Market timing model back-testing and optimization

Alpha Asset Allocation investment strategy was developed based on sector rotation model and bear market protection mode. In this post we will review the back-testing optimization results of the market timing model (also called bear protection model). You can read more about our optimization process and how we avoid over fitting in the Sector rotation model back-testing and optimization post.

The market timing model purpose is to get us out of the market (and into safe treasuries) in case of major bear market (ex: 2008 housing crises) and protect us from losing big chunks of our equity. So for testing this model we will use the Max draw-down result as a proxy for successful timing.

We alter the market timing model parameters (PARAM1 and PARAM2) the results measured by MAX SYSTEM DRAWDOWN%  does not show strong drift and remain within the range of 16.7% - 20.5% .

The low drift of the results is thanks to the adaptive nature of this model.


Sector rotation model back-testing and optimization

Alpha Asset Allocation investment strategy was developed based on sector rotation model and market timing model. In this post we will review the back-testing optimization results of the sector rotation model. For this purpose we will use 19 years of historical data from 1/1/1994 to 10/22/2013.

Many trading systems suffer from over-fitting, this is the reason big share of the trading systems in the market which shows very good back-testing results fail when it comes to real trading.

In order to avoid over-fitting our sector rotation model is built using only two parameters (the less parameters the less the chance for over-fitting) and also major parts of the model are adaptive, that  means they automatically adapt to market environment. On top of this the model shows good parametric stability, as seen in the 3D optimization chart when we alter the model parameters (PERIOD and LENGTH) the results measured by CAR (capital annual return) does not show strong drift and remain within the range of 18.7% - 23.7% .

Sector_rotation_backtesting.jpg


Reducing portfolio risk through active management

Warren Buffett’s first rule of investing is "Never lose money". Gosector investment strategy gives great emphasize on risk management, we combine diversification with more active risk management methods.

In the last two decades the strategy had a max worst draw down of ~17% vs. -55% of s&p500. So what are our ingredients of risk control?

1. Diversity - we invest in 3 different sector ETF's/Funds.

2. Market timing - we reduce market risk through a model called "Bear market protection”, this model takes the portfolio to cash or treasuries safety during major bear market

3. Negative volatility is one of the factor taken into account in the sectors ranking process, sector with high negative volatility are ranked lower.

4. Trend strength - sector with strong persistent trend has shown to be more prone to draw-downs.

5. Overbought conditions - is also one of the factors that lower the sector rank so the system avoids buying peaking sectors.

Contrarian Equity Allocation

Warren Buffet once said: “Be Fearful When Others Are Greedy and Greedy When Others Are Fearful”, so how can we implement his contrarian investment philosophy?  How can we combine it with our sector rotation strategy?

First we need to know when the average investor is fearful and when he is greedy, for this purpose we will look at the average investor’s equity allocation, when investors are fearful they tend to allocate more to bonds, treasuries, and cash when they are greedy they tend to risk more on equities.

Average investors equity allocation= Equity/(Bond+Treasuries+Cash)

Average investor allocation graph

Average investor allocation graph

When we look at the attached graph we can see that in times when investors were very greedy like in late 90`s there equity allocation went above 0.5 (more than 50% in equities) ,this over optimism condition was followed by major market crash, on the opposite in times when investors were very fearful like in the bottom of the 2008 crash the equity allocation was about 25% this was the start of a major bull market which still lasts today. So if we want to take a contrarian approach to improve our risk adjusted return we can allocate equity in an opposite way . Here is a simple way for choosing your allocation:

Equity weight=(1-Investors equity allocation)*Risk-Preference

  • Risk-Preference for Aggressive allocation is ~1-1.2
  • Risk-Preference for Moderate allocation is ~0.5-1
  • Risk-Preference for Conservative allocation is ~0.25-0.5
  • For example equity weight for aggressive portfolio today =(1-0.4)*1.2=72% the left 28% is held in Treasuries +Bonds +cash.

 

How we combine it with our sector rotation approach? In times when our sector rotation market timing is invested in sector equities we do not invest 100% of our portfolio in equities but rather use this model. In times when our market timing indicates bearish market conditions this model is not relevant we go 100% to cash and treasuries.

 

Hedging against market corrections

The recent market draw-downs reminds us why it is crucial to stay diversified. Our market timing system is designed to spot major bear markets conditions like the one prevailed during 2000-2002 and 2008, usually it means that the market (s&p500) can lose about 15% - 20% before we will exit the sector positions. 

Fast corrections like the recent one on January or extremely fast market crash like the one happened in 1987 are very hard to spot and act upon, trying to time the market in very high frequency often results in many whipsaws and lower return, so how can we reduce our portfolio draw-downs ?  Our preferred way is to diversify with assets which have negative or very loose correlation to the s&p500:

  1. Long term treasuries are excellent for hedging since they are usually negatively                    correlated to the market and also pay coupons (see attached chart).
  2.  Gold acts as an inflation Hedge.
  3. Short Term treasuries acts as a safe heaven.
S&P500 vs long term treasuries.

S&P500 vs long term treasuries.

How much to allocate to each asset is depending on your personal risk preference and your investment goals. We also use the Contrarian Equity Allocation   to help us estimate if the market is under/overvalued and hedge accordingly. An example of aggressive allocation:

  1. Gold weight= 10% of the portfolio.
  2. Equity weight=1-Average Investors Equity Allocation (can be found in the sector portfolio page) : about 50-75% of the portfolio weight.
  3. Long Term Treasuries: 100%-(gold weight+ equity weight) : about 15-40% of the portfolio.

During Bear market conditions when Alpha Asset Allocation signals to get out of all sectors we replace the equity holding with Short Term treasuries.

