A Dynamic New Approach to Pairs Trading

Alberto Micalizzi applies real options valuation to pairs trading

STUART FIELDHOUSE
1

Much has been written about the bad showing many quant managers put in over August. While they wanted to believe otherwise, many funds of hedge funds and institutional investors were less than impressed when program-based trading strategies came unstuck. That is why it is refreshing to find a manager with a market neutral fund that has been performing consistently, not just since the fund's launch in 2005, but right through this summer's credit crunch crisis too.

Dynamic Decisions is based in South Kensington in London, allowing the firm's founder, Dr Alberto Micalizzi, ready access to business partner and mentor Dr Nicos Cristofides, Professor of Operations Research and Head of the Centre of Quantitative Finance at Imperial College, London University. Micalizzi hails from a strong academic background. While he has acted as an external adviser to Morgan Stanley in the 1990s, he cut his quantitative teeth as an academic, earning his PhD in Finance from Imperial College, and going on to hold the chair in Finance at Bocconi University in Milan. In other words, he is not the product of a quant desk at an investment bank.

Studying under Cristofides, Micalizzi developed the theory of Growth Premium Analysis (GPA) that uses real options valuation to make selections within a pairs trading universe. "The concept of real options is still in its early adoption phase, although it is viewed by those familiar with the field as the most exciting development in the world of finance over the last few decades," says Micalizzi.

Back in 1997, it was Professor Robert C Merton who won a Nobel Prize in Economic Sciences for his new method of valuing derivatives. Part of his Nobel Prize Address, later published as a paper, 'Applications of Option Pricing Theory: 25 Years Later', outlined the emerging theory of real options, incorporating a different theory of risk in asset valuation. Its underlying principle is that a company's share price is composed of two items: the base value which investors pay for the company's existing assets, and the premium they pay in expectation of future growth, namely the option component of the stock price.

"These two components are intrinsically different with regard to the business volatility," explains Micalizzi. "While the higher volatility decreases the value of assets in place, in accordance with a standard bond-like valuation principle, it boosts the firm's growth opportunities in line with the known positive relationship between the premium of a call option and the volatility of the underlying assets."

In other words, Dynamic Decisions is using option theory to model and predict the behaviour of an option-like component of the share price. Its model compares the premiums implied in two similar stocks andindicates which of the two is more undervalued. Its comparison is not made purely on the implied premiums, but on the adjusted version. Only the unexplained difference between the adjusted premiums provides the firm with opportunities for arbitrage, ie. taking a long position in the stock with the low adjusted premium versus a short on the other half of the pair.

Not just a mathematical model

Micalizzi stresses that his GPA model is not just a mathematical model, it is a way of thinking. "What is happening in finance is a revolution in terms of understanding the role of risk in the pricing of companies," he explains. "A share price is partly an option – even if investors are not always conscious of it, they are paying for assets of a company already in place, and for its future growth. Any time you buy a stock, you are effectively buying an option for future growth – and these two parts have completely different dynamics."

The GPA model sits at the core of Micalizzi's pairs trading process. Although using the same, publicly available, variables as many other funds out there, Dynamic Decisions approaches these quite differently. First off, it takes the view that the volatility of a business may be a positive driver of stock price. The majority of earnings discount models used in the investment industry are based on what Micalizzi calls "a deterministic-negative relationship between price and risk."

In addition, the model also takes R&D as an input for the growth option implied in the stock price. Research and development expenditures are not computed in a linear relationship with the price. How much does R&D expenditure contribute to the growth premium, the system asks, and how does this compare with the same factor in a peer company?

"Companies cannot distinguish between money spent early on in the product development process, and money spent further down the road," Micalizzi argues. "In the 1990s there was much debate as to the impact of R&D announcements on a company's share price. In many cases the conclusion was that this produced little effect on the stock price. However, further work on this subject during this decade has proved that it is 'incremental' R&D and how it is viewed that market investors discount in the price over a period of time."

The GPA model also takes into consideration the company's cash position, but here, instead of treating it as a component of the book value of that firm, the model treats it as 'exercise capacity' – the degree to which a company can turn growth options into assets. In some respects there are similarities here with the way managers specialising in highly research-driven sectors, like biotechnology, analyse their investments. Applying it to highly traded large caps on a pairs basis, however, is breaking new ground.

The making of Dynamic Decisions

Dynamic Decisions, as a firm, was established in 2001 to provide advisory services for major institutional management houses, including independent equity research. In 2003-04 it was advising a Luxembourg SICAV on the construction of a market neutral strategy, long equities and short the NASDAQ 100 universe. With this fund achieving 16% in less than two years, it was becoming increasingly evident that Dynamic Decisions should try its hand at money management. "It was obvious that our stock selection had proved to be effective," Micalizzi says.
 

Thus, in 2004 he founded Dynamic Decisions Capital Management as an FSA-registered investment manager, having taken the decision to apply his theories to practical money management. Its first and thus far only fund, DD Growth Premium, was launched in January 2005 with $5 million. The fact that it now has over $310 million in assets under management, most of that received after it completed its 24 month trading track record, is testament to the growing confidence of the investor community in the Micalizzi formula. "August – and our lack of correlation with the market – has been a strong catalyst for us," he says.

