Momentum Investing

Do commonly used indicators reliably predict stock size, value and momentum strategy returns? In the June 2014 version of his paper entitled “A Comprehensive Look at Size, Value and Momentum Return Predictability”, Afonso Januario examines the abilities of 17 fundamental and technical indicators and indicator combinations to anticipate returns for these three factors. He defines factor portfolios based on market capitalization (size), book-to-market ratio (value) and return from 12 months ago to one month ago (momentum), reformed monthly, as follows:

1. Size = (SmallValue+SmallNeutral+SmallGrowth)/3 – (BigValue+BigNeutral+BigGrowth)/3
2. Value = (SmallValue+BigValue)/2 – (SmallGrowth+BigGrowth)/2
3. Momentum = (SmallWinners+BigWinners)/2 – (SmallLosers+BigLosers)/2

He selects the 17 indicators (such as book-to-market ratio, dividend yield, earnings-price ratio, return on equity, lagged return, short interest and implied volatility) from prior published research on predictive variables. He measures indicator values each month as the averages only for stocks in long or short sides (and the spread between them) of each of the above three factor portfolios. He applies linear regressions at monthly and annual frequencies to determine whether an indicator is more effective than the historical average factor portfolio return in predicting future factor portfolio returns. Using relevant sets of data for a broad sample of relatively liquid U.S. stocks from initial set availability (ranging from 1950 to 1995) through 2012, he finds that: Keep Reading

“Sticky Winner Asset Class ETF Momentum Strategy” tests whether limiting the trading of the “Simple Asset Class ETF Momentum Strategy” by holding onto the winner until it drops out of the top three boosts performance of the latter by reducing trading and thereby suppressing trading frictions. A subscriber proposed a more precise approach to limit trading: continue holding a past winner until it loses to a new winner by a significant margin. To investigate whether this approach (Buffered Winner) works, we compare it to the original strategy (Winner), which allocates all funds at the end of each month to the asset class exchange-traded fund (ETF) or cash with the highest total return over the last five months, as applied to the following nine assets:

PowerShares DB Commodity Index Tracking (DBC)
iShares MSCI Emerging Markets Index (EEM)
iShares MSCI EAFE Index (EFA)
SPDR Gold Shares (GLD)
iShares Russell 1000 Index (IWB)
iShares Russell 2000 Index (IWM)
SPDR Dow Jones REIT (RWR)
iShares Barclays 20+ Year Treasury Bond (TLT)
3-month Treasury bills (Cash)

Using monthly adjusted closing prices for the asset class proxies and the yield for Cash over the period July 2002 (or inception if not available then) through June 2014 (144 months), we find that: Keep Reading

Is there a practical way to apply momentum investing in a Modern Portfolio Theory (MPT) framework? In his June 2014 paper entitled “Momentum, Markowitz, and Smart Beta”, Wouter Keller constructs a long-only, unleveraged Modern Asset Allocation (MAA) model in three steps

1. Make MPT tactical by using short historical intervals to predict future asset returns (rate of return, or absolute momentum), return volatilities (based on daily returns) and return correlations (based on daily returns), assuming that measured behaviors will materially persist the next month. Assign zero weight to assets with negative returns over the historical measurement interval.
2. Simplify correlation calculations by relating daily historical returns for each asset to those for a single market return (the average return of all assets weighted equally) rather than to returns for all other assets separately.
3. Dampen errors in rapidly changing asset return, volatility and correlation predictions by “shrinking” them toward their respective averages across all assets in the universe, and dampen the predicted market volatility by “shrinking” it toward zero.

