The analysis of outlying data points by robust Locally Weighted Scatter Plot Smooth: a model for the identification of problem banks

Proportional Hazards Model of Bank Failure: Evidence from USA

“…Kimmel, Booth, and Booth (2010) employ a nonparametric robust regression using a locally weighted scatter plot smooth model. Gunsel (2010) adapt the Cox proportional hazards model using logistic transformation with 23 banks from Northern Cyprus over 19 years.…”

Can statistics-based early warning systems detect problem banks before markets?

“…From this, I conclude that the market for publicly traded commercial banks is highly, but not perfectly, semi-strong form efficient. Tables Table: 1 Martin (1977) ............................................................................................................... Table 2: Pettway & Sinkey (1980) ........................................................................................... Table 3: Lane et al (1986: Kolari et al (2000) ....................................................................................................... Table 5: Kimmel et al (2010) .................................................................................................... Table 6: Analysis of Merger Activity ........................................................................................ Table 7: FDIC Bank Failures 1986Through June 2010: CRSP Data for Commercial Banks ........................................................................ Table 11: Yields .......................................................................................................................... (2006) ............................................................ Table 20: Publicly Traded Banks (1996…”

“…They find the addition of market factors improve the predictive accuracy of the model. Kimmel, Booth and Booth (2010) use a nonparametric LOESS model with a single financial ratio to analyze the performance of 46 large banks over 3 years. The model is similar to…”

“…The models studied include logit analysis as typified by Martin (1977), MDA as typified by Pettway and Sinkey (1980), Cox PHM as typified by Lane et al (1986), trait recognition as typified by Kolari et al (2002), and robust regression as typified by the LOESS regression used by Kimmel et al (2010).…”

“…losses (or high loan loss provisions) compared to total assets respectively. In validating their model, Kimmel et al (2010) conclude that the model is successful if the problematic bank is one of the 5 banks with the largest (absolute value) "bad" residuals. Model 1 is the most adept at detecting the problematic banks and is the one used in this study.…”