Is the Sunk Cost Fallacy Actually Smart Business?
Skip to content
Economics Strategy May 5, 2013

Is the Sunk Cost Fal­la­cy Actu­al­ly Smart Business?

It is a mis­take — but a use­ful one

xubingruo via iStock

Based on the research of

Sandeep Baliga

Jeffrey Ely

Listening: Interview with Sandeep Baliga and Jeff Ely on Sunk Costs

0:00 Skip back button Play Skip forward button 11:50

Don’t throw good mon­ey after bad.” 

Add Insight
to your inbox.

We’ll send you one email a week with content you actually want to read, curated by the Insight team.

Any stu­dent of eco­nom­ics knows this basic rule, which states that ratio­nal agents should not take irrecov­er­able or sunk” costs into account when mak­ing deci­sions about present or future invest­ments. Nev­er­the­less, human beings break this rule all the time, suc­cumb­ing to a cog­ni­tive bias known as the sunk-cost fal­la­cy.” If you have ever sat through a bad movie because you did not want to waste” the mon­ey you paid for the tick­et or fin­ished a PhD pro­gram you lost inter­est in years ago because of all the work you had already done, you have made this mis­take. But what if it were not always a mis­take — what if, in cer­tain sit­u­a­tions, this fal­la­cy” were actu­al­ly an opti­mal deci­sion-mak­ing strategy?

Sunk Cost: Keep Calm and Car­ry On
That is the coun­ter­in­tu­itive the­o­ry that Sandeep Bali­ga of the Kel­logg Grad­u­ate School of Man­age­ment and Jef­frey Ely of North­west­ern University’s Depart­ment of Eco­nom­ics advance in their paper Mnemo­nom­ics: The Sunk Cost Fal­la­cy as Mem­o­ry Kludge.” The authors argue that human beings — even ratio­nal ones — have a lim­it­ed capac­i­ty to remem­ber the orig­i­nal rea­son­ing behind their deci­sions. If that capac­i­ty is exceed­ed, the infor­ma­tion could be lost — so we need a men­tal place­hold­er that can remind us of why we decid­ed some­thing, just as tying a string around your fin­ger reminds you that you need to pick up milk on the way home from work. This kind of ad hoc mem­o­ry device” is called a mnemon­ic. Mnemon­ic devices often encode some aspect of the rel­e­vant infor­ma­tion as well, just as the let­ters in the mnemon­ic name Roy G. Biv” stand for the first let­ters of the col­ors in the rainbow.

It’s a short­hand way of remem­ber­ing some­thing that you might oth­er­wise for­get,” Bali­ga explains. You look at it and you say, Ah ha!’ Our the­o­ry about sunk costs is like that. You made a deci­sion ear­ly on, and it cost you some mon­ey; lat­er, you for­get the details of why you made the deci­sion, but you remem­ber the cost. That’s the mnemon­ic.” The PhD stu­dent trudg­ing ahead to com­plete his stud­ies, for exam­ple, may not remem­ber why get­ting a degree felt so impor­tant to begin with, but he does know that it has cost him thou­sands of dol­lars and years of effort so far. There­fore, it must make sense to com­plete the degree, because that sunk cost” is a mnemon­ic device: it both reminds him of his orig­i­nal moti­va­tion and encodes some infor­ma­tion about it. He’d say, It cost me this much time and mon­ey, so it must be impor­tant, and I’m going to car­ry on,’” Bali­ga says.

Induc­ing For­get­ful­ness for Fun and Prof­it
As eco­nom­ic the­o­rists, Bali­ga and Ely could have laid out their intrigu­ing argu­ment (the title of their paper is a port­man­teau of mnemon­ic” and eco­nom­ics”) and let their empiri­cist col­leagues cre­ate mod­els to test it. Instead, we thought it would be fun to see our­selves if the the­o­ry held any water,” Bali­ga says. To sim­u­late the expe­ri­ence of mak­ing deci­sions and then los­ing the mem­o­ry of why they were made, Bali­ga and Ely devised a sim­ple two-stage invest­ment game. In the first, or ini­ti­a­tion, stage, the play­er is pre­sent­ed with two num­bers, each rep­re­sent­ing a pos­si­ble val­ue of a project to invest in. The game then presents the play­er with a numer­i­cal cost to ini­ti­ate the project and asks if he wants to invest at that cost or pass. The sec­ond, or com­ple­tion, stage shows the play­er the same infor­ma­tion — the two pos­si­ble val­ue out­comes of the invest­ment and the ini­ti­a­tion cost that was already sunk in the pre­vi­ous round — along with a new num­ber rep­re­sent­ing the cost to com­plete the project. The play­er can then choose to con­tin­ue invest­ing in the project or not.

