Bankruptcy risk, its financial determinants and reporting delays: Do managers have anything to hide?

Lukason, Oliver; Camacho-Miñano, María-del-Mar Article Bankruptcy risk, its financial determinants and reporting delays: Do managers have anything to hide? Risks Provided in Cooperation with: MDPI – Multidisciplinary Digital Publishing Institute, Basel Suggested Citation: Lukason, Oliver; Camacho-Miñano, María-del-Mar (2019) : Bankruptcy risk, its financial determinants and reporting delays: Do managers have anything to hide?, Risks, ISSN 2227-9091, MDPI, Basel, Vol. 7, Iss. 3, pp. 1-15, https://doi.org/10.3390/risks7030077 This Version is available at: https://hdl.handle.net/10419/257915 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ risks Article Bankruptcy Risk, Its Financial Determinants and Reporting Delays: Do Managers Have Anything to Hide? Oliver Lukason 1,* and María-del-Mar Camacho-Miñano 2 1Faculty of Economics and Business Administration, University of Tartu, Tartu 50409, Estonia 2Accounting and Finance Department, Complutense University of Madrid, Madrid 28223, Spain *Correspondence: oliver[email protected] Received: 22 May 2019; Accepted: 3 July 2019; Published: 7 July 2019   Abstract: The aim of this study was to investigate whether firms’ reporting delays are interconnected with bankruptcy risk and its financial determinants. This study was based on 698,189 firm-year observations from Estonia. Annual report submission delay, either in a binary or ordinal form, was used as the dependent variable, while bankruptcy risk based on an international model or the financial ratios determining it were the independent variables. The findings indicated that firms with lower values of liquidity and annual and accumulated profitability were more likely to delay the submission of an annual report over the legal deadline. In turn, firm leverage was not interconnected with reporting delays. In addition, firms with a higher risk of bankruptcy were more likely to delay the submission of their annual reports. Firms with different ages, sizes and industries varied in respect to the obtained results. Different stakeholders should be aware that when reporting delays occur, these can be conditioned by higher bankruptcy risk or poor performance, and thus, for instance, crediting such firms should be treated with caution. State institutions controlling timely submission should take strict(er) measures in cases of firms delaying for a lengthy period. Keywords: bankruptcy risk; financial ratios; reporting delays; private firms; SMEs; Estonia 1. Introduction Bankruptcy risk and information disclosure by companies are two “hot topics” in management, business and accounting literature due to their implications for stakeholders’ decisions. Bankruptcy is a critical issue for firms due to its negative social and financial consequences (Wu 2010). Risk information might not be fully disclosed in or understood from financial statements and, consequently, stakeholders’ decisions may be inappropriate (Linsley and Shrives 2006). Moreover, prior literature shows that financially distressed companies could have incentives to hide the reasons why they are not performing well (Singhvi and Desai 1971;Whittred and Zimmer 1984). Thus, the interconnection between financial distress (risk) and information disclosure is an ongoing important and promising area of research. Although academia has been working on developing a bankruptcy theory for more than 60 years, no consensus yet exists (du Jardin 2015;Bauweraerts 2016), even in the EU’s legal approach (Boon 2018). However, most researchers in the field agree that the main determinants of bankruptcy risk are liquidity, profitability and leverage, and these determinants are most frequently used in bankruptcy prediction models (Balcaen and Ooghe 2006;Tasc ó n Fern á ndez and Guti é rrez 2012;Lukason and Laitinen 2019). Thus, these determinants can condition the most when financial information is disclosed. In general, there are two opposite explanations for the information disclosed. From the point of view of external users, more information is demanded, although the type of information could be Risks 2019,7, 77; doi:10.3390/risks7030077 www.mdpi.com/journal/risks Risks 2019,7, 77 2 of 15 different depending on the stakeholders’ informational needs. Perhaps, certain information can be favourable for some users but detrimental to others (Inchausti 1997). From the point of view of internal users, due to the presence of competitive reasons, managers may be reluctant to disclose certain kinds of information. This conflict of interests among