A missing insurance market

By Eric Crampton 25/08/2015 4


I wonder if I’d be the only customer for this one.

Imagine a bolt-on to your existing home-owners’ insurance. It specifies that, in the event of a substantial earthquake,* the insurance company immediately buys your house from you for a pre-specified price. No inspections, no claims adjustment, nothing. Big enough quake, they own your house as-is where-is. Maybe you could set it as an option for the policy holder, maybe you could set it as an automatic thing. Take the option, and you have a big deposit in your account to let you start over somewhere else. The insurance company then has, say, six months after the roads to your house are passable by truck and the port or roads out of town are open to get the contents packed into shipping containers and delivered to the nearest functioning port facility.

Advantages for the insurer:

  • No messing around with finicky owners. The insurer runs the repairs that they think are necessary to on-sell the house afterwards with no hassles. The timing of repairs is entirely up to them. They can contract with larger scale firms to run rebuilds over larger parcels if they want too. Owners are often picky about who they want as builders (we were!). The insurer owning the house has no worries about whether an owner is trying to fix things to as-new or whether he’s trying to correct pre-existing damage.
  • The insurance on-sold home would be a sure-thing for future policies: everyone would know that it was fixed to insurer standards, so there would be no issues about the house’s future insurability.
  • Instead of a bunch of fragmented owners arguing over things like red-zoning, with flow-on consequence for the insurer, the insurer gets to have those conversations with the government.
Advantages for the insured:
  • A certain fast payout for anybody who wants to flee. No hassles, no arguments, no waiting, no living in limbo.
This seems an easy product to provide. I bet there’d be a lot of takers – or at least anyone who’s experienced Christchurch would give it a good look.

I would want a clause in there that this part of the insurance contract – either terms or premiums – cannot be changed by the insurer except with two-years’ notice: you wouldn’t want foreshocks leading to policy cancellation.

* This would have to be legally defined, but anything Christchurch 2011 scale upwards: substantial parts of downtown ruined, town a nightmare, services shut down for weeks… you know the drill.


4 Responses to “A missing insurance market”

  • Interesting idea, but almost certainly impractical. The first is defining the event, are you making it purely sized based, so anything over 6.0? The September event was 7.1, but the damage far less. What proximity – a distance from the hypocenter, or from the CBD? The Gisborne EQ in 2007 was a 6.7 but was deep enough and far enough away to only close the CBD for a short time.
    So if we look at Christchurch 2011 (2010 wouldn’t qualify?), then what areas will be paid out for? There are houses with bad damage quite a distance away that might be beyond a radius, but then houses very close (e.g. Riccarton) virtually unaffected – they would have the right to sell even if there was no damage?
    You could make it MM based, but that is subjective. Damage is easier (although not completely without issue) to identify, but a subjective qualifying test will simply lead to arguments as to whether it was substantial, for “long” enough – whatever qualifiers you use will end up in debate that ultimately lead to everyone being covered or not covered (assuming it was a common product, but inevitably you would end up with tweaks to the wording in the market and gaps).
    Now, you might think that all of that can be overcome; that someone with no damage can sell out and the insurer simply then has an undamaged asset to resell when they choose. But:
    a) the value involved would be too much. You would have to include the land, but land is not insured; and say an average land and building price of $475,000, and 100,000 houses in the zone, means insurers have to pay about $50Bn, then try to resell. In reality not everyone would claim, what if in the moment you wanted to repair; you could have a regular underlying policy that then paid for the repair. But then a big proportion of your neighbours move out and the city becomes a ghetto – a voluntary red zone.
    b) it turns the insurers into real estate agent and will flood the market depressing prices for years to come. Even if only 10% of houses had a claim, imagine putting 10,000 houses on the market in Christchurch over a 3 year period.

    I think the economic consequences of a wholesale shift in property ownership would be very bad, and I struggle to see how you define the product tightly enough given the wide disparity of effect we see in Christchurch from various events in various areas.

    • I’d take a first-cut definition something like:
      1) An earthquake causing at least 100 deaths in Wellington;
      2) An earthquake in the Wellington area triggering reinsurance payments of at least $5 billion;
      3) An earthquake resulting in Parliament’s reconvening or planning to reconvene outside of Wellington.

      Pretty much any of those would fit the bill for me. I don’t think any of those are hard to define: we’d have number of fatalities ballparked within days, or at least we’d know whether it was likely to exceed it by any reasonable margin. Reinsurance could take longer to sort. Parliament’s moving – that would happen reasonably quickly too, though it could be delayed if MPs couldn’t leave town (hence the “planning to reconvene”).

      I’d want to be able to sell out regardless of the damage to my property. I expect that relatively few people would take up this kind of policy: most people have very strong personal attachments to their homes and would prefer not to leave. And you could imagine that reinsurers would start to scale up the premiums for those policies if too many people started buying them – that can sort itself out.

      If 10,000 houses had come onto the Christchurch market quickly, that would have been a blessing, not a curse. Rebuild workers would have had places to live; those displaced due to quake repairs also would have had an easier time.

      I do not want the policy to trigger based on damage to my own house. I want to never have to wait for an insurance assessor. Basically, this is a life insurance policy that triggers on a sufficiently large

      • I accept that those are less subjective criteria, certainly the first two; although an event causing one fewer than the threshold could still be just as damaging – it just happened during the night or a long weekend. You could see different results, for the same damage based on some random quirk, and then you might feel aggrieved. (Or a small town devasted, say, Murchison, that could never achieve the threshold). But if we accept that there are big events and small events, and this is intended for a Wellington 1855, or Napier 1931, or Christchurch 2011; but not Murchison 1929 or Inangahua 1968 or Gisborne 2007 (not sure where Pahiatua 1942 or Edgecombe 1987 would fit) – then what is the radius, and from where (the hypocenter)? Again, that could throw up lots of disparities…
        I still think that the global market wouldn’t take it up because the modelling would show that the worst case scenario (Wellington EQ or Auckland Volcanic Eruption) would put so many houses into the frame that the accumulation would be above reinsurance possibilities.
        I also am not convinced that putting houses on the market in bulk would be good. It might have been good for the houses to be available for renting to workers etc., but they probably wouldn’t buy a house. You still damage the unaffected people when you dump the houses back on the market.

        • I’m pretty confident we could write up a contract specifying the triggers adequately.

          You could be right that reinsurers wouldn’t be keen. But would their liability be that much greater than they face when just insuring the house? Right now, they’re on the hook for the vast bulk of repair costs up to some stated value cap. So they face costs of (repairs + putting up homeowner elsewhere during repairs + assessment1), where assessment1 is the costs of assessing the damage and dealing with repairs when you have an annoying homeowner occupying the place. In my alternative, they face potential liability of (upfront stated value payment – resale value of property + assessment2), where assessment2<assessment1. Repair of the property is also likely cheaper where the insurer knows that the homeowner isn't trying to get betterments: the insurer provides repair up to the point where an additional dollar's repair gets an additional dollar of resale value for the house.

          There could maybe be policy reasons for not wanting a substantial portion of town to be on such contracts, but I'm not entirely convinced. And, again, if the reinsurers expected that this kind of contract would depress the post-quake market, shouldn't that just mean that ex ante premiums should be increasing in the number of such contracts?