Cowboys and Indians: Quasi-experimental evidence on cigarette brand loyalty

Despite much interest, there exists little economic research on cigarette brand loyalty. By contrast, there are many more such studies in the marketing literature, though the vast majority are inherently descriptive. In an attempt to move toward a more causal understanding of cigarette brand loyalty, we exploit a quasi-experiment that led to some cigarette brands becoming subject to a tax increase, while other similar brands were not. This quasi-random variation in cigarette price across similar brands was generated by a New York State (NYS) Supreme Court decision that allowed NYS to collect taxes from cigarette sales on Native American reservations (NARs) for the first time. This regime change effectively raised the price of cigarettes sold on NARs by roughly $4 per pack of twenty, though Native-produced cigarettes remained exempt from taxation. In other words, only non-Native brands faced a higher on-reservation price. There is much anecdotal evidence that these Native-produced cigarettes are highly similar to non-Native brands. Though considered non-premium, many Native-produced brands aim to replicate the taste and quality of popular premium cigarette brands by using comparable ingredients and manufacturing techniques; as a result, we consider them reasonable substitutes for premium cigarette brands. In this context, we ask the question: How did the loss of a source of untaxed premium brand cigarettes affect smokers’ choice of premium versus non-premiums brands?

Using data from the 2003-2012 waves of the New York State Adult Tobacco Survey, we model smokers’ choice of premium versus non-premium brand cigarettes in a generalized difference-in-differences framework. In particular, we allow the temporal impact of the policy change to vary by residential distance to a Native American reservation that sells cigarettes. Premium smokers who live closer to a NAR are more likely to respond to the policy regime change, due to their geographic proximity to untaxed Native-produced cigarettes. We find that the regime change led roughly one-quarter of premium brand smokers living close to a NAR to switch to non-premium brands, which include Native-produced cigarettes. Put differently, our results imply that roughly three-quarters of smokers residing close to a NAR did not switch from premium cigarettes, implying a high degree of brand loyalty given the proximity of reasonable substitutes at a much lower price. In particular, our main finding implies that roughly three-quarters of existing premium cigarette smokers were willing to pay at least $4 more to remain with their current premium brand. Our results are robust to many specification choices. Due to the short length of our post-period (i.e., 18 months), these results should be interpreted as inherently short-run in nature. Initial longer-run models, which extend the post-period until 2016, suggest that this high degree of brand loyalty in the short-run persists over longer periods of time. Our results may have implications for policies (e.g., plain packaging laws and laws that prohibit retailers from displaying cigarette brand information) that implicitly assume brand loyalty plays a strong role in smoking onset and duration.