users related to the demand for financial reporting, namely, information asymmetries, is part of the agency theory (Healy and Palepu 2001). The agency problem arises when investors and managers have non-aligned interests and relevant information is not disclosed (see Jensen and Meckling 1976). An example of these different interests is a bankruptcy situation where creditors and debtors both want to maximize their gains. In these circumstances, information delays could be a trade-offamong a firm’s stakeholders. Previous studies provide evidence that a significant number of bankrupt firms incurred delays in releasing their financial statements in the final year before bankruptcy (Ohlson 1980;Lawrence 1983;Lukason 2013). Moreover, delays in publishing financial statements were significant predictors of failure (Ahmed and Courtis 1999) and potential indicators of financial distress (Peel et al. 1986;Peel and Peel 1988). There may be several reasons causing reporting delays. According to Altman et al. (2010, p. 17), the late filing of accounts may be an intentional managerial decision in the case of companies which are facing financial difficulties, because they do not want to publish unfavourable information, “a by-product of the financial difficulties a firm faces” or even the result of a disagreement between auditors and directors due to the “true and fair view” of firms. The longer a company takes to file accounts after year-end, the more probable that it is encountering difficulties (Altman et al. 2010). Delays in filing accounts have direct negative consequences for privately held firms, as late filing is associated with lower credit ratings (Clatworthy and Peel 2016), increase of information asymmetries (Owusu-Ansah and Leventis 2006) and, in general, is a way to decrease the quality of stakeholders’ decision-making (Singhvi and Desai 1971). Thus, governments usually force minimum disclosure requirements for all companies through the use of penalties if those disclosure requirements are not met. The aim of this study was to investigate whether firms’ reporting delays are interconnected with bankruptcy risk and its financial determinants. As financial determinants, financial ratios portraying liquidity, profitability and leverage were implemented. We contextualized our research on a specific developed European country, Estonia, as factual whole population data about reporting delays were available. The main contributions of this paper are the following. First, in line with Healy and Palepu’s (2001, p. 411) research suggestions, we answered the question: which financial factors affect management’s disclosure choices in an interesting institutional setting—Estonia. Second, we conducted a comprehensive study of privately held firms’ financial reporting timeliness. Very little is known about information disclosure in this firm group (Luypaert et al. 2016), yet private firms constitute a major share of the European production capacity (Eurostat 2019). Moreover, private financial reporting differs significantly from listed companies’ reporting due to the differences in their financing sources and the availability of information (Hopwood 2000;Berger and Udell 2006). The next section discusses relevant prior literature and sets the research hypotheses. Section 3 describes the data and methods, while our empirical results with a discussion are presented in Section 4. The paper concludes in Section 5with a provision of limitations, practical implications and suggestions for further research. 2. Literature Review and Development of Hypotheses Disclosure regulation is, in many countries, motivated by the concern of reducing informational gaps between informed and uninformed stakeholders (Healy and Palepu 2001). Accounting disclosure is essential for decision-making purposes and could condition incorrect decisions and predictions for firms’ stakeholders (Botosan 1997). Agency theory considers accounting policy choice and disclosure as part of the contracting process between firms and its stakeholders (Smith and Watts 1992;Skinner 1993). Risks 2019,7, 77 3 of 15 Therefore, international accounting regulation highlights timeliness (i.e., non-delayed presentation) as one of the main characteristics of financial information (IASB 2010). Different incentives could influence the timing of the financial statements’ filing, such as the delay of unfavourable information that could affect the attainment of external financing. Some firms may delay the reporting of losses because they need to have “a good look” for obtaining more external financing (Givoly and Palmon 1982). In many