Trend Following vs Countertrend Trading Strategies

 

Introduction

A blog series to contrast the key distinctions between trend following and countertrend strategies during building, testing and trading. In this post we examine the effects of data integrity and simulated trade sample size on backtested performance.

Price Data Integrity

One of the major obstacles for traders looking to research trend following models is data. Since trend following models look to "cut losses short and let winners run", profitable trades can last for many months or even years. This inherent characteristic has two important implications. First, it results in much longer trade duration's and consequently fewer simulated trades from a backtest. Second, due to the strong positive skew in trade returns, a small number of highly rewarding trades contribute to the majority of the overall return. These characteristics combined mean that trend following strategies are very sensitive to potential data biases - they cannot tolerate data that has not been fully and properly adjusted for corporate actions and survivorship bias. "Garbage in, garbage out" aptly describes the effect of poor quality data on the backtesting process with respect to trend following. And you're out of luck if you think that you can simulate the effects of perceived data biases - the concentration of overall return, relatively low number of simulated trades and material impact of survivorship bias makes it near impossible to estimate the effects of known data shortcomings when employing poor quality data for trend following backtesting.

Unfortunately, few retail offerings provide the rigour needed to ensure properly adjusted price datasets. It's however possible to acquire data that has been professionally prepared for commercial entities in the asset management space, but these are costly and generally out of reach to the private investor.

Successful countertrend strategies on the other-hand are more short-term in nature, with trades lasting days as opposed to months. The shorter holds result in much higher number of simulated trades from a backtest. Another important distinction is that countertrend strategies have relatively low risk/reward ratios but high win rates, so their performance is not dependent on a few highly rewarding outcomes, but rather many small gains. These attributes – large number of historical trades with short duration's and low past trade return concentration - make countertrend strategies less sensitive to data integrity issues. One additional upside associated with the low trade return concentration (many trades contribute to the overall strategy return, as opposed to few trades as with trend following) is the ability to simulate some of the likely effects of the known data integrity issues on performance. For instance, we could remove the top 10% of most profitable trades from our simulated database to allow for survivorship bias and corporate actions and then rerun the test to determine the effect on overall performance. Essentially, we can emulate a test done on high quality data by massaging the performance numbers downward to allow for perceived data integrity issues.

Many retail offerings provide cheap end-of-day equity price data that are "good enough" to test countertrend strategies. For most retail traders, countertrend strategies are better suited to the data solutions currently available. If you do not have the budget nor understand the intricacies involved in testing long-term strategies, then short-term strategies, such as a countertrend approach, is likely a better place to start.

Simulated Sample Size

As discussed above, countertrend strategies generate a much larger number of simulated trades during a backtest relative to trend following strategies. This is one of the most desirable aspects of a short-term approach because sample size is the single most significant contributor to our confidence in estimating the future – the more simulated trades we have, the higher our confidence in future performance. Smaller samples are more susceptible to the effects of good or bad luck during a backtest, which can over or underestimate the underlying edge that a strategy exploits. Consequently, the expected performance in any given year for a trend following strategy is far less certain relative to a countertrend strategy – our confidence bands are set wider as a direct result of a smaller number of historical trades.

After data integrity, trade sample size from a backtest is the most effective metric to gauge the robustness of a strategy, and oddly enough the least spoken about in trading circles. Sample size is so powerful that it doesn't matter whether or not we understand why a given strategy works - as the trade sample increases, the probability that the strategy works due to chance alone decreases, and ultimately approaches zero. This fact alone is reason enough for most private investors to abandon research on long-term approaches and instead focus on short-term approaches.

Conclusion

Countertrend strategies, or short-term strategies in general, are much more forgiving when it comes to price data integrity issues. Regardless of whether you have high quality data or not, countertrend strategies always provide for higher levels of confidence in future performance due to the greater number of simulated trades relative to their trend following counterparts. For these reasons, most private investors will be better served by focusing their energies on developing short-term trading strategies as opposed to long-term strategies.

In my next post I'll explore and discuss the most appropriate markets for each approach. As always, I welcome your thoughts and suggestions.

Happy Trading,
PJ Sutherland BSc, CMT
CEO QuantLab (Pty) Ltd

    Digging Deeper with a Passive Portfolio

    This is my fourth and final post exploring the merits of passive investing. In my prior posts, I discussed the test results of holding a single ETF over the long-term, we then tried and succeeded in improving the buy-and-hold performance data with rand-cost averaging, and finally, we constructed a universe of ETF's that we can use to engineer a risk parity diversified portfolio that follow the allocation guidelines set by Ray Dalio, the founder of Bridgewater Associates. So in today's post I'm going to share and discuss the test results of our previously constructed portfolio.

