Predicting multilateral trade credit risks: comparisons of Logit and Fuzzy Logic models using ROC curve analysis

https://doi.org/10.1016/j.eswa.2004.12.016Get rights and content

Abstract

Employing pooled data of 3344 listed firms from seven Asia-Pacific countries, this is the first empirical study to classify and predict trade credit risks in the international trade context. In addition, this paper extends previous work by applying receiver operating characteristic (ROC) curve analysis to compare the model performance of Logit to that of Fuzzy Logic (FL). We are unaware of any other paper that has discussed the application of ROC curve analysis in the business and finance literature.

The results show that FL exceeds Logit in terms of overall classification accuracy and prediction accuracy. However, by incorporating measurement in the form of ROC curves, Logit is proven to outperform FL in classifying non-default firms. This suggests that though FL is superior in overall accuracy and in classifying default firms, Logit is preferable in situations where higher accuracy in classifying non-default firms is preferred. The stability of the models is also demonstrated.

Section snippets

Introduction and background to research

In recent years, intensified market competition in international trade has led to letter of credit (L/C) being replaced by open account (O/A). For exporters, however, sales on credit have greatly increased the amount of their foreign receivables and thus have had a severe impact on the operational risks, causing problems such as difficulties in capital turnover, fluctuations in exchange rate, bad debts, and decrease of profits due to the de facto discounts. The bad debt risk generated from

An overview of FL

Unlike conventional methods that entail an understanding of a system, exact equations, and precise mathematical values, FL incorporates an alternative way of thinking and reasoning, which allows anticipating or modeling the complex problem domain using a higher level of vagueness via our knowledge and experience. One of the characteristics of FL is its ability to be both linguistically tractable and mathematically sound. It resembles human decision-making with its ability to work from

Research design

In this section, we discuss our data sources, sample selection criteria, and variable selection process. The research design is depicted in Fig. 1.

Statistical results of variables

The statistics of skewness, kurtosis, and Kolmogorov–Smirnov Z indicate that the majority of variables, except FS5, FS6, FS9, FS10, FS15, BM2, IC3, and MF2, clearly violate the normality assumption. For the Mann–Whitney non-parametric test statistic, the dispersion of the two datasets or rather the homogeneity of the variance–covariance matrices is not the same.

Bartlett's Sphericity test statistic reveals statistically significant interrelations within the 36 raw variables (p-value=0.04).

Conclusions

Employing pooled data of 3344 listed firms from seven Asia-Pacific countries, this is the first empirical study to classify and predict trade credit risks in the international trade context. In addition, this paper extends previous work by applying ROC curve analysis to compare the model performance of Logit to that of FL. We are unaware of any other paper that has discussed the application of ROC curve analysis in the business and finance literature. Given the growing importance of

Acknowledgements

Financial support from the Taiwanese National Science Council, through Projects NSC 91-2626-H-150-001 and NSC 92-2626-H-150-002, is gratefully acknowledged.

References (28)

  • W.H. Beaver

    Financial ratios as predictors of failure

    Supplement to Journal of Accounting Research: Empirical Research in Accounting-Selected Studies

    (1966)
  • E.B. Deakin

    A discriminant analysis of predictors of failure

    Journal of Accounting Research

    (1972)
  • D. Dubois et al.

    Fuzzy sets and systems: Theory and applications

    (1980)
  • H. Frydman et al.

    Introducing recursive partitioning for financial classification: The case of financial distress

    The Journal of Finance

    (1985)
  • Cited by (68)

    • Estimation of jellyfish abundance in the south-eastern Spanish coastline by using an explainable artificial intelligence model based on fuzzy logic

      2022, Estuarine, Coastal and Shelf Science
      Citation Excerpt :

      Therefore, the collection of useful data and the obtaining conclusions requires a pre-processing of the information provided by the users and a subsequent meta-analysis characterised by the use of non-linearly correlated data associated with a high degree of uncertainty. Under these circumstances, techniques of explainable artificial intelligence such as fuzzy logic (FL) show better performance compared to the results obtained with others statistical procedures (Tang and Chi, 2005; Yi et al., 2016; Arrieta et al., 2020). In particular, FL mathematically formalises the way humans interact with the environment and make decisions.

    • Fuzzy risk analysis based on a similarity measure of fuzzy numbers and its application in crop selection

      2022, Engineering Applications of Artificial Intelligence
      Citation Excerpt :

      Hence, risk analysis is an important part of the study, which has to be thoroughly studied. Some of the studies involving risk analysis are Liu et al. (2010), Toledo et al. (2011), Wang and Elhag (2006) and Tang and Chi (2005). Risk analysis is a very important part in every field of applications, namely, decision-making, optimization, etc.

    • Fuzzy risk analysis in poultry farming based on a novel similarity measure of fuzzy numbers

      2018, Applied Soft Computing Journal
      Citation Excerpt :

      Kangari and Riggs [2] proposed a method of risk analysis using linguistic terms. Some of the studies involving risk analysis are Chen [3], Chen and Chen [4–6], Tang and Chi [7], Wang and Elhag [8], etc. Generally, it happens that the linguistic terms are expressed in terms of fuzzy numbers.

    View all citing articles on Scopus
    View full text