Munich Personal RePEc Archive Discretionary tax shocks in the United Kingdom 1945-2009: a narrative account and dataset Cloyne, James S University College London 2010 Online at https://mpra.ub.uni-muenchen.de/34913/ MPRA Paper No. 34913, posted 21 Nov 2011 15:58 UTC Discretionary Tax Shocks in the United Kingdom 1945-2009: A narrative 1 account and dataset James Cloyne2 First Draft: July 2010 This draft: November 2010 Abstract This paper constructs a narrative account of all legislated discretionary policy changes in the United Kingdom from 1945 to 2009. Following Romer and Romer (2009, 2010), evidence of the policymakers’ motivation is presented from U.K. official Budget documents together with technical notes, press releases, Acts of Parliament, the Budget speech by the Chancellor of the Exchequer and related entries in the parliamentary record (Hansard). The historical context in which the decision was made is also discussed. Using the given motives I isolate tax policy changes which were not responding to, or influenced by, current or prospective economic shocks. This ‘exogenous’ category is comprised of actions to improve long-run economic performance, those motivated by ideological or political reasons, rulings from external bodies such as courts, and fiscal consolidation measures based on long-run considerations. By contrast, the ‘endogenous’ changes are actions to manage demand, to stimulate production, to offset a debt crisis and those to fund spending decisions. For all the tax changes I collect information on the announcement, implementation and withdrawal dates as well as the type of the tax (such as income tax). The dataset contains nearly 2,500 tax changes and is aggregated into a quarterly series for analysis. In addition to creating a novel dataset this paper also contributes to the post-war history of U.K. taxation. 1 I am grateful for comments and advice from Wendy Carlin, Chris Carroll, Liam Graham, Nicola Pavoni and Morten Ravn, as well as seminar participants at University College London. I would also like to thank Carl Emmerson at the Institute for Fiscal Studies and the librarians at the London School of Economics and Her Majesty's Treasury. 2 Department of Economics, University College London, Gower Street, London, WC1E 6BT. E-mail: [email protected]. 1 Introduction Despite its importance for current macroeconomic policymaking, evidence of the macroeconomic effects of tax shocks in the United Kingdom is sparse. This gap is reflected in the newly created Office for Budget Responsibility’s first report from June 2010. The tax multipliers used by the OBR are derived, in part, from an IMF survey paper from 2009. Of the nineteen studies reviewed by the IMF only two specifically examine the U.K. The OBR's other multiplier assumptions come from common large-scale macro-econometric forecasting models which often crucially depend on modelling assumptions.3 While the academic literature has focused on the United States and cross country panel datasets, there is no consensus of the effects of tax changes. This reflects the difficulty of identifying tax policy shocks uncorrelated with, and uncontaminated by, other fluctuations. One major problem with common tax measures is simultaneity. Changes in taxes are likely to contemporaneously affect GDP but commonly used tax variables such as tax revenues are also contemporaneously driven by GDP. Despite the importance of resolving this identification problem, current approaches are limited. One popular method is that of Blanchard and Perotti (2002). This seeks to cyclically adjust the change in overall revenues, net of transfers, for changes in GDP. The method assumes policymakers are not informed about, or are unable to respond to, shocks within the same quarter. External information is used to calibrate the elasticity of output to GDP. A residual term can then be constructed and, under the timing assumptions, it can be interpreted as the discretionary policy decisions uncorrelated with other fluctuations. However, this approach is not without its problems. First, the timing assumptions may not hold in reality. Second, we need to be confident that revenues have been adjusted for all the possible cyclical influences. As many factors are likely to affect revenues, it is unclear what a comprehensive list would be. Third, legislated tax shocks are not simply shocks to revenues; they alter rates and liabilities, which themselves are likely to affect the output elasticity which Blanchard and Perotti (2002) assume to be constant. This paper addresses the identification problem directly by constructing a new measure of the tax policy shocks in the United Kingdom that should be uncorrelated with current or projected economic fluctuations. The source for this dataset is the narrative record in the United Kingdom. I therefore pursue a narrative identification approach following Romer and Romer (2010). Other ‘narrative approaches’ have also been used to identify government spending shocks (Ramey and Shapiro (1998); Ramey (2008)) and monetary policy shocks (Romer and Romer (1989, 2004)). To construct the narrative dataset, the first step is to collect direct measures of all the legislated tax policy changes in the United Kingdom between 1945 and 2009. The main source for these data is the official Budget documents. I then employ the Romer and Romer (2010) strategy of classifying tax changes by motivation. This allows me to identify those decisions that were taken for reasons uncorrelated with current or prospective economic conditions. I follow Romer and Romer (RR) in calling these actions ‘exogenous’ tax changes. Actions which do not satisfy this criteria are referred to as ‘endogenous’. There are nearly 2,500 discretionary policy changes over the period. Each Budget usually had both overall objectives as well as individual motivations for the specific measures. Motives for all discretionary changes are collected from sources such as the Budget speech by the Chancellor of the Exchequer and related entries in the parliamentary record (Hansard). A variety of motives are given for actions and I divide the exogenous and endogenous groups into subcategories. In doing so, I keep as close as possible to the stated motivation. This generates slightly different subcategories from those in RR. The ‘exogenous’ 3 Blanchard and Perotti (2002) argue “the evidence from large-scale econometric models has been largely dismissed on the grounds that, because of their Keynesian structure, these models assume rather than document a positive effect of fiscal expansions on output”. 2 category is split into actions to improve long-run economic performance, ideological changes related to party political or social causes, rulings from external bodies such as courts, and fiscal consolidation measures based on long-run considerations. The endogenous changes contain actions to manage demand, to stimulate production, to offset a debt crisis and those to fund spending decisions. In addition to classifying each discretionary policy action, I also collect information on their announcement dates, their implementation and end dates, their revenue effects and their type (e.g. income tax). In the narrative below I present evidence for the classification of all the major changes. I also provide an historical context, informed both by the policymakers’ own statements and the history literature. By constructing a new measure of tax changes for the United Kingdom this paper provides a rich new dataset dealing with the identification problem. The constructed series is so useful precisely because it isolates the policy changes which are uncorrelated with current and prospective economic shocks. This then facilitates further research for a country where surprisingly little is known about the macroeconomic effects of tax changes (see Cloyne (2010) for an analysis employing the new dataset). A number of factors make the U.K. ideal for a narrative approach. Firstly, the U.K. has made considerable use of fiscal policy post-war. Secondly, tax policy is highly centralised4 and, since the Budget is a major annual event, tax changes are largely saved for this announcement with implementation taking place throughout the year. Moreover, unlike in the United States, these announcements almost always become law. Third, detailed revenue forecasts are provided for all the Budget measures. Fourth, given the extensive political debate around the Budget, there is considerable discussion of the motives. The central contribution of this paper is to provide a new measure of tax shocks in the United Kingdom. The first part of this paper discusses the construction of the new dataset. I also discuss the more technical details, the assumptions made and how the 2,500 changes are aggregated. The second part of this paper then presents the narrative. The centrepiece of tax policy in the United Kingdom is the Budget; there was at least one a year over the period and the narrative is therefore presented chronologically by Budget. For each Budget, I consider the historical context in which the policymakers’ decisions were made. I then present evidence on the overall objectives and motivation, followed by evidence of the specific motives for individual measures. Instruments and implementation of tax policy in the United Kingdom The centrepiece of the British tax process is the annual Budget. This is a traditional and grand occasion which attracts extraordinary media coverage in spite of its technical nature. Part of the attraction is the rhetoric