class: center, top, title-slide # Rational Choice and Deterrence ## SOC371 ### Chuck Lanfear ### Jan 25, 2020<br>Updated: Jan 24, 2021 --- # Overview 1. Context for Deterrence and Rational Choice 2. Deterrence Theory * Certainty, Severity, Celerity * General vs. Specific 3. Rational Choice Theory 4. Matseuda, Kreager, & Huzinga (2006) --- # Context * Criticism of the classical school lead to rejection of rational actor ??? Went a little far--more structural, more deterministic theories--often called "oversocialized" perspectives. -- * Criminology dominated by subcultural theories well into 1970s. + Crime as irrational behavior resulting from socialization ??? These include differential association, differential opportunity, strain theory, etc. Subcultural theories usually focus on normative conflict or difference: Crime occurs because people are members of subcultures that reject conventional norms and have pro-crime norms. -- * Attacks on rehabilitation in 70's and 80's + Focus on deterring crime rather than fixing criminals ??? Move toward rehabilitative model in 1960s and onward was not associated with dropping crime People became pessimistic--"there's nothing to be done"--so just lock folks up and throw away key -- * James Q. Wilson's (1975) *Thinking About Crime* + Respond harshly so those on fence don't get idea that crime pays. ??? James Q. Wilson was influential political scientist and policy writer. Focused on crime a lot. Thinking About Crime introduced idea of incapacitation as mechanism for crime reduction in places with higher penalties. We'll talk about incapacitation more later. Also wrote the original Broken Windows article with George Kelling and was pro War on Drugs on moral grounds. -- .centernote[*Rather than focus on the intervening work first, we begin at **deterrence**.*] ??? Deterrence is still fundamental basis for most crime policy. If penalties go up, crime should go down. --- # Deterrence and Rationality Deterrence is framed now in terms of rational choice but has long been basis from criminal legal policy * Classical school foundation ??? As seen in week 1, deterrence goes back to Beccaria and Bentham. It entered US CJ policy at foundation of nation and hasn't gone anywhere. More or less same into Europe. -- * Gary Becker's (1968) Economic Model of Crime ??? First well-known general utility model of crime. Becker was focusing on optimal calibration of punishment in CJ system--mostly deterrence -- * Bringing agency back in ??? Swinging pendulum back to actors as being decision-makers and away from deterministic cultural theories -- * Rational choice reintroduced classical model in more realistic terms + People are *on average* rational + People respond to incentives + Structural position and context matter ??? People do not pick optimal choice every time but do on average Key is that people respond to incentives People's perceived rewards are idiosyncratic and taste-based; mainly just know that on average people want *more*, whatever it is -- * Since criminal behavior seems irrational, it is a "hard test" for a theory of rational actors ??? If rational choice can explain criminality well, it can likely explain more rational things as well --- # Deterrence Doctrine Deterrence is simply the idea that punishment can impact crime. + People *respond to incentives* -- Weaker assumptions than classical school Rational actor assumptions: * In **aggregate**, individuals act rationally ??? If you look at large groups together, on average they behave rationally -- * At **margin**, individuals act rationally ??? People may not act perfectly rationally, but they do respond If punishment increases, people will commit less crime; if gains increase, more. -- .centernote[*Crime can be controlled by increasing its cost*] --- # Marginal vs. Absolute * Absolute Deterrence + Deterrent effect of a penalty vs. no penalty + *Will decriminalizing drugs result in more drug use?* + Less research here, but evidence is fairly strong, depending on the crime ??? This is question of whether *any penalty at all* has an effect on a behavior. -- * Marginal Deterrence + Deterrent effect of a penalty compared to different penalty + *Does the death penalty reduce more crime than life sentences?