The usage of logistic regression and artificial neural networks for evaluation and predicting property-liability insurers' solvency in Egypt

Eid Elghaly Hassan; Diping Zhang; Eid Elghaly Hassan; Diping Zhang

doi:10.3934/DSFE.2021012

Data Science in Finance and Economics

2021, Volume 1, Issue 3: 215-234. doi: 10.3934/DSFE.2021012

Previous Article Next Article

Research article

The usage of logistic regression and artificial neural networks for evaluation and predicting property-liability insurers' solvency in Egypt

Eid Elghaly Hassan ^{1,2
,
,},
Diping Zhang ¹

1.
School of Science, Zhejiang University of Science and Technology, Hangzhou 310023, China
2.
Department of Applied Statistics and Insurance, Faculty of Commerce, Mansoura University, Mansoura 35516, Egypt

Received: 05 October 2021 Accepted: 19 October 2021 Published: 29 October 2021
JEL Codes: C45, G22, N27

Unlike prior solvency prediction studies conducted in Egypt, this study aims to set up a real picture of companies' financial performance in the Egyptian insurance market. Therefore, 11 financial ratios commonly used by NAIC, AM BEST Company, and S & P Global Ratings were calculated for all property-liability insurance companies in Egypt from 2010 to 2020. They have been used to measure those companies' financial performance efficiency levels by comparing these ratios with the international standard limits. The financial analysis results for those companies revealed that property-liability insurers in Egypt do not have the same level of financial performance efficiency where those companies are classified into three groups: excellent, good, and poor. Furthermore, this paper investigates using the stepwise logistic regression model to determine the most factors among these selected financial ratios that influence those companies' financial performance. The results suggest that only three ratios were statistically significant predictors: "Risk retention rate", "Insurance account receivable to total assets", and "Net profit after tax to total assets". Finally, this paper presents the multi-layers artificial neural network with a backpropagation algorithm as a new solvency prediction model with perfect classifying accuracy of 100%. The trained ANN could predict the next fiscal year with a prediction accuracy of 91.67%, and this percent is a good and favorable result comparing to other solvency prediction models used in Egypt.
- stepwise logistic regression,
- multi-layers artificial neural network,
- classification,
- solvency,
- financial ratios,
- property-liability,
- the Egyptian insurance market
Citation: Eid Elghaly Hassan, Diping Zhang. The usage of logistic regression and artificial neural networks for evaluation and predicting property-liability insurers' solvency in Egypt[J]. Data Science in Finance and Economics, 2021, 1(3): 215-234. doi: 10.3934/DSFE.2021012

Related Papers:

Abstract

Unlike prior solvency prediction studies conducted in Egypt, this study aims to set up a real picture of companies' financial performance in the Egyptian insurance market. Therefore, 11 financial ratios commonly used by NAIC, AM BEST Company, and S & P Global Ratings were calculated for all property-liability insurance companies in Egypt from 2010 to 2020. They have been used to measure those companies' financial performance efficiency levels by comparing these ratios with the international standard limits. The financial analysis results for those companies revealed that property-liability insurers in Egypt do not have the same level of financial performance efficiency where those companies are classified into three groups: excellent, good, and poor. Furthermore, this paper investigates using the stepwise logistic regression model to determine the most factors among these selected financial ratios that influence those companies' financial performance. The results suggest that only three ratios were statistically significant predictors: "Risk retention rate", "Insurance account receivable to total assets", and "Net profit after tax to total assets". Finally, this paper presents the multi-layers artificial neural network with a backpropagation algorithm as a new solvency prediction model with perfect classifying accuracy of 100%. The trained ANN could predict the next fiscal year with a prediction accuracy of 91.67%, and this percent is a good and favorable result comparing to other solvency prediction models used in Egypt.

