The drastic economic downturn accompanying the COVID-19 pandemic was met by an unprecedented fiscal response of the German government. On 3 June 2020, it announced a stimulus package worth €130 billion. To the general surprise of the public, this package included a six-month reduction of the standard value-added tax (VAT) rate from 19% to 16%, and a decrease of the reduced rate from 7% to 5%, at an estimated cost of €20 billion or 0.6% of GDP. As outlined by D’Acunto et al. (2020), this type of unconventional fiscal policy aims to temporarily reduce prices and stimulate consumption through inflation expectations. For this to work, however, firms need to reduce prices and pass on the tax reduction to consumers. D’Acunto et al. (2020) draw on the experience of a previous increase in the VAT rate from 16% to 19%, announced in 2005 and executed in 2007 which at the time raised inflation expectations and durable expenditure. They conclude from that episode that that the temporary VAT reduction implemented as part of the stimulus package in 2020 will successfully increase inflation expectations and expenditure. Benzarti et al. (2020) show, however, that pass-through is often asymmetric and that prices respond twice as much to VAT increases as to decreases. Their findings caution against drawing any conclusions from the 2007 tax increase for today’s tax decrease as an increase does not necessarily translate to a tax decrease. This is particularly true given the temporary nature of the stimulus measure.
Evidence on the pass-through of a temporary VAT reduction for diesel and gasoline
In a new study, we provide the first estimates of the pass-through rate of a temporary VAT reduction in a given sector. We estimate the pass-through rates for diesel and gasoline sold at fuel stations, a setting where price adjustments are costless and in fact happen frequently. This sector is worth 3% of GDP in Germany (Montag et al. 2020). We use a unique dataset that contains the universe of price changes at fuel stations in Germany and France for June and July 2020 and employ a difference-in-differences strategy comparing fuel stations in Germany and France before and after 1 July 2020. The German stations are treated starting 1 July whereas fuel stations in France are unaffected by the German policy change. We account for differences between stations and over time using station and date fixed effects. We also account for regional differences in demand over time by controlling for changes in mobility, using regional data from the Google COVID-19 Community Mobility Report.
Our initial results show that pass-through is fast and substantial, but it remains incomplete for all fuel types. Moreover, we find substantial differences in the extent to which the VAT rate reduction is passed on to consumers for different fuel types. We argue that these differences in pass-through rates are consistent with a different competitive pressure for different fuel types due to different characteristics of their respective consumers.
Pass-through differs greatly between fuel types sold at the same station
Fuel stations in Germany and France mainly sell diesel, E5 and E10. In Germany, diesel accounts for 44% of volume sold in the fuel market for passenger vehicles with combustion engines while gasoline accounts for 56%.1 E5 and E10 are two types of gasoline that differ with respect to their ethanol content (E5 contains 5% of ethanol, E10 10%). Whilst the former accounts for 82% of the gasoline market, the latter only has a share of 14%.2 In Germany, all fuel stations are required to sell all three types of fuel. We estimate pass-through rates separately for each of the three types of fuel.
We begin our analysis by estimating the pass-through rate for E5. We find that fuel stations pass on 40% of the VAT rate reduction to consumers. Figure 1 shows the evolution of the average price of E5 in Germany between 15 June and 31 July 2020.3 The dashed blue lines show the counterfactual evolution of the average price under full pass-through and zero pass-through.
Figure 1 Effect of the temporary VAT reduction on average prices, E5
Notes: The solid line shows the evolution of the daily weighted average price of E5 in Germany between 15 June and 31 July 2020. The dashed lines show the counterfactual evolution of the daily weighted average gasoline price in Germany under zero and full pass-through. The solid vertical line shows the beginning of the VAT reduction.
For E10, we find that fuel stations pass on 61% of the VAT rate reduction. Figure 2 describes the evolution of the average price of E10 in Germany, as well as counterfactual prices under full and zero pass-through.
Figure 2 Effect of the temporary VAT reduction on average prices, E10
Notes: The solid line shows the evolution of the daily weighted average price of E10 in Germany between 15 June and 31 July 2020. The dashed lines show the counterfactual evolution of the daily weighted average gasoline price in Germany under zero and full pass-through. The solid vertical line shows the beginning of the VAT reduction.
Finally, we estimate the pass-through rate for diesel. We find that fuel stations pass on 83% of the VAT reduction to consumers. Figure 3 plots the average price of diesel in Germany, as well as the prices if the VAT rate reduction was fully or not at all passed on to consumers.
Figure 3 Effect of the temporary VAT reduction on average prices, diesel
Notes: The solid line shows the evolution of the daily weighted average price of diesel in Germany between 15 June and 31 July 2020. The dashed lines show the counterfactual evolution of the daily weighted average gasoline price in Germany under zero and full pass-through. The solid vertical line shows the beginning of the VAT reduction.
Differences in pass-through rates are consistent with different propensities to shop for lower prices
The difference in the pass-through rates between fuel types is striking given that they are sold by the same stations. Thus, station characteristics, such as amenities or brand reputation, distance to competitors or supply structure are unlikely to be responsible for these differences.
