8 and the methods employed. Section 4 presents and discusses the empirical findings. Section
5 illustrates additional tests and Section 6 concludes the paper by outlining the contributions arising from this research. It also discusses limitations and avenues for further research.
2. Background, literature review and hypotheses development
2.1 RD reporting under IFRS Since 2005, all EU publicly traded firms need to report consolidated financial statements
under IFRS,
2
with IAS 38 governing the accounting treatment of intangible assets. IAS 38 requires the capitalisation of development expenditures which meet a specific set of criteria.
In order to capitalise the development costs a company should assess: the technical feasibility of the intangible asset; the intention to complete the asset and with the ability to
sell or use it; the availability of resources, technical or financial, to complete it; the ability to reliably measure the expenditure and the ability to justify that the asset will generate
future economic benefits paragraph 57. Based upon meeting these conditions, a company must then capitalise development
expenditure; it is not a discretionary choice. However, in establishing whether all these criteria are met, there is managerial judgement and hence effective discretion over the
capitalisation decision. As PricewaterhouseCoopers 2010 note, meeting the six conditions
“requires facts and some judgements” p. 7, emphasis added. Thus, if the company decides
that one of the conditions is not met, then it must expense the relevant cost incurred. Alternatively, and consistent with an earnings management and noisy signalling approach
Ahmed Falk, 2006, 2009; Cazavan-Jeny Jeanjean, 2006; Cazavan-Jeny et al., 2011; Ciftci, 2010; Dinh et al., 2015; Markarian et al., 2008; Prencipe, Markarian, Pozza, 2008,
2
In some jurisdictions e.g. Greece and Italy this requirement applied also to listed companies which published only individual accounts Tsoligkas Tsalavoutas, 2011.
9 a company may judge that all of the conditions have been met and effectively over-capitalise
development costs.
2.2 The RD capitalisation debate and academic evidence The accounting treatment of RD has been a controversial issue among standards setters,
financial statement preparers, users as well as academics. Under IFRS, certain development costs can be capitalised see above, with all other costs associated with RD expensed. In
contrast to IAS 38, under US Generally Accepted Accounting Principles GAAP
Statement of Financial Accounting Standards hereafter SFAS 2, all RD costs are expensed. However, Lev, Sarath, and Sougiannis 2005 argue that immediate expensing
leads to biases in earnings and mispriced stock prices and Amir, Guan, and Livne 2007 conclude that uniform expensing of RD costs is
‘overly conservative’ p. 245. One of the reasons behind this differential treatment are apparent concerns by Financial Accounting
Standards Board FASB regarding the uncertainty related to the future benefits expected by the assets capitalised. Additionally, they are concerned that management may use their
discretion to capitalise RD, resulting in a misrepresentation of RD accounting data Davies Wallington, 1999; Healy Wahlen, 1999. This is why Ahmed and Falk 2006
conclude that, standard setters, such as FASB, are concerned that ‘the cost of possible
misstatement to exceed the benefits of signalling’ p. 234 and mandate expensing all RD costs as a result.
3
In summary, within this debate, two issues dominate. Firstly, development costs effectively constitute investments which will result in future economic benefits and as a
result they should be capitalised i.e., not expensed to recognise their current value to the
3
The only exception to this, under US GAAP, relates to software development SD costs which can be capitalised once technological feasibility is established SFAS 86.
10 business and to provide a signal for future earnings arising from successful development
expenditure Lev, Nissim, Thomas, 2008. On the other hand, capitalisation can effectively be used as an earnings management vehicle. These inter-related effects are also highlighted
by Dinh et al. 2015 as follows: ‘RD capitalisation can be exercised by managers to signal
private information on future economic benefits to the market. It can, however, also serve as opportunistic earnings management
’ p. 1 and see Ciftci, 2010. Dealing with signalling, there is mixed evidence within the literature on the link between
capitalisation and future earnings. For instance, in relation to French companies in the pre- IFRS period in which firms were permitted to capitalise development costs under certain
conditions, Cazavan-Jeny and Jeanjean 2006 find a significant and negative relationship of capitalised development costs and share price. In line with our propositions, they suggest that
managers are more likely to take a more opportunistic approach to capitalisation, given the country contextual factors i.e., weak enforcement. Further, Cazavan-Jeny et al. 2011,
again in the same French context, find that decision to capitalise RD expenditures is generally associated with a negative or neutral impact on future performance, and hence
inconsistent with the proposition that managers capitalise RD to convey information about improved performance i.e., genuine signaling effect.