References

[1]	Affes Z, Hentati-Kaffel R (2019) Predicting US banks bankruptcy: logit versus Canonical Discriminant analysis. Comput Econ 54: 199-244. doi: 10.1007/s10614-017-9698-0
[2]	Altman EI, Haldeman R, Narayanan P (1977) ZETA Analysis, a New Model for Bankruptcy Classification. J Bank Financ 1: 29-54. doi: 10.1016/0378-4266(77)90017-6
[3]	Altman EI (1968) Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy. J Financ 23: 589-609. doi: 10.1111/j.1540-6261.1968.tb00843.x
[4]	Altman EI, Iwanicz-Drozdowska M, Laitinen EK, et al. (2020) A race for long horizon bankruptcy prediction. Appl Econ 52: 4092-4111. doi: 10.1080/00036846.2020.1730762
[5]	Atiya AF (2001) Bankruptcy Prediction for Credit Risk Using Neural Networks: A Survey and New Results. IEEE Tran Neural Networks 12: 929-935. doi: 10.1109/72.935101
[6]	Back B, Laitinen T, Sere K, et al. (1996) Choosing Bankruptcy Predictors Using Discriminant Analysis, Logit Analysis, and Genetic Algorithms. Technical Report No. 40. Turku Centre for Computer Science, Turku.
[7]	Baranoff EG, Sager TW, Shively TS (2000) A Semiparametric Stochastic Spline Model as a Managerial Tool for Potential Insolvency. J Risk Insur 67: 369-396. doi: 10.2307/253834
[8]	BarNiv R, McDonald JB (1992) Identifying Financial Distress in the Insurance Industry: A Synthesis of Methodological and Empirical Issues. J Risk Insur 59: 543-573. doi: 10.2307/253344
[9]	BarNiv R, Hershbarger RA (1990) Classifying Financial Distress in the Life Insurance Industry. J Risk Insur 57: 110-136. doi: 10.2307/252927
[10]	Beaver WH (1966) Financial ratios as predictors of failure. J Account Res 4: 71-111. doi: 10.2307/2490171
[11]	Berthold M, Hand DJ (2003) Intelligent data analysis. Vol. 2. Berlin: Springer.
[12]	Bradley AP (1997) The use of the area under the ROC curve in the evaluation of machine learning algorithms. Pattern Recognit 30: 1145-1159. doi: 10.1016/S0031-3203(96)00142-2
[13]	Brockett LP, Cooper WW, Golden LL, et al. (1994) A neural network method for obtaining an early warning of insurer insolvency. J Risk Insur 61: 402-424. doi: 10.2307/253568
[14]	Brockett LP, Golden LL, Jang J, et al. (2006) A comparison of neural network, statistical methods, and variable choice for life insurers. J Risk Insur 73: 397-419. doi: 10.1111/j.1539-6975.2006.00181.x
[15]	Burca AM, Batrinca G (2014) The Determinants of Financial Performance in the Romanian Insurance Market. Int J Acad Res Accounti Financ Manage Sci 4: 299-308.
[16]	Carson JM, Hoyt RE (1995) Life Insurer Financial Distress: Classification Models and Empirical Evidence. J Risk Insur 62: 764-775. doi: 10.2307/253595
[17]	Carson JM, Hoyt R (2000) Evaluating the Risk of Life Insurer Insolvency: Implications From the US for the European Union. J Multinatl Financ Manage 10: 297-314. doi: 10.1016/S1042-444X(00)00023-2
[18]	Cheng J, Weiss MA (2012) The Role of RBC, Hurricane Exposure, Bond Portfolio Duration, and Macroeconomic and Industry-wide Factors in Property—Liability Insolvency Prediction. J Risk Insur 79: 723-750. doi: 10.1111/j.1539-6975.2011.01452.x
[19]	Chen R, Wong KA (2004) The Determinants of Financial Health of Asian Insurance Companies. J Risk Insur 71: 469-499 doi: 10.1111/j.0022-4367.2004.00099.x
[20]	Choi RY, Coyner AS, Kalpathy-Cramer J, et al. (2020) Introduction to machine learning, neural networks, and deep learning. Transl Vision Sci Technol 9: 14.
[21]	Cummins JD, Grace MF, Phillips RD (1999) Regulatory Solvency Prediction in Property Liability Insurance: Risk-Based Capital, Audit Ratios, and Cash Flow Simulation. J Risk Insur 66: 417-458. doi: 10.2307/253555
[22]	Fernandes AAT, Figueiredo DB, Rocha EC, et al. (2021) Read this paper if you want to learn logistic regression. Rev Sociol Polít 28: e006.
[23]	Garson GD (2011) Multiple regression: Overview. Statnotes: Topics in Multivariate Analysis.
[24]	Garson GD (2014) Logistic regression: Binary and multinomial. Asheboro, NC.
[25]	Gross EP, Vozikis GS (2000) Prediction of Insolvency of Life Insurance through Neural Networks. In ECIS 2000 Proceedings.
[26]	Gschwind M (2007) Predicting late payments: A study in tenant behavior using data mining techniques. J Real Estate Portf Manage 13: 269-288. doi: 10.1080/10835547.2007.12089778
[27]	Hair JF, Black WC, Babin BJ, et al. (2009) Análise multivariada de dados. Bookman editora.
[28]	Hara K, Kataoka H, Satoh Y (2017) Learning spatio-temporal features with 3d residual networks for action recognition. In Proceedings of the IEEE International Conference on Computer Vision Workshops, 3154-3160.
[29]	Hosmer DW, Lemeshow S (2000) Applied Logistic Regression. New York: John Wiley & Sons.
