The $25b cut in agency debt purchases looks like its just a technical issue. It would make no sense for the fed to do a mini-exit like that for no reason
You can look at the Value Line Arithmetic Index, see http://finance.yahoo.com/q?s=^VAY. It is an equal weight index of around 1800 US stocks that is rebalanced daily, free of survivorship bias. It had a 60% max drawdown in 2008. The Blackstar stock system had half of that: 29.3%. This was out of sample, a true forward walk. The risk adjusted returns of this stock system are twice as good as buy and hold. BTW: No stock trendfollowing system will ever produce any iota of alpha in a straight up market as the mid 90s. It's impossible by definition.
We equal weighted the 3,000 most liquid stocks and rebalanced once per year. The result was lower returns, higher volatility, and greater drawdown relative to the S&P 500. It's not because of small caps stocks that the trend following system outperformed. There are several reasons for the outperformance, one of which is discussed in this paper: http://www.michaelcovel.com/pdfs/blackstar-sell.pdf We also recently applied a very simple, unoptimized channel breakout system to all stocks: http://www.michaelcovel.com/pdfs/DoesTrendFollowingWorkOnStocks_2.pdf
You mentioned there was several reasons for the outperformance, the data you provided(which I found interesting) tells me that system beats the market simply because it appears that there is an inneficiency in stock markets where badly performing stocks tend to be overpriced therefore being long the stock market ex-bad performers(those who had significant declines) beats the market with lower volatility, regardless of trends I look forward for your evidence on the other reasons why the breakout system works
It appears that the VAY has beaten the S&P500 by a wide margin over the last decade http://finance.yahoo.com/q/bc?s=^VAY&t=my&l=on&z=m&q=l&c=^GSPC It matched during 90's, then outperformed during 00's
Remember VAY is rebalanced daily without slippage nor transaction costs. It is not an investable index.
The Russell 2000 beat the S&P500 from 2000 to 2009 by quite a bit. That should explain at least part of the returns of that system
Yes obviously you are right. The SP500 is not a good benchmark. But: the maximum drawdown even against the VAY (which is a "fair" benchmark even though doesn't account for slippage nor for transaction costs!) was cut in half by the Blackstar Funds approach, out of sample (!). Isn't that amazing? IMO, it is. Why are you just looking at profits, not at drawdowns and standard deviation? I think I've said it in another thread once: Trendfollowing in my book was never designed to provide superior nominal returns. But I believe it can be utilized to achieve superior risk adjusted returns: participating in a good chunk of the upside/downside in many markets while limiting downside/upside tail risk.
Profits is what pay the bills, low drawdowns is nice though. Here is the thing if that system excess returns came from the small cap effect and avoiding stocks that had large declines, then it doesnt really support trend following. Even though avoiding stocks in large downtrends 'worked', that seems likely due the fact the large hedge funds have a hard time arbing those stocks due, borrowing difficulties, costs, liquidity and low profit potential compared to AUM, so it would be more like a anomaly that wasnt closed due technical factors rather than TF working. But I'm waiting for Eric's defense
IMO it does. If nominal profits is all you're after and you don't care about drawdowns you can use leverage on the above system and suddenly you have a system that has the same (historical) drawdown as the benchmark but double the profits. Most diversified futures trendfollowing systems are leveraged with exactly this rationale. They (try to ) neutralize tail risk out of markets and then feel more comfortable using leverage. It's like being long volatility (trend following) vs. being short volatility (counter trend, mean reversion). Here's an interesting short paper on the subject http://www.estlander-roennlund.de/binaer_view.asp?BinaerNr=70