Welfare Costs of Shopping Trips

Follow @yilmazkuday

 

Welfare Costs of Shopping Trips

One sentence summary: Welfare gains from removing traditional shopping-trip costs is about 4% for the average census block.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at the Annals of Regional Science.

The working paper version is available here.

 
Abstract

Using data on the number of visitors at the store level, this paper investigates the welfare costs of traditional shopping-trips for the U.S. census blocks. The investigation is based on an economic model, where individuals living in census blocks decide on which store to shop from based on the corresponding shopping-trip costs and idiosyncratic benefits. The implications of the model suggest that the welfare gains from removing shopping-trip costs in percentage terms can be measured for each census block as the weighted average of log distance measures between shopping stores and census blocks. The corresponding results show that the welfare gains from removing shopping-trip costs is about 4% for the average census block, with a range between 0.021% and 18% across census blocks that is further connected to their demographic or socioeconomic characteristics. Certain practical policy implications follow regarding how shopping-trip costs can be reduced to achieve higher welfare gains.

 

 

Non-technical Summary

Despite the increasing trend in online shopping, 88.5% of sales in the U.S. is still through traditional shopping, which may be due several reasons including its convenience or urgency of shopping. Since traditional shopping requires leaving home and walking/riding to a shopping store, it results in not only travel costs but also time and opportunity costs. Nevertheless, the literature lacks a quantitative investigation regarding the corresponding welfare costs of traditional shopping.

This paper attempts to measure the welfare costs of traditional shopping trips at the U.S. census block group level. The empirical investigation is based on an economic model, where individuals living in census blocks decide on which store to shop from based on the corresponding shopping costs (increasing with distance to the store) and idiosyncratic benefits. The implications of the model suggest that the welfare gains from removing bilateral shopping costs (that we consider as the welfare costs of traditional shopping in this paper) in percentage terms can be measured for each census block as the weighted average of log current bilateral distance measures between shopping stores and census blocks, where weights are the bilateral probabilities of individuals (living in certain census blocks) shopping at certain stores.

The model is empirically tested by using SafeGraph cellphone location data that provide information on the total number of visitors at the store level, where the census block group of visitors regarding their residence (home) is also given. The estimation results based on about 75 million observations show that the bilateral probabilities of individuals (living in certain census blocks) shopping at certain stores decrease with the corresponding distance measures. Quantitatively, the elasticity of shopping probability from a store with respect to distance is estimated around 0.0767.

The estimated distance effects on the bilateral probabilities of individuals (living in certain census blocks) shopping at certain stores are removed in a counterfactual investigation to measure the welfare costs of traditional shopping. This is achieved for each census block in the data set. The corresponding results show that the welfare gains from removing bilateral shopping costs is about 4% for the average (or median) census block, with a range between 0.021% and 18% across census blocks. The heterogeneity of welfare gains across census blocks is further investigated in a secondary analysis, where it is shown that the welfare costs of traditional shopping increase with cars per capita as census blocks with higher per capita number of cars currently make shopping trips to more distant stores.

Regarding heterogeneity of welfare gains across demographic or socioeconomic groups, it is depicted that census block groups with a higher share of Asian people would benefit the least from removing shopping costs, whereas those with a higher share of American Indian and Alaska Native people would benefit the most from it. When the relationship between welfare costs of traditional shopping and family income is investigated, it is shown that there is evidence for a hump-shaped relationship between family income and welfare costs. Finally, it is depicted that census block groups with a higher share of an educational attainment of an elementary school diploma would benefit the least from removing shopping costs, and those with a higher share of an educational attainment of a high school diploma would benefit the most from it. Based on the implications of the model used, these results suggesting that certain demographic or socioeconomic groups would benefit less from removing costs of traditional shopping can be explained by such groups currently making shopping trips to relatively close-by stores so that they gain relatively less when shopping costs are removed.