cases, the delay itself becomes a signal of possible poor performance (Elliott 1982). It is a fact that the pre-insolvency non-submission of annual reports is common (Lukason 2013), and thus, financially distressed firms might not want to show their poor situation. As a “vicious circle”, they are probably not going to receive more financial support if they obtained losses in previous accounting periods. Indeed, financially distressed companies are more likely to distort their financial communication. This idea is based on the agency perspective through asymmetric information, in line with Darrough and Stoughton’s (1990) theory of selective disclosure in the context of financial distress. There is strong evidence that firms with losses are less likely to disclose information than other firms (Ajinkya et al. 2005). According to the obfuscation hypothesis, managers may try to obfuscate failures (Courtis 1998;Clatworthy and Jones 2003) and the reporting delay may be a strategy to hide bad news. The latter is illustrated by multiple studies highlighting that managers are prone to release good results as soon as possible but tend to delay losses as much as possible (Kross 1982;Haw et al. 2000). Thus, relying on selective disclosure and obfuscation theories, we considered whether bankruptcy risk and its main determinants were interconnected with reporting delays and outlined the findings from the relevant literature. As an important cause of reporting delays is corporate financial distress (Impink et al. 2012), the determinants of this troubled financial situation should be taken into account. Indeed, the huge number of bankrupt firms during the worldwide recession in 2009 caused by the global credit crunch has rekindled studies aimed at detecting the first warning signs of financial distress ( Mselmi et al. 2017 ; Geambasu et al. 2013 ). Concretely, studies suggest that ratios measuring profitability, liquidity and leverage are the most relevant in predicting bankruptcy (Scott 1981; Laitinen 1991 ; Balcaen and Ooghe 2006 ; du Jardin 2015 ;Altman et al. 2017;Lukason and Laitinen 2019). Consequently, we propose the following hypotheses considering liquidity, annual and accumulated profitability, leverage and bankruptcy risk in interconnection with reporting delays. 2.1. Liquidity and Reporting Delays According to prior literature, bankrupt firms usually have liquidity problems (Lukason and Laitinen 2019). It is considered a technical insolvency (Laitinen and Suvas 2016), meaning that businesses cannot meet their current financial obligations. Consequently, the liquidity of the firm affects bankruptcy probability (Scott 1981) and empirical studies (e.g., Bunn and Redwood 2003) indicate that higher liquidity reduces the probability of failure. Therefore, firms with liquidity problems may not want to reveal their distressed situation, and deriving from this, the first hypothesis is as follows: Hypothesis 1 (H1). The higher the liquidity, the lower the likelihood of delaying annual reporting over the submission date set in law. 2.2. Profitability and Reporting Delays Common determinants of financial distress are annual and accumulated (aggregated) profitability (Altman et al. 2017;Lukason and Laitinen 2019). Keasey and McGuinness (1990) suggest that the profitability ratio is a significant indicator of failure for a number of years prior to the date of failure. Also, profitable companies have been noted to be less likely to delay annual accounts ( Dogan et al. 2007 ). The annual profitability is relevant for bankruptcy prediction models as it “captures the capacity of the firm to manage its assets efficiently and generate enough funds to meet its financial obligations” (Pindado et al. 2008, p. 998). In theory, an increase in profitability should reduce the likelihood of financial distress and failure (Chiaramonte and Casu 2017). Prior literature on this indicator Risks 2019,7, 77 4 of 15 related to timeliness highlights that late reporting exhibits less profitability (Hashim et al. 2013; Luypaert et al. 2016 ). Submission delays are often related to losses for the period, that is, companies experiencing losses are expected to have longer reporting delays than the ones reporting profits ( Givoly and Palmon 1982 ;Ismail and Chandler 2004), although the results are not fully conclusive ( Bonsón and Borrero 2011 ). Thus, we expect the following interconnection between annual and accumulated profitability and reporting delays: Hypothesis 2 (H2). The higher the accumulated profitability, the lower the likelihood of delaying annual reporting over the submission date set in law. Hypothesis 