Objective research to aid investing decisions
Menu
Value Allocations for June 2019 (Final)
Cash TLT LQD SPY
Momentum Allocations for June 2019 (Final)
1st ETF 2nd ETF 3rd ETF

Investing Research Articles

Commodity Futures Strategies Over the Very Long Run

Do momentum (nearest contract 12-month excess return), value (spot price change from one year ago to five years ago) and basis (12-month average ratio of nearest to next-nearest contract prices) commodity futures premiums hold up over the very long run? In their February 2019 paper entitled “Two Centuries of Commodity Futures Premia: Momentum, Value and Basis”, Christopher Geczy and Mikhail Samonov measure momentum, value and basis premiums with a 141-year sample of commodity futures contract prices, focusing on a previously untested old subsample. Specifically, they each month for each premium categorize each contract series as high, middle or low. They then measure gross performances of long-short (equally weighted high minus low) and long-only (equally weighted high) portfolios for each premium. They further assess diversification benefits by comparing a stocks-bonds portfolio with stocks-bonds-commodity futures portfolios. Using 25,595 nearest contract month returns (averaging 15.2 commodities per month for the full sample, but only 7.1 per month for the old untested subsample through 1959), U.S. stock and bond market returns and U.S. Treasury bill (T-bill) yield as the risk-free rate during 1877 through 2017, they find that:

Keep Reading

Machine Learning Factor?

What are potential monthly returns and alphas from applying machine learning to pick stocks? In their February 2019 paper entitled “Machine Learning for Stock Selection”, Keywan Rasekhschaffe and Robert Jones summarize basic concepts of machine leaning and apply them to select stocks from U.S. and non-U.S. samples, focusing on the cross-section of returns (as in equity factor studies). To alleviate overfitting in an environment with low signal-to-noise ratios, they highlight use of: (1) data feature engineering, and (2) combining outputs from different machine learning algorithms and training sets. Feature engineering applies market/machine learning knowledge to select the forecast variable, algorithms likely to be effective, training sets likely to be informative, factors likely to be informative and factor standardization approach. Their example employs an initial 10-year training period and then walks forecasts forward monthly (as in most equity factor research) for each stock, as follows:

  • Employ 194 firm/stock input variables.
  • Use three rolling training sets (last 12 months, same calendar month last 10 years and bottom half of performance last 10 years), separately for U.S. and non-U.S. samples.
  • Apply four machine learning algorithms, generating 12 signals (three training sets times four algorithms) for each stock each month, plus a composite signal based on percentile rankings of the 12 signals.
  • Rank stocks into tenths (deciles) based on each signal, which forecasts probability of next-month outperformance/underperformance.
  • Form two hedge portfolios that are long the decile of stocks with the highest expected performance and short the decile with the lowest, one equal-weighted and one risk-weighted (inverse volatility over the past 100 trading days), with a 2-day lag between forecast and portfolio reformation to accommodate execution.
  • Calculate gross and net average excess (relative to U.S. Treasury bill yield) returns and 4-factor (market, size, book-to-market, momentum) alphas for the portfolios. To estimate net performance, they assume 0.3% round trip trading frictions. 

They consider two benchmark portfolios that pick long and short side using non-machine learning methods. Using a broad sample of small, medium and large stocks (average 5,907 per month) spanning 22 developed markets, and contemporaneous values for the 194 input variables, during January 1994 through December 2016, they find that: Keep Reading

Joint Fundamental and Technical Analysis

What kinds of fundamental and technical indicators play well together? In their August 2018 paper entitled “When Buffett Meets Bollinger: An Integrated Approach to Fundamental and Technical Analysis”, Zhaobo Zhu and Licheng Sun test performance of six stock portfolios that jointly exploit one of three popular fundamental indicators and one of two popular technical indicators, as follows:

  1. Piotroski’s FSCORE – each quarter long (short) stocks having high (low) scores summarizing a composite of accounting variables.
  2. Standardized unexpected earnings (SUE) – each quarter long (short) the fifth of stocks with the highest (lowest) earnings surprises.
  3. Return on equity (ROE) – each quarter long (short) the fifth of stocks with the highest (lowest) ROEs.
  4. Moving averages (MA) – each month long (short) stocks with 20-day MAs above (below) 125-day MAs at the end of the prior month.
  5. Bollinger bands (BOLL) – long (short) stocks below (above) one standard deviation of daily prices below (above) the average prices over the past 20 trading days.

