Description

Recommendation: if you can’t find anything to buy and volatility matters to you, hold some cash.

“More money has been lost by investors preparing for corrections or trying to anticipate corrections than has been lost in the corrections themselves.” —Peter Lynch

If I could, I’d tattoo that on my computer screen. Now more than ever, “Lynch’s Law” is relevant. With a Shiller P/E of 27, the U.S. stock market is one of the three most expensive times since 1880. I don’t know a single professional investor who isn’t scared, if not secretly panicked, right now. To hedge or not? That question has caused a huge rift in the investment world. On one side, you have those who believe that the market is sky-high and you should protect yourself. And on the other side, you have those who believe you should always stay invested because timing the market is best left to people who read things like “Zimbabwe Copper: Is it the only safe haven?”

Complicating everything is the fact that absolutely everything seems to be in a bubble. There’s not only a bull market in stocks, there’s even a bull market in fear. If you’re holding the S&P 500, the cost to buy puts and insure your portfolio against all losses is 8% a year. 8%! So spend 8% of your wealth, and you’re assured of keeping the remaining 92% of it. And if you want to roll your hedge to the next year, that’ll be another 8%. With much respect to Nassim Taleb, tail-risk insurance isn’t for pessimists, or for value investors. It’s for the kind of optimist who goes to Las Vegas, loses everything and winds up living there.

So if buying portfolio insurance is out, what exactly is in? We think cash is highly underrated. We’ll examine its performance over the past as well the performance of shorting stocks to see if either can escape the gravitational pull of Lynch’s Law.

Is hedging a form of “timing” the Market?

As a threshold matter, we’d like to address the idea that hedging or holding cash is always an act of market timing. The traditional view in investing is that staying 100% invested beats “timing” the market, a view we largely agree with. Charts such as the one below seem to make the point in no uncertain terms.

Miss the best 30 days of the market from 1994 to 2013, and you’ll be dancing in the streets for spare change. However, while this chart may be true, we don’t believe it is the entire truth. The chart doesn’t take into account the absolute valuation of capital assets. What we’ve found is that absolute valuation correlates about -30% to -50% with future returns for value stocks. In other words, the higher the valuation, the lower the return. But the story doesn’t end there, and this is the significant part: we also found that at the extreme highs and the extreme lows of asset valuations, the correlation increases to -70%.

Now most people focus on the valuation of the market itself. We think that’s completely irrelevant to value investors because value investors don’t hold a market portfolio. They hold value stocks and thus should only be concerned with the valuation of value stocks. That’s our focus here: hedging only when value stocks get expensive.

Shorting stocks

We took a look at what shorting adds to a portfolio during crises using data from Ken French. (Note: we originally backtested it using our own data from Compustat, but if you really want to see if someone has robust results, you should have them recreate it on SOMEONE ELSE’S DATA.)

Using returns for the cheapest 10% of the market by P/E (“the value decile”), we compare two portfolios. One is long-only and simply holds all the stocks in the value decile. The other shorts high Price/book, low profitability “glamour” stocks. This short selling is triggered only when the Price/book of the value decile ranks above the 95^th percentile (in other words, when it’s more expensive than it has been 95% of the time). Shorting is done to reduce the net long position of the portfolio to zero and is continued until the Price/book of the value decile falls below the long-term average valuation of the value decile.

Side note: notice that the sum of the negative years for the value strategy is higher than it is for the S&P 500. No wonder people have such a hard time sticking with value investing! It’s not just emotionally painful over the short-term. It’s quantitatively painful too.

As you can see on the graph, shorting adds value twice, once in 1969 and another time 2008 when it cuts your drawdown from -40% from -18%. The Sortino ratio is basically a more intelligent Sharpe ratio. It measures return relative to downside volatility, and you can see the long-short portfolio knocks it out of the park on that front. This is nothing short of wonderful.

There’s a problem, however, and it’s the same problem all backtests have. We’ve cheated. We’ve used data from the 2000s to determine the rule that governs how we’d manage our portfolio in the 1960s. That information wouldn’t have been available to us, so we need to do something called “hold-out.” We need to divide the data into a learning data set and a testing data set. We use the learning data to derive our rule for when to begin shorting, and we use the testing data to see if the rule works.

