In my previous article on annuities, I explained that a life annuity is the only financial instrument or product that can give an individual a fixed income for life. So annuities sound like an attractive proposition, particularly if individuals are assumed to be rational and seek to smooth out consumption over their lifetimes. This proposition is even more attractive when we consider the higher payments due to the mortality premium.
And yet we see little voluntary investment in annuities in Australia or overseas. There are rational and behavioural reasons for this annuity puzzle. I will discuss the rational reasons here, and explore the behavioural reasons in my next article.
While this is an area of much research (I have drawn on work by Jeffrey Brown in the US and domestic researchers Michael Sherris, John Evans and Susan Thorp), reasons can also be drawn from industry experience.
The main reasons a person may sensibly choose not to annuitise include:
1. Cost – Many regard annuities as poor value. It is a difficult claim to make with confidence: the ‘money’s worth’ of annuities is a complex calculation. Money’s worth calculations compare the value of the expected payments from an annuity with the cost. This may sound simple but it isn't. For instance, it is hard to choose a discount rate and estimate life expectancy.
Much of the research points to annuities offering less than fair value (that is, a money’s worth ratio of less than 1). This shouldn’t surprise, as annuity providers must put capital aside and target acceptable shareholder returns.
This is not the end of the story. Annuity providers are exposed to adverse selection. This is a phenomenon whereby the life expectancy of those who choose to annuitise is actually higher than that of the broader population. This could be simply because those who are wealthy and seek financial advice often live longer. However it could also be because people who believe they have a longer than average life expectancy find annuities attractive. This is known as adverse selection.
The only way annuity providers can allow for this is through pricing. Research has shown that if ‘money’s worth’ calculations are based on the life expectancy of the annuitised population, then money’s worth is much closer to fair value.
Interestingly, even when prices are less than fair, the models suggest that individuals would still benefit from investing in annuities.
2. Age pension – Australians already have (conditional) exposure to a life annuity: the age pension. However the rational models suggest that full annuitisation is beneficial and can’t explain why people with different income levels choose not to annuitise (as the age pension will provide varying replacement rates across the population).
3. Default risk – Life companies have defaulted in the past. For instance, we saw several US life insurers fail in the early 1990s, including Executive Life Insurance, Mutual Benefit Life Insurance, and Confederation Life Insurance. APRA requires that life companies keep aside enough capital to withstand the events of the next year with a 99.5% probability of sufficiency. These standards could be thought of as a 0.5% chance of default (obviously life companies could hold more than the minimum capital, further reducing the risk of default). This may sound like a small risk, but if we consider that someone annuitising at 65 could live for another 30 years, then the probability of default over the annuitant's lifetime becomes 15%. And there is always the possibility that risk models fail to correctly estimate risk (surely not!). Unlike previous articles on credit investing, which have emphasised the benefits of diversification, it is difficult to diversify annuity provider exposure in Australia.
4. Bequests – For those with strong bequest intentions, full annuitisation is not rational. However partial annuitisation could be a rational choice.
5. Irreversibility – Typically annuities are irreversible contracts (though innovation by groups like Challenger has led to the introduction of exit clauses for reasonable time periods). The irreversibility takes away flexibility, which is difficult to value. An irreversible contract is not undesirable in the context of default risk and bequest motive issues previously outlined.
6. Deferral may be optimal – In the previous annuities article, I explained the concept of the mortality premium, which makes life annuities more attractive. Basically, because not everyone in the insured pool is expected to survive to the next period, a life insurer can afford to make higher payments compared with those received from the underlying securities (typically fixed interest securities).
Now consider the case where the probability of dying in the next period (say a year) is very low; the mortality premium associated with that first year will be low. The potential risk adjusted return from deferring annuitisation and instead investing in equities may be positive, so it may be rational to defer annuitisation. However this does not mean that we should not invest in annuities at all – there will come a point where the marginal mortality premium exceeds the risk-adjusted return expectation of the alternative investment.
7. Incomplete markets – We may not be offered the most attractively featured annuity products at reasonable prices. There are two broad reasons for this, as noted in the Henry Review: supply issues and barriers to innovation. Supply issues include the lack of market support for the hedging and sharing of mortality and longevity risk, and the availability of long-dated (particularly CPI-indexed) fixed income securities. Barriers to product innovation consist of the red tape burden imposed by various regulators. Deferred annuities are a case in point. They are interesting products from a financial planning perspective that are in effect ruled out by their tax treatment (proposed government changes could fix this problem).
8. Risk-sharing in couples – Couples effectively insure each other through bequests. However even in this context, life annuities still have a role to play among rational decision makers.
9. Financial advice models – Some suggest that financial planners who use planning models that rely on trailing commissions may be less likely to recommend life annuities. But I’m sure this statement cannot be applied to all financial planners.
So overall there are many reasons why a rational individual may choose not to invest in annuities. It should be pointed out that academic researchers have considered each issue and found that, in most cases, no reason carries enough weight on its own to justify excluding annuities altogether (although deferring or partial annuitisation may make more sense than full annuitisation in some cases).
And so the annuity puzzle remains for researchers, though many market practitioners can probably find enough cause to be put off annuities in the reasons listed above. Researchers are still determined to find behavioural explanations for the lack of annuitisation, and I will explore some of these reasons in a subsequent article.
David Bell’s independent advisory business is St Davids Rd Advisory. David is working towards a PhD at University of NSW.