Momentum Investing

Are positive carry and positive trend conditions consistently favorable across asset classes? In their March 2015 paper entitled “Carry and Trend in Lots of Places”, Vineer Bhansali, Josh Davis, Matt Dorsten and Graham Rennison employ futures prices to investigate whether the adages “don’t pay too much to hold an investment” and “don’t fight the trend” actually work across four major asset classes: equities, bonds, commodities and currencies. For testing, they select five liquid markets with relatively long futures histories within each asset class. They define carry as annualized excess return assuming that spot prices do not change. They define trend as positive (negative) if the futures price today is above (below) its one-year trailing moving average. They specify four states for each market:

1. Positive carry and positive trend (Carry + / Trend +).
2. Positive carry and negative trend (Carry + / Trend -).
3. Negative carry and positive trend (Carry – / Trend +).
4. Negative carry and negative trend (Carry – / Trend -).

They then calculate average subsequent daily excess returns for each market by state and annualize results. Using daily futures data as available and some simulated futures data (from spot prices) for 20 major markets across four asset classes during 1960 through 2014, they find that: Keep Reading

A subscriber hypothesized that combining short-term reversal with intermediate-term momentum would enhance momentum strategy performance. To investigate, we test a modification of the “Simple Asset Class ETF Momentum Strategy”, which each month allocates all funds at the end of each month to the one of the following asset class exchange-traded funds (ETF) or Cash with the highest total return over the past five months (Top 1):

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)

The modification each month first identifies the top three ETFs or Cash based on past five-month returns and then picks the one of these three with the lowest return over the past five trading days (Top 3 Loser). This approach should pick intermediate-term winners that tend to benefit (or at least not suffer) from any reversal of short-term movements. Using daily and monthly dividend-adjusted closing prices for the asset class proxies and for SPDR S&P 500 (SPY) and the yield for Cash during February 2006 (when all ETFs are first available) through February 2015 (109 months), we find that: Keep Reading

A subscriber requested corroboration of the findings in “Simple Debt Class Mutual Fund Momentum Strategy” with a universe restricted to a family of bond funds (such as Fidelity) to enable low-cost fund switching. We therefore apply the strategy to the following ten Fidelity mutual funds:

Investment Grade Bond (FBNDX)
Intermediate Bond (FTHRX)
Government Income (FGOVX)
Mortgage Securities (FMSFX)
GNMA (FGMNX)
Short-Term Bond (FSHBX)
Limited Term Government (FFXSX)
Convertible Securities (FCVSX)
Intermediate Government Income (FSTGX)
Fidelity New Markets Income (FNMIX)

Per the prior test, we allocate all funds at the end of each month to the fund with the highest total return over the past three months (3-1). We determine the first winner in May 1994 to accommodate momentum measurement interval sensitivity testing. Using monthly dividend-adjusted closing prices for the ten funds during May 1993 (as limited by FNMIX) through January 2015 (261 months), we find that: Keep Reading

Do stock return anomalies exhibit January and month-of-quarter (first, second or third, excluding January) effects? In his February 2015 paper entitled “Seasonalities in Anomalies”, Vincent Bogousslavsky investigates whether the following 11 widely cited U.S. stock return anomalies exhibit these effects:

1. Market capitalization (size) – market capitalization last month.
2. Book-to-market – book equity (excluding stocks with negative values) divided by market capitalization last December.
3. Gross profitability – revenue minus cost of goods sold divided by total assets.
4. Asset growth – Annual change in total assets.
5. Accruals – change in working capital minus depreciation, divided by average total assets the last two years.
6. Net stock issuance – growth rate of split-adjusted shares outstanding at fiscal year end.
7. Change in turnover – difference between turnover last month and average turnover the prior six months.
8. Illiquidity – average illiquidity the previous year.
9. Idiosyncratic volatility – standard deviation of residuals from regression of daily excess returns on market, size and book-to-market factors.
10. Momentum – past six-month return, skipping the last month.
11. 12-month effect – average return in month t−k*12, for k = 6, 7, 8, 9, 10.

