Tuesday 22 November 2011

Price of Peters

Folks on the iPredict discussion forums argue a bit about whether anybody's mucking about with prices on the Winston Peters contracts. I'd previously suggested prices on the Vote Share market might be a bit high, given the much higher cost of shorting contracts at $0.05 than going long. But some on the forum are now going the other way, arguing that evidence of Peters prices below 5 cents suggests big anti-Peters interests coming in to drive down the price as there's little potential return in it.

First, let's look at current polling. The first place I go for current polling summaries is KiwiPollGuy. Here's the latest:


The black line is a moving average of polls on New Zealand First, with black dots and lines representing individual poll results and their confidence intervals. The confidence interval of the moving average doesn't pass 4% and hasn't for a while. If iPredict prices are above 4 percent, either punters are predicting a late surge or somebody's a bit optimistic. iPredict has NZ First on 4.7%.

There's a ton of thickness on the ask side of the order book around the 5 cent mark. Where some on the forums take this as evidence of anti-Peters conspiracies, I take it rather as that a bunch of traders figuring out that some Peters fans keep making high volume trades trying to push the price up around the 5 cent mark and are leaving orders there waiting to take their money. At least that's why I've one of those ask orders in [full disclosure: I'm short both Peters contracts]. The deposit limits on iPredict are now fairly high. And if the fair price on Peters is really 3.5%, then selling contracts at 5 cents gets you a cent and a half on a 95 cent investment over a week - that's a pretty decent annualized return even if it's low compared to other potential iPredict contracts.

Traders also have Peters at 42% likely to return to Parliament. That's a bit lower than expected if his expected vote share is around the 4.75% mark, but on the high side if it's more likely to be 3.5% [and I'm consequently short and selling].

A fair bit of the grumpy vote that would have gone to Peters is likely to head over to the seemingly ridiculously well-funded new Conservative Party, which is also likely siphoning a bit of support from United Future [full disclosure: I'm buying the Conservatives at 1.6%, but would be selling at 2.5%]. I see at least as many Conservative billboards around as NZ First ones, which gives a first indication of support where the Conservatives aren't generally included in polling.

Bottom line: it's plausible to get manipulation stories resulting in NZ First prices being too high on the Vote Share market given how expensive it is to short low probability events. I can't see how it's possible to maintain manipulated low prices.
The maximum a user can deposit is $2,500.00 in a 6 month period. This is subject to a lifetime net contribution limit (defined as funds in minus funds out) of $10,000 in total.
Suppose an anti-Peters manipulator went all-in, planning well in advance to get a $10k nest egg all to be used in keeping Peters down. Each sell contract at $0.05 costs him $0.95 in frozen liquidity. So he can sell 10,526 contracts. How much would it cost to buy out his entire investment? $526. There are eight traders with total iPredict net worth north of $5k. Even if all of them had their entire net worth devoted to conspiring against Winston Peters, two new traders each depositing $1500 could pretty much clear them out. It's too expensive to maintain price manipulation either forcing down a low price contract or forcing up a high price contract.

No comments:

Post a Comment