Seeking Alpha

Marc Gerstein's  Instablog

Marc Gerstein
Send Message
Marc H. Gerstein (http://twitter.com/#MHGerstein or if already on Twitter, search for @MHGerstein) is an independent investment analyst/consultant specializing in rules-based equity and ETF investing strategies, with particular emphasis on small-cap equities and leveraged ETFs. Many of is views... More
My company:
Portfolio123 and Ariston Advisors
My blog:
Forbes Low-Priced Stock Report
My book:
Screening The Market
  • Sometimes, We Have To Just Throw Out The Past 0 comments
    Mar 29, 2010 2:38 PM

    By now, we've all heard "past performance is no assurance. . . blah, blah, blah" so often, it almost sounds liked a nursery rhyme. But it may be time to stop a moment, think about it, and possibly give the saying a new, broader, spin. Usually, the phrase is meant as a warning that cautions us against assuming the future will be as good as the past. Actually, though, this is a two sided coin. We need to also be careful about assuming the future will be as bad as the past.

    Crises, crises and more crises

    I suspect that years from now, many investors will still be telling stories about what happened, what they did, and what they didn't in the 2000s to those among the new generation who remain courteous enough to pretend to care. But what about the big picture? What will people think of it, if they think of it, a decade or so down the road?

    After all, how often do you think about October 1987 (how many even know the S&P 500 dropped more than 20% in a singe day)? How often do you think about late 1998 (do you recall the dangerous financial unraveling we faced, or do you just remember the early stages of the new-economy tech stock boom)? How often do you think about 1969-1972 (how many even know why I mention it), or 1974-75? For those of you who don't know what CLEC stands for, how often do you think about 2000-02?

    Every time a crisis occurs, people seem to want to connect it with what we think of as the big one, 1929-32 and the rest of that decade. So far, though, none of the subsequent ones have come anywhere close. There's much that has been and still can be said about what made the 1930s so distinct from the crises we've had since then. That's well beyond the scope of this post. The topic for today is how investors should consider crises when evaluating potential strategies. This is a particularly poignant topic for those who have access to backtesting, as do users of Portfolio123 and StockScreen123, and notice how badly results fare once we hit October 2008.

    The Prudence Police

    It's so easy to preach attention to worst-case scenarios. Nobody will ever accuse you of spreading hype or advocating recklessness. And those with 20-20 hindsight can definitely appreciate how well off everyone would have been had they listened to you in the first place (even if you hadn't started to articulate your warnings until well after the fact).

    We definitely don't want to ignore the possibility of more crises down the road. Here are two simple approaches that stand a good chance of protecting you or at least helping you keep losses to tolerable levels.

    Stay in cash

    This is the most extreme perma-bear solution and some will, or already have, adopted it. Most who read this will not want to go that far.

    Engage in some form of market timing

    This sounds exotic to some and like voodoo to others, but actually, it's more do-able than many realize. Something as simple as being bearish if the S&P 500 50-day moving average is below the 200-day moving average can be surprisingly helpful. (This is a nice approach because it's easy and you don't have to figure out why the crisis is here, just let price action tell you it is here.) If you're bearish, go all cash or in the alternative, consider staying in stocks but allocating 10% or so of your portfolio to a short or leveraged short ETF. If you're into options, this would be a cue to sell calls and/or buy puts.

    Strategies For All Seasons

    Few strategists or investors go all cash all the time. Some are willing to engage in market timing. But many others, especially those who have access to backtesting and simulation, work to develop good-for-all-seasons strategies that would be capable of withstanding crises as well as the best of times. Put another way, this is a pursuit of "robustness."

    Do you ever notice, though, how great robust strategies look on paper (That's why we say they're robust!) but how, when the next crisis comes around, they don't deliver? I suspect every investor who, thought at the start of 2008 that he or she was prepared for bad times is shaking his or her head and muttering.

    I believe robustness is a nice quest but like the pot of gold at the end of the rainbow, it's forever elusive. If you think you have a robust model, I suspect that what you really have is a four-letter word: L-U-C-K. Be glad you survived the last crisis or past crises. But don't get complacent about the next one.

