Gideon Ozik «funds with extensive media-coverage tend to underperform»

According to a recent research paper released by Gideon Ozik & Ronnie Sadka, US Hedge funds tend to underperform following extensive media coverage

Article also available in : English EN | français FR

Co-author of the study and former head of hedge-fund solutions at SG, Gideon Ozik, gives us more insights

What was the motivation of the study?

I have been actively involved in hedge fund management in recent years and have recently developed strong interest in research. I feel that my experience, brings a relevant perspective. As for the media-coverage story in hedge funds, Ronnie and I had two objectives; first, the internet intensifies the flow of information and is reshaping the way news is reported. Therefore, understanding how media (especially electronic) interacts with the fund management process is increasingly important. Secondly, we were motivated by similar evidence in equities, i.e. stocks with media-coverage tend to underperform no-media-coverage stocks. The explanation for this effect in equities is based on investors willing to pay a lower price (and expect to earn higher returns) for stocks that are “under the radar” (hidden gems). We find that funds exhibit similar patterns, and therefore, the explanation attributed to equities may not hold (because fund recognition should not determine its price rather the assets it holds).

How do you explain your results?

The results are quite interesting. Although the evidence is very strong, we were not able to quantitatively pinpoint a single reason for it. The main finding is that funds with extensive media-coverage tend to underperform. But we also found that high media-coverage is followed by outflow, and fund liquidation creates some pressure on prices. However, that on its own does not explain the entire underperformance of the high-media-coverage funds. Our paper reports that media-coverage is more intense in past extreme performers. Funds which have recently performed very well or very poorly, exhibit higher levels of media coverage. What is interesting is that media treats “star-funds” and bad-funds alike --- both when receive media-attention, perform worse than similar funds with low media-coverage. This study can teach us a lesson. Take the corporate world for example, academic research shows that super-start CEOs perform poorly after they receive extensive media-attention. One study even documents that after winning prestigious awards, CEOs significantly improve their golf handicap (they spend way too much time on the golf course). In hedge funds, where the management process requires a lot of time, attention and real-time risk monitoring, the great managers who turn celebrities may just stretch their attention to thin.

In addition to the negative effect on returns, we found that media-coverage helps predicting flow
Gideon Ozik

Can you give more details about the flows effects ? what about closed or locked-up funds?

In addition to the negative effect on returns, we found that media-coverage helps in predicting flow . As it turns out, hedge fund investors are attentive to the media and react to it by changing their exposures. When grouping funds by past flows, we find that most of the media-coverage prediction is concentrated among the middle funds not the extreme ones. The ones with extreme flows may actually have other things (other than media-coverage) affecting their returns.

As for locked-up funds, our study focuses on equity hedge funds (about 1,000 Long/Short Equity) which are typically among the more liquid in the universe of hedge funds. Nonetheless, we found that the media-coverage discount appears in both the locked-up and non-locked-up subsets. It will be interesting to investigate the media affect in less liquid hedge funds.

Your analysis focused on North American equity hedge funds, why such a geographical localization ? and moreover, what about long only or mutual funds ?

Theoretically, we should expect the same results regardless of the geography or maturity of the market. However, the news tool which we relied on aggregates English-based news items mostly from North America. Therefore, we had to restrict the study to North American funds. Once the news archive coverage expands, we would definitely attempt to test the media-coverage story in France and around the world. The note about mutual funds is actually very interesting because it provokes the question whether the phenomena is hedge-fund specific or more universal. Our study shows that media-coverage predicts lower hedge fund returns. Earlier studies show that media-coverage also predicts lower returns in the cross section of stocks. Furthermore, there is evidence that managers of companies (not necessarily financial-companies) who receive abnormal media-coverage tend to underperform their peers. Taken together, these evidence may point a universal pattern---increasing the likelihood that that mutual funds are no different. However, we have not yet checked it and it is definitely on our research agenda.

Regarding the data, were you able to isolate the effect (if any) of fund’s voluntary released news ?

For this study, we counted the number of news article per fund using Google News Archive. Google News Archive aggregates different news stories from various media-sources including press releases which are often pushed by the funds, and stories which are independently published by news outlets. Disentangling the sources is a natural avenue for further investigation. We also try to understand what determines media-coverage and found that large funds typically receive more media attention than small funds and that funds which charge higher fees generally get more media coverage. There are two points to note, first, the media-coverage spread holds in various size groups. For example, within the smaller funds which typically get less coverage on average, the ones with relatively higher coverage (within that group) underperform the lower-coverage funds. Secondly, the fees and size together provide a clue for fund’s budget. It would not be unrealistic to guess that high budget funds spend more on marketing and as a result get additional media-attention.

Given the current level of transparency requirement needed by investors, don’t you think that communication and to some extent media-coverage are now embedded in investment process?

Fund managers have regulatory requirements as well as client-communication responsibilities. Our research does not take a position of whether this communication is good or not. We simply point out that extensive media coverage predicts lower returns.

Media coverage is an important piece of the puzzle. Therefore, analysts should not ignore it
Gideon Ozik

Should fund analysts look at media coverage as an external input or as an internal variable in their work?

Media coverage is an important piece of the puzzle. Therefore, analysts should not ignore it. Importantly, our study finds that some well-known anomalities of hedge funds are more pronounced in high media funds. For example, hedge funds tend to exhibit return continuation (positive auto-correlation), our study finds that continuation is much stronger among media-coverage funds. These results hold on average and are especially strong within recent bad performers (who continue to do even worse). In earlier study, Ronnie and I show the “smart money” phenomena in hedge funds – that fund-flow predicts fund returns. In this media-coverage study we find that smart money stronger among high-media funds. Taken together, analysts, investors and fund of funds managers should take into account the particularity of investing in high-profile funds.

What about rumours or media manipulation?

The area of media-coverage and information is vast. In order to study rumors and manipulation of the media one has to have a finer data set, ability to infer sentiment and the knowledge of the origin (press release initiated by the hedge fund or independent media-reporting by the news outlet). In the current study we did not have access to some of these parameters but we are definitely working toward improving our technology hoping to answer some of these additional questions.

Next Finance March 2010

Article also available in : English EN | français FR

See online : Does Recognition Explain the Media-Coverage Discount? Contrary Evidence from Hedge Funds

Read also



In the same section