Micalizzi's operation feels more like an academic department than a traditional hedge fund operation, but it is these credentials that investors are buying into. Its quantitative research and development team is composed of seven individuals, including a unit of four PhDs headed by Professor Christofides at Imperial College, which focuses on quantitative and statistical research. A fundamental analysis team mainly based in Milan and led by Luigi Micalizzi looks after the firm's equity database, while Niall MacDougall the COO oversees the trading team, led by Paolo Belvederesi, and the back office in London. Finally, Marta Renzetti fulfils the CFO function, providing vital administrative and operational back up.

The firm's performance numbers have been solid, with a 10.73% net return for its US dollar class over the 12 months to July 2007, and 16.5% for its 1:1 US dollar levered class. It has a low correlation with the S&P 500, at +0.25.

Micalizzi sees market volatility and correlation among stocks as the two most important drivers of his fund's performance. For example, in January and May 2006, despite the opposite equity market trend, DD Growth Premium put out a 0.3% net return. The correlation among stocks was higher than average during these months – typically, during both strong bull markets as well as sell-off periods, investors buy or sell with less regard for corporate fundamentals. This higher correlation is taken as a negative factor for the fund's managers, whether the market is up, or down overall. The first two months' performance seems a little uncharacteristic given the fund's overall numbers. Micalizzi is quick to point out that this was the result of a flaw in the risk management process that did not take into consideration extraneous factors, namely M&A activity. "The initial strategy consisted of 20 pairs and the March and April [2005] performance was hit by a few unexpected earnings surprises," he says. "We decided to increase the number of pairs to 50 to diversify the risk, and more importantly, go 'event neutral'. Since these changes were implemented, the fund has not experienced a negative month."

Now, the fund uses analyst signals as indicators that it should stop trading a company because there is a rumour a company may be in play. It uses one analyst for each sector it covers within the universe, and at the first sign of a potential takeover, a pair will be taken out of the portfolio. This helps the strategy to preserve a high level of mathematical purity. "We try to hedge out anything that is not part of the business characteristics of the company. We hedge out the exogenous sources of risk in the market," explains Dynamic Decisions' founder.

DD Growth Premium requires plenty of gigabytes to run, as it has to rank 180 pairs in two hours. "Our turnover does not involve frenetic intra-day trading," says Micalizzi. "On the first day in the month, the fund rebalances." The current universe of pairs is composed of the largest stocks in the S&P 500 and the EuroStoxx600. These have the benefit of a high level of liquidity – the entire portfolio could be turned into cash in a couple of hours. The 175 pairs are ranked using a return per unit maximum drawdown, and the top 60 are chosen and weighted on a VaR adjusted basis. Each pair is tested against four 'neutrality policies', namely that it is sector neutral (based on Bloomberg's industry group attribution – eg. Nokia and France Telecom cannot be paired up); it is neutrally sized (the ratio between the market cap of the long and short leg of the pair is less than two); it is beta neutral; and it is earnings neutral.

"For companies reporting their earnings, we implement the fourth policy," says Micalizzi. "Discipline is most important here, there should be no discretionary element, no human intervention dictating changes. The important parameter is to minimise the drawdown."

The future

It seems as if Dynamic Decisions has done well out of August's quant meltdown. Micalizzi is positive about his firm's room for expansion too. For starters, DD Growth Premium is trading in a very liquid area, and can accommodate up to $2 billion in its master fund. However, given that it has tripled its assets under management this year, and is taking in between $30 million and $50 million every month, this may not take as long as some might think. In order to facilitate access to DD Growth Premium, Dynamic Decisions is also launching Luxembourg and Delaware feeder funds over the next couple of months.

In addition, the firm is also testing other strategies which it might consider applying to this universe, although these remain very much at the R&D stage with Dynamic Decisions' team of academics. A new launch, taking the form of a multi-strategy quant approach, could be in the offing. This will combine a number of the best strategies in an effort to improve on the current information and Sharpe ratio of DD Growth Premium.

Ultimately, Micalizzi is practising what he preaches. He firmly believes he can make Dynamic Decisions an established and recognised quantitative player on the hedge fund stage. He is also investing substantially in new talent from both the academic and investor worlds. He has taken his emphasis on R&D to heart, and is re-investing a large portion of his revenues in furthering the firm's internal product development efforts in order to continue to expand and bring more of his academic capital to the market.

Dr ALBERTO MICALIZZI is founder and chairman of Dynamic Decisions Capital Management, he is the author of the Growth Premium Analysis model. He has been Professor of Finance at Bocconi University (Milan), a researcher at Imperial College, and is the co-author of the book 'Real Options Applications.' Prior to founding Dynamic Decisions, he was the CEO of Real Options Group. He received his PhD in Finance from Imperial College and graduated with Honours in Economics from Bocconi University. He holds the IMC designation and is a member of UKSIP. He is also a CPA Charterholder.