It is good to be reminded of these investment basics. I have one question regarding market timing. What about studies which have shown that using a rule, such as the 200-day moving average, to be in or out of the market seem to reduce portfolio risk (i.e., volatility) while producing nearly the same return? Wouldn't this be an exception to the general idea that timing is not useful? Thanks.

Bill, first I am not aware of any academic study showing that. Second, it certainly would not work in taxable accounts which is where you should hold your equities if you have a choice.

In an article posted by Mark Hulbert on Marketwatch in May 2007, he cites a study conducted by Ned Davis Research that compares annualized returns from 1979-2007 for holding the S&P500 vs. annualized returns using the 200-day moving average to switch between the S&P500 and commercial paper over the same period. Holding the S&P500 produced a 10.2% annualized return while the trendfollowing approach produced an 11% return with a 13% lower standard deviation. Based on this research, Hulbert concludes that trendfollowing using the 200-day ma works to improve risk-adjusted returns. In another study in 2006, Mebane Faber reports that over the period of 1900-2005 the CAGR for the S&P500 was 9.75% with an SD of 19.9% vs. a CAGR of 10.66% with an SD of 15.4% for trendfollowing using the 10-month moving average to switch between stocks and 3-month treasuries. Are these studies flawed? I'm not particularly concerned about taxation since my portfolio is held within an IRA. The crux of the issue for me is whether or not simple timing using moving averages is a viable approach for reducing portfolio risk while not adversely impacting expected returns. Even if returns were reduced (which they do not appear to be), I'd still be interested in trendfollowing if risk-adjusted returns were improved. Your comments appreciated.

First, there is an old saying that you can torture the data until it confesses. You can always find some technical indicator that worked in past. But the evidence is that such indicators are highly unlikely to continue to work in the future. The very fact that people become aware of the data means that they will no longer work as people anticipate such strategies.

Gene Fama tells the story that his first job was to find strategies that worked using statistical analysis. He was always able to find strategies that worked in past but failed when he tried to implement them.

Also as I stated this would not work in taxable accounts--where if you have a choice you should hold your equity risks.BTW-this is a very common location error.

There are quite a few investment mavens pushing "Market Timing". Ben Stein and Phil DeMuth offer readers of their books to subscribe to their Market Timing service on line. See here:

www.hedgefund-index.com/markettiming.asp

As to Larry's saying that stocks belong in taxable accounts, it's b/c you can go for hi tax efficiency in index stock funds while you can't do it with bonds and bond funds. So you locate your bonds in tax deferreds accounts. The higher your taxrate, for a specific asset class that you're holding or selling, the more important for you is to follow this rule. The rule is based on the future net after-tax end value of one dollar at your retirement (or later) if placed in an asset class in each account. The higher the better. Step- up in cost basis option, after your death, is also a factor.

Larry- I share your concerns about data-snooping by relying on backtested results. I have no axe to grind; however, it occurs to me that one could invoke this concern to discount practically every approach to portfolio construction that is supported by backtesting, including the thesis you advocate that holding CCFs improves risk-adjusted returns (which I happen to agree with in spite of data-snooping). I'm not sure how to move beyond this problem, but Mebane Faber's study is quite interesting in this regard because he was able to confirm the utility of using the 10-month ma in numerous out-of-sample investigations, including several non-U.S. equity markets, other indexes such as the NASDAQ, and for non-equity asset classes such as Bonds, Commodities, REITS, and Gold as well. I've not seen many attempts to use out-of-sample confirmation elsewhere and it seems like a pretty good approach to trying to address the data-snooping concern about trendfollowing with moving averages. I'm not trying to defend trendfollowing but I just don't think the data-mining objection to it is compelling per se given Faber's results. However, it might be compelling in regard to using CCFs :-) Remember- I'm a fan! Thanks and I look forward to any additional observations you have on this.

Vig
First, re commodities, the reason commodities has produced such results is because they have negative correlation to stocks and bonds. That is a logical outcome since they are positively correlated to inflation and stocks and bonds negatively correlated. Also CCF often positively correlated to event risks that are negatively correlated to stocks and bonds (e.g.,we don't see sudden outbreaks of good weather but storms can create havoc). Thus there should be IMO some good logical reason to believe that a relationship should exist for you to believe it is likely to exist in the future. I cannot think of any such reason to beleive a 200 day moving average will work. Besides it produced no higher returns basically (though less volatility) and you could not do it in taxable accounts anyway-the tax costs would kill the strategy.

Second, yes there is some evidence of both positive and negative momentum in stocks. And DFA even uses that information in managing its portfolios. But I personally would not want to try to time the market based on any such strategies--and all of my equities are in taxable accounts so it would not work anyway.

To me this is not much better than what IMO is the foolish idea of seasonality in stock prices and relying on that to trade.

BTW-other poster: other reasons for holding stock in taxable accounts are also important--mainly the ability to harvest losses, the ability to use the foreign tax credit and also the ability to donate appreciated shares to charity, avoiding the tax altogether

Larry- The jury is still out for me on this and I appreciate your comments. But I do take intellectual issue with a couple of things you say in your last post. First, as you know, empirical findings can be valid per se and can provide actionable information. That said, because trendfollowing models of various durations seem to work historically across several different markets, one reasonable explantion to me seems to be that they help sidestep major losses due to downside cyclical price movements. It is important to note that the approach seems to function as a risk-reduction method not a return-enhancing one; i.e., the "efficient frontier" is moved left and up, which improves CAGR. Second, Jeremy Siegel investigated the use of the 200-day SMA with the DJIA from 1900. He found that when all transaction costs were included (taxes, spreads, commissions) the risk-adjusted return (but not absolute return) of the timing approach was superior to buy-and-hold. It is important to consider whether a portfolio is in accumulation or drawdown stage. If in drawdown stage, avoiding large losses is much more important than absolute return. It could be argued that Siegel's data support the use of trendfollowing even in taxable accounts, particularly in the drawdown phase.

Ibill
Assuming you do miss some of the downsides you also miss some of the upsides for the same reasons. And BTW-the chance of a large move up in a month (5% or more) is 70% greater than the chance of a large move down.
Again, I see no logical reason for the strategy to work--momentum is not a risk factor. On other hand it has worked for a long time--and may continue. But since everyone pretty much knows about it now, it should no longer work.
Even if you believed in it--for those in accumulation phase it would seem to be a poor strategy as can only spend returns, not risk adjusted returns (assuming that you can stay the course) but perhaps worthwhile in distribution phase. IMO still not good idea for reasons we have discussed--simply better IMO to lower your equity allocation accordingly and perhaps tilt more to size and value. That gives you lower dispersion of potential returns without lowering expected returns. You trade off opportunity for the fat right tail to have lower risk of left tail showing up.

I hope that helps and good luck

With reference to "Adhering to a plan requires that investors rebalance the portfolio, maintaining their desired asset allocation." What is the optimum plan for rebalancing a portfolio? And does that plan differ for a retired person? Thanks?