Each month, he sorts stocks into tenths (deciles) based on each anomaly variable and forms portfolios that are long (short) the decile with the highest (lowest) values of the variable. He updates all accounting inputs annually at the end of June based on data for the previous fiscal year. Using accounting data and monthly returns for a broad sample of U.S. common stocks during January 1964 to December 2013, he finds that: Keep Reading

A subscriber inquired whether the “Simple Asset Class ETF Momentum Strategy” (SACEMS) is a good diversifier of the U.S. stock market. This strategy allocates funds at the end of each month to the one (Top 1), equally weighted two (EW Top 2) or equally weighted three (EW Top 3) of the following 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)

To investigate, we first look at correlations between momentum strategy returns and those of SPDR S&P 500 ETF (SPY) and Vanguard Balanced Index Investor Shares (VBINX), with the latter maintaining an approximately 60% allocation to the broad U.S. stock market and a 40% allocation to the U.S. corporate bond market. We then generate return statistics for portfolios that hold equally weighted combinations of: (1) the Top 1 momentum strategy and SPY, and (2) Top 1 and VBINX. Using monthly dividend-adjusted returns for the specified funds and the monthly Treasury bills yield as a proxy for Cash during January 2003 through January 2015, we find that: Keep Reading

Are overnight trading motivations systematically different from those that drive trading during normal trading hours? In the January 2015 version of their paper entitled “Tug of War: Overnight Versus Intraday Expected Returns”, flagged by a subscriber, Dong Lou, Christopher Polk and Spyros Skouras (1) decompose abnormal returns associated with well-known stock return predictors into overnight and intraday components and (2) investigate whether differences between institutional and other traders account for differences. Using return, firm characteristic and institutional ownership data for a broad sample of U.S. stocks (excluding low-priced and the smallest fifth of stocks) during 1993 through 2013, they find that: Keep Reading

In reference to “Optimal Monthly Cycle for Simple Asset Class ETF Momentum Strategy?”, a subscriber asked about an optimal monthly cycle for the “Simple Debt Class Mutual Fund Momentum Strategy”. This latter strategy each month allocates the entire portfolio value to the one of the following 12 debt class mutual funds with the highest past total return (optimally over the last two months):

T. Rowe Price New Income (PRCIX)
Thrivent Income A (LUBIX)
Vanguard GNMA Securities (VFIIX)
T. Rowe Price High-Yield Bonds (PRHYX)
T. Rowe Price Tax-Free High Yield Bonds (PRFHX)
Vanguard Long-Term Treasury Bonds (VUSTX)
T. Rowe Price International Bonds (RPIBX)
Fidelity Convertible Securities (FCVSX)
PIMCO Short-Term A (PSHAX)
Fidelity New Markets Income (FNMIX)
Eaton Vance Government Obligations C (ECGOX)
Vanguard Long-Term Bond Index (VBLTX)

To investigate, we compare 21 variations of the strategy based on shifting the monthly return calculation cycle relative to trading days from the end of the month (EOM). For example, an EOM+5 cycle ranks funds based on closing prices five trading days after EOM each month. We use the historically optimal two-month fund momentum measurement interval. Using daily dividend-adjusted closes for the 12 funds during mid-December 1994 through mid-January 2015 (241 months), we find that: Keep Reading

Readers have suggested three alternative metrics for the strategy tested in the “Simple Sector ETF Momentum Strategy Performance”: (1) Sharpe Ratio over the past six months; (2) slope of price over the past six months; and, (3) average of three-month, six-month and 12-month past returns. Do these metrics outperform past six-month return in a momentum strategy applied to the following nine sector exchange-traded funds (ETF) defined by the Select Sector Standard & Poor’s Depository Receipts (SPDR), all of which have trading data back to December 1998:

