PUBLICATION

Refereed Publications

[25] Quantile-based Interpretable Neural Network Models: Mortality Forecasting and Actuarial Simulations

Qiao, Y.**, Zhang, J., Zhu, W., Wang, C.-W. (2026)

Insurance: Mathematics and Economics, forthcoming

Interpretable neural networks estimate mortality quantiles while retaining the structure of classic stochastic mortality models. The framework supports stochastic mortality simulation and longevity-linked derivative pricing.

DOISSRNPDF
  • Mortality, Health & Longevity
  • AI & Machine Learning in Insurance and Finance

[24] Air Quality, Health, and Economic Effects of Land Use and Land Cover changes in Southeast Asia in the 21st Century: A Modelling Study

Gu Y., Fang T., Xu Y., Duan X., Lee S. H. J. T., Zhu W., Zeng Y., Yim S. H. L. (2026)

Lancet Planetary Health, 10(4): 101457

This modelling study connects land-use and land-cover changes in Southeast Asia with air pollution, health outcomes and economic costs. It estimates excess deaths and associated losses through both biogeophysical and biogeochemical pathways.

DOISSRNPDF
  • Mortality, Health & Longevity
  • Climate Risk, Agriculture & Sustainability

[23] Enduring Impacts of El Niño on Life Expectancy in Past and Future Climates

Xu, Y.*, Zhu, W., Samanta, D.**, and Horton, B. P., (2026)

Nature Climate Change, 16, 148-154

The study quantifies persistent life-expectancy losses from El Niño across high-income Pacific Rim countries, estimating monetary-equivalent losses of US$2.6 trillion and US$4.7 trillion for the 1982–83 and 1997–98 events. Under moderate emissions pathways, projected losses reach US$35 trillion by 2100.

DOISSRNPDF
  • Mortality, Health & Longevity
  • Climate Risk, Agriculture & Sustainability
Persistence of El Niño effects on life expectancy.
Persistence of El Niño effects on life expectancy.
Economic cost of El Niño induced life expectancy losses.
Economic cost of El Niño induced life expectancy losses.

[22] Index-based Insurance Design for Climate and Weather Risk Management: A Review

Zhu, W., Zhang, J., Porth, L., and Tan, K. S. (2025)

in Springer Actuarial, Quantitative Risk Management and Insurance in Agricultural Business, edited by A. Hirbod, P. Liu and S. Wang

This chapter reviews the actuarial foundations, empirical evidence and practical design of index insurance for climate and weather risks. It also discusses potential applications of artificial intelligence and blockchain.

SourcePDF
  • Climate Risk, Agriculture & Sustainability
  • Financial Markets, Hedging & Risk Transfer

[21] Modeling Mortality with Kernel Principal Component Analysis (KPCA) Method

W*, S., Chen*, A., Xu*, Y., Pan, G., and Zhu, W. (2024)

Annals of Actuarial Science, 18(3):626-643

A kernel principal component extension of the Lee–Carter model captures nonlinear patterns in mortality data. Empirical tests examine forecasting performance and uncertainty, including during the COVID-19 pandemic.

DOI
  • Mortality, Health & Longevity
  • AI & Machine Learning in Insurance and Finance

[20] Epidemic Financing Facilities: Pandemic Bonds and Endemic Swaps

Huang*, S., Tan, K. S., Zhang, J. and Zhu, W. (2024)

North American Actuarial Journal, 28(3):626-657

The paper develops pandemic bonds and endemic swaps to finance severe pandemic outbreaks and recurrent epidemics. Coronavirus and dengue examples illustrate pricing with epidemiological and actuarial models.