3 (H3). The higher the annual profitability, the lower the likelihood of delaying annual reporting over the submission date set in law. 2.3. Leverage and Reporting Delays Another significant indicator of business failure is leverage, which has been noted to have predictive abilities even four or five years prior to failure (Merwin 1942). For SMEs (i.e., micro-, small- and medium-sized firms), private debt provided by trade creditors and banks is their main source of finance, because they do not have access to capital markets. Thus, those firms rely, almost absolutely, on private lenders who also require timely information for their decision making (Peek et al. 2010). This means that borrowing is, for SMEs, the main reason for preparing their financial statements, thus creating a demand for timely financial statements as a way to mitigate agency problems with debt ( Allee and Yohn 2009 ). However, the results in the literature are inconclusive on this topic. Some studies do not find evidence that heavier reliance on external debt financing will lead to more timely disclosure of information (Owusu-Ansah 2000;Hashim et al. 2013;Luypaert et al. 2016); however, contrary findings are also present (Carey and Clarke 2001). Empirically, prior research shows that leverage acts more as a proxy for distress and tends to be negatively related to timeliness (e.g., Impink et al. 2012 ). The monitoring cost theory postulates that highly geared companies are timely reporters (Owusu-Ansah and Leventis 2006). In turn, highly geared companies are expected to delay the presentation of their reports because the audit procedure of debt capital is more time consuming than that of equity capital (Carslaw and Kaplan 1991;Owusu-Ansah 2000;Conover et al. 2008). Thus, due to the controversial results in previous studies, we present the following hypothesis: Hypothesis 4 (H4). The level of leverage has no effect on the likelihood of delaying annual reporting over the submission date set in law. 2.4. Bankruptcy Risk and Reporting Delays The fact that financially distressed firms could delay reports has been empirically proven (e.g., Altman et al. 2010). Still, the majority of relevant studies used reporting delay as anindependent variable, therefore focusing on how long delay period should be used to best predict future failure. For predictive purposes, such a setting is reasonable, as many distressed firms do not report at all (Lukason 2013). When to consider the cause of delaying, then at first the distress situation occurs and consequently the decision to delay is taken by management. Several theories explain why distressed firms could start delaying reports, such as the theory of selective disclosure ( Darrough and Stoughton 1990 ) and the obfuscation theory (Courtis 1998). Still, the empirical evidence for these theoretical proposals is scant. Only one recent study by Luypaert et al. (2016), using firms from Belgium as an example, found certain evidence using a distress risk model from 1982—stating that higher distress risk leads to a longer delay in reporting—but the exact attributes of the distress risk model were not disclosed. From a practical viewpoint, it is logical that distressed firms may try to delay reporting bad news to their stakeholders, for instance, in order to avoid a reduction in trade credit and bank loans or customer trust. Thus, we would expect the following relationship with the last hypothesis: Risks 2019,7, 77 5 of 15 Hypothesis 5 (H5). The higher the risk of bankruptcy, the higher the likelihood of delaying annual reporting over the submission date set in law. 3. The Study’s Context, Data and Methodology 3.1. Context of the Study This study was based on Estonian whole population data from 2000 to 2014. The start period was the year 2000, as for earlier years, the annual report submission dates were unavailable, and the end period was the year 2014, as when conducting the analysis, no submission dates were available for later years. Also, since 2016, an abbreviated format of the annual report became into use for SMEs, which would not be fully comparable with earlier reports. The analysis was composed of 698,189 firm-year observations. For most of the studied years, the available data represented more than 90% of the active firm population, while for a certain proportion of firms, financial information was missing in our database. Estonian legislation obliges all firms to submit an annual report 0.5 years after the financial year has ended at the latest. For the overwhelming majority of firms, the financial and calendar years overlap, and thus, for almost