Comments on: Iowa Electronic Markets quiz

By: Dean Jens

Dean Jens — Fri, 30 Apr 2004 03:20:33 +0000

http://biz.yahoo.com/ifunds/040428/20040428_putcallparity_adv_bon_jb_1.html notes a dramatic violation in put-call parity on bond ETFs. Shares for shorting simply can’t be had, so there’s no arbitrage. When single stock futures were introduced, I was asking people whether the futures price of a potential short-squeeze might drift a bit below fair value for basically this reason. As with many of my questions, nobody seemed to care. At the time I looked at the options market for Krispy Kreme, but saw no significant effect. This is related to the previous point about the first question; if you can’t short the futures and pull the procedes out, and the IEM doesn’t pay interest on cash, it’s hard to see that there’s going to be a compelling force keeping the prices in line with interest rates.

By: dsquared

dsquared — Fri, 23 Apr 2004 18:52:11 +0000

In vaguely related news, the Economist’s Big Mac index has been claiming for the last ten years that a Big Mac is a homogeneous product around the world, despite no fewer than four letters from me telling them that it isn’t; fat content, calorific value and even weight differ significantly between local markets.

By: dsquared

dsquared — Fri, 23 Apr 2004 13:43:27 +0000

Question number 1, for instance, makes absolutely no sense to me above the jump but is perfectly clear below. Eh? Ahh. I typed it from memory, copied and pasted, then corrected the contract names above the fold but not below. Since the error improves comprehensibility, as you suggest, I’ll keep it. But practically speaking, it surely isn’t a coincidence that index funds that try to emulate the S&P 500 index do, in general, a very good job of duplicating the returns you’d get if you bought-and-hold (with rebalancing, etc.) the index. To put it differently, it’s not as if the strategies that index-fund managers use to track the index involve taking on market risk. Good point; it’s no more a coincidence than that bakers make bread (which is better for you than a buy-and-hold strategy with respect to flour and yeast). My point is, I think, your own. Vanguard does a much better job than you would be able to do yourself (trying to build your own replicating portfolio is a mug’s game). I’m less convinced than you are, though, that the outperformance built into Vanguard’s trading returns represents pure alpha (generated through execution) rather than market risk. I suspect that VFINX is taking on considerable market risk in the tails—in other words, I suspect that its replication strategies rely on good liquidity, which is a risk positively correlated to the market. So I don’t really think there’s a problem with studies that compare managed funds to a buy-and-hold strategy with the S&P Probably right from a statistical point of view. But it betrays a certain mindlessness (I’d also comment that I’ve seen studies published in decent journals that completely ignored the little matter of dividends!)

By: Steve Carr

Steve Carr — Fri, 23 Apr 2004 07:59:28 +0000

As with the IEM, the theoretical point that VFINX is not using a pure buy-and-hold strategy (which is mainly a function of the fact that the S&P 500 is itself in a perpetual process of rebalancing, adding and dropping stocks, etc.) is right. But practically speaking, it surely isn’t a coincidence that index funds that try to emulate the S&P 500 index do, in general, a very good job of duplicating the returns you’d get if you bought-and-hold (with rebalancing, etc.) the index. To put it differently, it’s not as if the strategies that index-fund managers use to track the index involve taking on market risk. More important, the reason we care about this is because we want to know if money managers can outperform the S&P. That question would be significantly less interesting—and perhaps unfair to money managers—if there were no way for investors to guarantee themselves the S&P’s return. But practically speaking, there is: buy VFINX. So I don’t really think there’s a problem with studies that compare managed funds to a buy-and-hold strategy with the S&P. At the same time, you’re absolutely right about the operational complexities involved. Vanguard’s index-fund money management has been exceptional at keeping down costs and capturing returns, and other index funds have been significantly less successful at that.

By: Anno-nymous

Anno-nymous — Fri, 23 Apr 2004 07:20:57 +0000

I think your questions #1 and 2 are written differently below than they are above. Question number 1, for instance, makes absolutely no sense to me above the jump but is perfectly clear below. Eh?

