Reply to Barnett and Smith
Iwould like to thank Richard Barnett and Peter Smith for their comment on my paper on English local authority finance, which I found both useful and stimulating. In their analysis they make two points. The first is that, given the system of finance described in the paper, the regression results reported on local authority spending are misspecified since they take no account of the kink in the budget constraint faced by local authorities. This is of course correct, and in subsequent work with James Gordon, we did use the econometrics of kinked budget constraints to look at the behavior of English local authorities in more detail.1 In the earlier paper, however, I was not attempting to build a model of local authority behavior. Rather, I was trying to illustrate the fact that local authorities did appear to alter their spending, depending on the differing tax prices they faced due to the system of local authority finance then in place; hence, the graphs of the relationship between ratable value and spending provided in the text. The regression results reported were simply meant to illustrate this relationship evident in the graphs, a point which I probably did not emphasize sufficiently in the text.
Barnett and Smith’s second point is that, while the system of local authority finance described in the text is accurate for the years before 1989/90, in 1989/90 itself (the year for which I analyzed spending data). the government administered the grant as a lump-sum payment. Hence, the tax price faced by the local authorities in that year was not related to the ratable value per head, but to the percentage of total ratable value attributable to domestic rate payers. As with most things connected with English local authority finance, the situation is somewhat complicated.
What occurred in 1989/90 was that the old grant system, described in my paper, was administered until July 1988. At this point, as part of the transition to the new system of finance, the grants were frozen and paid out as a lump sum. Hence, local authorities faced a change in the system of allocating central government grants in July 1988. It might therefore be expected that local spending would depend upon both the tax price associated with the old system of local authority finance administered until July 1988 and with the system in place after July 1988. A simple ordinary-least-squares regression (available from the author) for local authority spending in 1989/90 using my original data set confirms this intuition; both tax prices turn out to be significant explanatory variables.
While I accept that the regression is misspecified, for the reasons noted above, it would appear that local authority spending in 1989/90 depended on both tax prices, not merely the lump-sum payment regime in place after July 1988. This implies that the results I reported explain part, but only part, of the behavior of English local authorities in the year in question. Overall, the conclusion would appear to be that my choice of the fiscal year 1989/90 for an empirical analysis of local authority spending was an unfortunate one.