Chapter 4: Default prediction for audited and unaudited private firms under
4.2 Literature review
106 The results of the study provide compelling evidence showing that accounting information is crucial in separating defaulted private firms from non-defaulted ones under downturn conditions. The nominated input financial ratios are imperative because they represent some of the most important credit risk drivers, including, profitability, leverage, growth and liquidity. Also, this study indicates that the inclusion of macroeconomic variables improves model fit and the prediction performance of the default models. In line with this finding, Sheikh and Yahya (2015) and Hill, Perry and Andes (2011) indicated that the prediction results of bankruptcy forecasting techniques are improved by incorporating macroeconomic factors. This implies that firm-and- loan-characteristics, accounting-data and macroeconomic-information based models best explain the default probability for audited and unaudited Zimbabwean private firms. Further, the study reveals that the drivers of default risk for audited and unaudited Zimbabwean private firms are not the same.
The rest of the chapter is designed as follows: section 4.2 outlines the literature review and section 4.3 presents a brief overview of the methodology. In section 4.4, data and variables are described and section 4.5 is allocated to experimental results. Section 4.6 presents the discussion of results and section 4.7 outlines the robustness checks for the designed models. Section 4.8 concludes the analysis,provides the implications of the study and presents potential directions for future research.
107 analysis in examining corporate bankruptcy by forecasting bank failure and Ohlson (1980) became the first author to implement a logit model to analyse bankruptcy for non-financial sector corporates.
Although corporate default prediction has been receiving much attention in risk management, most studies focus on public firms in developed economies (Bauer and Agarwal, 2014; Tinoco and Wilson, 2013; Agarwal and Taffler, 2008; Shumway, 2001). Corporate default forecasting literature for developing economies (Kwak, Shi and Kou, 2012) and for privately-owned companies (Charalambakis and Garrett, 2019) is generally restricted. Limited evidence on private firm default prediction is substantially dedicated to advanced markets (Diekes et al., 2013; Cangemi, Servigny and Friedman, 2003; Falkenstein, Boral and Carty, 2000). Nevertheless, applying models designed for advanced countries to emerging economies does not always produce plausible results (see, Ashraf, Felix and Serrasqueiro, 2019) due to the fact that the economic structures of developed and undeveloped economies are considerably different (Liang, Tsai, and Wu, 2015; Fedorova, Gilenko and Dovzhenko, 2013). For instance, developing economies depend on primary commodity exports to generate foreign currency and they bank on imports for some basic food supplies and energy. Therefore, developing countries are exposed to more challenges when dealing with booms originating from commodity price surges and exogenous shocks which seem insignificant but habitually lead to high corporate default rates in undeveloped markets. Due to these structural differences, developing economies are characterised by limited historical statistical data, closed markets, institutional and legal barriers and reduced predictive power of market signals (Rylov, Shkurkin and Borisova, 2016), which affect how corporate default probability is modelled. Moreover, Waqas and Md- Rus (2018) argued that it is important to recognise that advanced economies are associated with clearer bankruptcy laws and procedures than undeveloped economies.
In the same vein, Altman (2018), Takahashi, Taques and Basso (2018) and Slefendorfas (2016) pronounced that each country has unique characteristics, and consequently models created explicitly for individual economies outperform general models.
Hayden (2011) propounded that the forecasting capability of statistical techniques is premised on the presumption that the past association between the predictor variables of the developed model and the default event will remain the same in the future.
108 However, this supposition may not remain unchanged over long periods given a wide range of possible events that can take place in financial markets, for example, changes in accounting policies of companies, financial and economic crises and the introduction of regulatory documents. (Takahashi, Taques and Basso, 2018; Singh and Mishra, 2016; Smaranda, 2014; Hayden, 2011). Owing to changes in time horizons, financial situations and economic conditions, the applicability and predictive performance of the existing default detection techniques under new settings is a practical inquiry that needs to be addressed in modern finance (Altman, 2018; Timmermans, 2014).