all firms, the final annual report submission date is 30 June in the year after the financial year. The Estonian regulation of six months is rather similar to other countries, where relevant research has been conducted, e.g., 7 months in Belgium (Luypaert et al. 2016) and 9 months in the United Kingdom (Clatworthy and Peel 2016). For each firm, information about the financial year end date and actual submission date is available, and thus, we can account for exactly how long the delay was for the report submission. The Estonian Business Code provides different penalties for firms that submit annual reports later than the final submission date. The simplest penalty is that a firm will be fined, the worst penalty, in turn, being that a firm will be deleted from the business register, which normally occurs after several consecutive years of non-submission. Still, according to Estonian law, firm deletion proceedings can already begin after one year has passed from the submission deadline. Deletion from the business register means that a firm cannot function anymore unless it is reinstated to that register. In the early 2000s, delays were quite common—occasionally more than half of firms delayed by at least one day; however, due to the implementation of a digital submission system in the late 2000s and better monitoring of submissions, the share of delaying firms dropped to around one-third of the population. Still, many of the delaying firms never submited a report due to the termination of activities (including permanent insolvency). The latter firms were not included in the sample, as financial information from the annual report was needed for testing the hypotheses. 3.2. Variables of the Study All the formulas for the variables are provided in Table 1, and herewith, their choice was motivated. The concept of “timeliness” can have a different meaning than the occurrence of a delay from a regulatory deadline in the reporting of financial statements, and therefore, there are many proxies of timeliness available. McGee (2006) defined it as the period between the company’s year-end and the date when the financial report was released for public view. Dyer and McHugh (1975) and Davies and Whittred (1980) remarked that reporting timeliness includes audit delay, which is the number of days between the balance sheet date and the date when external auditor’s report is signed. Soltani (2002) considers the financial statement issue delay as the number of days between the balance sheet date and the date of declaring the notice of the annual general meeting. We used a universal approach to define the delay, namely, whether firms exceeded the legal deadline for submission, as the sample firms were mainly non-audited private SMEs. Such a universal definition enables a future international applicability of the results. We coded the submission delay in this study in the simplest way, namely, whether a firm succeeded in submitting the report during the legal time requirements Risks 2019,7, 77 6 of 15 (that is, in half a year after the financial year; coded with 0) and whether it delayed the submission for at least one day (coded with 1). In the later analysis, this variable is named DELAY. Table 1. Coding of the variables. Dependent/Independent Variables Control Variables Code Calculation Code Calculation RETA Retained earnings/total assets SIZE Natural logarithm of total assets EBITTA Earnings before interest and taxes/total assets AGE Age (years) in the end of fiscal year WCTA (Current assets −current liabilities)/total assets AGRI NACE Rev.2 A dummy BVETD Book value of equity/total debt MANUF NACE Rev.2 B, C, D or E dummy ZSCORE Score from Altman et al. (2017) Model 2 * CONST NACE Rev.2 F or L dummy DELAY 0—non-delay, 1—delay SALES NACE Rev.2 G dummy MDELAY 0—non-delay, 1—moderate delay as ≤ 365 days, 2—severe delay as >365 days SERV NACE Rev.2 from H to K or from M to U dummy *Altman et al. (2017) Model 2 score was calculated as 1/(1 +e −L ), where L =0.035 − 0.495 × WCTA − 0.862 × RETA − 1.721 × EBITTA − 0.017 × BVETD, while for the meaning of WCTA, RETA, WCTA and BVETD refer to the calculation formulas provided in the table. NACE Rev.2 letters from A to U refer to the sections in the Statistical classification of economic activities in the European Community. The dataset was broken down into 453,776 non-delayed observations and 244,413 delayed observations. All firms that had delayed, had at some point in time still submitted their delayed reports, as otherwise it would be impossible to calculate the independent variables (i.e., financial ratios and bankruptcy risk) from the