What differs, though, are the customer groups for the different types of fuel. Evidence on driving patterns shows that, on average, diesel drivers drive twice as many kilometres per year as gasoline drivers. Thus, it seems likely that they are more willing to shop for lower prices. The fact that they experience the highest pass-through is consistent with this hypothesis.
Gasoline comes in two types, as indicated above. E5 is the standard type of gasoline (also referred to as 95 or Super). In 2011, E10 was introduced, with a higher share of ethanol. The aim was to decrease the amount of fossil fuel used in transportation and reduce greenhouse gas emissions. Although E5 and E10 are not taxed differently, E10 is usually around 4 cents cheaper than E5. This is partly due to the relative prices of crude oil and biofuels on the world market and also reflects the fact that stations are fined if they do not fulfil a minimum quota of biofuels.
Most drivers that buy E5 could also buy E10 but choose not to do so although E10 is cheaper. After the introduction of E10, a public discussion erupted about whether biofuels damage the engine. While this can happen for some cars, it is not the case for around 90% of gasoline-run vehicles, including all produced after 2012. The fact that most E5 customers could choose to buy the cheaper E10 gasoline but do not do so indicates that they are less price responsive than E10 customers. This suggests lower competitive pressure for E5 fuel which is consistent with the observation that pass-through is lower for E5.
Overall, our evidence indicates that fuel stations pass on more of the temporary VAT reduction to consumer groups who are on average more likely to be responsive to price differentials.
Anecdotal evidence from other markets corroborates these results. There are numerous reports of full pass-through of the VAT reduction by discount supermarkets where competitive pressure is strong.4 Starbucks Coffee instead, whose customers are less known for shopping for competitive prices, made headlines when it announced not to change its prices.5
Implications for unconventional fiscal policy
Past studies of pass-through of VAT changes in different contexts found mixed results. The literature includes evidence of under-shifting (e.g. Benzarti and Carloni 2019), full but slow pass-through (e.g. Benedek et al. 2019), full and fast pass-through (e.g. Büttner and Madzharova 2020) and even over-shifting (e.g. Besley and Rosen 1999). Relatively little is known about the determinants of different pass-through rates, however. Our unique dataset allows us to explore some of these determinants. In this first study, we focus on differences in fuel types and the characteristics of their customers.
While fuel markets are not the prime target of unconventional fiscal policy, this analysis allows us to learn about the mechanisms behind tax pass-through. Analysing pass-through for different fuel types, we find strong heterogeneities in the pass-through rate, even though price adjustment is costless for all types. In more competitive markets, pass-through is substantial and fast, whereas it is low in markets where consumers are less likely to shop for lower price. Hence, costless price adjustments are not a sufficient condition for high pass-through rates. This has implications for successful unconventional fiscal policy: while high pass-through is a necessary condition for unconventional fiscal policy to work, targeting this policy to competitive markets could significantly increase its cost-effectiveness.
Bibliography
Benedek, D, R D Mooij, M Keen, and P Wingender (2015), “Varieties of VAT pass through”, International Tax and Public Finance 27 (4): 890-930.
Benzarti, Y and D Carloni (2019), “Who really benefits from consumption tax cuts? Evidence from a large VAT reform in France”, American Economic Journal: Economic Policy 11 (1): 38-63.
Benzarti, Y, D Carloni, J Harju and T Kosonen (2020), “What Goes Up May Not Come Down: Asymmetric Incidence of Value Added Taxes”, Journal of Political Economy.
Besley, T. J. and Rosen, H. S. (1999), “Sales taxes and prices: An empirical analysis”, National Tax Journal, 52 (2): 157–78.
Büttner, T and B Madzharova (2020), “Unit sales and price effects of pre-announced consumption tax reforms: micro-level evidence from European VAT”, American Economic Journal: Economic Policy.
D’Acunto, F, D Hoang, and M Weber (2020), “Unconventional fiscal policy to exit the COVID-19 crisis”, VoxEU.org, 08 June.
Montag F, A Sagimuldina and M Schnitzer (2020), “Are temporary value-added tax reductions passed on to consumers? Evidence from Germany’s stimulus”, Working Paper, Ludwig- Maximilians University Munich, 19 August, 2020.
Endnotes
1 Our analysis focuses on the market for passenger vehicles only, as fuel stations do not report prices for truck diesel to the Market Transparency Unit.
2 The remaining volume share of 4.6% is gasoline with an octane rating of 98 or above (commonly referred to as Super Plus). Only some high-performance vehicles require Super Plus and it usually costs around 10% more than E5 or E10. Fuel stations do not report prices for Super Plus to the Market Transparency Unit.
3 The 15 June is a natural starting point, as it marks the beginning of the European Commission’s Re-open EU plan and is the date on which France and Germany lifted many of their travel restrictions and re-opened their borders.
4https://www.chip.de/news/MwSt-Senkung-bei-Aldi-Lidl-Netto-und-Co.-So-unt..., accessed 18 August 2020.
5https://www.faz.net/aktuell/wirtschaft/unternehmen/starbucks-gibt-gesenk..., accessed 18 August 2020.
