Contact Information
Email: alden15@mit.edu
Backup: alden1505@gmail.com
Embedded Files
Biography
Biography
I am a recent PhD graduate from MIT, with research interests in health economics, peer effects, and applied econometrics. I am currently an NBER postdoctoral research associate and Gies Health Initiatives Research Scholar, working from the Gies College of Business, University of Illinois. I am on the 2026-2027 job market.
Click here for my CV.
Education
PhD in Economics, 2023
Massachusetts Institute of Technology
BA in Economics, Applied Mathematics, and Statistics, 2016
University of California, Berkeley
Working Papers
Working Papers
Patient Peer Effects: Evidence from Nursing Home Room Assignments (with Martin Hackmann). Job Market Paper. NBER Working Paper 34538.
Patient Peer Effects: Evidence from Nursing Home Room Assignments (with Martin Hackmann). Job Market Paper. NBER Working Paper 34538.
Media Mentions: NBER Bulletin on Health; McKnights; Gies College of Business
Awards: Best Paper in Aging and Long-Term Care at American Society of Health Economists Conference (ASHEcon) 2026; Best Paper Award at the 2026 Economics of Mental Health Workshop
Presentations: ASSA Annual Meeting (scheduled); NBER Emerging Challenges in Long-Term Care and Caregiving Conference; NBER Summer Institute Workshop in Aging; NBER Economics of Chronic Disease; ASHEcon; Annual Health Economics Conference; Economics of Mental Health Workshop; Midwest Econometrics Group Conference; NBER Economics of AD/ADRD Meeting (poster)
Funding: NBER Research on the Economics of Alzheimer’s/Dementia (READ) Grant ($194,590; award pending). Co-PI with Martin Hackmann. The Economics of Roommate Assignment for Individuals with AD/ADRD: Implications for Spending, Resident Outcomes, and Care Delivery.
We provide causal evidence that peer effects among nursing home roommates generate mortality impacts comparable to provider quality differences. Using room availability at the time of admission as instruments for room assignments, we estimate that assignment to a roommate with dementia increases 90-day mortality by 7%, relative to being assigned no roommate. Effects differ sharply by patient type: patients with dementia benefit substantially from cognitively healthy roommates, whereas mortality of patients without dementia does not depend on roommates' cognitive health, implying a unidirectional positive externality. We find that room assignment policies that exploit peer effects can reduce mortality by 4%.
We provide causal evidence that peer effects among nursing home roommates generate mortality impacts comparable to provider quality differences. Using room availability at the time of admission as instruments for room assignments, we estimate that assignment to a roommate with dementia increases 90-day mortality by 7%, relative to being assigned no roommate. Effects differ sharply by patient type: patients with dementia benefit substantially from cognitively healthy roommates, whereas mortality of patients without dementia does not depend on roommates' cognitive health, implying a unidirectional positive externality. We find that room assignment policies that exploit peer effects can reduce mortality by 4%.
Presentations: University of Illinois Chicago; Imperial College London; Midwest Economics Association Conference; University of Massachusetts Amherst (online)
For over half a century, limited consumer responsiveness to provider quality has been viewed as a defining feature of health care markets. I show that this apparent exceptionalism partly reflects a failure to distinguish consumer preferences from provider-imposed constraints on access. Using administrative data covering the universe of nursing home residents in California, I construct quality measures based on risk-adjusted mortality that are robust to gaming of publicly provided quality measures and validate them using a distance-based IV strategy. I then use these quality measures as inputs into a structural demand model in which nursing homes may reject patients based on observable characteristics. Conventional demand estimation that treats all options as available yields precise null estimates of average demand for quality. Accounting for unobserved rejections instead reveals positive consumer demand, although average responsiveness remains lower than estimates from hospital settings and varies substantially across baseline demographic and health characteristics. The introduction of nursing home star ratings had limited effects on consumer demand, but counterfactual simulations indicate that financial steering incentives and improved hospital discharge processes may improve allocation and reduce mortality by 5 to 17 percent in spite of supply-side constraints. Finally, demand heterogeneity has important consequences for patient outcomes, explaining roughly 10 percent of the mortality-age gradient.