This paper contributes to the literature by measuring bilateral shopping costs between individuals (residing in census blocks) and stores by using the corresponding distance between them, where price faced at the store is also considered; the remaining factors such as quality of service or convenience are captured by idiosyncratic benefits at the store level for each individual. Measuring the corresponding welfare gains from removing shopping costs at the census block level is the key innovation in this paper, where connecting the heterogeneity of welfare gains across census blocks to certain demographic and socioeconomic characteristics is a further contribution.

 

The corresponding academic paper by Hakan Yilmazkuday is available as a working paper here.

 

 

Unequal Welfare Costs of Staying at Home across Socioeconomic and Demographic Groups

 Follow @yilmazkuday

 

Unequal Welfare Costs of Staying at Home across Socioeconomic and Demographic Groups

One sentence summary: There is evidence for unequal welfare costs of staying at home across socioeconomic and demographic groups.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Journal of Urban Sciences.

The working paper version is available here.

 
Abstract

Using daily census block group level data from the U.S., this paper investigates the welfare costs of staying at home due to COVID-19 across socioeconomic and demographic groups. The investigation is based on an economic model of which implications suggest that the welfare costs of staying at home increase with the stay-at-home probabilities of individuals. The empirical results provide evidence for significant heterogeneity across census block groups regarding the welfare effects of staying at home. This heterogeneity is further used to obtain measures of welfare changes for different socioeconomic and demographic groups at the national level.

 

Non-technical Summary

Staying at home is considered as one of the most effective ways to fight against COVID-19. Accordingly, several layers of government around the world have implemented lockdowns to mitigate the spread of COVID-19. Individuals have also reduced their mobility to protect themselves from COVID-19 in a voluntary way. Despite its success in reducing the spread of COVID-19, staying at home has resulted in many individuals having economic and psychological problems. Moreover, as individuals belonging to different socioeconomic and demographic groups have access to different employment, consumption or health-related opportunities, they have stayed at home in different amounts of time during COVID-19, suggesting that they might have been affected differently from COVID-19.

This paper attempts to measure the welfare implications of staying at home across alternative socioeconomic and demographic groups. The investigation is achieved by using the implications of an economic model, where both direct and indirect welfare effects of COVID-19 are considered. Specifically, the direct welfare effects are captured by the standard economic measures, namely the amount of consumption versus the amount of labor supplied, whereas the indirect welfare effects are captured by idiosyncratic benefits of being in another location (outside of home) versus the corresponding costs of mobility. In this context, idiosyncratic benefits of being in another location capture the welfare effects of having social interactions, whereas their absence captures the effects of mental distress, anxiety, worry, disinterest, depression, increased risks of suicide, domestic violence, obesity or poor general health perception. The corresponding costs of mobility capture not only the standard measures of traffic or the opportunity cost of time but also the effects of COVID-19 (e.g., the probability of getting sick) that is essential to measure the welfare effects of COVID-19.

In equilibrium, the model implies that the overall welfare effects of COVID-19 (discussed so far) can be captured by the changes in stay-at-home probabilities of individuals. This implication is used to measure the daily welfare changes in the U.S. at the census block group level. The measurement of mobility is achieved by using SafeGraph cellphone location data for each census block group (220,115 of them) for the daily period between January 1st and December 31st, 2020. The period between January 1st and February 29th, 2020 is considered as the pre-COVID-19 period, whereas the period between March 1st and December 31st, 2020 is considered as the COVID-19 period. The empirical results show that the median census block group has experienced a welfare loss of about 7.1% during the COVID-19 period. The corresponding nationwide welfare costs of COVID-19 (calculated as the weighted average across census block groups) is as much as 6.4%, with a daily average of about 2.1% during the sample period.

