17 I Introduction What determines tax policy? What motivations do governments follow in tax reforms? Do voters react to tax reductions and tax increases? Our knowledge of these questions is still very limited. Investigations of the determinants of tax policy and tax reforms are – especially in the empirical economic literature – very rare. This results largely from the lack of reliable data that makes it hard to test different economic and polit-economic theories of tax policy empirically. Based on a new data-set of tax reforms in Germany from 1964 to 2004 (including more than 1,000 new tax policy regulations, their fiscal impact, and their timing with respect to adoption and implementation) we want to contribute to filling this gap in the empirical literature. We offer empirical tests of central economic hypotheses on tax reform (derived from normative and positive theories of public finance) as well as of hypotheses on the influence of tax policy on elections. What can we learn about tax policy and taxation from our new data-set? What are the factors that determine tax policy? Is taxation just driven by expenditure needs? Or dominate partisan or electoral motivations? Does tax policy look different in times of divided government? And how does the electorate react to different tax policies? These are central questions of this book. Our analysis proceeds as follows. We start by a general presentation of the German tax system and its development (part II). Then we move on to a description of the data-set employed and test the reliability of the data (part III). Part IV analyzes tax reform patterns in the main German taxes from 1964 to 2004. In part V – the main part of our analysis – we derive hypotheses from different theoretical approaches to taxation which we test empirically based on our data-set. We start by discussing the influence of international tax competition on German tax reforms (V.1). Then we move on to normative theories of taxation. Here we analyze the influence of expenditure considerations, of macroeconomic stabilization goals, of cold progression, and of tax smoothing (part V.2). Part V.3 discusses and tests politeconomic hypotheses based on the theory of inertia, the theory of fiscal illusion, opportunistic government motivation, partisan considerations, and divided government. Finally we discuss and test the economic theory of voting based on our data (part V.4): how do voters react to different tax policies? Part VI concludes. In our analyses we are able to show that the data-set employed consists of reliable estimates of the fiscal effects of tax reforms, and we derive new insights into the pattern of tax reforms. We show that normative approaches have only a very limited explanatory impact for the tax reforms analyzed here. With respect to politeconomic theories we find that divided government matters – but in the opposite direction of the “gridlock-hypothesis”: tax reforms are larger and more frequent in times of divided government. In contrast to many qualitative studies we do not find evidence for partisan politics, but for opportunistic behavior of governments. How- 18 ever, the governments´ attempts to manipulate re-election probabilities by tax reductions in election years fail: our analyses show that voters react strongly to tax burden changes, but take the effects of tax reforms within the whole legislative period (and not just in election years) into account.