As a play­er, you infer from the high ini­ti­a­tion cost that the val­ue of the project must have been real­ly high — and so you fig­ure it makes sense to keep pay­ing in order to com­plete it.” — Sandeep Baliga

But Bali­ga and Ely insert­ed a catch: each play­er had to repeat the ini­ti­a­tion stage on twen­ty dif­fer­ent projects before mov­ing on to the com­ple­tion stage. When you come back to the com­ple­tion stage on the first project, you’ve made so many oth­er deci­sions on oth­er projects that you’ve for­got­ten the spe­cif­ic rea­son why you chose to ini­ti­ate it,” Bali­ga explains. How­ev­er, you can see exact­ly what you already spent.” After con­duct­ing their exper­i­ment on 100 first-year MBA stu­dents at the Kel­logg School, the authors found that a sig­nif­i­cant num­ber of them used their pri­or sunk costs as a mnemon­ic reminder of their ini­tial invest­ment strat­e­gy. If the sunk cost was low, they declined to com­plete the project; if it was high, they con­tin­ued to invest. As a play­er, you infer from the high ini­ti­a­tion cost that the val­ue of the project must have been real­ly high — and so you fig­ure it makes sense to keep pay­ing in order to com­plete it,” Bali­ga says.

Pro­rat­ed Fal­lac­i­es
But the authors’ exper­i­ment also uncov­ered a nov­el effect of lim­it­ed mem­o­ry on eco­nom­ic deci­sion mak­ing: a sub­tle vari­a­tion on the sunk-cost fal­la­cy that Bali­ga calls the pro rata bias.” This bias is a sim­i­lar­ly irra­tional” ten­den­cy to fig­ure sunk costs into cur­rent and future deci­sions, with the imag­ined goal of amor­tiz­ing those past sunk costs with inflat­ed vari­able costs in the present. Let’s say you’re a phar­ma­ceu­ti­cal com­pa­ny who spends a lot of mon­ey on R&D,” Bali­ga explains. Now let’s also say that a new drug dis­cov­ered by that R&D process costs pen­nies to pro­duce. That R&D mon­ey is gone and not com­ing back, but because of the pro rata bias, my stu­dents want to charge $5 per pill because of those sunk costs.”

Bali­ga and Ely were sur­prised to find that this pro rata bias appeared as a base­line effect in their exper­i­men­tal results — that is, even when the invest­ment-game play­ers were fur­nished with enough infor­ma­tion dur­ing the com­ple­tion stage to can­cel out the lim­it­ed-mem­o­ry effects, the pro-rata bias per­sist­ed. To Bali­ga, this sug­gests that sunk-cost think­ing may be hard­wired” into the human mind as a use­ful evo­lu­tion­ary adap­ta­tion. We’re pro­grammed by instinct — it’s like want­i­ng to eat fat­ty food or meat when­ev­er you see it,” he says. We have to exert an enor­mous amount of self-con­trol to avoid those ten­den­cies. In the same way, even if mem­o­ry were plen­ti­ful, we would still com­mit these sunk-cost fallacies.”

An Evo­lu­tion­ary Expla­na­tion?
So what of the fal­la­cy itself? If it is a use­ful men­tal short­cut passed down to us by our savan­nah-dwelling ances­tors, how can it also be a mis­take,” as Bali­ga calls it in no uncer­tain terms? We call it the the­o­ry of the sec­ond best,’” he says. The more you can remem­ber, the bet­ter it is, and you don’t have to com­mit this fal­la­cy at all. But if it’s a com­mon fea­ture of human beings that they do for­get stuff, then the sunk cost fal­la­cy is an opti­mal response to that for­get­ful­ness. Sunk costs can encode infor­ma­tion about deci­sions you made in the past, and if that’s the case you should take them into account, because if you didn’t, you’d make even worse decisions.”

