October 2018 - QUANTITATIVE RESEARCH AND TRADING

October 30, 2018

Combining Momentum and Mean Reversion Strategies

The Fama-French World

For many years now the “gold standard” in factor models has been the 1996 Fama-French 3-factor model:
Here r is the portfolio’s expected rate of return, R_f is the risk-free return rate, and K_m is the return of the market portfolio. The “three factor” β is analogous to the classical β but not equal to it, since there are now two additional factors to do some of the work. SMB stands for “Small [market capitalization] Minus Big” and HML for “High [book-to-market ratio] Minus Low”; they measure the historic excess returns of small caps over big caps and of value stocks over growth stocks. These factors are calculated with combinations of portfolios composed by ranked stocks (BtM ranking, Cap ranking) and available historical market data. The Fama–French three-factor model explains over 90% of the diversified portfolios in-sample returns, compared with the average 70% given by the standard CAPM model.

The 3-factor model can also capture the reversal of long-term returns documented by DeBondt and Thaler (1985), who noted that extreme price movements over long formation periods were followed by movements in the opposite direction. (Alpha Architect has several interesting posts on the subject, including this one).

Fama and French say the 3-factor model can account for this. Long-term losers tend to have positive HML slopes and higher future average returns. Conversely, long-term winners tend to be strong stocks that have negative slopes on HML and low future returns. Fama and French argue that DeBondt and Thaler are just loading on the HML factor.

Enter Momentum

While many anomalies disappear under tests, shorter term momentum effects (formation periods ~1 year) appear robust. Carhart (1997) constructs his 4-factor model by using FF 3-factor model plus an additional momentum factor. He shows that his 4-factor model with MOM substantially improves the average pricing errors of the CAPM and the 3-factor model. After his work, the standard factors of asset pricing model are now commonly recognized as Value, Size and Momentum.

Combining Momentum and Mean Reversion

In a recent post, Alpha Architect looks as some possibilities for combining momentum and mean reversion strategies. They examine all firms above the NYSE 40th percentile for market-cap (currently around $1.8 billion) to avoid weird empirical effects associated with micro/small cap stocks. The portfolios are formed at a monthly frequency with the following 2 variables:

Momentum = Total return over the past twelve months (ignoring the last month)
Value = EBIT/(Total Enterprise Value)

They form the simple Value and Momentum portfolios as follows:

EBIT VW = Highest decile of firms ranked on Value (EBIT/TEV). Portfolio is value-weighted.
MOM VW = Highest decile of firms ranked on Momentum. Portfolio is value-weighted.
Universe VW = Value-weight returns to the universe of firms.
SP500 = S&P 500 Total return

The results show that the top decile of Value and Momentum outperformed the index over the past 50 years. The Momentum strategy has stronger returns than value, on average, but much higher volatility and drawdowns. On a risk-adjusted basis they perform similarly. The researchers then form the following four portfolios:

EBIT VW = Highest decile of firms ranked on Value (EBIT/TEV). Portfolio is value-weighted.
MOM VW = Highest decile of firms ranked on Momentum. Portfolio is value-weighted.
COMBO VW = Rank firms independently on both Value and Momentum. Add the two rankings together. Select the highest decile of firms ranked on the combined rankings. Portfolio is value-weighted.
50% EBIT/ 50% MOM VW = Each month, invest 50% in the EBIT VW portfolio, and 50% in the MOM VW portfolio. Portfolio is value-weighted.

With the following results:

The main takeaways are:

The combined ranked portfolio outperforms the index over the same time period.
However, the combination portfolio performs worse than a 50% allocation to Value and a 50% allocation to Momentum.

A More Sophisticated Model

Yangru Wu of Rutgers has been doing interesting work in this area over the last 15 years, or more. His 2005 paper (with Ronald Balvers), Momentum and mean reversion across national equity markets, considers joint momentum and mean-reversion effects and allows for complex interactions between them. Their model is of the form where the excess return for country i (relative to the global equity portfolio) is represented by a combination of mean-reversion and autoregressive (momentum) terms. Balvers and Wu find that combination momentum-contrarian strategies, used to select from among 18 developed equity markets at a monthly frequency, outperform both pure momentum and pure mean-reversion strategies. The results continue to hold after corrections for factor sensitivities and transaction costs. The researchers confirm that momentum and mean reversion occur in the same assets. So in establishing the strength and duration of the momentum and mean reversion effects it becomes important to control for each factor’s effect on the other. The momentum and mean reversion effects exhibit a strong negative correlation of 35%. Accordingly, controlling for momentum accelerates the mean reversion process, and controlling for mean reversion may extend the momentum effect.

