Enforcement of Accounting Standards in Europe: Capital-Market-Based Evidence for the Two-Tier Mechanism in Germany

Abstract

On the background of regulatory initiatives that mandate the establishment of comparable enforcement systems in EU jurisdictions to ascertain consistent and faithful application of IFRS, this paper provides capital-market-based evidence on investor reactions for one specific institutional set-up: the two-tier enforcement system in Germany. In operation since 2005, the German enforcement mechanism consists of a private body, the DPR, which investigates compliance of published financial reports of firms listed on a regulated market segment and, upon error findings, involves the German securities regulator BaFin, which on a second level enforces disclosure of these findings to establish adverse disclosure (‘name and shame’). For a sample of error findings published in the period 2005–2009, we investigate short- and long-term market reactions to error announcements. Results for abnormal returns, abnormal trading volumes and abnormal bid‐ask spreads indicate that these announcements represent new, negative information and suggest that, despite an enforcement environment that is categorised as weak in the extant literature, the activities of the DPR/BaFin seem to penalise infringing firms and thus provide potential deterrence. Multivariate analyses yield weak evidence that the magnitude of the market value discount is positively associated with the severity of the errors, with the threat of subsequent litigation and with cases in which firms disagreed with the error findings of the DPR.

Acknowledgements

We gratefully acknowledge helpful comments and suggestions by Steve Young, the associate editor, and two anonymous reviewers. We would also like to thank Ulf Brüggemann, Willem Buijink, Dirk Hachmeister, Stephan Hollander, Olaf Korn, Laurence van Lent, Bernhard Pellens, Christian Stadler, Hung Tran, Herbert Meyer and Axel Berger (former president and vice-president, respectively, of the German Enforcement Panel DPR), two anonymous reviewers of the 72nd Annual Meeting of the German Academic Association for Business Research in Bremen 2010, the participants of the INTACCT workshops in Valencia (February 2009) and Varna (March 2010), of the Annual Congress of the EAA 2010 in Istanbul, of the 2010 AAA Annual Meeting in San Francisco and of the Ruhr-University Doctoral Colloquium in Accounting (October 2009). This research was supported by the Ruhr-University Research School. Jörg-Markus Hitz gratefully acknowledges the financial contribution of the European Commission Research Training Network INTACCT (Contract: MRTN-CT-2006-035850). All remaining errors are ours.

Notes

Similarly, a recent CESR review of the implementation of Standard No. 2 documents that only 9 out of 29 member states have adopted all of the principles on the coordination of enforcement activities (Berger, 2009).

For example, La Porta et al. (1998) assign a low rank to Germany in terms of anti-director rights, efficiency of the judicial system and rule of law. La Porta et al. (2006), who focus more narrowly on enforcement of securities and disclosure laws, classify Germany consistently as far below average in terms of liability standards and public enforcement. An aggregation of these measures and results is used by Leuz et al. (2003). Similarly, Hope (2003) takes some of the La Porta et al. (1998) variables, which he combines with data on auditor spending to arrive at an enforcement variable that again indicates below average quality for Germany.

Due to insufficient analyst coverage for small firms of which our sample mostly consists, we are not able to investigate the impact of DPR/BaFin error announcements on analyst forecast dispersion, analyst following (Dechow et al., 1996) or analyst forecast revisions (Feroz et al., 1991; Palmrose et al., 2004).

We estimate the market-model using the continuously compounded definition of daily stock returns based on a return index to take dividend payments explicitly into account. The non-continuously compounded return definition leads to slightly less normally distributed values of CAR. The empirical findings are largely equivalent but less significant by trend. We also run our tests using the CDAX^® (a value-weighted index of all publicly listed German firms), which leads to slightly significant negative mean and median cumulative abnormal returns for the three- and five-day event window. In line with Peterson (1989) and Brown and Warner (1985), we note that the equally weighted portfolio approach is more appropriate to address our research question as the correlations of the single stock return series and of the index are considerably higher on average.

We also use a definition of the relative trading volume that does not control for the rate of free float. The (untabulated) results are largely comparable but lead to more negative cumulative abnormal relative trading volumes by trend, especially for the event day and for the three-day event window.

Relative bid‐ask spreads are defined as the absolute difference of the daily closing bid and ask prices scaled by the midpoint. Bid‐ask spreads are only taken into account if they are based on positive trading volume. In line with Gassen and Sellhorn (2006), we note that intraday bid‐ask spreads are more suitable than closing price bid‐ask spreads because of the distortion at the end of the trading day. However, the greater overall availability of the closing price prompted us to use the first definition. If the shares of a firm are traded on more than one stock exchange, we select the main stock exchange, that is, the stock exchange with the highest trading volumes over the investigation period of the study.

Heckman et al. (1998) criticise PSM because it is not necessarily better than an archival approach in terms of reducing the variance of the resulting estimator. PSM can only control for differences between groups that can be identified by observable variables, but not for differences in unobservable variables. In spite of this, we opted for PSM because, in contrast to a simple matching approach, it provides a natural weighting scheme that yields unbiased estimates of the treatment effect (Dehejia and Wahba, 2002).

Overall, the respective sample sizes are comparable to those used in previous studies. For example, Feroz et al. (1991) analyze a sample of 58 firms subject to SEC Accounting and Auditing Enforcement Releases, and Peasnell et al. (2001) study characteristics of 47 cases of 43 firms publicly censured by the British FREP (Financial Reporting Enforcement Panel) for having produced erroneous accounting. Please note that we do not exclude financials from our analyses. Thus, balance sheet ratios are potentially biased. However, the major findings of our study also hold after a deletion of financials.

For example, Feroz et al. (1991) analyse a sample of 58 firms subject to SEC Accounting and Auditing Enforcement Releases, and Peasnell et al. (2001) study characteristics of 47 cases of 43 firms publicly censured by the British FREP (Financial Reporting Enforcement Panel) for having produced erroneous accounting.

Although the number of firms subject to SEC oversight is substantially higher than in Germany and the SEC enforcement activities are thus substantially larger and better staffed, Feroz et al. (1991) document only 188 Accounting and Auditing Enforcement Releases for the period April 1982–April 1989, and Dechow et al. (1996) investigate only 201 GAAP-related Accounting and Auditing Enforcement Releases for the period 1982–92. In contrast, the DPR/BaFin error announcements for the short period under inspection (effectively four years) already amount to 109 observations. Apparently, publication of SEC error findings represents a relatively rare event in comparison.

Consistent with this, Dechow et al. (1996) regard their sample of firms censured by the SEC via publication of an Accounting and Auditing Enforcement Release explicitly as ‘firms that manipulated earnings’.

Additional information

Notes on contributors

Jörg-Markus Hitz

Paper accepted by Salvador Carmona.

Jürgen Ernstberger

Paper accepted by Salvador Carmona.

Michael Stich

Paper accepted by Salvador Carmona.