Job Market Website

Amy R. Wagner

Contact Information

Department of Economics
Northwestern University
2001 Sheridan Road
Evanston, IL 60208

Phone: 847-749-7834

[email protected]


Ph.D., Economics, Northwestern University, 2017 (expected)
M.A., Economics, Northwestern University, 2013
B.Econ. (Hons I), University of Queensland, 2009
B.Sc., Mathematics, University of Queensland, 2008

Research Fields

Health Economics

Teaching Fields

Health Economics, Applied Econometrics, Industrial Organization

Curriculum Vitae

Download Vita (PDF)

Job Market Paper

“Effect of physician-hospital financial integration on health outcomes and spending”

Download Job Market Paper (PDF)

Physician-hospital financial integration rates have surged over the past few years. The Medicare fee rules permit integrated physicians to charge higher prices for office visits. Meanwhile, industry stakeholders claim that integration may promote care coordination and quality, each of which may in turn improve patient health. Economic theory provides mixed evidence about the consequences of physician-hospital financial integration. I examine the effect of physician-hospital financial integration on health outcomes and spending using patient-year level Medicare data, and physician-year level integration data. I exploit the granularity of this data to make more precise estimates of the effect of integration on health outcomes than previous studies made. I address selection on both patient and physician unobservables by using an instrumental variables model with physician fixed effects. Addressing these two important sources of selection allows me to identify the causal effects of physician-hospital financial integration. I find that having an integrated primary care physician (PCP) does not significantly affect average mortality risk, but does reduce the risk of less severe adverse health outcomes attributable to conditions that are treatable in primary care settings. I also show that attending an integrated PCP does not increase health care spending. Physician-hospital financial integration could potentially impact patient health via several mechanisms. Some of these mechanisms predict that the effect of integration should be heterogeneous over patient characteristics. I test these predictions in the data. My results indicate that inter-specialty coordination is not playing a large role in the beneficial effects of physician-hospital financial integration. I also show that the higher prices associated with integrated PCPs do not reduce access to primary care. I find suggestive evidence that integration increases mortality for low income patients but decreases mortality for high income patients.

Other Research Papers

“The association between patient safety indicators and medical malpractice risk: evidence from Florida and Texas” with Bernard S. Black, and Zenon Zabinski, 2016, conditionally accepted at the American Journal of Health Economics.

Association between PSIs and medical malpractice risk (SSRN webpage)

According to tort law theory, medical malpractice liability may deter negligence by health care providers. However, advocates of malpractice reform often argue that most malpractice claims are unrelated to provider performance. We study the connection between hospital adverse events and malpractice claim rates in the two states with public datasets on med mal claim rates: Florida and Texas. We use Patient Safety Indicators (“PSIs”), developed by the Agency for Healthcare Research and Quality, to measure rates for 17 types of adverse events. Hospitals with high rates for one PSI usually have high rates for other PSIs. We find a strong association between PSI rates and malpractice claim rates with extensive control variables and hospital fixed effects (in Florida) or county fixed effects (in Texas). Our results, if causal, provide evidence that malpractice claims leading to payouts are not random events. Instead, hospitals that improve patient safety can reduce malpractice payouts. We also study local variation in adverse event rates, in the spirit of the Dartmouth Atlas work on variation in treatment intensity. We find large variations, both across counties and across hospitals within counties. This suggests that many adverse hospital events are avoidable at reasonable cost, since some hospitals are avoiding them.

“Effect of physician-hospital financial integration on care-coordination opportunities”, working paper.

One of the biggest challenges facing the US health care system is improving care coordination. Stakeholders in the health care industry claim that physician-hospital financial integration may promote better care coordination. In this paper I investigate the effect of physician-hospital financial integration on several mechanisms that are hypothesized to provide opportunities for care coordination: referral patterns, electronic medical record (EMR) use, and case-management program implementation. I also test the effect of physician-hospital integration on a negative outcome that should be mitigated by improved care-coordination: adverse drug interactions. I use claims data from the Medicare 5% sample of fee-for-service patients, physician survey data from the company SK&A, and hospital survey data from the American Hospital Association. I use physician-year level regressions with physician fixed effects to examine the effect of physician-hospital integration on referral patterns and EMR uptake. I use hospital-year level regressions with hospital fixed effects to investigate the effect of physician employment on hospital participation in case management programs. To determine the effect of physician-hospital integration on adverse drug interactions I conduct patient-year level regressions of an indicator for a drug poisoning incident on the vertical integration status of the patient’s primary care physician (PCP). I include physician fixed effects and instrument for the integration status of the patient’s current PCP with the current integration status of their original PCP.

“Effect of provider continuity on health outcomes: evidence from physician retirements and relocations”, work in progress.

The US health care system is highly fragmented. Provider continuity directly integrates care, which acts to improve patient health. However, patients who always use the same provider may deprive themselves of the benefits of specialization, which can impair health. In this paper I test the effect of having a consistent outpatient care provider on patient costs, health outcomes, and treatment compliance. Previous studies of these relationships likely overestimate the benefits of provider continuity due to ignoring reverse causality, unobserved patient characteristics, or unobserved physician characteristics. I account for these potential sources of bias by focusing on physician changes due to physician retirements and relocations, and by including physician fixed effects. My measure of provider continuity is the share of the patient’s care that is provided by their most frequently attended physician. I instrument for this with an indicator for the patient’s primary physician from the current or previous year retiring or relocating. I identify “candidate” physician retirements by a sudden drop off in claims in the Health Care Cost Institute Claims data. I only count these “candidate retirements” as actual retirements if the physician is older than the expected retirement age. My physician age data comes from State Medical Board databases.


Teaching Assistant, Northwestern University, 2012-2013. Intermediate Microeconomics, Introductory Microeconomics, and Introductory Macroeconomics
Teaching Assistant, University of Queensland, 2009. Introductory Microeconomics, and Introductory Macroeconomics


Prof. David Dranove (Committee Chair)
Prof. Matthew Notowidigdo
Prof. Craig Garthwaite

Keywords and phrases: Amy Wagner, Amy Wagner Northwestern, Amy Wagner Economics, Amy R. Wagner, Health Economics, Healthcare Economics, Economics of Healthcare Markets, vertical integration, physician-hospital integration, physician-hospital mergers, fragmented health care, provider continuity