Momentum, Mean Reversion and Volatility

The presence of strong momentum and mean reversion in volatility processes provides a rationale for the kind of volatility strategy that we trade at Systematic Strategies. One sophisticated model is the Range Based EGARCH model of Alizadeh, Brandt, and Diebold (2002) . The model posits a two-factor volatility process in which a short term, transient volatility process mean-reverts to a stochastic long term mean process, which may exhibit momentum, or long memory effects (details here).

In our volatility strategy we model mean reversion and momentum effects derived from the level of short and long term volatility-of-volatility, as well as the forward volatility curve. These are applied to volatility ETFs, including levered ETF products, where convexity effects are also important. Mean reversion is a well understood phenomenon in volatility, as, too, is the yield roll in volatility futures (which also impacts ETF products like VXX and XIV).

Momentum effects are perhaps less well researched in this context, but our research shows them to be extremely important. By way of illustration, in the chart below I have isolated the (gross) returns generated by one of the momentum factors in our model.

October 23, 2018October 23, 2018

Developing Long/Short ETF Strategies

Recently I have been working on the problem of how to construct large portfolios of cointegrated securities. My focus has been on ETFs rather that stocks, although in principle the methodology applies equally well to either, of course.

My preference for ETFs is due primarily to the fact that it is easier to achieve a wide diversification in the portfolio with a more limited number of securities: trading just a handful of ETFs one can easily gain exposure, not only to the US equity market, but also international equity markets, currencies, real estate, metals and commodities. Survivorship bias, shorting restrictions and security-specific risk are also less of an issue with ETFs than with stocks (although these problems are not too difficult to handle).

On the downside, with few exceptions ETFs tend to have much shorter histories than equities or commodities. One also has to pay close attention to the issue of liquidity. That said, I managed to assemble a universe of 85 ETF products with histories from 2006 that have sufficient liquidity collectively to easily absorb an investment of several hundreds of millions of dollars, at minimum.

The Cardinality Problem

The basic methodology for constructing a long/short portfolio using cointegration is covered in an earlier post. But problems arise when trying to extend the universe of underlying securities. There are two challenges that need to be overcome.

The first issue is that, other than the simple regression approach, more advanced techniques such as the Johansen test are unable to handle data sets comprising more than about a dozen securities. The second issue is that the number of possible combinations of cointegrated securities quickly becomes unmanageable as the size of the universe grows. In this case, even taking a subset of just six securities from the ETF universe gives rise to a total of over 437 million possible combinations (85! / (79! * 6!). An exhaustive test of all the possible combinations of a larger portfolio of, say, 20 ETFs, would entail examining around 1.4E+19 possibilities.

Given the scale of the computational problem, how to proceed? One approach to addressing the cardinality issue is sparse canonical correlation analysis, as described in Identifying Small Mean Reverting Portfolios, d’Aspremont (2008). The essence of the idea is something like this. Suppose you find that, in a smaller, computable universe consisting of just two securities, a portfolio comprising, say, SPY and QQQ was found to be cointegrated. Then, when extending consideration to portfolios of three securities, instead of examining every possible combination, you might instead restrict your search to only those portfolios which contain SPY and QQQ. Having fixed the first two selections, you are left with only 83 possible combinations of three securities to consider. This process is repeated as you move from portfolios comprising 3 securities to 4, 5, 6, … etc.

Other approaches to the cardinality problem are possible. In their 2014 paper Sparse, mean reverting portfolio selection using simulated annealing, the Hungarian researchers Norbert Fogarasi and Janos Levendovszky consider a new optimization approach based on simulated annealing. I have developed my own, hybrid approach to portfolio construction that makes use of similar analytical methodologies. Does it work?

A Cointegrated Long/Short ETF Basket

Below are summarized the out-of-sample results for a portfolio comprising 21 cointegrated ETFs over the period from 2010 to 2015. The basket has broad exposure (long and short) to US and international equities, real estate, currencies and interest rates, as well as exposure in banking, oil and gas and other specific sectors.

The portfolio was constructed using daily data from 2006 – 2009, and cointegration vectors were re-computed annually using data up to the end of the prior year. I followed my usual practice of using daily data comprising “closing” prices around 12pm, i.e. in the middle of the trading session, in preference to prices at the 4pm market close. Although liquidity at that time is often lower than at the close, volatility also tends to be muted and one has a period of perhaps as much at two hours to try to achieve the arrival price. I find this to be a more reliable assumption that the usual alternative.

