## Sunday, July 15, 2012

I took a bit of time to visualize and quantify just how much the frictional costs of trading -- bid-ask spreads, commissions, and taxes -- are costing traders. There are many variables involved in answering this question so I had to make some simplifying assumptions.

I considered all possible combinations of:
• 3 different trade sizes (\$5000, \$25000, and \$125000)
• 5 different commission costs (\$3 per roundtrip trade -- meaning \$1.50 per individual trade), \$5, \$8, \$10, \$20)
• 7 different trade frequencies (1 roundtrip -- meaning 1 buy and 1 sell -- trade per month, 2, 3, 5, 10, 20, 40)
and for each of the 315 possible combinations I calculate the annual drag created by bid-ask spreads and commission costs.

I then repeated the same calculation while adding the drag imposed by capital gain taxes, assuming 25% federal tax bracket (ignoring state taxes) and 8% average market return.

Finally I color-coded the results to show which combinations of the above variables can be at least remotely plausibly overcome by a truly skilled trader and which ones have no chance. Here are the results, before and after taking taxes into account.

Each percentage number that you see above represents annual drag imposed by one intersection of the four variables we consider. For example the 1.9% you see in green at the top left of the before-taxes chart (cell D4) signifies 1.9% annual drag imposed by executing one roundrip trade per month with size of \$5,000 on security that trades at 0.1% bid-ask spread and with total roundtrip commissions of \$3 (meaning two trades -- one buy and one sell -- with commissions of \$1.50). For simplicity I assume that each trade happens on the entirety of the portfolio. In other words, each trade is either buying to be 100% invested or selling to be 100% in cash.

As you can see, if we ignore taxes (as would be the case for tax-sheltered accounts) a truly skilled trader might have a remotely plausible chance of beating the market in tax-sheltered account, especially with very liquid securities, infrequent trades, and low commission costs.

Once we get to taxable accounts, however, there's virtually no remotely plausible chance of beating the market. Meaning that even the generous allowance of 3% annual alpha would still be just barely enough to break even  in the most fortuitous of circumstances after accounting for the drag of taxes, bid-ask spreads, and commissions. If I stretch my definition of "remotely plausible" to include consistently beating the market by up to 6% annually then the picture looks a little brighter. Unfortunately even allowing consistent 3% annually in alpha is starting to push the boundaries of reality. Adding 6% of alpha annually puts one halfway to Warren Buffett status (he beat S&P500 by around 12% annually over his investing lifetime, depending on exactly what years you count). Pardon me for doubting your (or anyone else's) ability to achieve such lofty returns on consistent basis, year after year and decade after decade. Needless to say, that beating the market just once or twice is no proof of skill -- it's the same "skill" as calling a coin toss or two correctly. Only when you get into dozens and hundreds of trades over a long time period does it become plausible to say that beating the market is an indicator of skill. Also see http://www.longtermreturns.com/2012/03/benchmarks-for-traders.html

At the very least I hope the above charts clearly show why an average day-trader -- who might flip a few thousand worth of stocks every few days in a taxable account -- has no remotely plausible chance of beating the market in the long run. Even if he is truly skilled -- a huge if, if there ever was one -- the frictional costs are simply too high to overcome. Small, frequent trades in taxable accounts are a surefire way to lose money while enriching your broker (with commissions), market makers (with spreads), and Uncle Sam (with taxes).

#### 1 comment:

1. What a hard core analysis. Nice work. I recently found this article regarding mutual fund frictional costs and found it very illuminating.

http://online.wsj.com/article/SB10001424052748703382904575059690954870722.html