Home Political science
Fiscal Squeezes Identified and Compared
As explained in Chapter One, we conceive of fiscal squeeze as concerted effort by politicians and governments to impose losses on some individuals or groups by cuts in public spending and/or increases in state revenue to correct the public finances. In that chapter we distinguished that approach to 'austerity' from the more conventional focus on episodes when public debt or deficits were significantly reduced, given that our conception of fiscal squeeze represents effort rather than outcome in terms of fiscal correction. And we outlined two ways of identifying fiscal squeeze in that sense. One is to look at historical statistics for episodes of absolute or relative reductions in public spending or increases in revenue, distinguishing revenue from spending squeezes and 'hard' from 'soft' squeezes ('hard' squeezes meaning spending falling and/or revenue increasing both in absolute terms and relative to GDP). The other way of identifying fiscal squeeze is to look more qualitatively at the record of events (using items such as budget speeches, published histories, archival records, or interviews) to assess how much loss or deprivation is imposed on the population; the extent to which incumbent politicians have to expend their political capital or risk their reputations as they try to sell, broker, or fix fiscal squeeze policies with political parties, interest groups, or voters at large; and the degree of effort put in by the state apparatus in contriving ways to raise more revenue or reduce spending.
We turn to that latter approach in the following eight chapters and the conclusion. But this chapter uses historical statistics to identify episodes of spending reductions and tax increases as a proxy for the effort put into fiscal squeeze.
Any such venture immediately brings us to three unavoidable complications. First, 'the historical record' is often far from simple or unambiguous, not least because alternative statistical sources using different definitions of items such as 'public spending' may and frequently do show different results. As Richard Rose (1980: 205) puts it, 'Whether public spending is said to be rising or falling is a matter of definition, not fact [emphasis in original]. Nor are many statistical series fully consistent over more than a short time period, so we have to take into account changing definitions and conventions to make valid long-term comparisons. Further, statistical series are commonly altered retrospectively and new definitions emerge of key items such as 'deficit'. So (as, for example, in 1931) the picture of an episode that we get from retrospectively compiled numbers may diverge sharply from what decision-makers faced at the time. Even where that does not happen, reported numbers reflect the point at which spending restraint or revenue increases were implemented, not the point at which such changes were announced, and yet the politics of fiscal squeeze often centres as much on the latter as the former. That lag can matter, as we shall see later in the book.
Second, as we have already seen, squeezes may differ according to whether the pain is inflicted on the revenue side or on the spending side or both, and according to whether spending goes down or taxes up relative to GDP and/or in absolute (constant price) price terms. Is 'squeeze' more meaningful, difficult, or painful when it involves relative or absolute change? We think that could be argued either way a priori, and much seems to depend on whether GDP is growing, stable, or falling.
Third, following our discussion in the previous chapter about the need to differentiate fiscal squeeze from ordinary budgetary politics and to avoid making distinctions that are not meaningful given likely measurement error (for example, between a period when taxes rise by 0.001 per cent of GDP and another when they fall by a similarly infinitesimal amount), we need to specify thresholds for what is to be counted as a squeeze.
We deal with the first complication by comparing different data sources to see how much the timing, existence, or extent of fiscal squeezes varies between different sources. We deal with the second complication by analysing changes in revenue and expenditure both relative to GDP (the ratio method) and in constant prices (the 'levels' method). As already explained, we define a squeeze as 'hard' if spending falls or revenue rises on both of those measures; and 'soft' if spending falls or revenue rises on only one of them.1 And we deal with the third complication by setting a threshold of significance. We defined a squeeze on the ratio method as a fall in spending or rise in revenue, or both, recorded in historical statistics which was sustained for at least two years with an average annual change (fall for spending, rise for revenue) of not less than one percentage point of GDP. We defined a squeeze on the levels measure (that is, changes in spending and revenue in constant-price terms) as a fall in spending or rise in revenue of not less than
1 per cent for a single year or at least 1 per cent on average for two or more
Any such threshold of course involves some 'rule-of-thumb', but for the ratio measure at least it broadly follows the conventional literature on fiscal consolidation, and reflects a view once said to have been commonly held by the IMF, that a reduction of public spending by about 1 percentage point of GDP per year was the normal limit of politically feasible spending squeeze in most countries (Hood, Heald, and Himaz (2014): 11-12 and 29).
Within those analytic settings, we identify numerous periods of fiscal squeeze (of the sixteen types identified in Table 1.1 of Chapter One) over the century in the UK, and the Appendix to the book provides summary tables and graphs comparing those episodes. Tables A1 and A2 in the Appendix  
show that the instances of revenue and spending squeeze identified from financial outcome statistics are not highly sensitive to which particular historical data source is chosen, in that those squeezes show up, albeit with slight variations in depth, duration, and timing, in most of the available statistical sources. As for differences between the 'ratio' method and the 'levels' method of measuring squeeze (as discussed earlier), the episodes are roughly the same for spending squeezes following either method, as Table A1 in the Appendix shows, while Table A2 shows that the same does not apply on the revenue side.