In contrast and supportive of a capitalisation approach, Lev and Sougiannis 1996 in a US based study requiring full expensing of RD find that RD expenses are associated with
future income. In an Australian context, where companies were permitted to capitalise development costs under certain conditions under Australian GAAP, Ahmed and Falk 2006
in their study for the period 1992 to 1999, find that, ‘RD capitalised expenditure is
positively and significantly asso ciated with the firm’s future earnings’ p. 231. Thus, they
argue that managers are able ‘to credibly signal their superior information by either
11 capitalising successful RD investment or expensing unsuccessful RD investment
’ p. 259. The market relevance of capitalised development cost, and the signalling of managerial
information relevant to future earnings, is also noted by Shah et al. 2013 using UK data covering pre and post IFRS adoption,
‘that investors perceive the capitalisation of RD to be related to successful RD projects
’ ibid, p. 168 and see Oswald Zarowin, 2007; Tsoligkas Tsalavoutas, 2011. Finally, in relation to software development SD, in her US
study, Wolfe 2012 finds that the association between capitalised SD expenditures and future cash flows is stronger than the association between reported RD expense and future
cash flows for firms that capitalise. In relation to the potential manipulation of capitalisation associated with earnings
management, whilst much of the prior literature is based upon national GAAP rather than IAS, consistent findings have been reported. Cazavan-Jeny et al. 2011 contend that
managers may capitalise development costs to meet or beat earnings thresholdstargets or to avoid reporting losses. In their Italian study, based on the pre-IFRS period where similar to
France and Australia noted earlier companies were permitted to capitalise development costs under certain conditions, Markarian et al. 2008 conclude on the motivations for
capitalisation as forming part of earnings management for earnings smoothing. Finally, Dinh et al. 2015, in their study on Germany, which covers companies reporting under IFRS,
found that ‘pressure to beat past year’s earnings and analysts’ forecast of earnings increases
the probability of a firm capitalising RD in the current period. This evidence is in line with the notion of firms opportunistically managing earnings via RD capitalisation
’ p. 3. They contend that both the decision to capitalise and how much to capitalise are strongly and
positively associated with benchmark beating. Hence, they conclude that the presence of
12 earnings management counteracts the signaling value of capitalisation, a finding consistent
with the noted regulatory concerns in the US.
2.3 Hypotheses development Prior empirical research has, almost universally, found that corruption has an adverse impact
on business activity due to a lack of transparency and the misallocation and inefficient use of resources. For example, it is widely found to result in lower levels of foreign direct
investment and levels of economic growth Bryant Javalgi, 2016; Voyer Beamish, 2004. Additionally, companies in countries with high levels of corruption tend to have
lower levels of corporate social performance disclosures Baldini, Dal Maso, Liberatore, Mazzi, Terzani, 2016; Ioannou Serafeim, 2012. Finally, earlier research, has shown a
negative relationship between low levels of corruption and information transparency in general DiRienzo et al., 2007; Kimbro, 2002.
In a more accounting specific context, Malagueño et al. 2010 examine accounting and auditing quality and the perceived level of corruption across 57 countries. They find that
both accounting and auditing quality are negatively related to the level of perceived corruption. Fan et al. 2014 highlight the link of earnings opacity and corruption in China,
whereby managers distort accounting information to cover their expropriation of interests from common investors. Similarly, Picur 2004 and Riahi-Belkaoui 2004, in their multi-
country studies, find significant and positive relationships between the level of corruption and the level of earnings opacity, indicative of earnings opacity being susceptible to
corruption. Riahi-Belkaoui 2004, p. 73 asserts that ‘the variations in earnings opacity
internationally suggest the presence of local factors that act as major determinants of its level and change …. and that the level of corruption existing in a particular country is a major
13 determinant of the level of earnings opacity
’. Overall, this literature suggests that in more corrupt environments managers are more likely to distort earnings, reflecting managerial
opportunism. From this, and in line with the prior literature which suggests that managers take an
opportunistic approach to capitalisation of development costs Cazavan-Jeny Jeanjean, 2006; Cazavan-Jeny et al., 2011; Dinh et al., 2015; Markarian et al., 2008, it follows that
managers in a more corrupt environment would capitalise higher amounts of development costs. Hence, we test the following hypothesis:
H1: There is a positive relation between country level corruption and the amount of development costs capitalised.
Assuming that H1 holds, beyond that capitalising and not expensing development costs would result in better current-year earnings, managers falsely signal their private information
in relation to the future economic benefits associated with the capitalised development costs. Thus, one would expect that the contribution of development costs capitalised and
cumulative future earnings in the long-run would not be as high for firms in countries with higher corruption levels. This is because the capitalised development costs would not deliver
as high as signalled future economic benefits. Hence, we test the following hypothesis: H2. The higher the corruption in a country, the lower the association of the capitalised
development costs to future profitability. We are primarily interested in the overall effect of corruption on capitalisation of
development costs. Thus, H1 H2 are our main hypotheses. However, if it is the case that country level corruption influences capitalisation of development costs and the resulting
contribution to future earnings, then the effect of domestic corruption may be affected by the extent to which a particular firm is exposed to foreign norms and behaviours. Thus, we are
posing that the influence of domestic corruption on discretionary financial reporting choices
14 should diminish as firms become more exposed to foreign economic environments.