In our case, let’s make our learning data set the ten-year period from 1964-1973. 1969 was a peak for the valuations of value stocks, and some of you may recall that it was the year that Buffett closed his partnership because he felt there wasn’t anything worth buying. (He made the right call, by the way. The value decile fell 25% later that year before getting halved along with the rest of the market in 1973 and 1974.)

 Source: Ken French data library.

1969 was a peak for value stock valuations. It also ushered in a new era of valuation. As you can see, the valuation of value stocks went up in the late sixties and never reverted the pre-60s mean.

1969 was the previous peak. Now that honor goes to today. That’s right, the absolute valuation of value stocks is the highest since 1926. This is terrible news, but then the valuation for growth stocks is much, much worse. Growth stocks are back in tech bubble territory, both in terms of absolute valuation and valuation relative to the rest of the market (premium, chart on left).

To return our backtest, when we look at the price/book of the value decile  between 1964 and 1973, we find that it’s lower on average than it is between the full period we examined earlier: 1964-2014. The price/book at the 95^th percentile is .68 instead of .74. So for our “hold-out” test, we are going to start shorting glamour stocks when the price/book of the value decile hits .68 instead of waiting for it to reach .74. This is a sub-optimal rule, but we have to use it. It’s how we put ourselves in the position of an investor who has “learned” from the 1964-1973 period and is now applying the rules gleaned from it to the future. This is also important as a test of robustness: can we use the wrong rule and still get a decent outcome?

The results:

Now we see a different story. Shorting not only reduces your reduces your returns, it also adds volatility. That -44%, by the way, doesn’t come from the financial crisis. It comes from 1998, near the peak of the tech bubble. And you know what? If you’d been following this system, you still would have gotten lucky. The next year, value stocks fell in valuation, and so the portfolio happened not to be shorting at all, a good thing because glamour stocks were up 142%, and shorting them would have blown up your entire portfolio.

Of course we’ve been assuming you equal-weight the glamour stocks. Here’s what would have happened if you market-cap-weighted them.

Fortune ceases to smile upon you. In 1998, your portfolio would have blown up. Being long value and short glamour stocks is dangerous. It’s like being short crystal meth and long math homework. The harm of one consumes you before you get the benefit of the other.

These aren't volatile little micro-caps or small-caps either. Average firm size in 1998 was $3 billion. Today it's $10 billion. Our recommendation: heed the ancient warning that markets can stay irrational longer than you can stay solvent.

Which brings us, by default, to the final alternative: holding cash.

Cash

Cash is an utterly unsatisfying asset. It doesn’t do anything. It isn’t even a productive asset, much less an appreciating one. After inflation, it has negative returns. It’s like a girl that sits home alone, writing morbid poetry. Essentially it’s just a sad, little melting ice cube. Many famous investors have heaped scorn on holding cash, and understandably so. Retail investors who go in and out of cash invariably screw everything up. There was a famous study showing that investors who found the needle in a haystack and invested with Ken Heebner’s mutual fund during its incredible ten-year run STILL managed to generate negative returns by trading in and out of it. You can’t make this stuff up.

We went into this thinking cash would get killed. Below is the first backtest. We did the same thing we did with the long-short portfolio, except when the price/book of value stocks exceeds the 95^th percentile rank, we go 100% to cash instead of shorting.

These results, frankly, were surprising to us. We never would have imagined cash would outperform both long-short and long-only in terms of both return and volatility reduction. But again this backtest cheats by relying on information from the future to predict the past, so let’s get to the hold-out.

These results are pretty amazing. The reduction in volatility is incredible. There are no down years from 1995 to today. The Sortino ratio is through the roof, and the sum of the negative years is half that of the long-only portfolio. If you’re willing to sacrifice a little bit of return, you might be able to significantly reduce volatility for your investors.

But at the same time, we see here that holding cash and paying attention to absolute valuations is not a cure-all. Holding cash, even when value stocks are in bubble territory, hurts returns to the tune of 1.4% a year. Of course this assumes markets behave as they have for the last 50 years, which is to say they remain highly mean-reverting. In the event of a 1929 crash, where the market goes down and stays down, we’re not dealing with mere volatility but with a permanent loss of capital. In that scenario, cash would serve you incredibly. Also remember: value stock valuations, like valuations for the rest of the market, have gone up and stayed up since 1969. It’s not inconceivable at all that now that the great debt cycle has played out and our working-age population has stopped growing that valuations revert down to their long-term averages.

