Firm Characteristics, Peer-learning and Disclosure Quality under IFRS 15: A European construction-industry based research.

(1)

Firm Characteristics, Peer-learning and

Disclosure Quality under IFRS 15: A European

construction-industry based research.

MSc Accountancy - University of Groningen

Thesis submission date: 22-6-2020

Gerard Jan-Marten Hazekamp

s2940132 Johan de Wittstraat 63 9716CB Groningen +31636593193 g.j.hazekamp@student.rug.nl

Thesis supervisors

Prof. dr. R.L. ter Hoeven

R. van Duuren, MSc.

Wordcount: Abstract - 214 words Thesis – 11.520 words

(2)

Abstract

This paper examines the effect of three financial firm characteristics on IFRS 15 disclosure quality. Additionally, this paper examines whether there are peer-learning

effects in IFRS 15 disclosure quality between the financial years 2018 and 2019. As 2019 will be the second year after the application of IFRS 15, this paper will be one of the first to analyse peer-learning effects in disclosure quality. The setting for this

paper is the European construction industry, due to an increasing complexity in contracts within the industry. Unique for this paper is the self-composed disclosure index, which is used to measure a firm’s disclosure quality under IFRS 15. Disclosure

quality under IFRS 15 was found to be widespread. Although results indicated peer learnings effect to be insignificant, some good examples of peer-learning have been found. Using multiple linear regression models, this paper firstly finds, as expected, leverage to be significantly positively related with disclosure quality. Contrary to our

expectations, we found a significant negative relation between profitability and disclosure quality. Lastly. this paper was not able to find any significant results on the

relation between liquidity and disclosure quality. This paper can be seen as a guideline for further research on the determinants of IFRS 15 disclosure quality and

(3)

1 Firm Characteristics, Peer-learning and Disclosure Quality under IFRS 15:

A European construction-industry based research

1. Introduction

On May 28th of 2014 the IASB announced a new standard for revenue recognition, the IFRS 15: Revenue From Contracts With Customers. This standard has been put into use on January 1st, 2018 (Grant Thornton, 2017) and replaces the former IAS 11: Construction Contracts, IAS 18: Revenue and the related interpretations IFRIC 13, 15, 18 and SIC-31 (Roozen & Pronk, 2018). Until the introduction of IFRS 15, the recognition of revenue has been one of the gaps between IFRS and US GAAP. Due to the implementation of IFRS 15, this gap has been tightened, and the comparability of financial reports across both standards increased (PWC, 2019).

PKF (2015, p.1) states that: “the objective of IFRS 15 is to establish principles that an entity shall apply to report useful information to users of financial statements about the nature, amount, timing and uncertainty of revenue and cash flows arising from a contract with a customer”. Aarab, Bissessur & ter Hoeven (2015) add to that by stating that one of the main goals of IFRS 15 is to provide investors and financial analysts with additional information on revenue recognition to benefit them (i.e. higher the quality of disclosures).

This paper will focus on disclosure quality under the recently released IFRS 15 within the European construction industry. Thereby, this paper will focus on construction companies as the last few decades within whole Europe the complexity of contracts in the construction industry raised. Complexity indicates that projects are of a bigger size and scope. More projects accepted and on average last longer. Furthermore, there is an increasing uncertainty whether projects will be finished in time and will be profitable (Wood & Ashton, 2010).

IFRS 15 will have a continuous impact on several industries, as the disclosure requirements for revenue recognition are more extensive and stricter (KPMG, 2019). Especially, construction companies are socially relevant as most contracts within the construction industry do not only consist of building the construct, but also consist of designing and maintaining the construct (Masterman, 2013). This might result in multiple separate performance obligations to which the total transfer price needs to be allocated when revenue is being recognized. Revenue can be recognized under IFRS 15 when certain milestones within a single performance obligation are met. Each performance obligation recognized under IFRS 15, consists of a pre-arranged set of milestones (KPMG, 2019).

(6)

2

So, contracts within the construction industry can become quite complex in terms of revenue recognition. Therefore, it might be interesting to see how different construction companies gave shape to the requirements for revenue disclosure as set by IFRS 15. To capture those different interpretations on IFRS 15 revenue disclosure requirements, this paper makes use of a self-composed disclosure index in which construction companies are awarded a score based on the quality of the IFRS 15 disclosures. The disclosure index is based on the disclosure requirements set in the standard.

The main area of this paper will focus on financial firm-specific characteristics, forming determinants of disclosure quality under IFRS 15. Research (Abd-Elsalam, 1999; Agyei-Mensah, 2012; Omar & Simon, 2011) showed associations between specific firm characteristics and disclosure quality. The use of an index methodology is common when relating disclosure quality to a number of explanatory financial firm characteristics. Underlying financial firm characteristics differ

systematically across companies and have been found an explanatory factor of the extent and quality of a company’s disclosure (Soyemi & Olawale, 2019).

Most commonly profitability, liquidity and leverage are linked to disclosure quality (Omar & Simon, 2011). Those three financial firm characteristics could all be related to revenues, and thus IFRS 15 disclosure quality. Therefore, this paper expects that financial firm characteristics like a companies’ profitability, liquidity and leverage will influence the disclosure quality of a European construction company under IFRS 15.

First, companies with high profitability are more likely to voluntary disclose than companies with low profitability (Agyei-Mensah, 2012). Companies with a high profitability want to inform stakeholdersthat they are performing well. Stakeholders are aware of the increasing complexity and project-size in the construction industry (Financial Times, 2018). Due to the increased complexity of contracts, revenue recognition under IFRS 15 is subject to a high degree of significant judgements. This requires revenue clarification from the construction companies to their stakeholders. As

profitability is derived from revenues, it is likely that companies will show a higher disclosure quality under IFRS 15 if they show a high profitability.

Secondly, Omar & Simon (2011) found that companies tend to disclosure more information about their ability to meet short-term obligations. IFRS 15 disclosures form an important

communication instrument to stakeholders, as IFRS 15 requires construction companies to disclose more extensive and stricter on their revenue recognition. A phenomenon the last few years within the construction industry is the trend that construction companies are showing a negative or low net working capital in their annual year-reports (van Ommen & Karregat, 2020/1). To clarify this to their stakeholders, we expect construction companies to disclose on how revenue is recognized under IFRS 15, as net working capital and revenue are closely linked to each other.

(7)

3

Lastly, Galego-Álvarez & Quina-Custodio (2016) indicated a relation between the leverage level and disclosure quality of a company, in which a high leverage ratio indicates that a company is more likely to voluntary disclose information. Malone, Fries & Jones (1993) add to this by stating that a high leverage ratio encourages managers to disclosure more information to meet the interests of their stakeholders.

