An example of a time series would be the price of a stock over time in days or population over time in years. We’ll now turn our gold and gold miner relationship into a pair trading strategy. From a market-neutral strategy perspective, there are fundamental economic relationships that hold. We’ve discussed one with the price of gold and gold miners, but there are plenty of others. While the model breaking down is the primary risk, there are many risks with each type of statistical arbitrage. Gerry Bamberger developed the first arbitrage strategy using pair trades trading at Morgan Stanley in the mid-1980s.

statistical arbitrage

Arbitrageurs require a positive expected excess return over the risk free to compensate for risk. The potential loss must be acceptably small in order to qualify the strategy as arbitrage rather than simple investment. Although not all the academic literature reports it, trades always have take profit and stop loss features. The take profit identifies when a trade no longer offers positive expected excess returns. A take profit is triggered in case there is reversion to the mean or when the positive carry disappears .

So does this mean that for the average quantitative strategist investors trading strategy must remain an investment concept of purely theoretical interest? Firstly, for the investor, there are plenty of investment products available that they can access via hedge fund structures . This is but one way of adding machine learning methodologies to the mix. Even if a statistical model has been successful in the past, there is a chance that the market could change and render it useless in the future. Financial markets are always changing and evolving into different situations.

Today, most Currency Risk is conducted through high-frequency trading using a combination of neural networks and statistical models. Not only do these strategies provide liquidity, but they have also been largely responsible for some of the largest crashes we’ve seen in firms like LTCM in the past. As long as liquidity and leverage issues are combined, this is likely to continue making the strategy one worth recognizing even for the common investor. High-frequency trading is a relatively new development that aims to capitalize on the ability of computers to quickly execute transactions. Spending in the trading sector has grown significantly over the years and, as a result, there are many programs able to execute thousands of trades per second.


It is orthogonal to common risk factors, and exploits asymmetric local trend and reversion patterns. Our strategies remain profitable after taking into account trading frictions and costs. Our findings suggest a high compensation for arbitrageurs to enforce the law of one price. We show that the synthetic asset formed from the replicating asset and the original Berkshire A stock can give profitable entry and exit points for statistical arbitrage trading at different transaction costs. If investors want to get the maximum profits, they should look for the spread of pairs with high variance and strong mean reversion. The usual method is to construct a stationary, mean-reverting synthetic asset as a linear combination of securities .

statistical arbitrage

In the coming articles, I intend on delving deeper into the subject of quantitative trading strategies — if you have any requests or questions, please do not hesitate to share them with me in the responses below. Figure 2.4.1The residuals are the differences between the natural log of the price of stock A and the corresponding point on the regression line. Essentially they are turned into a residual series which also has to be stationary I. The way to do it would be by way of linear regression of the natural logarithms of the prices of the stocks A and B.

If the prices of the two stocks diverge, we short the winner and go long on the loser, hoping that these prices converge in the future. If the spread is mean reverting, it will revert to its historical mean. The “pairs” rule is sufficiently obvious that it could almost be implemented manually.

Tag: Statistical Arbitrage

That said, spreads of this kind can nonetheless be extremely volatile. If you want statistical arbitrage to work, you have to rely on your broker to execute your orders. Many times, you have to have split-second execution in order to profit from this type of trading strategy. In some cases, the broker will not be able to fill your trade and they will simply cancel the order.

  • In many countries where the trading security or derivatives are not fully developed, investors find it infeasible or unprofitable to implement statistical arbitrage in local markets.
  • His result also provides the optimal entry and exit points for arbitrage trading at a given transaction cost.
  • Thousands of strategies using tens of thousands of signals currently drive our live production, and our talented team of researchers from some of the best schools in the world inject new ideas into our system on an ongoing basis.
  • This outperformance was, however, found to be present in Sortino ratios only.
  • The take profit identifies when a trade no longer offers positive expected excess returns.

Even though it has the word “arbitrage” in its name, stat arb can be highly risky and lead to enormous and systemic losses, such as in the epic collapse of the hedge fund Long Term Capital Management . The instruments are highly correlated, so it is not likely that the ratio of prices will differ from mean by two standard deviations. The risk-adjusted performance of the strategy is consistently outstanding throughout the out-of-sample period from 2010. After a slowdown in 2014, strategy performance in the first quarter of 2015 has again accelerated to the level achieved in earlier years (i.e. with a Sharpe ratio above 4). The analysis in Appendix II suggests that the residual process is stable and Gaussian. In other words, the two-state Markov model is able to account for the non-Normality of the returns process and extract the salient autoregressive and moving average features in a way that makes economic sense.

