The OzEdge: A Brief Excursion into GeneraI Insurance History

Geoff Atkins recently contributed an article, The Aussie Advantage: How Our Actuaries Pioneered Insurance Reserving , whose thesis was that Australian actuaries tended to use loss reserving methodologies that are superior to those of our British cousins.

Perhaps I shall keep my own counsel on that opinion and confine myself to a discussion of the early general insurance (GI) developments mentioned by Geoff, and some further remarks on the conditions that stimulate technical development in the industry.

Livin’ in the 70s

As I read Geoff’s article while relaxing over morning coffee, I was jolted out of my early morning torpor by the mention of my name. It was very generous of Geoff to do this, but it caused an immediate flash of guilt because there were several others who made substantial contributions in those far-off days, and sharing of credit is appropriate.

In fact, I had no involvement of any substance in GI until 1975. As an academic, I had flirted with risk theory in the prior years, but it was in 1975 that I was suddenly thrust into the real world.

With a position in the UK Government Actuary’s Department, seconded to the Department of Trade, I was placed in a small group under the tutelage of the famous Professor R E Beard, and with the task of endeavouring to extract reserving value from recently prescribed run-off returns (claim triangles).

That’s a whole other story for another day, but one of its outcomes was to encourage my hope to participate in GI on my return to Australia in 1976. I discovered on my return that a fledgling actuarial involvement in GI did indeed exist. In consequence, I was able to ride comfortably on the coattails of predecessors who had hacked their way into the GI jungle, as it then was.

Here, I come to individual names, and I would single out Richard Cumpston and Roger Sawkins as the earliest and most intrepid explorers. Geoff mentions the “PP” methods, and indeed, to the best of my knowledge, the PPCI was first documented by Richard, and the PPCF and PPAC by Roger.

There were others. I mention particularly Bob Buchanan (of GI textbook fame), working with the Australian Government Actuary. David Hart (also of textbook fame), as well as Jim Gould, Trevor Matthews, John Ryder, John Trowbridge and Brent Walker. I have undoubtedly omitted others and, to them, I apologise.

These were times of invention

The new Insurance Act 1973 required the submission of Forms 11, which were the tabulations that provided the raw data from which claim triangles were constructed, and so data was accumulating.

Someone formed a Working Group, which I think was called simply the General Insurance Working Group, and it met frequently, at least monthly, to discuss matters of mutual interest, including technical matters. Richard Cumpston initiated a General Insurance Bulletin to which short-written articles were submitted for circulation.

Trevor Matthews organised the first General Insurance Seminar in Thredbo in 1978. This was organised by the participants, without assistance from the Institute, and spawned the sequence of GISs that persisted for more than 30 years.

They were times of invention because actuarial methodology was being built from ground up. There was no course of instruction in Australia, as GI was as yet not a recognised actuarial discipline. The UK situation was similar.

Of course, the Casualty Actuarial Society had been active for many years but, to my knowledge, had no Australian members, and no-one here had access to their Proceedings. To an extent, this was beneficial, as it required all thought to be independent. There was certainly no risk of slavishly following a received orthodoxy.

In retrospect, this resulted in some re-invention of the wheel. For example, some very influential papers appeared in the PCAS in the 70s (Bornhuetter and Ferguson, 1972; Fisher and Lange, 1973; Berquist and Sherman, 1977), and these remained unknown in Australia for some years. However, on the whole, I think the element of re-invention was not great, and there were some useful home-grown products, such as the PPs.

Punctuated equilibrium

Let me turn to a consideration of when and why the industry experienced technical development. I’ll commence by proposing a thesis that methodological progress (I am speaking here only of GI) proceeds in fits and starts and (this is the kicker) spikes in times of crisis or stress.

The crisis might be litigation; the stress might be the semi-adversarial debate over pricing of compulsory insurance products (which rely on models of the reserving type).

This calls to mind Stephen Jay Gould’s theory of punctuated equilibrium evolution, the theory that evolutionary change proceeds, for the most part, at a subdued rate, but with occasional punctuations of activity, limited in time but of high intensity.

In loss reserving, the punctuations are, or at least are initiated by, the times of crisis. These almost always correspond to events in which money changes hands between parties. This changes the whole complexion of the loss-reserving endeavour.

Indeed, one might say that there are two types of reserving, which we might characterise as the “Business Process” and “Demand for Accuracy” types. An example of the first of these might be quarterly reserving for internal management accounts; an example of the second might be reserving for the purpose of establishing net assets in an M&A transaction.

The stakes are higher in the second case as, here, any inaccuracy will create a winning party and a losing party to the transaction. In the first case, the penalty for inaccuracy is lower, perhaps a couple of percentage points in booked solvency. No money changes hands.

In the Demand-for-Accuracy case, the reserving actuary will certainly face interrogation, possibly hostile. This might consist of cross-examination in court. A less hostile environment might be that surrounding the pricing of compulsory insurance products mentioned above. These discussions are usually conducted with civility, but not necessarily without challenge.

The essential point is that the actuary’s analysis will be challenged down to the level of fine detail. To state that one has used this or that off-the-shelf method (complete with acronym), together with one’s professional judgment, may not constitute an adequate defence, and professional judgement may be regarded as a shibboleth.

The other party will naturally question the extent to which the actuarial model matches the data and whether that model’s structure is likely to produce a reliable forecast. This takes one to the mechanics of the model and whether or not these align well with the claim process. In the event of any failure of the actuary’s model to meet these criteria, it is likely to be viewed as dubious, or even inadequate.

Conditions such as these can create a pressure cooker for innovation. New forms of model or other fundamental changes of technique may be stimulated. These are the punctuations of equilibrium.

Seismic shifts are less likely in a Business-Process environment. This is not to say that the actuary does not seek accuracy in this case, but priority may be assigned to other factors, such as smoothness of progression from period to period, timeliness of reporting, and so on.