Regressive or Progressive? The Effect of Tobacco Taxes in Ukraine

Tobacco taxes are usually considered regressive, as the poorest individuals allocate larger shares of their budget toward the purchase of tobacco-related products. However, because these taxes also discourage tobacco use, some of the most adverse effects and their economic costs are reduced, including lower life expectancy at birth, higher medical expenses, increased years of disability among smokers, and the effects of secondhand smoke. This paper projects the effects of an increase in the tobacco tax on household welfare in Ukraine. It considers three price-elasticity scenarios among income deciles of the population. The results show that although tobacco taxes are often criticized for being regressive in the short run, in a more comprehensive scenario that includes medical expenses and working years, the benefits of tobacco taxes far exceed the increase in tax liability, benefitting in large measure lower income households. The results also indicate that lower health expenditure seems to be the main driver, because of the reduction in tobacco-related diseases that require expensive treatments. Tobacco taxes are also associated with positive distributional effects related to the higher long-term price elasticities of tobacco consumption.


Introduction
In the wake of the outbreak of the pandemic COVID-19, citizens in many countries are urged to stay at home and to reduce social contacts to a minimum. The degree to which economic activity is impaired by such social distancing measures largely depends on the capacity of firms to maintain business processes while many employees stay at home. The German industry with the highest share of workers registered for short time work in March 2020 was "Accommodation and Food Service Activities", one of the industries with the least scope for remote work. Conversely, the share of short-time work in the industry with the highest calculated capacity to work from home, "Financial and Insurance Activities", is close to zero. 1 At the individual level, the possibility to carry out work remotely can reduce the risk of exposure to both the disease and the economic shock associated with governmental actions against the pandemic.
In this paper we provide an estimate of the overall capacity of German employees to work from home (WfH) and document how the feasibility of remote work varies across occupations, industries and regions in Germany. We investigate which groups of employees are particularly vulnerable to the consequences of COVID-19 shutdowns, due to the fact that their jobs cannot be performed remotely. Finally, we identify work activities and job features that are most closely correlated with the feasibility of WfH.
Up until the COVID-19 pandemic outbreak, Germany's share of employees who work from home "sometimes" or "usually" was below the European average (Figure 1). In European countries in which remote work is most common, such as Sweden or the Netherlands, the proportion of employees working at least partially from home is about three times higher than in Germany. This disparity is arguably driven not only by cross-country heterogeneity in industry composition and technical viability, but also by cultural differences. Whatever the reason, in the course of the COVID-19 crisis, such hurdles to WfH will be quickly dismantled wherever possible. The policy-relevant questions now are: How many jobs can possibly be performed at home? How does the capacity to WfH vary across industries and regions? What groups of employees are particularly vulnerable as they lack the possibility to work from home? Figure 1 about here We find that roughly 56 percent of all jobs in the current German economy can plausibly be performed at home. In comparison, Dingel and Neiman (2020) and Del Rio-Chanona et al. (2020) calculate the WfH potential to be about 37 percent and 43 percent in the U.S. economy. In contrast to these studies, which base their measure on plausibility judgements, we compute a measure for WfH feasibility that relies on employees' own assessment concerning the feasibility to perform their jobs from home.

Data and Empirical Approach
Our measure for WfH feasibility builds on survey information from 17,160 employees (aged 18-65) from the 2018 wave of the BIBB/BAuA Employment Survey. More precisely, we suppose that a job cannot be performed at home if the respondent indicates that WfH is "not possible" even if the employer were to grant the option. 2 We subsequently aggregate this information (population weighted) to the occupation level using the 2-digit German Classification of Occupations (KldB 2010). We then combine this measure with administrative data from the Federal Employment Agency (BA) on occupational employment counts in the current German economy overall, by sector, and by county.
Similar to other studies, our results represent upper bound estimates for the actual WfH potential for two reasons: First, our measure cannot account for the fact that some jobs cannot entirely be performed at home. Second, to the extent that relocating work to peoples' homes is associated with costs (e.g. requirements for new technical equipment or decline in workers' productivity), a notable portion of the WfH capacity might not be exploited in equilibrium.

