2 Nov 2009

The Political Economy of Lobbying

Last week, I taught my students about different auction techniques. After I auctioned four books under different rules, I auctioned $1.00 for $3.75.

Now before you say "WTF?" (or perhaps I am too late!), let me explain.

First, it was an all-pay auction, which requires that all bidders pay their highest bid to the auctioneer, even if they do not "win."

Second, watch this video of the auction (4 minutes! drama! laughter!):

Third, what happened? There were four bidders. Mr. A bid $0.25; Ms. B bid $0.50; Ms. C bid $0.75; Mr. D bid $1.00. Then Ms. B raised her bid to $1.25 before Mr. D won the auction with a $1.50 bid. Mr. D got $1.00, and I got $3.75 ($2.75 net).

Nice money maker, that all-in-auction :), which is why you should see five, below.

Fourth, why did this happen? Consider this:
  1. If you bid $0.25 for $1.00, you stand to make $0.75.
  2. If someone raises you (to $0.50), you lose that $0.25.
  3. If you bid $0.75, then you still make (net) $0.25.
This set of incentives ("on the margin") makes it rational to keep raising the stakes given that you have bid because winning always has a higher payoff than losing. The "right strategy" is, of course to not bid at all. (As Joshua, the computer, says in Wargames: "Strange game. The only winning move is not to play.")

Fifth, all-pay auctions reproduce the dynamics of political lobbying in which the politician is auctioning the wording to some law, and lobbyists from both (many?) sides are contributing money, perks and attention to get their version of the law. All of the lobbyists pay, but only the politician wins.

Oh, and we lose as well. A politician who represented "the People" would write a law that maximized the positive change in social welfare, regardless of lobbying.

Bottom Line: Politicians benefit from lobbying; lobbyists compete to receive our money and rights; and citizens suffer. [I could soften this language to "some politicians" or "sometimes," but I these are structural institutionalized flaws.]


Jay said...

Nice lesson.

Bottom Line: Market mechanisms matter, and political markets are just another type of market.

Anonymous said...

Does your university IRB know about this?

David Zetland said...

@Anon -- It was a classroom demonstration, which I am happy to say is NOT covered by IRB.

Stephan said...

"The "right strategy" is, of course to not bid at all."

Not quite. The right strategy is to bid $0.01, hit your head on the table and yell out "as you can see, I am crazy, and if one of you mofos goes above $0.01 I will increase my bid straight to $1 and you are going down, suckers." Auction over, and you lose $0.99. The video would be shorter, but even more amusing.

michael webster said...

I like you trick of increasing the bids, and then telling people sunk costs don't matter which ratchets up the bids.

However, I doubt that this is an example of lobbyists throwing money at politicians. It is far more likely that the a group of lobbyists forms a coalition, writes the bill it wants, and doesn't pay enough for it!

As for not playing the game, it depends upon the auction mechanics, and the sophistication of the audience.

Stephan has an interesting approach, and I have a different observation written here:


Nice video, but why are all those people laughing?

Anonymous said...

Politicians benefit from lobbying; lobbyists compete to receive our money and rights; and citizens suffer.

Except that almost all citizens belong to special interest groups, which are represented by advocates, i.e., lobbyists.

See the 1st Amendment to U.S. Constitution, re: redress of grievance.

Anonymous said...

Anonymous said:
"Except that almost all citizens belong to special interest groups, which are represented by advocates, i.e., lobbyists."

True. But that doesn't mean it all comes out equal in the end. See Russ Roberts' great piece, "If You're Paying, I'll Have Top Sirloin," available here:

David said...

This also works well as a open cry, second bid, auction. (top bidder pays the 2nd highest bid for the "prize.")

By well, I mean the teaching points, not the IRR for the auctioneer. :-)

HoBs said...

I actually use this example in class too. Though it is not so simple.

The right strategy is NOT to not bid at all. Think about it, that can't be a Nash Equilibrium. Since if everyone else is playing that, then my optimal strategy is to bid. The optimal strategy is to bid once with some small probability and then stop.

Also, I realize in a classroom setting, people will bid just for fun, or on inducement by the professor. But actual people are closer to rational.

If you look at winning prices at Swoopo, a version of ebay that actually implements an All Pay Auction, you find that the final reveneues are not too far off from "fair" prices. Which means that revenue equivalence holds, which means that people actually bid surprisingly rationally in such situations.

Anonymous said...