* + Most research focused here--evidence is mixed ??? This is if increasing or decreasing penalties increases or decreases behavior. --- # Variables in Deterrence * **Certainty** + Probability of being sanctioned--likelihood of pay cost + Objective Measures: * Arrest or clearance rate * Proportion of offenders prosecuted + Subjective Measure: Perceived punishment probability + Tittle's "Shell of Illusion" ??? Tittle's Shell of Illusion: Those never committing a crime--and thus never getting away with it--overestimate sanction risk. This is related to the illusion most people have that certain types of law violation are incredibly rare; since at least Rome philosophers have stated that if people knew how often laws were broken, they too would break them. -- * **Severity** + Severity of the sanction + This is the *cost* + Objective Measures: Sentence length; fine amount + Subjective Measure: Perceived suffering ??? Subjective usually involves asking people "how bad would it be to be arrested / incarcerated" -- * **Celerity** (Swiftness) + Length of time between crime and sanction ??? Not often used Time discounting in economics--a loss or gain in the future is worth less than one now Not real relevant in US system due to everything being slow The slowness may actually be a severity component too; people know they'll spend a lot of time even waiting for a hearing or trial --- class: inverse # General vs. Specific Deterrence --- # General Deterrence *Does the threat of punishment deter people from crime?* + Perceived threat of sanctions deters individuals from crime + Based on perception, not personal experience + Punish so that others see they too will be punished + About reducing *crime in general* -- Research Approaches + Changes in crime rates before/after policy changes + Effects of perceived risk and severity + Effect of peers being arrested --- # Specific Deterrence *Does being punished deter people from committing more crime?* + Experience of punishment deters individuals from future crime + Based on experience, occurs only after punishment + About reducing *recidivism* ??? Specific deterrence is about keeping people from reoffending If they were punished once, they will realize they don't want that to happen again -- Research Approaches + Compare incarcerated then released vs. probationers + But those sent to prison may be more crime prone + Experimental design: Need equivalent treatment and control groups + Randomly vary punishment + Lab experiments + Randomized arrest (e.g. Sherman & Berk 1984) + Look at marginal cases--natural experiments + Same crime probation vs. incarceration + Harsher vs. more lenient judges ??? Specific deterrence can be hard to research--punishment often does more than just punish. Not really ethical to experimentally vary *real* punishments; can try in experiments --- class: inverse # Rational Choice --- # Rational Choice Model Deterrence focuses on the sanctioning. Rational choice is a *general* theory of behavior--not limited to crime. -- Full rational choice models recognize variation in gains. + Includes material and psychic gains + e.g. Piliavin, Thornton, Gartner, & Matsueda (1986) ??? Article found better support for rational response to rewards than to formal deterrence Later work often shows similar--may be due to low risk of apprehension -- .image-50[  ] --- # Formalization Rational choice models typically take a formal mathematical form. $$ E(U) = (1-p)U(y)+pU(y-F) $$ "Expected utility `\(E(U)\)` is function of the value of crime `\(U(y)\)`, the likelihood of being caught `\(p\)`, and the penalty if caught `\(F\)`" ??? This is conventional economic utility model Don't estimate exact values, but rather takes advantage of basic properties All else equal, reducing reward / increasing cost reduces value and thus likelihood of crime. -- Becker (1968) introduced rational model of crime + Little impact in sociology or criminology. + Assumed full information + Material gains compared to gains from legitimate job ??? First major economic treatment of crime as rational behavior Mainly focused on criminal justice policy -- Short & Strodtbeck (1964) -- decision to join a gang fight + Subjective probability of harm + Bounded rationality + Status and psychic rewards ??? Mixed methods work revealing participation in gang fights as rational behavior Gang leader decides whether or not to join a fight If he decides not to, likely he will likely suffer moderate status loss with gang If he does join fight, small chance he will need to shoot someone and end up in jail -- large cost Because likelihood of jail is small, he'll probably join fight--it is rational to do so --- class: inverse ### Matsueda, Kreager, Huizinga. 