Having established that, the analysis of squeeze episodes in the rest of this chapter is based on two separate sources of UK financial data, namely the well- known dataset compiled by Brian Mitchell (1988) for the period up to 1949 and data from the UK Office of National Statistics (ONS) (ONS 2014 and IFS 2014) for the post-1949 period. Similarly, having explored episodes using both the ratio and levels method, as shown in the Appendix, we base the timings of the various squeezes on the ratio method in the rest of this chapter.
Most of the episodes defined using the ratio method comprised expenditure and revenue changes well above the thresholds we specified earlier. Two cases were marginal. One is a fiscal squeeze under the post-World War II Labour Government, starting in the late 1940s, that straddles the two different historical datasets we used. If we combine those two different data sources (ONS (2014) and Mitchell (1988)) that squeeze ends in 1949/50, but if we use another data source (Middleton (1996)) it extends to 1951/52. We follow the combination approach here, but the point at which that squeeze ended is undeniably ambiguous and Chapter Five looks carefully at qualitative accounts of the Labour Government's efforts at fiscal restraint over this period. The other case, a spending fall in the late 1960s (involving major defence cuts) under Harold Wilson's Labour Government, falls right on the margin. We chose to include it here and discuss it in Chapter Six, but that is another of those line-ball categorization decisions.
Accordingly, Table 2.1 summarizes the fiscal squeezes (as defined above) that can be identified between 1900 and 2015. If we include all the times (derived from the ratio method) when there was a spending squeeze, revenue squeeze, or both (column 2) we can identify eighteen episodes in total.
a The ratio method refers to defining episodes by looking at percentage point changes to spending/GDP and revenue/GDP. The years refer to financial years. e.g., 1916-18 refers to 1916/ 17-1918/19. bTotal percentage point change during episode divided by the number of years the episode lasted.c Total percentage change during episode in constant prices divided by the numberofyearsthe episode lasted.d Coded yes or partly ifdeficit/GDP improved on average by more than 1 percentage point; No otherwise. eAsexplained in thetext, identified by changes in spending and revenue as proportions of GDP and in constant price terms over each squeeze episode.
Source: Spending: See TableA1 note a; Revenue: See TableA2 note a. There is a break in the data at 1949, due to differences in sources.
However, as indicated in column 1 of that table, some of those periods can more meaningfully be grouped together, as representing continuing effort by the same government (as with the Thatcher squeezes from 1979 to the late 1980s), or as responses to the same problem or policy initiative by more than one government (as with the implementation of the 'Geddes Axe' in the early 1920s or the responses to the deep financial crisis of the early 1930s). Grouping the episodes in that way produces twelve episodes over a century— slightly more than one a decade.
As Table 2.1 shows, those fiscal squeezes varied in length, depth, and type, raising the question of what accounts for that variation. It also raises the question of what makes governments put the stress on spending cuts or on increasing revenue, and why 'double hard' squeezes were apparently so rare over this period.
What Table 2.1 also shows (in the seventh column) is a marked difference between episodes of fiscal squeeze as identified from revenue increases or spending reductions and episodes of 'fiscal consolidation' as identified by significant reductions in deficit. In the previous chapter we underlined the difference between looking at 'austerity' in terms of deficit or debt reduction outcomes as against efforts to increase revenue and/or cut public spending. As can be seen from Table 2.1, while there were no episodes of significant deficit reduction that did not involve fiscal squeeze as defined earlier, by no means all episodes of fiscal squeeze in that sense involved deficit reduction. Indeed, as that table shows, if we define 'austerity' episodes as those involving 'substantial' deficit reduction (here taken to mean a reduction of not less than 1 percentage point of GDP), half of the eighteen fiscal squeeze episodes in Table 2.1 would disappear from view.
Moreover, as Table 2.2 shows, if we use deficit reductions as the criterion for identifying squeezes, the start and end times of the episodes as well as their length is markedly different, and in general the duration of those episodes is shorter. So it really does matter which approach we take to identifying episodes of austerity, and as we show in Chapter Eleven, conclusions about the electoral effects of austerity efforts can turn on whether we define such efforts in terms of substantial deficit reduction or of changes in revenue and spending.
In the following subsections we pick up two of the three analytic themes that Chapter One introduced and comment on apparent changes over time (we leave the third theme, on management of blame, to be considered in Chapter Eleven following our narrative chapters). One is the overall mix of expenditure cuts and revenue increases in fiscal squeeze, together with an apparent diminution of 'hard tax squeezes' towards the end of our period.
Table 2.2. Episodes of deficit reductions using various definitions and data sources 1900-2015
Note: Numbers in brackets refer to the average percentage point in deficit relative to GDP over the episode.
Source: a Derived from Hills, Thomas, and Dimsdale (2015), calendar years, b ONS public sector finances aggregates databankApril 2015. See http://budgetresponsibility.org.uk/data/#databank.
The other is the length and depth of squeezes, with an apparent shift from 'short and sharp' spending squeezes to longer and shallower ones over the period.