Mandating Information Disclosure vs. Banning Deceptive Contract Terms

Economists are very into the idea of mu­tu­ally benefi­cial ex­change. The stan­dard ar­gu­ment is that if two par­ties vol­un­tar­ily agree to a deal, then they must be bet­ter off with the deal than with­out it, oth­er­wise they wouldn’t have agreed. And if the terms of that deal don’t harm any third par­ties,* then the deal must be welfare-im­prov­ing, and any reg­u­la­tory re­stric­tions on mak­ing it must be bad.

One ob­jec­tion to this ar­gu­ment is that it’s not always clear what is and what is not “vol­un­tary.” I once has a well-pub­lished economist friend ar­gue that there are no gra­da­tions of vol­un­tari­ness: ei­ther a deal was made un­der some kind of com­pul­sion or it wasn’t. I asked him if he would be OK let­ting his then pre-ado­les­cent son make any school­yard deal he wanted as long as it was not made un­der any overt threat, and I think (but am not to­tally sure) that he has since backed off this po­si­tion. So there is an ar­gu­ment for purely pa­ter­nal­is­tic re­stric­tions on free­dom of con­tract.

Another ob­jec­tion, one which economists tend to take more se­ri­ously, re­lates to in­for­ma­tion. Speci­fi­cally, there is the idea that maybe one party to the con­tract is not fully in­formed about its terms. For this rea­son, many economists are will­ing to en­ter­tain poli­cies by which firms are re­quired to dis­close cer­tain in­for­ma­tion, and to do so in a way that is com­pre­hen­si­ble to con­sumers. So for ex­am­ple we now have “Schumer boxes” that gov­ern the ways in which credit card com­pa­nies pre­sent cer­tain in­for­ma­tion in pro­mo­tional ma­te­ri­als. This seems to many peo­ple to be a rea­son­able rem­edy: if the prob­lem was that one side of the trans­ac­tion was ig­no­rant, then a reg­u­la­tion that elimi­nates that ig­no­rance, while at the same time not in­terfer­ing with their free­dom to en­gage in mu­tu­ally benefi­cial ex­change, must be a good thing.

I think this rea­son­able-sound­ing po­si­tion is largely wrong. The stan­dard asym­met­ric in­for­ma­tion sto­ries with ra­tio­nal agents are sto­ries in which the uniformed party knows that it is un­in­formed, which in­fluences the con­tract terms that it is will­ing to ac­cept, which in turn ei­ther causes benefi­cial ex­change not to hap­pen (the “lemons” prob­lem) or causes con­tract terms to be dis­torted away from the effi­cient ones. They are gen­er­ally not sto­ries about un­in­formed con­sumers not un­der­stand­ing that they are un­in­formed and blithely march­ing into traps as a re­sult. But this is what we ac­tu­ally see all the time, one party tricks the other party into un­fa­vor­able terms. In­deed, very of­ten this is the real-world prob­lem to which pro­vid­ing bet­ter in­for­ma­tion is sup­posed to be the solu­tion! But for trick­ery to be the prob­lem, you usu­ally need a model in which some agents suffer from some limi­ta­tion on their ra­tio­nal­ity, such as my­opia.** And if you have that, then you have a differ­ent prob­lem from the prob­lem of asym­met­ric in­for­ma­tion, and there is no par­tic­u­lar rea­son for a differ­ent prob­lem to have the same solu­tion. If the prob­lem is that peo­ple are get­ting tricked, then pro­vid­ing more in­for­ma­tion is only go­ing to help if it is go­ing to cause them not to be tricked, and it is not at all ob­vi­ous whether and when this will be the case.