Specifically, for each of six fundamental-technical pairs, they each month reform a portfolio that is long (short) stocks with both fundamental and technical buy (sell) signals. For risk adjustment, they employ widely used 5-factor (market, size, book-to-market, profitability, investment) alpha. Using accounting data and stock returns for a broad sample of U.S. common stocks priced at least $5, plus monthly factor returns, during January 1985 through December 2015, they find that:

Keep Reading

Do High-dividend Stock ETFs Beat the Market?

A subscriber asked about current evidence that high-dividend stocks outperform the market. To investigate, from a practical perspective, we compare performances of five high-dividend stock exchange-traded funds (ETFs) with relatively long histories to that of SPDR S&P 500 (SPY) as a proxy for the U.S. stock market. The five high-dividend stock ETFs are:

iShares Select Dividend (DVY), with inception November 2003.
PowerShares Dividend Achievers ETF (PFM), with inception September 2005.
SPDR S&P Dividend ETF (SDY), with inception November 2005.
WisdomTree Dividend ex-Financials ETF (DTN), with inception June 2006.
Vanguard High Dividend Yield ETF (VYM), with inception November 2006.

For each of these ETFs, we compare average monthly total (dividend-reinvested) return, standard deviation of total monthly returns, monthly reward-risk ratio (average monthly return divided by standard deviation), compound annual growth rate (CAGR) and maximum drawdown (MaxDD) to those for SPY over matched sample periods. We also look at alphas and betas for the five ETFs based on simple regressions of monthly returns versus SPY returns. Using monthly total returns for five high-dividend stock ETFs and SPY over available sample periods through February 2019, we find that:

Keep Reading

Weekly Summary of Research Findings: 3/18/19 – 3/22/19

Below is a weekly summary of our research findings for 3/18/19 through 3/22/19. These summaries give you a quick snapshot of our content the past week so that you can quickly decide what’s relevant to your investing needs.

Subscribers: To receive these weekly digests via email, click here to sign up for our mailing list. Keep Reading

SACEMS with Three Copies of Cash

Subscribers have questioned selecting assets with negative past returns within the “Simple Asset Class ETF Momentum Strategy” (SACEMS). Inclusion of Cash as one of the assets in the SACEMS universe of exchange-traded funds (ETF) prevents the SACEMS Top 1 portfolio from holding an asset with negative past returns. To test full dual momentum versions of SACEMS equally weighted (EW) Top 2 and EW Top 3 SACEMS portfolios, we add two more copies of Cash to the universe, thereby preventing both of them from holding assets with negative past returns. The SACEMS universe thus becomes:

PowerShares DB Commodity Index Tracking (DBC)
iShares MSCI Emerging Markets Index (EEM)
iShares MSCI EAFE Index (EFA)
SPDR Gold Shares (GLD)
iShares Russell 2000 Index (IWM)
SPDR S&P 500 (SPY)
iShares Barclays 20+ Year Treasury Bond (TLT)
Vanguard REIT ETF (VNQ)
3-month Treasury bills (Cash)
3-month Treasury bills (Cash)
3-month Treasury bills (Cash)

We focus on the effects of adding two copies of Cash on compound annual growth rates (CAGR) and maximum drawdowns (MaxDD) of SACEMS EW Top 2 and EW Top 3 portfolios. Using monthly dividend adjusted closing prices for the asset class proxies and the yield for Cash during February 2006 (the earliest all ETFs are available) through February 2019, we find that: Keep Reading

