This dissertation contains three independent essays; each essay can be read in isolation. The first essay investigates the causal effect of criminal convictions on various labor market outcomes in young adults. The estimation method used is a nonparametric bounding approach intended to partially identify the causal effect. The data used for this essay comes from the 1997 National Longitudinal Survey of the Youth. The second essay reevaluates the causal effect of post-secondary schooling on unemployment incidence using historical data from the 1980 U.S. Census and information on cohort level Vietnam War conscription risk. Conscription risk is used as an instrument for endogenous post-secondary schooling in a specification that accounts for the discrete nature of the treatment and outcome of interest. The third essay investigates the underlying necessary assumptions needed for the monotone instrumental variable (MIV) assumption to have identifying power on both the upper and lower bounds of a treatment effect when the treatment of interest is binary. I show that if the treatment is monotonic in the instrument, as is routinely assumed in the literature on instrumental variables, then for the MIV to have identifying power on both the lower and upper bounds of the treatment effect, the conditional-on-received-treatment outcomes cannot exhibit the same monotonicity assumed by the MIV. Results are highlighted with an application investigating the effect of criminal convictions on job match quality using data from the 1997 National Longitudinal Survey of the Youth.