|
This article will discuss a simple method to use moving averages and moving average channels to increase a portfolio's return, while reducing risk, and avoiding over-trading or highly active management of a portfolio. Adjusting a portfolio on a monthly basis offers several disadvantages to a buy and hold portfolio. Namely, these disadvantages include difficulty in executing trades in a timely manner each and every month, and short term trades that eat away at returns in a taxable account and incur often unnecessary commissions on all accounts. Overtrading gives an investor a reason to doubt their strategy and the trades their strategy is executing each and every month, or possibly more often than this. Between doubting trades, and not wanting to take the time to execute every single month, investors often don't trade a complex strategy, and end up missing the best of what the strategy has to offer, and ultimately the results are often less than optimal. Overtrading also has very tangible negative consequences in taxable accounts, that is incurring short term gains instead of long term gains, thus increasing the amount of taxes you need to pay on your gains. 10% gains taxed at up to 39.6% for short term gains is worth a lot less than 10% gains taxed at the long term gains rates of 0-20% at the end of the year (US tax rates)! The simple, and default strategy for avoiding overtrading and short term gains in an account is a buy and hold strategy. If rebalanced yearly there is only 1 trade per year, and no short term gains. However, as we will see, a buy and hold strategy is susceptible to large downturns in the market. In order to gain exposure to different sectors upswing, we also have to be involved in their respective down turns using classical buy and hold methods. So the question comes to mind, "Can we avoid the down turns in a market, but be involved in the upswing of each market?". In our article we want to backtest a simple, and practical method for building a US sector portfolio and then seeing if applying a moving average and moving average channels can benefit this simple strategy. Of course these timing strategies will result in more trading than a buy and hold portfolio, so we will look at how often they trade and how many short term gains we expect to see. Portfolio Timing Logic The portfolio we will test from January 2000 to September 2016, the portfolio will consisting of equal parts XLP (Consumer Staples), XLY (Consumer Discretionary), XLE (Energy), XLU (Utilities), XLI (Industrial), XLB (Materials), XLF (Financial), XLK (Technology), and XLV (Health Care). When any of these funds falls below the 175 day simple moving average, we will de-invest in the respective fund, and instead invest in a short term bond, cash like fund. Portfolio Channels Logic We will use the same portfolio, and the same 175 day simple moving average. However we will only invest in a fund if it goes 2.5% above this moving average, and we will only de-invest and move our money to cash if a fund goes 2.5% below this moving average. The purpose of the Channels is to reduce the possibility of getting in and out of the market over and over again when it is not necessary.  Simple Buy and Hold Portfolio With Yearly Rebalancing Below are the results for rebalancing a buy and hold portfolio with these funds in equal amounts since 2000:    The results show 6.46% return per year, with yearly rebalancing there are no short term gains to report for a taxable account, but there are large drawdowns during poor market conditions (Maximum drawdown is 52.68%). Portfolio Timing with Cash (SHY) Below are the results for the portfolio timing strategy using a cash like fund (SHY) as the de-invested fund since 2000:     Timing improved the results from 6.46% return per year to 9.16% return per year while reducing drawdown from 52.68% to 19.11%. The average trade length shows most trades over 3% lasting over a year. A vast improvement just by adding 1 simple moving average filter, a simple rule with drastic benefits to return. Portfolio Timing with 7-10 Year Treasury Bond (IEF) Below are the results for the portfolio timing strategy using a 7-10 Year Treasury Bond (IEF) as the de-invested fund since 2000:     Using IEF instead of SHY resulted in a solid increase in returns to 10.71% return per year. Portfolio Channels with Cash (SHY) Below are the results for the portfolio channels strategy using a cash like fund (SHY) as the de-invested fund since 2000:     The Channels strategy is a vast improvement over buy and hold, and also better than the simple Portfolio Timing strategy. The average trade length shows almost all trades over 3% lasting over a year. This means less trading and less short term income to report on taxable accounts when compared to the simpler Portfolio Timing options. Portfolio Channels with 7-10 Year Treasury Bond (IEF) Below are the results for the portfolio timing strategy using a 7-10 Year Treasury Bond (IEF) as the de-invested fund since 2000:     The channels strategy with IEF obtains 10.9% return over the duration of the test, with the same longer trades as the SHY model.