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).

Another useful test procedure is to compare the strategy performance with that of a portfolio constructed using standard mean-variance optimization (using the same ETF universe, of course). The test indicates that a portfolio constructed using the traditional Markowitz approach produces a similar annual return, but with 2.5x the annual volatility (i.e. a Sharpe ratio of only 1.6). What is impressive about this result is that the comparison one is making is between the out-of-sample performance of the strategy vs. the in-sample performance of a portfolio constructed using all of the available data.

Having demonstrated the validity of the methodology, at least to my own satisfaction, the next step is to deploy the strategy and test it in a live environment. This is now under way, using execution algos that are designed to minimize the implementation shortfall (i.e to minimize any difference between the theoretical and live performance of the strategy). So far the implementation appears to be working very well.

Once a track record has been built and audited, the really hard work begins: raising investment capital!

October 22, 2018October 22, 2018

MATH-TWS: Connecting Wolfram Mathematica to IB TWS

Mathematica can now connect to Interactive Brokers Trader Workstation

At long last, it’s here!

MATH-TWS is a new Mathematica package that connects Wolfram Mathematica to the Interactive Brokers TWS platform via the C++ API. It enables the user to retrieve information from TWS on accounts, portfolios and positions, as well as historical and real-time market data. MATH-TWS also enables the user to place and amend orders and obtain execution confirmations from Mathematica.

In the following sections we will illustrate the functionality of the MATH-TWS package using the full functional form and show the abbreviated equivalent form in comments.

Conclusions: Connecting Mathematica to IB TWS

I have wanted a way to connect Wolfram Mathematica to Interactive Brokers’ Trader Workstation for the longest time. Now that it is finally available with MATH-TWS I am excited by the possibilities for Mathematica users.

The first release of MATH-TWS will be available within a couple of weeks. Anyone interested in licensing a copy should email algorithmicexecution@gmail.com with MATH-TWS in the subject line.

October 16, 2018

Algorithmic Trading

MOVING FROM RESEARCH TO TRADING

I have written recently about the comparative advantages of different programming languages in the context of research and trading (see here). My sense of it is that there is no single “ideal” programming language – the best strategy is to pick an appropriate tool for the job and there are usually several reasonable choices one could make.

If you are engaged in econometrics research, you might choose a package like RATS, Eviews, Gauss, or Prof. James Davidson’s excellent and inexpensive TSM, which I have used for many years and can recommend highly. For a latency-sensitive high frequency trading application, you will probably want to use something like C++, or possibly a 3rd party algo system like Apama or Tethys. But for algorithmic trading systems of intermediate frequency the choice appears almost unlimited.

The problem with retail trading tools like TradeStation, Multicharts, or Amibroker, is that they are designed primarily for single-asset strategies. That may be ok for futures trading,where more often than not the focus is on a single underlying, but in equities the opposite is true. Using one of these products to develop and implement a pairs trading strategy is a stretch. As for portfolio analytics – forget it.

This is where more general, high level languages like R, Matlab or Mathematica come in: their greater power and flexibility is handling large, multivariate data sets makes it much more straightforward to develop portfolio strategies. And they can often bridge the gap between R&D and implementation quite easily: code that was used in the research stage can often be quickly re-tooled to work in a production version of the system. As for production systems, there is now a significant cottage industry of traders who use Matlab in algo trading. R has a similar following (see here).

In addition to parallelizing the code (for use with the Parallel Computing Toolbox) to speed up the research phase, you might also want to implement a hybrid system by re-coding the slower routines in C++, to create a mex file (for details see here). Matlab’s Profiler is a useful tool for identifying code bottlenecks. In a recent piece of research in which I was evaluating over 30,000,000 cointegrated portfolios, I discovered to my surprise that the main code bottleneck was the multiple calls to Matlab’s std function, a problem easily fixed with a few lines of C++ code. The resulting hybrid program executed at more than twice the speed – important when your run time might be several hours, or even days.

HOOKING UP THE EXECUTION PLATFORM

The main challenge for developers using generic tools like Mathematica, Matlab or R is the implementation stage of the project. Providing connectivity to brokerage/execution platforms never seemed high on the list of priorities for Wolfram or Mathworks and things are similarly hit or miss with R.

Belatedly, Mathematica now offers a link to Bloomberg via its Finance Platform. Matlab, meanwhile, offers a Trading Toolbox, which supposedly offers connectivity , not only to Bloomberg, but also Interactive Brokers and Trading Technologies, amongst other platforms. Unfortunately, the toolbox interface to IB appears to rely on outdated 1990s ActiveX technology, which is flakey at best. In tests, I was unable to make progress past the ‘not connected’ error message.

