Managing retirement income: annuities

A couple of articles in last weekend’s newspapers got me thinking again about retirement income.

Diana Clements’ personal finance column in the Herald traversed the question of annuities. She begins:

There’s a gaping hole in the KiwiSaver concept: What happens to the money when people turn 65.

There is no need for people to do anything active with the accumulated Kiwisaver funds at 65, and in most cases the amounts involved are still quite small.  But when people have been in Kiwisaver for 20 or 30 years the amounts will be much larger.  And whether or not people do anything active with the funds at 65 they still face the question of how much of their wealth they can afford to spend each year when they do retire.

Clements reports that one company is launching a new hybrid annuity product: put your lump sum with the Retirement Income Group and you will receive an annuity for life, with the twist that if at death the total amount that has been paid out is less than the initial capital sum, the balance goes to the estate.  Of course, that additional insurance isn’t free: without it, the initial capital could afford to pay a higher annuity.

Annuities generally appeal to those with some reasonable level of wealth who think – or worry – that they will live longer than average (and to those without much of a bequest motive).  Actuarially fair annuities can be a great product for an older risk-averse person, especially if they come inflation-indexed.  They convert accumulated savings into a certain real income for life, and the recipient doesn’t have to worry about market investment risk or inflation for the rest of their life.  In that respect, they have some similarities to the best old defined benefit pension schemes.

But defined benefit pension schemes are now almost entirely a thing of the past – schemes have typically had no new contributors for decades –  and (as far as I’m aware) only government-associated schemes ever offered proper and full inflation indexing.  And annuities are hardly seen on the open market at all, and when they are offered they are not inflation-indexed.

The absence of inflation-indexing probably reflects the inability of markets to properly hedge against inflation risk.  That risk is particularly severe in a TTE tax system: when nominal interest is fully taxed at the time it is earned, an unexpected bout of inflation can do very nasty things to real returns.  Capital gains taxes also only rarely adjust for inflation in calculating the taxable gain.

Just recently, the British government announced it intended to scrap a requirement that accumulated funds in private defined contribution schemes be taken as annuities, a move described by one industry participant as “the final nail in the coffin of annuities”.  If so, that won’t be surprising.  Annuities fall foul of two things: adverse selection (when people can choose whether or not to take an annuity), and sharp falls in real long-term interest rates.

A couple of hundred years ago, annuities – and tontines – probably worked quite well (apparently the first known life tables for annuities date back to the third century AD).  Diagnostic medicine barely existed, and many people died from infections or illnesses with a random component.  An annuity provider could attract a pool for customers to whom it could provide reasonably actuarially fair annuities.  No one had the information to select into the pool in a way that would skew its long-term returns away from the provider.  (And expected inflation, and the interaction with the tax system, wasn’t a problem).

Now that is much less so.  Of course, there is still a random element to life expectancy, but it is the systematic element that creates problems for the provision of annuities.  People who come from long-lived families will be more inclined to like the idea of an annuity, but each $1m of capital will support a much lower annuity for them than it would for someone with a much shorter life expectancy.  If providers don’t, or can’t, differentiate by individual life expectancy, and instead set common annuity rates it will mostly be people who think (or fear) they will live for a long time will opt in, which will drive down the sustainable annuity rates, until fewer and fewer people opt for an annuity.    There will always be some, limited, market for the very risk averse, but it isn’t likely to be a very substantial market when people are free to make their own choices.  And even the risk-averse have to worry about the creditworthiness of the annuity provider – one reason why annuities and related products were often sold by governments (at least those with a pretty good track record of not defaulting on general creditors).

And, of course, the other problem is variable – but in trend decline over recent decades – real investment returns and real bond yields. As they have fallen, annuity rates (per $1m of capital) have fallen substantially.   In some respects, it is an inescapable problem. If real returns are low, there are no free lunches, and one can’t escape the constraint simply by not purchasing an annuity.  But timing can matter, potentially a lot.  With advanced country long-term bond yields so low, anyone forced by a government to take an annuity today risks having their purchasing power for the rest of their lives determined by today’s annuity prices.    That might be fine –  no better or worse than the alternative – if yields stay this low for the next 20 years (and who is to say confidently that they won’t)  but what if in three years’ time economies have snapped back and G7 long-term bond yields are averaging 3.5 per cent rather than something like 1 per cent at present?    A government that would force you to take an annuity now, just because you happen to turn 65 this year, would seem to be over-reaching itself.  Of course, real yields could fall further yet, but governments are no better-positioned than you or me to tell whether, and if so for how long, that will happen.

Probably none of us wants to see people in deep poverty in old age.  But the state is already providing an excellent product to cover that risk.  In New Zealand’s case, we have NZS, a modest and near-universal state welfare payment to those 65 and over, which is indexed to wages.  It is, in effect, a real indexed annuity, provided by a government with a pretty strong financial position, which has not defaulted its credit obligations for over 80 years.  NZS is not a particularly comfortable level of income, but it has been enough to ensure that New Zealand has one of the lowest levels of elderly poverty among OECD countries.  If one thinks of the role of the state as partly being a residual insurer, then NZS (combined with a public provision of health services, and resthome subsidies) pretty much has that role covered for the old.

It will be interesting to see how Ralph Stewart’s product goes. I wouldn’t be optimistic, but entrepreneurialism is partly about taking products to market and seeing if they work.  But we should resist the calls that will probably come, in various guises over the next few years as Kiwisaver matures, for anything like compulsory annuitisation of accumulated Kiwisaver balances at age 65.   Unlike the Australian compulsory private scheme, Kiwisaver was explicitly not envisaged as a replacement for the state pension, but as a top-up.  The complex Australian scheme needs all sorts of rules to avoid the state being rorted (probably more rules than it has).  Kiwisaver is just a moderately expensive, moderately distortionary, top-up to what is a pretty cheap and efficient state scheme, which serves what must surely be the main goal –  as it has been since 1898 – of avoiding the indignity of extreme poverty among our old citizens.  And it is a scheme that costs less, as a share of GDP, than most advanced country governments spend.  It isn’t perfect, but I’ll come back to that another day.

Some people might later regret blowing their lump sum Kiwisaver balance in a year or two after they turn 65, but people make choices that can look unwise  –  and that they might themselves regret later – at all sorts of phases of their lives.  Frankly, spending accumulated savings on a good time while they are still young enough to enjoy it, rather than spreading spending cautiously and evenly over decades (as I’m sure I will –  thank goodness for legacy DB pension schemes) seems like one of the less damaging choices people can make, and not really something for the state to concern itself with.

While on the subject of retirement, I noticed that the Retirement Commissioner is running a competition to come up with “catchy names” for different phases of retirement.  Why are we spending scarce public money on this patronising nonsense?  The amounts involved are no doubt small, but it is the mind-set that is concerning.   “Take care of the pennies…” was the old maxim.

Come to think of it, why do we need a Retirement Commissioner at all?   The Commissioner operates under the ambit of something now called the Commission for Financial Capability (costing almost $6m per annum).  But history suggests that governments themselves don’t have that much financial capability and are often worse  than the general public at making good financial choices –  which shouldn’t surprise us, since the people involved have less on the line.  In New Zealand, Think Big, Kiwirail, and a mismanaged Deposit Guarantee Scheme are just the first three central government activities to spring to mind.  And the Treasury is constantly warning us of what it regards as the somewhat perilous long-term position of even New Zealand government finances.  Individuals typically do less badly.