Contrary, Malone et al. (1993) stated that a low leverage ratio could induce managers to orient their disclosure policy more towards specific stakeholders such as shareholders, rather than to all stakeholders, which also includes lenders and creditors. So, because of contradictory views in earlier research, the sign of the relation between leverage and IFRS 15 disclosure quality is yet to be discovered. This all results in the following main research question:

As an additional area to this research question, this paper will focus on peer-learning effects between the financial years 2018 and 2019 concerning disclosure quality under IFRS 15. Kvaal & Nobes (2012) showed evidence that there is a learning process in adopting a new standard within IFRS (or IFRS as a whole). They showed that while getting into a more mature phase, more has been learned, and better results will be shown within the companies’ disclosures. In 2018, construction companies may still have been looking for the best practical implementation of the disclosure requirements of IFRS 15. While, in 2019, construction companies will have had one or two years of experience (depending on the implementation method) which we expect them to have learned from. This results in the following additional research question:

Alongside statistical evidence, we aim to show normative figures that give an indication of how construction companies have improved the disclosure quality concerning a specific IFRS 15 requirement. The use of both a statistical and normative approach is unique, as most research focusses on either one. The normative approach aims to provide more understanding to what has been derived from statistical research.

(8)

4 2. Contribution

This paper contributes to prior research on IFRS 15 by focusing on the construction industry while other researchers focussed on the telecommunication industry (Mattei & Paoloni, 2019), on real estate companies (Trabelsi, 2018) or did not focus on a specific industry (Roozen & Pronk, 2018). What makes the construction industry interesting, is that project contracts in the construction industry consist of multiple performance indicators, complex structures and long-term contract-components (Trabelsi, 2018). Due to this, revenue recognition under IFRS 15 within the construction industry is subject to a high degree of significant judgements.

Secondly, this paper will focus on the European construction market, as Masterman (2013) showed that within the European construction market, contract complexity has raised the last few decades. Furthermore, the European Union adopted IFRS as the required reporting standard for all European companies whose debt or equity securities trade in a regulated market in Europe in 2002. Since 2005, it is mandatory for listed companies in the European Union to use IFRS (IFRS, 2018). This makes Europe one of the main users of IFRS.

Furthermore, Financial firm characteristics like profitability, liquidity and leverage have been positively related with disclosure quality under various reporting standards (Abd-Elsalam, 1999; Agyei-Mensah, 2012; Omar & Simon, 2011). There has been no research yet on the impact of those firm characteristics on IFRS 15 disclosure quality. Also, Kvaal & Nobes (2012) found that companies show and disclose better results each year after adoption of a new IFRS standard. Indicating there is some kind of a peer-learning effect in adopting a new IFRS standard. There has been no research on peer-learning effects after the implementation of IFRS 15 yet.

Lastly, what makes this paper unique is that this paper uses a self-composed disclosure index to measure IFRS 15 disclosure quality. A disclosure index has been used to measure disclosure quality in various studies (Francis, Nanda & Olsson, 2008; Botosan, 1997), but not yet in studies on IFRS 15 disclosure quality. Also, as explained later, the disclosure index used in this study differs from earlier used disclosure indexes.

(9)

5 3. Theoretical framework

This section will start with background literature giving a short overview of the history of IFRS and IFRS 15. Thereafter, we will introduce the main relevant theories. Lastly, based on theory, we will develop and discuss the hypothesis used in this paper.

3.1 Background

3.1a IFRS history

January 1st, 2005, IFRS became mandatory for listed companies in the European Union. Listed companies had to applicate IFRS as a whole. Companies were not allowed to choose which elements of IFRS they want to implement or not (Paananen & Lin, 2009). Before the obligation to report under EU-endorsed IFRS, most listed companies in Europe followed a variety of US GAAP interpretations (Soderstrom & Sun, 2007). IFRS consists of the former IAS-standards, of which most have been restated. The two main objectives for IFRS are (1) increase comparability of financial reporting across countries and (2) increase the overall quality of financial reporting. The overarching objective is to increase the transparency of financial reporting (EY, 2017; Callao, Jarne, & Laínez, 2007).

3.1b IFRS 15

As of today, the IASB still publishes and continually improves IFRS standards. On May 28th of 2014 the IASB introduced the IFRS 15: Revenue From Contracts With Customers. IFRS 15 brings new requirements for the timing of revenue recognition. Technically, every aspect of IFRS 15 should be included in the financial statements, but in some circumstances not every disclosure requirement has to be fulfilled. For example, based on materiality considerations, a construction company might show unsatisfied disclosure requirements under IFRS 15. The five-step model (appendix 7.2) introduced by the IASB is used for the identification of contracts, the partition of contracts in performance obligations and to decide how revenue is recognized per performance obligation. Appendix 7.2 includes significant judgements construction companies have to make concerning the five-step model for revenue recognition.

Furthermore, IFRS 15 presents disclosure requirements (IFRS 15.110 up to and including 15.129) that are mandatory for companies that report under IFRS. These mandatory disclosure

broadlines require companies to disclose information regarding contracts with customers (IFRS 15.113 up to and including 15.122), significant judgements made in applying the standard (IFRS 15.123 up to and including 15.126) and any assets recognized from the costs to obtain or fulfil a contract with a customer (IFRS 15.127 and 15.128).

(10)

6

The quality of IFRS 15 disclosures differs between companies. This might be due to

differences in how companies give substance to the requirements set in IFRS 15 and due to more voluntary disclosure. There is no standard but only broadlines what the disclosure should include. Commonly companies are either benchmarking their disclosures to peers to enhance the quality of their disclosures or reflect on their own historical disclosure, and search for areas where they can improve.

To differentiate disclosure quality between companies, a distinction is made between

mandatory and voluntary disclosure. Mandatory disclosure is based on requirements by legislation and accounting rules (Einhorn, 2005). From an agency theory perspective, mandatory disclosure

legislation is necessary, as otherwise companies might choose to disclose nothing (Edelen, Evans & Kadlec, 2012). Voluntary disclosure focuses more on what information a company is providing in their financial statement disclosures on top of mandatory disclosed information. According to Cooke (1989) voluntary disclosures are those disclosures not mandatory required by a reporting standard.

Within IFRS 15, companies can choose how to present information within their disclosures, as long as they follow the mandatory requirements. Companies can disclose by a simple descriptive sentence, but also by extensive explanations or visualizing the information within a figure or table. Therewith they provide additional relevant information to users of the financial statements. This can be seen as voluntary disclosure (EY, 2017). A company might choose to voluntary disclose more information if they want to distinguish themselves from competitive companies in the same market (Kolsi, 2017). Furthermore, voluntary information disclosed could help reduce information asymmetry between the shareholders and the company (Watson, Shrives & Marston, 2002). The difference between mandatory and voluntary disclosure plays an important role within the self-composed IFRS 15 disclosure index (appendix 7.7) as the disclosure index is composed to identify relevant voluntary disclosures.

(11)

7 3.2 Relevant theories

This paper makes use of both agency- and signalling theory to clarify relations between the financial firm characteristics and IFRS 15 disclosure quality. Furthermore, this paper will make use of Kolb’s experiential learning theory (1984) to measure peer-learning effects and clarify the suggested positive relation between peer-learning effects and IFRS 15 disclosure quality.

