close× Call Us
close×
Title Obamacare Enrollment and Consumer Fairness
Post date 02/15/2015
C1 Background and Explanation of Rationale

"This experiment tests the effects of consumer fairness norms on policy uptake. In particular, we want to know if framing a policy as offering closely aligned costs and benefits will cause individuals to enroll in a policy more so than if the policy is framed as providing maximal individual benefits. We have partnered with Land of Lincoln Health, an Illinois healthcare insurance cooperative,that is federally funded as part of ""Obamacare,"" or the Affordable Care Act of 2010, to increase healthcare enrollment in the state. Land of Lincoln has distributed 10,000 mailers which we designed to to randomly selected Illinois residents. 5,000 of the mailers frame health insurance as providing aligned costs and benefits, while another 5,000 frame insurance as being in the best interest of the individual consumer. Land of Lincoln will make results available to us by the end of February, after the enrollment window has closed.

The general theory behind this experiment holds that citizens import norms from the consumer realm into the policy realm, and that the consumer fairness norm in particular can better explain policy uptake than self-interest frames. The design is indebted to Bertrand et al (2010), who examined the efficacy of advertising via a randomized mail design in South Africa. For that study, the authors partnered with a commercial lender. In this case, we will study the effects of advertising in a randomized experiment about public policy, in which we have partnered with a policy provider. Specifically, we are investigating whether consumer fairness frames that have been found to work in the consumer realm also affect policy uptake.

We are bringing the literature on consumer fairness, which has flourished in economics and social psychology, into political science. Consumer fairness is grounded in everyday consumer behavior which, with repetition, produces its own norms; the shared etymological origins of norms and normal are instructive in this regard. Scholarship has coalesced around a version of consumers fairness in which individuals expect the costs they pay to be roughly in line with the costs that firms bear, along with a profit margin that does not exceed the profits they perceive firms in similar contexts to be drawing. When consumers think fairness has not been achieved, they will penalize the unfair firm. Conversely, they are willing to pay higher prices when, contextually, they believe it is fair to do so.

Kahneman, Knetsch and Thaler (1986a) used survey data to show that concerns about fairness effectively prevented firms from maximizing profits. The vast majority of their respondents regarded a firm that would raise shovel prices after a snowstorm as engaging in unfair behavior; the same was true for a car dealership that raised prices for a particularly in demand yet hard-to-come-by automobile. One of their chief contributions was showing that consumers do not reflexively think of profit-making as unfair. In one vignette, survey respondents were told about a landlord whose rising costs forced him to raise the rent on a poor tenant, even though the increase might force the tenant to move. 75% of respondents found the action to be fair. Given certain conditions, consumers do not begrudge firms a profit motive. Respondents viewed salary cuts undertaken by a struggling company as fair, but unfair when undertaken by a profitable company. Ultimately, consumer fairness flows from a dual entitlement principle that binds both buyers and sellers. Buyers---consumers---are entitled to a purchase price roughly comparable to a reference transaction,"" or a transaction that recently occurred within both firm and customers' community or frame of reference. Sellers, meanwhile, are entitled to a profit within the limits of the reference transaction"" (Kahneman, Knetsch and Thaler 1986a; Kahneman, Knetsch and Thaler 1986b).

Evidence abounds that consumer fairness matters outside the laboratory. Consumers weigh the fairness of a particular price when deciding whether to purchase it (Martins and Monroe 1994). When they feel that a vendor is not being fair, consumers are more likely to be dissatisfied with the good sold, if not outright angry (Oliver and Swann 1989a and b). They are also likely to tell other consumers about the perceived unfairness, leading to negative costs for the vendor (Zeelenberg and Pieters 2004). For their part, vendors are acutely aware of consumers' perceptions about unfairness. Examples testifying to vendors' concerns with fairness are all around us. Why, for example, do events such as concerts and sporting events sell out well before all the demand has been captured and reflected in higher possible prices (Okun 1981; Kahneman, Knetsch and Thaler 1986a)? Why do popular restaurants not engage in dynamic pricing"" and charge extra for reservations during peak hours? In both cases, firms are willfully leaving excess demand---and thus profits---on the table. The likely explanation is that firms anticipate that such actions would generate accusations of unfairness, sowing the seeds of resentment and diminishing future profits. Indeed, recent experimental work has found that consumers do regard dynamic pricing as an unfair practice (Haws and Bearden 2006). "

C2 What are the hypotheses to be tested?

"H1: The mailers that frame healthcare as fair will be significantly more effective at causing Illinois residents to sign up for Land of Lincoln Health, compared to the mailers that frame healthcare in more traditional ways related to self-interest.

H2: Residents who receive the fairness mailer will enroll with Land of Lincoln at significantly higher rates than those who receive the traditional mailer, even when accounting for their proximity to the nearest medical centers that offer coverage under Land of Lincoln Health.

H4: Democrats who receive the fairness mailer will be most likely to sign up for Land of Lincoln. Republicans who receive the traditional mailer will be the least likely.

H5: Enrollment and recent voting participation will be positively correlated. Residents who have voted more often will be more likely to enroll, regardless of the mailer they receive."

C3 How will these hypotheses be tested? *

As a first pass, we will perform simple difference in means tests between the enrollment rates generated by the two mailers. We will also model enrollment as a function of treatment and available covariates (those that do not violate HPAA), including party identification, distance from medical facilities covered under Land of Lincoln, race, ethnicity, and recent voting history. To measure electoral history identification, we are relying on publicly available data about voting. To measure party identification, race and identify, we are relying on proprietary estimates provided by Catalist, the research firm. For party identity, we will use 2012 presidential preference as a proxy. Race and ethnicity are provided by Catalist. To measure distance between residents and the nearest Land of Lincoln facilities, we will rely on GIS data provided by Land of Lincoln about the location of their facilities.

C4 Country
C5 Scale (# of Units) not provided by authors
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number not provided by authors
C9 Date of IRB Approval not provided by authors
C10 Will the intervention be implemented by the researcher or a third party? Researchers
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? No
C13 JEL Classification(s) not provided by authors