London (HedgeNordic) – Systematic and quantitative are interchangeable terms. They involve using a model, formula, rule or algorithm, rather than discretion or judgement, to decide on a trade. Many discretionary managers do of course use systematic models as well, but they exercise judgement to pull the trigger on trades. Computers certainly allow for more sophisticated systematic strategies, but the first systematic traders used pens, ink and paper to draw charts, centuries before the first computer.
“The earliest systematic trading may have been in 18th century Japan, where candlestick chart patterns were used to trade futures on the Dojima Rice Exchange in Osaka, which started in 1730.”
The earliest systematic trading may have been in 18th century Japan, where candlestick chart patterns were used to trade futures on the Dojima Rice Exchange in Osaka, which started in 1730. The charts were reasonably advanced: they measured opening and closing prices, intraday highs and lows, and used black bars to show rising prices combined with white bars to show falling prices. Chart patterns identified included the shooting star, hanging man, and dark cloud cover. US futures markets started about a century later in 1848 with corn, and later wheat and soybeans traded on the Chicago Board of Trade, soon followed by cotton on the New York Cotton Exchange, while the London Metal Exchange was launched in 1877. However, systematic trading of futures in the West is generally held to have started in the 1970s, when the “Turtle traders” led by Richard Dennis, applied trend following to the growing universe of futures markets. The earliest CTAs included Campbell, Millburn, and Eckhardt, in the US, while firms such as AHL, which later led to Man Group, Winton and Aspect, started a bit later in the 1980s in Europe. The advent of currency futures on the Chicago Mercantile Exchange, after the Bretton Woods system of fixed exchange rates collapsed in 1971, had increased the number of markets to trade, and the associated end of the gold standard also paved the way for trading gold, silver and platinum futures. Financial futures in the form of Treasury bond and equity index futures came later. The universe of futures contracts is growing every year as more contracts are launched on both western and emerging market exchanges, where China has been especially active, though trend following can of course also be applied to non-futures markets.
“A different approach – factor investing – has since the 1970s has been based only on fundamental data.”
If systematic trading of futures emphasized trends, in equities one of the earliest approaches – statistical arbitrage – did the opposite: it was about short term mean reversion and pairs trading. If two stocks in the same sector moved in the opposite direction, a simple stat arb approach would short the riser and buy the faller, expecting them to reconverge. Whereas trend following generally made most of its money from a few big winners in any given year, statistical arbitrage was more about compounding up lots of small profits The quantitative trading group at Morgan Stanley led by Nunzio Tartaglia was one of the first to trade stat arb in the 1980s, and some members of this group, such as David Shaw, started their own firms. Over time statistical arbitrage has often become a higher frequency strategy pursued by specialist proprietary trading firms as well as hedge fund managers.
The application of systematic approaches to credit markets has been much slower, partly because they have taken much longer to move towards exchange and electronic trading. Over the counter (OTC) trading of credit markets for many years meant that quantitative analysts did not have good enough data to build models. Different counterparties quoted different prices and credit derivative contracts were also rather bespoke. Since 2000, several developments have steadily paved the way for systematic credit strategies. TRACE has since 2002 recorded corporate bond prices, CDS has been standardized since 2009, clearing of credit derivatives also helps to mitigate counterparty risks, and some markets can now be traded electronically. The assets in systematic credit are still tiny compared with equities, but this is a growing space that has been a rich source of alpha generation and is evolving fast. Historically the fact that different corporate bond issues could have different features was seen as an obstacle to systematic credit strategies, but now artificial intelligence can be taught how to read a bond prospectus or indenture and pick out at least some key features.
FUNDAMENTAL AND ALTERNATIVE DATA
Some of the first systematic earliest approaches – both trend following and statistical arbitrage – used only technical or price data, and ignored fundamental data such as economic figures, company profits, or valuation. A different approach – factor investing – has since the 1970s has been based only on fundamental data, such as using the value factor to buy a basket of companies with lower valuations, higher growth rates, more predictable cashflows or other qualities. Quantitative fundamental approaches have often run market neutral portfolios of equities, based on a range of signals. In macro investing, the largest and most famous systematic macro fund is Ray Dalio’s Bridgewater Associates. Increasingly, alternative data, such as sentiment, news, social media, or satellite pictures, is also being used as an input for systematic strategies – and machine learning statistical techniques are also being used to analyse the data. Some systematic managers are purely technical, purely fundamental or purely based on alternative data while others will blend two or three data types.
Though some systematic strategies – and especially higher frequency ones – do require electronic execution this is not an essential feature. In the early years, CTAs used models to generate trading signals, but trade execution would be carried out manually, probably by traders shouting and screaming at each other in the pit. The growing electronification of financial markets means that trade execution is now automated using algorithms, with the exception of OTC (over the counter) markets. Some of the most profitable trend following CTAs in recent years, including Man AHL Evolution, Systematica Alternative Markets, Gresham Alternative Commodities, and Brummer affiliate Florin Court, trading “alternative markets, which may be entirely systematic in terms of their models,” need teams of people to execute trades with large numbers of counterparties. It remains to be seen if artificial intelligence will eventually be deployed to automate the trading of OTC markets.
Title Pic: (c) Joseph Sohm—shutterstock.com