[PROTCT – Protector Forsikring] Insider Buying; Reserve Hits; Low Cost vs. Cost Advantage

In the insurance industry, growth stories are to be treated with caution.  When I come across an underwriter, like Protector Fosikring, that is profitably growing premiums by 20%/year when its peer group is growing 1-4%/year, the first flag I raise is colored red, not green.  More likely than not, it is underpricing risk or accepting risk that its competitors are wise enough to avoid. 

As I wrote in my post on Chubb:

The investment income you immediately earn by attracting premium flows today is near certain; the existential consequences of doing so at any price are not.  Given the temptation to roll the dice, underwrite for volume, and pray for good weather, a capable operator with capital allocation discipline is table stakes for withstanding the inevitable natural disaster and eeking out excess returns…in a base case, an insurer typically earns the returns of a levered high-grade bond fund (0% underwriting margin + a/t portfolio yield x invested assets / equity)”

This doesn’t have to be true.  An operator can offer insurance at a discounted price and sustain excess returns if it has a low cost advantage.  The canonical example here is GEICO, which offered auto insurance to low risk government employees, replaced agents and branch offices with direct marketing, and scaled the latter into unassailable cost leadership.  From 1945 to 1950, GEICO compounded its written premiums by 37%.

Like GEICO, Protector is growing premiums way faster than its peers:

But unlike GEICO, Protector relies on third party distribution and competes with incumbents for the same risk pools (mostly short-tail motor and property risk for municipalities and SMEs in Norway, Sweden, Denmark, and increasingly the UK).  Nevertheless, with a gross expense ratio that is roughly half that of peers, it seems that Protector has something interesting going on:

[Gross expense ratio = operating expenses / gross premiums earned]

[Differences in business mix will drive differences in expense and loss ratios.  But even adjusting for this – for instance, isolating Gjensidige’s commercial business in Norway – Protector’s expense ratios are still lower]

But having a superior expense ratio is not the same thing as having a cost advantage.  Protector’s management team points to cultural factors as driving the gap, taking pains to cultivate the narrative of a scrappy, irreverent underdog taking on the big incumbents – “Protector will be the challenger”.  That Protector’s growth has been entirely organic over its 14-year history (aligning values and behaviors across acquired companies who have evolved along different paths is complicated); that its bombastic CEO has been with the company from the very start; and that he has made a point of calling out the cultural challenges in expanding to the UK (where org structures are more hierarchical than they are in Norway) when most others would have talked about tactical or market obstacles, somehow makes the culture argument more believable and salient to me.  But even so, cost advantages almost always have their roots in scale economies. Cost discipline may reinforce scale economies – as it does in the case of Ryanair, Fastenal, GEICO, and Wal-Mart – but it is no substitute.

Another possible, more readily available reason that Protector has lower operating expenses than peers is that it relies exclusively on third party brokers for premiums, whereas incumbents invest in proprietary distribution channels.  Gjensidige (the largest P&C insurer in Norway with 29% share of commercial premiums vs. 6% for Protector), for instance, only sources 20% of its Norwegian commercial premiums through brokers. 

[Nov 12, 2019: the above claim is inaccurate.

i recently met with Protector’s CEO and asked him what he thought of my hypothesis. he told me it was wrong. the in-house distribution that incumbents have built out relates to their consumer business, where protector does not participate. for commercial and public risk, the categories of risk relevant to Protector, the entire industry sources through brokers.

so, if protector’s distribution does not differ from that of incumbents, what explains the lower expense ratio? it seems to boil down to two things.

the first is that protector leverages more modern IT systems than incumbent peers, who run outdated cobol-based backends and whose processes are still significantly paper-based (CEO Sverre Bjerkeli, who has a background in IT, made technology a priority during the company’s early days). protector’s superior systems allow it to make more thorough use of data in underwriting risk compared to peers, who depend somewhat more on finger-in-the-air risk assessments from expensive teams of people.

(aside: at around the time of protector’s founding, the Norwegian regulators, in order to spark competition in its oligopolistic insurance industry, granted big tax breaks to protector and other new entrants who invested in modern IT systems, tax breaks that are no longer available to newcomers).

the second factor relates to a culture of cost discipline and high service levels. the “culture arbitrage” point sounds squishy, but i think it’s real.

because protector operates with lower expense ratios, it can afford to underwrite the same risk at lower prices and provide superior service to its clients. brokers like this and direct volumes to protector…hence protector has been growing premiums by high-teens/year vs. low/mid-single digits for incumbents.

the incumbents in this space are well-run but stodgy institutions focused on maximizing underwriting profits, limiting volatility, and protecting dividends in any given year. Protector, on the other hand, focused on maximizing the total stream of profits – underwriting + investment – over many years.

It’s worth asking, what’s harder…building your own dedicated distribution or relying on brokers?  Surely these disciplined Nordic incumbents, who as a group post some of the most impressive underwriting margins I have ever seen, are not needlessly burdening their cost structures and overlooking an obvious margin opportunity just for kicks.  Nor does Protector have privileged access to the broker channel.  In the 4-5 years leading up to 2010, a bunch of Nordic insurance upstarts breached the market by tapping third party brokers, grabbing share from incumbents by offering aggressively priced policies.

Why build out proprietary distribution when a faster and easier route is readily available?  Better to forgo this investment and just use the cost savings to undercut competitors on price, build market share, and scale operating costs, right?  But of course, there are trade-offs to relying on brokers, who represent the client’s interest, not that of the carrier, and who are incented predominantly on volume, not underwriting performance.  An insurer who goes this route may have limited purview of the underwriting risks it is assuming, which may lead to higher loss ratios.

And indeed, Protector’s combined ratios looks far less impressive:

[Net combined ratio = expense ratio + net loss ratio…and so, based on Protector’s middling combined ratio and its superior expense ratio, we can infer that the company has a much higher than average loss ratio. 

A net loss ratio (or “claims ratios”) is calculated by dividing net underwriting claims (gross claims less the reinsured portion) incurred into net premiums earned].

When I compare combined and loss ratios by country and by risk exposure (where I can), the disparities between Protector and its peers are consistent with the above table and in most cases, even more gaping.  For instance, the commercial segments of Gjensidige and Tryg have loss ratios well below those of Protector; ditto for If (at least from 2015 to 2017, the years where I could find disclosure).

Likewise, Protector’s 75%-80% loss ratio in Sweden (its second largest geography after Norway, accounting for ~1/4 of premiums) are ~5 to 15 points worse than those of Tryg and If, and its loss ratio is Denmark is ~25 points above Tryg’s.