Generalizing earlier works of Delbaen & Haezendonck [5] as well as [18] and [16] for given compound mixed renewal process S under a probability measure P, we characterize all those probability measures Q on the domain of P such that Q and P are progressively equivalent and S remains a compound mixed renewal process under Q with improved properties. As a consequence, we prove that any compound mixed renewal process can be converted into a compound mixed Poisson process through a change of measures. Applications related to the ruin problem and to the computation of premium calculation principles in an insurance market without arbitrage opportunities are discussed in [26] and [27], respectively.

Credit ratings are one of the primary keys that reflect the level of riskiness and reliability of corporations to meet their financial obligations. Rating agencies tend to take extended periods of time to provide new ratings and update older ones. Therefore, credit scoring assessments using artificial intelligence has gained a lot of interest in recent years. Successful machine learning methods can provide rapid analysis of credit scores while updating older ones on a daily time scale. Related studies have shown that neural networks and support vector machines outperform other techniques by providing better prediction accuracy. The purpose of this paper is two fold. First, we provide a survey and a comparative analysis of results from literature applying machine learning techniques to predict credit rating. Second, we apply ourselves four machine learning techniques deemed useful from previous studies (Bagged Decision Trees, Random Forest, Support Vector Machine and Multilayer Perceptron) to the same datasets. We evaluate the results using a 10-fold cross validation technique. The results of the experiment for the datasets chosen show superior performance for decision tree based models. In addition to the conventional accuracy measure of classifiers, we introduce a measure of accuracy based on notches called "Notch Distance" to analyze the performance of the above classifiers in the specific context of credit rating. This measure tells us how far the predictions are from the true ratings. We further compare the performance of three major rating agencies, Standard $\&$ Poors, Moody's and Fitch where we show that the difference in their ratings is comparable with the decision tree prediction versus the actual rating on the test dataset.

Volatility is a key variable in option pricing, trading and hedging strategies. The purpose of this paper is to improve the accuracy of forecasting implied volatility using an extension of genetic programming (GP) by means of dynamic training-subset selection methods. These methods manipulate the training data in order to improve the out of sample patterns fitting. When applied with the static subset selection method using a single training data sample, GP could generate forecasting models which are not adapted to some out of sample fitness cases. In order to improve the predictive accuracy of generated GP patterns, dynamic subset selection methods are introduced to the GP algorithm allowing a regular change of the training sample during evolution. Four dynamic training-subset selection methods are proposed based on random, sequential or adaptive subset selection. The latest approach uses an adaptive subset weight measuring the sample difficulty according to the fitness cases errors. Using real data from SP500 index options, these techniques are compared to the static subset selection method. Based on MSE total and percentage of non fitted observations, results show that the dynamic approach improves the forecasting performance of the generated GP models, specially those obtained from the adaptive random training subset selection method applied to the whole set of training samples.

Education has traditionally been classroom-oriented with a gradual growth of online courses in recent times. However, the outbreak of the COVID-19 pandemic has dramatically accelerated the shift to online classes. Associated with this learning format is the question: what do people think about the educational value of an online course compared to a course taken in-person in a classroom? This paper addresses the question and presents a Bayesian quantile analysis of public opinion using a nationally representative survey data from the United States. Our findings show that previous participation in online courses and full-time employment status favor the educational value of online courses. We also find that the older demographic and females have a greater propensity for online education. In contrast, highly educated individuals have a lower willingness towards online education vis-\`a-vis traditional classes. Besides, covariate effects show heterogeneity across quantiles which cannot be captured using probit or logit models.

Higher life expectancy and rapidly aging populations in developing countries, especially in the last three decades, have created the need for policymakers to introduce pension programs in developing countries. China launched the New Rural Pension Scheme (NRPS) in 2009 to ease internal demographic pressures and concerns about old-age poverty. Using data from the introduction of the NRPS in China, we estimate the effects of pension benefits, due to participation in the NRPS, on individual cognition, measured by episodic memory and intact mental status, among individuals aged 60 and above. We find large negative effects of the provision of pension benefits on cognitive functioning among the elderly. We detect the most substantial impact of the program to be on delayed recall, a measure implicated in neurobiological research as a significant predictor of the onset of dementia. We show suggestive evidence that the program leads to stronger negative impacts among the female sample. Our findings support the mental retirement hypothesis, that decreased mental activity results in atrophy of cognitive skills. We show that retirement plays a significant role in explaining cognitive decline at older ages.