Takahashi, Taques and Basso (2018), Smaranda (2014) and Hayden (2011) revealed that it is necessary to regularly re-validate and re-calibrate the bankruptcy forecasting models in the wake of new events in order to guarantee that their detection capacity does not decrease. Further, Sign and Mishra (2016) re-estimated the Z-score (Altman, 1968), Y-score (Ohlson, 1980), and X-score (Zmijewski, 1984) models and posited that the coefficients of these models are responsive to time horizons and changes in financial situations.
Charalambakis and Garrett (2019) collecting default data and information for private corporations is a difficult task because their shares are not bought and sold on stock exchanges. Therefore, financial statements and records of firm borrowers pooled from banks are the main sources of default data and information for private corporations.
Accounting information used in detecting default probability for private firms is derived from audited and unaudited financial statements. In several economies, there are no legal demands for privately-owned corporates to disclose their financial results and generate audited financial statements even though some do (Minnis and Shroff, 2017). Bratten et al. (2013) and Minnis (2011) proffered that audited financial records guarantee that there are no material mistakes or misstatements in the results. Further, Bratten et al. (2013) and Dechow, Ge and Schrand (2010) posited that by reducing the misrepresentations of financial records, a credible audit guarantees reliable financial reporting.
Auditors perform information and insurance roles. Investors believe that companies audited by huge firms have plausible earnings and are less risky. Accordingly, audited corporates benefit from low-interest rates on borrowed funds and low returns anticipated by investors (Cenciarelli, Greco and Allegrin, 2018). Huq, Hartwig and Rudholm (2018), Cassar (2011) and Minnis (2011) posited that firms that present
109 audited financial statements to creditors are characterised by lower costs of debt than corporates that do not. The cost of debt indicates the probability of default associated with the borrower. Corporates with high cost of debt are associated with high default risk, and those with low cost of debt are characterised by low default risk.
Consequently, Cenciarelli, Greco and Allegrin (2018) and Gul, Zhou and Zhu (2013) indicated that corporations with audited financial statements pose lower default risk to creditors than their unaudited counterparts. Further, Cenciarelli, Greco and Allegrin (2018) posited that auditing firms can avert corporate bankruptcies. They can address the issues related to accounting frameworks’ inadequacies and financial regulations’
mediocrity. Inadequate accounting frameworks, mediocre financial regulations and suboptimal productivity are some of the major causes of corporate failure (Jahur and Quadir, 2012). Cenciarelli, Greco and Allegrin (2018) and references therein argued that big auditing firms have the skills and prowess to analyse bankruptcy and advise firms on how to deal with it. Based on the examination of the financial ratios, Hamzani and Achmad (2018) proposed that small-to-medium enterprises (SMEs) complying with the accounting standards have higher profitability than their non-complying counterparts. Given that profitability is negatively associated with bankruptcy, the finding of Hamzani and Achmad (2018) seems to support the supposition that audited firms are associated with low rates of default.
Since the economic downturns are associated with high default frequencies (see, for example, Mihalovic 2016; Canals-Cerda and Kerr, 2015a; Leow and Crook, 2014), the AIRB approach and IFRS 9 provide new impetus for banks to design new default detection models under distressed economic and financial conditions (see International Accounting Standards Board 2014; Basel Committee on Banking Supervision, 2011, 2006). Jensen, Lando and Medhat (2017) further stated that a private firm default model premised on only firm-specific factors is not proficient in describing the cyclical nature of the witnessed defaults. Using Cox models, Jensen, Lando and Medhat (2017) posited that macroeconomic factors and accounting ratios are crucial in default forecasting for Danish privately-owned firms. The authors proposed that the effects of firm-specific factors remain robust to the addition of the macroeconomic variables.
Charalambakis and Garrett (2019), Crook and Bellotii (2013) and Bellotii and Crook, (2009) observed that the inclusion of macroeconomic factors improves the model fit and the forecasting ability of default models.
110