# Dr. Iain D. Currie

It is with great sadness that we note the passing of our long-term collaborator, Dr. Iain D. Currie, on 24th May 2022.

Iain's dedication to education was immense — whenever we hosted client seminars, it seemed that a large proportion of every audience had once been a student at one of his lectures. This also applied to me, as I first encountered Iain in 1988 when he taught an undergraduate class for Generalized Linear Models (GLMs). However, my own personal educational debt to Iain is larger than most — he introduced me to survival models in 2005, then encouraged me to co-author my first journal paper with his then-PhD student, James Kirkby, in 2006. This was followed by another joint paper, which turned into my own PhD (for which Iain was also one of the supervisors).

Iain made a great many scientific contributions, but I will highlight three that stand out for actuaries:

1. **Enhancement**. When building a statistical model, the improvement in fit from including explanatory variables \(X_1\) and \(X_2\) together can exceed the sum of the improvements in fit from including \(X_1\) and \(X_2\) on their own. Currie and Korabinski (1984) showed that this phenomenon of enhancement is actually rather common, and explained why this is the case. Long experience of fitting survival models to pensioner data sets shows that enhancement is very much the rule in mortality modelling, rather than the exception.

2. **Smoothing GLMs**. GLMs form an important class of statistical models. Currie (2013) extended the fundamental GLM fitting algorithm to simultaneously apply smoothing and identifiability constraints, rather than as separate steps. This allows a range of stochastic projection models to be fitted with a single method. Iain was too modest to accept praise for what may turn out to be a landmark extension of iteratively reweighted least squares.

3. **Value-at-risk**. At the time Solvency II was introduced throughout the European Union, there was no established means of expressing longevity trend risk in a one-year view. The now-widespread approach of doing this via simulation and recalibration was Iain's idea, and is described in Richards, Currie & Ritchie (2014).

Iain's intellectual output seemed to increase after retirement, even extending to co-authoring a book (Macdonald, Richards & Currie, 2018). Of course, there was a lot more to Iain than mathematics and statistics, including a strong artistic side. At a dinner after the presentation of one of our joint papers he surprised both me and the audience by thanking our hosts in verse specially composed for the occasion. I didn't know that mathematicians were allowed to write poetry, let alone be good at it!

I wouldn't want the preceding text to give the impression that Iain was solely focused on work, for he had a strong family orientation. When he was bestowed with the academic title of Reader he joked that it was probably recognition of his most important role at that point in time, namely as the reader of "*The Jumping Frog*" to his grandchildren. Iain was a man of many talents. He will be greatly missed by all those who had the good fortune to know him.

**References: **

Currie, I. and Korabinski, A. (1984) Some comments on bivariate regression, *Journal of the Royal Statistical Society, Series D (The Statistician)*, **33(3)**, pages 283–293, doi: 10.2307/2988232

Currie, I. D. (2013) Smoothing constrained generalized linear models with an application to the Lee-Carter model, *Statistical Modelling*, **2013;13(1)**, pages 69–93. doi:10.1177/1471082X12471373

Macdonald, A. S., Richards, S. J. and Currie, I. D. (2018) Modelling Mortality with Actuarial Applications, *Cambridge University Press*, ISBN 9781107051386, doi:10.1017/9781107051386

Richards, S. J., Currie, I. D. and Ritchie, G. P. (2014) A value-at-risk framework for longevity trend risk, *British Actuarial Journal*, **19(1)**, pages 116–139. doi:10.1017/S1357321712000451

## Previous posts

### Reheating a Cold Case

In criminal investigation, it is well known that passing time obscures the facts, making what happened more difficult to discern. Eventually, the case turns cold - unlikely to be solved unless we discover new evidence. In some ways for over a century, epidemiologists have been dealing with just such a cold case, picking through the rubble of the 1918 Influenza pandemic and trying to make sense of what they find. But as we will see, debate continues in a number of areas.

**Tags:**Filter information matrix by tag: coronavirus, Filter information matrix by tag: pandemic, Filter information matrix by tag: influenza, Filter information matrix by tag: cardiovascular

### Virus evolution

Humanity has suffered from many pandemics in the past, but the SARS-Cov-2 virus is the first to have its genome studied so extensively while the pandemic is ongoing. In a previous blog I looked at how the Delta variant displaced all other variants in the UK due to its increased infectiousness. Unfortunately, the increased infectiousness of Delta was not accompan

**Tags:**Filter information matrix by tag: coronavirus

## Add new comment