 

The empirical results also provide evidence for significant heterogeneity across census block groups regarding the welfare effects of staying at home due to COVID-19. This heterogeneity is further used to obtain measures of welfare changes for alternative socioeconomic and demographic groups at the national level, where the American Community Survey data on socioeconomic and demographic characteristics (at the census block group level) are used to aggregate across census block groups. The corresponding results based on race/ethnicity show that the average (across days) welfare costs have been experienced by the Asian population, followed by the Hispanic population, the white population, the black population and the native population. The results based on education level suggest that the average (across days) welfare costs have been experienced by the master's degree holders, followed by bachelor's degree holders, doctorate degree holders, elementary school graduates, high school graduates, and middle school graduates. Finally the results also that the average (across days) welfare costs have increased by the income level of individuals.

 

The results can be explained by individuals belonging to different socioeconomic and demographic groups having access to different employment opportunities. Specifically, the heterogeneity in welfare changes due to COVID-19 based on race/ethnicity can be explained by the Hispanic and black populations not being able to work from home compared to the white or Asian populations. This is reflected in welfare calculations of this paper as the white or Asian populations staying at home more compared to the Hispanic and black populations and thus experiencing higher welfare costs of COVID-19. Similarly, the heterogeneity in welfare changes due to COVID-19 based on education or income levels can also be explained by higher-educated or higher-income individuals being able to work from home. This is reflected in welfare calculations of this paper as higher-educated or higher-income individuals staying at home more compared to lower-educated or lower-income individuals and thus having experiencing welfare costs of COVID-19.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Journal of Urban Sciences.

The working paper version is available here.
 

 

 

Nonlinear Effects of Mobility on COVID-19 in the U.S.: Targeted Lockdowns Based on Income and Poverty

 Follow @yilmazkuday

 

Nonlinear Effects of Mobility on COVID-19 in the U.S.: Targeted Lockdowns Based on Income and Poverty

One sentence summary: The positive effects of mobility on COVID-19 increase with certain demographic or socioeconomic characteristics.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Economic Studies.

The working paper version is available here.

 
Abstract

This paper investigates nonlinearities in the relationship between mobility and COVID-19 cases or deaths. The formal analysis is achieved by using county-level daily data from the U.S., where a difference-in-difference design is employed. Nonlinearities in the relationship between mobility and COVID-19 cases or deaths are investigated by regressing weekly percentage changes in COVID-19 cases or deaths on mobility measures, where county fixed effects and daily fixed effects are controlled for. The main innovation is achieved by distinguishing between the coefficients in front of mobility measures across U.S. counties based on their demographic or socioeconomic characteristics. The results suggest that the positive effects of mobility on COVID-19 cases or deaths increase with population, per capita income, or commuting time as well as with having certain occupations, working in certain industries, attending certain schools, or having certain educational attainments. Important policy implications follow regarding where mobility restrictions would work better to fight against COVID-19 through targeted lockdowns.

 

 

Non-technical Summary

The relationship between the spread of COVID-19 and social interactions through mobility is well established. Accordingly, several governments have employed lockdowns to slow down the spread of COVID-19. However, this relationship by itself does not suggest anything related to targeted lockdowns that can be useful when policy makers face trade-offs between health-related concerns and economic slowdown as certain group of people or certain communities can be more vulnerable to the spread of COVID-19.

This paper investigates how the relationship between mobility and the COVID-19 spread changes with demographic or socioeconomic characteristics. The formal investigation is achieved by using daily county-level data from the U.S., where a difference-in-difference approach is employed. The nonlinear relationship between mobility and COVID-19 cases or deaths is investigated by regressing weekly percentage changes in COVID-19 cases or deaths on mobility measures, where county fixed effects and daily fixed effects are controlled for; accordingly, county-specific factors that are constant over time and day-specific factors that are common across U.S. counties are already controlled for. The main innovation is achieved by distinguishing between the coefficients in front of mobility measures across U.S. counties based on their demographic or socioeconomic characteristics that we utilize as threshold variables.
 

Several demographic or socioeconomic characteristics of U.S. counties are considered for investigating the nonlinear relationship between mobility and the COVID-19 spread. These include 45 different variables based on the categories of population characteristics, economic variables, occupations, employment in industries, school attendance, educational attainment, and race. The motivation behind including these potential threshold variables comes from the existing literature, where several studies have shown how the spread of COVID-19 is related to these demographic or socioeconomic characteristics. 

