Bryan Caplan  

Johnson Bet

PRINT
Why the new old Keynesians are... Power corrupts......
I have bet Daniel Filan $100 at even odds that Gary Johnson will receive less than or equal to 5% of the 2016 U.S. Presidential popular vote. I hope to lose, since I (a) consider Johnson far less bad than either Clinton or Trump, and (b) think a high Johnson share reduces the risk that the GOP goes even further down the path of anti-foreign bias.  But since the odds are even, I obviously expect to win.

Filan has prepaid me, so if I lose I will pay him a gross of $200.

Update: In the comments, Matthew Moore writes:

What private / different information are you using here, Bryan? Pretty much every model and market I can find is substantially over the 5% mark, e.g.:

538's three models have Johnson at 7.3, 8.7 and 5.7

Ladbrokes are offering 150/1 for Johnson to *win*.

Pivit.io predictions net out at 95% chance of >5%

Why are you so relatively pessimistic? What information is missing from these markets?

Answer:

I discount the polls because several insiders have confessed to me that Libertarian Party candidates do much better in polls than actual elections. 

I discount the betting markets because extremist bettors seem to noticeably tilt the odds for extreme long-shots.  This is stable due to moderate transaction costs and thin markets: If Johnson's expected vote share is 3%, a market prediction of 6% is not a major money-making opportunity.

Update #2: Bill Friedman has made the same bet with me.


Comments and Sharing






COMMENTS (19 to date)

Will you vote in order to increase your chances of winning?

Cheers!

Jason writes:

I'd say you got favorable odds on that one.

Kitty_T writes:

What odds might you offer that Johnson will receive electoral college votes?

Khodge writes:

Bryan, bad bet. Ross Perot got 18% of the vote in an election where each candidate was (possibly) ideologically flawed but not personally compromised. Surely, with that precedent, Johnson can do better than 5% with such personally and ideologically flawed candidates that we currently have.

Bill Friedman writes:

I'll take that bet, too, if you're offering. I would be shocked if he got as little as 5% of the vote, though I obviously think he has less than a 5% chance of winning the election.

Bill

Jonathan writes:

I think Steven Landsburg argues to bet against your favorite team in the Armchair Economist. It's a simple argument, you pay to see your favorite team win, you make money to see your favorite team lose.

Unless both Trump and Clinton get extremely ill, die and/or go to jail, this argument doesn't work perfectly in this case. However, is it safe to say that greater than 5 percent of the vote would be a win for Libertarians?

Also, I remember Bryan posting his betting record in the past. Do any of his past bets reflect this line of thinking?

Daniel Filan writes:

For the record, I bet based on fivethirtyeight's polls-plus forecast, whose median projection has Johnson getting 5.5% of the vote. My probability of winning this bet is around 65% - less than certain, but more than even odds.

Charlie Hughes writes:

Would you have been less confident if it was the combined vote of Stein and Johnson (and new entrant Evan McMullin), or does that not move the needle for you?

Floccina writes:

If he gets over 3% I will be happy. It is interesting that Americans love to talk about freedom but want government to control so much.

Matthew Moore writes:

What private / different information are you using here, Bryan? Pretty much every model and market I can find is substantially over the 5% mark, e.g.:

538's three models have Johnson at 7.3, 8.7 and 5.7

Ladbrokes are offering 150/1 for Johnson to *win*.

Pivit.io predictions net out at 95% chance of >5%

Why are you so relatively pessimistic? What information is missing from these markets?

Gordon Blewis writes:

538's polls-only model forecasts 7.3% for Johnson. The polls-plus model forecasts 5.7%.
http://projects.fivethirtyeight.com/2016-election-forecast

[broken link fixed--Econlib Ed.]

Matthew Moore writes:

What private / different information are you using here, Bryan? Pretty much every model and market I can find is substantially over the 5% mark, e.g.:

538's three models have Johnson at 7.3, 8.7 and 5.7

Ladbrokes are offering 150/1 for Johnson to *win*.

Pivit.io predictions net out at 95% chance of >5%

Why are you so relatively pessimistic? What information is missing from these markets?

Matthew Moore writes:

Thanks for the above-the-fold response Bryan.
Your answer makes sense to me. It's the same in the UK: people feel comfortable to take a free hit against the establishment in surveys, but not in the polling booth. That said, 538 etc should try to account for this.

Now that I think about it, it might be somewhat rational for voters to act this way, since voting for fringe candidates in surveys marginally pressures other mainstream candidates to change their public positions, whereas the actual vote has no such effect, and only the tiny marginal probability of selecting from among the existing platforms. Although candidates knowing this discounts the polling pressure, unless it reveals voters' private information. Might be cool for someone to model as a two-stage game. I personally adopt extreme positions in attitude surveys in order to drag the average in my preferred direction.

Rimfax writes:

History is on your side:

* 2012 - Johnson peaked at 5.3%. Final 1%
* 2008 - Barr peaked at 8%. Final: 0.4%
* 2004 - Badnarik peaked at 1.5%. Final: 0.3%
* 2000 - Browne peaked at 1.1%. Final 0.4%
* 1996 - Browne ???. Final 0.5%

Nathan L Ashby writes:

[Comment removed pending confirmation of email address. Email the webmaster@econlib.org to request restoring this comment. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Scott Sumner writes:

FWIW, I predicted 4% a few weeks back---and only that high because Trump is an unusual candidate (and Hillary's not very popular either.)

BZ writes:

As a dedicated L Party muck-a-muck, Johnson delegate, party history buff, and so forth, I have to agree with Dr. Caplan. Rimfax stole my thunder by posting the numbers, but in the end it takes a great deal of emotional fervor for a candidate to overcome the Lesser of Two Evils reasoning, something which Perot approached, but our candidates generally don't, and Johnson will not.

We are seeing a lot of signaling in the polling right now that goes our way. I love it! It helps motivate the press to put our ideas in front of voters at a time when voters are thinking about politics, which is precisely what I think the LP does best. However, when people enter the booth later this year, the less personally indulgent reasoning that Dr. Caplan has mentioned elsewhere will take over, and people will vote for the R or D they believe to be less destructive.

All this said, it will be a record year for the LP and I am as happy as a pig in a poke over every bit of it.

J. Goard writes:

"I discount the polls because several insiders have confessed to me that Libertarian Party candidates do much better in polls than actual elections."

By "insiders", do you mean Wikipedia? It's rather hard to think of any more public, easily accessible information!

Anton Maier writes:

Although five thirty eight forecasts something different, i'd say you're right.

Comments for this entry have been closed
Return to top