    Quantifying a Passive Approach

    To refresh our memories, here are Ray's suggested allocations:

    Gold: 7,5%
    Commodities: 7,5%
    Stocks USA Large Cap: 12.5%
    Stocks USA Mid Cap: 5.5%
    Stocks USA Small Cap 3%
    Stocks International: 6%
    Stocks Emerging Market: 3%
    Long-term bonds: 40%
    Intermediate bonds: 15%

    And the list of ETF's that I selected to clone the above allocations:

    Gold: SPDR Gold ETF NYSE:GLD)
    Commodities: PowerShares DB Commodity Index Tracking (NYSE:DBC)
    Stocks USA Large Cap: SPDR S&P 500 (NYSE:SPY)
    Stocks USA Mid Cap: Vanguard Mid-Cap ETF (NYSE:VO)
    Stocks USA Small: Vanguard Small Cap ETF (NYSE:VB)
    Stocks International: Vanguard Europe Pacific ETF (NYSE:VEA)
    Stocks Emerging Market: Vanguard Emerging Markets Stock ETF(NYSE:VWO)
    Long-term bonds: Vanguard Long-Term Bond ETF (NYSE:BLV)
    Intermediate bonds: Vanguard Intermediate-Term Bond ETF (NYSE:BIV)

    To start, I did not include the effects of transaction costs. These would be minimal, but would however erode some of the performance. Further, there are a number of ways we could construct the portfolio, and a number of different ETF's that could be used for inclusion. In my last post, I quoted the test results (CAGR of +8% with worst annual loss of -4%) achieved by the group of analysts in Tony Robbins' book – Money, Mastering the Game. I however realised a very different set of test data, the cause of which would be attributed to the investment vehicles used and their respective allocations. Notwithstanding this, my aim with this bit of research was to gain a broad and general view of passive investing and present the results to you. The passive investing space has grown significantly over the last couple of years and continues to draw capital from the active space, but I'm not certain that retail clients receive the full picture when advised to go the passive route. Yes, passive is cheaper, has tax benefits and the majority of active mangers underperform. However, there remain the unique few that deliver on the promise of active investing, which is absolute returns, or positive returns in every and any market environment. QuantLab falls into the active space and our aggregate performance to date proves that active investing can deliver on its promise: we have been positive every year and have totalled +25% after QuantLab fees vs the JSE Top 40's +5%, or five times outperformance on a pure return basis. As you'll learn in a minute, we have also outperformed a famed passive approach developed by one of the brightest minds in the industry. So next time you're told that passive investing is the best option for you, be sure to question the returns and drawdown statistics, especially recovery periods, before committing your hard earned funds.

    The Test Results Are In

    Let's start by taking a look at the performance data relative to the S&P 500 total return index. The test period ran from July 2007, the furthest lookback to include all the ETF data, to the end of April 2016. At the end of each year, the portfolio is re-balanced according to our desired allocation percentages.

    The test sample includes the 2008 financial crisis, providing an excellent view of how the portfolio behaves under stress. Given the large allocation to bonds, we expect to see our risk parity portfolio better control losses during market crisis, which is indeed the case. However, the risk-adjusted measures are basically the same, which means we could halve our allocation to the S&P index and enjoy the same return and risk, while having 50% of our capital free for use elsewhere.

    Below I've included the monthly returns and equity curve for our risk parity portfolio. There's no question that our portfolio is smoother through time, a desirable attribute, but since the risk-adjusted measures are the same we could easily replicate this performance as mentioned above by simply halving our allocation to the index. I was hoping to see better numbers here, and I'm certain we could find those with the correct portfolio, but too much scratching around will likely lead to over-optimisation and meaningless performance data. I personally don't see enough here to launch a full on investigation into passive investing, nor would I risk my money on this approach.

    Conclusion

    Passive investing has its place, but be careful of misrepresentations by vendors looking to earn commission by driving clients into the multitude of robotic ETF advisors that have hit the market by storm. Passive investing requires a big dose of patience and psychological fortitude to weather the often steep and extended drawdowns. Don't be too quick to write-off the active space, you may be required to dig a little deeper, but it's my view that your effort will be rewarded. QuantLab is a great place to start, if you'd like to trial our platform simply respond to info@quantlab.co.za.

    Welcome your comments, thoughts and suggestions. Happy Trading!

