Company rating with support vector machines
-
Russ A. Moro
,
Wolfgang K. Härdle
Abstract
This paper proposes a rating methodology that is based on a non-linear classification method, a support vector machine, and a non-parametric isotonic regression for mapping rating scores into probabilities of default. We also propose a four data set model validation and training procedure that is more appropriate for credit rating data commonly characterised with cyclicality and panel features. Tests on representative data covering fifteen years of quarterly accounts and default events for 10,000 US listed companies confirm superiority of non-linear PD estimation. Our methodology demonstrates the ability to identify companies of diverse credit quality from Aaa to Caa–C.
Funding source: Deutsche Forschungsgemeinschaft
Award Identifier / Grant number: SFB 649 “Economic Risk”
Funding statement: The authors gratefully acknowledge the support of this project by the Risk Management Institute of the National University of Singapore (RMI NUS). We thank RMI NUS for providing access to their database of financial statements and default events. Russ A. Moro was financially supported by the RMI NUS and Wolfgang K. Härdle by Deutsche Forschungsgemeinschaft through SFB 649 “Economic Risk”.
Acknowledgements
We are grateful to Jin-Chuan Duan and Oliver Chen of RMI NUS and Laura Auria and Ralf Körner of the Deutsche Bundesbank for their irreplaceable assistance and valuable suggestions. We also would like to thank participants of the International Risk Management Conferences for their comments and discussion.
References
[1] E. Altman, Financial ratios, discriminant analysis and the prediction of corporate bankruptcy, J. Finance 23 (1968), no. 4, 589–609. 10.1111/j.1540-6261.1968.tb00843.xSuche in Google Scholar
[2] E. Altman and A. Saunders, Credit risk measurement: Developments over the last 20 years, J. Banking Finance 21 (1998), no. 11–12, 1721–1742. 10.1016/S0378-4266(97)00036-8Suche in Google Scholar
[3] R. E. Barlow, J. M. Bartholomew, J. M. Bremmer and H. D. Brunk, Statistical Inference Under Order Restrictions, John Wiley & Sons, New York, 1972. Suche in Google Scholar
[4] E. Carrizosa, A. Nogales-Gómez and D. R. Morales, Strongly agree or strongly disagree? Rating features in support vector machines, Inform. Sci. 329 (2016), 256–273. 10.1016/j.ins.2015.09.031Suche in Google Scholar
[5] A. H. Chen, N. Ju, S. C. Mazumdar and A. Verma, Correlated default risks and bank regulations, J. Money Credit Banking 38 (2006), 375–398. 10.1353/mcb.2006.0028Suche in Google Scholar
[6] C.-C. Chen and S.-T. Li, Credit rating with a monotonicity-constrained support vector machine model, Expert Syst. Appl. 41 (2014), no. 16, 7235–7247. 10.1016/j.eswa.2014.05.035Suche in Google Scholar
[7] W.-H. Chen and J.-Y. Shih, A study of Taiwan’s issuer credit rating systems using support vector machines, Expert Syst. Appl. 30 (2006), no. 3, 427–435. 10.1016/j.eswa.2005.10.003Suche in Google Scholar
[8] J. de Leeuw, K. Hornik and P. Mair, Isotone optimization in R: Pool-adjacent-violators algorithm (PAVA) and active set methods, J. Stat. Softw. 32 (2009), no. 5, 1–24. 10.18637/jss.v032.i05Suche in Google Scholar
[9] J.-C. Duan, J. Sun and T. Wang, Multiperiod corporate default prediction – A forward intensity approach, J. Econometrics 170 (2012), no. 1, 191–209. 10.1016/j.jeconom.2012.05.002Suche in Google Scholar
[10] D. W. Dwyer, A. E. Kocagil and R. M. Stein, Moody’s KMV RiskCalc v3.1 model: Next-generation technology for predicting private firm credit risk, White Paper (2004), Moody’s KMV Company. Suche in Google Scholar
[11] B. Engelmann, E. Hayden and D. Tasche, Measuring the discriminative power of rating systems, Discussion Paper Series 2: Banking and Financial Supervision, Deutsche Bundesbank, 2003. 10.2139/ssrn.2793951Suche in Google Scholar
[12] E. Falkenstein, A. Boral and L. Carty, Riskcalc for private companies: Moody’s default model, Report 56402, Moody’s Investors Service, New York, 2000. 10.2139/ssrn.236011Suche in Google Scholar
[13] J. E. Fernandes, Corporate credit risk modeling: Quantitative rating system and probability of default estimation, preprint (2005), http://econwpa.repec.org/eps/fin/papers/0505/0505013.pdf. 10.2139/ssrn.722941Suche in Google Scholar
[14] T. V. Gestel, J. A. K. Suykens, G. Lanckriet, A. Lambrechts, B. D. Moor and J. Vandewalle, Bayesian framework for least-squares support vector machine classifiers, Gaussian processes, and kernel Fisher discriminant analysis, J. Neural Comput. 14 (2002), no. 5, 1115–1147. 10.1162/089976602753633411Suche in Google Scholar
[15] D. Glennon and P. Nigro, Measuring the default risk of small business loans: A survival analysis approach, J. Money Credit Banking 37 (2005), no. 5, 923–947. 10.1353/mcb.2005.0051Suche in Google Scholar
[16] S. J. Grotzinger and C. Witzgall, Projections onto simplices, Appl. Math. Optim. 12 (1984), no. 1, 247–270. 10.1007/BF01449044Suche in Google Scholar
[17] P. Hajek and K. Michalak, Feature selection in corporate credit rating prediction, Knowledge Based Syst. 51 (2013), 72–84. 10.1016/j.knosys.2013.07.008Suche in Google Scholar
