ETFReplay Portfolio for May

The Portfolio holdings have been updated for May 2014.  I previously detailed here and here how an investor can use to screen for best performing ETFs based on momentum and volatility.

The portfolio begins with a static basket of 15 ETFs. These 15 ETFs are ranked by 6 month total returns (weighted 40%), 3 month total returns (weighted 30%), and 3 month price volatility (weighted 30%). The top 4 are purchased  at the beginning of each month. When a holding drops out of the top 5 ETFs it will be sold and replaced with the next highest ranked ETF.

Start Algo Trading Here

In addition, ETFs must be ranked above the cash ETF SHY in order to be included in the portfolio, similar to the absolute momentum strategy I profiled here. This modification could help reduce drawdowns during periods of high volatility and/or negative market conditions (see 2008-2009), but it could also reduce total returns by allocating to cash in lieu of an asset class.

The top 5 ranked ETFs based on the 6/3/3 system as off 4/30/14 are below:

DBA PowerShares DB Agricultural Commodities
VTI Vanguard Total U.S. Stock Market
LQD iShares iBoxx Invest Grade Bond
PCY PowerShares Emerging Mkts Bond

The portfolio maintains positions in DBA, VNQ, and LQD for the month of May. HYG was sold for a gain of 0.97% after purchasing it originally on 10/31/13. The proceeds were used to purchase Vanguard Total U.S. Stock Market (VTI).

Beginning in 2014 we track both the 6/3/3 strategy (same system as 2013) as well as the pure momentum system, which will rank the same basket of 15 ETFs based solely on 6 month price momentum. There is no cash filter in the pure momentum system, volatility ranking, or requirement to limit turnover – the top 4 ETFs based on price momentum will be purchased each month. The portfolio and rankings will be posted on the same spreadsheet as the 6/3/3 strategy.

The top 5 six month momentum ETFs are below:

6 month Momentum
DBA PowerShares DB Agricultural Commodities
VTI Vanguard Total U.S. Stock Market
TLT iShares Barclays Long-Term Trsry
LQD iShares iBoxx Invest Grade Bond

For May the portfolio maintains positions in VTI, VNQ and DBA. EFA was sold for a gain of 1.82% and purchase date of 12/31/13. Proceeds were used to purchase iShares Barclays Long-Term Treasury ETF (TLT).

More on this topic (What's this?)
Core ETF Report
US Market ETF Trading Map
Read more on Exchange Traded Fund (ETF) at Wikinvest

Hedging the Graham Value Stock Portfolio

A reader asked if hedging helped reduce drawdowns in the Graham Value Stock portfolio. The portfolio, like many long strategies, suffered significant drawdowns in 2008.

I used Portfolio123 to back-test two modifications to the strategy.  All tests re-rank stocks every 3-months, selling those that no longer qualify and replacing them with the highest rated stocks, assume .50% slippage, and cover the time period of 1/2/99 – 4/27/14.

First, let’s review the baseline backtest for the un-hedged version of the stratgegy.  The test below has different re-balance dates than the portfolio update from mid-April and therefore slightly different results. However, it serves as a consistent comparison for the two hedged portfolios:


Next, we sell all stocks in the Graham Value strategy and goes to cash when the 50 day moving average of the SPDR S&P 500 ETF (SPY) crosses below its 200 day moving average. Stocks are re-purchased when the 50 day moving average of SPY is above its 200 day moving average on the re-rank date. This hedge “re-ranks” on the same 3-month frequency as the stocks in the strategy. Thus, it could conceivably keep an investor in cash for 3 months. Interim signals (50 day SMA crossing over/under 200 day SMA) are ignored and only honored on the re-rank date:


The last strategy uses a 15% entry-based stop loss. Stocks are sold when they trade at or below -15% of their purchase price. The resulting proceeds are held as cash until the next re-rank date.  Thus, the portfolio could hold significant amounts of cash if several stocks are stopped out:


There are a number of variations which could be applied to hedge the portfolio. The two listed above are not intended as recommendations or optimal, but as a demonstration of the material impact risk-management can have on an investment strategy. In both cases drawdowns were reduced, and in the crossover strategy total returns and sharpe ratio were also negatively impacted. The stop-loss strategy had slightly lower total returns but a higher sharpe ratio.

