Home » Posts tagged 'instrumental variables'
Tag Archives: instrumental variables
The Case for Clemency: Differential Impacts of Pretrial Detention on Case and Crime Outcomes
by George Rateb
Abstract
About half-million of individuals in US jails are detained pretrial while legally presumed
innocent. Using data on quasi-randomly assigned bail judges in the third-largest court system in
the U.S., we study the impact of pretrial detention on defendants’ court and crime outcomes
between 2008 and 2012. We supplement our primary analysis to document patterns on bail
amounts and how they differentially impact Black defendants relative to their white and Hispanic
counterparts. Instrumental variable estimates suggest that pretrial detention increases the
likelihood of being found guilty, mainly driven by the uptake of guilty pleas, especially for
minorities. By linking court and jail data, we provide mechanistic evidence that jail time is
positively correlated with the uptake of these guilty pleas. To the best of our knowledge, these
findings have not been empirically documented due to a lack of previous data availability.
Professor Bocar Ba, Faculty Advisor
Professor Michelle Connolly, Faculty Advisor
JEL Codes: C26; J15; K14
Exceeding the Threshold: Analysis of Public Information Transfer using Instrumental Variables
By Gabrielle Inder
This paper examines how information transfer about contamination levels found at brownfield sites capitalizes into nearby property values. More specifically, a hedonic model is used to test the impact on housing transaction prices when a binary measure (i.e. exceeding a threshold or not) or a continuous measure (i.e. chemical levels) is used. In the analysis, I exploit the variation in the contaminant thresholds, caused by regulatory conditions defined by the state of Massachusetts, holding the contaminant level constant. As thresholds are tied to neighborhood attributes in areas surrounding brownfields, threshold exceedance is potentially correlated to unobserved factors that impact housing values. An instrumental variables approach is used to create variation in threshold
exceedance through the use of an instrument that measures the presence to underground aquifers. After instrumenting for threshold exceedance, my estimates indicate that a 10.8% decrease in housing values occurs when a contaminant threshold is exceeded, while the continuous measures of toxicity indicate a negative but insignificant effect. These findings suggest that policy makers should consider information conveyance when creating policies to inform homeowners of pollution presence, as improved information provision may increase public awareness about local environmental concerns.
Advisor: Christopher Timmins, Michelle Connolly, Lala Ma | JEL Codes: C26, Q5, Q53 | Tagged: Brownfields, Hedonic Analysis, Housing Markets, Instrumental Variables, Pollution
