Some data suggest that stock pickers are having a harder time outperforming the market. Each year between 1995 and 2007, for example, on average, 50% of mutual funds focusing on large, fast-growth companies beat the Russell 1000 Growth Index, a benchmark for that category, according to Morningstar Inc. Over the past year, only about 24% of those funds beat that index.Gotta love that. The folks who evaluate mutual funds for a living, (who don't even include all the funds that go out of business,) say managers who pick amongst large market capitalization "growth" stocks are as good as a coin flip when times are good. Lately, they look much worse. So, what's the problem, and should we attribute this to the death of stock picking?
(Remember, John Bogle, the founder of Vanguard, will always remind you mutual funds are not on average able to beat the market. On average, they are the market, and they have to pay fees to themselves and their brokers.)
Active portfolio management (stock picking, as opposed to index funds) imposes biases in portfolios in favor of smaller companies relative to comparable indexes. The S&P500 and most others weight individual stocks by their market caps. This means the largest companies make up the vast majority of the risk in any particular index. For example, GE's weight in the index is significantly greater than that of MMM.
The distribution of market caps has very high skew: There are many large outliers relative to the number predicted by a normal distribution, or bell curve. Got that? Relatively speaking, too many big companies. (No, that's not a political statement. It's an observation about an empirical distribution!!)
Unlike indexes, active managers don't weight stock picks based on market cap. Active managers weight positions based on information. What does that mean? In the simplest sense, their "best idea" will be their biggest position, second best idea, second biggest, etc. [This does hold even for complex quant modeling, but "ideas" and "biggest position" translate into "signal quality" and "deviation from benchmark."] Those positions will have a distribution across market capitalization too.
We have no reason to expect this distribution will have the same skewness as an index. I suspect, for most managers, positions are distributed uniformly by market cap, or even skewed in the reverse direction. I say this because discovering valuable information of performing better analysis ought to be easier in areas with less attention--smaller companies. Which company seems more fertile ground for overlooked analysis Allstate or National Security Group?
What does this mean for stock picking's demise? When small companies outperform large, stock picking has a tailwind. This tailwind can help massively at times. From the beginning of 2001 to the end of 2006, the Russell 2000 outperformed the S&P 500 by more than 50%. With that kind of tailwind, what's not to love? about stock picking? The problem is, since 2007, the two indexes are basically neck-and-neck.
Think about this exercise: Monkeys throwing darts at the stock listings in the newspaper amounts to equal weight random portfolio management. (Every stock has the same chance of landing in the portfolio if the dart hits.) The monkeys will have a better chance of winning when smaller stocks outperform. That doesn't mean monkeys are smart. It doesn't even mean they're lucky!
Let's not forget about value biases and correlation...to be continued...
With the degredation from 50% to 24% of those funds beating the index, doesn't this indicate that picking has somehow moved from no better than random to actively worse than random? Perhaps the best fund structure is to lock the manager in a room as a kind of anti-oracle, and set up a shadow fund that just inverts his positions.
ReplyDelete