3.2a Agency theory

Agency theory examines the relationship between business principals (shareholders/investors) and their agents (management) (Jensen & Meckling, 1976). An agency problem is a conflict of interest between principal and agent were one party is expected to act in the other party’s best interests instead of their own.

Furthermore, Jensen & Meckling (1976) assume that both agent and principal will try to maximize their own utility. This can result in agents that will act in their own interest instead of the principal’s best interests. Agents can act in their own interest due to the gap in possession of available information between them and the principals. So, there is information asymmetry and a conflict of interest between the agent and the principal from which an agency problem arises. To solve agency problems, the interests of the agent should be aligned with the interest of the principal. To mitigate agency problems, agents can provide a credible signal by expanding resources, while principals can expand to engage more in monitoring (Jensen & Meckling (1976).

Hooghiemstra, Hermes & Emanuels (2015) found that agency problems can be reduced by an increased level of disclosure. This implies that when the amount and quality of disclosures are improving, the information asymmetry between agent and principal is descending. Due to lower information asymmetry it is unlikely that agency problems will occur. Sarikhani & Saif (2017) found results similar to Hooghiemstra et al. (2015). They both conclude by stating that financial disclosures are often used to solve agency problems regarding information sharing between the company and their stakeholders.

Agency costs and disclosure quality can differ per company due to voluntary disclosure. Construction companies can choose to provide additional voluntary information regarding revenue recognition under IFRS 15 to prevent for agency problems arising from information asymmetry (Watson et al., 2002). Voluntary disclosures should provide a clear view to stakeholders about how the construction company will deal with the timing of revenue, how revenue is recognized and how revenue is disaggregated to performance obligations under IFRS 15. Voluntary disclosures should be good enough to satisfy the needs of various stakeholders in order to prevent for agency problems (Shehata, 2013).

(12)

8 3.2b Signalling theory

Signalling theory is the idea of an agent credibly conveying information about itself to a principal. Initially, signalling theory was about employees’ ability to signal their abilities to their employer. A credible signal by the employee, can help convince employers of the competency of their employees (Spence, 1973). The idea of signalling theory has been adapted to many other domains.

Within the domain of economics and business economics signalling is more formed around the idea of asymmetric information. To decrease the amount of asymmetric information signalling theory states that the company sends a signal to their stakeholders that reveals relevant information to them. Furthermore, Spence (2002) used his own developed signalling theory in an economical perspective. Spence (2002) stated that the quality of signals by the companies to their investors differ due to how much time, energy or money is spend on the signal. Also, investors can have different interpretations about the signal. Some investors might not trust the signal to consist honest information, while others would trust the signal.

Connelly, Certo, Ireland & Reutzel (2011) state that signalling theory can be used to describe behaviour when two parties have access to different information. The sender must choose whether and how to signal information to the receiver. The receiver must choose how to interpret the signal

communicated to them by the sender. The receiver will send their interpretations of the signal as feedback to the sender of the signal. According to Connelly et al. (2011) reducing information asymmetry between the sender and the receiver is a fundamental concern of signalling theory.

The aim of IFRS 15 is to increase transparency regarding the recognition of revenues. (EY, 2017). A transparent company should share as much as possible about their performance with their stakeholders. In order to be transparent, construction companies should in line with Watson et al. (2002) convince stakeholders that they are acting optimally when it comes to disclosures about revenue recognition under IFRS 15. In case of high revenues, the construction company should give stakeholders a signal on how this revenue has been achieved. Also, in case of low revenues, the construction company should give stakeholders a signal why the revenue is lower compared to average revenue. Also, the company should give the stakeholders an idea about how future revenues might look like.

Birjandi, Hakemi & Sadeghi (2015) used industry as one of their explanatory variables for voluntary disclosure, as characteristics specific to a certain industry can cause differences in policies of communication. Because of the complexity of contracts within the construction industry the construction industry might require more voluntary disclosure, to send out a signal of willingness to share information to their stakeholders. Voluntary disclosure helps to enhance the credibility of the management (Watson et al. 2002).

(13)

9 3.2c Kolb’s experiential learning theory

Kolb’s experiential learning theory helps to make learning from experience measurable. Kolb (1984, p.38) defines learning as: “The process whereby knowledge is created through the

transformation of experience”. Experiential learning theory can be represented by a learning cycle with four different stages. Learning within the experiential cycle is effective when passing all four stages (Kolb & Fry, 1975).

In the first stage, the learner is having a concrete new experience or encounters a new situation. This can also involve a reinterpretation of existing experience. In the second stage, the learner is reflecting on the concrete experience they have had. In the third stage, the learner actually learned from the concrete experience, and was able to understand it. In the last stage, the learner applies their new ideas to see what happens (Kolb, 1984). As the four stages form a cycle, the learner is likely to gain new experiences from applying their previous ideas.

We expect that in 2018 construction companies (but also auditors, controllers and financial advisors) were still looking for the best practical implementation of the IFRS 15 disclosure

requirements. In 2019, we expect to see a peer-learning effect due to the concrete experience (1) that companies had with IFRS 15. Due to this concrete experience, companies could have come up with new ideas how to implement the required disclosures. We expect companies to have looked at their own, and their peers’ disclosures from 2018 to come up with ideas how to voluntary improve

disclosure quality (2). When the companies actually come up with ideas (3) and implement them (4), we expect that the required disclosures might have changed, and disclosure quality thus improved.

Dani, Magro, Martinez, Lourenço & Branco (2018) found that disclosure quality increases when a company adopts IFRS. Lopes & Lopes (2019) add to this by stating that adopting to a new standard within IFRS also has positive effects on disclosure quality from year to year. Furthermore, Kvaal & Nobes (2012) found evidence suggesting that there is a learning process in the adoption of IFRS or a new standard within IFRS. They found that companies show better disclosures from year to year after adoption of a new standard.

We expect to find peer-learning results similar to Kvaal & Nobes (2012) as there are no companies with concrete IFRS 15 experiences in 2018, besides the companies that opted for an early adoption. The adoption of IFRS 15 on January 1st, 2018 can be seen as a normative isomorphic change for disclosures on revenue recognition. Normative isomorphic change is driven by pressures brought by professions. A professional standard is set, which becomes mandatory for companies to follow (Dimaggio & Powell, 1983).

(14)

10

Alongside this mandatory disclosure, companies often follow a voluntary approach to better inform their shareholders or potential investors. By disclosing more voluntary than other companies, companies can set a standard. Potential shareholders will expect from the under-disclosed companies to follow the level of voluntary disclosure set by the companies who disclosure more. Dimaggio & Powell (1983) names this incentive coercive isomorphism. Companies are forced to change due to external forces like their shareholders or potential investors.

To be ahead of coercive isomorphism, a lot of companies make use of mimetic isomorphism. This refers to the tendency of an organization to imitate what other organizations are doing better than they do (Dimaggio & Powell, 1983). By facing experiences with ideas imitated from other companies, companies also learn. We expect companies to learn from those experiences following Kolb’s

experiential learning theory (1984) and thus increase their disclosure quality under IFRS 15.

