Health Policy

• “Medicaid and Mortality: New Evidence from Linked Survey and Administrative Data,” by Sarah Miller, Sean Altekruse, Norman Johnson, and Laura R. Wherry. July 2019. NBER #26081.

How should we evaluate health insurance policy initiatives? One obvious metric is whether increases in the population covered by health insurance reduce mortality. A previous Working Papers column (Summer 2019) reviewed a paper that described the statistical difficulties in ascertaining whether publicly subsidized health insurance expansions reduced mortality. Because most people are insured and mortality rates among the non-elderly are low, the detection of an incremental mortality reduction from increased insurance coverage in the entire population would require a sample size of 40 million people. Because such a study is not feasible, the use of general mortality rates to evaluate the effects of policies to expand coverage is also not feasible.

This paper modifies that conclusion in an important way. If the analysis could be confined to a smaller population in which many people were not insured and the existing mortality rate was higher, the sample size required to detect an effect would be much smaller. This paper studies only individuals ages 55–64 who had incomes equal to or lower than 138% of the poverty level or whose educational attainment was less than high school. The annual mortality rate in this sample was 1.4%.

The paper compares mortality rates in those states that expanded Medicaid coverage under the 2009 Affordable Care Act to those states that did not. Prior to Medicaid expansion, the mortality rates among those in the expansion-eligible population were parallel in the two groups of states and not statistically distinguishable.

After Medicaid expansion, the mortality rate in the near-retirement-age low-income population in the expansion states fell to 1.27%. Given an estimate of 3.7 million people in expansion states ages 55–64 whose incomes were less than 138% of the poverty line and thus eligible for Medicaid under expansion, about 4,800 fewer people died annually because of Medicaid expansion.

Mutual Funds and Antitrust

• “Common Ownership Does Not Have Anti-Competitive Effects in the Airline Industry,” by Patrick Dennis, Kristopher Gerardi, and Carola Schenone. July 2019. SSRN #3423505.

Institutional investors that sponsor index funds — Vanguard, BlackRock, Fidelity, etc. — are now among the largest shareholders of most publicly traded companies. They frequently hold significant stakes in all the firms within a market. Critics of this pattern of “common ownership” theorize that such institutional investment could reduce market competition and increase consumer prices. There are empirical papers in the published literature that ostensibly demonstrate such harm is occurring.

A previous article in Regulation (“Calm Down about Common Ownership,” Fall 2018) criticized the common ownership thesis on logical grounds. The diversified shareholders of index funds not only own all airline stocks or bank stocks (the industries examined in two empirical articles that concluded that consumers are hurt by common ownership), they also own all other firms, including firms whose profits would be hurt by high prices for airline and banking services. So why would shareholders want to hurt themselves by exercising market power in some economic sectors that would reduce profits in (many) other sectors?

Another logical argument in the Regulation article noted that the mutual fund companies that offer index funds also offer managed funds, which differ in their ownership stakes of firms within industries. One Vanguard managed fund, for example, owns similar amounts of American, Delta, and United Airlines but no Southwest, while another Vanguard fund owns Southwest but no American, Delta, or United. Thus, the economic interests of the mutual fund companies regarding intra-industry competition are not obvious.

Finally, the article argued that the measure of common ownership used in the empirical studies (the “modified Herfindahl–Hirschman Index” [MHHI]) was not a pure measure of common ownership, but also of market share. Thus, increased demand could cause both prices and market share — and this measure of common ownership — to increase.

This article explores that statistical problem in detail. The conceptual problem is that the MHHI includes both ownership and control terms as well as airline market shares. This is important because the widespread interpretation of the empirical results in the literature is that increased common ownership by institutional investors increases airline ticket prices. If the results are instead driven by variation in airline market shares, then the policy implications are unclear.

The authors conclude that the positive relationship between the measure of common ownership and airline ticket prices is the result of variation in airline market shares rather than variation in common ownership among institutional investors. The paper demonstrates this result by constructing two alternative “placebo” measures of common ownership: one in which variation in market shares is muted by design while variation in common ownership is retained, and a second measure in which variation in common ownership is muted by design while variation in market shares is retained. The authors find that the first placebo measure is uncorrelated (or negatively correlated in some specifications) with average prices, while the second measure is positively correlated with prices. Thus, it is variation in market shares, not ownership, that drives the correlation between this measure of common ownership and prices.

Consumer Credit

• “Does Price Regulation Affect Competition? Evidence from Credit Card Solicitations,” by Yiwei Dou, Geng Li, and Joshua Ronen. March 2019. SSRN #3361050.

• “The Economic Consequences of Bankruptcy Reform,” by Tal Gross, Raymond Kluender, Feng Liu, et al. September 2019. NBER #26254.

