Debating Austerity Measures

In last year’s July issue of the Newsletter (no. 150) we published a critical view of the UK government’s current austerity measures, based to a large degree on an article by Victoria Chick and Ann Pettifor. Given its provocative nature (see the letter on p.17) we were somewhat surprised at the lack of reaction until Philip Booth and Len Shackleton took issue with Chick and Pettifor in this year’s April issue. Their reply follows below.

Other items related to this theme, published in the Newsletter, include:
LSE blog questions UK budgetary policy (July 2010)
Are we suffering from deficit fetishism? (April 2011)
Surely you're joking, Mr Keynes ( October 2011)
Fiscal stimulus improves solvency in a depressed economy (April 2012)
Understanding the crisis: clarity on measurement, clarity on policy (April 2012)
In partial defence of austerity (October 2012)

We have been struck by the fact that so few economists have responded to the editor’s challenge to come out in support of the government’s austerity policies. While we do not agree with their arguments, we are therefore grateful to Philip Booth and J R Shackleton (hereafter B and S) for raising their heads above the parapet (in Newsletter no. 152, April 2011) and permitting this fuller debate.

B and S challenge our evidence on three points: that we ignore transfers, which have risen considerably in the century we study; that we do not deal with the lags implicit in our theory; and that our averaging process is suspect. We deal with these in turn and then comment on their approach more widely.


We did not ignore transfers. We excluded them quite deliberately, for two reasons, one theoretical and one statistical. The theoretical reason stems from our purpose: we were interested in the direct impact of changes in government spending on aggregate income and the debt ratio. The ‘G’ of C + I + G is government expenditure on goods and services. The statistics treat government expenditure slightly differently: there, ‘G’ is government current expenditure, and the government’s contribution to I is reported under investment but can be extracted. Our measure combines these two elements to conform to the theoretical concept of G. (Recall that Keynes used the expression ‘public works’, not the modern ‘government expenditure’, which has misled B and S to argue that we should have included transfer payments.)

An increase in G at less than full employment raises income and taxes and reduces expenditure on unemployment benefit and income support, thus reducing the need to borrow. The numerator of the debt ratio may go down and the denominator will certainly go up. This is the immediate, first-round effect. We expect there to be further, reinforcing multiplier effects on income and borrowing which will further reduce the debt ratio over time. 

Transfer payments do not stimulate output directly; they do what they say: transfer income from taxpayers or government bond-buyers to others. Their effect on aggregate income depends on differences in the marginal propensity to consume between the ‘transferors’ and the ‘transferees’ and will certainly act only with a lag.

The second, statistical, reason for excluding transfers is that the amount spent on transfers is governed, in the case of interest payments, by the amount of debt and, in the case of benefits, by the level of economic activity. The causality between our ‘dependent’ and ‘independent’ variables would be complex and uncertain if transfers were included. (For the same reason, we did not analyse taxes or measure the fiscal stance by the deficit.)

We looked at transfer payments (separately from the main analysis) in the second version of our paper, published after the original RES Newsletter piece on which B and S comment (see; go to ‘publications’).

In this second version we separated out interest payments on government debt (where the effect on the aggregate mpc today is ambiguous and in earlier years may have constituted transfers to the relatively well-off) from other transfers, mainly benefits, where the effect of increased payments is likely to be stimulatory. It would be quite wrong to lump these together, let alone add them to G in the analysis of the effects of government spending.

The treatment of lags

It was to deal with lags of uncertain length that we decided to take averages over the periods when G is increasing or decreasing (or, since the Second World War, increasing at a faster or slower rate), hoping that most of the effect would take place within each period.


There is nothing in principle wrong with taking averages. From the statistical point of view you lose degrees of freedom but reduce random variation in the data. A relevant precedent is Phillips’s original article (Economica, 1958). He used grouped data to eliminate the cycle, while we are using them to highlight the direction of change of policy. We also reported the results using unaveraged, annual data in an appendix. B and S do not refer to this. The correlation is less strong but the impact of G on the debt ratio, as measured by the slope, is even stronger than for the averaged data.

Other issues

Outside of these three points, we would like to comment on a number of other contentious statements that B and S make. First, B and S charge Keynes with no understanding of monetary considerations, which, while conventional wisdom, is the reverse of the truth. As we detail in our paper, Keynes devised various mechanisms for the financing of public works expenditures, including short-term loans at very low rates of interest (Bradburys and TDRs), and for the reduction of the long-term rate of interest. In doing so, Keynes’s fiscal and monetary policies were complementary. The present strategy sets the two main levers of policy against each other – not normally a virtue.

Second, B and S appeal to the need to appease financial markets. Evidence from recent experience in Portugal, Greece and Ireland demonstrates the gulf between the announcement of an austerity package and its imposition. The former is universally met with euphoria in financial markets. However, once the package is imposed, the actual cuts in government demand damage activity and employment. By contrast, the performance of the German economy in particular must be seen against the backdrop of an ongoing expansion of government demand and other state-devised mechanisms to support private activity. Even for the UK, the impact of austerity lies in the future.

Third, while B and S call for substantially lower public expenditure and taxes, we ask them to look again at our analysis of developments since the Second World War. The statistics that we study in some detail show clearly that economic conditions have deteriorated substantially ever since the move to the more liberal environment; which began even before the famous Healey/Callaghan surrender to the IMF in 1976, especially in terms of the monetary conditions. This deterioration is not only apparent from the perspective of ‘real’ outcomes: after 1976 the uninterrupted improvement in the government debt ratio came to a halt.

We are most grateful to the editor of the RES Newsletter for the opportunity to raise these matters. We hope that others will now contribute to the fuller debate demanded by what Keynes called the ‘facts of experience’.

From issue no. 154, July 2011, pp.13-15.

Page Options