Intrade Manipulation Fail

Brad Plumer at the Wonkblog discusses a recent attempt to manipulate Intrade.

On Monday night, after the debate, Barack Obama was leading Romney on Intrade by around 60 percent to 40 percent. But at around 10:00 a.m. on Tuesday morning, Romney surged to 48 percent. Was this evidence that the conventional wisdom was wrong? Had Romney actually won the debate handily? Or, alternatively, was the nosedive in the stock markets putting a dent in Obama’s re-election chances?

Neither. As economist Justin Wolfers pointed out on Twitter, the huge swing toward Romney appears to have been driven by a single trader who spent about $17,800 buying up Romney shares and pushing the Republican candidate’s chances on Intrade up to 48 percent. But the surge only lasted a few minutes before other traders whittled the price back down to what they saw as a more accurate valuation. Romney’s odds of winning are currently back at around 41 percent.

…As Wolfers pointed out, this mysterious trader ended up overpaying by about $1,250 for shares that quickly collapsed in value. Was this just someone who made a bad trade? Or was somebody trying to influence Intrade odds in order to sway perceptions of the race? And if so, was it worth $1,250 to jolt the markets for less than 10 minutes?

Plumer quotes me from 2008 discussing an earlier attempted manipulation:

This supports Robin Hanson’s and Ryan Oprea’s finding that manipulation can improve (!) prediction markets – the reason is that manipulation offers informed investors a free lunch.  In a stock market, for example, when you buy (thinking the price will rise) someone else is selling (presumably thinking the price will fall) so if you do not have inside information you should not expect an above normal profit from your trade.  But a manipulator sells and buys based on reasons other than expectations and so offers other investors a greater than normal return.  The more manipulation, therefore, the greater the expected profit from betting according to rational expectations.

Addendum: Justin Wolfers offers more comment.


Does the $1,250 have to be reported to the FEC as an independent expenditure? Or $17,800?

Why would any reporting take place in the first place?

Also note that, the 'correction' pushed the valuation up to 41%, a higher steady-state value than previously. This means the $1250 was sufficient to give a 1% increase in the 'odds'.

(So much for efficient markets?)

Also interesting if you look at the chart is that the valuation was resting at 40 and then there was a flurry of activity bringing it up to 41 (testing the waters?), before a huge leap an hour later, causing a spike, adaptation, and bringing it back to a resting level of who's going to make a 'killing' betting that it goes back down to 40?

Maybe the right level is 41. Maybe there are transaction costs.

What do we think would happen if the Romney camp kept up a sustained attack on Intrade of one million dollars a day from today until the election day? Maybe the activity would change perceptions and fading it would not seem as wise as it did on the single spike this morning.

Two issues I see.

1. At $1M a day what could a campaign do? If you paid $200 per person per day, you could have 5,000 workers going door to door in swing states talking to maybe 20+ voters a day every day until election time. It doesn't seem plausible that the best use a campaign could make with an extra $1M a day is dump it on Intrade in the hopes a rising price in betting markets will turn into optimism and more real votes in Nov. The ROI on that doesn't seem very good.

2. Would the manipulation be coherent? Right now Obama trades at about $5.80 on Intrade, but contracts that the Dems will get 270+ electorial votes is at $6.4 and 280+ at $6.0 or so. That means one of these contracts has to be undervalued and others over to some degree....but the margin is probably not that great. If Romney spent $1M a day pumping up the Romney contract, would he also disperse money in the individual state markets to create a plausible electorial win map? If not investors could simply sell Romney contracts to Romney and buy up state contracts and take the 'gift'.

3. It may not be manipulation, it may simply be someone with lots of Obama contracts who wants to hedge that a bit by buying a big block of Romney. This tells me the market is not very deep, which is a bad thing, but at 41M contracts a day trading @ $10 that's a lot of money.

Good points.

Speaking of hedging on Intrade, why don't those businesses who claim an Obama (or Romney) victory will hurt them financially hedge their risks on Intrade? You could say, well, the market is too thin. But that logic cuts both ways: the thinness of the market is evidence that few businesses are hedging bets on Intrade.

There's also transaction costs, investment opportunity costs, and counterparty risk. Not to mention that the businesses most effected (USA) can't participate legally.

