Western society is ageing: the consequences for incomes, savings, taxes and pensions are uncertain. Recent work by David Miles illustrates how economists can assess the impact of demographic change and its implications for public policy.

It is well known that there will be significant changes in the demographic structures of virtually all developed countries over the next few decades. In the absence of massive immigration or catastrophic new illnesses, by the middle of the next century, the ratio of people of working age to those of retirement age will have halved since 1961 – very roughly, from four to one to two to one.

Such dramatic demographic change could have a powerful impact on savings behaviour in both the public and private sectors. It is also likely to affect capital formation, labour supply, tax rates and real wages. But estimates of the likely size of these effects vary significantly, depending on what kind of evidence is used.

In a new CEPR Discussion Paper, David Miles argues that simulations based on ‘calibrated general equilibrium’ models are likely to provide the most reliable method for assessing the impact of ageing on the UK and the rest of Europe. How governments respond to shifts in saving and in the burden of state pensions is a vitally important issue: such models can be used to assess a range of public policy options.

Calibrated ‘overlapping-generations’ models take as their starting point the decisions of individual agents, calculated on the basis of explicit decision procedures. The outcomes for the economy as a whole are obtained by adding up these individual decisions. Such models have several attractions as a tool for analysing demographic shifts: they are based on strong theoretical foundations; and they can be used for policy simulations, for calculating the implications of varying degrees of myopia or of different degrees of individuals’ rationality; and, by varying a small number of parameters, for assessing the aggregate implications of a wide range of individual behaviour patterns.

Miles develops such a model, featuring agents who choose their labour supply, consumption and wealth accumulation in an environment where state pensions and labour taxes may be changing over time. He uses the model to generate projections of how shifts in population structure will affect savings, wealth, interest rates and taxes over the next 50 years.

These projections reveal that saving rates will tend to fall over the next 50 years. In addition, real wages for the working population will be higher than they would be with an unchanging demographic structure. Workers will become relatively scarce, the capital to labour ratio will rise and the marginal product of labour and the wage will increase.

Miles notes that this real wage effect is unlikely to be particularly powerful. He estimates that real wages by 2040 will turn out to be only a little more than 3% higher than if the population structure remained unchanged. And although real wages will be higher, per capita GDP will follow a path that is well below what would happen if the labour force remained a fixed proportion of the population. According to these projections, the average European level of per capita GDP will be just under 25% lower by 2050 than it would be with an unchanged demographic structure.

Of course, per capita GDP is not a measure of welfare, and falling saving rates per se should certainly not be seen as a ‘problem’ (any more than a decline in an individual’s saving rate or a fall in their labour income late in life are necessarily a problem). Nonetheless, governments mayfeel they should respond to low savings with tax breaks on wealth accumulation or, perhaps more likely (and certainly more helpful in terms of reducing public sector deficits and taxes), by sharply reducing the value of the state pension.

Miles considers an experiment involving a steady reduction in the average European replacement rate, that is, the ratio between the state pension and the average gross earnings of workers. In his experiment, the replacement rate is reduced by 0.01 a year from its assumed starting value in 2010 of 0.4 – when the state pension is 40% of earnings. Miles sets the starting date for cutting the relative value of pensions some time in the future – the precise date is arbitrary, but it allows individuals to begin adjusting now to the prospect of lower benefits in the future. And an annual reduction of 0.01 from 2010 implies that the state pension will be completely abolished by 2050.

The simulations reveal that a fall in the replacement rate has a dramatic impact: with the complete phasing out of state pensions, the decline in the aggregate saving rate is much reduced. In the base case (where the replacement rate remains constant at 40%), the saving rate falls from around 16% in 1995 to 5% by 2045; with phasing out of the state pension, the saving rate stabilizes at 10% by 2050.

Higher savings imply a higher wealth to income ratio and a higher capital to labour ratio: consequently, there would be lower rates of return on savings. And so with gradual abolition of the pension, the real interest rate would fall from just over 5% in 1995 to 4.2% by 2050.

The implication of these simulations is that if behaviour consistent with life cycle models of saving is important (in essence, that savings rise and fall over an individual’s lifetime with most saving done between ages 30 and 65), then demographic changes in nearly all developed countries will mean significantly lower aggregate private sector saving in the medium term.

Miles draws two broad conclusions from the results of the simulations. First, that there may well be substantial swings in private saving rates over the next 50 years, but that savings in the longer term are likely to fall well below recent levels as the proportion of the population aged over 65 rises to levels never seen before.

The second conclusion is that the impact of the reduced saving rate on rates of return on capital may be relatively muted because a lower saving rate (which, other things being equal, would lower the path of the capital to labour ratio) is likely to be offset by a smaller workforce (which obviously works in the opposite direction). In the simulations Miles conducts for the European economy as a whole, the saving rate will be less than half its recent level by 2030, while the real interest rate will fall by 40 basis points (from around 5%), a proportionate decline of under 10%.