bachmann24jan.jpg
VoxEU Column Industrial organisation

Firms and collective reputation: The Volkswagen emissions scandal as a case study

There has been limited research on how groups’ collective reputations are affected by the misbehaviour of individual members. This column uses the Volkswagen emissions scandal as a natural experiment to explore group reputation externalities. German auto manufacturers that weren’t implicated in the scandal suffered significant declines in sales, stock returns, and public sentiment in the US. Volkswagen’s malfeasance appeared to materially harm the group reputation of German car engineering.

A group’s collective reputation can influence the outcomes for its members independently of their individual characteristics or behaviour. In a seminal paper, Tirole (1996) develops a theoretical framework for modelling collective reputation, showing how stereotypes can arise through history dependence so that an original sin by elder group members can have long-lasting effects on a group. Tadelis (1999), writing in the large theoretical literature on firm reputation, highlights the importance of names as intangible assets conveying reputation. In a recent paper, we combine these two perspectives and empirically investigate a case in which firms have a collective reputation by virtue of their association with a particular country (Bachmann et al. 2017).

Despite the theoretical interest in the issue, there is limited empirical evidence that group reputation and reputational externalities matter economically at the level of the firm. This evidence is limited by many factors. First is the scarcity of large and prominent shocks that directly implicate only a subset of group members, so that the existence of reputational spillovers on the other group members can be demonstrated. Second is the difficulty of identifying natural and reputationally salient groups of firms. Third is the rarity of direct measures of group reputation, which requires researchers to make indirect inferences about the effects of reputation. To overcome these issues, we use the 2015 Volkswagen (VW) emissions scandal, which directly implicated VW but not the other German automakers, as a natural experiment. We study the scandal’s effects on their vehicle sales, stock returns, and social media sentiment. In doing so, we provide empirical support for the theoretical literature and for the existence of group reputation externalities in an important setting.

The Volkswagen emissions scandal

On 18 September 2015, the US Environmental Protection Agency served a Notice of Violation to the VW Group, which consists of Volkswagen proper, plus Audi and Porsche. The notice alleged that approximately 500,000 VW and Audi diesel engine cars sold between 2009 and 2015 in the US contained a ‘defeat device’ that allowed these cars to comply with emissions regulations in the test box, while having higher on-road emissions. This date marks the public eruption of one of the largest industrial scandals in recent history, with a prolonged legal fallout in the US.

A natural experiment

Several features of the scandal make it an appealing natural experiment.

First, for the general public the scandal was a clear surprise, and it immediately generated extensive media coverage. We show that the mentions of “Volkswagen” in US newspapers, newswires, journals, and magazines more than tripled to 5,500 in September 2015, compared to the months prior to the scandal. There was also no concurrent notion in the public that other German car manufacturers – BMW, Mercedes-Benz, and Smart – were manipulating vehicle emissions. For instance, on 22 September 2015, CNN published an article stating: “But before you start worrying about the complete collapse of the German auto industry, it’s worth repeating that – at least for now – the scandal is limited to Volkswagen. Other German automakers such as Daimler, which owns Mercedes-Benz, and BMW have said they’re not affected” (Petroff 2015). Only in the summer of 2017, after our study period, was it suggested that Mercedes-Benz had also manipulated emissions (Zeit Online 2017), although Mercedes-Benz never admitted wrongdoing in the US. We show that Google searches skyrocketed for Volkswagen but not for the other German automakers, supporting the notion that the others were not directly implicated by the scandal in the public eye.

Second, German auto manufacturers featured the notion of ‘German car engineering’ prominently in their US advertising, creating a natural reputational group. Additionally, we show that pre-scandal trends in reputation and business outcomes appear to have been similar among automakers from all countries, suggesting that non-German automakers can serve as a control group for understanding the effects of a German-specific shock.

Third, individual automotive makes are salient to consumers, enabling us to use novel company-specific data on US Twitter sentiment and Google internet searches to directly establish the existence of reputational externalities.