 

The two aspects of investor’s financial risk

Usually when speaking on investors risk we think of it as the probability of losing part or all of our investment but more general definition for risk will be the chance that an investor will not meet his financial goals. We can look at risk in two different ways:

  1. Chances of losing some or all of the original investment.

  2. Chances of not meeting our return target.

 

risk-reward.jpg

After we defined the risk more clearly we can deal with each aspect of it in a specific way. First let’s try not to lose. We will take the following measures to reduce the chances of losing big part of our investment:

1. Diversification: Never put all your eggs in one basket, even if you are completely sure you found the holy grail of investments. Diversified assets should exhibit low or negative correlation (Known also as hedging with risk parity) to one another for example stocks and long term treasuries. An example of bad diversification would be holding a portfolio of few stocks from the same sector/industry. You can diversify between stocks, real estate, commodities, bonds, treasuries, futures, options, forex etc. but it would be more wise and easy to diversify between no more that 5-6 asset classes. I consider hedging as part of diversification with low/negatively correlated assets, so for example buying long term treasuries and hedging for inflation using gold is considered as diversification. You can also use geographical diversification like holding equity of emerging and developed markets in different locations, but you need to remember it exposes you to forex exchange risks, for example you can hold a strong stock with positive momentum in the Japanese market but if the Yen (JPY/USD) is becoming weaker you can end up losing.

2. Timing: Find a strategy that will help you successfully time the asset price movement and keep you safely out from big moves that can reduce your equity substantially. Note that not all assets can be timed successfully and even if it can be timed it is not always a simple task. Another thing to understand is that we cannot predict asset prices, at least not for the long term, but we can let the price and fundamental data show us the direction where we have better chances of winning.

3. Liquidity: Holding assets with low liquidity can be more risky since we can have a problem when trying to get out of the position, so even if you find value, momentum or any other edge in low liquidity asset think twice before entering the position.

4. Volatility: High volatility assets tend to be more risky then low volatility assets, before taking this risk be sure that the risk is compensated with higher expected returns.

After briefly learning how to keep our money safer we need to learn how to make it work for us and meet our expected return goals (reasonable return goal..) :

1. Taking the risk and the reward: Over conservative investors should understand that not taking risks (for example holding 100% short term treasuries) will not be very beneficial in the long run. In order to keep pace with inflation and also gain some real benefit of your hard earned money you have to take the risk of holding non cash assets, the compensation of taking this risk will usually be higher reward on the long run. Of course we need to choose carefully what risk we are willing to take and what is the reward we expect, for example we can set a goal of 10% annual return and 15% maximum portfolio drawdown, and then look for ways to implement this targets.

2. Limit your hedging weight: Your hedging position might help you reduce the overall risk in your portfolio but they also carry a cost which results in overall lower return. For example holding few stock and buying put options or short futures on the market will definitely reduce the market risk component of your stock portfolio but you will pay the cost of holding an hedging position that decrease in value on the long run and on top of this you will have less available money to put on the desired stock equities. Generally I do not believe in hedging with instruments that have high costs like put options and market shorts, I rather buy bonds and treasuries that also pay some interest to the holder although they do not have the perfect negative correlation that we seek in hedging positions.

3. Find a strategy with an edge and stick to it: If your target is to outperform the general market you need to do your share of investigation and find yourself a time proven strategy that has some edge on the market , you can use value investing , momentum , trend following, sector rotation , timing etc. The most important thing is once you found a good reliable strategy, stick to it, do not jump from one strategy to another, and do not let your greed/fear emotions run your investment decisions.

As you can see the two aspects of risk sometimes contradicts each other so our job is to find the balanced approach that is tailored for our own needs, our investing style and our psychology.

Market timing -The Golden Cross

The golden cross is one of the most famous methods of market timing, it is based on 50 day and 200 day moving average. The rules are simple: 

  • Buy if the 50 day moving average crosses above 200 day moving average.
  • Sell if the 50 day moving average crosses below 200 day moving average.

For this test we will use the S&P500 index. Back-testing this simple market timing strategy for the last 5 decades (since 1961) resulted in:

  • Annual return of 6.69% (vs. 6.79% buy and hold)
  • Max draw-down of -33% (vs. -56% of buy and hold)
  • Ulcer index of 8.16 (vs. 16.2 of buy and hold)
  • Sharpe of 0.31

We can clearly see that while there was almost no change in annual return we reduced draw-down risk almost by half. Implementing the same strategy on a monthly basis will help us reduce whipsaws and improve performance, these are the results when I used 2 month and 10 month moving average cross.

  • Annual return of 7.11% (vs. 6.79% buy and hold)
  • Max draw-down of -27% (vs. -56% of buy and hold)
  • Ulcer index of 7.1 (vs. 16.2 of buy and hold)
  • Sharpe of 0.35

Nice improvement!

Initial jobless claims - the one macro indicator you should watch

Initial jobless claim report tracks how many new people have filed a claim to receive state jobless benefits in the previous week. The report issued by the U.S. Department of Labor on a weekly basis is a good gauge of the U.S. job market and is highly (negatively) correlated with the economy.

In the graph you can see the ratio of total U.S work force to 4 weeks moving average of initial jobless claims, the shaded areas represent recessions. We can see how this ratio has successfully indicated recession times. his ratio has also high correlation to the stock market which can help us in our market timing decisions. Right now the ratio shows very good positive trend which indicates a strong bullish market.

You can get the data used to construct the graph here:

http://research.stlouisfed.org/fred2/graph/?s[1][id]=IC4WSA#

Initial-jobless-claims.jpg