Abe Dunn

Designing effective public policies depends on accurate and timely economic measurement. Big data promises substantial benefits for improving economic measurement but also presents significant challenges. This paper introduces a framework for quantifying the usefulness of big data for specific applications, relative to official statistics. We weigh the potential benefits of additional granularity and timeliness while examining the accuracy associated with any new or improved estimates, relative to comparable accuracy produced in existing official statistics. We apply the methodology to improving timely regional employment estimates. We find that using data from a payroll processor reduces out‐of‐sample error in state employment estimates by 11%. Additionally, we produce new county‐level estimates offering more timely, granular insights than previously available. Applying a novel test, we cannot reject the hypothesis that the new county estimates have an accuracy in line with official measures. An application to COVID highlights the practical benefits of these estimates.

Economic geography data are typically reported using political units, such as counties, which often do not match economic units. Commuting zones (CZs) group counties into labor markets. However, CZs are not the most appropriate grouping for other economic activities. We introduce consumption zones (ConZs), groupings of counties appropriate for the analysis of consumption. We apply CZ methodology to payment card data, which report spending flows across US counties for fifteen industries. We find that different industries have different market sizes. Grocery stores have more than five times the number of ConZs as entertainment. We apply ConZs to measuring industry concentration.

We investigate alternative methods for constructing quality-adjusted medical price indexes both from a theoretical perspective and empirically usingmedical claims data. A method based on utility theory produces the most robust and accurate results, while the alternative methods used in recent work overstate inflation. Based on claims data for three medical conditions, we find declining prices across each condition when properly adjusted for quality. The methodology that is applied and assumptions made in the formation of the index have substantive effects on the magnitude of the quality-adjusted price index.

We examine the impact of Medicare Part D on mortality for the population over the age of 65. We identify the effects of the reform using variation in drug coverage across counties before the reform was implemented. Studying mortality rates immediately before and after the reform, we find that cardiovascular-related mortality drops significantly in those counties most affected by Part D. Estimates suggest that up to 26,000 more individuals were alive in mid-2007 because of the Part D implementation in 2006. We estimate the welfare benefit from lives saved to range from $1.5 to $4.8 billion.

We introduce a new source of detailed data on spending by medical condition to analyze US health care spending growth in the period 2000–14. We found that thirty conditions, which represented only 11.5 percent of all conditions studied, accounted for 42 percent of the real growth rate in per capita spending during this period, even though they accounted for only 13 percent of overall spending in 2000. Primary drivers of spending growth included the use of new technologies, a shift toward the provision of preventive-type services, and an aging and more obese population. The health benefits of many new technologies appeared to outweigh the associated expenditures on treatment, which indicates that these are cost-effective and provide a net value to society. However, while these technologies may be of value, new treatments are often more expensive than older ones.

We study the impact of competition among physicians on service provision and patients' health outcomes for the U.S. commercial market. We focus on cardiolo-gists treating patients with a first-time heart attack treated in the emergency room. Physician concentration has a small, but statistically significant effect on service utilization. Cardiologists in more concentrated markets perform more intensive pro-cedures, particularly, diagnostic procedures–services in which the procedure choice is more discretionary. Higher concentration leads to fewer readmissions but no effect on mortality. These findings suggest that changes in organizational structure, such as a merger of physician groups, not only influence the negotiated prices of services, but also service provision.