Conclusion A simple buy and hold portfolio has a simple once a year update schedule with no short term trades, at the expense of lower performance, more drawdown, and more volatility. The Portfolio Timing with longer term bonds shows the much better performance, but has a number of short term trades that could negatively impact a taxable account as well as make portfolio management an ongoing task each month. The Portfolio Channels results shows performance numbers even better than the simple moving average crossover of the Portfolio Timing Tool, with less short term trades, and less trading overall, making this option a possible compromise between a moving average strategy and a pure buy and hold portfolio. Overtrading was reduced by using a channel method instead of a simple moving average crossover strategy. Overtrading can cause psychological difficulties for investors as the market moves up and down, doubt and busyness get in the way of executing strategies correctly. Overtrading also increases your short term tax burden in taxable accounts, reducing trades to a minimum eliminates short term trades and taxable events in an account. Rotation strategies have often been touted as a way to increase returns, reduce drawdown, and avoid the nasty drawdowns associated with buy and hold portfolios and improving risk adjusted returns for holding single funds or stocks. Bear markets and corrections have proven difficult for buy and hold strategies, resulting in drawdowns, and less than stellar performance in great markets. Bonds surged while the stock market crashed in 2008, but many bonds faltered for much of 2013 while the stock market soared. If you recall 2008, the S&P 500 lost around 55% of it's value, but Gold didn't miss a beat until it has lost 1/3 of it's value from 2011-2013. Recently the stock market has been volatile but the bonds markets have surged upward. Why do we hold onto funds in our portfolios that are performing so poorly and are very volatile instead of just removing them from our portfolio or only investing in the best funds? One obvious answer is to just get rid of funds that are doing poorly and wait for them to do better again before getting back into that fund. That is the idea rotation investing attempts to accomplish quantitatively instead of just relying on emotions or guesses about when funds are doing poorly. Rotation Based Strategies The idea behind rotation strategies is simple, each month (or week or quarter...) you check a list of stocks, ETFs and/or mutual funds to see which is stronger and buy the top 1 or more funds and sell the old funds. This results in a portfolio constructed of only the strongest funds each month. These type of strategies have proven very effective in avoiding bear markets in the past while gaining exposure to uptrends. So instead of holding onto funds when they are dropping or very volatility like buy and hold lets test out rotating or swapping out these funds for stronger funds. Many rotation strategies have suffered in 2015, however my adding mean reversion into strategies there has been a lot of success in the past few years. Building a high return strategy using RotationInvest.com's Tools Tool: RotationInvest.com's Advanced Rotation Tool Funds: QLD, MIDU, and TMF Settings: Choose top 2 funds, 40% 3 month momentum, 60% 10 day mean reversion.      The performance for this rotation strategy is: 2040% Total Return over the period 49.99% CAGR 1.39 Sharpe Ratio 33.33% Volatility 21.94% maximum monthly drawdown No single year had a negative return Monte Carlo Warning The Monte Carlo simulation allows users to run a random sampling of possible performances and stress test a portfolio's results. This shows users possible outcomes given a return and volatility. What this simulation shows is for this particular strategy the higher volatility numbers mean a much larger range of possible outcomes, anywhere from ~$2M to more than $25M! The Monte Carlo simulation reveals an average of $8,050,000 average account balance after only 10 years with $100K starting balance and $1K monthly deposit, however the 90th Percentile and 10th Percentile (really good vs. really bad outcomes) vary dramatically due to strategy volatility! This difference in best possible outcome vs. a bad outcome shows how a high return strategy like this can result in vastly different outcomes even when they perform well in the future.  
Strategies presented anywhere on this site are ideas only, and only for education. This is not investment advice, or suggested financial information.
