Strategic Allocation

What factors make asset class diversification work? To investigate empirically, we consider the following mix of exchange-traded funds (ETF) as asset class proxies (the same used in “Simple Asset Class ETF Momentum Strategy”):

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)

We calculate the monthly gross return-risk ratio (average monthly return divided by standard deviation of monthly returns) for an equally weighted, monthly rebalanced portfolio of all nine asset class proxies. We then recalculate the return-risk ratio nine times, each time excluding one of the assets, and relate the resulting return-risk ratios to three characteristics of the respectively excluded assets: (1) average monthly return; (2) standard deviation of monthly returns; and, (3) average (pairwise) cross-correlation of monthly returns with the other eight assets. The objective is to determine whether any of these three characteristics explain asset contribution to diversification benefit.  We ignore trading frictions associated with monthly rebalancing, which would be similar for all combinations. Using dividend-adjusted monthly returns for the above nine asset class proxies during September 2006 (so that monthly returns for all assets are available in equal-weight calculations) through April 2014 (92 monthly returns), we find that: Keep Reading

How well do long-short stock strategies work, after accounting for all costs? In their February 2014 paper entitled “Assessing the Cost of Accounting-Based Long-Short Trades: Should You Invest a Billion Dollars in an Academic Strategy?”, William Beaver, Maureen McNichols and Richard Price examine the net attractiveness of several long-short strategies as stand-alone investments (as for a hedge fund) and as diversifiers of the market portfolio. They also consider long-only versions of these strategies. Specifically, they consider five anomalies exposed by the extreme tenths (deciles) of stocks sorted by:

1. Book-to-Market ratio (BM) measured annually.
2. Operating Cash Flow (CF) measured annually as a percentage of average assets.
3. Accruals (AC) measured annually as earnings minus cash flow as a percentage of average assets.
4. Unexpected Earnings (UE) measured as year-over-year percentage change in quarterly earnings.
5. Change in Net Operating Assets (ΔNOA) measured annually as a percentage of average assets.

For strategies other than UE, they reform strategy portfolios (long the “good” decile and short the “bad” decile) annually at the end of April using accounting data from the prior fiscal year. For UE, they reform the portfolio at the ends of March, June, September and December using prior-quarter data. They highlight cost of capital, financing costs and rebates received on short positions, downside risk and short-side contribution to performance. They assume that the same amount of capital supports either a long-only portfolio, or a portfolio with equal long and short sides (with the long side satisfying Federal Reserve Regulation T collateral requirements for the short side). They account for shorting costs as fees for initiating short positions plus an ongoing collateral rate set at least as high as the federal funds rate, offset by a rebate of 0.25% per year interest on short sale proceeds. They estimate stock trading costs as the stock-by-stock percentage bid-ask spread. They consider two samples (including delistings): (1) all U.S. listed stocks; and, (2) the 20% of stocks with the largest market capitalizations. Using accounting data as described above for all non-ADR firms listed on NYSE, AMEX and NASDAQ for fiscal years 1992 through 2011, and associated monthly stock returns during May 1993 through April 2013, they find that: Keep Reading

Under what conditions is periodic rebalancing a successful “volatility harvesting” strategy? In his February 2014 paper entitled “Disentangling Rebalancing Return”, Winfried Hallerbach analyzes the return from periodic portfolio rebalancing by decomposing its effects into a volatility return and a dispersion discount. He defines:

• Rebalancing return as the difference in (geometric) growth rates between periodically rebalanced and buy-and-hold portfolios.
• Volatility return as the difference in growth rates between a periodically rebalanced portfolio and the equally weighted average growth rate of its component assets.
• Dispersion discount as the difference in growth rates between a buy-and-hold portfolio and the equally weighted average growth rate of portfolio assets.