same report. We considered two other options when coding the variable DELAY. First, the usage of DELAY in a continuous form is problematic, as for non-delayed firms, the management can freely choose a date inside the legal requirements, therefore making firms submitting right after financial year unfoundedly more correct than those doing it for instance after 180 days. Second, DELAY could be used in an ordinal form, for instance, by coding non-delayed firms with 0 and delays in ascending order to represent different lengths of delays. Such an approach is also challenging, as a grouping of delays based on their length is highly subjective, as no proper scientific and practical guidelines exist on how to implement it. Still, in the robustness test of the results, delays were also used in an ordinal form, where 0 denoted non-delay, while 1 and 2, respectively, delays ≤ 365 days and >365 days (variable MDELAY). The motivation for breaking delays into mild and severe delays was the Estonian legislation (see Section 3.1), according to which a delay of more than one year can lead to more severe consequences. For coding the independent variables, we used the formulas used in the Altman et al. (2017) study, which is the extension of the initial Altman (1968) model on European non-listed firms. The formulas based on Altman et al. (2017) are as follows: (a) retained earnings to total assets (RETA) reflecting accumulated profitability (H2); (b) earnings before interest and taxes to total assets (EBITTA) reflecting annual profitability (H3); (c) working capital (i.e., current assets minus current liabilities) to total assets (WCTA) reflecting liquidity (H1); and (d) book value of equity to total debt (BVETD) reflecting leverage (H4). We acknowledge that there are different ratios available to study these financial dimensions, but our aim was to apply the most widely used ones (e.g., literature review by Dimitras et al. 1996). The second model in the Altman et al. (2017) study (a logistic regression-based model named “Model 2”) includes the latter four variables, thus, we chose it for testing H5. This failure prediction model included Estonian firms for the model estimation and has a high classification accuracy in Estonia, and thus, is suitable for use in this country. In Estonian scientific literature there are so far no multi-sectorial prediction models available, thus, the usage of the Altman et al. (2017) second model is a good solution for calculating bankruptcy risk in Estonia irrespective of firm sector. For each observation, the values of RETA, EBITTA, WCTA, BVETD and Z-score based on the Altman et al. (2017) second model (coded Risks 2019,7, 77 7 of 15 as ZSCORE) were calculated. We used the bankruptcy risk in a transformed logit format, namely, 0 ≤ ZSCORE ≤ 1, where larger values indicate higher bankruptcy risk. All four ratios were winsorized before usage in statistical analysis, as classical statistical analysis methods can be quite sensitive to outliers, which are common in case of financial ratios. 3.3. Methodology of the Study For testing the hypotheses, logistic regression analysis (logit) (Stata 15.1 command “logistic”) was used with DELAY as the dependent variable. Hypotheses from H1 to H4 were assessed together with four independent variables (RETA, EBITTA, WCTA, BVETD), and for the H5, the ZSCORE was the only independent variable, thus, resulting in two sets of logit models (Models 1.1, 1.2 and 1.3 to assess hypotheses from H1 to H4 and Models 2.1, 2.2 and 2.3 to assess H5). The testing of H1 to H4 and H5 were separated into two sets of models, as the variable ZSCORE was calculated based on the other four independent variables, which would lead to serious multicollinearity issues. When Models 1.1 and 2.1 were base models without control variables, two additional sets of models were then composed to study how the characteristics of firms portrayed with different control variables altered the initial results. In Models 1.2 and 2.2, two control variables, namely, firm size (coded as SIZE), measured as a natural logarithm of total assets, and firm age (coded as AGE), measured as years from foundation to the end of the specific financial year, were included. In Models 1.3 and 2.3, the latter two control variables were supplemented with firm sector control variables. The industry dummies (the respective NACE Rev.2 sections with proportions in the dataset in brackets) were coded as follows: (1) primary sector firms—AGRI (A; 5%); (2) industrial firms—MANUF (B, C, D, E; 11%); (3) construction firms—CONST (F, L; 20%); (4) sales firms—SALES (G; 22%); and (5) service