A Value Strategy That Doubles Down When Stocks Get Cheaper

Value investors get compensated for taking contrarian bets, particularly when it is agonizing to do so. Sam Zell, known as "the grave dancer" for his penchant for buying companies as they emerged from bankruptcy, sold his Equity Office Properties Trust near the peak of the real estate market in 2007. In the depths of the financial crisis, Warren Buffett took a large stake in Goldman Sachs, which was desperate for capital at the time.

It is very difficult to sell stable companies and buy unstable companies during periods of uncertainty. Index funds are supposed to take emotion out of the equation and remove our behavioral biases. But traditional value indexes don't do a great job of replicating the behavior of contrarian investors. As a result of its market-cap-weighting approach, the Russell 1000 Value Index actually increased its exposure to relatively stable companies such as  Exelon Corp (EXC),  Intel (INTC),  Bristol-Myers Squibb (BMY), and  CVS Health Corp (CVS) in the wake of the 2008 financial crisis. At the same time, it decreased exposure to the most beaten-down value stocks, such as  General Electric (GE) and  Citigroup (C). 

Market-cap weighting can prevent traditional value indexes from behaving like a contrarian value manager. When a value stock appreciates in price, it will become a large part of the value index even if its price appreciation means that it is less of a bargain. However, a non-market-cap-weighted value index is not forced to increase its weight in such stocks.  

For instance, the FTSE RAFI US 1000 Index, which  PowerShares FTSE RAFI US 1000 ETF (PRF) tracks, weights its holdings according to book value, sales, cash flows, and dividends, rather than market capitalization. This unique approach caused the fund to cut its exposure to companies that weathered the 2008 market downturn better than their peers such as  Wal-Mart (WMT),  Exxon Mobil (XOM),  Verizon (VZ),  Johnson & Johnson (JNJ),  Procter & Gamble (PG), and  Pfizer (PFE). It used the proceeds to fund a dramatic increase in exposure to beaten-down stocks such as  Bank of America (BAC), Citigroup, General Electric,  Ford (F),  Dow Chemical (DOW),  Alcoa , and  Altria Group (MO).

To illustrate how a market-cap index works compared with a fundamentally weighted index, consider the following indexes based on two companies. Company A has a market cap of $120 and a book value of $100 for a book/price ratio of 0.83. Company B has a market cap of $60 and a book value of $80 for a book/price ratio of 1.33. Company B exhibits stronger value attributes as it sells at a higher book/price than Company A. Yet, it represents a smaller portion of the market-cap-weighted index than the book value weighted index, as the table below illustrates. Suppose that as we move to period 2, Company A's price has appreciated to $140, but its book value is unchanged. Company A's book/price is now 0.71 and it now exhibits weaker value characteristics. However, it is now a bigger part of the market-cap-weighted index because its price has appreciated. The market-cap-weighted index now has a less attractive book/price overall. Also in period 2, Company B's book value increased from $80 to $90, but its price is unchanged. It is now a bigger part of the book value weighted index and that index will have a lower value overall.

Source: Morningstar Analysts.

While most fundamentally weighted indexes exhibit a similar value tilt to traditional value indexes, their dynamic rebalancing approach may allow them to more effectively profit from mean reversion in valuations. To compare the value characteristics of the FTSE RAFI US 1000 Index with those of the Russell 1000 Value Index, I regressed their monthly returns against a multifactor model that included a value factor based on data from Andrea Frazzini's Data Library. The value factor is defined as a portfolio long a basket of stocks with high book/price values and short a basket of expensive stocks with low book/price values. The more sensitive the portfolio's returns are to the value factor, the deeper its value tilt. In the table below, I display only the value factor, even though this was a multifactor regression.

During the entire period (September 2003 to August 2014), the FTSE RAFI index had a value beta of 0.35, only slightly less than the 0.40 for the Russell 1000 Value Index. This means that it tended to increase in value by 0.35% for each 1 percentage point that deep-value stocks outperformed expensive stocks included in the value factor portfolio, holding everything else constant. This represents the average beta during the entire period. However, if we break it up into two periods, before and after the financial crisis, we see substantial variation in the value beta. The value beta for the FTSE RAFI index increased from 0.30 in the first period to 0.44 in the second. Meanwhile, the value beta for the Russell 1000 Value Index fell in the second period. Yet value stocks performed much better in the second period, as stocks in the long-short value stock basket returned 9.03% in the second period compared with negative 3.79% in the first period. Value stocks performed better, in part, because they were selling at a larger discount to the market. So the FTSE RAFI index had a higher sensitivity to value stocks just as value stocks were performing better. This may be because its alternative weighting approach causes it to bet big when valuations are most attractive.

Sources: Morningstar Analysts, Andrea Frazzini's Data Library.

  Sources: Morningstar Analysts, Andrea Frazzini's Data Library.

A value investor should be willing to take bigger value bets when stocks become cheaper. Certain strategic beta funds may be better able to accomplish this than traditional market-cap-weighted funds. Some of the ETFs that use non-market-cap weighting include PRF, Guggenheim S&P 500 Equal Weight ETF (RSP), WisdomTree LargeCap Dividend ETF (DLN), and Guggenheim S&P 500 Pure Value ETF (RPV).