For over half a century, limited consumer responsiveness to provider quality has been viewed as a defining feature of health care markets. I show that this apparent exceptionalism partly reflects a failure to distinguish consumer preferences from provider-imposed constraints on access. Using administrative data covering the universe of nursing home residents in California, I construct quality measures based on risk-adjusted mortality that are robust to gaming of publicly provided quality measures and validate them using a distance-based IV strategy. I then use these quality measures as inputs into a structural demand model in which nursing homes may reject patients based on observable characteristics. Conventional demand estimation that treats all options as available yields precise null estimates of average demand for quality. Accounting for unobserved rejections instead reveals positive consumer demand, although average responsiveness remains lower than estimates from hospital settings and varies substantially across baseline demographic and health characteristics. The introduction of nursing home star ratings had limited effects on consumer demand, but counterfactual simulations indicate that financial steering incentives and improved hospital discharge processes may improve allocation and reduce mortality by 5 to 17 percent in spite of supply-side constraints. Finally, demand heterogeneity has important consequences for patient outcomes, explaining roughly 10 percent of the mortality-age gradient.
From Peer Effects to Network Evolution: Evidence from the Illinois Workplace Wellness Study (with Damon Jones, David Molitor, and Julian Reif)
From Peer Effects to Network Evolution: Evidence from the Illinois Workplace Wellness Study (with Damon Jones, David Molitor, and Julian Reif)
Presentations: CausNetS, Satellite of NetSci 2026; Economic Science Association World Meetings; Advances in Field Experiments
We randomly assigned eligibility and financial incentives for a two-year workplace wellness program to 4,834 employees. Friendship network data were collected at baseline and again one and two years after randomization. Participation in health screenings rose by 4.2 percentage points for each additional coworker assigned to treatment, and by 6.4 percentage points for each additional coworker induced to participate as a result of treatment. The program also shaped network dynamics: eligibility increased the probability of being listed as a coworker by others in subsequent years by 9 percent, and the intervention increased network homophily along some dimensions. Behavioral spillovers amplified the effect of the intervention on screening participation by 6 percent, while dynamic spillovers through network evolution increased future participation by about 0.8 percent of baseline.
We randomly assigned eligibility and financial incentives for a two-year workplace wellness program to 4,834 employees. Friendship network data were collected at baseline and again one and two years after randomization. Participation in health screenings rose by 4.2 percentage points for each additional coworker assigned to treatment, and by 6.4 percentage points for each additional coworker induced to participate as a result of treatment. The program also shaped network dynamics: eligibility increased the probability of being listed as a coworker by others in subsequent years by 9 percent, and the intervention increased network homophily along some dimensions. Behavioral spillovers amplified the effect of the intervention on screening participation by 6 percent, while dynamic spillovers through network evolution increased future participation by about 0.8 percent of baseline.
Presentations: NBER Health and Healthcare Variations Across Populations Conference; Midwest Health Economics Conference; University of Illinois Urbana-Champaign (Econometrics Seminar); Seoul National University; Yonsei University; ASHEcon; University of Massachusetts Amherst (online)
Racial minorities are disproportionately admitted to lower-quality nursing homes, but the forces generating segregation across facilities need not be the same as those generating racial gaps in quality. Using nationwide administrative data on nursing home residents in the U.S. from 2000–2010, I show that neighborhood sorting explains an important share of across-facility segregation, yet substantial disparities remain in the quality of nursing homes that different racial groups are admitted to, even conditional on fixed effects for zip code of prior residence. I then document three additional mechanisms that may drive across-facility segregation and disparities: (i) race-based admissions policies, where nursing homes become less likely to admit minority residents when capacity is strained; (ii) in-group preferences, whereby shocks to the minority share at nursing homes lead to persistent changes in the racial composition, and (iii) heterogeneity in responsiveness to quality, where information shocks lead to larger changes in where White residents are admitted. To quantify the contributions of these explanations, I estimate a two-sided matching model with latent choice set constraints in three large states, using distance and short-run within-facility occupancy shocks for identification. Counterfactual simulations show that neighborhood sorting and race-based admissions policies are complementary drivers of segregation: while the sum of their individual contributions accounts for just over one-half of segregation, eliminating both simultaneously reduces segregation by 85–90 percent. In contrast, the drivers of disparities do not always align with drivers of segregation and vary substantially across states as well as between Black-White and Hispanic-White comparisons within states, implying that policies which reduce segregation need not reduce—and may even worsen—disparities for some minority groups.