[30]	Hsiao SH, Whang TJ (2009) A Study of Financial Insolvency Prediction Model for Life Insurers. Expert Syst Appl 36: 6100-6107. doi: 10.1016/j.eswa.2008.07.024
[31]	Huang CS, Dorsey RE, Boose MA (1994) Life Insurer Financial Distress Prediction: A Neural Network Model. J Insur Regul 13: 131-167.
[32]	Kennedy P (2008) A guide to econometrics. John Wiley & Sons.
[33]	Kohonen T, Hynninen J, Kangas J, et al. (1996) LVQ PAK: The LearningVector Quantization Program Package. Helsinki University of Technology, Technical report.
[34]	Kristóf T (2004) Mesterséges intelligencia a csõ delõ rejelzésben [Artificial Intelligence in BankruptcyPrediction]. Jö võ tanulmányok 21. Budapest: Future Theories Centre, BUESPA.
[35]	Krogh A (2008) What are artificial neural networks? Nat Biotechnol 26: 195-197.
[36]	Lee SH, Urrutia JL (1996) Analysis and Prediction of Insolvency in the Property-Liability Insurance Industry: A Comparison of Logit and Hazard Models. J Risk Insur 63: 121-130. doi: 10.2307/253520
[37]	Maas AL, Hannun AY, Ng AY (2013) Rectifier nonlinearities improve neural network acoustic models. In Proc. Icml, 30: 3.
[38]	Menard S (2002) Applied logistic regression analysis. Sage.
[39]	Mollahosseini A, Chan D, Mahoor MH (2016) Going deeper in facial expression recognition using deep neural networks. In 2016 IEEE Winter conference on applications of computer vision (WACV), IEEE, 1-10.
[40]	Negnevitsky M (2011) Artificial intelligence: a guide to intelligent systems. Third Edition.
[41]	Ohlson JA (1980) Financial ratios and the probabilistic prediction of bankruptcy. J Account Res 18: 109-131. doi: 10.2307/2490395
[42]	Ooghe H, Claus H, Sierens N, et al. (1999) International Comparison of Failure Prediction Models from Different Countries: An Empirical Analysis. Department of Corporate Finance, University of Ghent.
[43]	Park HA (2013) An introduction to logistic regression: from basic concepts to interpretation with particular attention to nursing domain. J Korean Acad Nurs 43: 154-164. doi: 10.4040/jkan.2013.43.2.154
[44]	Pinches GE, Trieschmann JS (1974) The Efficiency of Alternative Models for Solvency Surveillance in the Insurance Industry. J Risk Insur 41: 563-577. doi: 10.2307/251955
[45]	Ping W, Peng K, Gibiansky A, et al. (2017) Deep voice 3: Scaling text-to-speech with convolutional sequence learning. arXiv preprint, arXiv: 1710.07654.
[46]	Pottier SW, Sommer DW (2002) The Effectiveness of Public and Private Sector Summary Risk Measures in Predicting Insurer Insolvencies. J Financ Serv Res 21: 101-116. doi: 10.1023/A:1014325802171
[47]	Sharpe IG, Stadnik A (2007) Financial Distress in Australian General Insurers. J Risk Insur 74: 377-399. doi: 10.1111/j.1539-6975.2007.00217.x
[48]	Stephenson B, Cook D, Dixon P, et al. (2008) Binary response and logistic regression analysis. Available from: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.142.6636&rep=rep1&type=pdf.
[49]	Szandała T (2021) Review and Comparison of Commonly Used Activation Functions for Deep Neural Networks. In: Bhoi AK, Mallick PK, Liu CM, et al., Bio-inspired Neurocomputing. Springer, Singapor, 203-224.
[50]	Trieschmann JS, Pinches GE (1973) A Multivariate Model for Predicting Financially Distressed P-L Insurers. J Risk Insur 40: 327-338. doi: 10.2307/252222
[51]	Vanstone B, Finnie G (2009) An empirical methodology for developing stockmarket trading systems using artificial neural networks. Expert Syst Appl 36: 6668-6680. doi: 10.1016/j.eswa.2008.08.019
[52]	Virág M, Kristóf T (2005) Neural networks in bankruptcy prediction-A comparative study on the basis of the first Hungarian bankruptcy model. Acta Oecon 55: 403-426. doi: 10.1556/aoecon.55.2005.4.2
[53]	Walker J (1996) Methodology Application: Logistic Regression the Using CODES Data. No. HS-808 460. Available from: https://rosap.ntl.bts.gov/view/dot/4427.
[54]	Were K, Bui DT, Dick Ø B, et al. (2015) A comparative assessment of support vector regression, artificial neural networks, and random forests for predicting and mapping soil organic carbon stocks across an Afromontane landscape. Ecol Indic 52: 394-403. doi: 10.1016/j.ecolind.2014.12.028
[55]	Zekić-Sušac M, Šarlija N, Has A, et al. (2016) Predicting company growth using logistic regression and neural networks. Croat Oper Res Rev 7: 229-248. doi: 10.17535/crorr.2016.0016
[56]	Zhang L, Nielson N (2015) Solvency Analysis and Prediction in Property-Casualty Insurance: Incorporating Economic and Market Predictors. J Risk Insur 82: 97-124. doi: 10.1111/j.1539-6975.2013.12012.x

Reader Comments

Your name:*

Email:*
© 2021 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)