Cigarette taxes and older adult smoking: Evidence from recent large tax increases

Abstract

While recent evidence casts some doubt, it is generally accepted that the price sensitivity of smoking varies inversely with age. We investigate the responsiveness of older adult smoking using variation from recent historically large cigarette tax increases in the United States. Using data from the Behavioral Risk Factor Surveillance System from 2000 to 2005, we find consistent evidence that higher taxes reduced smoking participation by older adults, especially those who are less educated and live in low-income households. Our findings run contrary to existing evidence which suggests that cessation behavior by older adults is not sensitive to price. Since a large literature suggests smoking cessation even later in life reduces morbidity and increases longevity, our findings may represent substantial gains in health among tax-induced quitters.

Introduction

While recent evidence casts some doubt, it is generally accepted that smoking by youth and young adults is more price sensitive than that of older adults. This consensus is based on empirical evidence and on conventional wisdom that individuals with more experience with an addictive good, like cigarettes, are less sensitive to changes in its price.¹ Despite these findings, adult smoking behavior remains of interest, since a large literature implies that even late-in-life cessation can yield substantial improvements in mortality and morbidity.

Using data from the Behavioral Risk Factor Surveillance System (BRFSS) from 2000 to 2005, we investigate the price-responsiveness of older adult smoking using variation from recent historically large cigarette tax increases in the United States. Given the repeated cross-sectional nature of these data, we estimate standard two-way fixed effects models that include controls for other state-level smoking policies implemented over this time period. We find consistent evidence that higher taxes reduced smoking participation by adults aged 45–59 and a slightly older group, suggesting that cessation behavior is price responsive. In particular, our findings imply that a $1 increase in the cigarette excise tax leads to a 1.0–1.5 percentage point decrease in daily smoking participation among these groups. Off a base of roughly 18%, this translates into a reduction in smoking prevalence of between 6 and 8%. We also find that tax-induced reductions in smoking are fairly equal for men and women, but are especially large for relatively less-educated, lower-income and less-healthy individuals.

The paper proceeds as follows. The next section reviews existing literature on the impact of price on older adult smoking. In general, most studies rely exclusively on cross-sectional variation in taxes or prices and, to our knowledge, we are first to exploit the recent large tax increases that occurred in the U.S. after 2001. These increases provide ample variation for the identification of tax effects, relative to earlier periods. We also provide motivation for the study, which rests in part on the findings of others which suggest that even later-in-life smoking cessation can yield substantial health improvements. Section 3 discusses our data, paying particular attention to key measures and providing detailed information on our analysis sample. In addition to focusing on the size and breadth of these historically large tax increases, we present evidence which suggests that they were driven by state budget shortfalls, rather than state-specific collective preferences regarding tobacco. Section 4 discusses our empirical strategy which involves a standard two-way fixed effects framework. Section 5 presents our results, along with estimates from several robustness checks, which together imply that older adult smoking participation fell in response to the large post-2001 tax increases. We also present evidence of substantial heterogeneity in the impact of cigarette taxes on older adult smoking and, separately, discuss possible health implications of our findings. Section 6 concludes the paper.