3.3 Hypotheses development

Based on the introduced theories, hypotheses will be elaborated to specify the two research questions. By performing further literature analysis on the financial firm characterises used within this paper and on peer-learning effects, hypotheses will be elaborated.

3.3a Profitability

When profitability increases managers will have multiple incentives to disclose more about the company (Omar & Simon, 2011). Based on agency theory, companies with a higher profitability often show a higher level of disclosure. Gallery et al. (2008) state that managers of profitable companies will use voluntary disclosures to support their current position and their compensation arrangements.

A company’s profitability can influence the amount of voluntary disclosure in multiple ways. From an agency theory perspective, Watson et al. (2002) shows similarities to Gallery et al. (2008) by stating that managers of profitable companies will use voluntary disclosures to secure their

compensation arrangements and their position within the company. Although, Watson et al. (2002) adds to that by stating from an agency theory perspective that more profitable companies will show more voluntary disclosure in order to avoid external regulation caused by a greaterpublic scrutiny for profitable companies. Voluntary disclosures are used by profitable companies as some kind self-regulation mechanism.

(15)

11

According to signalling theory, higher profitability ratios are often shared by companies using voluntary disclosures (Omar & Simon, 2011). Companies want to signal shareholders and potential investors about their good performance, by voluntary disclosing more information about their profitability. Watson et al. (2002) shows similarities to Omar & Simon (2011) by stating that companies aim to inform investors when they are performing well. So, Wang, Sewon & Claiborne (2008) summarizes from a signalling theory perspective that profitability is positively related with disclosure quality because of the managers incentive to signal good performance to investors.

Contrary, Omar & Simon (2011) argue that companies with a low profitability or losses are disclosing more information to shareholders, as a part of their responsibility to their shareholders. Abd-Elsalam (1999) adds to this that companies have the incentive to show higher disclosures for low profitability to maintain the reputation of the company. Meanwhile, Omar & Simon (2011) directly withdraw these arguments by stating that companies with a poor performance are more likely to engage in earnings management instead of disclosing poor performance.

Caylor (2010) adds that managers have the incentive to falsely inflate revenue and thus profitability to gain a higher bonus. Companies manage earnings to meet benchmarks set by financial analysts. An often-used method that both Akers, Giacomino & Gissel (2007) and Caylor (2010) mention is to manage earnings through the process of revenue recognition. Practically, Nagar and Tatiparti (2016) found evidence that within the construction industry, management is managing earnings, using for instance accruals for unbilled revenues.

So, now IFRS 15 is in use, managers should be able to better justify their revenue recognition policies through mandatory and further voluntary disclosure. As disclosures on identification of customer contracts, the identification of performance obligations and the allocation of transfer-prices to those performance obligations have become mandatory under IFRS 15, it will become harder for managers to manage earnings through the process of revenue recognition. voluntary disclosures regarding IFRS 15 might be beneficial for investors, as stakeholders better understand the performance of the company due to a better understanding of the underlying revenue recognition policies.

Concluded from both agency- and signalling theories we expect that highly profitable

companies have a bigger incentive to make use of more voluntary disclosure, thereby increasing their disclosure quality. We expect profitability to be one of the financial firm characteristics that will positively influence IFRS 15 disclosure quality. Thus, the first hypothesis is:

(16)

12 3.3b Liquidity

Liquidity represents a companies’ ability to meet short-term liabilities. Following the

traditional approach to liquidity, there are three key figures to measure liquidity: the (1) current ratio, (2) quick ratio and (3) net working capital (Emery & Cogger, 1982). What makes these ratios relevant, is that they are all composed of current liabilities and current assets in which current contract assets and current contract liabilities are captured (Deloitte, 2015). Within this paper, liquidity will be measured using net working capital because of the phenomenon of construction companies showing a negative or low net working capital in their year-reports (van Ommen & Karregat, 2020/1). The current- and quick ratio will be used for robustness checks. Net working capital is essential for the continuation of operational processes of construction companies, it should be sufficient to pay

suppliers, salaries and other operational costs on a short-time basis (van Ommen & Karregat, 2020/1).

Furthermore, net working capital is intertwined with bills still to be received from projects (debtors) and bills related to costs still to be payed for projects (creditors). Debtors and creditors are not only of significant influence for net working capital but are closely related to revenue and the recognition of revenue as most companies calculate revenue as income plus debtors minus costs minus creditors (Schilling, 1996). Debtors and creditors are also related to the contract assets and -liabilities under IFRS 15. When a construction company meets a certain milestone in a project, they are allowed to take revenue or form a contract asset. Contract assets can be formed from outstanding debtors recognized regarding the project. For contract liabilities this is vice versa, they can be formed from outstanding creditors recognized regarding the project (Deloitte, 2019).

Liquidity levels can differ from year to year as construction companies are working with projects (van Ommen & Karregat, 2020/2). Irregularity in the average debt collection period can cause significant differences in net working capital at year-end. It makes a lot of difference if a debtor pays their debt slightly before or after year-end (van Ommen & Karregat, 2020/2). By effectively managing liquidity, construction companies can be guaranteed to have sufficient liquidity throughout the year (van Ommen & Karregat, 2020/3).

Based on signalling theory, companies sharing their liquidity levels often show a higher disclosure quality, as they credibly want to convey the information about their liquidity level to their stakeholders (Sarikhani & Saif, 2017). Gallery et al. (2008) add to this, that in uncertain industries, like the construction industry, liquidity disclosures are highly wanted by stakeholders. Stakeholders want to make sure the company is able to meet its short-term obligations. To send out a good signal to stakeholders, high liquidity levels are higher disclosed. Furthermore, low liquidity levels are also higher disclosed, to justify low liquidity levels to their stakeholders (Naser, Al-Khatib & Karbhari, 2002).

(17)

13

Adb-Elsalam (1999) and Naser et al. (2002) support the relation between a companies’

liquidity level and a high level or quality of financial disclosure, from an agency theory perspective. Adb-Elsalam (1999) states that from an agency theory perspective companies with a low liquidity level, discloses more information to reduce potential agency conflicts with their shareholders and creditors. Naser et al. (2002) adds to that by stating that high liquidity levels are more extensively disclosed to reduce the information asymmetry between the stakeholders and the firm.

In conclusion, this paper expects to see a positive relation between a construction companies’ liquidity and IFRS 15 disclosure. Thus, this results in the following hypothesis:

3.3c Leverage

Leverage results from using borrowed capital as a source of funding. When a company is highly leveraged, it means that is has more debt than equity. Based on signalling theory, highly leveraged companies prefer to voluntary present private information in their disclosures to creditors and other stakeholders to increase transparency within their financial statements (Zarzeski, 1996).

Banks and creditors would be primarily interested in debt covenant related leverage ratios (Chen, Hung & Mazumdar, 1995). This implies that higher leveraged companies incur more monitoring costs. According to Sengupta (1998) monitoring costs decline by voluntary disclosing more information regarding leverage. Banks and creditors would not necessary be interested in IFRS 15 disclosures. Although, they might be interested in disclosures regarding significant changes in contract asset or liability balances (IFRS 15.118), risks regarding payment terms (IFRS 15.119b) and the recognition of unsatisfied performance obligations (IFRS 15.120). These disclosure requirements might unravel potential risks in the process to generate sufficient money to pay back debt obligations, which might have not been included in debt covenants.