At that point I turned to Yair Altman’s IB-Matlab product. Happily, this uses IB’s Java api, which is a great deal more robust than the ActiveX platform. It’s been some time since I last used IB-Matlab and was pleased to see that Yair has been very busy over the intervening period, building the capabilities of the system and providing very comprehensive documentation for it. With Yair’s help, it took me no time at all to get up and running and within a day or two the system was executing orders flawlessly in IB’s TWS. The relatively few snags I ran into were almost all due to IB’s extremely terse error messaging, which often gives almost no clue as to what the issue might be. Fortunately, Yair is very generous with his time in providing support to his users and his responses to me questions were fast and detailed.

EXECUTION ALGOS

With intermediate systems trading at frequencies of, say, 5-minutes to daily, one has a choice to make as regards execution. Given that the strategy is not very latency sensitive, it is certainly conceivable to develop one’s own execution algos in Matlab. However, platforms like TWS are equipped with native algos, not only from IB, but also other providers like Credit Suisse and Jeffries.

Actually, I have found several of IB’s own algos such as Scaletrader and Accumulate/Distribute to be very effective. Certainly IB seems very proud of them – IB CEO Thomas Peterffy has patented at least one of them. Accumulate/Distribute, for instance, is quite sophisticated, allowing the user to randomize and slice the size and interval between individual orders, use passive or aggressive order types, and pause execution on a news alert, or when the price falls below a moving average, or outside a specified range.

There is much to be said for using algos native to the execution platform rather than reinventing the wheel, providing the cost is reasonable. So, while it is perfectly feasible to build execution algos in Matlab, it typically isn’t necessary – in most cases standard algos will suffice.

There are exceptions, of course. IB doesn’t offer the kind of basket-trading capabilities that are available in advanced algo platforms like Tethys or RediPlus. In those systems, for example, you can set the level of long/short imbalance in the portfolio that you are willing to tolerate and the algo will speed up or slow down execution of trades in individual components of the basket to maintain the dollar imbalance within that tolerance. You can also manage the sector risk dynamically during execution.

Those kind of advanced capabilities don’t come cheap and you wont find them at IB, or any other retail platform. If you need that kind of functionality, for example, because you are trading a long/short equity portfolio within a universe of 200-300 names, your best option is probably to switch to a different execution platform. Otherwise you will need to code a custom algo in your language of choice.

For many quantitative strategies, (at least the low frequency ones) IB’s standard algos are often good enough. The Accumulate/Distribute algo, for instance, will show a visual representation of the progress of the execution of individuals legs of a pairs trade, and it is easy enough to identify a potential imbalance and adjust the algo parameters in real time. If you are only trading pairs, or small portfolios of cointegrated securities, it probably isn’t worthwhile to develop the sophisticated logic that would be required to handle the adjustment of the execution of individual legs of a trade in a fully automated way. A large portfolio would be a different matter, however.

MATLAB EXAMPLE

I thought it might be instructive to take a look at how you might implement the execution of a strategy in Matlab, using IB algos. In the Matlab code fragment below, the (2 x nTickers) array tradeActions contains, in the first row, the action we wish to take (1 = BUY, -1 = SELL, -2 = SELL SHORT) and in the second row the (absolute value of) the number of shares we wish to trade for tickers i =1:nTickers. We break each order up into hundred lots and odd lots, routing the former via IB’s Accumulate/Distribute algo and the latter as passive REL orders (note that A/D will typically randomize the timing of each sub-order, while REL orders are posted directly into the market). The Matlab function AccumulateDistribute implements the most important features of IB’s A/D algo, including random size and time slicing of the order. Orders are submitted as passive REL orders with zero offset (so they will sit on the current bid or ask) – obviously you would typically want to consider allowing some non-zero offset for less liquid securities. It is not hard to envisage how one might further enhance the algo to monitor the progress of the execution and speed up or slow down certain orders accordingly.

A couple of IB api “gotchas” to be aware of:

(i) IB requires unique and monotonically increasing orderIds for each order. One way to do this, suggested by Yair, is to use orderId = round((now-735000)*3e5); This fails when you are submitting a number of orders sequentially at high speed (say in a for loop), where the time increments are sub-second, so you need to pass the orderID back and force a minimal increment, as I have in the code below.

(ii) It is very important to specify the primary exchange of each security: securities with identical tickers can be found trading on different exchanges. Failing to specify the primary exchange in such a case will result in IB rejecting the order with a typically cryptic api message.