    PJ Sutherland 

      A Passive Diversified ETF Portfolio

       

      Introduction

      In my last two posts we explored a purely passive approach that holds a single ETF long, in our case it was the Satrix Top 40 ETF which tracks the JSE Top 40, and rand-cost-averaging, which makes period contributions to an ETF as opposed to a single lump sum investment. Our conclusions were 1) holding an ETF on its own requires one to weather significant losses and endure long recovery periods that makes it unattractive as an investment approach and 2) rand-cost-averaging is a highly effective technique that helps reduce volatility and boost risk-adjusted returns. However, both tests focused on a single ETF and therefore neither enjoyed in the benefits of diversification. So our next task then is to investigate the effects of diversification, across both ETF's and asset classes, within a purely passive approach. Once we've built our diversified portfolio of ETF's, it will be interesting to investigate whether we can further improve our performance by including a rand-cost-averaging overlay. But before we commence with either of these tests, we need to first engineer an appropriate diversified ETF universe, as well as explore the target allocations within our portfolio for each asset class. This will be the topic of today's post.

      The All Weather Strategy

      Most of our work here has in fact already been done for us by one of the brightest and most successful long-term hedge fund managers of all-time, Ray Dalio. Ray founded Bridgewater Associates, the biggest hedge fund firm in the world, and many years ago came up with a portfolio that he believed would perform well across all environments. He aptly named the portfolio the All Weather portfolio, and his work started the Risk Parity movement, which focuses on the allocation of risk rather than capital.

      Ray realised long ago that he was not able to consistently predict the future movements of the economy (for a quick reality check, one need only examine the success rate of economists) nor the market. He further believed that he couldn't trust his own experience: anyone's lifetime is too narrow a perspective. So he needed to develop an approach that did not rely on predictions about the future. How? Well he and his team researched the market for twenty-five years and their genius came in the conclusion that there are four things that move the price of assets:

      1. inflation,
      2. deflation,
      3. rising economic growth,
      4. declining economic growth.


      In Ray's view, these are the four possible economic seasons that will ultimately affect whether investments will go up or down. And knowing all the possible future market environments means that we can build a portfolio that will survive through each season, one that does not rely on prediction because it already has something for every market condition.

      Asset Allocations

      One of the key differentiating factors to Ray's approach is the way that he diversifies across asset classes. He took a very different view – at the time this was pioneering and original work – in that he allocates risk as opposed to capital. He focuses on the amount of risk in each asset component rather than the specific rand amounts invested in each component. This is in stark contrast to traditional allocation methods that are based on holding a certain percentage of investment classes, such as 60% stocks and 40% bonds, within one's investment portfolio. Today this technique is well-known and referred to as Risk Parity allocation. So let's take a look at Ray's suggested asset allocations:

      Gold: 7,5%
      Commodities: 7,5%
      Stocks USA Large Cap: 12.5%
      Stocks USA Mid Cap: 5.5%
      Stocks USA Small Cap 3%
      Stocks International: 6%
      Stocks Emerging Market: 3%
      Long-term bonds: 40%
      Intermediate bonds: 15%

      The first observation above is the relatively low allocation to stocks (30% in total), which is the direct result of Risk Parity allocation. Stocks are much riskier than bonds, as measured by their volatility, so in order to balance our portfolio based on risk they receive a much lower allocation. This is a purely passive approach and the performance of this portfolio through time is remarkable – over the past 30 years it's achieved about +8% pa with the worst year being -4% in 2008 and a standard deviation of 8%. This particular strategy does however take a very long-term point of view and looks to protect and preserve wealth against any and all crises from depressions, wars, to hyperinflation.

      ETF Universe

      Now that we know the allocations to each asset class we can construct a portfolio of ETF's that clone the All Weather portfolio in the following way:

      Gold: SPDR Gold ETF NYSE:GLD)
      Commodities: PowerShares DB Commodity Index Tracking (NYSE:DBC)
      Stocks USA Large Cap: SPDR S&P 500 (NYSE:SPY)
      Stocks USA Mid Cap: Vanguard Mid-Cap ETF (NYSE:VO)
      Stocks USA Small: Vanguard Small Cap ETF (NYSE:VB)
      Stocks International: Vanguard Europe Pacific ETF (NYSE:VEA)
      Stocks Emerging Market: Vanguard Emerging Markets Stock ETF(NYSE:VWO)
      Long-term bonds: Vanguard Long-Term Bond ETF (NYSE:BLV)
      Intermediate bonds: Vanguard Intermediate-Term Bond ETF (NYSE:BIV)

      There are a multitude of other variations in terms of ETF's that can be used, but this will suffice for our discussion here. Moreover, you could easily adjust the allocations within the portfolio to suit your needs.

      In my next post we'll build and test our portfolio above and discuss annual re-balancing.

