A Proposal for Defeating Moloch in the Prison Industrial Complex


I’d like to in­creas­ing the well-be­ing of those in the jus­tice sys­tem while si­mul­ta­neously re­duc­ing crime. I’m miss­ing some­thing here but I’m not sure what. I’m think­ing this may be a worse idea than I origi­nally thought based on com­ment feed­back, though I’m still not 100% sure why this is the case.

Cur­rent State

While the prison sys­tem may not con­sti­tute an ex­is­ten­tial threat, At this mo­ment more than 2,266,000 adults are in­car­cer­ated in the US alone, and I ex­pect that be­ing in prison greatly de­creases QALYs for those in­car­cer­ated, that fur­ther QALYs are lost to vic­tims of crime, fam­ily mem­bers of the in­car­cer­ated, and through the con­tin­u­ing effects of in­sti­tu­tion­al­iza­tion and PTSD from sen­tences served in the cur­rent sys­tem, not to men­tion the brain­power and man-hours lost to any pro­duc­tive use.

If you haven’t read these Med­i­ta­tions on Moloch, I highly recom­mend it. It’s long though, so the ex­ec­u­tive sum­mary is: Moloch is the per­son­ifi­ca­tion of the forces of com­pe­ti­tion which per­verse in­cen­tives, a “race to the bot­tom” type situ­a­tion where all hu­man val­ues are dis­carded in an effort to sur­vive. That this can be solved with bet­ter co­or­di­na­tion, but it is very hard to co­or­di­nate when per­verse in­cen­tives also pe­nal­ize the co­or­di­na­tors and re­ward dis­sen­ters. The prison in­dus­trial com­plex is an ex­am­ple of these per­verse in­cen­tives. No one thinks that the cur­rent sys­tem is ideal but in­cen­tives pre­vent pos­i­tive change and in­crease ab­solute un­hap­piness.

  • Poli­ti­ci­ans com­pete for electabil­ity. Con­victs can’t vote, pris­ons make cam­paign con­tri­bu­tions and jobs, and ap­pear­ing “tough on crime” ap­peals to a large por­tion of the voter base.

  • Jails com­pete for money: the more pris­on­ers they house, the more they are paid and the longer they can con­tinue to ex­ist. This in­cen­tive is strong for pub­lic pris­ons and dou­bly strong for pri­vate pris­ons.

  • Po­lice com­pete for bonuses and pro­mo­tions, both of which are given as re­wards to cops who bring in and con­vict more criminals

  • Many of the in­mates them­selves are mo­ti­vated to com­mit crim­i­nal acts by the small num­ber of non-crim­i­nal op­por­tu­ni­ties available to them for fi­nan­cial suc­cess, be­sides crim­i­nal acts. After be­com­ing a crim­i­nal, this num­ber of op­por­tu­ni­ties is fur­ther nar­rowed by back­ground checks.

The in­cen­tives have come far out of line with hu­man val­ues. What can be done to bring in­cen­tives back in al­ign­ment with the com­mon good?

My Proposal

Us­ing a model that pre­dicts re­ci­di­vism at sixty days, one year, three years, and five years, pre­dict the ex­pected re­ci­di­vism rate for all in­mates at all in­di­vi­d­ual prison given av­er­age re­ci­di­vism. Sixty days af­ter re­lease, if re­ci­di­vism is be­low the pre­dicted rate, the prison gets a small sum of money equal­ing 25% of the pre­dicted cost to the state of deal­ing with the pre­dicted re­ci­di­vism (in­clud­ing lawyer fees, court fees, and jailing costs). This is re­peated at one year, three years, and five years.

The statis­ti­cal mod­els would be read­justed with cur­rent data ev­ery years, so if this model causes re­ci­di­vism to drop across the board, jails would be com­pet­ing against ever higher stan­dard, com­pet­ing to cre­ate the most in­no­va­tive and ground­break­ing coun­sel­ing and job skills and restora­tive meth­ods so that they don’t lose their edge against other pris­ons com­pet­ing for the same money. As it be­comes harder and harder to edge out the com­pe­ti­tion’s ad­vanced meth­ods, and as the prison pop­u­la­tion is re­duced, ad­di­tional in­cen­tives could come by end­ing state con­tracts with the bot­tom 10% of pris­ons, or with any pris­ons who have re­ci­di­vism rates larger than ex­pected for mul­ti­ple years in a row.

