![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
Using a sample of Portuguese audit firms and their client companies, this study examines the association between gender composition of the partnership structure of audit firms and audit quality. Audit quality is measured through the earnings quality of audit clients. We find that gender diversity in the partnership structure of audit firms, per se, has no association with audit quality. In turn, we find significant evidence that a predominant presence of female Certified Public Accountants (CPAs) in partner positions in audit firms is associated with higher audit quality. In particular, the results indicate that audit firms with female-dominated partnership structures are negatively related with aggressive accounting practices in audit clients.
Utilizando una muestra de firmas auditoras portuguesas y sus clientes, este trabajo analiza la relación entre la composición de género de la estructura de propiedad de las firmas auditoras e la calidad de auditoría. La calidad de auditoría se mide a través de la calidad de la información financiera de los clientes. Los resultados sugieren que la diversidad de género en la estructura de propiedad de las firmas auditoras no está asociada con la calidad de auditoría. Por otra parte, evidencia significativa sugiere que la supremacía femenina en las estructuras de propiedad de las firmas auditoras se asocia con una mayor calidad de la auditoría. Las firmas auditoras con estructuras de propiedad donde predominan las mujeres, presentan una relación negativa con las prácticas de manipulación contable agresivas de los clientes de auditoría.
Acknowledgements
We thank Professor Ana Carvalho for her useful suggestions and the participants of the seventh EARNet Symposium, 2013, for their helpful insights. We also appreciate the support of the Editor, Professor Beatriz García Osma, and two anonymous reviewers, since their comments and suggestions have significantly improved the quality of our work.
ORCID
Tânia Menezes Montenegro 
http://orcid.org/0000-0003-1071-4368
Notes
1. The Deloitte (DTTL, Citation2013) report Women in the boardroom: A global perspective provides details about countries’ legislative and initiatives to promote gender diversity in top corporate levels.
In the European context, the European Commission proposed legislation with the aim of attaining a 40% objective of the female non-executive board-member positions in publicly listed companies, with the exception of small/medium enterprises. In Portugal a Government Resolution of 8/03/2012 requires state-owned companies to adopt gender equality plans to promote gender balance in management and executive positions. It has also recommended that listed companies adopt such plans [http://ec.europa.eu/justice/newsroom/gender-equality/news/121114_en.htm].
3. Portuguese audit firms are generally small professional service firms with small hierarchical structures, owned by either CPA partners or non-CPA partners. However, the Portuguese law makes it mandatory that CPA partners have a qualified majority in the ownership of the firm, and requires that only the CPA partners are responsible for the supervision and execution of the financial audit services. Non-CPA partners are a minority in Portuguese audit firms (in our samples, the larger percentage of non-CPA partners in an audit firm is lower than 3% of the total partners).
5. Data on CPA partners’ gender were collected from the annual membership lists issued by the Portuguese Institute of Statutory Auditors (Ordem dos Revisores Oficiais de Contas, OROC). Although the first regulation of the audit profession occurred in 1972, only from 1973 on has OROC been providing two different annual membership lists. One gives an overview of the names and addresses of registered auditors (each auditor has an id number), the date of first registration as a CPA, the audit firm they are associated with (as partners) and the CPA professional status (active or suspended). The other provides information of the name and address of the registered audit firms (each firm has an id number), the date of the first registration at OROC, the name of the CPAs enrolled with the firm (identifying the partners and non-partners) and the firm status (active or suspended). The CPA partners’ gender was unambiguously described by his/her first name (and second when applicable) available in the annual membership OROC lists.
6. A measure of full dominance by female CPA partners (percentage of female CPA partners equal to 100%) was not tested because in the full panel only 11 companies are audited by audit firms fully dominated by female CPA partners.
7. To capture the overall magnitude of the male and female CPA partners’ general experience in auditing of an audit firm, based on Madsen, Mosakowski, and Zaheer (Citation2002), the MGenExp and FGenExp variables are computed as the average experience of an audit firm’s male and female CPA partners. The CPA partner’s amount of general experience is defined as the number of years since the auditor has obtained his/her certification as CPA and which his/her enrolment in the annual membership lists of the OROC is valid (the time periods during which the CPA has the professional status suspended are eliminated).