This is the important point. We can't legally bet here in the states so most of these bets are coming from overseas people who know nothing of what is going on here. Obama is in deep trouble and is behind substantially in major polls.

Last election, Intrade claimed that someone was hedging. This was the guy who lost a lot of money by trading large amounts once each day. Intrade talked to him to clear up accusations of manipulation.

Intrade is basically illegal for Americans to use. I expect that this binds corporations a lot more than individuals.

I think a whole lot of people would make a whole lot of money.

First, InTrade is illegal for a US citizen to buy into per the IRA.

Secondly, do you realize how small the InTrade market is? You can't hedge significantly on the 2012 US Presidential election. There aren't enough shares to buy or sell. At one point a few month past I looked and there were less than 100 'shares' for sell. Granted, you can make your own 'shares', but someone has to buy them.

Would Plumer have any proof that this is "manipulation", and not just a sweep of the book? The guy/gal bough roughly 4K contracts. Intrade does not display the whole book, but it probably gets much thinner away from the current price. Of course a conspiracy theory sounds juicier, but what evidence is there that the buyer does did not honestly want to put a certain dollar amount into the contract, and Obama supporters did not quickly swoop in to manipulate the price back down (the latter part of the question is in jest)?

A move after the first debate probably would have been more effective. He probably came up with the idea then.

Sheldon Adelson????

Those touting gold investment often say: 1) Gold is a good investment because its value is immutable. 2) The price of gold would be much higher, and will be one day soon, but manipulators are holding the price down right now. Maybe those two points aren't as contradictory as they seem.

Look at the idiot, he keeps pushing up the price, only to see it fall back down. What a sucker! Let's go short next time he does it and make some free money.

Intrade has investigated and concluded that it was not manipulation.

See comments:

Unless...this goes all the way to the top!

Exactly. Since when is a large purchase a market manipulation? Rule 10b-5 is malleable, but even it can't stretch to cover this. The purchaser would have to either simultaneously sell the Romney shares as he is buying (to give the impression of market movement) or be releasing some kind of false information aside from the purchase to create the impression that there is new information in the market.

Otherwise, the price of the shares go up when the purchaser buys, and then back down when the purchaser sells. There has to be something else going on to keep the shares up for the manipulator to profit.

Exactly. Since when is a large purchase a market manipulation?

It's not classic manipulation which is done for a profit motive. In that case the buyer does need some type of 'scam' to offset his large purchases such as false information that will allow him to then dump his share at an inflated price. But if the purpose of manipulation is to create the impression of Romney 'momentum' or simply an individual's arrogant desire to see himself move a multimillion dollar market just for the fun of it, then it may be termed manipulation even if not technically illegal.

What's kind of interesting is that if you could somehow achieve a public image of momentum and that happens to cause Romney to win then the cost of the manipulation turns into a profit as the shares will all go to $10 when Romney wins. But as I pointed out before if someone has millions of dollars they want to use to help Romney, one would think a lot of other tactics would have a lot more bang for the buck. It's a leap to conclude more will vote for Romney simply because betting markets improve his odds.

What IS manipulation is when the buyer is buying at a price much higher than the price at which people are currently selling. I see it happening on InTrade all the time, and the vast majority of the time, It's Romney's stock. There's no reason to over-spend on his shares like this, unless your primary goal is to make Romney look better.

If people are selling their shares at $4.00, and someone comes in out of nowhere and buys a ton of stock at $5.00, that looks like manipulation to me. Unless that guy likes spending an extra dollar per share for absolutely no reason.

I'm not sure that would be very 'rational manipulation'. If I wanted to drive up Romney shares and I had tens of thousand of dollars to play with, I'd start buying shares for around the market price. I'd rather buy more shares and gain control of a supply of Romney shares rather than blowing all my money by paying $1 or $2 extra per share. If nothing else that would allow my buying operation to last longer and maybe get others to join in so I don't have to spend as much. Otherwise I'll blow all my money really fast only to make a very quick 'blip' in the market that goes down so far it will be easy for others to write it off as someone playing games rather than 'market wisdom'.

So it sounds like a mini boom-bust. The market moved and other participants gathered that "smart money" had placed a bet and followed it. Markets usually move before the news so traders may have assumed they were following an insider ahead of breaking news. Or maybe there are technical traders on Intrade who follow momentum.