 

The results of the nonlinear investigation suggest that the positive effects of mobility on COVID-19 cases or deaths increase with population, per capita income, or commuting time as well as with having certain occupations, working in certain industries, attending certain schools, or having certain educational attainments. Since mobility restrictions to fight against COVID-19 would work better in counties where the positive effects of mobility on COVID-19 cases or deaths are bigger, it is implied that policy makers can consider targeted lockdowns based on the threshold variables identified in this paper.

 

 

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Economic Studies.

The working paper version is available here.

 

Welfare Costs of COVID-19: Evidence from U.S. Counties

 Follow @yilmazkuday

 

Welfare Costs of COVID-19: Evidence from U.S. Counties

One sentence summary: The average (across days) welfare reduction during COVID-19 is about 11% for the average U.S. county and up to about 46% across counties.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Regional Science.

The working paper version is available here.

 
Abstract

Using daily U.S. county-level data on consumption, employment, mobility and the coronavirus disease 2019 (COVID-19) cases, this paper investigates the welfare costs of COVID-19. The investigation is achieved by using implications of a model, where there is a trade-off between consumption and COVID-19 cases that are both determined by the optimal mobility decision of individuals. The empirical results show evidence for about 11% of an average (across days) reduction of welfare during the sample period between February and December, 2020 for the average county. There is also evidence for heterogeneous welfare costs across U.S. counties and days, where certain counties have experienced welfare reductions up to 46% on average across days and up to 97% in late March, 2020 that are further connected to the socioeconomic characteristics of the U.S. counties.

 

 

  

Non-technical Summary

The coronavirus disease 2019 (COVID-19) has resulted in not only numerous casualties but also unprecedented reductions in economic activity. Since both COVID-19 cases and economic activity are positively related to mobility, individuals have faced trade-offs regarding the optimal amount of mobility that they should have. It is implied that investigating the welfare changes due to COVID-19 requires taking into account the mobility of individuals.

Based on this background, this paper investigates the welfare costs of COVID-19 by considering the interaction between COVID-19 cases, economic activity and mobility of individuals. A multi-region model is introduced to motivate the empirical investigation, where individuals optimally decide on their mobility that further determines their current consumption and future COVID-19 cases. The parameters and unknown variables of the model are estimated by using daily U.S. county-level data on consumption, employment, mobility and COVID-19 cases.

The estimation results confirm that economic activity (measured by either consumption or employment) increases with mobility of individuals. The estimation results also confirm the positive relationship between mobility and COVID-19 cases. These results are robust to the consideration of county-specific factors that are constant over time and time-varying nationwide factors that are common across counties.
 

The implications of the model are further used to investigate welfare costs of COVID-19 and its components based on economic activity and COVID-19 cases. The corresponding model implications suggest evidence for about 11% of an average (across days) reduction of welfare during the sample period between February and December, 2020 for the average U.S. county. 

 

When welfare costs are decomposed into those due to each model component, it is shown that COVID-19 cases contribute the most to welfare reductions in early months of COVID-19, whereas they have similar contributions with consumption/employment starting from about May 2020. Mobility contributes negatively to welfare in a steady way during the sample period, whereas other factors have been more effective in early months of COVID-19. In terms of the contribution of each welfare component as an average across days, increases in COVID-19 cases reduce welfare by about 6.7% for the average county (up to 14.2% across counties), whereas consumption reductions contribute to welfare costs by about 3.7% for the average county (up to 43.2% across counties). The contribution of mobility (with respect to other factors) is much more on average (across days) during the sample period.

 

 

 

The empirical results of this paper also provide evidence for heterogeneous welfare costs across U.S. counties and days, where certain counties have experienced welfare reductions up to 46% on average across days and up to 97% in late March, 2020. These results are robust to the consideration of alternative data sets as well as alternative parameter values considered in the model.

