Would reducing earthquake premiums lead to greater market penetration?

A couple of posts ago (Why ‘just-in-time’ insurance would be a non-starter) I answered a question that I get asked at least a few times a year: ‘Why not allow consumers to purchase insurance just before a foreseeable loss is about to occur or immediately after it has occurred?’ I laid out several reasons why a regime of ‘just in time’ insurance would not be workable.

Not long after that, another question was posed to me by an acquaintance in the earthquake home retrofit business: ‘Why don’t insurers reduce their premiums in order to encourage more people to buy earthquake insurance?’

Again, many suggestions involving insurance that come my way almost always are from folks who know virtually nothing about how insurance works and, thus, tend to be unrealistic and unworkable. But in order to understand why arbitrarily slashing the price of earthquake insurance in order to sell more policies would likely be futile (and probably even very risky), it is probably best to start with a basic theory of economics: price elasticity of demand (i.e. the change in demand for a product or service that occurs when its price changes – either upward or downward).

The question with elasticity of any product or service is: If a seller reduces prices, will demand always increase? The answer, it turns out, depends on several factors, including the type of product in question, the need for the product, supply and demand, and the personal situation of the buyer. With these and other things to consider, there are three (really four) main things that might happen with demand, depending on the type of elasticity that is in play.

Under a regime of unitary elasticity, demand changes in response to price changes in a way that leaves total revenue the same. So, if a seller doubles prices, she sells half as much or if she cuts her price in half, she sells twice as much. Either way, she brings in the same revenue.

Under a regime of elasticity, demand goes up when the price goes down (with a corresponding increase in revenue) and demand goes down as the price goes up (with a corresponding decrease in revenue). Incidentally, there is a subset of elasticity know as infinite elasticity wherein a price increase could cause demand to drop to zero. According to one source, “Cases of infinite elasticity are rare and usually confined to a single locale or a special situation.”

Finally, under a regime of inelasticity, demand goes up when the price goes down but revenue also goes down. That is, price goes down, sales go up, but revenue still goes down.

So the question then becomes: Is earthquake insurance pricing unitary elastic, elastic or inelastic?

While the question has not yet been definitively settled as very little work has gone into looking at the price elasticity of earthquake insurance (more has been done on the price elasticity of flood insurance), some interesting research has been published.

In a 2010 study sponsored by The RAND Corporation, researchers concluded that earthquake insurance is, indeed, price elastic (i.e. demand goes up when price goes down and vice versa).

According to the researchers: “When one controls for other factors, including indicators likely to be correlated with the within-zone variation in perceived risk, premium price appears to have a significant and negative effect on demand. Since the model is specified as a log-linear relationship, the estimated coefficient of –0.4814 represents the price elasticity of demand. In other words, a 10-percent decline in price would result in about a 4.8-percent increase in the number of households purchasing earthquake insurance. This relationship is based on holding other factors constant, including the level of perceived risk, household income, housing values, and demographics, such as age, race, and other ZIP Code characteristics.”

However, contrary to The RAND paper published just a few months prior, a 2011 study conducted by Athavale & Avila concluded that demand for earthquake insurance is driven by something other than price: “There is an economically and statistically insignificant relationship between the price of earthquake insurance and the demand for earthquake insurance that is not explained by risk. Specifically, we obtain price elasticity estimates of 0.005, suggesting that the demand for earthquake insurance is almost perfectly inelastic. Our results indicate that the price of earthquake insurance coverage does not explain the demand for earthquake coverage beyond that which can be explained by the risk of occurrence. The results of the empirical analyses are consistent with the notion that insurance companies price risk in the underwriting process and that the demand for earthquake insurance is driven by earthquake risk.”

Along with arriving at different conclusions (The RAND study found earthquake insurance to be price elastic while the Athavale & Avila study found it to be “almost perfectly inelastic”), another difference between the two studies comes with how each measured demand. The RAND research did not use take-up of insurance as the measure of demand but, instead, used the Insured to Value ratio (ITV) defined by researchers as the total structural coverage minus the total of policy deductibles for earthquake insurance divided by the total residential housing structural value in the area under study. Conversely, Athavale & Avila used the traditional measure of take-up – i.e. the percentage of residential policies that had an earthquake endorsement – as a proxy for the demand for earthquake insurance.

Both papers, however, are very similar in concluding that price is just one factor deciding whether people purchase earthquake insurance. Other factors include income, ethnicity, location in relation to risk, consumers’ perception of risk, personal loss experience (including the amount of time that has lapsed between earthquake events), and homeowners’ knowledge/awareness of the peril.

Part of the issue – and the challenge of conducting such research – is to understand that while there are similarities between earthquake insurance and other types of catastrophe coverage (such as flood insurance), there are also important differences. Athavale & Avila explain: “Catastrophic earthquakes have been infrequent; earthquake insurance is not covered under a standard homeowner policy and is not required even in areas prone to frequent earthquakes. Further, earthquake insurance is available through the private insurance market and not at a subsidized rate through a government sponsored entity. Homeowners may also be reluctant to purchase earthquake insurance in the expectation that losses may be covered by disaster recovery and relief programs. While prior research suggests that the demand for earthquake insurance would be negatively related to the price of coverage, the differences between earthquake insurance and other types of catastrophic coverage as enumerated above may suggest otherwise.”

So, while it is not entirely clear whether demand for earthquake insurance would go up if prices were reduced or vice-versa, it is very clear in both papers that price is not the only driver of earthquake insurance demand. Indeed, it may play far less a role in demand for earthquake insurance than has been assumed.

The bottom line is that while lowering premiums in order to encourage more people to buy earthquake insurance may work (if the RAND research is correct and earthquake insurance really is price elastic), the gamble that reducing pricing wouldn’t impact demand at all is too great. What’s more, it is likely that the practice wouldn’t pass muster regarding accepted actuarial practice, and may not survive scrutiny by regulators and reinsurers.

But these issues must be left for another day.