interest. The implication of this is that the attention-grabbing or most exciting stocks will likely attract a lot of activity and capital flows. In my view, these stocks can pose significant risks to investors. A well-known episode which encapsulates this risk is the bursting of the dotcom bubble in 2000.”

Ilia Dichev - Professor of Accounting at Emory University’s Goizueta Business School.

Ilia Dichev
Professor of Accounting at Emory University’s Goizueta Business School.

Some of the finer points on the quality of earnings are not appreciated enough

Part of your research has also focused on earnings quality. What have you gleaned from it?

“There is a lot of good research that shows that while markets are quite efficient, some of the finer points on the quality of earnings are not appreciated enough. And these effects can drive returns. One of the most well-known ones is the so-called accrual effect.2 Basically companies that have high accruals typically have poorer earnings quality and tend to earn lower returns. But the question about a firm’s quality of earnings can be refined a lot more.”

“For instance, it’s one thing if your earnings are purely based on your fundamental business. Let’s say you run a company that sells cars and its earnings improve on higher prices and volumes. But it’s an entirely different issue if your earnings increase because you have sold a division or due to some tax-related shenanigans. It is therefore important to determine if reported earnings are solid, or if they resemble Swiss cheese and are full of holes. I've conducted a fair amount of research3 in this area and it seems as if investors are willing to pay a premium for higher earnings quality.”

More broadly, what are the potentially exciting areas of research in the field of finance?

“In my opinion, a topic which has not been covered enough in the academic literature is equity duration or cash flow duration. Intuitively, a company with a longer duration of cash flows is riskier. I mean, if most of its cash flows will only be realized in 10 or 20 years, who knows what might happen before then? From an academic perspective, this is one of the most important research topics that have yet to be extensively explored.”

“Aside from that, there is a lot of interest in variables that can predict returns. But with hundreds of documented anomalies, it's become a free-for-all discussion. In fact, a lot of these anomalies are either variations of each other, unstable, time-specific, or dependent on the methodology used. Individually, the discovery of each new anomaly seems okay, but when you put them all together, the picture is unsatisfying.”

“Somewhat paradoxically, the advances in research methods and computing power have made this whole problem worse. The point is that this powerful machinery is used by academics or practitioners to find patterns in the data and it often does so. The key question is whether these patterns are meaningful and repeatable. So, given the huge number of documented anomalies and papers written on them, I think a healthy dose of skepticism and robust academic analysis are required to separate the wheat from the chaff.”

Stay informed on our latest insights with monthly mail updates

Receive our Robeco newsletter and be the first to read the latest insights and build the greenest portfolio.

Stay updated

And what is your general view on machine learning?

“The opportunities presented by machine learning are very exciting. In some sense, I think there is no going back in terms of data analysis. But now that the genie is out of the bottle, I do think that deeper theoretical thinking is required to validate and organize the findings beyond just recognizing empirical patterns.”

Ilia Dichev - Professor of Accounting at Emory University’s Goizueta Business School.

Ilia Dichev
Professor of Accounting at Emory University’s Goizueta Business School.

Investor returns are both lower and riskier than investment returns

Lastly, which are the most intriguing topics in your research agenda?

“One interesting project that I am working on is related to the volatility of investor returns (or dollar-weighted returns). This is important as it looks at the risk associated with these returns which can shed more light on the investor experience. Based on the ongoing research, we find that dollar-weighted returns also have a higher variance versus buy-and-hold returns. So basically, investor returns are both lower and riskier than investment returns. I think this finding is not yet fully appreciated.”

“Another paper that I am busy with looks at the relationship between accrual duration and earnings quality. Simply put, the idea is that the longer it takes between an accrual and its associated cash flow, the poorer the quality of earnings. For instance, line items that have short duration like accounts receivables are typically fairly valued as they are collected within a few months. But if you think of long-term accruals such as pension obligations or environmental obligations which stretch over decades, then it is tougher to evaluate their value. So, earnings based on such unreliable accruals are low quality.”

“For example, General Electric (GE) has a lot of long-term health obligations for its GE Capital unit. But no one really knows what its health obligations will be in 20 years. These depend on numerous factors, such as the rate of medical inflation. Therefore, it is very difficult to determine the value of these accruals. In its earnings calculations, GE essentially incorporates the assumptions which are built into this 20-year vehicle. Ultimately, this has an impact on the quality of its earnings.”

“To sum up, these long-term accruals are bound to result in lower earnings quality as they involve long-term estimates. This brings to mind the example of Enron, where revenues were booked based on long-term gas contracts. This required the estimation of gas prices 20 years into the future. But given the lack of certainty on gas prices that far into the future, it afforded Enron employees the room to reference a wide range of prices. So this is what I mean by low quality accruals.”