This is one of a series of columns that were produced for Moneyweb Investor in which Stuart Theobald explores the intersection of philosophy of science and finance. This followed an earlier series for Business Day Investors Monthly on the same theme. This column was first published in February 2016.
Treasury has been forced into another two year delay in its efforts to change the treatment of provident funds on retirement. The main issue is that savers will be restricted to a lump sum payment of one third of their savings at retirement, with the balance used to purchase an annuity. While purchasing annuities generally makes sense, government’s move has come at a cost to provident fund savers. Analysing why provides important insights on the nature of insurance, which even the wealthy can learn from.
Provident funds are the retirement vehicle of choice particularly for lower income workers, so unions, including Cosatu affiliates as well as Amcu and Numsa, have been fighting hard over the change. The problem for their members is that their provident funds serve as security for borrowing. Currently, provident savers can withdraw the full balance of their savings on retirement. So the drop from the full amount to one third dramatically reduces their ability to offer security to lenders.
I am sympathetic to the unions’ point. From the point of view of option theory, provident fund members are losing optionality that has a value, value they are currently effectively able to monetize by obtaining cheaper finance. While the change includes a carrot in that contributions are going to become tax deductible, most such savers are below the tax threshold anyway. So for them, the change is entirely negative (though in a questionable strategic move for government, the tax deduction is not subject to the two year delay, so that carrot is gone in future negotiations over annuitisation).
All insurance is irrational in a sense. The probability of you incurring the insured loss weighed up by the value of the loss is always lower than the premium you will be paying (imagine a R100 000 car you have a 10% chance of writing off – the premium would be more than R10 000 else the insurer would make no profit). Usually we make investment decisions on the basis of expected returns, by which we mean the probability weighted return discounted for the time value of money. For instance, we’d pay some amount less than R50 000 for a 50% chance of getting R100 000; if you had to pay more than R50 000 it would be an irrational investment, yet that is effectively what we are doing when we pay insurance premiums.
But just how important is it to buy an annuity? The answer is, very. An annuity is effectively a reverse life insurance policy: it pays out if you don’t die. That’s a flippant way of saying they are good for dealing with longevity risk – the risk that you will outlive your savings in retirement. An annuity continues paying you an income, no matter how long you live. Of course, the inverse is also true – if you die early, you effectively are denying the next generation an inheritance by using your savings to buy an annuity, a less plausible argument the unions have also made. Insurers will adjust the monthly annuity payments according to their perception of risk – if you are a smoker, for instance, you’ll get a bigger monthly cheque than if you are not.
Some insurance does actually provide us with economic value. An insurance company is able to pool risks in order to get portfolio benefits and has a large balance sheet so it can withstand surprises. We have small balance sheets and large assets are a big part of our wealth, so the volatility of cash flows around insurable events is much higher than for the insurance company. Some negative outcomes would wipe us out. It is best, then, to think of all insurance as against economic ruin rather than the risk of any particular insured thing. One obvious insight is that it never makes sense to insure small value losses. An extended warranty on a TV for instance, is an absolute waste of money. You will not be financially ruined if your TV, phone, fridge or other appliance dies. It also follows that the wealthier you are, the less insurance you need. At a certain point of base wealth it makes no sense to insure your own car for example (though 3rd party cover, subject to a high excess, may make sense because those losses could theoretically be infinite). Of course, this doesn’t apply if you know that you are a much higher risk than the average – though your insurer is likely to figure that out eventually.
It’s worth thinking about this in the context of your own investing behaviour. If you have a pool of investments, you could get a better expected return by cancelling insurance policies and redirecting the premiums to your portfolio, effectively self-insuring. If an insurable event happens, you can settle it from your portfolio, but the probabilities are in your favour, just as they are in any other investment decision. You get the profits your insurance company would be getting.
When you apply these lessons to longevity risk, the case to be insured becomes much clearer. You have a single risk, that of outliving your savings. Your ability to self-insure declines as your savings are disbursed in living costs. Unless your savings are very large, such that expenditure falls below what you could spend such that you could afford to live beyond the most optimistic life span, an annuity makes sense.
For low income retirees, it is unlikely that expenditure will be lower than what their savings can support. So the benefit to being insured through an annuity clearly increases as they age. It is one type of insurance that absolutely makes sense.
That is true whether government makes it law or not. But the law removes the option of using savings as loan security. Provident savers effectively sell that option to lenders in return for cheaper debt, but are repaid it when they settle the debt. No retiree should reach retirement with debt, but the ability to sell the option makes debt cheaper for them earlier in their lives. Government has delayed the new rules for two years. It needs to rethink them