2006. "Deterring Delinquents: A rational choice model of theft and violence." *American Sociological Review* --- # Premise Rational choice hypothesis: Perceived cost/benefit *impacts* future crime * Crime = Pr(Reward) `\(*\)` Value(Reward) > Pr(Cost) `\(*\)` Value(Cost) ??? That is, crime occurs when reward weighted by likelihood of getting it is higher than sanction risk weighted by likelihood. -- Research Questions: * How are perceived risks formed? * How do perceived risks and rewards influence crime? -- &zwj;Data: Denver Youth Survey * Annual panel survey of high-risk youth * Self-reported delinquency and crime * Perceived risk of arrest * Perceived rewards of crime * Wide range of crimes + Focus here is violence and theft ??? 1526 youth surveyed over 10 waves--only first five have rat choice questions Very detailed questions on involvement in crime, perceptions of it, background --- # Perceived Risk Theory &zwj;Deterrence: Perceptions should be rooted in reality * Experiential Learning + Baseline estimate of perceived risk is overestimated (Tittle 1980) + New information used to update ??? Perceived risks come from real experiences Those who have never committed a crime think they're likely to be caught When they get away with it (or get punished), they update perceptions--this is a Bayesian updating model -- * Measure for Perceived Arrest Risk + Experienced certainty: Arrests/Crimes for each person + Unsanctioned offenses: Number of crimes for non-arrested + Crimes of peers + Peers getting away with crime is information ??? DYS has actual arrests and committed crimes for respondents Know number of unsanctioned offenses Also know crimes of their peers--we assume peers getting away with it (or being sanctioned) gives info --- ## Formation of Perceived Arrest Risk .image-full[  ] ??? Baseline risk is perceived risk of arrest at last survey New info is unsanctioned offenses, arrests, and delinquency of peers *since last survey*. Updated risk is current survey. The model is somewhat more complicated: Preceding neighborhood and individual characteristics also predict all three. --- # Perceived Risk of Arrest .image-full[  ] ??? Find support for shell of illusion: High risk perceived by those not comitting Then risk increases with experienced certainty. Looks basically same for violence. People who never committed a crime perceive risk of sanctions similarly to those who did and were always caught --- # Rational Choice Theory * Rewards + Excitement + Being seen as cool ??? Focus here on psychic gains, both direct (excitement) and from status (being seen as cool) -- * Cost + Arrest ??? Main analyzed cost here is arrest but they also investigated trouble with parents, risk of ending up in jail or reform school Similar but weaker effects compared to arrest -- * Certainty (Probability) + "What are the chances of getting (excitement/coolness/arrested) for (theft/violence)?" ??? Measured as a percentage in 10 point increments -- * Severity (Value) + "How good or bad is this?" ??? 5 point scale The cost or benefit ends up as a variable equal to the certainty times the value. -- * Opportunity ??? Also address reality that crime can't be committed without the opportunity to do so, so asks if they had the opportunity to commit crimes --- # Rational Choice and Crime .image-full[  ] ??? Original participation in crime predicts perceived risks and gains as well as future crime Future crime predicted by these perceived risks and gains --- # Findings * Small deterrent effect + 10% risk increase -> 3% theft, 5% violence reduction + 10% reduction in coolness -> 6% theft, 7% violence reduction ??? Evidence for modest deterrent effect, but relationship much stronger with rewards Youth more sensitive to the gains than the costs -- * US already has high arrest and imprisonment rates + Would be draconian to increase substantially + Marginal deterrence is weak ??? Not practical to increase risk sufficiently to deter these crimes -- * Some level of arrest and incarceration essential + Absolute deterrence likely important ??? Complete removal of threat likely to increase crime substantially -- * Deterrence not a panacea for crime problem + Need to address rewards to crime ??? Can't punish our way out of this. Better to address, in particular, psychic gains of crime. --- class: inverse # Questions --- # For Next Time * Sherman, Lawrence W., and Richard A. Berk. 1984. "The Deterrent Effect of Arrest for Domestic Assault." Pp. 357-361 in *Classics of Criminology*. Edited by J. E. Jacoby. Prospect Heights: Waveland * Cohen, Lawrence and Marcus Felson. 1979. "Social Change and Crime Rate Trends: A Routine Activity Approach." *American Sociological Review* 44(4): 588-608 ### Things to pay attention to: * Different costs involved in deterrence (e.g. arrest vs. incarceration) * Relationship between routine activities, deterrence, and rational choice * Potential causes of changes in crime that don't involve offender motivation * Policy implications for deterrence and routine activities