firms—SERV (from H to K, from M to U; 42%). As service firms hold the largest proportion in the sample, that sector was used as the base category and not included in Model 1.3 and Model 2.3 estimations. In the case of both models, the marginal effects were presented and multicollinearity tests conducted. In addition, the robustness of the results in respect to the dependent variable choice was reassessed for Models 1.1 and 2.1 using the dependent variable MDELAY in ordered logistic regression (Stata 15.1 command “ologit”). The models assessed with either logistic regression (dependent variable DELAY) or ordered logistic regression (dependent variable MDELAY) were as follows (idenotes firm and tperiod): Model 1.1: DELAYit =b0+b1×RETAit +b2×EBITTAit +b3×WCTAit +b4×BVETDit +εit Model 1.2: DELAY it =b 0 +b 1× RETA it +b 2× EBITTA it +b 3× WCTA it +b 4× BVETD it +b 5× SIZEit +b6×AGEit +εit Model 1.3: DELAY it =b 0 +b 1× RETA it +b 2× EBITTA it +b 3× WCTA it +b 4× BVETD it +b 5× SIZEit +b6×AGEit +b7×AGRIit +b8×MANUFit +b9×CONSTit +b10 ×SALESit +εit Model 2.1: DELAYit =b0+b1×ZSCOREit +εit Model 2.2: DELAYit =b0+b1×ZSCOREit +b2×SIZEit +b3×AGEit +εit Model 2.3: DELAY it =b 0 +b 1× ZSCORE it +b 2× SIZE it +b 3× AGE it +b 4× AGRI it +b 5× MANUFit +b6×CONSTit +b7×SALESit +εit Model 1.1 (reassessed): MDELAY it =b 0 +b 1× RETA it +b 2× EBITTA it +b 3× WCTA it +b 4× BVETDit +εit Model 2.1 (reassessed): MDELAYit =b0+b1×ZSCOREit +εit 4. Results and Discussion The descriptive statistics of variables are documented in Table 2. It can be seen that the means and medians of WCTA, EBITTA, RETA, and BVETD were higher in the case of non-delayed firms, therefore presenting better financial situations for non-delayed firms than for delayers. Also, the ZSCORE obtained higher mean and median values in the case of delayed firms, therefore indicating a higher risk of bankruptcy. These results are in line with prior literature, e.g., Courtis (1976), Abdulla (1996), Owusu-Ansah (2000). Out of the population applied, 42% of firms were from the service sector, followed Risks 2019,7, 77 8 of 15 by firms dealing with sales and construction. The median age of firms was 6.9 years, therefore reflecting an adolescent company, and the median size was 43.5 thousand euros (the natural logarithm of size was 10.68), indicating a micro firm. Thus, our paper contributes to the scant literature on firms with a small size (Luypaert et al. 2016). Six logit models were composed, which are documented in Tables 3 and 4with respective marginal effects and multicollinearity statistics (i.e., variance inflation factors, VIFs). Additionally, two ordered logistic regression models are presented in Table 5to demonstrate the robustness of the results in respect to the dependent variable choice. Table 2. Descriptive statistics of the variables. Variable Non-Delayed Firms (N=453,776) Delayed Firms (N=244,413) All Firms (N=698,189) Mean SD Median Mean SD Median Mean SD Median WCTA 0.30 0.49 0.31 0.22 0.54 0.24 0.27 0.51 0.29 EBITTA 0.05 0.35 0.04 0.01 0.39 0.03 0.03 0.37 0.04 RETA 0.23 0.71 0.32 0.09 0.85 0.21 0.18 0.77 0.29 BVETD 9.36 16.04 1.69 8.44 15.76 1.14 9.04 15.95 1.49 ZSCORE 0.36 0.22 0.35 0.40 0.24 0.39 0.37 0.23 0.36 SIZE 10.87 2.09 10.70 10.76 2.14 10.64 10.83 2.11 10.68 AGE 8.21 5.18 7.15 6.98 4.54 5.90 7.78 5.00 6.70 Note: Descriptive statistics are not reported for sectoral dummies, as they are not informative in the case of binary variables. Instead, the sectoral composition of the sample is described in Section 3.3. The p-value of the two-sample t-test and Brown-Forsyth ANOVA was <0.001 for all independent variables in Table 2. Correlations of the variables are documented in Appendix A. Models 1.1, 1.2 and 1.3 indicated that three of the four financial ratios were significantly related to DELAY, namely, a rise in liquidity, annual profitability and accumulated profitability decreased the likelihood of a reporting delay, thus enabling to accept H1, H2 and H3. Based on the composed models (1.1 and 1.2), the relationship between leverage and reporting delays can be considered insignificant, leading to the acceptance of H4. Still, it must be acknowledged, that in Model 1.3, including the largest amount of control variables, leverage was significant, which could be caused by some sector-specific features. The marginal rise in accumulated profitability mattered the most, which could be explained by a strong managerial intent to hide dropping of it below a certain threshold, i.e., accumulated profits becoming very low or even negative. Firms seem to be more afraid of different sanctions concerning low (or