Background, Purpose, & Research Questions: High school exit exams are meant to standardize quality and ensure that students graduate with basic knowledge and skills. However, policies that expand exit exams are associated with increased dropout rates, particularly among non-white students (e.g. Papay, Murnane, and Willet, 2010; Reardon and Kurlaender, 2009). Some states offer an alternative pathway to graduation where students complete projects instead of passing exams. In a state-level panel study, Hemelt and Marcotte (2013) found that those implementing non-test alternatives did not experience declines in completion rates otherwise associated with exit exams. This study is the first to investigate whether students who choose an alternative pathway lose valuable human capital by foregoing exams. Using data from Maryland, we ask: (1) whether these students graduate with similar levels of human capital to those who pass the exit exams; and (2) whether they have better postsecondary outcomes than non-completers. Setting & Intervention: Since 2005, Maryland has both required high school exit exams and offered the Bridge Program for High School Validation (aka Bridge). Bridge empowers each school district to create criteria for students to graduate by demonstrating mastery of tested subjects through projects and portfolios. Bridge is intended to be "no easier" than passing an exam, and completers receive the same high school diploma as other graduates. During the period studied, students were eligible to Bridge in a subject only after attempting and failing the exit exam twice. Students could Bridge in any or all of three required subjects (algebra, biology, and English) and were also free to attempt exam retakes while simultaneously pursuing Bridge. Population & Data: We use administrative data from the Maryland Longitudinal Data Systems Center, a state agency and research center that combines data from Maryland public PK-12 schools, post-secondary institutions, and Unemployment Insurance system. We follow the full population of public high school students in four 9 grade cohorts from 8 grade through four years of high school and two years of post-secondary outcomes. Available measures include demographics, attendance, suspensions, exit exam scores, Bridge completion, high school completion, college enrollment anywhere in the US, and UI-eligible employment and wages in Maryland. We focus our analysis on the subset of students who failed an exit exam twice making them eligible for the non-test alternative. Approximately 25% of students were eligible to Bridge in at least one subject, and 12% used Bridge to graduate. The subset of Bridge-eligible students is 56% Black, 67% FRPL, and 19% SPED. Research Design: Our objective is to estimate the effect of Bridge on post-secondary outcomes while addressing the problem of student self-selection. While we can perfectly observe eligibility for treatment based on test attempts and scores, there is no available instrument to identify the causal effects of Bridge that overcomes self-selection. Instead, we use observable characteristics to create matched comparisons between bridge completers and students who are similar across characteristics that influence both the decision to bridge and post-secondary outcomes. We first divide the subset of students who are eligible to Bridge in each subject into three mutually exclusive groups: (1) Bridge completers, (2) those who passed a retake, and (3) those who failed to complete high school. We then employ two matching strategies to compare bridgers to test-passers, and then to compare bridgers to non-completers. Our primary matching strategy (Angrist, Pathak, and Walter, 2013) sorts students into cells that attended the same school, are demographically identical, and received similar failing exam scores. Regression with cell fixed effects isolates the unique effect of Bridge on observably similar students. We also conduct propensity score matching based on a broad set of observables. In both strategies, we exploit pre-high school, 8 grade measures of academic performance and school engagement that are exogenous to exit exam performance. We conduct this analysis separately for each exit exam subject, as well as for subgroups at high risk of non-completion. Results are highly consistent across matching strategies. Findings: Comparing bridge completers to those who pass exam retakes: (1) Bridge completers are 2-4 percentage points less likely to attend 2-year college and 2-3 percentage points less likely to attend 4-year college; (2) No difference in probability of employment; and (3) Wages are comparable for students who bridge in biology and English. Wages are 10-15% lower for students who bridge in algebra. Comparing bridge completers to those who do not complete high school: (1) Bridge completers are more likely to attend 2-year college by 6-8 percentage points. There is no difference in the probability of attending 4-year college; (2) Bridge completers are more likely to be employed by 6-9 percentage points; and (3) Wages for Bridge completers in biology and English are 19-26% higher. There is no difference in algebra. Conclusions: We find that high school graduates who complete a non-test alternative have slightly lower college-entry rates than similar students who graduate by passing exams. However, post-high school employment prospects are similar, except that graduates who pass algebra test make higher wages. Graduates who complete a non-test alternative are substantially better off than non-completers including higher enrollment in two-year colleges, higher rates of employment, and higher wages. States with exit exams and high dropout rates should consider adding non-test alternatives with consideration of the apparent importance of passing an algebra exam and with efforts to support the transition to college.