    Crises in perspective

    Here's a simple, but quite revealing, screening exercise. It involves back-dating so you’ll probably want to try it on StockScreen123.com, which is in beta and, hence, free.

    Create a simple two rule screen to look for stocks with market caps of at least $5 billion that declined 20% or more in the past four weeks. Use StockScreen123’s “Free Form” rules interface. You can copy each of these rules into your clipboard and paste them into the screener (create two separate rules):

     

    MktCap>=5000
    Pr4W%Chg<=-20

     

    Click on Totals. Note the percent of stocks meeting the market cap requirement that pass. As of this writing, it was one out of 775, or 0.13%.

    Change the as-of date and run this same screen as of 7/31/02, when the dot-com explosion was really in full force. I got 101 out of 521, or 19.4%.

    Now set the as-of date to 10/31/08 and run the screen again. I got 443 out of 547, or 80.1%.

    Change the size and percent-decline thresholds. No matter how you slice and dice it, you'll see October 2008 was one heck of a bad month, even by bad-month standards.

    Figure 1 provides a bigger-picture perspective. It's compiled by backtesting the above screen starting on 3/31/01 and rebalancing every four weeks. The graph shows the number of stocks falling 20% or more in the past four weeks as a percent of those that pass the market-cap test.

    Figure 1

    We see that by counting every four weeks from 3/31/01, I was able to capture an even worse period in 2002. we also see that the post 9/11/01 period was quite awful. Notice, though, how October 2008 stands out as being one of a kind, at least since 3/31/01.

    Pause for a moment and think about that: approximately 80% of big (supposedly respectable) stocks dropped 20% or more in a single month! Now do you understand why your strategies didn't work? Nothing worked, except for L-U-C-K. That wasn't a stock market we experienced, where some do well and others don't based on some sort of criteria we can try to discover. That was a widespread dumping of equities as an asset class. If it could be sold, there was an 80% chance it would be dumped.

    Do you really think it's worthwhile to try to build a "robust" model that will outperform the benchmark even in October 2008? Yes, with enough "data mining," you could probably come up with something. But why bother? What chance do you think any such model would have helping you in the future. It might have to be so conservative as to make you wonder why you're even bothering to deal with equities at all. Or it might be an instance of that four-letter L word (which tends to be what happens when you engage in data mining).

    Market timing, as mentioned above, will probably serve you better - seriously. Figure 2 represents a backtest of a strategy of being in the S&P 500 SPDR ETF (NYSEARCA:SPY) when the index's 50-day moving average is above the 200-day average, and being out of the market at other times.

    Figure 2

    Figure 3 offers a different perspective. It plots all rolling four-week declines in the S&P 500 going back to January 30, 1950.

    Figure 3

    Finally, figure 4 shows all rolling four-week S&P 500 percent changes since early-1950.

    Figure 4

    So again, do you really think it's worthwhile to try to come up with a good-for-all-seasons strategy that will protect you from the sort of thing that occurred in October 2008? Is it even possible?

    Ignoring The Prudence Police

    When it comes to formulating a strategy, the prudence police will definitely wag their fingers if you ignore October 2008. FINRA and the SEC would probably do likewise, or worse, if you're a professional and creating a backtest that you want to use as part of your marketing efforts.

    But for your own consideration, when deciding on what kind of an all-in equity strategy (i.e. aside from market timing) you want to create and use, I think the best way to deal with October 2008 is to ignore it. Throw it out. There's noting useful you can learn from it. Ditto for other exceptional crises, such as September 2001, October 1987, etc. (If you're a pro using test results to market, you'll have to show the data, but don't be shy about explaining it.)

    "Respectable" types will talk against market timing and advocate robust modeling. Don't believe it. Market timing can help you. Robust modeling can fool you.

    Disclosure: No positions.
Back To Marc Gerstein's Instablog HomePage »

Instablogs are blogs which are instantly set up and networked within the Seeking Alpha community. Instablog posts are not selected, edited or screened by Seeking Alpha editors, in contrast to contributors' articles.

Comments (0)
Track new comments
Be the first to comment
Full index of posts »
Latest Followers

StockTalks

More »

Latest Comments


Most Commented
Posts by Themes
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.