Investors and shareholders are also interested in leverage ratios, as they want to reduce the amount of investment risk they take. Companies carrying lots of debt, are considered a risky investment (Temple, 2007). A company disclosing more voluntary information regarding revenue recognition under IFRS 15, might be interesting for investors and shareholders. From a signalling theory perspective, companies could justify high leverage ratios by disclosing higher revenues, which indicate that companies would be able to pay back debts (Zarzeski, 1996).

(18)

14

Furthermore, according to Malone et al. (1997) a low leverage could induce manager to orient their voluntary disclosures more towards investors and shareholder than to banks and creditors. Reasoning behind this is that a low leverage implies that banks and creditors are less likely to face potential risks regarding their outstanding debt not covered by debt covenants. On the other hand, a lower leverage level could be potentially interesting for investors and shareholders, as they face less investment risks investing in a low leveraged company.

From an agency theory perspective, companies make use of voluntary disclosure on their leverage to reduce agency costs (Sarikhani & Saif, 2017). Furthermore, Edelen et al. (2012) also state that agency costs can be reduced by voluntary disclosure to mitigate information asymmetry. Edelen et

al. (2012) adds to this that voluntary disclosure also lowers the cost of capital caused by information

asymmetry. This implies that companies showing higher leverage ratios, will increase their disclosure quality, and thus reduce their cost of capital and agency costs.

However, Eng & Mak (2003) state that the relation between leverage, and disclosure quality might be negative. Higher leverage ratios reduce voluntary disclosure by companies due to a reducing need for disclosure, as a higher leverage ratio restricts free cash flows. Omar & Simon (2011) adds to this by stating that leveraged companies share more private information to their creditors and other relevant stakeholders, which is not included in their statements. By sharing this private information, there would be less need to disclose the same information.

The same contradictory view is found within research focusing on the construction industry. While some research found a negative relation between a construction companies leverage and disclosure quality (Kolsi, 2017), others found a positive relation (Gulko, Hyde & Seppala, 2017). Furthermore, as shown by the debt-to-equity ratio of 7 in this studies sample, construction companies can be considered highly leveraged. This makes construction companies interesting to test the

aforementioned contradictory views on.

In conclusion, this paper does expect that there is an association between a construction companies leverage and IFRS 15 disclosure quality. The sign of this association is yet to be discovered as there are contradictory views within existing literature. Thus, the third hypothesis is:

(19)

15 3.3d Peer-learning

Following Kolb’s experiential learning theory (1984) and earlier research, we expect the implementation of IFRS 15 to follow the experiential learning cycle as defined by Kolb (1975). We expect to see changes in disclosure quality from year to year, in which this paper expects that companies will show an improved disclosure quality in 2019, the year after the adoption of IFRS 15, compared to 2018, the year of IFRS 15 adoption.

Tan, Wang & Welker (2011) found a positive relation between peer-learning and disclosure quality under IFRS. Salewski, Teuteberg & Zülch (2016) add to this by stating that IFRS adoption in an early stage lowers the company’s results and disclosure quality, while getting into a more mature phase within the application of IFRS the disclosure quality increases significantly due to learning effects of preparers and users of the financial statements. The aforementioned research specifically focused on the first-time adoption which might also be of relevance for the adoption of a new IFRS standard. Evidence of this was found by Lopes & Lopes (2019). They found that the adoption of a new IFRS standard has positive effects on disclosure quality over time due to learning effects.

Furthermore, Kravet & Muslu (2013) found that disclosures do not change continually from year to year. Disclosure quality does increase, but not continually. In accordance with Kravet & Muslu (2013) we expect that disclosures under IFRS 15 will also not change continually from year to year. We do expect to see changes in disclosure quality due to peer-learning in some financial statements though. Alongside statistical research, we will also make use of normative research to point out positive learning effects within the adoption and application of IFRS 15 that will increase the disclosure quality. Further details on this normative part can be found in the method section.

Construction companies faced significant changes regarding the disclosure requirements for the recognition of revenues under IFRS 15 due to the long-term contracts in the industry (Aarab et al., 2015). For example, performance obligations had to be identified in order to recognize revenues. As IFRS 15 has just been adopted, there is still a lack of best practices and benchmarks within the industry. Construction companies might still be searching for the best practical implementation of the IFRS 15 disclosure requirements. Therefore, we expect construction companies to alter some

disclosures, due to reviewing and learning from their own disclosures, and make use of mimetic isomorphism. Thus, the fourth, and last hypothesis is:

(20)

16 4. Research method

Within the research method section, we will discuss the measurement of variables and the collection of data. This paper will use quantitative data to perform statistical research. The sample consists of 37 European listed construction companies from 10 different countries. The database used for selecting the sample of companies is the ORBIS database. The firm-specific variables to determine disclosure quality will be manually collected from the 2018 and 2019 financial statements of the construction companies in the sample. Data will be collected by using a data-collection manual created and used by a cooperation of multiple researchers within a study-group.

From the ORBIS database, the study-group selected 37 construction companies using some criteria restrictions. First of all, companies have been selected from two specific subsections of level 2 NACE Rev. codes, which has been further disaggregated in four different level 4 NACE Rev. 2-codes. The distribution over the different NACE Rev. 2-codes can be found in appendix 7.3. From the selection left, we excluded companies with no recent financial data and public authorities, states or governments. Thereafter, we selected the 37 construction companies based on highest turnover. We defined construction companies within the sample as project developers. Project developers are companies that develop land through construction and seeks profit from the development of the land.

For eleven of the construction companies, we were unable to collect data, and thus they were not included in the sample. Three companies in the sample used a broken financial book year starting July 1st_{. Therefore, the companies did not applicate IFRS 15 in their 2018 financial statements.} Therefore, only the 2019 financial statements were included. The other eight companies did not publish their 2019 financial statement yet at the moment of data-collection, therefore only the 2018 financial statements were included. This sets the sample size of financial statements over two financial years to 63. The distribution of the sample can be found in appendix 7.3. The amount of data is

(21)

17 4.1 Dependent variable: IFRS 15 disclosure quality

Disclosure quality can be defined in terms of the information characteristics identified by the IASB. Information should be comparable, understandable, relevant and reliable (Botosan, 2004). Disclosure quality can be increased by providing more additional relevant voluntary information and disclosed information increases disclosure quality if it increases the readability of the financial statements (Aburaya, 2017). IFRS 15 disclosure quality will be measured using a self-composed disclosure index. Francis et al. (2008) and Botosan (1997) also created disclosure indexes to measure disclosure quality or voluntary disclosure.

Within this disclose indexes each element results in a specific score, by combining these scores and ranking them an ordinal measure of the companies’ disclosure quality can be obtained (Botosan, 1997). The disclosure index for this research will be created during meetings with the study-group and is expected to show similarities with the disclosure indexes by Francis et al. (2008) and Botosan (1997) in terms of scoring the elements and combining those scores.

