The Case for Momentum in Expensive Markets

Charlie Bilello analyzed the relationship between market valuation and future returns (over various time horizons) in a recent post Valuation, Timing, and a Range of Outcomes. The post contained some very insightful tables, such as the one below, where he shows that valuations matter... if you pay less for stocks, you will generally be provided with higher returns (on average) over almost all time frames.

The Case for Momentum

In a previous post Valuations Do Matter (Even Over Shorter Time Frames) / Momentum Driven Valuation Timing, I highlighted a similar point and in addition took a look at how stocks performed at various valuation levels when 12-month returns had been positive or negative. The takeaway (highlighted in the table below) was that market returns were generally strong when stocks were:

• Cheap (with positive or negative momentum)
• Expensive with positive momentum

To bring this full circle, the tables below replicate Charlie's analysis, but also compares those results with the average forward return for a momentum strategy with the following rules (note the returns in the table below differ slightly from Charlie's - not sure what data he used, but I used data from Ibbotson's):

• If the CAPE of the S&P 500 was within the bottom 50th percentile (which happens to be less than 17x), allocate to stocks; otherwise...
• If the S&P composite had a one year backward looking return that was positive, allocate to stocks 
• If the S&P composite had a one year backward looking return that was negative, allocate to cash (t-bills)

Click the table for a larger view

The case for trend following when markets are expensive becomes abundantly clear when viewed in chart form... when markets are cheap, an allocation to stocks resulted in returns that were on average exactly the same as a buy-and-hold strategy over the short-term and returns that were broadly in-line with those of a buy-and-hold over longer periods. When markets were stretched, momentum protected an investor from severe drawdowns over shorter periods and allowed returns to on average compound over longer periods.

At a current CAPE in the U.S. approaching a level that would put it in the top 10% of most expensive levels over this time frame, a momentum strategy may prove to be a good mechanism for investors to time exposure to an expensive U.S. market.

The Brutal Math of a 60/40 Portfolio

Think only a bear market can keep returns of a 60/40 near 0%... think again.

Given the huge opportunity cost of allocating to cash or bonds at current yield levels, even generally optimistic return assumptions for stocks are enough to keep portfolio level returns near 0% real. The goal of this post is to set the stage for a future post where I hope to share potential solutions that may improve potential returns with a similar risk profile as a traditional 60/40 and to set proper expectations of what a 60/40 allocation dragged down by low yields may provide.

After-tax real return forecasts (see below for the formula used in the calculation)

Stocks

• Let's say you assume stocks will return 6% nominal going forward. 
• After tax returns (assuming gains are taxed at the more favorable 20% capital gains tax rate) = 4.8% (6% x [1 - 20%]) 
• After tax after inflation returns assuming a forecasted 2% inflation rate = ~2.8% (4.8% - 2.0%) 

Bonds 

• Bonds will generally (best case scenario) return their yield (current yield to worst of the Barclays Aggregate Bond index = 2.0%) 
• After tax returns assuming the less favorable rate applied to coupons (and a 35% tax rate) = 1.4% (2% x [1 - 35%]) 
• After tax after inflation returns assuming the forecasted 2% inflation rate = ~-0.6% (1.4% - 2.0%) 

60/40 

• 60% Stocks = 2.8% x 60% = 1.68% contribution 
• 40% Bonds = -0.6% x 40% = -0.24% contribution 
• Total return = 1.45% real 

Throw on the ~1% fees many financial advisors charge and/or the lower yields many investors are accepting by taking less duration risk / diversifying U.S. bond exposure to even lower yields abroad and an investor may break through the 0% threshold even with a 6% stock forecast. This coming from an allocation that has a historical standard deviation of roughly 10% over time.


Initial takeaways 

The math above outlines the importance of:

• Shielding returns from taxes whenever possible 
• Keeping fees as low as possible (or ensuring you get something for your fees)
• Seeking alternative sources of return (whether through allocation or alternative asset classes that now have a very low hurdle rate relative to bonds to be included)
• Minimizing an allocation to negative real return asset classes