Professor Paul van Rensburg has formed his own asset management company, Salient Quants, as of 1 July, taking with him his long/short equity

The SalientQuants SA Hedge Fund employs a statistical arbitrage strategy to invest in Johannesburg Stock Exchange-listed equities. It has returned 23.45% since inception last August, with no down months (see South AfricaHedge July/August 2006).

The fund now has R81 million under management, including seeding from Hedgehog Capital, which has until now had sole access. Van Rensburg will begin talking to other investors this month and is looking to soft-close the strategy at around R400 million.

The fund uses multifactor stock selection models, diversified across more than 15 factors at any one time, to accumulate a range of small positions. Van Rensburg will typically have up to 100 active positions spread over both the long and short sides.

Van Rensburg's proprietary quant process is the result of years of ongoing research. He is the University of Cape Town's professor of finance and has been active in asset management since 1998, including as head of research at Futuregrowth before joining Kagiso as senior portfolio manager and head of equities.

Kagiso continues to use Van Rensburg's quant system under licence on the long-only side, with more than R12 billion under management. The Kagiso Active Quants Fund remains the best-performing unit trust in the country of the 65 funds listed in the general equity sector.

 Kagiso Active Quants Fund: Leading general equity fund over 12 months

The Kagiso Active Quants Fund, classified as a domestic general equity fund, was launched in April 2004. It was the first South African 'quants' fund available to retail investors. The fund manager, Paul van Rensburg bases his stock picking decisions on a statistical model of share returns. Despite good performances from the fund the assets under management of the fund are currently only R10.4 million.

Equinox: Could you tell us a little about your investment experience?

Paul van Rensburg: I started my career working in the quants division at RMB in 1998, and then spent four years as Head of Research at Futuregrowth. I joined Kagiso as Head of Equities in January 2004. Simultaneously, I have had an academic career and am the Professor of Finance at the University of Cape Town. Indeed, my many years of academic research underlies the quantitative multifactor stock pricking model that I use.

Equinox: Could you tell us more about Kagiso Asset Management?

Paul van Rensburg: Until fairly recently Kagiso Asset Management was jointly owned by Coronation Fund Managers and the Kagiso Group. Within the last three months things have changed and the Kagiso Trust owns 70% of the company, with the staff owning the balance. We are an empowered company with 90% black shareholders. We have R4.3 billion under management, most of which is institutional money. We have made no serious efforts to target or advertise to retail investors.

Equinox: What is the mandate and benchmark of your fund?

Paul van Rensburg: The benchmark that the fund tries to outperform is the average performance of general equity funds over a rolling 12 month period. We use equally weighted average holdings in the other general equity sector as a starting point for the construction of our portfolio and take active positions around that. Clearly, the fund is not a passive quants fund and we aim to covert our view, if correct, to moving up the performance rankings within the General Equity sector.

Equinox: Do you use the average positions of the general equity unit trusts as the starting point of some of your institutional funds as well?

Paul van Rensburg: On the institutional side each client typically specifies their own benchmark and we try to outperform whatever they choose. No other funds use this particular benchmark.

Equinox: Your fund is currently the top performing general equity fund over a 12 month period while competitor quants funds in the general equity sector are, according to our records, placed 25th, 34th and 35th over the same time period. Could you explain why you think your fund has achieved this performance?

Paul van Rensburg: I would attribute our success to good stockpicking. Regarding our 'competitors', I am not sure but I would guess that they are either passively managed quants funds, or have recently changed their mandates and do not have a full 12 month performance history of active management. A passive fund aims to track an index, where a fund like ours that is actively managed aims to outperform an index. We see our real competitors as the top performing active managers.

Equinox: Could you clarify exactly what you mean by tracking error?

Paul van Rensburg: Tracking error is the standard deviation of the difference in returns between your fund and its benchmark. Tracker funds, for example, tend to have a very low tracking error while active funds may not even care about any benchmark. However in my case, the investment strategy is different: I hug my chosen index during volatile or unpredictable investing environments and deliberately push it out, take more risk relative to the average general equity fund when we are more confident. For, the last few months (which were like a washing machine) my tracking error has been reduced to 3.5% p.a. and only recently have I started to increase it up to 2004 levels again.

Equinox: How do you look for good share prospects?

Paul van Rensburg: The multifactor quant model predicts the next months' on abroad range of JSE listed shares returns and if something looks good here we take a closer look at it. We have done some recent screening rule research and have filters that alert us to possible opportunities

Equinox: Do you overlay your quants-based stock picking with qualitative information or are you a quants purist?

Paul van Rensburg: We definitely overlay our final decisions with some qualitative insight. The numbers cannot give you an indication of a corporate event, for example. I would say our models are responsible for between 80 and 90% of the tilt of the portfolio with the balance being a qualitative overlay. After managing portfolios such as this for a while you get a feeling about when to bet on your numbers and when to doubt them.

Equinox: Do you generally run a fully invested portfolio?

Paul van Rensburg: Yes. As you know we are allowed 25% of the portfolio in cash but we keep as little cash as possible.

Equinox: Your marketing material says that Kagiso uses 'leading-edge' quantitative investment thinking. Must we assume from this that there are some quantitative methods that have become outdated or unfashionable?

Paul van Rensburg: I wouldn't say that certain quants techniques become outdated, but there are certainly ongoing refinements, improvements and new ideas. However our research is always ongoing – I found out this week that we were lucky enough to win the latest Investment Analysts Journal best paper prize and our research is always in some way relevant to building better models.

Equinox: How would you seek to reduce or minimize negative returns in a falling market?