DOISSRNPDF
  • Mortality, Health & Longevity
  • Financial Markets, Hedging & Risk Transfer

[18] Managing Weather Risk with a Neural Network-Based Index Insurance

Chen, Z., Lu, Y., Zhang, J., and Zhu, W. (2024)

Management Science, 70(7):4306-4327

A neural network embedded in a utility-maximization framework designs weather index insurance and endogenously determines premiums and demand. The approach captures nonlinear weather–loss relationships to reduce basis risk and improve welfare.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability
  • AI & Machine Learning in Insurance and Finance
  • Financial Markets, Hedging & Risk Transfer
Payoffs of index insurance and actual losses. The plot shows an NN-based index insurance contract payoffs.

[17] Machine Learning in Long-term Mortality Forecasting

Qiao, Y.*, Wang, C., and Zhu, W. (2023)

The Geneva Papers on Risk and Insurance - Issues and Practice, 49:340-362

This mortality forecasting framework combines neighbouring prediction, model ensembling and tree boosting. Tests using the Human Mortality Database examine long-horizon prediction and the benefits of pooling information across populations.

DOISSRNPDF
  • Mortality, Health & Longevity
  • AI & Machine Learning in Insurance and Finance

[16] Cyber Risk Modeling: A Discrete Multivariate Count Process Approach

Lu, Y., Zhang. J., and Zhu, W. (2023)

Scandinavian Actuarial Journal, 2024(6), 625-655

A count-based frequency–severity framework models cyber events and data breaches, accounting for dependence, zero inflation and heavy tails. The estimation approach is applied to the Privacy Rights Clearinghouse dataset.

DOISSRNPDF
  • Financial Markets, Hedging & Risk Transfer

[15] Storm CAT Bond: Modeling and Valuation

Huang*, S., Zhang, J. and Zhu, W. (2023)

North American Actuarial Journal. 28(4): 718-743

The paper models and prices storm catastrophe bonds using loss distributions and spatial–temporal dependence. Florida county-level data illustrate how the proposed bonds can hedge reinsurers’ storm losses.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability
  • Financial Markets, Hedging & Risk Transfer

[14] A Deep Factor Model for Crop Yield Forecasting and Insurance Ratemaking

Zhu, W. (2023)

North American Actuarial Journal. 28(1): 57-72

A deep factor model combines an autoencoder and convolutional neural networks to forecast crop yields. An application to Iowa data evaluates forecasting accuracy and implications for crop insurance ratemaking.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability
  • AI & Machine Learning in Insurance and Finance

[13] Neighbouring Prediction for Mortality

Wang C.-W., Zhang, J., and Zhu, W. (2021)

ASTIN Bulletin, the Journal of the IAA, 51(3), 689-718

A neighbouring prediction framework treats mortality tables as two-dimensional images, enabling convolutional neural networks to learn intricate nonlinear patterns across ages and years. Evaluations cover 41 countries and regions, with an extension to multiple populations.

DOISSRNPDFOnline AppendixDemo code
  • Mortality, Health & Longevity
  • AI & Machine Learning in Insurance and Finance
Neighbouring prediction for m(x,t).

[10] Dynamic Bayesian Ratemaking: A Markov Chain Approximation Approach

Li, H., Lu, Y. and Zhu, W. (2021)

North American Actuarial Journal, 25(2), 186-205

A hidden Markov chain approximation makes Bayesian premiums tractable in insurance models with dynamic unobserved policyholder risk. Numerical and real-data examples evaluate its accuracy and pricing flexibility.

DOISSRNPDF
  • Financial Markets, Hedging & Risk Transfer

[9] The Design of Weather Index Insurance Using Principal Component Regression and Partial Lease Squares Regression: The Case of Forage Crops

Boyd, M., Porth, C. B., Porth, L., Tan, K. S., Wang*, S., and Zhu, W. (2020)

North American Actuarial Journal, 24(3), 355-369

The study compares principal component regression and partial least squares regression for constructing multivariable weather insurance indices. Ontario forage data illustrate how dimension reduction can address multicollinearity and basis risk.