The Credit Card Accountability Responsibility and Disclosure Act (CARD Act) was enacted in 2009 as an ostensible enhancement of consumer protection. A previous Working Papers column (Winter 2014–2015) reviewed a paper that concluded that consumers benefited from the legislation because fee reductions mandated by the law occurred without any offsetting interest rate increases.

Another provision of the law prohibits interest rate increases on new transactions within the first year of opening an account and on existing balances in the first year (except when the rate on existing balances was explicitly described as a time-limited introductory rate). The first of these papers examines how these provisions affected consumers.

The authors examine credit card mail solicitation offers from 2001 through 2016, comparing offers to consumers, which are regulated, to offers to small businesses, which are not. The authors use the solicitations to examine the responsiveness of offers to those of other companies. Given that the CARD Act regulated only interest rate increases, how did firms respond to interest rate decreases of other firms? The paper concludes that issuers’ offered interest rates to consumers were about 1.3 percentage point higher because of reduced responsiveness to competitors’ interest rate offers in the post–CARD Act period.

Relatedly, the consumer bankruptcy rate rose from 0.3% of households in the early 1980s to 1.5% in the early 2000s. In 2005, Congress passed the Bankruptcy Abuse Prevention and Consumer Protection Act, which made filing for bankruptcy more difficult and expensive. Reform supporters argued that the savings from reduced bankruptcy (the greater amount of loans repaid to banks) would be passed on to consumers in the form of lower interest rates.

The second of these papers finds that reform reduced the bankruptcy rate by 50%, resulting in roughly 1 million fewer bankruptcies in the two years after reform (even after accounting for the dramatic increase in bankruptcy filings just before the reform took effect). The authors conclude that 60–75% of the savings from reduced bankruptcies were passed on to consumers in the form of lower interest rates.

Franchise Laws

• “Trouble Brewing? Impact of Mandated Vertical Restraints on Craft Brewery Entry and Production,” by Jacob Burgdorf. September 2018. SSRN #3392392.

Manufacturers and distributors need each other, but distributors often believe that manufacturers make money at distributors’ expense through what economists describe as opportunism. In this narrative, distributors engage in costly, irreversible investments that help the sales of their product, only to have their distribution relationship subsequently terminated in an attempt by the manufacturer to expropriate the profits from the distributors. Or manufacturers force distributors to purchase products during sales slumps so that distributors, rather than manufacturers, suffer wealth losses.

In some industries, distributors use the political system to mandate the nature of the relationship between manufacturers and distributors. The contentious history between auto manufacturers and dealers led dealers to lobby successfully in all 50 states for legislation that prohibits manufacturers from forcing dealers to accept unwanted cars, protects dealers against termination of franchise agreements, and restricts additional franchises in a franchised dealer’s relevant market area. The laws also prohibit a manufacturer from selling directly to the public through vertical integration. Tesla has run afoul of this provision in its attempts to sell cars directly to the public, bypassing the existing dealer network. (See “Tesla and the Car Dealers’ Lobby,” Summer 2014.)

This paper describes an analogous history in the beer industry. Almost all states have beer franchise laws. They were passed in the 1970s as wholesalers became concerned about increased consolidation among national brewers. The laws make renewal of brewer distributor contracts more or less automatic and the termination of such contracts extremely difficult. Twenty states, though, allow brewers to distribute their own beer without restriction. This freedom puts an important check on the role of distributors.

The role analogous to Tesla in the beer story is played by craft brewers. In states with beer franchise laws but with no possibility of self-distribution, craft brewers had to use distributors that were oriented toward serving the large national brands. Craft brewers often felt ill-served by these mandated distributor relationships.

Bell’s Brewery, a highly regarded craft brewer in Kalamazoo, MI, illustrates the defects of mandated franchise restrictions. In 2006, Bell’s Chicago wholesaler was owned by National Wine and Spirits (NWS), which planned to sell the rights to distribute Bell’s products to another wholesaler. Bell’s opposed the sale, worrying that its beers would not be promoted well by the new wholesaler. Rather than engaging in a costly legal battle trying to end the wholesale contract, Bell’s pulled distribution of its beer out of the entire state of Illinois, despite the state comprising over 10% of Bell’s sales. Exiting the state allowed Bell’s to end its contract with NWS. Bell’s returned to Illinois nearly two years later, once NWS lost its wholesale license and the right to sue.

This paper examines craft brewer entry from 1984 through 2016. The average was 0.880 breweries per million people per year. But in states that had franchise laws that restrict brewery distribution and require the use of independent wholesalers, net entry decreased by 30–60% (0.320 to 0.518 breweries per million people per year). In the states in which brewers could self-distribute, the decrease was just 0.16 breweries per million people per year.