Bidding $0.25 is not a "sunk cost" because a dollar is worth a dollar. A "sunk cost" would be the cost invested in an outcome much larger than the one-off auction, such as the entertainment value of playing such a silly game. So if the opportunity cost is, say, a $2 video on iTunes then perhaps you could bid up to that and not "lose."

Kevin said...

In the context of this game, there is one winning move: be the first to bid, and bid $0.99.

You stand to earn $0.01, making you a profit. For any other potential bidder, they can, at best, bid $1.00, breaking even, and risk losing that investment if someone else outbids them. Therefore, it is not rational for anyone to beat your bid.

Of course, there is the possibility of a jerk bidder, who is willing to spend money to ensure you don't earn any, but if one of your opponents is willing to spend X dollars to prevent your profit of X/100 dollars, he's likely out of your league.

Josh said...

That's interesting, but I don't see how lobbyists lose out.

If a bill allows for your business costs to decrease by, say, $100 million, annually, then you should be willing to pay a certain percentage of that to help guarantee its passage. Even if you pay $200 million, you get that back within two years.

Second, if you DON'T lobby, then your interests can be trampled. So, it's more akin to insurance.

The best, of course, would be to allow for equity of speech, by limiting donations.

David Zetland said...

Thanks for your comments folks!

@Stephan -- unless there's another crazy mofo there.

@mw -- no it's not a trick. I reminded them of sunk costs. Lobbyists, by defn (=special interest group), do not form coalitions.

@Anon(10:08) -- wow, you need to catch up on reality. Start with the ethanol program.

@Anon(10:34) -- totally.

@David -- I did 2nd price earlier in the class, but that's not what all-pay is about.

@HoBs -- (speaking of ethanol). Your NEq idea is falsified by the class AND by lobbying. Further, bidding $0.01 will only "work" if nobody else bids. Since you would bid more ($0.98 profit), you get the result (out of eq, if you insist) of paying more than $1. Come on Ben, you can do better than that.

@Anon(14:42) It's sunk as soon as it's bid. That's b/c sunk costs CANNOT be recovered. Even a new, higher bid merely adds to sunk costs.

@Kevin -- THAT's a good strategy. Unfortunately, people tend to get greedy and lowball, and THEN lose to someone else who gets in the game.

@Josh -- if lobbyists are NOT in a cartel (i.e., as is normal), then some will win and some will lose. Pols win either way. Of course, there are examples of unity (ethanol, defense) where citizens lose. So, yes, this example doesn't apply ALL the time.

HoBs said...

I can do better. The exact equilibrium is going to depend on the bidding rules and the bidding increment, but the Nash Equilibrium can't be bid nothing. This comes from discussions with classmates who are really good game theorists, one of whom is at MIT and the other of who is at Northwestern whose dissertation was on All Pay Auctions.

Again, think of the definition of nash. if everyone does nothing, then i certainly should bid.

But what about a strategy where I bid with probability 5%. Then it could be optimal to bid if everyone played this strategy, because there is some probability that no one outbids me.

Also, I've run this in class 4 or 5 times now, and I realize that classes will bid because they figure out you want them to (you can see it in your video, you were encouraging them). This is why lab experiments are ideally conducted by RAs who don't know what the experiment is about (double blind in effect).

Also, we've debating lobbying before, it is something I've studied. Baron has a paper that models lobbying as effectively an all pay auction to explain why there is so little money in US politics. This is not just what I think, this is what the literature says.

Anonymous said...

Your definition of "sunk cost" as simply non-recoverable is incomplete. As I said, your auctioned dollar is in fact worth a dollar. The cost of bidding on it is the cost of something else, not the dollar itself. You are employing a rhetorical trick when you do not reveal what that sunk cost is actually paying for (what I am calling the "entertainment value").

Anonymous said...

Mr. D here, Kevin's description of the "jerk bidder" basically describes my motivation in this experiment. $0.50 was a small price to pay for the amusement of other people losing more money than me.

David Zetland said...

@HoBs -- As you know, I am not a fan of lit, and certainly not theory (which is just ex post rationalization dressed up with math). I'd say that there is NO NE in most experiments with strategy (like all-pay, or trust game, etc.) and that's b/c people are not math models.

I doubt that the students were "doing what I want," since only 4 of 60+ bid, and -- as Mr D says just above -- some wanted to screw around.

Finally, this auction *does* reproduce "harmful competition" -- read THAT lit :)

@Anon -- you're saying that money spent on one thing has an opp cost in terms of another thing? Great. And? Try a different tack: compare this auction to a regular auction (pay if win) and then tell me that there's no difference.