DOI
  • Climate Risk, Agriculture & Sustainability

[8] Remote Sensing Applications for Insurance: An Example of Predicting Forage Crop Loss in the Presence of Systemic Weather Risk

Porth C. B., Porth, L., Zhu, W., Boyd, M., Tan, K. S., and Liu, K. (2020)

North American Actuarial Journal, 24(2), 333-354

The paper compares thirteen pasture production indices, including satellite-derived vegetation and biophysical measures, for predicting forage losses. Alberta pasture observations provide a validation testbed for remote-sensing applications in insurance.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability

[6] Agricultural Insurance Ratemaking: Development of a New Premium Principle

Zhu, W., Tan, K. S., and Porth, L. (2019)

North American Actuarial Journal, 23(4), 512-534

A multivariate weighted premium principle reweights historical agricultural losses using external risk information. Manitoba reinsurance data illustrate its pricing implications and allocation of loadings across risk layers.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability

[4] Spatial Dependence and Aggregation in Hedging Systemic Weather Risk: A Lévy Subordinated Hierarchical Archimedean Copulas (LSHAC) Approach

Zhu, W., Tan, K. S., Porth, L., and Wang, C.-W. (2018)

ASTIN Bulletin, the Journal of the IAA, 48 (2), 779-815

A Lévy subordinated hierarchical Archimedean copula (LSHAC) model captures spatial dependence in weather risks to support agricultural hedging. The analysis examines basis risk, downside protection and the effects of spatial aggregation.

DOISSRNPDF
  • Climate Risk, Agriculture & Sustainability
  • Financial Markets, Hedging & Risk Transfer

[3] Modeling Multi-population Longevity Risk with Mortality Dependence: A Lévy Subordinated Hierarchical Archimedean Copulas (LSHAC) Approach

Zhu, W., Tan, K. S., and Wang, C.-W. (2017)

Journal of Risk and Insurance, 84, 477-493

A Lévy subordinated hierarchical Archimedean copula (LSHAC) framework models mortality dependence across countries and is evaluated using data from thirteen countries. A survival index swap example illustrates the role of dependence in longevity risk and population basis risk.

DOISSRNPDFOnline Appendix
  • Mortality, Health & Longevity
  • Financial Markets, Hedging & Risk Transfer
Illustrative Grouping Results of 13 Countries.

[1] A credibility-based Erlang mixture model for pricing crop reinsurance

Porth, L., Zhu, W., Tan, K. S. (2014)

Agricultural Finance Review, 74(2), 162 - 187

An Erlang mixture model and a modified credibility approach are used to price crop reinsurance. Canadian crop insurance data are used to assess tail fit and liability-weighted loss-cost experience.

DOIPDF
  • Climate Risk, Agriculture & Sustainability

Other Publications (e.g., Technical Reports, etc.)

[13] Cyclones in the Asia Pacific: A View to Future Risks

Timothy A. Shaw, Erandani Lakshani, Adam D. Switzer, Yanbin Xu, Shinichi Kamiya, Wenjun Zhu, 2026.

Global Asia Insurance Partnership (GAIP) Living Lab Report. June 2026

A technical report on cyclone risk in the Asia Pacific and its future outlook.

Source
  • Climate Risk, Agriculture & Sustainability

[12] Climate-Adjusted Mortality Modelling in Asia

C. K. Liang, Zhu, W., S. Kamiya, C. Nie, Mar 2026.

Global Asia Insurance Partnership (GAIP) Living Lab Report. March 2026

The report proposes a Climate Risk Adjustment Factor (CRAF) to incorporate climate effects into mortality tables, supporting insurers’ climate stress testing and forward-looking risk assessment in Asia.

Source
  • Mortality, Health & Longevity
  • Climate Risk, Agriculture & Sustainability

[10] Inland Flood Risk Modeling in Southeast Asia

Xu, Y., Zhu, W., and Kamiya, S., 2025.

Global Asia Insurance Partnership (GAIP) Living Lab Report. May 2025

A technical report on modelling inland flood risk in Southeast Asia.

Source
  • Climate Risk, Agriculture & Sustainability

Note: * denotes graduate students; ** denotes research fellows.