Please read our Terms and Conditions. Summary RotationInvest.com now allows using any of our backtesting tools with historical data going back as far as 1924. In this article we will look at the performance of a simple rotation strategy between 3 asset classes, US Large Cap, US Small Cap, Global Stocks, US Financial Sector, US Intermediate Term Treasury Bonds, and US Intermediate Term Total Bond Market based on their momentum since 1969. Example Strategy Funds Selected: US Large Cap, US Small Cap, Global Stocks, US Financial Sector, US Intermediate Term Treasury Bonds, and US Intermediate Term Total Bond Market Position Score: 3 Month Momentum. Note: we found anywhere from 2 to 12 months work almost as well.  [Equity Chart without a Log Plot]  Performance shows 12.36% returns since 1969, with a 0.92 Sharpe Ratio and a maximum month to month draw down of 24.6%. Beside the strategies performance you can see that US Large Cap had 9.23% returns during the same period, but had a 50.65% drawdown and only a 0.59 Sharpe Ratio.  Below you can see how often the strategy traded into each equity or bond option:  Looking at the yearly return we see 1973 was the worst year with a 12.3% loss on the year, while 1982 was the strongest year with 42.3% gains for the year.
See https://www.rotationinvest.com/blog/historical-data-since-1924 for more information about our historical data options inside our tools. Disclosure: This data is provided partially by St. Louis Fed Web Services [FRED], read their terms of service before using: https://research.stlouisfed.org/docs/api/terms_of_use.html This product uses the FRED® API but is not endorsed or certified by the Federal Reserve Bank of St. Louis. Introduction Building a portfolio strategy that has performed well over the past 10-15 years using ETFs (Exchange Trade Funds) is a simple matter, there is a ton of information widely available that shows performance numbers for a short section of stock market history. But does this strategy perform over the long haul, does it perform over a significant period of history or does it just happen to work for a little while? In this article we backtest a simple strategy over more than 60 years of historical data to see how one of the simplest strategies has performed during that period. Strategy Our core portfolio will be composed of 40% US Mid Cap Value Stocks, 40% US Small Cap Growth Stocks, and 20% Total World Stocks. When any one of these categories is below the 175 day moving average we will invest in Intermediate Term US Bonds. Investing in a bond fund instead of equities will allow us to limit our risk. Results During the period from November 1954 to today, US Large Cap stocks have provided 9.16% average yearly return, our portfolio without timing has given 10.51% returns if you rebalance it yearly. However our portfolio has also had 57.4% drawdown and only a 0.72 Sharpe Ratio over that period. Adding in portfolio timing to this portfolio is an active strategy that's goal is to avoid draw downs by de-investing in funds as they fall below moving averages, and re-invest when the fund goes back above the moving average. Our strategy is to apply the 175 day (~8 month) moving average to this portfolio, and only invest in funds that are above this moving average. We will be able to see if this strategy is profitable over the long term, or if this management style only works over smaller sections of history.  Results: 11.26% CAGR, 1.11 Sharpe, 6.86% daily volatility, 29.36% maximum daily drawdown. Below we see the strategy is invested ~70% of the time, while ~30% of the time the strategy invests in the bond fund.  Below is a chart that shows time in trades vs. return. If your account is a taxed account, this is a major concern to make sure trades are held at least 1 year. Below shows 1 year is the minimum length of time any trade making more than 15% occurs, although there are several trades in the 10-15% range that last for shorter periods of time.   Conclusion