 

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Journal of Regional Science.

The working paper version is available here.

  

COVID-19 and Housing Prices: Evidence from U.S. County-Level Data

 Follow @yilmazkuday

 

COVID-19 and Housing Prices: Evidence from U.S. County-Level Data

One sentence summary: The effects of COVID-19 cases on housing prices are negative and significant after controlling for other factors.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Review of Regional Research.

The working paper version is available here.

 
Abstract

This paper investigates the effects of coronavirus disease 2019 (COVID-19) on housing prices at the U.S. county level. The effects of COVID-19 cases on housing prices are formally investigated by using a two-way fixed effects panel regression, where county-specific factors, time-specific factors, and mobility measures of individuals are controlled for. The benchmark results show evidence for negative and significant effects of COVID-19 cases on housing prices, robust to the consideration of several permutation tests, where the negative effects are more evident in counties with higher poverty rates. Exclusion tests further suggest that U.S. counties in the state of California or the month of May 2020 are more responsible for the empirical results, although the results based on other counties and months are still in line with the benchmark results.

  

Non-technical Summary

The coronavirus disease 2019 (COVID-19) has resulted in not only a health crisis through its direct effects but also an economic one through its indirect effects. These indirect effects are reflected in housing prices within the U.S. in an unequal way across counties, where housing prices have increased by about $1,408 on average across counties on a monthly basis (between February 2020 and August, 2021), with a range between $1,979 of a reduction and $14,963 of an increase. Within this context, what is the contribution of COVID-19 cases on this heterogeneity representing unequal changes in housing prices across U.S. counties? The answer to this equation depends on several channels that affect the housing market at the local (U.S. county) level.

This paper investigates this heterogeneity representing unequal changes in housing prices across U.S. counties due to COVID-19 cases. The formal investigation is achieved by using a two-way fixed effects panel regression, where county-specific and time-specific factors are controlled for. Several mobility measures of individuals are also considered as control variables as they not only represent the overall economic activity at the U.S. county level over time but also the developments in the housing sector at the U.S. county level over time due to staying at home as a housing-demand shifter.

The benchmark empirical results show evidence for negative and significant effects of COVID-19 cases on housing prices, and they have been confirmed by several robustness checks based on permutation tests, exclusion tests, or interactions with other variables. 

 

When the channels of causality are further investigated, poverty is shown to be an important factor. Specifically, the U.S. counties with higher rates of poverty have experienced more reductions in housing prices due to COVID-19, whereas those with lower rates of poverty have experience almost no changes (or sometimes increases) in housing prices. Therefore, there is evidence for unequal effects of COVID-19 on housing prices across U.S. counties due to poverty differences.

 

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Review of Regional Research.

The working paper version is available here.

 

 

 

Drivers of Global Trade: A Product-Level Investigation

Follow @yilmazkuday

 

Drivers of Global Trade: A Product-Level Investigation

One sentence summary: Supply-side factors, capturing production and exporting costs in source countries, are responsible for about 85% of changes in global trade between 1995 and 2018.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Economic Journal.

Working paper version is available here.

Abstract
This paper investigates the drivers of global trade at the six-digit product level by using the implications of a model that are consistent with a large class of trade models. The drivers of global trade at the product level are identified first by estimating the product-level bilateral trade implications of the model and second by aggregating the fitted estimation results across bilateral countries using Taylor series. The empirical results suggest that supply-side effects (capturing production or exporting costs in source countries) contribute to changes in global trade more than six times the demand-side effects (capturing economic activity or preferences in destination countries) and more than ten times the effects of bilateral trade costs (capturing bilateral protectionism measures). Several product-level implications follow.

