Life reinsurers are set to reform the use of data in the industry

Reams of newsprint have been dedicated to the technology issues faced by the property/casualty sector, but it appears that the life insurance industry has technological acne too. In real terms, this means that the technology problems that submerge the non-lifers are just as real in the life industry too.

The truth is that wherever you look in the insurance game, all that appears is an unresolved, and increasingly unstable, volcanic vista of unjoined-up data. But unlike the p/c people, some clued-up thinking in the life industry is whispered to be leading to a cross-industry initiative.

Consolidation pain

The life industry has, in some respects, an even bigger problem than the p/c sector with regards to its data. It's true that both sectors share the IT dysfunctionality and pain caused by decades of industry consolidation; dozens of legacy systems under one roof, so to speak, all storing years of data in differing technology languages in different formats. It's a toxic equation for all. But add to this that the (literally) countless millions of life policies are in constant flux - people die, people give up smoking, people take up smoking, people get diabetes, sickness, and take up dangerous sports and, of course, they get older - and it's clear that the net amount of risk also changes. If you're in the life industry, the chances are that 70% of your policies will need alteration every year; in a North American market estimated to have over 300 million policies in force, the hard cold truth is that administrators believe that more than 200 million policies change every year. Whatever the number is, it's intimidating, and certainly enough to give the most die-hard keyboarder repetitive strain injury.

Equally, though, you might argue that amending the policies is the job of the agents; spread the workload around thousands of agents and you're okay. Add to this that the life insurance companies in North America have some pretty good systems in place after years of investment and you can argue that the problem is not only contained but well sorted. This is true, but only to an extent.

The North American life industry is heavily reinsurance oriented. The latest estimates indicate that retention levels are only about 10%. In many respects, the direct writers of life policies act in a key distribution role, more so than being providers of capital. The reinsurers, too, pass on a chunk of their books to the large retrocessionaires. It is estimated that 5% of this reinsurance premium was being retroceded (a further 10 million policies). That, in total, is an awful lot of lives being reinsured, and an awful lot of data too.

Inevitably, the data which is needed for pricing and adjusting these books of business accurately is the essence of the life industry. And, of course, the precision achieved in monitoring, adjusting and amending pricing in face of the millions of changes each month or quarter will effectively define the ratios. Hence, as in the non-life sector, the data is the lifeblood, and the effectiveness of how it is pumped to the right parts of the body, at the right time, defines the health of the industry.

And that's where the problems kick in. The troubles do not lie with those at the top-end of the distribution chain. Data flows to the direct writers' systems, and they have a good grasp of their position. The panic sets in further down the line. Inevitably, the systems of the life insurers work in splendid isolation from the rest of the real world. That real world contains no more than a dozen professional life reinsurers, plus five or so retrocessionaires.

The silo nature of the direct writers' systems means that they produce their data in their own in-house formats. Each of hundreds of direct writers therefore passes a set of data in its own format to each of the reinsurers.

It's not just the data format that then needs to be mapped by each reinsurer to their own format, but the media may be paper. Or if not sent on paper, the files will be in an alien computer language.

To recap, adjustments to 200 million policies, in hundreds of different formats, in dozens of media and IT languages which need inputting accurately, fast, into 17 different companies' systems. Massive cost, massive potential for inaccuracy, slow processing, and, above all, questionable bases for underwriting.

But, you might argue, ultimately in an industry with hundreds of thousands of agents and hundreds of direct writers, who really cares about the problems of 17 or so companies down at the end of the distribution ladder?

Well, it's a pyramid, if you like, and, like any pyramid, the further down each brick is, the more important it is to the structure. The 17 companies that sit right at the bottom of the North American distribution ladder are the foundation of the industry, which could not survive without them. According to Munich Re's draft figures for 2003, those 17 companies collectively assumed $1.3trn (that's $1,300,000,000,000) of new business.

Take away their ability to trade successfully and you are looking at a lot of disappointed policyholders.