Racial minorities are disproportionately admitted to lower-quality nursing homes, but the forces generating segregation across facilities need not be the same as those generating racial gaps in quality. Using nationwide administrative data on nursing home residents in the U.S. from 2000–2010, I show that neighborhood sorting explains an important share of across-facility segregation, yet substantial disparities remain in the quality of nursing homes that different racial groups are admitted to, even conditional on fixed effects for zip code of prior residence. I then document three additional mechanisms that may drive across-facility segregation and disparities: (i) race-based admissions policies, where nursing homes become less likely to admit minority residents when capacity is strained; (ii) in-group preferences, whereby shocks to the minority share at nursing homes lead to persistent changes in the racial composition, and (iii) heterogeneity in responsiveness to quality, where information shocks lead to larger changes in where White residents are admitted. To quantify the contributions of these explanations, I estimate a two-sided matching model with latent choice set constraints in three large states, using distance and short-run within-facility occupancy shocks for identification. Counterfactual simulations show that neighborhood sorting and race-based admissions policies are complementary drivers of segregation: while the sum of their individual contributions accounts for just over one-half of segregation, eliminating both simultaneously reduces segregation by 85–90 percent. In contrast, the drivers of disparities do not always align with drivers of segregation and vary substantially across states as well as between Black-White and Hispanic-White comparisons within states, implying that policies which reduce segregation need not reduce—and may even worsen—disparities for some minority groups.
Avoiding Regulation: Evidence from Nursing Homes (with Riley League). NBER Working Paper 35849. Under Review.
Avoiding Regulation: Evidence from Nursing Homes (with Riley League). NBER Working Paper 35849. Under Review.
We study strategic responses by firms to threshold-based regulations. We document extensive bunching in response to additional staffing requirements for California nursing homes with at least 100 beds. To estimate the regulation’s effect on staffing, we develop a novel treatment effects estimator for threshold-based regulations under precise manipulation, by decomposing firms into response types based on counterfactual and realized staffing and size, separating causal effects from selection. We estimate that higher staffing above the threshold is nearly entirely due to selection, with low-staffing facilities manipulating their size to avoid the regulation. These size distortions limit access, particularly for disadvantaged patients.
We study strategic responses by firms to threshold-based regulations. We document extensive bunching in response to additional staffing requirements for California nursing homes with at least 100 beds. To estimate the regulation’s effect on staffing, we develop a novel treatment effects estimator for threshold-based regulations under precise manipulation, by decomposing firms into response types based on counterfactual and realized staffing and size, separating causal effects from selection. We estimate that higher staffing above the threshold is nearly entirely due to selection, with low-staffing facilities manipulating their size to avoid the regulation. These size distortions limit access, particularly for disadvantaged patients.
