A secondary but still important issue raised by Peter Saunders in his CIS paper is that of the lifetime impacts of redistribution. Given that the same person may be poor in one period, and rich in another, is anything gained if we tax the rich and give to the poor?

To think sensibly about this question, we need to distinguish a number of different processes that may be at work in changing people’s relative income positions over time.

First, there are life-cycle effects. A person might spend their first adult years as a student with a low income, then experience rising incomes over their working life, then retire and have a lower income. In the presence of that magical beast, the perfect capital market, they would simply smooth their consumption by borrowing and saving and there would be no reason for governments to do anything. In practice, things aren’t so easy and there are good reasons why governments should help students and assist people to save for retirement. But, for people whose income is generally going to be high, there’s a good case for doing this in an individualised fashion rather than through general taxes and transfers. The HECS system provides an example. Rather than giving everyone free education and taxing everyone (or all high-income earners) to pay for it, individuals get assistance and then repay it through the tax system. Since HECS repayments reduce an individual debt, they shouldn’t, in principle, involve a significant disincentive to work. And systems like HECS avoid perverse redistributions. Similar points can be made for a ‘three-pillar’ retirement system in preference to universal income-related social security.

Second, there are random fluctuations in income arising from things like unemployment. In the absence of perfect capital markets, there is a strong case for governments to provide some form of insurance against these risks, independent of concerns about redressing inequality in lifetime income. Moreover, increases in the general level of income have no real effect on the argument. What matters is the riskiness of income. Most of the labour market policies promoted by the CIS are designed to reduce the security of incumbent employees and generate a more dynamic market with higher levels of both job creation and job destruction. Inevitably, this means more risk and strengthens the case for publicly provided insurance.

Finally, there is economic mobility, which I’ll define to mean a shift from one kind of lifetime income trajectory to another. We can look at mobility within one person’s lifetime as they change careers, or from one generation to the next. If there’s lots of mobility within, or between lifetimes, then the case for redistribution appears less compelling. Its often assumed that ‘dynamic’ economies like that of the US have lots of mobility, while ‘static’ economies, like those in Europe have less. In fact the reverse is true. When you think about it, that’s not surprising. If you have high levels of inequality and highly privatised consumption of human capital services like education ,it’s very difficult to catch up from an initial disadvantage, and relatively easy for parents to pass on advantages to their children.

Thinking about poverty and inequality in a lifetime sense, rather than in terms of a single-period snapshot gives us a more sophisticated understanding of the issues. In some respects, a life-cycle perspective weakens traditional arguments for the welfare state; in others, the two reinforce each other.