A Theory Of Herding And Instability In Bond Markets

The new portfolio underperforms the index during 2014, but with lower volatility and average drawdown. The idea is to examine the characteristics of the returns process and assess its predictability. In the previous post I outlined some of the available techniques used for modeling market states. The following is an illustration of how these techniques can be applied in practice. But it is also clear that there are many other significant correlations between non-conjugate pairs.

statistical arbitrage

Remember, most stock market crashes arise from issues with liquidity and leverage—the very arena in which statistical arbitrageurs operate. Stat arb algorithms have also been blamed in part for the “flash crashes” that the market has started to experience over the past decade. A flash crash is an event in electronic securities markets wherein a rapid sell-off of securities leads to a negative feedback loop that can cause dramatic price drops over a matter of minutes. Unlike traditional statistical arbitrage, risk arbitrage involves taking on some risks. The largest risk is that the merger will fall through and the target’s stock will drop to its pre-merger levels.

Building A Topic Modeling Pipeline With Spacy And Gensim

On the other hand, the arbitrageur buys a treasury bond, with the same maturity as the swap, with the money borrowed through a repurchase agreement known as repo. Entering this part of the trade the arbitrageur earns the treasury rate TR and pays the repo rate r t . The overall cash flow of the trade is ( L t − r t ) − ( S R − T R ) where S R − T R is the fixed interest rate component and L t − r t is the floating rate part which needs to be rolled periodically .

This reset will prevent the prices of the two paired assets from drifting too far apart. Specifically, at the beginning of each year, investors will recalculate the theoretical value of the replicating asset and compare it with the market value of Berkshire A stock and take the appropriate positions for the next cycle . In this paper, a high frequency and dynamic pairs trading system is proposed, based on a market-neutral statistical arbitrage strategy using a two-stage correlation and cointegration approach. The proposed pairs trading system was tested for out-of-sample testing periods with high frequency stock data from 2012 and 2013. Our trading strategy yields cumulative returns up to 56.58% for portfolios of stock pairs, well exceeding the S&P 500 index performance by 34.35% over a 12-month trading period.

Statistical Arbitrage Risk Premium

Similarly, if the implied volatility is higher, the trader can sell the option and hedge with the underlying security to make a delta-neutral portfolio. The key to success in risk arbitrage is determining the likelihood and timeliness of the merger and comparing that with the difference in price between the target stock and the buyout offer. Some risk arbitrageurs have begun to speculate on takeover targets as well, which can lead to substantially greater profits with equally greater risk.

Machine Learning For Portfolio Diversification

This restriction created more meaningful pairs, because all sample companies are now matched within the same sectors. Their results show that the pairs trading strategy remains profitable, albeit at much more modest levels. Cohen and Frazzini find substantial customer-supplier links in the U.S. stock market that allow for return predictability in the context of this strategy. Bertram derives the entry/exit time and analytical formula for the trading thresholds for synthetic assets formed by pairs, whose price assumptions follow the Ornstein Uhlenbeck process.

In a real world application, that discovery could only be made in real time, when the unknown, future ETFs prices are formed. I have traded pairs successfully using all of the techniques described in the first part of the post (i.e. Ratio, Regression, Kalman and Copula methods). Equally, I have seen a great many failed pairs strategies produced by using every available technique. One often finds that a pair that perform poorly using the ratio method produces decent returns when a regression or Kalman Filter model is applied. From experience, there is no pattern that allows you to discern which technique, if any, is gong to work. In this example I have selected a universe of the Dow 30 stocks, together with a sample of commodities and bonds and compiled a database of daily returns over the period from Jan 2012 to Dec 2013.

Bondarenko’s SA is a trading strategy which can have negative payoffs, as long as the average payoff is non-negative for given augmented information set. Key in the definition is the introduction of the augmented information set, which, in addition to the market information triangular arbitrage at time t, also includes the knowledge of the final price. Hogan et al. provide an alternative definition of SA which focuses on long horizon trading opportunities. Hogan’s SA is a long horizon trading opportunity that, at the limit, generates a risk-less profit.

Pairs Trading

In many cases, just the existence of the model will have an impact on the market. If enough people execute the same strategy, it can impact the securities that are being traded. For example, if enough people know that two stocks are correlated, they will alternately purchase orders at the same time, when the model recommends it. This can increase or decrease the prices of the securities and affect the profitability.

Author: Ian Sherr