The evolution of actuarial software reflects a long history of technological innovation. As documented by Lewin et al. (Reference Lewin, Evans, Goodare and Packer1989), the development of calculating devices has influenced actuarial work since the 1600s. Early actuaries relied on compound interest tables, which appeared in Europe around 1600. By the 1700s, the invention of logarithms significantly reduced the burden of manual computation and enabled the construction of more sophisticated life tables, some of which are reproduced in Campbell-Kelly & Croarken (Reference Campbell-Kelly and Croarken2007). The introduction of electronic desk calculators in the 1950s and their widespread use in actuarial offices during the 1960s marked another major step toward automation in actuarial practice.

The historical relationship between computation and actuarial work has been examined from multiple perspectives. Yates (Reference Yates1997) explored the interaction between tabulating machinery and life insurance operations from 1890 to 1950, observing that proprietary technologies, though advantageous in the short term, sometimes led to technological dead ends as new systems emerged. Similarly, Jackson & Heabich Reference Jackson and Heabich1981) described one of the earliest implementations of actuarial calculations on a microcomputer in 1980. With hardware limited to 64K of memory and double-density 5-inch floppy disks, their experience underscored the challenges posed by inadequate documentation and the inherent complexity of actuarial notation.

As computing capabilities expanded, more specialized actuarial software began to appear. Kaas (Reference Kaas1992) reviewed the software landscape of the early 1990s, highlighting the tools for stop-loss premium calculations, modeling members of the GB2 family of distributions, and performing tasks related to reserving, rate-making, and reinsurance. Around the same time, Pryor et al. (Reference Pryor, Evans, Foley, Garner, Hilary, Skinner, Shapland, Staff and Ranser2006) conducted a survey revealing that software was already central to actuarial practice, particularly in reserving and pricing, and that 98% of respondents regularly used Excel.

The emergence of open-source platforms in the 2000s marked another transformative phase. Goulet (Reference Goulet2008) developed one of the first comprehensive R packages for actuarial modeling, providing flexible tools for data analysis and simulation. Building on this foundation, Spedicato (Reference Spedicato2013) introduced additional R packages that expanded computational capabilities and made complex actuarial methods more accessible to researchers and practitioners alike.

Although Insurance: Mathematics and Economics published the first paper related to actuarial software (Kaas, Reference Kaas1992), Annals of Actuarial Science appears to be the first actuarial journal to formally expand its scope to include actuarial software as a distinct category. In 2021, Antonio et al. (Reference Antonio, Dutang and Tsanakas2021) introduced three papers – Tseung et al. (Reference Tseung, Badescu, Fung and Lin2021), Hu et al. (Reference Hu, Murphy and O’Hagan2021), and Pesenti et al. (Reference Pesenti, Bettini, Millossovich and Tsanakas2021) – in the editorial for the special issue on Insurance Data Science. These papers presented new software tools for actuarial applications: Tseung et al. (Reference Tseung, Badescu, Fung and Lin2021) developed a Julia package LRMoE.jl for modeling insurance loss frequencies and severities using the Logit-weighted Reduced Mixture-of-Experts model; Hu et al. (Reference Hu, Murphy and O’Hagan2021) introduced an R package mvClaim for modeling multivariate insurance claim severities; and Pesenti et al. (Reference Pesenti, Bettini, Millossovich and Tsanakas2021) proposed an R package SWIM for stress testing simulation models. Although published as regular research papers, these contributions laid the groundwork for recognizing the actuarial software as a dedicated research area within the journal.

Since then, Annals of Actuarial Science has continued to expand this category, publishing a diverse set of actuarial software contributions across multiple domains. In the area of claims and loss modeling, Avanzi et al. (Reference Avanzi, Taylor and Wang2022) developed the R package SPLICE to simulate the evolution of case estimates of incurred losses over a claim’s lifetime. Pittarello et al. (Reference Pittarello, Luini and Marchione2024) introduced the Python package GEMAct, which implements the collective risk model for applications in risk costing, loss aggregation, and reserving, while Mildenhall (Reference Mildenhall2024) proposed the Python package Aggregate, featuring a fast Fourier transform (FFT)-based algorithm for approximating compound distributions. Complementing these efforts, van Jaarsveldt et al. (Reference van Jaarsveldt, Ames, Peters and Chantler2023) presented the Python package AdvEMDpy for empirical mode decomposition, with applications in both claims and mortality modeling.

In the field of investment and portfolio analysis, Marupanthorn et al. (Reference Marupanthorn, Peters, Ofosu-Hene, Nikitopoulos and Richards2024) developed the R package DivFolio for multi-period portfolio selection incorporating environmental, social, and governance (ESG) considerations, and van Jaarsveldt et al. (Reference van Jaarsveldt, Peters, Ames and Chantler2024) introduced the Python package CovRegpy for covariance regression factor models and dynamic multi-period asset allocation.

Other contributions have focused on statistical and computational tools that enhance actuarial modeling. Willame et al. (Reference Willame, Trufin and Denuit2024) developed the R package BT, which implements boosted Poisson regression trees for insurance studies, while Bladt et al. (Reference Bladt, Mueller and Yslas2025) contributed the R package matrixdist, providing tools for the statistical analysis of matrix distributions, which are useful for modeling heavy-tailed loss data, log returns, and joint claim frequency – severity structures. In mortality modeling, Ungolo et al. (Reference Ungolo, Garces, Sherris and Zhou2024) introduced the R package AffineMortality, which implements affine mortality models for parameter estimation, goodness-of-fit testing, simulation, and projection of future mortality rates.

Together, these developments trace a remarkable journey from the manual construction of actuarial tables to today’s high-performance and open-source computational environments. The papers featured in this special collection continue that trajectory, illustrating how the actuarial profession is embracing modern computational methods to enhance accuracy, efficiency, and reproducibility in an increasingly data-rich world.

Looking ahead, the future of actuarial software will likely be defined by deeper integration of artificial intelligence, high-performance computing, and open-source principles. Advances in machine learning, probabilistic programming, and cloud-based computation are creating powerful new tools for modeling complex risks at scale. At the same time, the growing emphasis on transparency, reproducibility, and collaboration is reshaping how actuarial models are developed, validated, and shared. As the profession continues to evolve, maintaining a balance between computational sophistication and interpretability will be essential to ensure that the technological progress supports better decisions, sound risk management, and sustained public trust.