Materials Select Sector SPDR (XLB)
Energy Select Sector SPDR (XLE)
Financial Select Sector SPDR (XLF)
Industrial Select Sector SPDR (XLI)
Technology Select Sector SPDR (XLK)
Consumer Staples Select Sector SPDR (XLP)
Utilities Select Sector SPDR (XLU)
Health Care Select Sector SPDR (XLV)
Consumer Discretionary Select SPDR (XLY)

The three alternative strategies are, at the end of each month, allocate all funds to the sector ETF with the highest: (1) monthly Sharpe Ratio over the past six months (SR6-1); (2) monthly price slope over the past six months (Slope6-1); and, (3) average of past three-month, six-month and 12-month past total returns (3-1;6-1;12-1). For comparison, we include the strategy of monthly allocation to the sector ETF with the highest total return over the past six months (6-1). Using monthly dividend-adjusted closing prices for the nine sector ETFs over the period December 1998 through December 2014 (193 months), we find that: Keep Reading

In an effort to generate more responsive exchange-traded fund (ETF) momentum switching, a subscriber proposed a version of the “Simple Asset Class ETF Momentum Strategy” (SACEMS) that measures ETF returns from the lowest daily close within the momentum measurement interval rather than the monthly close at the beginning of the momentum measurement interval. To investigate, we run a competition between these alternative ways of measuring momentum as applied 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)

Specifically, the baseline strategy allocates all funds at the end of each month to the ETF or cash with the highest total return over the past five months (5-1). The alternative strategy allocates all funds at the end of each month to the ETF or cash with the highest return measured from its low during the last 105 trading days (about five months) to the end of the current month (Max 5-1). Using daily dividend-adjusted closing prices for the asset class proxies and the monthly yield for Cash during July 2002 (or inception if not available then) through December 2014 (150 months), we find that: Keep Reading

Does a cross-asset class, momentum-driven, simplified version of Modern Portfolio Theory (MPT) offer reliably strong performance over the long run? In their December 2014 paper entitled “A Century of Generalized Momentum; From Flexible Asset Allocations (FAA) to Elastic Asset Allocation (EAA)”, Wouter Keller and Adam Butler present an asset allocation strategy based on five concepts:

1. MPT is a sound framework for portfolio construction.
2. Momentum, a form of trend measurement, is a generally effective way to estimate key inputs to MPT: asset returns (R), return volatilities (V) and return correlations (C).
3. Crash protection based on excluding assets with negative past returns is a reasonable corollary of reliance on trends.
4. Tractability requires compromise to strict MPT, such as calculating return correlations relative to a single index (the equally weighted average returns of all assets).
5. Recognition of differences in import among inputs means weighting R, V and C inputs differently according to their elasticities (how much small changes in R, V and C affect the optimal portfolio weight for the asset).

The fifth concept is the innovation relative to the Flexible Asset Allocation (FAA) predecessor (see “Asset Allocation Combining Momentum, Volatility, Correlation and Crash Protection”), which weights expected R, V and C inputs based on a simple scoring system. The new Elastic Asset Allocation (EAA) strategy each month scores all assets in a universe by: (1) calculating expected R, V and C for each asset as geometrically weighted averages of past values; and, (2) weighting the expected values of R, V and C by their respective elasticities. For R, they use average total monthly excess (relative to the 13-week U.S. Treasury bill yield) returns over the last 1, 3, 6 and 12 months. For V and C, they use the last 12 monthly returns. To test the EAA strategy, they each month reform a long-only portfolio of the top-ranked assets weighted by their respective scores. They replace a fraction of the portfolio with 10-year U.S. Treasury notes (selected empirically as the best “cash” asset) according to the fraction of assets in the universe with non-positive excess returns. They apply a nominal one-way index switching friction of 0.1%. They consider three universes of 7, 15 and 38 asset classes. They emphasize Calmar ratio (focusing on drawdown) as a key optimization metric, but also consider Sharpe ratio. To mitigate data snooping, they optimize elasticity parameters during April 1914 through March 1964 and test it out-of-sample during April 1964 through August 2014. Using monthly returns for the three sets of financial asset indexes as available during April 1914 through August 2014, they find that:

Keep Reading