Akrasia, hyperbolic discounting, and picoeconomics

Akra­sia is the ten­dency to act against your own long-term in­ter­ests, and is a prob­lem doubtless only too fa­mil­iar to us all. In his book “Break­down of Will”, psy­chol­o­gist Ge­orge C Ainslie sets out a the­ory of how akra­sia arises and why we do the things we do to fight it. His ex­traor­di­nary pro­posal takes in­sights given us by eco­nomics into how con­flict is re­solved and ex­tends them to con­flicts of differ­ent agen­cies within a sin­gle per­son, an ap­proach he terms “pi­coeco­nomics”. The foun­da­tion is a cu­ri­ous dis­cov­ery from ex­per­i­ments on an­i­mals and peo­ple: the phe­nomenon of hy­per­bolic dis­count­ing.

We all in­stinc­tively as­sign a lower weight to a re­ward fur­ther in the fu­ture than one close at hand; this is “dis­count­ing the fu­ture”. We don’t just ac­count for a slightly lower prob­a­bil­ity of re­ciev­ing a more dis­tant award, we value it at in­her­ently less for be­ing fur­ther away. It’s been an ac­tive de­bate on over­com­ing­bias.com whether such dis­count­ing can be ra­tio­nal at all. How­ever, even if we al­low that dis­count­ing can be ra­tio­nal, the way that we and other an­i­mals do it has a struc­ture which is in­her­ently ir­ra­tional: the weight­ing we give to a fu­ture event is, roughly, in­versely pro­por­tional to how far away it is. This is hy­per­bolic dis­count­ing, and it is an em­piri­cally very well con­firmed re­sult.

I say “in­her­ently ir­ra­tional” be­cause it is in­con­sis­tent over time: the rel­a­tive cost of a day’s wait is con­sid­ered differ­ently whether that day’s wait is near or far. Look­ing at a day a month from now, I’d sooner feel awake and al­ive in the morn­ing than stay up all night read­ing com­ments on less­wrong.com. But when that evening comes, it’s likely my prefer­ences will re­verse; the dis­tance to the morn­ing will be rel­a­tively greater, and so my hap­piness then will be dis­counted more strongly com­pared to my pre­sent en­joy­ment, and an­other groggy morn­ing will await me. To my hor­ror, my fu­ture self has differ­ent in­ter­ests to my pre­sent self, as surely as if I knew the day a mur­der pill would be forced upon me.

If I knew that a mur­der pill re­ally would be forced upon me on a cer­tain date, af­ter which I would want noth­ing more than to kill as many peo­ple as pos­si­ble as grue­somly as pos­si­ble, I could not sit idly by wait­ing for that day to come; I would want to do some­thing now to pre­vent fu­ture car­nage, be­cause it is not what the me of to­day de­sires. I might at­tempt to frame my­self for a crime, hop­ing that in prison my abil­ity to go on a kil­ling spree would be con­tained. And this is ex­actly the be­havour we see in peo­ple fight­ing akra­sia: con­sider the al­co­holic who moves to a town in which al­co­hol is not sold, an­ti­ci­pat­ing a change in de­sires and de­liber­ately con­strain­ing their own fu­ture self. Ainslie de­scribes this as “a re­la­tion­ship of limited war­fare among suc­ces­sive selves”.

And it is this war­fare which Ainslie analy­ses with the tools of be­havi­oural eco­nomics. His anal­y­sis ac­counts for the im­por­tance of mak­ing re­s­olu­tions in defeat­ing akra­sia, and the rea­sons why a re­s­olu­tion is eas­ier to keep when it rep­re­sents a “bright clear line” that we can­not fool our­selves into think­ing we haven’t crossed when we have. It also dis­cusses the dan­gers of willpower, and the ways in which our in­tertem­po­ral bar­gain­ing can leave us act­ing against both our short-term and our long-term in­ter­ests.

I can’t re­ally do more than scratch the sur­face on how this anal­y­sis works in this short ar­ti­cle; you can read more about the anal­y­sis and the book on Ainslie’s web­site, pi­coeco­nomics.org. I have the im­pres­sion that defeat­ing akra­sia is the num­ber one pri­or­ity for many less­wrong.com read­ers, and this work is the first I’ve read that re­ally sets out a mechanism that un­der­lies the strange bat­tles that go on be­tween our shorter and longer term in­ter­ests.