Delta Boosting Machine with Application to General Insurance

In this article, we introduce Delta Boosting (DB) as a new member of the boosting family. Similar to the popular Gradient Boosting (GB), this new member is presented as a forward stagewise additive model that attempts to reduce the loss at each iteration by sequentially fitting a simple base learner to complement the running predictions. Instead of relying on the negative gradient, as is the case for GB, DB adopts a new measure called delta as the basis. Delta is defined as the loss minimizer at an observation level. We also show that DB is the optimal boosting member for a wide range of loss functions. The optimality is a consequence of DB solving for the split and adjustment simultaneously to maximize loss reduction at each iteration. In addition, we introduce an asymptotic version of DB that works well for all twice-differentiable strictly convex loss functions. This asymptotic behavior does not depend on the number of observations, but rather on a high number of iterations that can be augmented through common regularization techniques. We show that the basis in the asymptotic extension differs from the basis in GB only by a multiple of the second derivative of the log-likelihood. The multiple is considered to be a correction factor, one that corrects the bias toward the observations with high second derivatives in GB. When negative log-likelihood is used as the loss function, this correction can be interpreted as a credibility adjustment for the process variance. Simulation studies and real data application we conducted suggest that DB is a significant improvement over GB. The performance of the asymptotic version is less dramatic, but the improvement is still compelling. Like GB, DB provides a high transparency to users, and we can review the marginal influence of variables through relative importance charts and the partial dependence plots. We can also assess the overall model performance through evaluating the losses, lifts, and double lifts on the holdout sample.     

“Analysis of Incremental Returns of Canadian Mutual Funds”, J. F. Carrière; C. F. Hill,April 2001

Abstract

The determination and allocation of economic capital is important for pricing, risk management, and related insurer financial decision making. This paper considers the allocation of economic capital to lines of business in insurance. We show how to derive closed-form results for the complete markets, arbitrage-free allocation of the insurer default option value, or insolvency exchange option, to lines of business for an insurer balance sheet. We assume that individual lines of business and the surplus ratio are joint log-normal although the method we adopt allows other assumptions. The allocation of the default option value is required for fair pricing in the multiline insurer. We discuss and illustrate other methods of capital allocation, including Myers-Read, and give numerical examples for the capital allocation of the default option value based on explicit payoffs by line.     

Localized poaching and skills shortages of manufacturing employees among MNEs in China

Using a mixed-method approach, we examine the experiences and responses among foreign-invested enterprises (FIEs) of employee poaching and skills shortages in a prominent Chinese electronics manufacturing hub (Suzhou Industrial Park). This highlights strategic challenges MNEs face in Asia Pacific economies as more advanced subsidiaries outgrow their earlier low-cost, low-skill vocations. In examining poaching empirically, we also extend the field by introducing a local labour market perspective into the study of HRM in China. Third, we advance theorization of employee poaching beyond mainstream Western approaches to capture the interactions between internal and local labour markets that increasingly mark FIE manufacturing in fast developing economies.     

Swedish engineering employers: The search for industrial peace in the absence of centralised collective bargaining

The Swedish engineering employers’ association recently shifted away from the dramatic decentralisation of bargaining it had pioneered. Behind this shift were employer desires for better guarantees of industrial peace. The authors conclude by stressing the importance of the relationships between employer strategy, historical contingency and debates on bargaining structure.