Note that this pro­posal makes no policy recom­men­da­tions or value judge­ment be­sides chang­ing the in­cen­tive struc­ture. I have opinions on the san­ity of cer­tain laws and poli­cies and the pri­vate prison sys­tem it­self, but this spe­cific pro­posal does not. Ideally, this will re­duce some amount of par­ti­san bick­er­ing.

Us­ing this added suc­cess in­cen­tive, here are the mod­ified mo­ti­va­tions of each of the ma­jor ac­tors.

  • Poli­ti­ci­ans com­pete for electabil­ity. Con­victs still can’t vote, pris­ons make cam­paign con­tri­bu­tions, and ap­pear­ing “tough on crime” still ap­peals to a large por­tion of the voter base. The poli­ti­cian can promise a re­duc­tion in crime with­out mak­ing any spe­cific policy or pro­gram recom­men­da­tions, thus shield­ing them­selves from crit­i­cism of be­ing soft on crime that might come from en­dors­ing restora­tive jus­tice or psy­cholog­i­cal coun­sel­ling, for in­stance. They get to claim suc­cess for pro­grams that other peo­ple, are in charge of ad­minis­trat­ing and de­sign­ing. Fur­ther, they are sav­ing 75% of the money pre­dicted to have have been spent ad­minis­trat­ing crim­i­nals. Pri­sons love get­ting more money for do­ing the same amount of work so cam­paign con­tri­bu­tions would stay sta­ble or go up for poli­ti­ci­ans who sup­port re­duced re­ci­di­vism bonuses.

  • Pri­sons com­pete for money. It costs the state a huge amount of money to house pris­on­ers, and the net profit from hous­ing a pris­oner is small af­ter pay­ing for food, cloth­ing, su­per­vi­sion, space, re­pairs, en­ter­tain­ment, ect. An ad­di­tional 25% of that cost, with no ad­di­tional ex­pen­di­tures is very at­trac­tive. I pre­dict that some amount of book-cook­ing will hap­pen, but that the gains pos­si­ble with book cook­ing are small com­pared to gains from ac­tual im­prove­ments in their prison pro­gram. Small differ­ences in pris­ons have po­ten­tial to make large differ­ences in post-prison be­hav­ior. I ex­pect hav­ing an on-staff CBT psy­chi­a­trist would make a big differ­ence; an ad­dic­tion spe­cial­ist would as well. A new ca­reer field is born: ex­pert con­sul­tants who travel from pri­vate prison to pri­vate prison and make recom­men­da­tions for what changes would re­duce re­ci­di­vism at the low­est pos­si­ble cost.

  • Po­lice and judges re­tain the same in­cen­tives as be­fore, for bonuses, pres­tige, and pro­mo­tions. This is good for the sys­tem, be­cause if their in­cen­tives were not run­ning counter to the pris­ons and jails, then there would be a lot of pres­sure to cook the books by look­ing the other way on crim­i­nals til af­ter the 60 day/​1 year/​5 year mark. I pre­dict that there will be a cou­ple scan­dals of cops found to be in league with pris­ons for a cut of the bonus, but that this method isn’t very prof­itable. For one thing, an en­tire po­lice force would have to be cor­rupt and for an­other, crim­i­nals are mo­bile and can com­mit crimes in other precincts. Po­lice are also mo­ti­vated to work in safer ar­eas, so the gen­eral pro­gram of re­ward­ing re­duced re­ci­di­vism is to their ad­van­tage.


If it could be shown that a model for pre­dict­ing re­ci­di­vism is highly pre­dic­tive, we will need to cre­ate an­other model to pre­dict how much the gov­ern­ment could save if switch­ing to a bonus sys­tem, and what re­duc­tion of crime could be ex­pected.

Halfway houses in Penn­syl­va­nia are already re­ceiv­ing non-re­ci­di­vism bonuses. Is a pi­lot pro­ject us­ing this pric­ing struc­ture fea­si­ble?