But if either of my above speculations are correct it would mean that the hard version of the EMH doesn't hold -- although the soft version does. This matters tremendously, because many academic studies and court cases (such as shareholder lawsuits) use the strong version of the EMH when determining what moved a market.

Obviously the strong version doesn't hold. Certainly not on inTrade. I used to snap up contracts on inTrade that differed substantially from Pinnacle prices. If prices normalized (which they often did, and quickly) I would sell out my position at a profit.

A smart intrade player would buy contracts that pay if Obama wins on Intrade and place bets in the UK that Romney wins. Look at the arbitrage opportunity!

That makes me doubt intrade as a real market. Seems too thin, even on such a large issue, to be able to arbitrage properly.

The basic problem is that even if you're "right" on intrade, if you have to put money in up front for your position and then sit on it for months, you will tend to make less money than you can get in less risky investment interest outside intrade. So people with serious money don't invest in intrade contracts.

Aren't all these prediction markets based in the UK (or Gibraltar or the Channel Islands)? I think they are technically illegal in the United States.

Buying Obama and Romney for < 100% is not an arb.

At one point I was looking at the 2012 US Presidential election and the combined Obama+Romney win was 94%. I'm not sure what the odds are that someone else could theoretically still win, but it's got to be a lot lower than 6%. However, there are enough transaction costs when dealing with InTrade and the market is so small, that it didn't strike me as too odd to see the variance.

It's not exactly months at this point. And you aren't getting very much risk free, so if there is truly arbitrage to be had here it's indeed surprising that no one in the UK has taken advantage.

It might be worth it if one traded with a deflating currency, gold, stocks, bonds, etc.

I would think that the inTrade with Obama as the favorite is suspicious. Romney has pulled decisively ahead and the possibility of a rout is real.

Not if you look at state by state polls, which is what matters for getting electoral votes. Romney has a solid lead in Florida, but is still behind in states like Ohio, Pennsylvania, Wisconsin, and Michigan. Virginia is a toss-up, certainly not a slam dunk for Romney.

His national numbers have closed b/c he's gained more support in states that he was going to win anyways (e.g. the Gallup poll had him way ahead in a national poll b/c he had a 22 point lead over Obama in the southeast).

Not saying that Obama's a lock - just saying that Intrade's odds are not nearly as suspicious as your claim that "Romney has pulled decisively ahead."

No candidate has won the popular vote by more than 1 % and lost the electoral vote. It looks like Romney is on track to win by more than 1%. Keep in mind that state polls trail national polls by 1-2 weeks.

The swing states are getting polled constantly right now, and with much better resolution than the national polls. Romney is behind. Only delusional partisans claim otherwise.

Whatever kook told you that the swing state polls are lagging by 2 weeks just before the election is... well... they weren't telling the truth, and you were incredibly foolish for believing them. But most partisans like to believe things that feel good, since the truth threatens your fragile identities.

Actually, InTrade is telling us that there is about a 43% chance that Romney is ahead. But YOU know that he is behind? No delusional partisan, you!

Think about it this way - If I tossed a coin 100 times and it came up heads 43 times, how certain would I be that the coin was not "fair"?

"No candidate has won..."

Ooh, ooh, care to put your money where your mouth is? Please?

Somebody's been the right-wing media spin a bit too seriously...

Does it matter if Obama loses Georgia by 55-45 or 65-35?

No. Romney gets 16 EV either way. National polls are simple for the rubes.

I don't think so. I think that the correlations between states are significant. If Georgia moves heavily for Romney it suggests that North Carolina and Virginia will move in that direction as well. Another example, as of today one poll showed Obama leading Ohio by 5%, one showed it tied, and another showed Obama by 3%. Obama carried Ohio last time by a little under 4.6%. Given that his poll numbers are down everywhere over 2008, what are the chances that Ohio is an outlier? Particularly given that Pennsylvania and Indiana and West Virginia have moved pretty dramatically toward Romney (not enough to win in Penn., but much tighter). So I'm supposed to believe Obama is +5% in Ohio? More likely tied.