He reforms the MAA portfolio monthly at the first close. His baseline historical interval for estimation of all variables is four months (84 trading days). His baseline shrinkage factor for all variables is 50%. His principal benchmark is the equally weighted (EW) “market” of all assets, rebalanced monthly. He assumes one-way trading friction of 0.1%. He considers a range of portfolio performance metrics: annualized return, annual volatility, maximum drawdown, turnover, Sharpe ratio, Omega ratio and Calmar ratio. Using daily dividend-adjusted prices for assets allocated to three universes (10 exchange-traded funds [ETF], 35 ETFs and 104 U.S. stocks/bonds) during December 1997 through December 2013, he finds that: Keep Reading

The “Simple Asset Class ETF Momentum Strategy” each month allocates all funds to the one of the following eight asset class exchange-traded funds (ETF), or cash, with the highest total return over the past five months:

PowerShares DB Commodity Index Tracking (DBC)
iShares MSCI Emerging Markets Index (EEM)
iShares MSCI EAFE Index (EFA)
SPDR Gold Shares (GLD)
iShares Russell 1000 Index (IWB)
iShares Russell 2000 Index (IWM)
SPDR Dow Jones REIT (RWR)
iShares Barclays 20+ Year Treasury Bond (TLT)
3-month Treasury bills (Cash)

“Simple Asset Class ETF Momentum Strategy Robustness/Sensitivity Tests” shows that, among uniform ranking intervals, five months is optimal. Citing the optimality of a three-month ranking interval in “Simple Debt Class Mutual Fund Momentum Strategy”, a subscriber inquired whether using a three-month ranking interval just for TLT might improve Simple Asset Class ETF Momentum Strategy performance. To investigate more generally, we compute net terminal values for 108 variations of the strategy by letting the ranking interval for each asset range from one to 12 months, while holding the ranking interval for all other assets at five months. In order to compare ranking intervals of different lengths, we use the average total return per month for ranking. For example, the average monthly total return for a five-month ranking interval is the five-month total return divided by five. Using monthly dividend-adjusted closes for the asset class proxies and the yield for Cash during July 2002 (or inception if not available then) through May 2014 (141 months), we find that:

Keep Reading

A subscriber suggested an alternative to the “Simple Asset Class ETF Momentum Strategy” that weights asset class ETFs according to five-month past return ranking (such as 35-25-20-10-4-3-2-1) rather than allocating all funds to the winner. Do the diversification benefits of this alternative outweigh the loss of momentum purity? To investigate, we return to the following eight asset class exchange-traded funds (ETF), plus cash:

PowerShares DB Commodity Index Tracking (DBC)
iShares MSCI Emerging Markets Index (EEM)
iShares MSCI EAFE Index (EFA)
SPDR Gold Shares (GLD)
iShares Russell 1000 Index (IWB)
iShares Russell 2000 Index (IWM)
SPDR Dow Jones REIT (RWR)
iShares Barclays 20+ Year Treasury Bond (TLT)
3-month Treasury bills (Cash)

As one benchmark, we allocate all funds at the end of each month to the asset class ETF or cash with the highest total return over the past five months (5-1). As another benchmark, we maintain an equal-weighted (EW), monthly rebalanced portfolio of all nine asset classes. As alternatives, we test two momentum rank-weighted (RW), linearly-scaled combinations of all nine classes, one steep across ranks and one shallow. We also test EW combinations of the Top 5, Top 4, Top 3 and Top 2 momentum ranks. Using monthly adjusted closing prices for the asset class proxies and the yield for Cash over the period February 2006 (the earliest all ETFs are available) through May 2014 (100 months), we find that: Keep Reading

We have updated the detail at “Momentum Strategy” to incorporate the historical data changes described in “Simple Asset Class ETF Momentum Strategy Data Changes”. The principal effects are to decrease the performance of the Top 1 portfolio and increase the performance of the Top 2 portfolio.

Over the coming week, we will accordingly update much of the supporting research listed at “Momentum Strategy”.