  

Non-technical Summary

Global merchandise trade has increased by more than 6 trillion U.S. dollars between 1995-2018. This increase is mostly accounted for by products such as machinery/electrical (with a contribution of about 26%), mineral products (with a contribution of about 18%), chemicals and allied industries (with a contribution of about 11%), and transportation (with a contribution of about 11%). Across broad economic categories, trade of intermediate inputs account for about 66% of this increase, whereas trade of capital goods and intermediate inputs account for about 18% and 16%, respectively. Although these statistics provide useful information on products or categories that drive the global trade, policy making requires knowledge on the economic forces that are responsible for the contribution of these products or categories.

This paper investigates the economic drivers of global trade by using six-digit product level data covering the years 1995-2018. These economic drivers are identified by using the implications of a large class of trade models, where bilateral trade between any two countries depend on source prices, bilateral trade iceberg costs, and a measure of economic activity at the destination country. Based on this motivation, a simple trade model is introduced of which implications are used for decomposing the changes in global trade into those due to supply-side factors (capturing source prices and thus production or exporting costs in source countries), demand-side factors (capturing economic activity or preferences in destination countries), and bilateral trade costs (capturing bilateral protectionism measures).

The knowledge of the decomposition of changes in global trade is important especially countries focusing on export-led growth, because if supply-side factors are effective in explaining changes in global trade, source countries may want to invest more into their production technologies, infrastructure, financial depth, operational costs of exporting, costs related to entering foreign markets, or modifying their products for individual foreign markets. In contrast, if demand-side factors are effective, source countries may want to invest in removing information barriers (e.g., through advertising their products) to affect preferences of destination countries. Finally, if bilateral trade costs are effective, source countries may want to get involved in negotiations to reduce trade barriers (e.g., through free trade agreements).

Regarding the methodology, the decomposition of changes in global trade is achieved first by estimating the product-level bilateral trade implications of a trade model and second by aggregating the fitted estimation results across bilateral countries using Taylor series to obtain global product-level measures. This methodology results in identifying the contribution of supply-side factors, demand-side factors and bilateral trade costs to changes in product-level global trade between 1995 and 2018. The corresponding results suggest that supply-side effects have contributed to changes in global trade by about 85%, followed by demand-side effects with a contribution of about 13% and by bilateral trade costs with a contribution of about 8%. The corresponding contribution of residuals by only about -6% capturing unexplained part of the data by the model implications or approximation due to using Taylor series further supports the investigation.

 

Across products, supply-side effects explain cumulative changes in product-level global trade between 47% (for Textiles) and 97% (for Chemicals & Allied Industries). In comparison, demand-side effects explain cumulative changes in product-level  global trade between 3% (for Stone/Glass) and 45% (for Animal & Animal Products). Finally, bilateral trade costs contribute to product-level global trade between 3% (for Chemicals & Allied Industries or Wood & Wood Products) and 24% (for Textiles). Across broad economic categories, supply-side factors contribute to global trade between 62% (for consumption goods) and 89% (for intermediate goods), demand-side factors contribute to global trade between 8% (for intermediate goods) and 29% (for consumption goods), and bilateral trade costs contribute to global trade between 7% (for intermediate goods) and 13% (for consumption goods).

 

Since supply-side factors are shown to be the main drivers of global trade, it is implied that rather than purely focusing on reducing bilateral trade costs through trade negotiations, one additional way for source countries to increase their exports is to reduce their production costs, say, by investing more into technology, infrastructure, or financial depth, while another way is to reduce their export-related costs such as operational costs of exporting, costs related to entering foreign markets or modifying their products for individual foreign markets. 

 

 

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Economic Journal.

 

Working paper version is available here.

Inflation Convergence over Time: Sector-Level Evidence within Europe

 Follow @yilmazkuday

 

Inflation Convergence over Time: Sector-Level Evidence within Europe

One sentence summary: Average half-life of inflation differentials across European countries has decreased from about 15 months to about 8 months within the last two decades.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Finance.

Working paper version is available here.