      References

      Bridgewater Associates All Weather Strategy http://www.bwater.com/Uploads/FileManager/research/All-Weather/All-Weather-Story.pdf
      Money Mastering The Game by Tony Robbins

        A Better Approach to Passive Investing

         

        Introduction

        Following on from my last post, this week I'm going to explore the effects of rand-cost averaging when applied to a passive ETF approach. I'll focus on a portfolio with a single ETF for testing, namely the Satrix Top 40 ETF. I'm interested to learn how the performance of rand-cost-averaging compares with investing a single lump sum at the start of the period. But before we get to the results, what is rand-cost averaging?

        What is Rand-Cost-Averaging?

        Rand-cost averaging is the process of making periodic contributions to an ETF as opposed to an initial lump sum investment. Basically, we split our initial investment into equal portions which we then invest periodically, for instance monthly or quarterly. By spreading our investment entry points through time we are able to realise an average entry price which is less affected by market timing. Therefore, the primary benefit of rand-cost-averaging is to mitigate the risk of unlucky market timing. For example, if you invest a lump sum at the start of a major secular bear market you will likely have to endure many years of under-performance and negative returns, however, if you adopted rand-cost-averaging instead, by systematically investing smaller portions periodically, you will have reduced the impact of the bear market on your portfolio because of the better entry prices achieved by buying at lower levels. Moreover, your portfolio will be well-positioned to capitalise on any subsequent recovery because by default it will own more shares in the ETF as a direct result of buying at lower prices. Therefore, rand-cost averaging works best during market downturns and high volatility. Said differently, when markets rise with low volatility, rand-cost-averaging may not be optimal. Notwithstanding this, most investors will be best served by following a disciplined and proven approach such as rand-cost-averaging rather than attempt to time the markets. Let's take a look at the performance numbers.

        Performance Results

        The first test I ran covers the period 2001 through 2015. For the test I allocated R1000 at the start of each new month to the Satrix Top 40 ETF. In total there were 180 such allocations/months bringing our total cash investment to R180 000 during the test period. Because the allocation is made slowly through time, a large portion of our investable equity initially remained in cash. As a result, I included a simple cash return of +5% per annum for the portion in cash. Interest earned was then added to the R1000 monthly contribution for investment in the ETF. The buy-and-hold test simply invested the total contributions made with the rand-cost-averaging approach at the start of the period. Dividends were not considered in either test.

        rand-cost-averaging-satrix40

        Despite slightly lower annual returns, we see a strong improvement in our maximum drawdown statistic, but not enough to improve our risk-adjusted MAR ratio to above that of buy-and-hold (the MAR ratio is simply the absolute value of the CAGR divided by the Max DD. Higher numbers are better.) Therefore, during this time-frame, rand-cost-averaging performed more-or-less in line with a buy-and-hold approach on a risk-adjusted basis. Over longer time-frames we expect this to be the case because we initially have less of our capital exposed to the equity market. But let's see what happens when we apply this approach to a bear market, and there's no better example than the 2008 financial crisis.

        What About Bear Markets?

        rand-cost-averaging-satrix40

        Now we start to see the true power of rand-cost-averaging. Over the four-year period from 2007 to 2010, averaging into the market not only generated superior returns, but did so with half the amount of risk as measured by the maximum drawdown. As a result, the MAR ratio is almost three times greater than a simple buy-and-hold approach. A significant result in favour of this approach.

        And a Bull Market?

        For completeness, let's take a look at the performance over a four-year bull trend.

        rand-cost-averaging-satrix40

        As expected, rand-cost-averaging slightly under-performed on a purely return basis, but in this instance it did so with lower levels of risk, boosting its MAR ratio to above that of buy-and-hold.

        In Conclusion

        In short, rand-cost averaging works. It positions a portfolio to significantly outperform during periods of high volatility and uncertainty. More importantly, it does a very good job of reducing the impact of major bear markets on one's portfolio. I for one am in favour of this approach, and believe this to be a solid approach for investors in general that want to buy some market exposure. That said, we could likely improve on these numbers further. For instance, what about employing some form of timing filter that switches a portfolio to rand-cost-averaging during bear markets and buy-and-hold during bull markets? Moreover, there are an endless number of permutations related to the way the contributions can be made: what about quarterly, bi-annual or annual contributions instead? Or, we could instead restrict our contributions to months when the ETF closes negatively? There are many other interesting ways to exploit the power of rand-cost-averaging, but I leave these to your creativity.

        In my next post I'm going to explore the effects of diversifying across a portfolio of ETF's instead of a single ETF as with this study. We'll start by discussing a reasonable ETF universe to include in the portfolio and then move onto the testing where we'll examine the concept of re-balancing a portfolio.

        As always, I welcome your thoughts and questions.