Context: Since the ratification of the WHO Framework Convention on Tobacco Control (WHO FCTC) in 2004, Pakistan has taken various measures of tobacco use prevention and control. We examine how household level tobacco consumption and tobacco expenditure evolved after 2004.

Methods: We use nationally representative household survey data of 2004-05 (base year), 2007-08, 2011-12, and 2015-16 to estimate the likelihood of a household incurring tobacco expenditure on any-, smoking-, and smokeless-tobacco and compare the adjusted odds ratios for respective years. We estimate the adjusted differences in average tobacco expenditure shares relative to the base year for all households (extensive margin) and for tobacco user households only (intensive margin). We also analyze concentration curves and indices generated by plotting the cumulative tobacco expenditure shares at each percentile over percentiles of household consumption per capita to examine the how burden of tobacco expenditure share changed from 2005 to 2016.

Results: We find that the odds ratio for consuming smoking tobacco, compared to 2005, gradually declines from 1.01 in 2008 to 0.70 in 2016; whereas for smokeless tobacco, it increases from 0.60 to 0.76 during the same period. Average expenditure share decreases by 21.4% and 16.7% at the extensive margin and increases by 6.1% and 10.6% at the intensive margin, respectively for smoking and smokeless tobacco from 2005 to 2016. The concentration curves, depicting the burden of tobacco expenditure share, shifts upward, suggesting that the households at the bottom deciles are bearing greater burden of tobacco expenditure share in 2016, compared to that in 2005.

Conclusion: Our study demonstrates considerable differences in changes in smoking and smokeless tobacco consumption, and tobacco use by economic status. These findings could inform policies for special attention on smokeless tobacco control measures, and for effective strategies to further reduce tobacco consumption of poor households.