Over more than 60 years of backtesting adding a timing aspect to a portfolio results in superior results with less drawdown and a greater risk/return ratio. The drawdown over a long period of time was reduced from 57.4% to 29.36%, while increasing the annual percent gain from 10.51% to 11.26%. A simple portfolio strategy of investing in funds when they are above a moving average, and switching to a bond fund when they are below a moving average worked over a long period of time to reduce strategy drawdown and increase portfolio returns. Disclosure: Data is provided partially by St. Louis Fed Web Services [FRED], read their terms of service before using: https://research.stlouisfed.org/docs/api/terms_of_use.html This product uses the FRED® API but is not endorsed or certified by the Federal Reserve Bank of St. Louis. Monte Carlo simulations are run on every backtest our tools perform, this tool provides the ability to test long term expected portfolio growth based on savings and account size, and portfolio survival based on withdrawals, and if a portfolio can survive the planned withdrawals during the retirement years. What the Monte Carlo simulation does is take the expected returns rate from the backtest and the volatility values from the backtest, and use these values to construct a normal distribution of expected returns and randomly choose outcomes up to 1000 times to give an expected portfolio return. Since results are based on the backtests return % and volatility % the results are only as accurate as the backtest is an accurate representation of future results. We also have stand alone Monte Carlo simulators for savings and withdraw, where users can manually input their expected return % and volatility %. Calculation of the normal distribution has density: f(x) = 1/(√(2 π) σ) e^-((x - μ)^2/(2 σ^2)) where μ is the mean of the distribution and σ the standard deviation. After each backtest the tool automatically runs a Monte Carlo simulation on the result, and allows the user to interact with changing the inflation, advisor fees, duration of simulation, and account size and deposit/withdraw amounts. To change from 'Savings' to 'Withdraw' mode select those options from the dropdown. In savings mode the strategy will allow users to simulate contributing to an account, and in withdrawal mode the strategy will allow users to simulate withdrawing from account each month.  We also have stand alone Monte Carlo simulation tools, called the Retirement Savings and Retirement Withdrawal tools, they allow users to input their expected % return and their expected volatility %. To understand the purpose of each input in the Monte Carlo simulation tool please read our How to articles on the stand alone tools: Please keep in mind the results are only as accurate as the information you put into the tools, therefore if you put in values for return and volatility and other entries that do not reflect future market conditions this tool will not show accurate information. These tools only shows possible futures, and is not a guarantee of future account success. RotationInvest.com takes no responsibility for the accuracy of results.
A very exciting new feature at RotationInvest.com is our ability to run backtests on data as far back as 1924 on a daily basis. The historical data uses a mix of information from St. Louis Fed Web Services [FRED], mutual fund data, and index data. All of our backtesting tools have the ability to use this Historical Data and allow the full functionality of our tools using this data. So just to be clear, this data is not locked away and only available for limited backtesting, it is fully available in all of our backtesting tools with all of the settings and algorithms available. This includes weighting and adaptive asset allocation using momentum, Sharpe Ratio, mean reversion, volatility and the rest! Step 1: Select Long Term Historical Test from the Data Source for Backtesting dropdown at the top of each tool.  Step 2: Select from the list which funds you would like to choose or invest in. The categories of funds are pre-populated once you select Historical Data.  Below is a list of fund categories that are available to select from for Historical Data: US Markets and Sectors and Commodities US Large Cap Index this data is based on: Russell 1000 [Data Since 1924] US Large Cap Value Index this data is based on: Russell 1000 Value [Data Since 1931] US Large Cap Growth Index this data is based on: Russell 1000 Growth [Data Since 1925] US Mid Cap Index this data is based on: Russell Midcap [Data Since 1938] US Mid Cap Value Index this data is based on: Russell Midcap Value [Data Since 1949] US Mid Cap Growth Index this data is based on: Russell Mid Cap Growth [Data Since 1935] US Small Cap Index this data is based on: Russell 2000 [Data Since 1956] US