Mean­while, what about cor­po­ra­tions, with mem­o­ries” lim­it­ed only by their employ­ees’ prowess at record keep­ing? Accord­ing to Bali­ga, com­pa­nies and busi­ness­es fol­low sunk-cost bias­es as often as indi­vid­u­als do. So why would a com­pa­ny like IBM or Apple need the same mnemon­ic kludge” against for­get­ting that a sin­gle fal­li­ble per­son does?

The mem­o­ry-lim­i­ta­tion prob­lem is actu­al­ly big­ger in orga­ni­za­tions,” Bali­ga asserts. Costs on a bal­ance sheet are easy to keep track of. But strate­gic vision — why was this a good idea in the first place’ — is not. As man­age­r­i­al turnover occurs, an orga­ni­za­tion for­gets. You may inher­it a pre­vi­ous CEO’s projects, but you can’t direct­ly access his vision or strate­gic rea­son­ing. What you can do is say, The pre­vi­ous guy was smart, and I’m going to car­ry on because I can see what costs he sunk into it.”

Bali­ga cau­tions that the sunk-costs-as-mnemon­ic the­o­ry does not imply that every incom­ing CEO should blind­ly march ahead on what­ev­er plans they inher­it. Steve Jobs, for exam­ple, famous­ly scrapped most of Apple’s then-cur­rent prod­uct lines when he retook the com­pa­ny reins in 1997 and led the orga­ni­za­tion to record prof­its. In that sit­u­a­tion, though, Jobs had clear evi­dence of what wasn’t work­ing in the busi­ness, so he changed course,” Bali­ga explains. In con­trast, he says, cur­rent Apple CEO Tim Cook is con­tin­u­ing on the path set by Jobs, even though he may not lit­er­al­ly share the vision that start­ed it all. It’s up for inter­pre­ta­tion, because that type of soft infor­ma­tion’ can nev­er be ful­ly record­ed,” Bali­ga says. Our research sug­gests that it may be pru­dent not to ignore the deci­sions made by your pre­de­ces­sors. They might encode things you don’t know, and may nev­er know.”

Featured Faculty

Sandeep Baliga

John L. and Helen Kellogg Professor; Professor of Managerial Economics & Decision Sciences

Jeffrey Ely

Morrison Professor of Economics (Weinberg), Professor of Managerial Economics & Decision Sciences (courtesy)

About the Writer

John Pavlus is a writer and filmmaker focusing on science, technology, and design topics. He lives in Portland, Oregon.

About the Research

Baliga, Sandeep, and Jeffrey C. Ely. 2011. “Mnemonomics: The Sunk Cost Fallacy as a Memory Kludge.” American Economic Journal: Microeconomics, 3(4): 35–67.

Read the original

Suggested For You

Most Popular


How Are Black – White Bira­cial Peo­ple Per­ceived in Terms of Race?

Under­stand­ing the answer — and why black and white Amer­i­cans’ respons­es may dif­fer — is increas­ing­ly impor­tant in a mul­tira­cial society.


Why Warmth Is the Under­ap­pre­ci­at­ed Skill Lead­ers Need

The case for demon­strat­ing more than just competence.

Most Popular Podcasts


Pod­cast: Our Most Pop­u­lar Advice on Improv­ing Rela­tion­ships with Colleagues

Cowork­ers can make us crazy. Here’s how to han­dle tough situations.

Social Impact

Pod­cast: How You and Your Com­pa­ny Can Lend Exper­tise to a Non­prof­it in Need

Plus: Four ques­tions to con­sid­er before becom­ing a social-impact entrepreneur.


Pod­cast: Attract Rock­star Employ­ees — or Devel­op Your Own

Find­ing and nur­tur­ing high per­form­ers isn’t easy, but it pays off.


Pod­cast: How Music Can Change Our Mood

A Broad­way song­writer and a mar­ket­ing pro­fes­sor dis­cuss the con­nec­tion between our favorite tunes and how they make us feel.