Continue reading “Algorithmic Trading”

October 10, 2018October 11, 2018

Cointegration Breakdown

The Low Power of Cointegration Tests

One of the perennial difficulties in developing statistical arbitrage strategies is the lack of reliable methods of estimating a stationary portfolio comprising two or more securities. In a prior post (below) I discussed at some length one of the primary reasons for this, i.e. the lower power of cointegration tests. In this post I want to explore the issue in more depth, looking at the standard Johansen test Procedure to estimate cointegrating vectors.

Johansen Test for Cointegration

Start with some weekly data for an ETF triplet analyzed in Ernie Chan’s book:

After downloading the weekly close prices for the three ETFs we divide the data into 14 years of in-sample data and 1 year out of sample:

We next apply the Johansen test, using code kindly provided by Amanda Gerrish:

We find evidence of up to three cointegrating vectors at the 95% confidence level:

Let’s take a look at the vector coefficients (laid out in rows, in Amanda’s function):

In-Sample vs. Out-of-Sample testing

We now calculate the in-sample and out-of-sample portfolio values using the first cointegrating vector:

The portfolio does indeed appear to be stationary, in-sample, and this is confirmed by the unit root test, which rejects the null hypothesis of a unit root:

Unfortunately (and this is typically the case) the same is not true for the out of sample period:

More Data Doesn’t Help

The problem with the nonstationarity of the out-of-sample estimated portfolio values is not mitigated by adding more in-sample data points and re-estimating the cointegrating vector(s):

We continue to add more in-sample data points, reducing the size of the out-of-sample dataset correspondingly. But none of the tests for any of the out-of-sample datasets is able to reject the null hypothesis of a unit root in the portfolio price process:

The Challenge of Cointegration Testing in Real Time

In our toy problem we know the out-of-sample prices of the constituent ETFs, and can therefore test the stationarity of the portfolio process out of sample. In a real world application, that discovery could only be made in real time, when the unknown, future ETFs prices are formed. In that scenario, all the researcher has to go on are the results of in-sample cointegration analysis, which demonstrate that the first cointegrating vector consistently yields a portfolio price process that is very likely stationary in sample (with high probability).

The researcher might understandably be persuaded, wrongly, that the same is likely to hold true in future. Only when the assumed cointegration relationship falls apart in real time will the researcher then discover that it’s not true, incurring significant losses in the process, assuming the research has been translated into some kind of trading strategy.

A great many analysts have been down exactly this path, learning this important lesson the hard way. Nor do additional “safety checks” such as, for example, also requiring high levels of correlation between the constituent processes add much value. They might offer the researcher comfort that a “belt and braces” approach is more likely to succeed, but in my experience it is not the case: the problem of non-stationarity in the out of sample price process persists.

Conclusion: Why Cointegration Breaks Down

We have seen how a portfolio of ETFs consistently estimated to be cointegrated in-sample, turns out to be non-stationary when tested out-of-sample. This goes to the issue of the low power of cointegration test, and their inability to estimate cointegrating vectors with sufficient accuracy. Analysts relying on standard tests such as the Johansen procedure to design their statistical arbitrage strategies are likely to be disappointed by the regularity with which their strategies break down in live trading.

October 9, 2018

Successful Statistical Arbitrage

I tend not to get involved in Q&A with readers of my blog, or with investors. I am at a point in my life where I spend my time mostly doing what I want to do, rather than what other people would like me to do. And since I enjoy doing research and trading, I try to maximize the amount of time I spend on those activities.

As a business strategy, I wouldn’t necessarily recommend this approach. It’s just something I evolved while learning to play chess: since I had no-one to teach me, I had to learn everything for myself and this involved studying for many, many hours alone.

By contrast, several of the best money managers are also excellent communicators – take Roy Niederhoffer, or Ernie Chan, for example. Having regular, informed communication with your investors is, as smarter managers have realized, a means of building trust and investor loyalty – important factors that come into play during periods when your strategy is underperforming. Not only that, but since communication is two-way, an analyst/manager can learn much from his exchanges with his clients. Knowing how others perceive you – and your competitors – for example, is very useful information. So, too, is information about your competitors’ research ideas, investment strategies and fund performance, which can often be gleaned from discussions with investors. There are plenty of reasons to prefer a policy of regular, open communication.