The disclosure index used within this paper will focus around the three broad themes set within the IFRS 15 disclosure requirements: (1) contracts with customers, (2) significant judgements in applying the standard and (3) costs to fulfil a contract. These themes represent all aspects present in the IFRS 15 disclosure requirements. Disclosure quality is determined by judging the readability of disclosures in two aspects.First, the qualitative aspect, which gives a signal whether disclosed information is useful. Secondly, the quantitative aspect, which gives a signal how disclosures are presented. The more quantitative, the more structured disclosures are presented. A company earns two points if they voluntary disclose on a requirement, one point if they disclose generic information, but fulfil the requirement, and zero points if the requirement is not fulfilled. The self-composed disclosure index and guidelines can be found in appendix 7.7.

The disclosure index consists out of 52 points that can be scored within the three broad themes. After computing all scores, the scores will be ranked to ensure that no item is more important than any other item. Lastly, after ranking the scores each company got a percentage score. A higher percentage score indicates a more complete disclosure, and thus a higher disclosure quality.

Every member of the study-group collected an equal amount of data, randomly selected to them. This might cause differentiations in judgement while scoring the aspects within the disclosure index. To ensure high quality data, the collected data is evaluated and discussed during various meetings with the study-group. 19 of the 63 year-reports have been controlled by study-group

members twice, to control for the risk of differences in judgement. Significant differences in allocated scores have been discussed. Based on the outcome of the discussions other scores were subsequently re-evaluated.

(22)

18 4.2 Independent variable: Profitability

Profit is defined by Nickels, McHugh & McHugh (1997) as the money a company earns above and beyond what it spends for salaries and other costs. Profitability is defined by Amah & Ahiauzu (2011) as a state of producing a profit or the degree to which a company is profitable. Profitability can be seen as the primary goal for all profit-orientated companies. Without profit they will not survive in the market. To measure profitability, we will use EBIT-margin, which is calculated by EBIT divided by total revenue following Pazarskis, Alexandrakis, Notopoulos & Kydros (2011). Thereby, EBIT is calculated as net income plus income taxes and interest expenses. According to Pazarskis et al. (2011) and Berger & Hann (2007) EBIT-margin is one of the most accurate measures of profitability. All relevant data to compose EBIT-margin has been abstracted from the year-reports, no data was missing.

4.3 Independent variable: Liquidity

Liquidity is defined by Sukiennik (2012, p.339) as: “an enterprise’s ability to transform assets into cash in the shortest possible time and without the loss of value”. As mentioned before, this paper will make use of net working capital to measure liquidity as van Ommen & Karregat (2020/1) stated that there is a phenomenon within the construction industry in which they tend to show a negative or low net working capital. Net working capital is measured by the difference between current assets and current liabilities following Tu and Nguyen (2014). To check for alternative specifications of liquidity and their potential influence on the outcomes of the model, a robustness check will be performed to control for the quick- and current ratio (appendix 7.6). All relevant data to compose net working capital has been abstracted from the year-reports, no data was missing.

4.4 Independent variable: Leverage

Leverage is defined as an investment strategy of using borrowed money to increase the potential return of an investment. Leverage is based on the book-value of debt and assets following Jermias (2008) because, when discussing financial leverage, financial managers tend to think in terms of book-value rather than market-value ratios. Within this paper leverage will be measured using the debt-to-EBITDA ratio. The debt-to-EBITDA ratio is composed following Izzi, Oricchio & Vitale (2012) as net-debt divided by EBITDA. In this formula net-debt is measured as current- and non-current liabilities minus cash and cash equivalents. EBITDA is measured as operating income plus depreciation and amortization. The debt-to-EBITDA ratio is used quite regularly within the

construction industry to indicate whether the revenue a company makes is sufficient to be able to pay back their debts. All relevant data to compose the debt-to-EBITDA ratio has been abstracted from the year-reports, no data was missing.

(23)

19 4.5 Independent variable: Peer-learning

Peer-learning is defined by Topping (2005, p.631) as: “The acquisition of knowledge and skill through active helping and supporting among status equals or matched companions”. Within this paper, peer-learning is proxied by experiential learning. Experiential learning is defined as: “a knowledge creation process through which new experiences are integrated into prior experiences and transformed into relevant, durable and retrievable knowledge suitable for use in the learners’

environment” (Kolb, 1984). Experiential learning effects will be measured by comparing the 2018 disclosure quality score with the 2019 disclosure quality score, with the expectation that the disclosure quality score increased in 2019 compared to 2018.

Besides measuring the effect of experiential learning statistically, we will also measure peer-learning effects in a normative manner. This normative measure explicitly looks deeper into what has been positively changed in the IFRS 15 disclosures of 2019 compared to 2018. In appendix 7.8 normative changes between 2018 and 2019 are shown for certain interpretations on the IFRS 15 disclosure requirements by construction companies.

4.6 Control variables: Firm size and Country

Two control variables will be introduced. First of all, Alfaro, Asis, Chari & Panniza (2019) state that firm size influences the decision usefulness of a financial report. Bigger companies report and disclose more relevant information than smaller companies. Agyei-Mensah (2013) adds to this by stating that firm size is positively associated with disclosure quality using IFRS. To control for the unknown effect of firm size we will use the log of total assets to measure firm size, equal to research by Alfaro et al. (2019). Firm size is measured differently in various industries (Berg & Stylianou, 2009). The log of total assets is taken to normalize the data and can be seen as a good measure for firm size within the construction industry.

Secondly, Kvaal & Nobes (2012) showed that IFRS disclosure quality varies among countries in Europe. We will not include the country variable directly in our models, as we consider the amount of construction companies per country too low to significantly say something about the influence of country. As shown in appendix 7.3, the 37 construction companies in the sample are distributed over 10 different countries. Some countries include one or two companies, which might harm achieved results. Therefore, to control for the unknown effect of country on IFRS 15 disclosure quality we will perform a robustness check for those countries that involve at least three companies. Results can be found in appendix 7.6.

(24)

20 4.7 Data adjustments

To achieve higher quality data, some adjustments have been made to the collected data. Firstly, the sample consists out of construction companies from various European countries. The currency used to calculate the financials within this paper is the Euro. Most companies in the sample use Euro as the currency in their financial statements. Companies from both Sweden and Great Britain do not use the Euro, but respectively the Swedish krona and the Pound sterling. Therefore, financials from Sweden and Great Britain have been restated to Euro using the year-end exchange rates in appendix 7.4.

Secondly, the maximum score on the disclosure index is 52 points if all criteria are met. Although, this maximum score can be lower in case a criteria in the disclosure index has been rated as ‘not applicable’. Some criteria can be rated as ‘not applicable’ due to their nature. For instance, IFRS 15.118 requires a company to provide an explanation of significant changes in the contract assets or -liabilities balances during the reporting period. We defined a significant change as >10% of last year. When the change is <10% for both contract assets and -liabilities, we rate the criteria as ‘not

applicable’. Also, IFRS 15.129 requires the company to disclose if they elect to use the practical expedient in 15.63 (existence of a significant financing component) or 15.94 (incremental costs of obtaining a contract. When the company does not use either practical expedients, the criteria has been rated as ‘not applicable’. The maximum and achieved score for construction companies in the sample can be found in appendix 7.5.