Measuring the Value Premium with Value and Growth ETFs

Do popular style-based exchange-traded funds (ETF) offer a reliable way to exploit the value premium? To investigate, we compare differences in returns (value-minus-growth, or V – G) for each of the following three matched pairs of value-growth ETFs:

  • iShares Russell 2000 (Smallcap) Growth Index (IWO)
  • iShares Russell 2000 (Smallcap) Value Index (IWN)
  • iShares Russell Midcap Growth Index (IWP)
  • iShares Russell Midcap Value Index (IWS)
  • iShares Russell 1000 (Largecap) Growth Index (IWF)
  • iShares Russell 1000 (Largecap) Value Index (IWD)

To aggregate, we define monthly value return as the equally weighted average monthly return of IWN, IWS and IWD and monthly growth return as the equally weighted average monthly return of IWO, IWP and IWF. Using monthly dividend-adjusted closing prices for these ETFs during August 2001 (limited by IWP and IWS) through February 2019, we find that: Keep Reading

ISM PMI and Future Junk Bond Returns?

A subscriber asked about the validity of the assertion in “The Daily Shot” of February 26, 2019 (The Wall Street Journal) that “recent weakness in the ISM [Institute for Supply Management] Manufacturing PMI [Purchasing Managers’ Index] index points to downside risks for high-yield debt.” Such a relationship might support a strategy of switching between high-yield bonds and cash, or high-yield bonds and U.S. Treasuries, based on PMI data. To investigate, we consider the following two pairs of funds:

  1. Vanguard High-Yield Corporate (VWEHX) and Vanguard Long-Term Treasury (VUSTX) since May 1986 (limited by VUSTX).
  2. iShares iBoxx High Yield Corp Bond (HYG) and iShares 7-10 Year Treasury Bond (IEF) since April 2007 (limited by HYG).

We consider both statistical tests and strategies that each month (per the PMI release frequency) holds high-yield bonds or cash, or high-yield bonds or Treasuries, according to whether the prior-month change in PMI is positive or negative. We use the 3-month U.S. Treasury bill (T-bill) yield as a proxy for return on cash. Using fund monthly total returns as available and monthly seasonally adjusted PMI data for January 1950 through January 2016 from the Federal Reserve Bank of St. Louis (discontinued and removed) and from press releases thereafter, all through February 2019, we find that: Keep Reading

Measuring the Size Effect with Capitalization-based ETFs

Do popular capitalization-based exchange-traded funds (ETF) offer a reliable way to exploit an equity size effect? To investigate, we compare the difference in returns (small minus big) between:

  • iShares Russell 2000 Index (Smallcap) Index (IWM), and
  • SPDR S&P 500 (SPY)

Using monthly dividend-adjusted closing prices for these ETFs during May 2000 (limited by IWM) through February 2019, we find that: Keep Reading

Consumer Inflation Expectations Predictive?

A subscriber noted and asked: “Michigan (at one point) claimed that the inflation expectations part of their survey of consumers was predictive. That was from a paper long ago. I wonder if it is still true.” To investigate, we relate “Expected Changes in Prices During the Next Year” (expected annual inflation) from the monthly final University of Michigan Survey of Consumers and actual U.S. inflation data based on the monthly non-seasonally adjusted consumer price index (U.S. All items, 1982-84=100). The University of Michigan releases final survey data near the end of the measured month, and the long-turn historical expected inflation series presents a 3-month simple moving average (SMA3) of monthly measurements. We consider two relationships:

  • Expected annual inflation versus one-year hence actual annual inflation.
  • Monthly change in expected annual inflation versus monthly change in actual annual inflation.

As a separate (investor-oriented) test, we relate monthly change in expected annual inflation to next-month total returns for SPDR S&P 500 (SPY) and iShares Barclays 20+ Year Treasury Bond (TLT). Using monthly survey/inflation data since March 1978 (limited by survey data) and monthly SPY and TLT total returns since July 2002 (limited by TLT), all through January 2019, we find that: Keep Reading

Daily Email Updates
Login
Research Categories
Recent Research
Popular Posts