Presentations: International Health Economics Association Webinar
In regression discontinuity designs with multiple running variables (MRD), units are assigned different treatments based on whether their values on several observed running variables exceed known thresholds. In such designs, applied work commonly analyzes each running variable separately, estimating a single-dimensional RD design in the first running variable after limiting the sample to the set of individuals qualifying on the second threshold, and vice versa. In this paper, I propose a new estimator for MRD designs using thin plate splines that improves upon the applied practice in two ways. First, the estimator can be used to estimate the conditional average treatment effect at every point on the boundary separating treated and untreated units, and second, it provides efficiency gains by using the entire sample. I also develop analogous estimators for multidimensional regression kink (MRK) and multidimensional regression discontinuity/kink (MRDK) designs. I establish theoretical properties for these estimators, before presenting simulation results showing that they perform well in finite samples. Finally, I demonstrate the performance of my MRD estimator with two empirical applications: Londoño-Vélez, Rodríguez, and Sánchez (2020) on the effect of financial aid on college enrollment, and Keele and Titiunik (2015) on the effect of political ads on election turnout. Open-source software is available for implementing the proposed methods.
In regression discontinuity designs with multiple running variables (MRD), units are assigned different treatments based on whether their values on several observed running variables exceed known thresholds. In such designs, applied work commonly analyzes each running variable separately, estimating a single-dimensional RD design in the first running variable after limiting the sample to the set of individuals qualifying on the second threshold, and vice versa. In this paper, I propose a new estimator for MRD designs using thin plate splines that improves upon the applied practice in two ways. First, the estimator can be used to estimate the conditional average treatment effect at every point on the boundary separating treated and untreated units, and second, it provides efficiency gains by using the entire sample. I also develop analogous estimators for multidimensional regression kink (MRK) and multidimensional regression discontinuity/kink (MRDK) designs. I establish theoretical properties for these estimators, before presenting simulation results showing that they perform well in finite samples. Finally, I demonstrate the performance of my MRD estimator with two empirical applications: Londoño-Vélez, Rodríguez, and Sánchez (2020) on the effect of financial aid on college enrollment, and Keele and Titiunik (2015) on the effect of political ads on election turnout. Open-source software is available for implementing the proposed methods.
Selection on Unobservables in Discrete Choice Models
Presentations: Indiana University Bloomington
Selection on unobservables is an important concern for causal inference in observational studies, and accordingly, previous papers have developed methods for sensitivity analysis for OLS, binary choice models, instrumental variables, and movers designs. In this paper, I develop methods for sensitivity analysis for a setting that has not been previously studied — discrete choice models. In particular, I derive bounds for the omitted variables bias under an assumption about how much the consumer values the omitted variable(s) relative to the included control variables, and about the relationship between the omitted variable and the variable of interest. I provide theoretical results for my bounding procedure, and demonstrate its performance in simulations. Finally, I show in several empirical applications that my procedure produces economically meaningful bounds.
Selection on unobservables is an important concern for causal inference in observational studies, and accordingly, previous papers have developed methods for sensitivity analysis for OLS, binary choice models, instrumental variables, and movers designs. In this paper, I develop methods for sensitivity analysis for a setting that has not been previously studied — discrete choice models. In particular, I derive bounds for the omitted variables bias under an assumption about how much the consumer values the omitted variable(s) relative to the included control variables, and about the relationship between the omitted variable and the variable of interest. I provide theoretical results for my bounding procedure, and demonstrate its performance in simulations. Finally, I show in several empirical applications that my procedure produces economically meaningful bounds.
Past Work
Past Work
Media Mentions: Boston Globe
The issue of fake news has been hotly debated in recent years, with some claiming that it played a role in 2016 U.S. presidential elections. Despite these claims, there has been limited evidence to date linking fake news directly to voting behavior. In this project, I seek to provide credible evidence on this question by using big college football games as an instrument for fake news consumption, during the lead-up to the 2016 presidential election. During this period, I find that search volumes for pro-Trump fake news terms were lower in counties close to college football teams that played a big game when a piece of pro-Trump fake news went viral, and also that these counties were less likely to vote for Trump. The magnitude of these estimates suggests that a one-standard deviation increase in search volume for pro-Trump fake news terms increased Trump’s vote share in 2016 by about 5.4–7.4 percentage points. In addition, I find that the relationship between fake news exposure and voting behavior is approximately linear, by developing and estimating a novel nonparametric IV estimation method.