@Mr D -- good job. I don't care what your motivation was -- you helped make a good point (and only paid $0.50 to do it :)

Eric said...

Could someone talk about the multi-auction strategy, where this particular auction is an example of an auction that occurs once a week and the bidder is involved in the auction every week for two years. Some of the auctions have much more worth to the bidder than others do.

What might be an optimal strategy for a particular bidder then? In addition, consider that certain bidders, cannot afford to lose more than a certain amount over the two year time scale.


HoBs said...

actually, in point of fact, people are math models, in that people are made of particles, and thus in principal there is an equation that describes their behavior in the same way Newton's laws describe the behavior of all physical systems.

or less snarkily, the assumptions required for nash equilibrium are pretty weak. whether we correctly model what the utility function looks like is a different question, but so long as people have preferences, there is a Nash equilbrium.

but sure, there are various reasons why this demonstration lacks external validity, one is the jerk bidder, but you were clearly prodding people to bid.

that's the benefit of video tape. that should be standard experimental practice actually.

reminds me of the milgrom experiment videos.

David Zetland said...

@Eric -- Interesting question. Without explaining every aspect (b/c I can't :), I'd say that reputation (is this person's threat to outbid me real?) would matter. A limited budget would be hard to "optimize" without a fwd view of what's for sale over time AND how time changes things. That's why the stock market is so hard to "game."

@HoBs -- yes, I agree that people have preferences but those are "known unknowns" that can neither be modeled ex ante nor disaggregated ex-post.

Sam Rothstein said...

The problem with all-pay auctions is that they do not have a pure strategy NEq. When the rest of the bidders bid anything less than the value of the object, your best response is to match their bid plus something if that bid still remains below the object´s value. That´s your best response (even in an english auction). The all-pay feature only adds to the phenomena of "over-bidding", since the all-pay feature transforms the value of the object into its intrinsic value plus your highest bid.

But there must be a Nash Equilibrium in mixed strategies if the set of strategies is finite. Here I found a nice paper that summarizes the main points.



David Zetland said...

@Sam -- thanks! HoBs can comment on that paper :)

dcp said...

This is just another instance of applying highly contrived stories to a much more complicated reality. First, it perpetuates the confusion between lobbying and contributing. Lobbying can be defensive (trying to prevent or minimize the negative effects of some rule or regulatory change that harms the firm unnecessarily) or predatory (trying to use the political system to hurt rivals). One benevolent (and observationally equivalent) view of lobbying is that it simply reflects the cost of information. If I ask you to analyze the benefits/cost of legislation x, it would involve a lengthy process and a hefty consulting fee. But, in most cases there will not be one answer. Why should I trust yours? So, the political process lets each party propose their version of costs/benefits - which is an expensive proposition. Which brings me to your other assumption: that there is ONE item being auctioned off. At the end of the day, the information providers (Lobbyists - not auctioneers) walk away with the dollar. In reality, (I know a far cry from game theory) firms lobby for many reasons. Some are 'winners' in that the legislation gets changed reflecting the poorly worded parts; some are loosers in that the legislation might not get changed. Perhaps all are winners. There is no way to tell. Finally, even with contributions (note the distinction), 'winners' are hard to define. Campaign coffers bulge, but these can only be used for political activities (i.e., work). A very strange form of utility. In the end it is, at best, a 'cute' but largely irrelevant story. At worst, it is an unethical way for the professor to take money from students...

Anonymous said...

dcp, there is no practical difference between lobbying and contributing unless you are dependent on process-point distinctions that make no end-result difference.

When I compare my ability to influence my Congresscritters against Blue Cross/Blue Shield's ability, it matters not whether BC/BS is just throwing $$$, or helping write legislation. In either case, my Congresscritter isn't listening to me -- he's listening to BC/BS. The only way I can beat that situation is to have as much influence as BC/BS, and that influence exists because of BC/BS's financial leverage equating with political leverage. In the eyes of the Congresscritter, "jobs" and "economic growth" promised by BC/BS are more important than me being entitled to public health care. My influence is foreclosed from the start. It matters not which point in the chain of influence we're examining -- BC/BS wins every time.

Your argument is very similar to the process-maven, technocratic, meritocratic sort of analysis that displays knowledge of details, but a lack of discussion on the relevance of those details.

Post a Comment

Note: only a member of this blog may post a comment.