| dc.identifier.uri | http://hdl.handle.net/11401/77415 | |
| dc.description.sponsorship | This work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree. | en_US |
| dc.format | Monograph | |
| dc.format.medium | Electronic Resource | en_US |
| dc.language.iso | en_US | |
| dc.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dc.type | Dissertation | |
| dcterms.abstract | In chapter 1, we analyze the concept of credibility in two generalized count models: Mittag Leffler and Weibull count models which can handle both underdispersion and overdispersion in count data and nest the commonly used Poisson model as a special case. A correct specification of the model for determining individual risk premium is important since without a proper pricing mechanism one is simply not competitive in the insurance industry. We find evidence using data from Danish Insurance Company that the simple Poisson model can set the credibility weight to one as a result of large heterogeneity among policyholders and thus, breaks down the credibility model. We propose parametric estimators for the structural parameters in the credibility formula using the mean and variance of the assumed distributions and a maximum likelihood over a collective data. As an example, we show that the proposed parameters from Mittag Leffler provide weights that are consistent with the idea of credibility, while a simulation study is carried out to investigate the stability of the maximum likelihood estimates from the Weibull count model. Finally, we extend the analyses to multidimensional lines and show how our approach can be adopted to cross selling application of the credibility model. Chapter 2 shows that multinomial logit model (MNL), previously found to be a good tool in predicting performance in the Indian stock market, can only be implemented at the industry level but not the entire U.S. market if the appropriate financial ratios are selected. When prediction is of ultimate importance, we design a multilayer perceptron (MLP) neural network to show that as an alternative tool for the U.S. market with overall average accuracy rate at about 57.6% and 59.4% in training and testing samples respectively. The results obtained reveal a firm’s ability to pay its short term obligations and its efficient use of cash in generating sales revenue are highly predictive and important when predicting performance in a probabilistic framework. | |
| dcterms.abstract | In chapter 1, we analyze the concept of credibility in two generalized count models: Mittag Leffler and Weibull count models which can handle both underdispersion and overdispersion in count data and nest the commonly used Poisson model as a special case. A correct specification of the model for determining individual risk premium is important since without a proper pricing mechanism one is simply not competitive in the insurance industry. We find evidence using data from Danish Insurance Company that the simple Poisson model can set the credibility weight to one as a result of large heterogeneity among policyholders and thus, breaks down the credibility model. We propose parametric estimators for the structural parameters in the credibility formula using the mean and variance of the assumed distributions and a maximum likelihood over a collective data. As an example, we show that the proposed parameters from Mittag Leffler provide weights that are consistent with the idea of credibility, while a simulation study is carried out to investigate the stability of the maximum likelihood estimates from the Weibull count model. Finally, we extend the analyses to multidimensional lines and show how our approach can be adopted to cross selling application of the credibility model. Chapter 2 shows that multinomial logit model (MNL), previously found to be a good tool in predicting performance in the Indian stock market, can only be implemented at the industry level but not the entire U.S. market if the appropriate financial ratios are selected. When prediction is of ultimate importance, we design a multilayer perceptron (MLP) neural network to show that as an alternative tool for the U.S. market with overall average accuracy rate at about 57.6% and 59.4% in training and testing samples respectively. The results obtained reveal a firm’s ability to pay its short term obligations and its efficient use of cash in generating sales revenue are highly predictive and important when predicting performance in a probabilistic framework. | |
| dcterms.available | 2017-09-20T16:52:39Z | |
| dcterms.contributor | Montgomery, Mark R | en_US |
| dcterms.contributor | Centorrino, Samuele | en_US |
| dcterms.contributor | Zhao, Jake | en_US |
| dcterms.contributor | Rachev, Svetlozar (Zari) | en_US |
| dcterms.contributor | Kim, Aaron. | en_US |
| dcterms.creator | Asamoah, Kwadwo | |
| dcterms.dateAccepted | 2017-09-20T16:52:39Z | |
| dcterms.dateSubmitted | 2017-09-20T16:52:39Z | |
| dcterms.description | Department of Economics. | en_US |
| dcterms.extent | 106 pg. | en_US |
| dcterms.format | Monograph | |
| dcterms.format | Application/PDF | en_US |
| dcterms.identifier | http://hdl.handle.net/11401/77415 | |
| dcterms.issued | 2015-12-01 | |
| dcterms.language | en_US | |
| dcterms.provenance | Made available in DSpace on 2017-09-20T16:52:39Z (GMT). No. of bitstreams: 1
Asamoah_grad.sunysb_0771E_12625.pdf: 2700705 bytes, checksum: c2d2c2c4a19bc9d8073dcfbf706d28ed (MD5)
Previous issue date: 1 | en |
| dcterms.publisher | The Graduate School, Stony Brook University: Stony Brook, NY. | |
| dcterms.subject | Economics | |
| dcterms.subject | Artificial neural network, Credibility, Financial ratios, Mittag-Leffler count Model, Multinomial logit, Weibull count model | |
| dcterms.title | Two Essays on Actuarial and Financial Econometrics | |
| dcterms.type | Dissertation | |