[18] D. T. Hamilton and R. Cantor, Measuring corporate default rates, Report 100779, Moody’s Investors Service, 2006. Suche in Google Scholar
[19] D. Hand, Measuring classifier performance: A coherent alternative to the area under the ROC curve, Mach. Learn. 77 (2009), no. 1, 103–123. 10.1007/s10994-009-5119-5Suche in Google Scholar
[20] J. A. Hanley and B. J. McNeil, The meaning and use of the area under a receiver operating characteristic (ROC) curve, Radiology 143 (1982), no. 1, 29–36. 10.1148/radiology.143.1.7063747Suche in Google Scholar
[21] Z. Huang, H. Chen, C.-J. Hsu, W.-H. Chen and S. Wu, Credit rating analysis with support vector machines and neural networks: A market comparative study, Decis. Support Syst. 37 (2004), no. 4, 543–558. 10.1016/S0167-9236(03)00086-1Suche in Google Scholar
[22] A. Karatzoglou, D. Meyer and K. Hornik, Support vector machines in R, J. Stat. Softw. 15 (2006), no. 9, 1–28. 10.18637/jss.v015.i09Suche in Google Scholar
[23] K. Kim and H. Ahn, A corporate credit rating model using multi-class support vector machines with an ordinal pairwise partitioning approach, Comput. Oper. Res. 39 (2012), no. 8, 1800–1811. 10.1016/j.cor.2011.06.023Suche in Google Scholar
[24] J. T.-Y. Kwok, Moderating the outputs of support vector machine classifiers, IEEE Trans. Neural Networks 10 (1999), no. 5, 1018–1031. 10.1109/IJCNN.1999.831080Suche in Google Scholar
[25] J. T.-Y. Kwok, The evidence framework applied to support vector machines, IEEE Trans. Neural Networks 11 (2000), no. 5, 1162–1173. 10.1109/72.870047Suche in Google Scholar
[26] Y.-C. Lee, Application of support vector machines to corporate credit rating prediction, Expert Syst. Appl. 33 (2007), no. 1, 67–74. 10.1016/j.eswa.2006.04.018Suche in Google Scholar
[27] H. B. Mann and D. R. Whitney, On a test of whether one of two random variables is stochastically larger than the other, Ann. Math. Stat. 18 (1947), no. 1, 50–60. 10.1214/aoms/1177730491Suche in Google Scholar
[28] M. J. Manning, Exploring the relationship between credit spreads and default probabilities, Working Paper 225, Bank of England, 2004. 10.2139/ssrn.641262Suche in Google Scholar
[29] D. Martens, B. Baesens, T. van Gestel and J. Vanthienen, Comprehensible credit scoring models using rule extraction from support vector machines, Working Paper 878283, Social Science Research Network, 2006. 10.2139/ssrn.878283Suche in Google Scholar
[30] D. Martin, Early warning of bank failure: A logit regression approach, J. Banking Finance 1 (1977), no. 3, 249–276. 10.1016/0378-4266(77)90022-XSuche in Google Scholar
[31] R. C. Merton, On the pricing of corporate debt: The risk structure of interest rates, J. Finance 29 (1974), no. 2, 449–470. 10.1142/9789814759588_0003Suche in Google Scholar
[32] J. A. Ohlson, Financial ratios and the probabilistic prediction of bankruptcy, J. Accounting Res. 18 (1980), no. 1, 109–131. 10.2307/2490395Suche in Google Scholar
[33] J. C. Platt, Probabilities for SV machines, Advances in Large Margin Classifiers, The MIT Press, Cambridge (2000), 61–73. 10.7551/mitpress/1113.003.0008Suche in Google Scholar
[34] M. D. Richard and R. P. Lippmann, Neural network classifiers estimate Bayesian a posteriori probabilities, J. Neural Comput. 3 (1991), no. 4, 461–483. 10.1162/neco.1991.3.4.461Suche in Google Scholar PubMed
[35] B. Schölkopf and A. J. Smola, Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond, MIT Press, London, 2002. 10.7551/mitpress/4175.001.0001Suche in Google Scholar
[36] W. Sjur, S. Hol and N. van der Wijst, Capital structure, business risk, and default probability, Discussion Paper 975302, Social Science Research Network, 2007. Suche in Google Scholar
[37] V. M. Tikhomirov, The evolution of methods of convex optimization, Amer. Math. Monthly 103 (1996), no. 1, 65–71. 10.1080/00029890.1996.12004701Suche in Google Scholar
[38] V. Vapnik, The Nature of Statistical Learning Theory, Springer, New York, 1995. 10.1007/978-1-4757-2440-0Suche in Google Scholar
[39] L. Yu, X. Yao, S. Wang and K. K. Lai, Credit risk evaluation using a weighted least squares svm classifier with design of experiment for parameter selection, Expert Syst. Appl. 12 (2011), 15392–15399. 10.1016/j.eswa.2011.06.023Suche in Google Scholar
© 2017 Walter de Gruyter GmbH, Berlin/Boston
Artikel in diesem Heft
- Frontmatter
- A double clustering algorithm for financial time series based on extreme events
- Improved algorithms for computing worst Value-at-Risk
- Testing for asymmetry in betas of cumulative returns: Impact of the financial crisis and crude oil price
- Company rating with support vector machines
- Loan pricing under estimation risk
Artikel in diesem Heft
- Frontmatter
- A double clustering algorithm for financial time series based on extreme events
- Improved algorithms for computing worst Value-at-Risk
- Testing for asymmetry in betas of cumulative returns: Impact of the financial crisis and crude oil price
- Company rating with support vector machines
- Loan pricing under estimation risk