Based on mathematical derivations with some approximations, he concludes that: Keep Reading

What drives the performance of risk parity asset allocation, and on what asset classes does it therefore work best? In their January 2014 paper entitled “Inter-Temporal Risk Parity: A Constant Volatility Framework for Equities and Other Asset Classes”, Romain Perchet, Raul Leote de Carvalho, Thomas Heckel and Pierre Moulin employ simulations and backtests to explore the conditions/asset classes for which a periodically rebalanced risk parity asset allocation enhances portfolio performance. Specifically, they examine contemporaneous interactions between risk parity performance and each of return-volatility relationship, return volatility clustering and fatness of return distribution tails (kurtosis). They then compare different ways of predicting volatility for risk parity implementation. Finally, they backtest volatility prediction/risk parity allocation effectiveness separately for stock, commodity, high-yield corporate bond, investment-grade corporate bond and government bond indexes (each versus the risk-free asset). They optimize volatility prediction model parameters annually based on an expanding window of historical data. They forecast volatility based on one-year rolling historical daily return data. Using daily total returns in U.S. dollars for the S&P 500 Index during 1980 through 2012 and for the Russell 1000, MSCI Emerging Market, S&P Commodities, U.S. High Yield Bond, U.S. Corporate Investment Grade Bond and U.S. 10-Year Government Bond indexes and the 3-month U.S. Dollar LIBOR rate (as the risk-free rate) during January 1988 through December 2012, they find that: Keep Reading

Is there a tractable way to combine momentum investing with Modern Portfolio Theory (MPT)? In their December 2013 paper entitled “Tactical MPT and Momentum: the Modern Asset Allocation (MAA)”, Wouter Keller and Hugo van Putten present a tactical, simplified, long-only version of MPT that applies momentum to estimate future asset returns. Specifically, they:

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

They reform the MAA portfolio monthly at the first close. Their baseline historical interval for estimation of all variables is four months (84 trading days). Their baseline shrinkage factor for all variables is 50%. Their benchmark is the equally weighted (EW) “market” of all assets, rebalanced monthly. They assume a one-way trading friction of 0.1%. They consider a range of portfolio performance metrics: annualized return, annual volatility, maximum drawdown, Sharpe ratio, Omega ratio and Calmar ratio. Using daily dividend-adjusted prices for assets allocated to nine universes (of seven to 130 assets, generally consisting of asset class proxy funds) during November 1997 through mid-November 2013, they find that: Keep Reading

Is there a simple compromise between easy-to-implement market weights and more diversified equal sector and equal stock weights? In their December 2013 paper entitled “A Simple Diversified Portfolio Strategy”, Bernd Hanke and Garrett Quigley present a stock portfolio construction approach that blends market weights, equal stock weights and equal sector weights. The objectives of the approach (relative to market weights) are: (1) higher returns (by capturing more of the diversification premium); (2) lower risk (via increased diversification); and, (3) competitive capacity and rebalancing frictions (by limiting the tilt toward small, illiquid stocks). In testing this approach, they form and rebalance annually regional (U.S., European and Japanese) portfolios of relatively liquid stocks. They ignore rebalancing frictions. They define sectors via the broadest Global Industry Classification Standard level (ten sectors). Using total (dividend-reinvested) returns, market capitalizations and sector memberships for a broad sample of relatively liquid stocks during January 1992 through March 2013, they find that: Keep Reading

Are there easily implementable life cycle investing strategies reliably superior to the conventional glidepath from equities toward bonds? In their June 2013 paper entitled “The Glidepath Illusion… and Potential Solutions”, flagged by a subscriber, Robert Arnott, Katrina Sherrerd and Lillian Wu summarize flaws in the conventional glidepath approach and explore simple alternatives that address some of the flaws. Specifically, they compare the follow six strategies:

1. 80–>20: the conventional linear glidepath from 80% stocks-20% bonds to 20% stocks-80% bonds at retirement, with market capitalization weighting.
2. 20–>80: inverse of the conventional linear glidepath.
3. 50-50: constant 50% stocks-50% bonds, with market capitalization weighting.
4. Dynamic Bond Duration: the 50-50 strategy, but: (a) hold 20-year bonds for the first 21 years; (b) shift linearly to 10-year bonds during the next ten years; and, (c) shift linearly from 10-year bonds to T-bills during the last 10 years before retirement.
5. Dynamic Value/Low Beta: the 50-50 strategy, but: (a) stocks are weighted by book value for the first 21 years (from the 1,000 U.S. stocks with the highest book value); and, (b) shift linearly to low-volatility stocks (the 1,000 largest U.S. companies by market capitalization, weighted by inverse volatility) during the next 20 years.
6. Dynamic Combined: the 50-50 strategy, but use Dynamic Bond Duration and Dynamic Value/Low Beta for bonds and stocks, respectively.

Comparison tests assume that: (1) each individual makes inflation-adjusted $1,000 annual contributions to a retirement portfolio over a 41-year career; and, (2) portfolio rebalancing is annual, frictionless and tax-free. Using simulations based on long-term samples of U.S. stock index, bond index and U.S. Treasury bill (T-bill) returns through the end of 2011, they find that: Keep Reading

Is diversification across stock and bond factors superior to diversification across asset classes? In their August 2013 report entitled “Investing in Systematic Factor Premiums”, Kees Koedijk, Alfred Slager and Philip Stork measure the gross performances of widely used stock and bond factors and pit portfolios diversified across those factors against portfolios diversified across asset classes. For equities, they examine market, size, value, momentum and low-volatility factors. For bonds, they examine market, term spread, credit spread, high-yield, short-term credit yield and short-term government yield factors. They consider both U.S. and European data as available. They take an institutional perspective and therefore restrict consideration to simple, long-only portfolios. For asset class diversification, they consider stocks-bonds and stocks-bonds-commodities-real estate. They ignore all trading frictions involved in constructing factor portfolios and in rebalancing multi-asset and multi-factor portfolios. Using monthly prices for U.S. and European stocks, bonds, Real Estate Investment Trust (REIT) indexes and a common global commodity index as available through mid-to-late 2012, they find that: Keep Reading

What is the global financial asset allocation? In their November 2013 paper entitled “The Global Multi-Asset Market Portfolio 1959-2012”, Ronald Doeswijk, Trevin Lam and Laurens Swinkels construct the aggregate portfolio of all investors encompassing market capitalizations for eight asset classes: equities, private equity, real estate, high-yield bonds, emerging markets debt, investment-grade credits (corporate bonds and mortgage-backed securities), government bonds and inflation-linked bonds. They exclude human capital (earned income streams), durable goods (such as cars), residences and family businesses. They exclude commodities because the net position in commodity futures is zero. They suggest that these aggregate allocations represent a natural benchmark portfolio for financial investors. Further, they trace the evolution of allocations to the eight asset classes during 1990 through 2012, and the evolution of allocations to equities, real estate, non-government bonds and government bonds during 1959 through 2012. Using a variety of data sources and estimation methodologies, they find that: Keep Reading

Are implied volatility futures good diversifiers of underlying indexes? Do implied volatility futures for different indexes represent a reliable pair trading opportunity? In their November 2013 paper entitled “Investment Strategies with VIX and VSTOXX Futures”, Silvia Stanescu and Radu Tunaru update the case for hedging conventional stock and stock-bond portfolios with near-term implied volatility futures for the S&P 500 Index (VIX) and the Euro STOXX 50 Index (VSTOXX). For this analysis, they use data for the U.S. and European stock market indexes, associated implied volatility futures and U.S. and European aggregate bond indexes from March 2004 for U.S. assets (VIX futures inception) and from May 2009 for European assets (VSTOXX futures inception), both through February 2012. They also investigate a statistical arbitrage (pair trading) strategy exploiting a regression-based prediction of the trend in the gap between VIX and VSTOXX during the last six months of 2012. Using daily data for the specified indexes and implied volatility futures contracts, they find that: Keep Reading