 
Abstract

This paper investigates inflation convergence among European countries by using sector-level data for the period between 1997:M1 and 2019:M12. Panel unit root tests at the country-sector level are conducted by using moving windows, which is useful to analyze changes in inflation convergence and the corresponding speed of convergence over time. The results suggest evidence for inflation convergence for the majority of sectors within Europe, although disruptions have been experienced by certain countries, especially during the 2008 financial crisis. Regarding the speed of inflation convergence, the average half-life across European countries has decreased from about 15 months to about 8 months during the sample period. Important sector-level implications follow for European Union (EU) candidate countries and non-euro EU member countries regarding the Maastricht Treaty.

Non-technical Summary

Inflation convergence is one of the important criteria in the Maastricht Treaty to ensure price stability and integration within the European Union (EU). This criterion not only requires member countries to have a high degree of price stability but also calls for a price performance that is sustainable for the adoption and continuous circulation of euro. Accordingly, when candidate countries are considered for EU membership or the Euro Area (EA), part of the evaluation is achieved through inflation convergence. Moreover, even when a country is an EU member or within EA, its performance of price stability is evaluated over time for sustainability. It is implied that an investigation of inflation convergence within Europe over time is essential for the price stability and continuous integration of EU.

This paper achieves such a time-varying investigation for inflation convergence among European countries. The formal analysis is conducted by using four-digit sector-level inflation data from 34 countries covering the monthly period between 1997:M1-2019:M12, where five year (i.e., sixty months) moving windows are considered to have a time-varying investigation. Panel unit root tests are used to investigate the convergence of inflation rates at the country-sector level. In particular, country-sector specific panel estimations are achieved by comparing sector-level inflation rates of each country with those of other countries within Europe; i.e., the cross-sectional dimension of the panel unit root tests consist of countries at the sector level.

Having a sector-level investigation is essential to avoid any aggregation bias. This type of an investigation is also useful to obtain sector-specific policy implications, especially for EU candidate countries and non-euro EU member countries, as such an investigation can reveal the sectors that are responsible for non-convergence (if any). Moreover, different from country-level analyses where evidence for only convergence versus non-convergence can be obtained, having a sector-level investigation results in obtaining information on the total expenditure share of sectors for which there is evidence for inflation convergence.

 

When there is evidence for convergence (if any) for a particular sector in a particular country, the corresponding speed of convergence is further investigated; this is convenient to observe how the speed of convergence has changed over time at the country-sector level. The corresponding results show that inflation convergence is achieved for all sectors in several countries for most of the sample period, although the total expenditure share of sectors experiencing convergence is as low as about 75% across countries.

 

Once estimations are achieved at the country-sector level, the corresponding results are further aggregated across sectors (of each country) to have country-specific results for inflation convergence. These country-specific results suggest that there is evidence for stability over time for most countries except for certain time periods that mostly coincide with the 2008 financial crisis. In particular, countries such as Bulgaria, Estonia, France, Ireland, Iceland, Lithuania, Latvia and United Kingdom have experienced disruptions in their sector-level inflation convergence processes during the 2008 financial crisis, whereas countries such as Switzerland, Hungary, Italy, Poland, Slovakia and especially Turkey have experienced disruptions in their sector-level inflation convergence processes starting from around 2015. Regarding the speed of convergence, the average half life across countries has decreased from about 15 months between 1997:M1-2001:M12 to about 8 months between 2015:M1-2019:M12.

 

Sector-level half-life estimates for the median country suggest that before the official circulation of the euro (i.e., between 1997:M1 and 2001:M12), half-life estimates have an average (across sectors) of about 17 months, with a range between 10 months (for "food and non-alcoholic beverages") and 26 months (for "restaurants and accommodation services"). By the latest period of 2015:M1-2019:M12, the average half-life estimate (across sectors) have reduced to about 11 months, with a range between 5 months (for "clothing and footwear") and 35 months (for "restaurants and accommodation services"). 

 

 

It is implied that especially "restaurants and accommodation services" is responsible for not having any further reductions in half-life estimates over time; as indicated in earlier studies, this can be fixed by having more labor mobility across countries, product diversification and trade openness.