Adding to its appeal as a natural experiment, the scandal occurred within an important setting. The car manufacturing industry is large and important in Germany. In 2014, the year prior to the scandal, cars amounted to 18% of Germany’s total exports according to the German Federal Statistical Office (Destatis 2015), making it Germany’s largest export category. Also as of 2014, Germany captured 22.7% of world car exports in dollar terms, by far the largest share of world car exports in UN trade statistics (United Nations 2017).

German vehicles also represent a large share of the US market – in 2014, German car manufacturers comprised 8.1% of all US light vehicle sales, making Germany the second-largest source for foreign-branded vehicles.

Additionally, the scandal’s reputational consequences were amplified by the damage the excess emissions caused to the public. Oldenkamp et al. (2016) estimate that the excess emissions caused by VW diesel cars cost 45,000 disability-adjusted life years, with a value of life lost of approximately $39 billion. Barrett et al. (2015) estimate 59 premature deaths and a social cost of $450 million; Holland et al. (2016) estimate similar numbers.

Finally, the scandal also sparked a widespread public discussion regarding the mechanism at the centre of our paper – country-related reputational spillovers. Following the scandal, media attention to ‘German engineering’ spiked, with 130 media articles mentioning the term in September 2015, a five-fold increase over the preceding months. A recurring theme in this coverage was the notion that the scandal might tarnish the broader reputation of German manufacturing firms. As part of this coverage of the scandal, Reuters published an article on 22 September 2015 titled “VW scandal threatens ‘Made in Germany’ image” (Chambers 2015). A day later, Reuters doubled down with an article entitled “Volkswagen could pose bigger threat to German economy than Greek crisis” (Nienaber 2015), which included the claim that “The broader concern for the German government is that other car makers such as Mercedes-Benz and BMW could suffer fallout from the Volkswagen disaster.”

Results

We find that, in our preferred specification, the VW scandal reduced the sales growth rate for non-VW German automakers by 10.4 percentage points in the twelve months following the scandal (as measured by unit sales data from Ward’s Automotive). This decline translated to a loss of 76,000 vehicle sales worth roughly $3.7 billion of revenue for non-VW German automakers, along with a decline in their stock returns relative to expected market outcomes – the cumulative abnormal returns for BMW and Daimler were 6.4 percentage points lower in the two days following the scandal. We arrive at these numbers using a difference-in-differences approach (e.g. Angrist and Krueger 1999) that compares how key outcomes changed over time for the treated group of non-VW German automakers versus the control group of non-German automakers. The differential responses to the scandal provide causal evidence on the scandal’s economic consequences.

Our interpretation of these results is that the scandal harmed the collective reputation of German automakers in the US. To support this interpretation, we proceed in several steps.  

  • First, we document that the scandal reduced the sales of each non-VW German automaker individually – a 15.1 percentage point lower growth rate for BMW in the twelve months following the scandal, a 6 percentage point lower rate for Mercedes-Benz, and a 30.8 percentage point lower rate for Smart.
  • Second, we document a deterioration in positive Twitter sentiment toward the non-VW German automakers. Prior to the scandal, 12.3% of Twitter posts mentioning these automakers displayed positive sentiment. In the two weeks following the scandal, that share declined by 3.5 percentage points. 
  • Third, we show that our results are not primarily driven by diesel cars, despite the scandal’s origins in the diesel market. The non-diesel cars of non-VW German automakers experienced a sales growth rate decline of 9.6 percentage points, close to our baseline estimate of 10.4 percentage points.
  • Finally, we show the robustness of our baseline result across a variety of control groups, including US cars versus non-US, non-German cars; and low-price makes versus high-price makes, etc.

Our results substantiate the opening claim in Tirole (1996) that “[c]ollective reputations play an important role in economics and the social sciences. Countries, ethnic, racial or religious groups are known to be hard-working, honest, corrupt, hospitable or belligerent.” Further, we show that the actions of one member of a group can materially damage the group’s reputation, producing negative reputation externalities from the standpoint of individual firms.

Broader implications

Our study is related to four additional themes.

First, a recent literature in international macroeconomics (e.g. di Giovanni and Levchenko 2012, di Giovanni et al. 2014, 2015) – emphasises granularity and the importance of large international firms; our results suggest that misbehaviour at such firms can damage the collective reputation of particular national powerhouse industries.