Blinkered vision

The industry has not, of course, been blind to the importance of finding a way through this data jungle. Initially the strategy for doing this was through agreement of common data standards and formats to which each company would subscribe. But the initiators had always abjectly failed, falling at the very first hurdle. The ability to agree whose standards to implement - ACORD/ANSI, SOA, TAI, or even some extant proprietary standard used by one of the bigger players in the market - was a bridge too far.

If these standards had been agreed and uniformly adopted, theoretically a free flow of life data between systems would have been possible so long as each company in the industry updated its systems to exactly the same standards, and did so uniformly without ever changing or amending the data fields (which the direct writers of life insurance do with regularity).

The cost alone of one such modernisation could run into hundreds of man months, and many companies which grew through acquisition still have multiple systems that would need such work. Not surprisingly it didn't happen.

After more than ten years of wrestling with the problem of how to share data in a single agreed structure and format, and how to implement, the reaction of the industry could have been to shrug it shoulders and live with the mess or to get smart.

For followers of the insurance industry - life, non-life, US, global - the clever money would normally load up on the 'do-nothing' side of the betting equation. And, one hears, some companies have indeed reverted to type. Others, though, are apparently moving forward fast with a lateral line of thinking.

The question that has been asked is this: what if an electronic transformer or hub was created to sit between all the companies? It would be here that the data from the life insurers would be received in its multiple structures and formats. It would restructure it and reformat it before posting it on to the systems of the reinsurers, so that they received all their data from multiple sources in one usable, analysable format.

They, in turn, would send their output on to the retrocessionaires using the same system. No company would need to adjust its own systems, but would merely connect itself once to the central utility.

Meantime, as the insurers continuously change their data layouts, a single team within the central utility could track, translate and implement these alterations just once on behalf of the two dozen or so reinsurers. This could free up resources and would make a great deal of sense for the entire industry by increasing accuracy and speed of service.

Could a 'do-it-once' integration allow companies to communicate data to and from a public utility? Could this serve the whole industry and format data to the requirements of each party?

The first real hurdle has been that, until recently, the technology to make a sensible, affordable solution happen simply did not exist. Now though, the arrival of XML and web services technologies has allowed companies to dare to believe that, if some kind of consensus could be reached, the feasibility of building such a utility for the industry would no longer be a major obstacle. It would be possible, it could be affordable, and it should be done relatively quickly.

But the question remains: if a viable solution were to be launched, would the industry support it?

Such a utility could well work, but aside from the technology, there are the human and corporate factors. Technology fatigue is rife from top-to-bottom of the industry. The cost and effort of past attempts has dulled the appetite of all but the most resilient. Corporate officers have grown intolerant of IT consultant feeding frenzies. Overcome these not insignificant barriers, and then you face the issues of self-interest.

Issues abound over who on earth would pay for a public utility.

Apportioning cost by value and benefit by industry segment, by company, is highly divisive. This, therefore, has an effect on who jumps over the cliff first, and who is prepared to hold hands in doing so. Personal career risk and corporate reputation are at stake too, against a background of zero successful initiatives in the life industry and a busted boom. And then again, some companies would argue that they have a small advantage over their competitors by managing the current chaos more efficiently, so why give up such advantage?

The introspection has legs on it still. Imagine you can persuade an entire industry to set aside self-interest, to ignore small but unimportant differentiators, to work harmoniously on the creation of a common public infrastructure.

What then about governance and control? What about ownership and pricing?

Little wonder perhaps that no solution has been implemented despite years of talk and a clear requirement. Much to the delight of the naysayers, indeed.

That, though, was then. This is now. Today's buzz humming around the life sector in North America is that a meaningful consensus has now been reached and that a core group is intent on progress. More crucially, it is said that within weeks an announcement will be made launching an industry utility. That should lay down the gauntlet to the non-life sector.

After all, if reform can happen within one industry, the regulators might be tempted to ask why it can't happen in the other. Hold tight. It looks like an exhilarating ride, whichever side of the insurance industry you sit on.