Divergence between InTrade and BetFair political prices??

Chris F. Masse January 3rd, 2008

I have been inundated by e-mails asking me to comment on the Crooked Timber posting on arbitrage possibility between InTrade and BetFair on US presidential elections. My answer: I would rather see such a study on long series —as Justin Wolfers and Eric Zitzewitz did for their “Prediction Markets” paper (if my memory is correct). Their conclusion (again, from memory): there isn’t any arbitrage possibility.

Now that economist Mike Giberson (just like the rest of our readers, according to yesterday’s web stats) is out of Christmas hibernation, he will jot down a comment on that, I’m sure.

Tags: arbitrage, BetFair, InTrade

Analysis (Data) , Analysis (Data) - Market Arbitrages , Exchanges & Markets , Market Prices & Probabilities
Comments(3)

3 Responses to “Divergence between InTrade and BetFair political prices??”

# Michael Giberson on 03 Jan 2008 at 8:41 am

Sure, it is possible that there are arbitrage opportunities. The prices quoted diverge, and on the assumption that the contracts traded are identical (and therefore will be expired in the same manner by each exchange), a profit opportunity may exist.
.
But just quoting divergent prices isn’t enough to convince me: show me the money! (Or, at the very least, a convincing numerical example.)
.
The CT poster provides an alleged example of pure arbitrage profits in comment #7 (which assumes, in effect, infinite liquidity at the quoted prices — i.e. that one could trade $500 on IEM and $1500 at Intrade without moving prices — and ignores transactions costs). However, comment #13 points out an error in the math and shows that the alleged arbitrage is actually a risky bet that wins if Guiliani wins and loses money otherwise.
.
Sure, if one can buy at 22.4 cents in one market and sell at 28.5 cents in another market, that is a sure-fire 6.1 cents arbitrage profit. In this particular case, of course, you don’t get your 6.1 cents for at least another 11 months, and you may pay transactions costs in both markets. And if we don’t assume infinite liquidity in a real money market, then we must expect that after a few trades at 6.1 cents our profit margin gets squeezed to 6 cents, then 5.9 cents, and so on.
.
So, how much profit is out there for the arbitrageur?
# Chris. F. Masse on 03 Jan 2008 at 8:47 am

Good idea to put a dot (”.”) to force the system to mark a blank line between paragraphs. I, too, dislike that with the new WP theme. If one day I find a way to hack the code, I’ll do.
.
Oh, and I like your comment.
# Chris. F. Masse on 06 Jan 2008 at 6:15 pm

JONATHAN GALE:

There is no arbitrage. If you want to do the arb, you need to take both
sides. Taking both sides requires using capital, approximately $10 for 1
Intrade contract. If you tie up $10, you better make at least the risk free
rate, approx 5%. That being said, a difference of a couple of points is not
an arb. You need at least 5 points.

Of course this is assuming the capital you tie up with a prediction market
is not earning interest. Intrade provides interest on your account if it’s
big enough. Betfair does not. So you need to take that into account in the
math.

Oh yeah, there is also transaction costs.