How And Why Do Consumers Choose Their Payment Methods PDF Download

My dissertation consists of three chapters that address important questions in the payment card industry. Chapter 1 provides an overview of the institutional background, and introduces the problem of regressive distributional effects generated from credit card pricing. The regressive distributional effects arise when merchants pass on their costs associated with credit cards to all consumers by raising the retail prices. Since merchants typically do not differentiate prices across payment methods, these additional costs are cross-subsidized by cash and debit users. This induces a regressive transfer from low-income to high-income consumers because credit card usages tend to increase with income. Chapter 1 contributes to the literature by developing a measurement to quantify the regressive transfers made by the consumers to the merchants in a micro level. Using a unique shopping diary data conducted by Bank of Canada in 2013, I show that non-credit card users on average made a regressive transfer that is more than twice of that made by credit card users per transaction. The ratio of regressive transfers to transaction amount also decreases monotonically with income. These results suggest that how consumers choose between payment methods to make transactions have important implications on the distribution of regressive transfers, which motivates a structural estimation on consumer's payment method choices. Chapter 2 constructs a structural model of consumer adoption and usage choices, and uses the parameter estimates to simulate the counterfactual outcomes on the distributions of regressive transfers under various institutional changes. The model is built upon Huynh et al. (2021), which features a two-stage process where consumers first choose which payment bundle to adopt, then choose which payment method to use upon transaction. Heterogeneous preferences across consumer groups are estimated using a discrete-type of consumer demand model. Unlike most of the literature which ignores consumers’ choices between the issuer banks, the model considers consumers’ issuer bank choices among credit cards. Simulation results suggest that the model fits the observed data well, and generate reasonable demand elasticities of consumer usage and adoption probabilities. I conduct three policy experiments using the model estimates: a hypothetical removal of cash, a monopoly setting, and a perfect competition setting in the issuer banks. The results show that the regressive distributional effects are reduced under all three scenarios. Particularly, the monopoly setting has the strongest effects in the redistribution of regressive transfers, where it reduces the per-transaction and per-transaction value regressive transfers made by non-credit card users and low-income consumers, while increases those made by credit card users and high-income consumers. On the other hand, welfare comparisons show that perfect competition renders the highest increase in consumer surplus, while the monopoly setting and removal of cash on average hurt the consumers in terms of consumer surplus. This is the first paper to my knowledge that studies the regressive distributional effects with a structural demand model, and contributes to the literature by investigating the potential outcomes from changes in the market structure of the payment card industry. Chapter 3 builds upon the previous chapters and introduces dynamics into consumer's payment method choices. In particular, I ask how consumer awareness on merchant acceptance affects consumer's adoption and usage choices, and how information diffusion drives the adoption and usage curve over time. I extend the model developed in Chapter 2 by considering consumer awareness that varies between payment methods. Using the parameter estimates, I conduct policy experiments where I introduce a hypothetical new payment instrument in the market, assuming different consumer inform probabilities for existing instruments and the new instrument. Simulation results on post-introduction adoption and usage probabilities show that there is a large impact of consumer awareness on consumers' adoption decisions, with a bigger impact when assuming different inform probabilities for the new instrument. To understand how consumers' adoption and usage decisions change over time when consumer awareness evolves, I borrow the literature of diffusion and simulate a diffusion process of consumer awareness using the Bass Diffusion model (Bass (1969)). The simulation results show that the adoption and usage of new payment instrument exhibits an S-shaped curve after its introduction in the market, where it takes over six years to reach the convergence. Welfare analyses show that consumer surplus initially drops after the introduction, due to the lack of information, and gradually increases when consumers become more informed. This suggests that there is an impactful welfare loss associated with information failure, and it is important for the policy makers to develop measurement that ensures a quick diffusion of information when introducing a new payment method.