Following on from this, in more international firms the contribution of the capitalised development costs to future profitability will not be as reduced due to domestic corruption
influences compared to those less international firms. Effectively, we examine if the relation between corruption and capitalisation of development costs, as well as their contribution to
future earnings, is moderated by the degree of internationalisation measured at the firm level. The rationale of these additional hypotheses is as follows.
One of the strategic choices that firms undertake is the internationalisation of their b
usiness activity. Internationalisation can be defined as ‘the process of increasing involvement in international operations’ Welch Luostarinen, 1988, p. 84. Drawing on the
strategic management and international business literature, Reid 1983 highlights the importance of export market characteristics, such as economic and social norms, becoming
recognised and part of a firm’s decision-making processes. As such, firms will adapt and adopt an increasingly market centred, as opposed to domestic, approach to international
business. Johanson and Vahlne 1990 process model of internationalisation further highlights the need for firms to make adaptations in their ways of performing business to
signal their credibility to international partners and networks Ford, 1979; Turnbull Valla, 2013. Moreover, as part of the internationalisation process, Murtha et al. 1998 argue that
managerial thinking and attitudes towards issues such as accountability and global values are critical for strategic change. In line with this, Segaro et al. 2014, in their research
examining SME behaviour, argue that through the process of internationalisation, stewardship and the long-term orientation of the firm become more embedded in managerial
thinking and attitudes. The resource diversity of internationalisation, compared to the domestic environment, positively influences management attitudes and mindsets in a global
15 environment Nadkarni Perez, 2007; OGrady Lane, 1996. From this, it derives that the
more international the firm is the more long term-orientation and stronger elements of stewardship it has Murtha et al., 1998; Segaro et al., 2014.
With regard to corruption in particular, Sandholtz and Gray 2003 assert that as international trade augments international norms, the effect of corruption is reduced on those
involved with trade, consistent with a ‘significant inverse relationship between international
trade and corruption levels ’ p. 765. From the foregoing, it follows that the more
international a firm becomes, compared to those more domestic firms, the less it will be influenced by local corruption c.f., Ades Di Tella, 1999.
These arguments are also informed by the evidence in the accounting literature which provide wider evidence that the impact of country-level corruption at a firm-level is reduced
through internationalisation. For example, Dauth, Pronobis, and Schmid 2017, p. 71 provide evidence that,
‘top management internationalisation mitigates the level of managerial discretion in financial reporting
’. Additionally, Hope, Kang, Thomas, and Yoo 2008 argue that if a domestic firm derives most of its revenues from overseas operations, or if the firm is
cross-listed, then the firm ‘is less likely affected by domestic norms – such as the extent of
secrecy in the country ’ p.361 than other, less internationally-oriented firms.
4
They find that firms in secretive environments are more likely to employ a non-Big 4 auditor with negative
implications for the quality of financial statements. Thus, they suggest that ‘...multinational
firms are less affected by home country cultural norms than are domestic firms ’ p. 371.
In an earnings management context in particular, Lang, Lins, and Miller 2003 and Lang, Raedy, and Yetman 2003 employ cross-listing as a measure of internationalisation and find
4
Whilst they employ secrecy, as a measure of culture Hofstede, 1980, secrecy is commonly associated with a lack of transparency and accountability Gray, 1988, consistent with corrupt environments. In line with this,
as a robustness test, we use secrecy as an alternative proxy for corruption.
16 that cross-listed firms have better, more transparent, information environments and would
appear to be less aggressive in terms of earnings management. Moreover, In fact, ‘multinationals tend to carry out less income-increasing earnings management than domestic
firms ’ Prencipe, 2012: 693. This is particularly relevant to our context given that
capitalisation of development costs results in income increasing earnings management as discussed earlier.
From the foregoing, we would argue that companies with an international focus are less driven by local factors due to internationalisation, and as a result, the influence of domestic
corruption would diminish. In our context, the short-term opportunism by over-capitalising development costs that is influenced by local norms of corrupt behaviour would be less
pronounced. Thus, we hypothesise that the effect of corruption on the amount of capitalised development costs would be lower for those companies that are more international.
H3. The amount of capitalised development costs will be more influenced by domestic corruption in firms with lower levels of internationalisation.
From this and on reflection of H2, it derives that the expected positive effect of capitalised development costs to future earnings is moderated by the extent of domestic corruption when
the company is less international. Thus, we also test the following hypothesis: H4. The association of the capitalised development costs to future profitability is more
negatively influenced by the level of domestic corruption in less international firms compared to more international firms.
17
3. Research design