Paul van Rensburg: We are a general equity fund and we would rise and fall in line with the performance of equities. The fund is not an absolute return fund.

Equinox: Do you think your quants-based portfolio construction methods will hold, or be sustainable if the fund were to grow substantially in size?

Paul van Rensburg: Definitely, although certain small-cap opportunities would unavoidably be lost. As mentioned, our main clientele tends to be larger institutional mandates.

Equinox: Please could you describe the details of your performance fee structure.

Paul van Rensburg: There is an annual service fee of 1.14% (including VAT), and the fund manager earns a performance fee of 15% of the out performance of the fund if the performance of the fund exceeds the benchmark. The annual fee reduces to 0.79% (including VAT) if the fund manager underperforms the benchmark. The performance fee is capped at 2.5%. This means that the total fee will no be greater than 3.5% p.a. (1% base fee +2.5% performance fee) excl VAT.

MONEYWEB: Our guest market commentator this evening is Paul van Rensburg. He’s head of research at Futuregrowth Asset Management. Paul, I understand that you were lecturing at the University of Cape Town not too long ago. Everyone talks about the intense volatility that we have these days. The whole investment philosophy – do you think it has changed in what you are teaching at this point in time?

PAUL VAN RENSBURG: As you know, the markets are in constant change and in the research we try and identify those sort of salient facts that persist. But yes, definitely we are in the situation where certain characteristics tend to be associated with better performance than others, as is always the case. We can apply this type of research to the markets at the moment and identify what characteristics are associated with better performance. At the moment I think high dividend yield stocks tend to be rewarded pretty well. I would go for the smaller stocks rather than the larger, particularly in the industrial sector. I think that some of the larger stocks have larger betas and bigger responsiveness to the Dow, and I think the Dow is still not over its woes. With the shares with high momentum, that’s good price behaviour over the last 12 months, with growth those factors tend to be rewarded. As well as positive [indistinct] etc, which are more salient factors that tend to be associated with higher performance.

MONEYWEB: And there’s this big debate that always goes on about efficient markets and they say that our markets are always efficient in the end. I ‘m sure you’ve done a lot of research on the issue. What do you think? Are markets as efficient as people like to believe?

PAUL VAN RENSBURG: Yes, we’ve done a tremendous amount of testing. As one example, I recently supervised a guy’s doctorate, testing to see what characteristics are associated with outperformance. And I believe, if you look at the data, the reality is quite different from what we teach with higher betas and higher returns, as in the capital asset pricing model. People I don’t believe are rational and there are cases of over-reaction and mean reversion in the longer term. People tend to over-react, which results in momentum type of effects and that later leads them to value type of effects.

MONEYWEB: What do you mean by momentum and value effect? Start off with momentum effect – what does that mean?

PAUL VAN RENSBURG: Momentum effects would be, for example, the shares that have done well over the recent past and that continue to do well. And that would be measured by the past price performance for example in the last 12 months. So there is a habit that they tend to run, and that would be related to the over-reaction type of phenomenon.

MONEYWEB: Some people would also call it “herd investing”?

PAUL VAN RENSBURG: Oh yes, there’s definitely a herd effect. I mean, people see how their shares have done over the last while and, if they’ve done very poorly, will they tend to sell. And if they have done pretty well, they tend to carry on and invest more. Typically, the efficient market argument would say that, if there is an anomaly, the rapid reaction of investors should remove that. But you can imagine that if there is a price momentum effect – i.e. that shares that carry on running while trying to exploit that effect would merely exacerbate that effect. To a certain degree our research definitely shows that that exists. We have developed a quantitative investment process that tries to exploit simultaneously a number of these effects.

MONEYWEB: Are you getting it right?

PAUL VAN RENSBURG: Oh, yes. Our performance has been excellent. For two years we have been running a quantitative process. I must make the distinction between passive quantitative, which is essentially index-tracking, and active quantitative, which is actually trying to take best, based on these statistical models.

MONEYWEB: What are you benchmarking yourself against, to say that you are doing well?

PAUL VAN RENSBURG: We have various benchmarks, but generally, if the client gives us a benchmark, if we’ve outperformed that particular benchmark, then we will say we are doing well. And we generally measure whether we are going well by outperformance relative to the tracking error with regard to that benchmark. All funds that we’ve run under the quantitative process have outperformed with information ratios, that’s outperformance divided by tracking error of between 0.5 and 1.5.

MONEYWEB: Banking stocks – quite a few market commentators on this programme reckon that they are quite undervalued, time to pile into those. What’s your view on that?

PAUL VAN RENSBURG: People can see now the peak in interest rates and people have been punting banking stocks for a long time. Some of them like Absa might be a little undervalued, etc. With the strengthening of the rand, this is the right time of the cycle. But I reckon there’s better value and our model tells us to tilt elsewhere in the portfolio.

MONEYWEB: Like where?

PAUL VAN RENSBURG: At the moment, I would prefer stocks that have the characteristics that I mentioned earlier, this high dividend yield, small with good momentum, particularly in the industrial sector. After the crash in ’98, the small caps really got hammered and those survivors, particularly those ones that are starting to show some price strength, are still under-valued. And if I had to name some stocks, I would like shares – and we have really long positions in these shares – shares like Cashbuild, City Lodge, Group Five, Wilson Bayly, Rainbow had some good results, Omnia, Netcare, MediClinic, Hudaco. Those type of shares, they are the cheaper, high dividend yield, industrial-type shares. They’ve got the characteristics that the market is rewarding at this particular moment in time.

MONEYWEB: A long list of shares that Paul van Rensburg fancies. He’s the head of Research at Futuregrowth Asset Management.

PAIA Manual