 

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at International Finance.

Working paper version is available here.

 

Fighting Against COVID-19 Requires Wearing a Face Mask by Not Some but All

 Follow @yilmazkuday

 

Fighting Against COVID-19 Requires Wearing a Face Mask by Not Some but All

One sentence summary: Causal effects of social interaction on COVID-19 are statistically eliminated when more than 85% of people "always" wear a face mask.

The corresponding academic paper by Hakan Yilmazkuday has been accepted for publication at Transportation Research Interdisciplinary Perspectives.

is available as a working paper here.

 
Abstract

This paper investigates the effects of wearing a face mask on fighting against coronavirus disease 2019 (COVID-19). The formal analysis is achieved by using a difference-in-difference design, where U.S. county-level data on changes in COVID-19 cases or deaths are regressed on lagged changes in social interaction of people measured by Google mobility. The main contribution is achieved by distinguishing between the effects of social interaction on COVID-19 in mask-wearing versus non-mask-wearing counties determined by Dynata surveys. After controlling for county-specific and time-specific factors, the results show that social interaction causally increases both COVID-19 cases and deaths across U.S. counties. Wearing a face mask starts working to fight against COVID-19 only if more than 75% of people in a county "always" wear a face mask, while the effects of social interaction on COVID-19 are statistically eliminated when more than 85% of people in a county "always" wear a face mask.

  

Non-technical Summary

Social interaction between people is accepted as one of the key determinants for the spread of viruses leading to infections, including coronavirus disease 2019 (COVID-19). Despite its leading effects on COVID-19, social interaction is still necessary to prevent the corresponding societal and economic costs. Accordingly, wearing a face mask in public has been suggested by several studies to be able to continue having social interactions during the COVID-19 era as mask wearing reduces the transmissibility per contact by reducing transmission of infected droplets in both laboratory and clinical contexts.

 

Despite the consistency in the recommendation that especially symptomatic individuals should use face masks, discrepancies have been observed in the general public and community settings regarding face-mask wearing. Accordingly, this paper investigates the effects of wearing a face mask on the causal relationship between social interaction and COVID-19 cases or deaths. This initially requires confirming the causal relationship between social interaction and COVID-19 cases or deaths. This confirmation is achieved by using daily data from U.S. counties on COVID-19 cases or deaths as well as social interaction measures based on Google mobility for the period between February 15th, 2020 and August 30th, 2020. The formal analysis is achieved by using a difference-in-difference design, where U.S. county-level data on changes in COVID-19 cases or deaths are regressed on lagged changes in social interaction of people after controlling for county-specific and time-specific factors. The results of this initial investigation confirm that higher social interaction leads to higher COVID-19 cases and deaths across U.S. counties.

 

After confirming the causal relationship between social interaction and COVID-19 cases (or deaths), we continue with a secondary investigation regarding the effects of wearing a face mask on this relationship. In order to do so, we categorize U.S. counties as mask-wearing counties versus non-mask-wearing counties by using Mask-Wearing Survey Data collected by Dynata at the request of New York Times from 250,000 survey respondents. This categorization of U.S. counties results in splitting the effects of social interaction on COVID-19 cases/deaths into those in mask-wearing counties versus non-mask-wearing counties.

 

 

The results of this secondary investigation reveal that wearing a face mask starts working to fight against COVID-19 cases or deaths only if more than 75% of people in a county "always" wear a face mask, while the effects of social interaction on COVID-19 are statistically eliminated when more than 85% of people in a county "always" wear a face mask. Therefore, it is possible to continue having social interactions without any statistically significant effects on COVID-19 cases if a community-wide wearing of face masks can be achieved. This result is important as it provides insights about how societal and economic costs due to COVID-19 can be prevented by wearing a face mask by not some but all.

The corresponding academic paper by Hakan Yilmazkuday is available as a working paper here.