The Holy Grail in equity markets

The 'problem' with all long-only equity portfolios is that no matter how good your portfolios manager is or what strategy you choose, the portfolio will suffer from the equity market cycles. A low-risk portfolio will fall less in a bear market, but it will fall.

We at DNB Systematic Equity Team just went live with our well-prepared implementation of a European Equity Market Neutral portfolio. This portfolio aims to generate strong, stable absolute returns with as low correlation as possible to equity markets. It's a portfolio that delivers outperformance that is not 'contaminated' with the market return.

Despite all the critics against active portfolio management, it is not very hard to identify portfolio managers with long and strong track records. You can find managers with many years of stable Information Ratios and alphas. Yet, the Holy Grails, the outstanding Market Neutral portfolios, are very hard to find.

I have no doubt that there are portfolio managers out there that generate significant true alpha. Some of them are individuals, some are well-organized teams, and some create alpha with quantitative methods. I am a quantitative portfolio manager, but contrary to what many would expect, I don't think quantitative managers are better at generating alpha than traditional managers, in general. I do think quants are better at risk management. And risk management is a real challenge in constructing a market-neutral equity portfolio. I think you can become a successful portfolio manager without sophisticated risk control. –I really don't think it is possible to make a true market-neutral equity portfolio without serious efforts to control risk.

Most active portfolio management requires an alpha model to generate outperformance and risk management to give the portfolio the characteristics it is meant to have.

Alpha model

The alpha model we implement for the European market neutral has been tweaked on for more than 25 years. It was published in a research article in 1995 and has been proprietary since. The last tweak is to take a model built to run on individual markets and implement it on Europe as if it was one market. European markets have become more synchronized, and the results were strong.

We use around 60 fundamental company characteristics to identify the investor-preferences that can explain the performance differences found in a period. We do this on monthly numbers to forecast what the preferences will be next month. An investor that can predict market preferences in the next month can exploit this by overweighting favorably positioned companies. –And the investor can underweight companies that don't fit expected market preferences. In a market-neutral portfolio, we sell them short. This model has proved to work well historically, in tests and long-only portfolios. One of the advantages of this model is that we cover all listed companies. There are no analysis- or analyst bottlenecks. To put it simply, we have an expected return for every listed company in the eligible investment universe. That is a huge advantage when it comes to market-neutral portfolios. Traditional managers will have to prioritize what companies to cover and probably cover only interesting or large companies. This creates a bias in itself, even before the analysis begins.

Risk model

Since everybody who discusses long-short portfolios always uses the term 'market-neutral,' one would think that we can at least agree on what it means. Most people seem to think it is enough to sell short the same $ value as you go long. That seems fair. It is balanced, neutral, money-wise. But what happens if your beta is different in the long- and the short-leg of the portfolio? If the sensitivity (beta) to market movement is different, you will have a bias towards- or against market movements. Your absolute return profile will most likely have a strong (positive or negative) correlation to market performance. It will appear as if you were not balanced money-wise.

A lot of people agree that beta needs to be controlled for this reason. Then there is volatility exposure. Markets tend to peak when volatility is low, and vice versa. So if your long- and short- portfolios have different volatility exposure, that would probably lead to correlation issues as well. Not so many managers control this because now this is getting complex, maybe at a level where traditional tools cannot help you. And there may be other factor exposures as well, that basically cause these sorts of problems.

It takes a capable implementation model and optimizer to control for such exposures. And you probably have to build it yourself, because commercial optimizers usually do not offer this functionality. Building one takes years, rather than weeks and months. To control for all these exposures, the model requires a lot of companies in the portfolios, which in turn demands comprehensive coverage by the alpha-model.

We have spent more than 25 years developing our implementation-model to do this. It is well tested. Our global alpha analysis coverage is more than 40.000 companies. That means that we have an updated expected return for all companies in the selection universe every month. We do balance our portfolios money-wise. We control for beta (sensitivity to stock market) and differences in volatility-exposure. We also control for factor-exposures, company characteristics, that we have found can cause problems similar to beta- and volatility.

We are excited to have started our European Market Neutral portfolio to join the quest for the Holy Grail. The strategy is now one of the ingredients in the "DNB Fund Multi Asset" product.