Capitalising or Expensing Development Costs? – Mixed Methods Evidence on the Determinants and Motives of the Accounting Policy in the context of UK Private Companies

ABSTRACT

Following calls for further research, this study evaluates the determinants and motives behind private companies’ decision to capitalise development costs by using mixed methods. While prior literature and expert interviews indicate initially that private firms may be motivated opportunistically, subsequent archival analyses show that development costs are capitalised to meet benchmarks and ameliorate poor profitability. Additionally, interview evidence emphasises that debt covenant violation avoidance and increasing merger and acquisition (M&A) values are important drivers for capitalisation, whereas management compensation schemes do not seem to influence their accounting policy. Moreover, findings imply a negative association between firm size and the capitalisation of development costs. Expert interview evidence indicates that smaller companies are more likely to have financing needs, suggesting that capitalisation is employed to signal future economic benefits to investors. Conversely, the motivation for larger companies which are more likely to expense may be grounded on risk avoidance from future impairments.

Keywords:

• Private firms
• accounting policy choice
• R&D investment
• development costs capitalisation

JEL Classifications:

• D82
• M40
• M41
• O3

Acknowledgements

We gratefully acknowledge helpful comments and suggestions received from the Editors (Araceli Mora and Andrei Filip), two anonymous reviewers, toghether with those from Jill Collis, Francesco Mazzi and Ioannis Tsalavoutas. Additionally, we appreciate valuable feedback from seminar participants at the University of Bamberg and the Brunel University London as well as from session chair and participants of the 14^th Workshop on European Financial Reporting EUFIN 2018 (Stockholm, August 2018), the British Accounting & Finance Association (BAFA) Annual Conference 2019 (Birmingham, April 2019) and the 42^nd Annual Congress of the European Accounting Association (EAA, Paphos, May 2019). Finally, we thank our anonymous interview partners for their willingness to contribute to our research project. All errors remain our own.

Disclosure Statement

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

Notes

1 Tashakkori and Creswell (2007) consolidated the mixed methods approach by defining it broadly as a type of research for which the researcher in a single study: gathers and analyses quantitative as well as qualitative data; integrates the results; and finally derives conclusions. Building on this strategy, we decided to combine the quantitative and qualitative approach, which has the advantage of compensating for the weaknesses in each individual method with the counter-balancing strengths of the other (Jick, 1979). Our key objectives in using this mixed methods design were to obtain a better understanding about the research issue than we would have done with a single method (Creswell and Plano Clark (2007)); to extend findings beyond those achievable and observable using an isolated approach (Grafton et al., 2011); and to validate our results not only by illuminating the research issue from different perspectives, but also by considering the convergence of findings from different research designs (Denzin (1989)).

2 Details on the interview design are available in Appendix 2.

3 Prior interview studies already address R&D accounting treatment, but with different focuses. For instance, Ball et al. (1991) conduct 30 interviews with small companies from the Unlisted Securities Market, the Third Market and the OTC market to analyse the influence of R&D accounting conventions on internal decision-making of companies. Furthermore, Entwistle (1999) examines the R&D disclosure environment within Canadian technology-intensive firms by interviewing 12 analysts and 21 firm executives. Mazzi et al. (2019b) interview key stakeholders of public, IFRS-adopting firms like preparers, auditors and investors to gain insights into the capitalisation of R&D, related disclosure and its decision usefulness. More recently, Mazzi et al. (2022) explore contemporary buy-side and sell-side equity investor views on accounting for R&D costs under IAS 38, mainly adopted by public firms.

4 See information about the research project at https://www.frc.org.uk/accountants/accounting-and-reporting-policy/research/intangibles-how-can-business-reporting-do-better.

5 The scope of our study covers large, medium-sized and small private companies, which are required to prepare their consolidated and individual financial statements in accordance with FRS 102 or voluntarily according to FRS 101 or EU-adopted IFRS. Please note that companies meeting the qualifying criteria according to the micro-entities regime as set out in The Small Companies (Micro-entities’ Accounts) Regulations 2013 (SI 2013/3008) can optionally apply FRS 105. See Collis et al. (2017) for an overview of the UK GAAP.

6 For further information on tax reporting of R&D expenditures, see HMRC internal manual CRID81450, available at https://www.gov.uk/hmrc-internal-manuals/corporate-intangibles-research-and-development-manual/cird81450, and CRID81700, available at https://www.gov.uk/hmrc-internal-manuals/corporate-intangibles-research-and-development-manual/cird81700.

7 Earnings management is defined as a manager’s accounting and transaction decision aimed at amending financial reports either to mislead stakeholders about the company’s true economic performance or to influence contractual outcomes depending on reported accounting numbers (Healy & Wahlen, 1999).

8 It should be noted, that we decided to quote all interviewee verbatim for the sake of authenticity.

9 If a company expenses its development costs, the operating cash flow decreases because any expenses incurred reduce the net cash provided by operating activities. In the following periods, the operating cash flow will remain constant. In the case of development costs capitalisation, expenses do not affect the operating cash flow but that of investment. Subsequently, the operating cash flow is constant over the entire useful life of the internally generated intangible asset, as neither the recognition nor the later reversal of deferred tax liabilities nor the amortisation to be offset have an effect on cash outflows. In summary, although the option to capitalise development costs has no influence on the actual amount of a company's overall cash flow, it does affect the allocation within the cash flow statement as capitalisation leads to a redesignation from operating to investment cash flow. See e.g. Coenenberg and Meyer (2003); Wohlgemuth (2007) or Cazavan-Jeny et al. (2011).