The issue of fake news has been hotly debated in recent years, with some claiming that it played a role in 2016 U.S. presidential elections. Despite these claims, there has been limited evidence to date linking fake news directly to voting behavior. In this project, I seek to provide credible evidence on this question by using big college football games as an instrument for fake news consumption, during the lead-up to the 2016 presidential election. During this period, I find that search volumes for pro-Trump fake news terms were lower in counties close to college football teams that played a big game when a piece of pro-Trump fake news went viral, and also that these counties were less likely to vote for Trump. The magnitude of these estimates suggests that a one-standard deviation increase in search volume for pro-Trump fake news terms increased Trump’s vote share in 2016 by about 5.4–7.4 percentage points. In addition, I find that the relationship between fake news exposure and voting behavior is approximately linear, by developing and estimating a novel nonparametric IV estimation method.
Public awareness of mental health issues has grown in recent years, and there is a common perception that mental health in the population is worsening, concerns that are supported by descriptive evidence. However, changes in public attitudes towards mental health make the interpretation of these trends challenging: low response rates to mental health surveys make their results sensitive to changes in sample selection bias, and even diagnosis rates in comprehensive data sets such as the Medicare data are a function of individuals' willingness to seek professional help. In this paper, I address these measurement issues using a comprehensive data set on all nursing home residents, so it does not suffer from sample selection bias. Similar to Medicare data, it contains information on whether each resident was diagnosed with depression in the recent past, but in addition, it contains a rich set of psychosocial measures, which provides us with a detailed picture of the resident's underlying mental health. I find that while depression diagnoses at admission increased from 19 to 24 percent for residents admitted to a nursing home in California between 2000 and 2010, underlying mental health of these residents (based on observed psychosocial behavior as well as machine-learning predictions using hundreds of covariates) was roughly constant over the same period. These results illustrate the perils of inferring mental health trends from survey evidence or diagnosis trends alone.
Public awareness of mental health issues has grown in recent years, and there is a common perception that mental health in the population is worsening, concerns that are supported by descriptive evidence. However, changes in public attitudes towards mental health make the interpretation of these trends challenging: low response rates to mental health surveys make their results sensitive to changes in sample selection bias, and even diagnosis rates in comprehensive data sets such as the Medicare data are a function of individuals' willingness to seek professional help. In this paper, I address these measurement issues using a comprehensive data set on all nursing home residents, so it does not suffer from sample selection bias. Similar to Medicare data, it contains information on whether each resident was diagnosed with depression in the recent past, but in addition, it contains a rich set of psychosocial measures, which provides us with a detailed picture of the resident's underlying mental health. I find that while depression diagnoses at admission increased from 19 to 24 percent for residents admitted to a nursing home in California between 2000 and 2010, underlying mental health of these residents (based on observed psychosocial behavior as well as machine-learning predictions using hundreds of covariates) was roughly constant over the same period. These results illustrate the perils of inferring mental health trends from survey evidence or diagnosis trends alone.
Cigarette Consumption and Tax Salience
This paper studies how cigarette consumption responds over time to changes in tax rates. Using a panel of state data, I estimate that the cumulative effect of an excise tax rise on consumption is larger than the cumulative effect of an increase in sales tax, in line with a theory of tax salience. In addition, I find that consumption falls in advance of an excise tax hike, whereas it only falls in the year after a sales tax increase. The pattern of consumption response to sales taxes is also consistent with consumer learning over time.
Cigarette Consumption and Tax Salience
This paper studies how cigarette consumption responds over time to changes in tax rates. Using a panel of state data, I estimate that the cumulative effect of an excise tax rise on consumption is larger than the cumulative effect of an increase in sales tax, in line with a theory of tax salience. In addition, I find that consumption falls in advance of an excise tax hike, whereas it only falls in the year after a sales tax increase. The pattern of consumption response to sales taxes is also consistent with consumer learning over time.
Page updated
Google Sites
Report abuse