How we extrapolate the past…

Guido Balthussen

Author: Guido Baltussen, Head of Quantitative Research & Strategy Fixed Income and Multi-Asset at NN Investment Partners, and Associate Professor, Erasmus University Rotterdam.

We fall prey to the extrapolation bias, which means we have the tendency to extrapolate recent trends. Therefore, the returns experienced in the past shape our decisions about the future. Recent research reveals, this bias determines to a great extent the fund investments we make. But, as with most behavioural biases; it does not really help us in the end…

Imagine you have some money to invest and you would like to pick a fund. Which fund category, and which fund within that category do you pick? When focusing on Europe alone, already over 18,000 funds are available spread over 166 fund categories (like Equities Technology, or Equities Japan) across the Equity and Bonds asset classes.

One driver in such decisions is the extrapolation bias. We chase those fund categories (and funds) which experienced a good recent performance. After all, we tend to believe, outperforming funds should have displayed a good performance and a poor recent performance means it will likely underperform.

A recent study I conducted in cooperation with Jan Jaap Hazenberg, Head of Product Development at NN IP and Willem Van Der Scheer, Senior Strategic Marketing Manager at NN IP confirms that we form our believes in this way. The study, focusing on the European fund market, argues that the expectations we have about equity and fixed income mutual fund categories are driven by their recent performances over the past 1 to 12 months. Consequently, the allocations to Equities, Bonds, or sub-asset classes like Mid-cap Europe Equities increase if their recent returns were good.

Moreover, the stronger the recent returns are, the more we tend to believe the fund category will outperform, and the more we tend to allocate to the funds in that category. A return of 10% over the past 3 months, leads to inflows of roughly 60 million euro over the next month, or about 780 million per annum.

The chasing of recent returns in our fund allocations are stronger for those fund categories we tend to feel less knowledgeable about or which are more driven by growth prospects. In general, assets driven more by growth prospects are harder to value. In general, we are then more inclined to rely on past price trends as a source of information. In fact, the study shows that compared to local investments the chasing of recent returns is about double as strong for growth categories, like Emerging and Frontier Communications Markets, Small and Mid-Caps, Biotechnology and Technology/IT.

Is this extrapolative behaviour in our fund investments smart? Our study addresses this question as well. At first sight it seems that investors have slightly benefitted from the chasing of past returns. The reason for this; various asset classes displayed trending behaviour over the studied period (2002-2015). However, further analysis reveals that this could well have been different. It turns out that the extent of our extrapolation bias varies over time. Simply put, in some years we chase past returns to a greater extent than in other years. However, in those years that we extrapolate the recent past the most, our portfolios tend to perform materially worse than average (by about 5%).

In summary, returns in the past shape our investments. Our money tends to chase past returns. Sadly, the well-known finance disclaimer applies here; “Past results are no guarantee for the future”. But they definitely shape our decisions about the Forewarned is fore armed!