As a case in point, I was surprised to learn from comments on another research blog that readers drew the conclusion from my previous posts that pursuing the cointegration or Kalman Filter approach to statistical arbitrage was a waste of time. Apparently, my remark to the effect that researchers often failed to pay attention to the net PnL per share in evaluating stat. arb. trading strategies was taken by some to mean that any apparent profitability would always be subsumed within the bid-offer spread. That was not my intention. What I intended to convey was that in some instances, this would be the case – some, but not all.

To illustrate the point, below are the out-of-sample results from a research study applying the Kalman Filter approach for four equity pairs using 5-minute data. For competitive reasons I am unable to identify the specific stocks in each pair, which result from an exhaustive analysis of over 30,000 pairs, but I can say that they are liquid large-cap equities traded in large volume on the US exchanges. The performance numbers are net of transaction costs and are based on the assumption of a 5-minute delay in execution: meaning, a trading signal received at time t is assumed to be executed at time t+5 minutes. This allows sufficient time to leg into each trade passively, in most cases avoiding the bid-offer spread. The net PnL per share is above 1.5c per share for each pair.

While the performance of none of the pairs is spectacular, a combined portfolio has quite attractive characteristics, which include 81% winning months since Jan 2012, a CAGR of over 27% and Information Ratio of 2.29, measured on monthly returns (2.74 based on daily returns).

Finally, I am currently implementing trading of a number of stock portfolios based on static cointegration relationships that have out-of-sample information ratios of between 3 and 4, using daily data.

October 2, 2018

A Comparison of Programming Languages

Towards the end of last year I wrote a post (see here) about the advent of modern programming languages, including the JIT compiled Julia and visual programming language ADL from Trading Technologies. My conclusion (based on a not very scientific sample) was that we appear to be at the tipping point, where the speed of newer, high level languages languages is approaching that of the fastest compiled languages like C/C++.

Now comes a formal academic study of the topic in A Comparison of Programming Languages in Economics, Aruoba and Fernandez-Villaverde, 2014. Using the neoclassical growth model, the authors conduct a benchmark test in C++11, Fortran 2008, Java, Julia, Python, Matlab, Mathematica, and R, implementing the same algorithm, value function
iteration with grid search, in each of the languages. They report the execution times of the codes in a Mac and in a Windows computer and briefly comment on the strengths and weaknesses of each language.

The conclusions from the study mirror my own thoughts on the subject very closely. The authors find that:

C++ and Fortran are still considerably faster than any other alternative, although one needs to be careful with the choice of compiler.
C++ compilers have advanced enough that, contrary to the situation in the 1990s and some folk wisdom, C++ code runs slightly faster (5-7 percent) than Fortran code.
Julia delivers outstanding performance. Execution speed is only between 2.64 and 2.70 times slower than the execution speed of the best C++ compiler.
Baseline Python was slow. Using the Pypy implementation, it runs around 44 times slower than in C++. Using the default CPython interpreter, the code runs between 155 and 269 times slower than in C++.
Matlab is between 9 to 11 times slower than the best C++ executable.
R runs between 475 to 491 times slower than C++. If the code is compiled, the code is between 243 to 282 times slower.
Hybrid programming and special approaches can deliver considerable speed ups. For example, when combined with Mex files, Matlab is only 1.24 to 1.64 times slower than C++ and when combined with Rcpp, R is between 3.66 and 5.41 times slower. Similar numbers hold for Numba (a just-in-time compiler for Python that uses decorators) and Cython (a static compiler for writing C extensions for Python) in the Python ecosystem.
Mathematica is only about three times slower than C++, but only after a considerable rewriting of the code to take advantage of the peculiarities of the language. The baseline version of the algorithm in Mathematica is considerably slower.

C++ still represents the benchmark for speed, but not by much. It is barely faster than the old stalwart, Fortran, and only 1.5 – 3 times faster than up-and-coming rivals amongst the higher level languages (especially when you allow for hybrid programming to speed up the slowest algorithms).

So, as regards developing financial models and trading systems, my questions are (as before):

Why would anyone prefer Python, given that there is a much faster, free alternative in Julia, which is just as easy a language to program in?
What justification is there for preferring R to Matlab, other than cost?
Why does anyone bother with Java? If speed is the critical issue, there are faster alternatives. If you like the relative simplicity of the syntax, Julia is cleaner, simpler and just as fast in execution.

When you reach a point where a high level language like Matlab is only around 1.5x – 2x slower than C++, you really have to question whether the latter is an appropriate choice. Yes, of course, in mission-critical applications where you need access to the hardware layer for speed purposes, C++ is the way to go. But for so many applications, that just isn’t the case.

