Capital market response to high quality annual reporting: evidence from UK annual report awards

Abstract

We examine the capital market response to the publication of annual reports shortlisted for corporate reporting awards. We find weaker capital market reactions to the publication of shortlisted annual reports compared with a matched sample of non-shortlisted annual reports, consistent with shortlisted reports containing similar or less price sensitive information relative to non-shortlisted reports. Further analysis shows that firms publishing shortlisted reports are more likely to release information to investors in a timelier manner throughout the financial year. We complement our archival empirical analysis with interview evidence from FTSE350 executives and consultants to shed light on the motives for investing in high-quality annual reports. Collectively, our results support the view that high quality annual reporting reflects superior firm-level investor communication processes and that the broader corporate reporting cycle shapes the information role of firm reporting.

Keywords:

• capital market
• event study
• annual report
• quality awards
• corporate reporting cycle
• timeliness

JEL Codes:

• M40
• M41
• M48

Acknowledgments

We are grateful for comments and advice from Mahmoud El-Haj, Paul Rayson, and workshop participants at the University of Saskatchewan and the European Accounting Association Annual Congress 2017. We thank Martin Walker and Thomas Schleicher for helping us with the collection of the interview data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00014788.2022.2106542.

Notes

1 The International Accounting Standards Board (IASB) does not provide a formal definition of the annual report. International Standard on Auditing (ISA) 720 (Revised) describes the annual report as ‘a document, or combination of documents … An annual report contains or accompanies the financial statements and the auditors’ report thereon and usually includes information about the entity’s developments, its future outlook, a risks and uncertainties statement by the entity’s governing body, and reports covering governance matters’ (IAASB 2015, p. 7, para. 12a).

2 Report Watch is an international ranking for which we have considered U.K. firms that rank at B⁺ or higher.

3 Public disclosures may also change the investor base (i.e., foreign, government, institutional holdings).

4 The advantage of a continuous ranking variable is ease of matching high- to low-quality reports. Absent this ranking information, we consider firms that are shortlisted during the sample period to be potentially high-quality and hence, eliminate observations for shortlisted firms for the four years around the shortlisting. This mitigates against inclusion of highly ranked, yet non-shortlisted, reports in the control sample (see Section 4.1 for further details).

5 Controlling for firm characteristics in linear regression models does not mitigate bias arising from misspecification of the (potentially non-linear) relation between firm characteristics and narrative reporting quality (Shipman et al. 2017). The accounting literature uses the Heckman model and matching to mitigate selection bias. Although the Heckman model is effective at reducing selection bias, its criteria for application are stringent and accounting models generally do not meet these requirements (Lennox et al. 2012).

6 Although endogeneity is mitigated using matching, it is not eliminated because unobservable or unidentified determinants (i.e., characteristics affecting both high-quality reporting and capital market outcomes) may exist that are not accounted for in the model (Caliendo and Kopeinig 2008). To mitigate the possibility that correlated omitted variables bias our results we undertake Rosenbaum bounds analysis. Shipman et al. (2017) also identify limitations specific to PSM and its implementation including systematic exclusion of treatment observations with large treatment effects, and propensity to match treatment and control observations close to the cut-off (i.e., matching award-shortlisted observations to high-quality but non-shortlisted observations). We mitigate sample selection and generalizability concerns by matching with replacement, eliminating firm-year observations where a firm has been shortlisted in the previous or subsequent two years, and comparing PSM results to entropy balancing.

7 We match on propensity score with replacement to minimise bias from elimination of control observations while avoiding poor matches. We also test the sensitivity of outcomes to repeated control observations as recommended by Stuart (2010), who notes that repeated observations in the control sample violate the assumption of independence. Austin (2011), using Monte Carlo simulations in the medical field, finds that an optimal caliper of 0.2 times the standard deviation of propensity scores mitigates at least 99% of bias in confounding variables. We follow Austin (2011), which in our setting results in a caliper of 0.045.

8 We do not use firm fixed effects in Equation (1)(1) $\log {\displaystyle \frac{p_t}{1-p_t}=\propto +\sum _{\mathrm{i}=1}^{\mathrm{I}}{\beta }_i\mathrm{FIRM}\mathrm{CHARACTERISTIC}{S}_t+\sum _{j=1}^J{\delta }_j{\mathrm{TIME}}_t+\sum _{k=1}^K{\gamma }_k{\mathrm{IND}}_t.}$(1) as award-shortlisting tends to be sticky over time, hence including firm fixed effects would subsume other explanatory variables.

9 In robustness tests we adjust the time window used to calculate capital market outcomes to check the sensitivity of our results to this research design choice.

10 PwC: http://www.pwc.co.uk/building-public-trust-awards/index.jhtml; IR Society: http://www.irs.org.uk/events/dinner; Communicate Magazine: https://www.communicatemagazine.com/awards/; Accountancy Age: http://www.britishaccountancyawards.co.uk/; Report Watch: http://www.reportwatch.net/

11 The dataset of annual reports that have been identified as having been shortlisted for an annual report award is available at doi:10.17635/lancaster/researchdata/534.

12 As an example of overlap, in 2009 PwC and IR Society shortlisted Great Portland Estates, PwC and Accountancy Age shortlisted HSBC, and IR Society and Accountancy Age shortlisted Marks & Spencer. Therefore, PwC, IR Society and Accountancy Age each had two observations that overlapped with another awarding body.

13 Testing differences in covariates is controversial because discarding matched observations reduces power thereby inflating statistical significance independent of covariate balance (Austin 2008). Hansen (2008) argues, however, that both match validity and outcome analysis suffer from reduced power, therefore in the event of a type I error in match validity (rejecting the null hypothesis that the treatment and control samples are the same) arising from over-trimming in the sample, the reduction in power is sufficient to undermine causal effects in outcome analysis. Caliper tests are another way of assessing the quality of match through case and control group overlap. Austin (2011) argues that using a caliper of 0.2 times the standard deviation mitigates at least 99% of bias in confounding variables. (The standard deviation of shortlisted firms is 0.224 and therefore we use a caliper of 0.045.)

14 Brown and Tucker (2011) posit that year-on-year annual report modification is associated with market reactions. Therefore, to ensure the robustness of our timeliness results, we control for year-on-year annual report modifications (untabulated). Specifically, we add ΔWORDS, the log change in number of words, ΔSECTIONS, the change in the number of sections reported, and ΔPAGES, the change in the number of pages in the timeliness specification. Results for this test are in line with the results in Table 10.

15 Interviews took place in London over the period June to September 2017.

16 Customers and NGOs may also use the annual report (CF1; CF2)

17 For example, the earnings announcement includes commentary on performance while strategy days (often co-occurring with the half-year results release) review business model and strategic objectives.

Additional information

Funding

Funding was provided by the Economic and Social Research Council [contracts ES/J012394/1, ES/K002155/1, ES/R003904/1 and ES/S001778/1], Institution of Chartered Accountants in England and Wales’ charitable trusts, and the Institute of Chartered Professional Accountants of Saskatchewan.