By: dsquared

dsquared — Fri, 23 Apr 2004 01:00:08 +0000

Wesley has it right; Vanguard (and other trackers) do all sorts of things to make their money; they also play some games with their execution which look to me to be for all the world like a bank prop trading desk. Given that Vanguard doesn’t buy and hold, because it couldn’t both buy and hold and track the SP500, one comes to the conclusion that it has to generate quite material outperformance of the SP500 just to track it. Academics often give absolutely no thought whatever to the operational complexities of conceptually simple strategies; I used to sit near someone who did the Royal Dutch/Shell arbitrage for a living, and I can tell anyone who asks that arbitrage trading is neither “simple” nor “risk-free”, and that 99.9% of people who tried to do it would get their faces ripped off. But I agree; comparing VFINX to a mutual fund is a legitimate comparison, as long as you don’t make the mistake of assuming that VFINX is equivalent to buy and hold. If you sneaked a look at the last chapter of Campbell, Lo & McKinlay’s “A Non-Random Walk Down Wall Street”, you’ll see their conclusion is that it is possible to sustainably outperform the market, but that the gains which can be statistically shown to be there come from execution and market impact, not investment strategy.

By: wesley phoa

wesley phoa — Thu, 22 Apr 2004 22:48:33 +0000

Vanguard outperforms the S&P 500 by engaging in security lending; if you take out the excess return due to that, the return is lower than the “naive” S&P 500 return. For example, if they could lend 100% of their securities they would be picking up about 30 bp, I think. Also, I believe that security lending reduces the after-tax return on a stock portfolio because a “payment in lieu of dividends” received on lent shares does not qualify for the lower dividend tax rate.

By: Steve Carr

Steve Carr — Thu, 22 Apr 2004 18:45:15 +0000

Doesn’t an index fund do the rebalancing for you? Vanguard 500, for instance, has captured just about all of the S&P’s returns over the past three years (actually, it outperformed the index by a tiny bit: .01% in 2003, .12% in 2002, and .04% in 2001. So if you buy the Vanguard 500 and hold it, you’re effectively buying-and-holding the S&P 500. Comparing the performance of managed funds to the performance of VFINX is a straight apples-to-apples comparison: you can buy Mutual Fund X or you can buy VFINX.

By: dsquared

dsquared — Thu, 22 Apr 2004 18:35:05 +0000

I think not; 1 Bush plus 1 Clinton on the contract specified is a cash-and-carry arbitrage, so it’s the risk-free rate, shurely?

By: Hipocrite

Hipocrite — Thu, 22 Apr 2004 18:25:21 +0000

One issue- 1. “Risk-adjusted TVM.”

By: dsquared

dsquared — Thu, 22 Apr 2004 17:48:26 +0000

Re Iowa, you’re probably right. Re SP500, my bugbear is studies which refer to “a Buy and Hold strategy on the SP500 index”. You can’t “buy and hold” the SP500. It’s an index that gets regularly rebalanced. If you made the same trades, you’d typically find a cash shortfall between the value of the stocks you sold and the ones you bought and the index does not reflect the losses one would sustain on rebalancing. You could buy and hold the SP500 future, but then you don’t get the dividends and that kills your long term total return.

By: Steve Carr

Steve Carr — Thu, 22 Apr 2004 16:40:43 +0000

Daniel, your point about paying close attention to the specifications of the contracts in time series is an excellent one. But I don’t understand your reference to studies of mutual-fund performance vs. the S&P. What is it that academics aren’t paying attention to in these studies? Also, with regard to the question of the time value of money, you’re right as a matter of theory. But in practice, aren’t the sums involved in the IEM so small that most investors feel free to neglect time value? Even if you invest the maximum in a contract, and hold it for a year—something few IEM investors do, I imagine—you’re talking about two dollars (that could have been earned at the risk-free rate). At that price, it’s probably not worth spending the time adjusting your forecast to take into account the discount rate.