News

Five Questions for Thomas Graeber

What are the topics you are currently working on?

Be­hav­ioral eco­nom­ics has pro­duced a kind of “bias zoo”. By that I mean we’ve found hun­dreds of ways peo­ple make de­ci­sions that aren’t per­fect­ly ra­tio­nal, and most of them are stud­ied sep­a­rate­ly. My work, rather than adding to this list, ex­plores whether deep­er prin­ci­ples might joint­ly ex­plain many of these be­hav­iors. One re­cur­ring puz­zle across do­mains is that peo­ple’s de­ci­sions tend to be much less re­spon­sive to changes in key pa­ra­me­ters than eco­nom­ic mod­els pre­dict. For ex­am­ple, val­u­a­tions of lot­ter­ies don’t shift near­ly as much as the­o­ry sug­gests when prob­a­bil­i­ties change from 60% to 70% to 80%. Sim­i­lar­ly, peo­ple’s will­ing­ness to pay now for a de­layed pay­ment changes re­mark­ably lit­tle whether the de­lay is three, four, or five years. Tra­di­tion­al­ly, these and many sim­i­lar-look­ing pat­terns have been treat­ed as sep­a­rate phe­nom­e­na with dis­tinct ex­pla­na­tions. In a re­cent pa­per, we show that in each of these de­ci­sions, most peo­ple are un­cer­tain about what their best choice real­ly is. Think of how much you’re will­ing to pay for a lot­tery tick­et that pays 1,000 francs with a 70% chance. Maybe you say 600 francs, but are you sure it’s ex­act­ly 600? Not 550, or 620? We ar­gue that this kind of “cog­ni­tive” un­cer­tain­ty can help ex­plain why peo­ple’s choic­es are of­ten less sen­si­tive to changes in prob­a­bil­i­ties, de­lays, and oth­er de­ci­sion pa­ra­me­ters. If you’re un­sure what’s best, you end up treat­ing near­by sit­u­a­tions too much alike.

What is the main finding of this work?

Across more than 30 ex­per­i­men­tal set­tings – span­ning de­ci­sions un­der risk, over time, in proso­cial con­texts, fi­nan­cial choic­es, be­lief for­ma­tion, fore­cast­ing, strate­gic games, and many oth­ers – we find that cog­ni­tive un­cer­tain­ty about what’s best strong­ly pre­dicts how re­spon­sive peo­ple are to key de­ci­sion pa­ra­me­ters. The main take­away is that many seem­ing­ly dif­fer­ent be­hav­iors may be re-in­ter­pret­ed through a com­mon lens: how peo­ple sys­tem­at­i­cal­ly re­spond to the feel­ing of be­ing un­sure of the right course of ac­tion.

Is there another important specific work of yours that you would like to mention?

Along­side the pat­tern of peo­ple be­ing less re­spon­sive than stan­dard the­o­ry pre­dicts to key de­ci­sion pa­ra­me­ters, there is a broad­er pat­tern of re­spon­sive­ness to spe­cif­ic fac­tors that the mod­el of homo eco­nom­i­cus would not pre­dict. In par­tic­u­lar, ex­ten­sive ev­i­dence high­lights the role of com­par­i­son points such as goals, an­chors, ref­er­ence points, norms, ex­pec­ta­tions, and so­cial com­par­isons. In on­go­ing work, we ex­plore the idea that re­liance on these com­par­i­son points may, sim­i­lar­ly, part­ly re­flect un­cer­tain­ty about what is op­ti­mal. This per­spec­tive also helps ex­plain the mixed ev­i­dence on when com­par­i­son points lead de­ci­sions to shift up­ward ver­sus down­ward.

You have recently joined UZH, what motivated you to come to Zurich?

UZH has an ex­cep­tion­al group of peo­ple work­ing in my field. I strong­ly be­lieve that re­search is a col­lec­tive en­deav­or: it is not pro­duced alone in a room, but through con­stant in­ter­ac­tion with col­leagues and stu­dents. Along­side the peo­ple, the re­search in­fra­struc­ture is also cru­cial for work like mine. Few in­sti­tu­tions world­wide match the in­fra­struc­ture of lead­ing U.S. uni­ver­si­ties, UZH is def­i­nite­ly one of them.

Who is the inspiration behind your work?

I wouldn’t say I had a sin­gle guid­ing in­spi­ra­tion from the start, but over time I’ve come to ap­pre­ci­ate how much of what I work on res­onates with ideas that Her­bert Si­mon ex­plored decades ago. For ex­am­ple, most re­cent­ly, when I be­gan work­ing on how peo­ple de­com­pose com­plex eco­nom­ic prob­lems, I re­al­ized how di­rect­ly it builds on Si­mon’s re­flec­tions about cog­ni­tion and de­com­pos­abil­i­ty.

Back
Twitter logo