negative) retained earnings, rather than being concerned about the consequences of a delay. The control variables in Models 1.2 and 1.3 indicated that, ceteris paribus, larger firms are at a higher risk and older firms are at a lower risk of delay. Our results are opposite to Owusu-Ansah (2000) , who found an inverse relationship between timeliness and the size of the company. A possible explanation for our results could be connected with the composition of the firm population, namely, the large proportion of micro firms. Smaller micro firms may have a very simple accounting, namely, so few items to report, that they can manage the composition of their reports quickly. Moreover, with the increase in size, the probability that a firm is using leverage increases, and thus, such firms could be more prone to hiding negative results from lenders as long as possible. However, for older firms, this relationship is opposite (i.e., delays become less frequent), in line with Iyoha (2012). This may be caused by the fact that older firms have established a position in the market and do not want to hurt their prestige by delaying their reports. Firms in primary and sales sectors are more likely to delay than in the reference group of service firms. In turn, manufacturing and especially construction firms do not differ from service firms. There are two additional findings concerning the age and size context based on the models presented in Table 3. Larger firms have to be more liquid and profitable, so that their likelihood to present annual reports in time would be the same as for their smaller counterparts. Larger firms have more stakeholders and there is wider public interest in them, which could lead to higher willingness to postpone unfavourable financial developments. At the same time, for instance in the case of a one-individual owned micro firm, where the owner is also the manager, they may not have any such Risks 2019,7, 77 15 of 15 Owusu-Ansah, Stephen. 2000. Timeliness of corporate financial reporting in emerging capital markets: Empirical evidence from the Zimbabwe Stock Exchange. Accounting and Business Research 30: 241–54. [CrossRef] Owusu-Ansah, Stephen, and Stergios Leventis. 2006. Timeliness of corporate annual financial reporting in Greece. European Accounting Review 15: 273–87. [CrossRef] Peek, Erik, Rick Cuijpers, and Willem Buijink. 2010. Creditors’ and shareholders’ reporting demands in public versus private firms: Evidence from Europe. Contemporary Accounting Research 27: 49–91. [CrossRef] Peel, Michael J., and David A. Peel. 1988. A multilogit approach to predicting corporate failure—Some evidence for the UK corporate sector. Omega 16: 309–18. [CrossRef] Peel, Michael J., David A. Peel, and Peter F. Pope. 1986. Predicting corporate failure—Some results for the UK corporate sector. Omega 14: 5–12. [CrossRef] Pindado, Julio, Luis Rodrigues, and Chabela de la Torre. 2008. Estimating financial distress likelihood. Journal of Business Research 61: 995–1003. [CrossRef] Scott, James. 1981. The probability of bankruptcy: A comparison of empirical predictions and theoretical models. Journal of Banking & Finance 5: 317–44. Singhvi, Surendra S., and Harsha B. Desai. 1971. An empirical analysis of the quality of corporate financial disclosure. The Accounting Review 46: 129–38. Skinner, Douglas J. 1993. The investment opportunity set and accounting procedure choice: Preliminary evidence. Journal of Accounting and Economics 16: 407–45. [CrossRef] Skinner, Douglas J. 1994. Why firms voluntarily disclose bad-news. Journal of Accounting Research 32: 38–60. [CrossRef] Skinner, Douglas J. 1997. Earnings disclosure and stockholder lawsuits. Journal of Accounting and Economics 23: 249–83. [CrossRef] Smith, Clifford W., and Ross L. Watts. 1992. The investment opportunity set and corporate financing, dividend, and compensation policies. Journal of Financial Economics 32: 263–92. [CrossRef] Soltani, Bahram. 2002. Timeliness of corporate and audit reports: Some empirical evidence in the French context. The International Journal of Accounting 37: 215–46. [CrossRef] Tasc ó n Fern á ndez, Mar í a T., and Francisco J. Castaño Guti é rrez. 2012. Variables and models for the identification and prediction of business failure: Revision of recent empirical research advances. Spanish Accounting Review 15: 7–58. Whittred, Greg, and Ian Zimmer. 1984. Timeliness of financial reporting and financial distress. The Accounting Review 59: 287–95. Wu, Wei-Wen. 2010. Beyond business failure prediction. Expert Systems with Applications 37: 2371–76. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).