Traders' Roundtable

I computed the terminal wealth and compounded annual rate of return for all stocks that traded on the NYSE, AMEX and NASDAQ between the years 1983 and 2006 (including delisted stocks). I did this on a total return basis (dividends re-invested). Attached are the resulting distributions. I am somewhat surprised by the results. I'm curious as to what the rest of you think.

Terminal wealth** = ( last closing price / first closing price ) - 1
Annual compounded return** = (( last closing price / first closing price ) ^ ( 1 / ( number of days / 220 ))) - 1

**data proportionately back-adjusted (on a percentage of stock price basis) for all data points prior to the ex-dividend date

--
Eric Crittenden
Director of Research
Blackstar Funds, LLC
2375 East Camelback Road Fifth Floor
Phoenix, AZ 85016
eric@blackstarfunds.com

Looks to me as if 'survivorship bias' is much larger than generally accepted ...

Interesting graphs.

This is a huge issue for buy and hold "investors" and those who trade long-only funds but not so much for one who trades a mechanical system which will get you out based on price behavior.

- Forum Mgmnt

Assuming that a trend-follower has the ability to short stocks as well, survivorship bias would imply higher returns in live trading compared to back-test (probably higher risk-adjusted returns as well).

nickmar wrote:Assuming that a trend-follower has the ability to short stocks as well, survivorship bias would imply higher returns in live trading compared to back-test (probably higher risk-adjusted returns as well).

This is true. But, this is usually nullified (for most people) because they are backtesting on data that has not been adjusted for cash dividends (which overstates the profitability of short selling).

We spent a long time trying to come up with ways to short stocks profitably. We couldn't come up with anything that worked over long stretches of time (>10 years) nor could we come up with a fundamentally sound reason to expect to. Stocks are basically a perpetuity with unlimited upside and guaranteed containment on the downside. Approx 50% of stocks have historically ended their existence with a negative terminal wealth relative, but their potential declines have always been capped at -100%. The other 50% of stocks enjoyed potentially unlimited gains, many exceeding 500%. From a short sellers perspective the math simply doesn't yield a positive expectancy.

The best we could ever come up with was -4% compounded annual return, which is better than -12% one gets from simply shorting the indexes. Still it's hard to get excited about losing less money than something else.

ec

in that -4% figure, did you include the t-bill return from the broker on the short sales? institutional accounts can usually get a short rebate. . .

budonk wrote:in that -4% figure, did you include the t-bill return from the broker on the short sales? institutional accounts can usually get a short rebate. . .

Our true cost of borrow is approx 50 basis point. We can borrow at libor + 25 bps, short stocks which brings in cash, invest the cash at libor - 25 bps. The above figure does not account for these borrowing costs.