But there is a big­ger prob­lem with the stan­dard way that economists usu­ally think about these prob­lems, which is that they com­pletely ig­nore the fact that when peo­ple are be­ing tricked, the virtues of vol­un­tary ex­change are ab­sent and so there is no rea­son for a strong pre­sump­tion against in­terfer­ing with it in the first place. And some­times the very ex­is­tence of cer­tain con­tract terms is an in­di­ca­tion that the con­tract is a trick. Think about the con­tro­ver­sial terms of­ten found in credit card con­tracts, such as pro­vi­sions by which be­ing one day late with a pay­ment or be­ing one dol­lar over your credit limit jumps your in­ter­est rate to 29.99% for­ever, or in which cards with mul­ti­ple bal­ances at differ­ent in­ter­est rates pay off the low rate bal­ance first. What should be in­ferred from the fact that these terms ex­ist? Is it at all plau­si­ble that there is some sub­tle but very im­por­tant rea­son why these terms must be pre­sent, and that if they were banned lots of mu­tu­ally benefi­cial deals would not be made? Is it not much more likely that that these terms ex­ist pre­cisely be­cause many con­sumers don’t un­der­stand them and will be tricked by them? Have you ever heard of such terms be­ing in con­tracts ne­go­ti­ated be­tween so­phis­ti­cated par­ties? Shouldn’t this cause you to be much less wor­ried about the con­se­quences of sim­ply ban­ning them?

There is a very good pa­per by Gabaix & Laib­son (2006) that pro­vides a for­mal model in which firms “shroud” rele­vant in­for­ma­tion in or­der to trick my­opic agents. The neat thing about their pa­per is that they show that this per­sists in equil­ibrium: com­pet­ing firms turn out to have no in­cen­tive to march in and ex­pose the shroud­ing and offer trans­par­ent pric­ing in­stead. But you don’t need a fancy (and re­cent) pa­per to have known that the afore­men­tioned terms in credit card con­tracts are only there to trick peo­ple. And if that’s true, then what you re­ally want is to get rid of con­tracts with those terms. Man­dat­ing in­for­ma­tion dis­clo­sure is only a good rem­edy in­so­far as it causes those terms to dis­ap­pear. Gone is the economist’s no­tion that the right solu­tion is to make sure ev­ery­one knows the score and then to step out of the way. If you man­dated dis­clo­sure and then saw those con­tracts con­tin­u­ing to ex­ist, the con­clu­sion you should draw was that the dis­clo­sure was in­effec­tive, not that the terms were effi­cient.

One could ob­ject to heavy-handed reg­u­la­tion on the ba­sis of a slip­pery-slope ar­gu­ment. While there are some clear-cut cases like the credit card con­tracts, a gov­ern­ment with lots of reg­u­la­tory power, even a well-in­ten­tioned one, may end up get­ting overzeal­ous and in­terfer­ing in ways that will have un­in­tended and nega­tive con­se­quences for effi­ciency. And Gabaix & Laib­son take pains to point out that they are not ad­vo­cat­ing lots of reg­u­la­tion. What­ever the mer­its of this ar­gu­ment (I un­der­stand the fear of reg­u­la­tory over­reach but worry about it less than a lot of other peo­ple do), the main point of this post re­mains. There are im­por­tant in­stances in the world we live in where the un­aware sim­ply get tricked and screwed. That is not, at root, a prob­lem of asym­met­ric in­for­ma­tion among ra­tio­nal agents, and there is no rea­son to think that the ap­pro­pri­ate rem­edy is the same as if it were. More im­por­tantly, in this world the virtues of vol­un­tary ex­change are ab­sent, and so the economists’ defer­ence to it is mis­placed.

There has been an im­por­tant real-world de­vel­op­ment on this front. It seems that the Fed­eral Re­serve did a bunch of con­sumer test­ing to see how well peo­ple un­der­stood var­i­ous terms in credit card con­tracts un­der differ­ent dis­clo­sure re­quire­ments, and con­cluded that they were sim­ply too com­pli­cated for most peo­ple to un­der­stand. They there­fore de­cided to go ahead and sim­ply pro­hibit cer­tain prac­tices. Which I say is good news for the good guys.

*A weaker ver­sion of this con­di­tion is that any third par­ties that are hurt are hurt less than the con­tract­ing par­ties are helped.

**I say “usu­ally” be­cause there are a few spe­cial mod­els in which fully ra­tio­nal agents can nev­er­the­less be tricked.