I think that reputation matters when exit matters. That is, if people
will switch suppliers based on word of mouth, then reputation will be
This sounds eminently plausible, but it's misleading. Imagine a world where we all choose our health insurance provider when we're twenty; then we're locked in for life. If we take Arnold literally, reputation would be powerless in this world, because no one can exit. But the mere fact that insurers have to get initial voluntary consent gives them a strong incentive to treat all of their locked-in customers well. Otherwise, it's going to be hard to attract new customers.
I'm not saying that reputation in this "choose-once" regime would be as effective as it normally is. But I'm still confident that it would work fairly well. And in truth, many major markets - such as residential construction - are close to choose-once regimes. Yes, if you don't like your house, you can move; but if the builder did a bad job, it will be reflected in price you're able to get.
What about Arnold's three specific doubts about reputational incentives in health insurance?
1. Most people get their health insurance from an employer. The
individual cannot switch companies. The most the individual can do is
voice complaints (about non-payment of claims for example) to their
human resources department, and hope that the human resources people
use the threat of exit to change the insurance company's behavior.
My reply: Yes, health insurance is bundled with jobs. But markets bundle lots of products, and I see no evidence that this bundling undermines reputational incentives in the least. Think about a typical restaurant. It bundles many kinds of food, plus service, decor, cleanliness, and location. But restaurants' concern for their reputations leads them to maintain quality on all of these dimensions. Admittedly, it may be hard for any one customer to get exactly the combination of food, service, decor, cleanliness, and location he wants. But if the restaurant can make a significant number of customers happier at a reasonable price, it will.
2. Turning to the individual health insurance market: In many
states, such as Maryland where I live, there are not enough suppliers
to have meaningful competition and opportunities for exit. The state
regulations for "community rating," "must-carry" and mandated coverage
have driven most companies away.
Lack of competition raises prices. But why should it reduce cariers' desire to be known for high-quality service? Even a profit-maximizing monopolist-by-law has to worry about the reputation of its products. After all, the monopoly price for a well-regarded product is a lot higher than the monopoly price for junk.
3. The insurance industry does not have a solution for the
pre-existing condition problem, so many people face lock-in. If John
Cochrane's idea of health status insurance were implemented, that would address the problem of lock-in.
I agree that Cochrane's idea would be better than the status quo. But again, if your firm has a reputation for mistreating locked-in customers, people aren't going to want to become your customers in the first place.
Can reputational incentives really be as robust as I say? I know it's hard to believe. But think about all of the firms you interact with that have the same problems that Arnold attributes to health insurers: Restaurants, grocery stores, apartment complexes, builders, car companies, employers, etc. How often is it in their short-term interest to mistreat you? How often are you in fact mistreated? Aren't reputational incentives the only reasonable explanation?