What matters, far, far more, are the months of costly and laborious programming effort that is often required to reproduce basic functionality that is already embedded in higher level languages like Matlab or Mathematica. Not only that, but the end result of a C++ /Java development effort is likely to be notoriously inflexible by comparison. That’s a huge drawback. Rarely, if ever, does a piece of research translate flawlessly into production – it requires one to iterate towards a final solution, often making significant changes to the design of the system in the light of practical experience.

If I had to guess, based on my experience, I would say that 80% or more of development tasks in quantitative research and trading would produce a superior result if preference was given to using a higher level language for the initial development. When the system is sufficiently stable to put into production, you simply create a hybrid application by recoding any mission-critical components for which speed is an issue in C++.

Finally, where does that leave my beloved Mathematica? To be fair, while you don’t have the joys of strong typing to contend with, Mathematica’s syntax is just as demanding and uncompromising as C++ – a missed comma or incorrectly placed bracket is just as critical. But, the point is, while in C++ the syntactical rigor is just annoying, in Mathematica it’s worth putting up with because the productivity is so much greater. A competent programmer can produce, in a single line of Mathematica code, a program that would require hundreds, if not thousands of lines of C++ code to accomplish. Sure, he might get the syntax wrong at first: but it’s only a single line of code and the interactive gui interface makes debugging very simple.

That said, while Mathematica can be very tedious to use for procedural programming, it excels in three areas:

1. Symbolic programming. Anything involving mathematical symbols and equations – Mathematica is #1

2. User interface. In Mathematica, it is trivial to build a sophisticated, dynamic gui in no time at all, again, often in 1-2 lines of code

3. Functional programming. Anything that can be thought of as a function, Mathematica handles extremely well. We are not talking about finding a square root here: I mean extremely complex functions that, again, might take hundreds of lines of code in another language.

It is also worth pointing out that Mathematica comes supplied with functionality that Matlab provides only through numerous, costly add-on packages.

CONCLUSION
Before I allow a development team to start mindlessly coding up a system in Java or C++, I want to hear their reasons why they aren’t going to do it 10x faster in another, higher level language. “We always use C++/Java for production” is not a reason. Specifically, which parts of the system require the additional 1.5x speed-up, and why can’t they be coded as dlls (Matlab mex functions)?

Finally, on a cost-benefit basis, ask yourself how much you might benefit if the months and tens (or hundreds) of thousands of dollars wasted on developing in C++ were instead spent on researching and developing new trading ideas.

October 2, 2018October 2, 2018

Quant Strategies in 2018

Quant Strategies – Performance Summary Sept. 2018

The end of Q3 seems like an appropriate time for an across-the-piste review of how systematic strategies are performing in 2018. I’m using the dozen or more strategies running on the Systematic Algotrading Platform as the basis for the performance review, although results will obviously vary according to the specifics of the strategy. All of the strategies are traded live and performance results are net of subscription fees, as well as slippage and brokerage commissions.

Volatility Strategies

Those waiting for the hammer to fall on option premium collecting strategies will have been disappointed with the way things have turned out so far in 2018. Yes, February saw a long-awaited and rather spectacular explosion in volatility which completely destroyed several major volatility funds, including the VelocityShares Daily Inverse VIX Short-Term ETN (XIV) as well as Chicago-based hedged fund LJM Partners (“our goal is to preserve as much capital as possible”), that got caught on the wrong side of the popular VIX carry trade. But the lack of follow-through has given many volatility strategies time to recover. Indeed, some are positively thriving now that elevated levels in the VIX have finally lifted option premiums from the bargain basement levels they were languishing at prior to February’s carnage. The Option Trader strategy is a stand-out in this regard: not only did the strategy produce exceptional returns during the February melt-down (+27.1%), the strategy has continued to outperform as the year has progressed and YTD returns now total a little over 69%. Nor is the strategy itself exceptionally volatility: the Sharpe ratio has remained consistently above 2 over several years.

Hedged Volatility Trading

Investors’ chief concern with strategies that rely on collecting option premiums is that eventually they may blow up. For those looking for a more nuanced approach to managing tail risk the Hedged Volatility strategy may be the way to go. Like many strategies in the volatility space the strategy looks to generate alpha by trading VIX ETF products; but unlike the great majority of competitor offerings, this strategy also uses ETF options to hedge tail risk exposure. While hedging costs certainly acts as a performance drag, the results over the last few years have been compelling: a CAGR of 52% with a Sharpe Ratio close to 2.

