Tax-benefit model of Budget shows new perspectives

The distributive impact of Budget 2002 is markedly different, not only from its immediate predecessors, but from most budgets…

The distributive impact of Budget 2002 is markedly different, not only from its immediate predecessors, but from most budgets over the past 15 years.

Typically the greatest gains during those years have been for those in the middle or at the top of the income distribution, with small gains, if any, for those on the lowest incomes.

By contrast, the impact of Budget 2002 is highly progressive, favouring those with the lowest incomes. This is one of the key findings of an investigation of the distributive impact of the Budget using Switch, the ESRI tax-benefit model.

This analysis differs in a number of respects from other post-budget examinations of the distributive impact of budget policy. The tax-benefit model is based on the incomes and circumstances of a large-scale national survey of households.

READ MORE

As a result, it reflects the diversity and complexity of real households in a way that is not possible with a limited set of selected illustrative households. It can answer questions about the overall impact of the budget in a more systematic way than was possible heretofore.

The other major difference is in the benchmark against which budgetary impact is assessed. The benchmark most widely used is that provided by the conventional opening budget - no change, in nominal terms, in taxes or welfare.

This is deeply flawed as a yardstick for assessing the distributional effects of budgets. If the opening budget were actually implemented, welfare payments, tax credits and the width of tax bands would be frozen in nominal terms.

This would be far from neutral in its distributive effects: welfare recipients would lose out in real terms, while "fiscal drag" would raise effective tax rates. Measuring distributive effects against this backdrop is not, in my view, a sensible procedure.

The alternative benchmark proposed in ESRI model-based work is a budget which simply indexes tax and welfare parameters in line with expected earnings growth. This benchmark has the attraction of being "distributionally neutral" i.e. if implemented, disposable income would grow at the same rate across all income groups and family types. The distributive impact of actual budgets can be measured more accurately and more clearly against this yardstick.

The table presents results based on the ESRI tax-benefit model. It shows that the poorest one-fifth of families gain about 4 per cent as against gains of under 1 per cent at the top. There are above-average gains for the bottom 60 per cent of families and below-average gains for the top 40 per cent.

How has this come about? Several features combine to give these results.

First, the balance of the budgetary package is tilted firmly away from tax reduction towards increased benefits, including child benefit. Over most of the past 14 years, total welfare payments have amounted to no more than 10 per cent of the net cost of the budgetary package (over and above wage indexation).

In Budget 2002, however, the net cost of welfare changes represents almost 80 per cent of the total net cost.

While the amount devoted to tax changes is significant in absolute terms, much of it is required simply to bring tax credits and bands into line with expected earnings growth.

Second, the substantial rise in child benefit gives the same absolute increase in income to similar families, irrespective of their income levels; but the proportionate boost to income is, of course, greatest at low incomes.

Third, general rates of increase in welfare payments are often in excess of 10 per cent, well above expected earnings growth of 6 to 7 per cent.

On the tax side, the concentration of resources on raising the personal tax credit also helps to focus benefits on the middle, rather than the top, of the income distribution.

What about the impact of Budget 2002 on families of different types? Does it, as some have argued, favour two-earner families without children over single-earner couples with children? Our findings indicate that this is not the case. Relative to the neutral benchmark, two-earner couples without children gain half a percentage point in income from the Budget. But one- and two-earner couples with children gain about one and a half percentage points.

The largest gains are for those depending on welfare. Pensioners and unemployed persons without children gain from 2 to 4 per cent, while unemployed couples with children gain an average of about 7 per cent.

The analysis set out above does not cover indirect tax changes, nor does it take account of potential changes in labour supply behaviour arising from the tax/welfare changes.

Even if such factors could be included, however, it is likely that the distributive impact of Budget 2002 would remain in sharp contrast with that of earlier budgets.

Tim Callan is a research professor at the Economic and Social Research Institute and co-author, with Mary Keeney, Brian Nolan and John Walsh, of Reforming Tax and Welfare, ESRI Policy Research Series Paper No 42