This is the third in a series of posts written by Frank Smietana, an expert guest contributor to QuantStart. In this detailed post Frank examines the different algorithmic trading strategies carried out by quantitative hedge funds. Click for parts one and two.
Institutional asset managers specialize in a particular asset class, style, sector, or geography, based on their expertise or domain knowledge. This is reflected in the investment products they offer their clients. For example, stock funds managed by style include growth and value offerings, geographic-based funds would hold only US, European or emerging market equities, and sector offerings might focus on technology, healthcare or financials.
A parallel exists for quant funds that build trading systems for various timeframes (millisecond, intraday, and multi-day), asset classes and instruments (individual stocks, index options, ETFs, or soybean futures) and style (trend following or mean-reverting).
This leaves quants with hundreds of potential combinations to choose from when developing systematic strategies. Additionally, a robust trading idea designed for a particular timeframe, instrument, and asset class can sometimes be “cloned”, and then applied to a different timeframe and/or asset class.
This type of fund attempts to detect and confirm valid market trends, then initiate and maintain positions in those markets as the trend moves in their favor. Perhaps one of the oldest and conceptually simple trading ideas in existence, trend following funds tend to hold positions over timeframes ranging from days to months. Trend following strategies generally don’t have exceptionally high win-rates, but catching a few big trends each year can handily erase the many small losses.
While emotional fortitude and high frustration tolerance are required to be a successful trend following trader, a fund that follows this strategy requires investors that also share those traits. Equity and rates markets have been ideal for trend follower’s post-2008, underpinned by implicit backstops from central banks. With prices seemingly only going up, aside from a few one day risk-off events, trend followers have been amply rewarded with minimal drawdowns.
There are numerous approaches for building systematic trend following systems, including moving average crossovers, volatility breakouts and oscillators; all topics for future blog posts. Trend following also requires tradeable instruments that can be used to profit from downtrends as seamlessly as uptrends. Shorting equity shares is a cumbersome process that is impossible to back test accurately. Shares may not even be available for shorting, or brokerages holding borrowable shares may demand a significant premium to do so. Futures and options are far more appropriate instruments, but ETFs and inverse ETFs are also increasingly viable instruments.
These types of strategies attempt to pick tops and bottoms, or take short term positions during price reversals from the dominant market trend. Timeframes tend to be shorter than trend following trades, but the intensity of a countertrend move is often fast and violent, especially when the predominant trend results in a crowded trade. The sudden reversal shakes out weak conviction and undercapitalized traders, adding fuel to the countertrend move.
Systems for countertrend trading are typically based on extremes in common technical indicators, including RSI, stochastics, and Bollinger Bands.
“Stat Arb” provides a number of viable system development ideas for quantitative traders. These strategies have broad appeal, in that they can be implemented across a broad range of asset classes and instruments, including equities, ETFs, convertible bonds, futures, and options. Furthermore, some implementations (mostly futures-based) offer significantly lower margin requirements than trend following strategies.
Essentially, stat arb seeks to exploit a mean-reverting trend in prices between related securities, or baskets of securities. A number of variations exist, also referred to as equity market neutral, spread trading or pairs trading strategies. A common approach is to select two groups of equities, with the “long” group expecting to outperform and the “short” group expecting to underperform. The groups are then simultaneously bought or shorted, and held until some degree of price convergence takes place. Note that each group can consist of a single equity, hence the reference to pairs trading. The groups could also involve trading an index against an ETF or a single stock against an index future such as the FTSE100 or the S&P500. Properly constructed, this trend minimizes exposure to market, geographic, sector and currency risks, which explains the “market neutral” label.
Key to systematic implementation of pairs trading is the notion of co-integration. Closely related to correlation, co-integration is a measure of how closely two securities move together. An accessible primer with sample R code can be found here. When this spread exceeds some historic threshold, the strategy signals a trade entry. Efficiently finding tradeable co-integrated pairs within a collection of liquid, short-able securities is an interesting computational exercise. A detailed approach can be found here. [MHM: QuantStart has articles on time series analysis and cointegration using the R environment.]
This strategy employs convertible bonds, which is a misnomer, in that these bonds are actually hybrid securities that can be converted into common stock at a pre-determined time and price. An arbitrage opportunity exists when traders spot convertible bonds that are mispriced relative to the firm’s common stock. The strategy exploits this opportunity by shorting the stock while buying the convertible bond. If the stock falls, the short sale is profitable; while the bond declines less than the stock, yielding a net profit. If the stock rises, the bond is converted into stock at a preferential price, offsetting the loss in the short position. Convertible arbitrage is not as simple as it sounds, nor is it risk free. Realizing a profit can involve significantly long holding periods, tying up capital that could be utilized for shorter term opportunities. A detailed overview can be found here.
This strategy can be implemented at scale in mature government bond markets with liquid underlying derivatives such as interest rate swaps or futures. Using US Treasuries as an example, the “basis trade” involves selling/buying US Treasury Futures, and buying/selling a corresponding amount of the deliverable Treasury Bond. Conceptually simple, this strategy involves a fair amount of calculation to implement correctly. Learning the details and pitfalls of this strategy is well worth the price of the book.
Another popular fixed income strategy is the yield curve flattener/steepener. To streamline implementation and minimize trading costs, this strategy is typically implemented using Treasury Futures. This strategy strives to capture changes in yield curve differentials, typically between the front end of the curve (2 or 5 year Treasuries) and the back end (10 or 30 year Treasuries). A flattener is profitable when the yield differential narrows, and is implemented by selling the spread (shorting the front end whilst buying the back leg). Conversely, the steepener makes money when the yield differential widens, implemented by buying the front end and shorting the back end.
The CME Group has posted an accessible primer on the calculations required to implement this strategy successfully.
These are another great source of systematic trading ideas, and can be efficiently implemented with some degree of scalability using liquid futures contracts. These strategies attempt to capture mean reversion profits by exploiting seasonal trends and economic anomalies. Note that these tend to be long term trades, typically requiring weeks or months to play out, and some may only be tradeable once a year due to seasonality. A brief overview can be found here. Three variants exist:
While most quant funds will focus on particular markets and asset classes; based on capital, system development capacity and domain expertise, it’s important to separate strategies from asset classes. Whether a fund focuses on emerging markets, US equities or global macro opportunities, they are utilizing one or more strategies, discussed above, to implement trades in those asset classes.
Emerging Markets generally exhibit greater volatility and geopolitical risk, but also significant “first-mover” advantages for funds that can spot new arbitrage opportunities in mispriced assets or enter trends before they become validated by massive capital inflows. Emerging markets often suffer from liquidity constraints and under-developed derivative markets, making it more difficult to efficiently implement common “First-World” trading strategies.
Global Macro funds focus on macroeconomic conditions and disparities, exploiting interest rate and currency differentials. These funds often use the popular “carry trade”, implemented by taking a long position in a relatively higher yielding security financed by a short position in a lower yielding security. “Carry” refers to the yield spread.
This article has provided a brief overview of popular quant funds and strategies, and the instruments and asset classes typically utilized in those strategies. While all of these strategies are conceptually simple, gaining a deep understanding of the details is critical to success. Patience is also important. Some trade ideas only provide an occasional entry opportunity. Larger quant funds with sufficient capital and development resources typically deploy a portfolio of the strategies described above, allowing them to keep capital deployed continuously as opportunities present themselves.comments powered by Disqus
You'll get instant access to a free 10-part email course packed with hints and tips to help you get started in quantitative trading!
Every week I'll send you a wrap of all activity on QuantStart so you'll never miss a post again.
Real, actionable quant trading tips with no nonsense.