More on this topic (What's this?)
Chris Hedges: “America is a Tinderbox”
New ETF Hedges Foreign Exposure
Read more on Value Investing, Hedging at Wikinvest

Weekend Investment Readings

Below is my weekend reading list:

Stop Sabotaging Your Investments – MarketWatch

Software is Eating Investment Management & Active vs Passive – Abnormal Returns

Do Commodities Belong in Your Allocation? Advisor Perspectives

Dare to be Great II (pdf) John Mauldin

Larry Swedroe ( Fixed Income’s Low-Risk Anomaly, Looking at Value Premiums Part 1 and 2.

4 Keys to Picking the Right ETF –

Still Believe in Efficient Markets? Pension Partners

Silver Forecast – The Gold and Oil Guy


Graham Value Stock Portfolio Update

In January 2012 I announced a new portfolio, a Benjamin Graham “inspired” value stock portfolio.  The purpose of the hypothetical portfolio is to track returns for a portfolio of 15 stocks selected based on a variety of valuation metrics. I originally intended to update the portfolio monthly; however, in the spirit of creating a lower turnover, value-driven portfolio it is now updated quarterly.

I have also added an additional criteria to limit turnover in the portfolio (see below). The Graham portfolio is an attempt to add a value strategy to Scott’s Investments, which is otherwise focused on momentum, trend, income and market timing strategies.

The criteria used to select the stocks are listed below.  The tool used to perform the screen and backtests are courtesy of  Portfolio123 (“P123″).

The actual screen factors are below:

  • Liquidity filter: No OTC Stocks
  • Market capitalization > $100 million
  • Eliminate companies classified in the Miscellaneous Financial Services Industry, most of which are investment companies and funds and not the kind of stocks this all-star tended to seek
  • Current ratio must be at least 1.5
  • Long-term debt must be no higher than 10% above working capital
  • EPS must be above breakeven in each of the last four quarters and in each of the last five annual periods
  • Trailing 12 month EPS most be above EPS in the latest annual period
  • EPS in the latest annual period must be above EPS in the prior year and five years ago
  • The company must have paid common dividends in the last 12 months

The ranking system used as a basis for selecting the top 15 based among those stocks that pass the Graham screen are below:

  • Valuation – 60% of total
  • Trailing 12 month P/E (15% of this category)
  • Price-to-Book (15% of this category)
  • Price-to-Tangible Book Value (35% of this category)
  • Operating P/E, defined as Market Capitalization divided by Business Income, which is Sales minus Cost of Goods sold minus Selling, General & Administrative Expense and omits unusual items (35% of this category)
  • Earnings – 40% of total
  • 5-year EPS Growth Rate (50% of this category)
  • EPS Stability, defined as the standard deviation of EPS over the past 16 quarters, lower being better (50% of this category)

Stocks will now be sold when they drop below the 75th percentile ranking based on the ranking system above. Improvements in the screening and testing platform (via Portfolio123) allows a change in the sell/turnover rule from previous updates.

I began tracking this portfolio real-time on January 13th, 2012. As of April 15th, 2014 it is up over 43%. A real-world application of this portfolio could also utilize stop losses in order to prevent large drawdowns in single positions. However, for the purposes of tracking the portfolio results, all positions are bought and held until rebalancing.

Below is a 14+ year backtest results for this screen  using a quarterly rebalance and .50% slippage to help account for bid/ask spreads. Backtests include the 75th percentile sell rule (stocks will only be sold when they drop below the 75th percentile ranking):

(test data courtesy of Portfolio123)