Lastly, Independent variables have been calculated from the collected data. Furthermore, to control for outliers, all independent variables, have been 95% winsorized at the top and bottom 5%. This means that extreme upper and lower values have been replaced with the 5th_{or 95}th_percentile. Also, the log of total assets has been taken to make data conform to normality.

(25)

21 4.8 Models

To measure the main research question concerning the effects of specific financial firm characteristics on IFRS 15 disclosure quality the model below will be used. This formula will be implemented for both financial years 2018 and 2019.

To measure the additional research question concerning peer-learning effect on IFRS 15 disclosure quality we aim to show both statistical- and normative evidence. Statistically, we will make use of Wilcoxon signed-rank tests to test for significant differences between disclosure quality in 2018 and 2019. Normatively, we will show best practices of differences between the disclosure quality in 2018 and 2019 for a company.

The underlying variable and proxy of the abbreviations in the model can be found in table 1.

Abbreviation Variable Proxy

DISC Disclosure Score IFRS 15 disclosure index

PROF Profitability EBIT-margin

LIQ Liquidity Net Working Capital

LEV Leverage Debt-to-EBITDA ratio

SIZE Firm size Log of total assets

(26)

22 5. Results

5.1 Descriptive statistics

The descriptive statistics, as shown in table 2, give an indication of the sample size (N), mean, standard deviation (Std. Dev) and the minimum and maximum of each variable. To provide more information to the total descriptive statistics, a distinction has been made between the financial years 2018 and 2019. The disclosure score represents the obtained disclosure score as a percentage of the total score a company was able to score on the IFRS 15 disclosure index. The lowest score was 0.26 (26%) whereas the highest score was 0.75 (75%), this indicates a large gap of 0.49 (49%) between the lowest- and highest achieved disclosure score. This large gap is also shown by the standard deviation of 0.106 (11%). The average disclosure score, as indicated by the mean, is 0.508 (51%). In 2018 the average score equals 0.502 (50%). Compared with the average score in 2019 of 0.514 (51%), there is minimal improvement in disclosure quality showed within the sample.

Profitability, as measured by the EBIT-margin shows a minimum of -0.029 and a maximum of 0.266. This indicates that within the sample, both loss-making and profitable companies are included. On average, the companies within the sample are profitable (0.071).

The minimum liquidity, as measured by net working capital, is -2,740 million and the maximum liquidity is 4,520 million. This indicates that there are companies within the sample showing a low liquidity, but there are also companies showing a high liquidity. On average, liquidity is 612 million. This indicates that on average, companies within the sample have a slightly positive net working capital.

Leverage, as measured by the debt-to-EBITDA ratio, shows a minimum of -38.445 and a maximum of 31.445. This indicates that the company with a ratio of -38.445 has almost no debt, and a lot of cash, while the company with a ratio of 31.445 has a lot of debt, which might be too heavy. On average the debt-to-EBTIDA ratio for the sample is 7.189. This indicates that it would take companies within the sample on average 7 years to pay back debts, if net-debt and EBITDA stay constant.

The minimum firm size, as measured by the log of total assets, is 20.470, while the maximum is 24.396. On average firm size is 22.413. This indicates that there are no huge differences in firm size within the sample.

(27)

23

Variable Year N Mean Std. Dev Minimum Maximum DISC (in %) 2018 34 .502 .102 .30 .70 2019 29 .514 .111 .26 .75 Total 63 .508 .106 .26 .75 PROF 2018 34 .061 .064 -.029 .205 2019 29 .083 .087 -.015 .266 Total 63 .071 .076 -.029 .266

LIQ (in millions)

2018 34 438 1300 -2430 3460 2019 29 816 1810 -2740 4520 Total 63 612 1550 -2740 4520 LEV 2018 34 6.684 13.936 -38.445 31.445 2019 29 7.780 5.319 0.371 17.248 Total 63 7.189 10.791 -38.445 31.445 SIZE 2018 34 22.434 1.102 20.470 24.392 2019 29 22.388 1.087 20.616 24.396 Total 63 22.413 1.087 20.470 24.396 Table 2: Descriptive statistics

5.2 Correlation model

To check for correlations between variables, the Pearson correlation matrix is used. Results, divided by financial year, can be found in table 3. In 2018, only one correlation is found, there is a significant negative correlation between profitability and disclosure quality (-0.618**). In 2019, more correlations are found. Equal to 2018, there is a significant negative correlation between profitability and disclosure quality (-0.601**). Secondly, the is a significant positive correlation between leverage and disclosure quality (0.600**) Also, there is a significant positive correlation between liquidity and profitability (0.597**). Lastly, there are significant negative correlations between leverage and profitability (-0.565**) and leverage and liquidity (-0.417*).

Furthermore, variables within the model are tested for multicollinearity. Results indicate that there is no sign of multicollinearity. The average VIF-value in 2018 is 1.07 with all individual variable values around 1. For 2019, all individual values are between 1 and 2, with an average VIF-value of 1.61.

(28)

24

Correlations 2018 (N = 34)

DISC PROF LIQ LEV SIZE DISC 1 PROF -0.618** (0.000) 1 LIQ 0.024 (0.889) 0.156 (0.379) 1 LEV 0.080 (0.654) -0.012 (0.947) -0.022 (0.901) 1 SIZE -0.102 (0.565) 0.214 (0.225) -0.078 (0.663) 0.189 (0.284) 1

*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

Correlations 2019 (N = 29)

DISC PROF LIQ LEV SIZE DISC 1 PROF -0.601** (0.001) 1 LIQ -0.208 (0.279) 0.597** (0.001) 1 LEV 0.600** (0.001) -0.565** (0.001) -0.417* (0.025) 1 SIZE -0.061 (0.752) 0.150 (0.437) -0.068 (0.726) 0.110 (0.570) 1

*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed). Table 3: Pearson correlation matrix

(29)

25 5.3 Regression analysis

To determine relationships between variables, linear regression analyses are performed for both 2018 and 2019. Results can be found in table 4. The second column in table 4 represents the expected sign (Exp. Sign) of the relations. For leverage (LEV) there is no sign given (X), as the sign has yet to be determined. The first model (1) includes the control variable, and none of the

independent variables, to indicate whether there is a relationship between the control variable and disclosure quality. The models two (2) up to and including four (4) give an indication of the

relationship between disclosure quality and each independent variable, including the control variable. These three models are used to test H1 up to and including H3. Model five (5) includes all

independent variables and the control variable, to give an indication of the relation between all variables and disclosure quality.