8. The correlation between the abnormal accruals values (absAWCA) obtained from the Jones Modified model and the DD model is 0.52 (significant at the 1% level).
9. The modified version of the Jones (Citation1991) model that we use focus exclusively on the working capital component of accruals because it appears that working capital accruals are more likely than other accruals to be subject to manipulations (e.g., Ashbaugh, LaFond, & Mayhew, Citation2003; Burgstahler & Dichev, Citation1997).
10. Ball and Shivakumar (Citation2006) use four different proxies for the existence of gains or losses in the current year, three of them financial reporting (“book”) measures (level of cash flows, change in cash flows and industry-adjusted cash-flows) and the fourth is based on stock market returns.
11. SABI (Sistema de Análise de Balanços Ibéricos) is a database from BUREAU VAN DIJK – Electronic Publishing, represented in Portugal by COFACE-MOPE – Serviços de Informação e Gestão de Cobranças, SA, that provides standardised financial statements data.
12. As described by Guerreiro, Rodrigues, and Craig (Citation2012), in Portugal, the Sistema de Normalização Contabilística (SNC) (Portuguese Accounting Standards) came into force in January 2010. SNC is the IFRS adaptation to the Portuguese context and to be followed by unlisted companies. In the period 2005–2009, Portuguese law allowed listed companies to use IFRS in their individual accounts and unlisted companies in either their consolidated and their individual accounts. However, the voluntary adoption of IFRS in individual accounts of listed and unlisted companies was under two conditions: first, that companies maintained another accounting system for individual accounts in accordance with the national accounting system; and second, that companies must have had their accounts publicly certified. In this study we use individual financial statements to infer the quality of the reported earnings of firms to the four-year period from 2003 to 2006. Thus, transition from Portuguese GAAP to IFRS does not affect our measures of earnings quality.
13. Examples of studies that estimated the cross-sectional or time-series versions of the Jones-type models that required a minimum of about 10 observations for efficiently estimating the regression parameters include Dechow et al. (Citation1995), Dowdell and Krishnan (Citation2004) and Ferguson et al. (Citation2004).
14. In 2013, the total percentage of female CPAs reached 23.75%, and in Portuguese audit firms the total percentage of female CPAs occupying partner positions was about 20.4%.
15. Greene (Citation2007) indicates that both fixed-effect and random-effect models control for omitted variables in panel data, and that a fixed-effect model is a special case of the more general random-effect model. A Hausman specification test shows that a fixed-effect model is more appropriate to test Model (1). However, given the short time series of four years in our data, and the low within-variation of the test variables and some of the control variables, we chose a firm-level random-effect model to test Model (1).
In Sample A only 100 firms have within-variation in GDIV (mean value of the variation (standard-deviation) 0.55); 61 have within-variation in FGD (mean value of the variation 0.58); 1062 have within-variation in FEXP (mean value of the variation 2.96) and 2138 firms have within-variation in MEXP (mean value of the variation 1.63). Sample B presents similar results. In the FE approach, observations of audited firms without/or with low within-variation in the GDIV, FGD, FEXP and MEXP variables do not contribute to the identification of effects on abnormal accruals. Nevertheless, we run FE regressions (results not tabulated) and the coefficients on GDIV, FGD, FEXP and MEXP present the same signal and statistical significance as in OLS and RE regressions. Thus, our results are consistent in OLS, RE and FE estimations.
16. We do not include GDIVER in the main analysis because it is highly correlated with GDIV (0.5, significant at the 1% level).
17. Untabulated descriptive statistics reveal that in the subsample of companies whose audit firms are dominated by female CPA partners, the mean value of experience for male CPA partners is 5.4 years (median 7.5 years), while for female CPA partners it is 3.8 years (median 2.5 years). It turns out that in these audit firms the female CPA partners’ experience is considerably lower than that of the male CPA partners’.