We perform a large scale analysis of a list of fintech terms in (i) news and blogs in English language and (ii) professional descriptions of companies operating in many countries. The occurrence and co-occurrence of fintech terms and locutions shows a progressive evolution of the list of fintech terms in a compact and coherent set of terms used worldwide to describe fintech business activities. By using methods of complex networks that are specifically designed to deal with heterogeneous systems, our analysis of a large set of professional descriptions of companies shows that companies having fintech terms in their description present over-expressions of specific attributes of country, municipality, and economic sector. By using the approach of statistically validated networks, we detect geographical and economic over-expressions of a set of companies related to the multi-industry, geographically and economically distributed fintech movement.

In the presence of monotone information, stochastic Thiele equations describing the dynamics of state-wise prospective reserves are closely related to the classic martingale representation theorem. When the information utilized by the insurer is non-monotone, classic martingale theory does not apply. By taking an infinitesimal approach, we derive generalized stochastic Thiele equations that allow for information discarding. The results and their implication in practice are illustrated via examples where information is discarded upon and after stochastic retirement.

We study the value and the optimal strategies for a two-player zero-sum optimal stopping game with incomplete and asymmetric information. In our Bayesian set-up, the drift of the underlying diffusion process is unknown to one player (incomplete information feature), but known to the other one (asymmetric information feature). We formulate the problem and reduce it to a fully Markovian setup where the uninformed player optimises over stopping times and the informed one uses randomised stopping times in order to hide their informational advantage. Then we provide a general verification result which allows us to find the value of the game and players' optimal strategies by solving suitable quasi-variational inequalities with some non-standard constraints. Finally, we study an example with linear payoffs, in which an explicit solution of the corresponding quasi-variational inequalities can be obtained.

We experimentally evaluate the comparative performance of the winner-bid, average-bid, and loser-bid auctions for the dissolution of a partnership. The analysis of these auctions based on the empirical equilibrium refinement of Velez and Brown (2020) arXiv:1907.12408 reveals that as long as behavior satisfies weak payoff monotonicity, winner-bid and loser-bid auctions necessarily exhibit a form of bias when empirical distributions of play approximate best responses (Velez and Brown, 2020 arXiv:1905.08234). We find support for both weak payoff monotonicity and the form of bias predicted by the theory for these two auctions. Consistently with the theory, the average-bid auction does not exhibit this form of bias. It has lower efficiency that the winner-bid auction, however.

Extracting previously unknown patterns and information in time series is central to many real-world applications. In this study, we introduce a novel approach to modeling financial time series using a deep learning model. We use a Long Short-Term Memory (LSTM) network equipped with the trainable initial hidden states. By learning to reconstruct time series, the proposed model can represent high-dimensional time series data with its parameters. An experiment with the Korean stock market data showed that the model was able to capture the relative similarity between a large number of stock prices in its latent space. Besides, the model was also able to predict the future stock trends from the latent space. The proposed method can help to identify relationships among many time series, and it could be applied to financial applications, such as optimizing the investment portfolios.

We propose a stochastic model allowing property and casualty insurers with multiple business lines to measure their liabilities for incurred claims risk and calculate associated capital requirements. Our model includes many desirable features which enable reproducing empirical properties of loss ratio dynamics. For instance, our model integrates a double generalized linear model relying on accident semester and development lag effects to represent both the mean and dispersion of loss ratio distributions, an autocorrelation structure between loss ratios of the various development lags, and a hierarchical copula model driving the dependence across the various business lines. The model allows for a joint simulation of loss triangles and the quantification of the overall portfolio risk through risk measures. Consequently, a diversification benefit associated to the economic capital requirements can be measured, in accordance with IFRS 17 standards which allow for the recognition of such benefit. The allocation of capital across business lines based on the Euler allocation principle is then illustrated. The implementation of our model is performed by estimating its parameters based on a car insurance data obtained from the General Insurance Statistical Agency (GISA), and by conducting numerical simulations whose results are then presented.

In this paper we continue the research of our recent interest rate tree model called Zero Black-Derman-Toy (ZBDT) model, which includes the possibility of a jump at each step to a practically zero interest rate. This approach allows to better match to risk of financial slowdown caused by catastrophic events. We present how to valuate a wide range of financial derivatives for such a model. The classical Black-Derman-Toy (BDT) model and novel ZBDT model are described and analogies in their calibration methodology are established. Finally two cases of applications of the novel ZBDT model are introduced. The first of them is the hypothetical case of an S-shape term structure and decreasing volatility of yields. The second case is an application of the ZBDT model in the structure of United States sovereign bonds in the current $2020$ economic slowdown caused by the Coronavirus pandemic. The objective of this study is to understand the differences presented by the valuation in both models for different derivatives.