Results
We find that overall, roughly 56 percent of jobs in the current German economy can plausibly be performed at home. Figure 2 reports the capacity to WfH as well as the share of employees actually working from home (frequently or occasionally) overall and broken down by industry. Industries are displayed in descending order according to their contribution to GDP. The figure shows that in the pre-pandemic German economy less than half of the overall WfH capacity was exploited.

Figure 2 about here
Moreover, there is a considerable variation in WfH potential across industries. Sectoral WfH capacity ranges between 37 and almost 90 percent. The magnitude of a sector's "untapped WfH capacity" provides information about the potential adjustment costs in a sector that would occur in order to exploit the full WfH capacity. For example, while almost 90 percent of employees in "Financial and Insurance Activites" could in fact work from home, only 38 percent did so in 2018. Hence, the number of employees working remotely would have to more than double to exploit the full WfH potential in this industry, which might put a strain on corporate IT systems. Table A2 in the Appendix reports the results for each industry at the 2-digit NACE level. Figure 3 depicts the geographic distribution of WfH capacity across German counties. The map reveals a clear divide between East and West and between urban and rural regions. While on average 59 percent of employees in West Germany (including Berlin) can perform their job from home, in East Germany (excluding Berlin) only 50 percent of employees can do so. Even more striking are the urbanrural differences in WfH capacity: WfH capacity amounts to roughly 65 percent in cities with 500,000 inhabitants or more, versus on average 53 percent in the rest of the country. The breakdown of WfH capacity by employees' education, income, gender, and domestic childcare duties is displayed in Figure 4. It is striking that the feasibility of Wfh increases strongly with higher education qualification and income.
Women exhibit an about 9 percentage points higher WfH capacity than men, mainly due to women's larger untapped WfH capacity. Within gender groups, employees with young children (below age 11) in the household show a 3-4 percentage point higher WfH capacity than employees without such domestic childcare duties.  Table A1 in the Appendix reports the corresponding numbers. At the top of the WfH-feasibility distribution range occupations in computer science and ICT as well as occupations in advertising and marketing. Examples for jobs with rather low WfH capacity are drivers and operators of vehicles and transport equipment or occupations in the field of interior construction.

Figure 5 about here
In addition to the calculation of the WfH capacity in the overall German economy, we use our measure for WfH feasibility at the employee level to identify job tasks and characteristics that are most correlated with WfH feasibility. This exercise may be helpful to calculate WfH capacity in circumstances in which an employerbased assessment is unavailable or to validate previous efforts to calculate WfH feasibility based on idiosyncratic subjective judgements at the task level. To this end, we regress the available information on job tasks on employee's assessment of WfH feasibility in a simple logit model. Average marginal effects from this model are reported in Figure 6. We find that the top-3 job characteristics that are significantly positively associated with the feasibility of working from home are "Using computers, the internet or e-mail processing", "Developing, researching, constructing" and "Working in seated position". In contrast, the features "Working standing up", "Transporting, storing, shipping" and "Nursing, caring, healing" are significantly negatively correlated with WfH feasibility.

Discussion and Conclusion
Due to emergency confinement measures in the current COVID-19 crisis, employees' capacity to work from home has become an important instrument for firms to prevent business failure. Hence, our estimate of the overall capacity to work from home among German employees is highly relevant for researchers and policymakers attempting to predict economic activity during the COVID-19 shutdowns. Additionally, the breakdown of WfH capacity by industries and counties can be helpful in directing assistance to the most deprived industries and regions as well as to evaluate where a selective and gradual exit from economic shutdown is most/least urgent. Once public health considerations are taken into account, one might argue that those industries and regions displaying high WfH capacities should have lower priority when reducing restrictions.
Our results document which groups of employees are most vulnerable to the consequences of COVID-19 shutdowns, due to the fact that their jobs cannot be performed from home. We find disproportionately high levels of vulnerability for the low-skilled and low-wage earners. Gender differences are less stark but indicate higher WfH capacity for women and for employees with young children in the household. To some extent, women with small children might hence be temporarily sheltered from income losses due to COVID-19 restrictions. However, these employees are also more likely to be affected by increased stress levels due to the necessity to reconcile work and childcare, especially if their partners do not have the possibility to work from home.