Second, the international economics literature has examined the extent to which taste shocks for domestic versus foreign goods can explain the co-movement of international business cycles (Stockman and Tesar 1995). Our results suggest that the misbehaviour of large multinational firms might generate such taste shocks through reputational spillovers.

Third, our results provide a case study for the recent macroeconomic literature on customer capital; our evidence shows how customer capital can decline through reputational spillovers and quantifies the economic consequences of such a loss (Drozd and Nosal 2012, Gourio and Rudanko 2014).

Fourth, in what Berry et al. (1995) call the ‘characteristics approach’ to demand estimation, researchers must specify what properties of products enter into consumers’ preferences. Our results suggest that country of origin might be an important such attribute, consistent with an existing marketing literature (Newburry 2012).

Concluding remarks

The substantial spillovers we document suggest the need to understand what, if any, policy steps are required to address the centrality of national companies to the reputation of the country as a whole. Our results could provide an argument for policy instruments that would incentivise large companies to internalise their potential reputation spillovers.

References

Angrist, J and A Krueger (1999), “Empirical strategies in labor economics”, in O Ashenfelter and D Card (eds), Handbook of Labor Economics 3, Elsevier, Ch 23: 1277–1366.

Bachmann, R, G Ehrlich and D Ruzic (2017), “Firms and collective reputation: The Volkswagen emissions scandal as a case study”, CEPR Discussion Paper No. 12504.

Barrett, S, R Speth, S Eastham, I Dedoussi, A Ashok, R Malina and D Keith (2015), “Impact of the Volkswagen emissions control defeat device on US public health”, Environmental Research Letters 10(11): 114005.

Berry, S, J Levinsohn and A Pakes (1995), “Automobile prices in market equilibrium”, Econometrica 63(4): 841–890.

Chambers, M (2015), “VW scandal threatens ‘Made in Germany’ image”, Reuters, 22 September.

Destatis, (2015), “Automobilindustrie liefert wichtigstes Exportgut”, Destatis, 14 October.

di Giovanni, J and A Levchenko (2012) “Country size, international trade, and aggregate fluctuations in granular economies”, Journal of Political Economy 120(6): 1083–1132.

di Giovanni, J, A Levchenko and I Mejean (2014), “Firms, destinations, and aggregate fluctuations”, Econometrica 84(2): 1303–1340.

di Giovanni, J, A Levchenko and I Mejean (2015), “The micro origins of international business cycle comovement”, NBER Working Paper No. 21885.

Drozd, L and J Nosal (2012), “Understanding international prices: Customers as capital”, American Economic Review 102(1): 364–395.

Gourio, F and L Rudanko (2014), “Customer capital”, Review of Economic Studies 81(3): 1102–1136.

Holland, S, E Mansur, N Muller and A Yates (2016), “Damages and expected deaths due to excess NOX emissions from 2009 to 2015 Volkswagen diesel vehicles”, Environmental Science & Technology 50: 1111–1117.

Newburry, W (2012), “Waving the flag: The influence of country of origin on corporate reputation”, in T Pollock and M Barnett (eds), The Oxford Handbook of Corporate Reputation, Oxford University Press.

Nienaber, M (2015), “Volkswagen could pose bigger threat to German economy than Greek crisis”, Reuters, 23 September.

Oldenkamp, R, R van Zelm, and M Huijbregts (2016), “Valuing the human health damage caused by the fraud of Volkswagen”, Environmental Pollution 212: 121–127.

Petroff, A (2015), “What Volkswagen means to the German economy”, CNN Money, 23 September.

Stockman, A, L Tesar (1995), “Tastes and technology in a two-country model of the business cycle: Explaining international comovements”, American Economic Review 85(1): 168–185.

Tadelis, S (1999), “What’s in a name? Reputation as a tradeable asset”, American Economic Review 89(3): 548–563.

Tirole, J (1996), “A theory of collective reputations (with applications to the persistence of corruption and to firm quality)”, Review of Economic Studies 63(1): 1–22.

United Nations (2017), UN comtrade data base.

Zeit Online (2017), “Daimler soll mehr als eine Million Fahrzeuge manipuliert haben”, Zeit, 12 July.

1,470 Reads