Do individuals who actively reallocate funds within their pension accounts outperform passive counterparts? In the March 2014 update of their paper entitled “Individual Investor Activity and Performance”, Magnus Dahlquist, Jose Vicente Martinez and Paul Soderlind examine the activity and performance of individual participants in Sweden’s Premium Pension System. This system allows individual participants to reallocate among available mutual funds on a daily basis with no switching fees/impediments. Information about the 1,230 funds offered during the sample period includes type (fixed income, balanced, life-cycle and equity), return and risk measured at several horizons, fee and major holdings. Most are equity funds, about half of which invest primarily in international equities. The government assigns individuals who make no choice to a default fund. Using daily net returns, fund trades and demographics for 70,755 individuals (from a random draw of individuals in the system over the entire period) and contemporaneous returns for several benchmarks during September 2000 through May 2010, they find that: Keep Reading

Does the return to momentum investing reliably reflect a reward for taking some risk? In the March 2014 version of her paper entitled “Asymmetric Risks of Momentum Strategies”, Victoria Dobrynskaya examines the performance of past winner and loser stocks, stock indexes and currency exchange rates during global equity market advances and declines. She focuses on the difference between downside and upside market betas (asymmetric beta) of momentum portfolios that are long past winners and short past losers (winner-minus-loser, or WML). She bases momentum rankings on total returns from 12 months ago to one month ago, with a month skipped before portfolio formation to avoid short-term reversal. She considers momentum portfolios formed from: a broad sample of U.S. stocks (January 1927 through July 2013): global and world regional stocks (November 1990 through August 2013); 40 country stock indexes (as they become available during January 1983 through August 2013); and, 45 currency exchange rates relative to the U.S. dollar (as they become available during October 1983 through August 2013). She also examines U.S. short-term reversal stock portfolios sorted by prior-month return (January 1927 through July 2013) and U.S. long-term reversal stock portfolios sorted by past five-year return (January 1931 through July 2013). She relies on Kenneth French’s data library for stock-level portfolio returns. Using total return data for the specified portfolios and for associated market indexes, she finds that: Keep Reading

Is relative momentum an effective way to switch between stocks and bonds? In his May 2014 paper entitled “Simple and Effective Market Timing with Tactical Asset Allocation”, Lewis Glenn compares the performances of two tactical asset allocation strategies:

1. SPY-TLT:  simple relative momentum strategy that each month holds SPDR S&P 500 (SPY) or iShares Barclays 20+ Year Treasury Bond (TLT) depending on which has the higher total return over the last three months.
2. Ivy 5: diversified trend following strategy that initially allocates equal amounts to each of five funds: exchange-traded funds (ETF) representing U.S. stocks, non-U.S. large stocks, 7-10 year U.S. Treasuries and real estate, and a mutual fund representing commodities.Thereafter, each allocation remains in the fund (goes to cash) when the fund is above (below) its 10-month simple moving average (SMA10) the prior month.

He includes “nominal” transaction fees for trades but does not account for bid-ask spread or impact of trading frictions. Using monthly dividend-adjusted closes of the specified funds during 2004 (plus prior data for momentum/trend calculations) through 2013, he finds that:

Keep Reading

Do relative momentum and trend filters operate differently on value and growth stocks? In their May 2014 paper entitled “When Growth Beats Value: Removing Tail Risk from Global Equity Momentum Strategies”, Andrew Clare, James Seaton, Peter Smith and Stephen Thomas investigate the effects of relative momentum and trend filters on portfolios of developed and emerging market broad, value and growth stock indexes. Their relative momentum filter each months picks either the top five indexes (Mom5) or top quarter of indexes (Mom25%) based on volatility-adjusted past 12-month return (return divided by standard deviation of monthly returns) at the end of the prior month. Their trend filter each month invests in an index or U.S. Treasury bills (T-bills) according to whether the index is above or below its 10-month simple moving average (SMA10) at the end of the prior month. Using monthly levels of broad, value and growth stock indexes for 23 developed country markets (since 1976) and 21 emerging country markets (since 1998) through 2012, they find that: Keep Reading