Small Cap Value Index this data is based on: Russell 2000 Value [Data Since 1968] US Small Cap Growth Index this data is based on: Russell 2000 Growth [Data Since 1946] US Micro Cap Index this data is based on: Wilshire US Micro-Cap Total Market Index [Data Since 1978] Gold Index this data is based on: London Bullion Market Association (LBMA) Gold Price [Data Since 1968] US Convertibles/Preferred Stocks Index this data is based on: BofA Merrill Lynch - Convertible Bonds All Qualities [Data Since 1956] US Sector - Communications Index this data is based on: S&P 1500 Telecom Services [Data Since 1984] US Sector - Financial Index this data is based on: S&P 1500 Financials [Data Since 1963] US Sector - Health Index this data is based on: S&P 1500 Health Care [Data Since 1981] US Sector - Natural Resources Index this data is based on: S&P North American Natural Resources [Data Since 1969] Sector - World Metals/Mining Index this data is based on: MSCI World - Metals & Mining [Data Since 1956] US Sector - US Real Estate REIT Index this data is based on: Wilshire US Real Estate Investment Trust Total Market Index [Data Since 1977] US Sector - Technology Data Since 1948 US Sector - Utilities Index this data is based on: S&P 1500 Utilities [Data Since 1948] World and Emerging Markets Total World Stocks (including US) Data Since 1954 World [Ex. US] Large Cap Index this data is based on: MSCI ACWI Ex USA [Data Since 1961] World [Ex. US] Large Cap Value Index this data is based on: MSCI ACWI Ex USA Value [Data Since 1981] World [Ex. US] Large Cap Growth Index this data is based on: MSCI ACWI Ex USA Growth [Data Since 1981] World [Ex. US] Mid/Small Cap Value Index this data is based on: MSCI World Ex USA SMALL-MID Value [Data Since 1991] World [Ex. US] Mid/Small Cap Growth Index this data is based on: MSCI World Ex USA SMALL-MID Growth [Data Since 1988] Emerging Markets (Diversified) Index this data is based on: MSCI EM [Data Since 1989] European Equities Index this data is based on: MSCI Europe [Data Since 1986] Chinese Equities Index this data is based on: MSCI China [Data Since 1992] Pacific [Ex. Japan] Equities Index this data is based on: MSCI AC Far East Ex Japan [Data Since 1989] Japanese Equities Index this data is based on: MSCI Japan [Data Since 1962] Latin American Equities Index this data is based on: MSCI EM Latin America [Data Since 1991] Global Real Estate (REIT) [ex. US] Index this data is based on: S&P Global REIT [Data Since 1989] Bonds US Government Bonds - Long Term Index this data is based on: Barclays US Government Long [Data Since 1983] US Government Bonds - Intermediate Term Index this data is based on: Barclays US Government [Data Since 1969] US Government Bonds - Short Term Index this data is based on: Barclays Government 1-5 Yr [Data Since 1975] US Treasury Inflation Protected Securities (TIPS) Index this data is based on: Barclays US Treasury US TIPS [Data Since 1988] US Total Bond Market - Long Term (Investment Grade & Government) Index this data is based on: Barclays US Govt/Credit Long [Data Since 1973] US Total Bond Market - Intermediate Term (Investment Grade & Government) Index this data is based on: Barclays US Agg Bond [Data Since 1954] US Total Bond Market - Short Term (Investment Grade & Government) Index this data is based on: Barclays US Govt/Credit 1-5 Yr [Data Since 1971] US Total Bond Market - Ultra-Short Term (Investment Grade & Government) Index this data is based on: Barclays Govt/Corp 1 Yr Duration [Data Since 1983] US High Yield Corporate Index this data is based on: BofA Merrill Lynch US High Yield Master II [Data Since 1935] Emerging Market Bonds Index this data is based on: JPM EMBI Global [Data Since 1994] Municipal Bonds - High Yield Index this data is based on: Barclays Municipal 10 Yr 8-12 [Data Since 1976] Municipal Bonds - Long Term Index this data is based on: Barclays Municipal 20 Yr 17-22 [Data Since 1976] Municipal Bonds - Intermediate Term Index this data is based on: Barclays Municipal 10 Yr 8-12 [Data Since 1976] Municipal Bonds - Short Term Index this data is based on: Barclays Municipal 3 Yr 2-4 [Data Since 1977] Data is provided partially by St. Louis Fed Web Services [FRED], read their terms of service before using: https://research.stlouisfed.org/docs/api/terms_of_use.html
This product uses the FRED® API but is not endorsed or certified by the Federal Reserve Bank of St. Louis. |
|||