It sounds like your hypothesis is if Georgia goes from pro-Romney to super-pro-Romney then some of that should spill over into Ohio. In other words as goes Georgia so goes Virgina, NC and OH. That might be true but it would require looking at past elections and not only observing whether Georgia more often than not lines up with VA, NC and OH but also does it lead or follow those states over time before elections? Suppose Georgia is behind Ohio rather than in front of it. In that case the Romney bounce from the first debate might have hit Ohio first and Georgia now....leading you to speculate that maybe Georgia is leading the way that will pull Ohio red rather than blue. But the opposite could also be true, Georgia is following Ohio's lead in becoming a bit more red.


Not at all. All I'm saying is that polls in one state are not irrelevant to the ultimate results in other states. Causality don't enter into it. Just correlation. A poll is is subject to random error and a certain amount of model error too (what, for example, constitutes a likely voter?). The polling companies tell you about the calculated random error, but not the model error (which I don't think they don't estimate). Since a state's results are correlated with national results (and not just because the individual state is included in the national result), the national results provide some information regarding the likely outcome of the state. Same for neighboring states, and to some extent even for states far removed. In this election the polls show that Obama will underperform his 2008 results with just about every demographic group and in just about every locale. So if someone does a poll of Ohio and it show him outperforming 2008, do we throw out everything else we know and simply adopt the polling result? Not if we are interested in getting good estimates.

Think about it another way. Suppose Candidate X is trouncing Candidate Y in 49 states, but we don't have a poll in the 50th state. Would you assume that the 50th state is a toss-up? Not if you are being rational you don't.

Remember that voting % != votes. Obama could do better in Ohio than he did last time in terms of percentages even if more Ohioians dislike him in 2012 than 2008. That could easily happen if overall turnout is lower or if turnout among pro-Obama people is a bit higher.

You're saying that Obama won Ohio by 4.6% last election so he should do worse now. But look at Nationwide Obama only won about 4% but KY and WV, both states that border Ohio, went for McCain by huge margins...nearly Texas size margins approaching 60%. Being super-Red didn't make Ohio red. Or if it did it didn't make it red enough to turn it it red overall.

"Or was somebody trying to influence Intrade odds in order to sway perceptions of the race? And if so, was it worth $1,250 to jolt the markets for less than 10 minutes?"

Considering that people often spend a couple of hours for a statistically non-existent chance of swaying the actual election, is the economics of this so bad?

IMO this goes back to ROI. Even $1250 can probably translate into a few additional votes in a swing state where it might matter. It's a leap of faith to think that spending $1250 to make a prediction market blip will create some type of peer pressure sensation that will cause a single vote to change.

It might be worth $1250 for someone to get an ego boost at being able to see their actions change a market. Wonder if the same type of thing happens in larger financial markets?

This really looks nothing like an attempt to manipulate the market for any sort of political reason. Who would be silly enough to think that a one shot bid of such a small volume would move the market for long? If a part of your election strategy was to convince the other side that the election was already over you would want to dedicate some much larger sum (probably millions) and then stack orders of various innocuous appearing sizes and see if you could withstand the onslaught before running out of money. Not spike the market all at once and give up.

Given how many people I have encountered in comments sections posting something like "Obama is down to 58% on Intrade! Romney is crushing him!" in apparent sincerity, I would not rule out silly enough.

Or people who are saying "Obama is at 57%! It's over! He won!"?

I'm an active participant on Intrade, and I have to say that I LOVE it when this happens...although I'm disappointed to have missed this one. Last time around (in 2008) there was a lot of discussion on the Intrade forums about a mysterious deep-pocketed McCain buyer (it always seems to be the GOP partisans who are price insensitive!) who would come along every couple of days and hit the ask up 5% or so. A pretty profitable strategy was to place limit orders 3-5 pts off the market price and just wait for the buyer to come along...then take a profit a few minutes later or hold to expiry, depending on your political assessment.

This likely was multiple individuals making big bets on the race as their checks cleared (it can take a couple weeks to fund your account). This time it was probably someone willing to pay up to 50/50 odds and not really caring about the trading frictions. Or someone new to financial markets who didn't understand the impact they'd have.

But amusingly as more people hear of Intrade in the run-up to the elections you start seeing a lot of these, and it's funny how with the press picking up on it the price action can sometimes become a bit self-reinforcing.

