How do high deductibles influence consumer behavior?
Zarek C. Brot-Goldberg, Amitabh Chandra, Benjamin R. Handel, and Jonathan T. Kolstad│ April 14, 2019
Policy brief written with the assistance of Emma Fernandez
Summary of: What Does a Deductible Do? The Impact of Cost-Sharing on Health Care Prices, Quantities, and Spending Dynamics. Brot-Goldberg, Z. C., Chandra, A., Handel, B.R. & Kolstad, J.T. (2017). The Quarterly Journal of Economics vol. 132, no. 3.
Curbing U.S. health care spending
Spending on health care in the United States is growing rapidly. In 1960 spending was equivalent to 5% of the GDP, but by 2014, it had risen to 17.5%. Increasingly, policymakers are wrestling with the difficult question of how to constrain ballooning spending without reducing the quality of medical care accessed by consumers. In recent years, demand-side interventions, or policy shifts that expose consumers with insurance to more of the cost of health care, have been popular with employers and policymakers. The percentage of consumers with employer insurance who paid deductibles over $1,000 nearly doubled from 22% in 2009 to 41% in 2014.
In theory, high deductible plans reduce spending by using the price of health care as a mechanism to get consumers to cut back on unneeded medical expenses. Past evidence has indicated that higher health care costs do reduce spending on health care, but hasn’t shown how this reduction occurs. Without this missing evidence, it is difficult to know whether consumers understand the price incentives of their plans or whether they are reducing needed as well as unneeded care. The fundamental theory of high-deductible plans rests on the assumption that the increased share of costs paid by consumers will cause them to make informed trade-offs – reducing lower-value care and “price shopping” for providers offering the same services at cheaper prices. But whether or not this actually occurs will have crucial implications for the success of this type of plan.
How do high deductible plans affect consumers?
To understand how high deductible plans change consumer spending, we studied a shift in health insurance options at a large U.S. firm. The firm we studied had between 35,000 and 60,000 employees, with 105,000 to 200,000 people covered under its health care plans when dependents were counted. (To preserve the anonymity of this firm, we only provide a range for our descriptive statistics.) Originally, the firm offered a free insurance plan, which almost all employees enrolled in. During our observation period, the firm discontinued this option and moved to a high-deductible insurance plan. The major shift under the new plan was in the pricing structure, but employees still had access to the same providers and medical services. We were able to observe extensive data on the firm’s employees for four years before the switch and two years after, including individual health claims, characteristics of the employees, and HSA and 401(k) contributions. Overall, employees of the firm were high-income, well-educated and tech-savvy. Post-switch, they also had access to a sophisticated online tool to help consumers search for lower medical prices. In other words, this scenario represents close to the “ideal” situation for educated consumer choices.
Consumers do not price shop
After the firm changed their insurance plan, consumers did reduce health care spending. Overall, there was a spending reduction between 11.1 and 15.4%. However, our analysis found no evidence of consumer price shopping. Essentially, all consumer spending reduction was achieved through quantity reduction. In other words, consumers accessed fewer health care services overall, rather than spending less money on the services they accessed. To examine this in detail, we observed the total spending over time on consumers’ top 30 procedures. Rather than seeing reductions in specific procedures, indicating that consumers targeted expensive services, we saw spending reductions across the board for procedures. We also found that consumers reduced spending both on non-essential or non-effective treatments (“low-value” care) and on “high-value” care like preventive care, mental health care, physical therapy, and drugs for diabetes, cholesterol, depression, and hypertension. Not all consumers reduced spending equally; in fact, the majority of the overall spending reduction across the firm came from the sickest three-quarters of those covered. We also found that consumers do not respond to the complex benefit structure the way one would if they were rational and forward-looking (i.e. behaved as standard economic theory predicts). Instead, we find that consumers reduced their health care spending by 42.2%, about three times more than the overall average reduction, in months when they were under the deductible. This indicates that consumers responded to the immediate out-of-pocket price when making health care decisions, rather than to the overall pricing scheme of their insurance plan.
Health Care Spending Over Time
Our results paint a picture of demand-side health care policies for reducing costs as blunt instruments even in the best settings. Our findings provide clear evidence that consumers do not respond to these policy changes as anticipated. This indicates that, though they do achieve the cost savings intended, these policies have negative impacts on consumer health care choices and are therefore unlikely to effectively improve overall efficiency in the health care system. Employers who have already implemented high-deductible plans for cost-saving purposes should also expect to see effects on the health and experience of their employees. We would also expect to see these effects in the high-deductible plans offered under the Affordable Care Act. Our study provides a counter to the traditional economic theory’s model of the outcomes of high-deductible plans. Our detailed analysis of administrative data allowed us to obtain a more nuanced understanding of this model based on real behavior. Our results show that moving forward, it is crucial to develop insurance policy design with real behavior, rather than unrealistic rational modeling, at its heart.