The first regression model (1) analysis the effect of firm size on disclosure quality, in which firm size is measured by the log of the total assets. Results for 2018 show a negative coefficient (-0.009) with no significant relation (0.564). Results for 2019 are similarly showing a negative coefficient (-0.006) with no significant relation (0.752). The negative coefficient for both financial years is not equal to the expected positive sign, although the negative coefficients are not significant.

The second regression model (2) analysis the effect of profitability on disclosure quality. The effect of firm size as a control variable is included in this model. Results for 2018 show a negative coefficient (-0.985**) with a significant relation (0.000) for profitability. Firm size shows a positive coefficient (0.003), with no significant relation (0.831). Results for 2019 are similarly showing a negative coefficient (-0.776**) with a significant relation (0.001) for profitability. Firm size is showing a positive coefficient (0.003), with no significant relation (0.854). The negative significant profitability coefficient in both financial years is not equal to the expected positive sign. A relation between profitability and disclosure quality is found, but vice versa. Therefore, H1 is rejected, as H1 indicated a positive relation.

The third regression model (3) analysis the effect of liquidity on disclosure quality. The effect of firm size as a control variable is included in this model. Results for 2018 show a positive coefficient (0.000) with no significant relation (0.925) for liquidity. Firm size is showing a negative coefficient (-0.009) with no significant relation (0.577). Results for 2019 are similarly showing a positive

coefficient (0.000) with no significant relation (0.277) for liquidity. Firm size is showing a negative coefficient (-0.008) with no significant relation (0.695). The positive coefficient for both financial years is equal to the expected positive sign, although the relation between liquidity and firm size is not significant. Therefore, H2 is rejected.

(30)

26

The fourth regression model (4) analysis the effect of leverage on disclosure quality. The effect of firm size as a control variable is included in this model. Results for 2018 show a positive coefficient (0.001) with no significant relation (0.574) for leverage. Firm size is showing a negative coefficient (-0.011) with no significant relation (0.506). Results for 2019 differ. They show a positive coefficient (0.013**), but with a significant relation (0.001) for leverage. Firm size is equally to 2018 showing a negative coefficient (-0.013) with no significant relation (0.415) in 2019. So, there is a positive significant relation between leverage and disclosure quality in 2019. Therefore, H3 is

accepted, despite the fact that the relation is insignificant in 2018. This might be due to the presence of outliers for leverage in the dataset in 2018. Even though we tried to control for outliers, 2018 still does not show a clear pattern in the underlying data for leverage. Furthermore, validity as measured by R2 for leverage in 2019 (0.377) is much larger than in 2018 (0.021). This implies that leverage in 2019 carries much more explanatory value than in 2018.

(31)

27

Regression models 2018 (N = 34) Exp. Sign (1) (2) (3) (4) (5) PROF (+) -0.985** (0.000) -0.998** (0.000) LIQ (+) 0.000 (0.925) 0.000 (0.384) LEV X 0.001 (0.574) 0.000 (0.641) SIZE (+) -0.009 (0.564) 0.003 (0.831) -0.009 (0.577) -0.011 (0.506) 0.003 (0.832) Constant 0.714* (0.058) 0.499 (0.102) 0.711* (0.065) 0.749* (0.053) 0.490 (0.122) R2 _0.011 _0.383 _0.011 _0.021 _0.404 Adjusted R2 _-0.020 _0.343 _-0.053 _-0.042 _0.322 *. significant at the 0.10 level (2-tailed).

**. significant at the 0.05 level (2-tailed).

Regression models 2019 (N = 29) Exp. Sign (1) (2) (3) (4) (5) PROF (+) -0.776** (0.001) -0.684** (0.017) LIQ (+) 0.000 (0.277) 0.000 (0.131) LEV X 0.013** (0.001) 0.009** (0.028) SIZE (+) -0.006 (0.752) 0.003 (0.854) -0.008 (0.695) -0.013 (0.415) -0.001 (0.958) Constant 0.655 (0.149) 0.511 (0.170) 0.699 (0.125) 0.710* (0.057) 0.507 (0.146) R2 _0.004 _0.362 _0.049 _0.377 _0.513 Adjusted R2 _-0.033 _0.313 _-0.024 _0.329 _0.432 *. significant at the 0.10 level (2-tailed).

**. significant at the 0.05 level (2-tailed). Table 4: Regression analysis results

(32)

28 5.4 Validity analysis

The bottom three rows in table 4 show the constant (α0), R2_{and adjusted R}2_{. The fifth model} (5) analysis the effect of all independent variables and the control variable on disclosure quality. Therefore, this can be considered the best model to interpret validity. Validity indicates how accurate a method measure something. As found in table 4, R2_{for model 1 is 0.011 in 2018. In comparison, R}2 for model 5 is 0.404 in 2018. The increment of R2_{in model 5 indicate that the explanatory value of the} model increases, due to adding the independent variables.

Results in 2019 are similar as in 2019 model 1 shows a R2_{of 0.004. By adding the} independent variables in model 5, R2_{increased to 0.513. The increment of R}2_{in model 5 is even} higher than in 2018. This indicates that the explanatory value of the model increases even more in 2019, due to adding the independent variables. Overall, based on R2_{the explanatory power of model 5} can be considered as moderate, there might still be other variables determining disclosure quality.

Subjectivity can harm the validity and reliability of results. Within this paper the dependent variable, IFRS 15 disclosure quality is subject to subjectivity. The data for this paper is collected by multiple members of the study-group. Due to this, subjectivity might rise, as criteria within the disclosure index can be judged differently by members of the study-group. To control for subjectivity, the Cronbach’s alpha has been calculated for the data underlying to the disclosure score. Underlying data includes either a score of 0, 1 or 2 for IFRS 15.113a up to and including IFRS 15.129. As shown in table 5, the Cronbach’s alpha for this paper is 0.761. Therefore, based on Cortina (1993) stating that a Cronbach’s alpha of 0.7 or higher is acceptable, the underlying data for DISC is considered valid and reliable.

(33)

29

Item Cronbach’s alpha Item Cronbach’s alpha

IFRS 15.113a 0.742 IFRS 15.123a 0.765

IFRS 15.113b 0.761 IFRS 15.123b 0.753

IFRS 15.114 / IFRS 15.115 0.737 IFRS 15.124 0.749

IFRS 15.116a1 0.765 IFRS 15.126a 0.758

IFRS 15.116a2 0.760 IFRS 15.126b 0.753

IFRS 15.116a3 0.752 IFRS 15.126c 0.759

IFRS 15.116b 0.757 IFRS 15.127a 0.755

IFRS 15.116c 0.744 IFRS 15.127b 0.745

IFRS 15.117 0.760 IFRS 15.128a 0.760

IFRS 15.118 0.752 IFRS 15.128B 0.755

IFRS 15.119a 0.740 IFRS 15.129 0.772

IFRS 15.119b 0.744

IFRS 15.119c 0.765 TOTAL 0.761

IFRS 15.120a 0.746 IFRS 15.120b 0.741 Table 5: Decomposition of Cronbach's alpha

Firm Characteristics, Peer-learning and Disclosure Quality under IFRS 15: A European construction-industry based research.