Anyway, I've already tied up all my margin. Disclosure: long 2012.Pres.Obama. Also long Romney.Tax.Ret.5Nov with 20-1 and feeling good gambling on 20-1 odds.

I suspect it was "manipulation" but not in the sense many think. In all financial markets some traders will try to 'bully' the market into one direction or another, especially surrounding fundamentally important events that might lead to volatile expectations. Intimidating other speculators into dumping their positions is a tried and true technique. Markets that have some liquidity but far from perfect liquidity are perfect for this. I doubt this person was trying to manipulate Americas expectations about the election. Maybe he or she was just trying to make money.

I watched the debate last night, and then checked intrade. I was suprised and it left me wondering if someone had some inside information that I didn't . Thanks MR for clearing up the confusion.

It seems to have settled back at that level now. Obama at 57.1 when I just checked at Real Clear Politics.

We checked this out for potential manipulation - it certainly fit the pattern at first glance.

However, during the period you reference above 40 individuals bought Romney shares and no individual bought more than 15% of the shares traded during that period. A look at those 40 accounts shows nothing suspicious.

So we don't believe this was manipulation but more a run on Romney at a moment when the market was unusually thin.

Carl Wolfenden
Exchange Operations Manager
Intrade - The Prediction Market

By looking at the "fundamentals," such that they are, I have reached a similar yet tentative conclusion, as detailed on my own blog.

The daily Rasmussen presidential tracking poll gets released at 9:30 AM and the daily Rasmussen swing state poll gets released at 10:00 AM (I think). The presidential tracking poll came out at Obama/Romney at 46/50% and the swing state poll came out a 45/50% both bullish for Romney/bearish for Obama and probably took some people by surprise. These polls combined with the thin markets at the time probably led to the short term volatility. This is similar to what happens in the financial markets during an economic release (e.g. non-farm payrolls).

It looks like the manipulation was not a total failure. Obama's numbers are further down today than they have been on days with polling much worse for Obama. A small market like Intrade has a limited capability to dissipate this kind of maneuver.

This reminds me of a parliamentary byelection in the UK in 1973 in the Isle of Ely . The Liberal candidate was a well known writer and broadcaster Clement Freud . The Liberals had not even fought the seat in the previous election and the bookmakers priced his chances at 33/1 . In his autobiography , he tells how he made several bets to win £ 10,000 for £ 300 stake and the price dropped rapidly to 8/1 . The Daily Telegraph commented , that the smart money seems to be going on the Liberals . Freud says in fact it was the Freud money that was going on Freud . He won the election by a narrow margin of 3%

This morning it's 54.6-45.3... another manipulation, or perhaps the fact all but one national polls giving Romney ahead has something to do with it?

Could'nt be! For a poll to be any good it must show Obama winning. Same with InTrade. If it shows Romney moving up it might be manipulation. If it shows Obama moving up it's just market efficiency.

Simple, really.

Perhaps Alex should do a further update. Intrade, Betfair, and the Iowa market have all moved towards Romney. Perhaps someone is manipulating all 3 markets? Or maybe there are more reasonable explanations - bad news for Obama today (e.g., Benghazi email), a fat finger or some other new trader not familiar with how Intrade works, a person that wanted to put a large bet out quickly, a person that was very confident in their prediction, some combination of the above, etc.

It may have been more prudent for Alex to say there is possible manipulation, but also give a balanced discussion of what could or couldn't otherwise cause sharp movements in Intrade prices.

There were rumors at that time of that story, so don't think manipulation at all.

But a manipulator sells and buys based on reasons other than expectations and so offers other investors a greater than normal return.

That is my explanation of why I have been able to make money buying put on green firms. Some green investors look at investing green as charity.

Pepito has it right, this was an overzealous reaction to the pending Trump announcement by a very naive trader (who could have bought at 43 instead of 48 by shorting Obama instead), not attempted manipulation:

Is Intrade legal or illegal in the U.S.? I know funding it with a credit or debit card issued by a U.S. bank is not legal, but can you send a check to fund an Intrade account?

Pepito, I am totally agree with your thoughts. Keep doing these type of work.

Comments for this post are closed