Conclusion: holding cash can add a huge amount of value to managers who care about volatility and/or a permanent loss of capital.

One caveat: the long-cash portfolio holds cash and no productive assets for years at a time. For instance, it would have been 100% in cash for four years from 2005 to 2008. No one will invest with a manager who does this, which is why we’d make the obvious recommendation to hold substantially less than 100% cash no matter how expensive things get. We run one long-only strategy and one strategy that goes to cash when valuations are terrible. Currently the second strategy holds about 50% cash.

Going forward

The price/book for value stocks is the highest on record. We think that for those interested in reducing volatility and/or avoiding a permanent loss of capital scenario a la 1929, going into cash is a smart idea. The results of the Long-cash portfolio are pretty compelling. The fact is clients hate volatility, and it does them no good to offer them a strategy that whips the market if they can’t stick to it. Lowering volatility offers them a way to do this.

More broadly, what we’ve found is that over the last 50 years when value stocks get expensive, returns suffer. And at times, the valuations get so high that the stocks don’t earn out them out. When Warren Buffett says that a stock is just like a bond insofar as the more you pay for it, the less it yields, we’d say he’s about 70% right.  

But don’t trust us—or our backtests. For any backtest or any theory to be valid, it must work out-of-sample. Almost no one in the financial world checks anything out-of-sample, but in our case, we have the best out-of-sample data possible, and that’s real-world results. From Steve Romick’s FPA Fund and Seth Klarman’s Baupost Group.

FPA is one of the best mutual funds in the world. Steve Romick was on the short-list for “Mutual Fund Manager of the Decade” several years ago. FPA has compounded capital at 10.9% a year versus 9.3% a year for the S&P 500 since inception, but it’s done so with less risk. FPA carries over 50% cash when valuations are high, which is why FPA was down only 21% when the market fell 37% in 2008.

Seth Klarman, meanwhile, is one of the best investors to have ever lived. He’s compounded capital at 17-20% a year since the early 80s while taking substantially less risk than the market. (Like FPA, he almost always holds a lot of cash, and in 2008 he was down only in the low teens.)

Both funds are not only outperforming the market, they’re doing so with less risk. This is the opposite of what CAPM and those pointy-headed little misers in academia tell us should happen. But we think it is because Steve Romick and Seth Klarman are focused on absolute value and taking less risk that they are outperforming to such an extent and doing so in a way that their clients can handle.

We believe paying attention to the absolute valuation of value stocks is a sort of meta value investing. You're value investing in value stocks themselves. The fact that valuation matters even for value stocks is yet another example of the power of value investing. There’s no need to get fancy. One only needs to control what one pays, and be willing to fearful when others greedy, and a whole new opportunity set opens up.

Limitations of this analysis

It’s always useful to “know your error,” so I’d like to end with some caveats about where all this analysis could go wrong. Recall this chart showing the price/book over time for the value decile:

The problem with value investing is that it depends on mean reversion, and the world simply isn't 100% mean-reverting all the time. Value investing makes very little room for the paradigm shifts that permanently raise or lower the mean. Imagine if you thought the valuations prevalent in the 1940s were the norm. You would have gone into cash in the 80s and never left it, thereby missing out on decades of return.

We believe the past is the best way to predict the future because the future is almost always more similar to the past than not. But as Mark Twain said, “History doesn’t repeat, but it does rhyme.” There is the possibility that the range of valuations for value stocks that has existed for the last 50 years will be different in the future. If backtesting could tell you what exactly to do in the face of an unknown like this, then mathematicians would rule the world. They don’t because we live in a world of rough approximations of the truth rather than precise measurements of it.

What I mean to say is we think it’s important to adhere to Graham’s concept of “margin of safety.” Currently our model says go 100% to cash. We’ve gone 50% to cash not just to satisfy clients, but to gives ourselves a margin of safety in the event we’re wrong. In our opinion, it’s almost always best to err on the side of hedging way too little and much too late. 

Catalyst

None