10 Compared to large private companies, small private firms in the United Kingdom can opt for an audit exemption (cf. Companies Act 2006 Section 475 and 477).

11 Interview evidence identifies the risk of future impairments from capitalised costs as one of the main reasons for expensing development projects regardless of a company’s size. This is in line with interview study findings for listed companies from Entwistle (1999) as well as Mazzi et al., (2019b), which show that executive managers oppose R&D capitalisation to avoid potential impairment losses in future accounting periods. Furthermore, I6 noted that companies capitalise the costs in order to attract funding early in the development phase whereas impairments follow relatively late and, therefore, move into the background of their considerations. Beyond that, I3 remarked that large companies in particular are additionally able to use strategic tricks concerning the redefinition of cash generating units to avoid impairment losses.

12 Additionally, following prior research (e.g. Markarian et al., 2008; Dinh et al., 2016; Kreß et al. 2019; Mazzi et al., 2019a; Mazzi et al., 2019b) we also ran a Tobit model to analyse determinants for the value of R&D investment a company capitalises during the year divided by total sales (RDCAP). These results are discussed in the robustness tests (see section 5.2.4).

13 Based on prior R&D studies, which also contain details on the rationale and theoretical justifications for the inclusion of main as well as control variables, e.g. Aboody and Lev (1998); Smith et al. (2001); Wyatt (2005); Markarian et al. (2008); Oswald and Zarowin (2007); Oswald (2008); Cazavan-Jeny et al. (2011); Eierle and Wencki (2016); Dinh et al. (2016); Kreß et al. (2019); Mazzi et al. (2019a); Mazzi et al. (2019b); as well as Dargenidou et al. (2021).

14 Evidence from prior research about the interplay of accounting earnings management (capitalisation of development costs) and the real earnings management counterpart (cutting R&D spending) is discussed in section 3.2. Following Dinh et al. (2016), we expect a negative association with the capitalisation of development costs.

15 It should be noted that according to FRS 102.1.15 any amendments with regard to the small entities regime are effective for accounting periods beginning on or after January 1, 2016, but early adoption is permitted for accounting periods beginning on or after January 1, 2015.

16 In order to consider the different size classifications, we followed requirements from of the Companies Act 2006 section 465. Due to size class adjustments during the observation period, we calculated all descriptive and multivariate analyses with the qualifying conditions required in 2015 and 2016. Hence, a company qualifies as large (small and medium-sized) if it exceeds (does not exceed) two or more of the following thresholds: 1. Turnover: £25.9m/36m, 2. Balance sheet total: £12.9m/18m, 3. Number of employees: 250. The use of both old and new thresholds, provided us with the same results (see section 5.2.4).

17 Naturally, the highest correlation coefficients of above 0.8 can be observed between our dependent variables related to the capitalisation decision CAP and CAP_RATIO. In order to demonstrate that multicollinearity is not of a major concern in our regression models, we additionally report Variance Inflation Factors (VIF) as already explained in section 4.1.

18 Even weaker are the findings for the variables GROWTH, AGE, IMPACT_INTANG, RD_INT and INDEPEND since we find only two or less significant coefficients for the total 4 calculated correlations (Spearman and Pearson correlations for CAP and CAR_RATIO).

19 Moreover, it is noticeable that descriptive statistics (see Table 3 Panel B) and correlation results (see Table 4) show a statistically positive association between a company’s impairment losses from capitalised projects and the capitalisation of development costs. Furthermore, additional regression analysis with the amount of development costs a company capitalises show a statistically positive coefficient at the 10% significance level (untabulated). Hence, it can be assumed that companies determine the amount of development cost capitalisation at its realised impairment loss of R&D assets to compensate for this damage, implying that impairments of capitalised development costs also play a role in a company’s capitalisation decision, albeit in a subordinate form.

20 In contrast to the signalling strategy, one interview partner cited the disclosure of sensitive information as a reason for not capitalising development costs: ‘You might speculate that keeping proprietary information confidential is another [reason for expensing].’ (I2). Likewise, prior literature (e.g. Verrechia, 1983; Wagenhofer, 1990; Dye, 2001; Wagenhofer and Ewert (2015); Eierle & Wencki, 2016) suggests that disclosing information about R&D projects could be used to the detriment of the capitalising company.

21 For the classification in the size classes see section 5.1 and footnote 16. Detailed variable definitions are available in Appendix 1.

22 In doing this, we focus on the variables BEAT_BENCH, ROA as accounting earnings management motives and CUT_RD as real earnings management variable.

23 For instance, survey evidence from Ball et al. (1991) indicates that expensing companies claim inter alia to react to their auditors’ advice. Moreover, our interviewees could also imagine that auditors tend to advise clients to expense development outlays; for example, I2 said: ‘I would also speculate that a lot of private firms […] are influenced by their auditors. And their auditors would tend to push them towards expensing on the grounds that it saves them work, it saves exposure on why did you [decide a development project] to be capitalised and then it was written off two years later.’ (I2).

24 In this context, previous research should be mentioned. For instance, prior studies indicate that the capitalisation of development costs leads to greater audit efforts and consequently increases audit costs (see e.g. George et al., 2013; Cheng et al., 2016; Kuo & Lee, 2018; Kreß et al., 2019; Mazzi et al., 2019b). In contrast, Krishnan and Wang (2014) conclude that the capitalisation of software development costs induces auditors to lower the assessment of audit risk in case of clients’ lower earnings management incentives.