F/X Strategies

One of the common concerns for investors is how to diversify their investment portfolios, especially since the great majority of assets (and strategies) tend to exhibit significant positive correlation to equity indices these days. One of the characteristics we most appreciate about F/X strategies in general and the F/X Momentum strategy in particular is that its correlation to the equity markets over the last several years has been negligible. Other attractive features of the strategy include the exceptionally high win rate – over 90% – and the profit factor of 5.4, which makes life very comfortable for investors. After a moderate performance in 2017, the strategy has rebounded this year and is up 56% YTD, with a CAGR of 64.5% and Sharpe Ratio of 1.89.

Equity Long/Short

Thanks to the Fed’s accommodative stance, equity markets have been generally benign over the last decade to the benefit of most equity long-only and long-short strategies, including our equity long/short Turtle Trader strategy , which is up 31% YTD. This follows a spectacular 2017 (+66%) , and is in line with the 5-year CAGR of 39%. Notably, the correlation with the benchmark S&P500 Index is relatively low (0.16), while the Sharpe Ratio is a respectable 1.47.

Equity ETFs – Market Timing/Swing Trading

One alternative to the traditional equity long/short products is the Tech Momentum strategy. This is a swing trading strategy that exploits short term momentum signals to trade the ProShares UltraPro QQQ (TQQQ) and ProShares UltraPro Short QQQ (SQQQ) leveraged ETFs. The strategy is enjoying a banner year, up 57% YTD, with a four-year CAGR of 47.7% and Sharpe Ratio of 1.77. A standout feature of this equity strategy is its almost zero correlation with the S&P 500 Index. It is worth noting that this strategy also performed very well during the market decline in Feb, recording a gain of over 11% for the month.

Futures Strategies

It’s a little early to assess the performance of the various futures strategies in the Systematic Strategies portfolio, which were launched on the platform only a few months ago (despite being traded live for far longer). For what it is worth, both of the S&P 500 E-Mini strategies, the Daytrader and the Swing Trader, are now firmly in positive territory for 2018. Obviously we are keeping a watchful eye to see if the performance going forward remains in line with past results, but our experience of trading these strategies gives us cause for optimism.

Conclusion: Quant Strategies in 2018

There appear to be ample opportunities for investors in the quant sector across a wide range of asset classes. For investors with equity market exposure, we particularly like strategies with low market correlation that offer significant diversification benefits, such as the F/X Momentum and F/X Momentum strategies. For those investors seeking the highest risk adjusted return, option selling strategies like the Option Trader strategy are the best choice, while for more cautious investors concerned about tail risk the Hedged Volatility strategy offers the security of downside protection. Finally, there are several new strategies in equities and futures coming down the pike, several of which are already showing considerable promise. We will review the performance of these newer strategies at the end of the year.

Go here for more information about the Systematic Algotrading Platform.

October 1, 2018

ETF Pairs Trading with the Kalman Filter

I was asked by a reader if I could illustrate the application of the Kalman Filter technique described in my previous post with an example. Let’s take the ETF pair AGG IEF, using daily data from Jan 2006 to Feb 2015 to estimate the model. As you can see from the chart in Fig. 1, the pair have been highly correlated over the last several years.

Fig 1. AGG and IEF Daily Prices 2006-2015

We now estimate the beta-relationship between the ETF pair with the Kalman Filter, using the Matlab code given below, and plot the estimated vs actual prices of the first ETF, AGG in Fig 2. There are one or two outliers that you might want to take a look at, but mostly the fit looks very good.

Fig 2 – Actual vs Fitted Prices of AGG

Now lets take a look at Kalman Filter estimates of beta. As you can see in Fig 3, it wanders around a lot! Very difficult to handle using some kind of static beta estimate.

Fig 3 – Kalman Filter Beta Estimates

Finally, we compute the raw and standardized alphas, being the differences between the observed and fitted prices , i.e. Alpha(t) = AGG(t) – b(t)* IEF(t) and kfAlpha(t) = (Alpha(t) – mean(Alpha(t)) / std(Alpha(t) I have plotted the kfAlpha estimates over the last year in Fig 4.

Fig 4 – Standardized Alpha Estimates

The last step is to decide how to trade this relationship. You might, for example, trade the portfolio in proportion to the standardized deviation (i.e. the size of kfAlpha(t)). Alternatively, you might set a threshold level, say +/- 1 Sd, and trade the portfolio when kfAlpha(t) exceeds this the threshold. In the Matlab code below I use the particle swarm method to maximize the likelihood. I have found this to be more reliable than other methods.

Continue reading “ETF Pairs Trading with the Kalman Filter”