LUND UNIVERSITY

PO Box 117221 00 Lund+46 46-222 00 00

Essays on the Politics of Taxation

Andersson, Per F.

2017

Document Version:Publisher's PDF, also known as Version of record

Link to publication

Citation for published version (APA):Andersson, P. F. (2017). Essays on the Politics of Taxation. Lund University.

Total number of authors:1

General rightsUnless other specific re-use rights are stated the following general rights apply:Copyright and moral rights for the publications made accessible in the public portal are retained by the authorsand/or other copyright owners and it is a condition of accessing publications that users recognise and abide by thelegal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private studyor research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal

Read more about Creative commons licenses: https://creativecommons.org/licenses/Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will removeaccess to the work immediately and investigate your claim.

Essays on the Politics of Taxation

Per F. Andersson

DOCTORAL DISSERTATION

by due permission of the Faculty of Social Science, Lund University, Sweden. To be defended at Hornsbergssalen. Date 2017-05-19 and time 10:00.

Faculty opponent

Reader Lucas Leemann Essex University

Organization LUND UNIVERSITY Department of Political Science Box 52 SE-221 00 Lund, Sweden

Document name DOCTORAL DISSERTATION Date of issue

May 2017

Author(s) Per Fredrik Andersson

Sponsoring organization

Title and subtitle Essays on the Politics of Taxation

Abstract Taxation is a key activity of any state and a constant point of political struggle. The structure of taxation is continuously changing and evolving, and its size has grown dramatically during the last two hundred years. Many of the cross-national patterns we observe today are the result of centuries old conflicts and challenges, hence we need to take history into account if we want to understand contemporary tax systems. This dissertation is concerned with the evolution of taxation in the last two centuries. During this period modern parliamentary democracy developed and spread, and it is during this period that the contemporary party systems crystallized and the broad lines of conflict between the left and the right emerged. Thus, this period is crucial for our understanding of the effects of political institutions and ideology on policy making. Because of a lack of comparative information on taxation with a long time scale, previous research has been constrained to a small number of mostly European countries. In this dissertation I present a novel dataset (a collaboration with Thomas Brambor) over government tax revenues covering 31 countries from 1800 to 2012. The dataset is unprecedented in both temporal and geographic scope and includes countries from Europe, North America, South America, as well as Australia, New Zealand, and Japan. The first paper is concerned with the notion that democracy itself has an impact on taxation by extending influence to previously excluded groups of poor citizens. I present evidence for the argument that the effects of democracy depend on urbanization. Democratization in an urbanized state leads to higher taxes on income and lower taxes on consumption. In contrast, democratization in rural countries is associated with lower taxes on property. The second paper investigates the influence of ideology on taxation. A puzzle in the literature is why left-wing parties are associated with regressive taxation (e.g., on consumption). I argue that how left-wing governments tax depend on the institutional environment. In countries using majoritarian/plurality electoral systems the left relies more on income tax, and in countries using proportional representation systems the left relies more on consumption tax. In the third paper I investigate the mechanism behind left-wing tax strategy in more detail by studying reforms of consumption taxation in post-war United Kingdom and Sweden. I find that strategic considerations related to how the political system concentrates power in the United Kingdom affected the Labour Party’s attitude towards the value-added tax and its decision not to adopt the tax. The left-wing government in Sweden on the other hand, operating in a different institutional context, decided to introduce a similar tax. The fourth paper, which is coauthored with Johannes Lindvall, contrasts political investments, of which taxation is one example, with short-term crisis management. We present a game theoretic model in which institutions that concentrate power are better at handling sudden crises but worse at making policy with short-run costs and long-term gains. Power-sharing institutions, on the other hand, are better at resolving inter-temporal dilemmas, but perform worse when faced with sudden crises.

Key words Taxation, democracy, economic history, redistribution, political economy Classification system and/or index terms (if any) Supplementary bibliographical information Language

English ISSN and key title ISSN:0460-0037; Essays on the Politics of Taxation

ISBN 978-91-7753-229-3 (print) 978-91-7753-230-9 (pdf)

Recipient’s notes Number of pages 247

Price

Security classification I, the undersigned, being the copyright owner of the abstract of the above-mentioned dissertation, hereby grant to all reference sources permission to publish and disseminate the abstract of the above-mentioned dissertation.

Signature Date 2017-04-18

Essays on the Politics of

Taxation

Per F. Andersson

Cover photo by the author

© Per F. Andersson

Faculty of Social Sciences and Department of Political ScienceISBN ���-��-����-���-� (print)ISBN ���-��-����-���-� (PDF)ISSN ����-����

Typeset in LATEX

Printed in Sweden by Media-Tryck, Lund UniversityLund ����

Acknowledgements

It is almost a tradition to begin this section by noting that the processof writing a dissertation is a lonely one. My experience has beendi�erent. From day one I have been working closely together with themembers of the Reform Capacity of Governments team. Working withmy research in this group has undoubtedly made the dissertation betterand the process more much more enjoyable than working alone. Thehead of the group, Johannes Lindvall, was also my main supervisor anddeserves the largest share of my gratitude budget. With his feedback,intellectual curiosity, and great wisdom Johannes has been a constantsupport during these years.

I have had the privilege to have Jan Teorell as my second supervisor.His support and input have improved my work tremendously. HannaBäck has read and commented on my work several times and her advicehas always been very helpful.

I also owe many thanks to Thomas Brambor, a member of thereform capacity team and collaborator on the database on which muchof this dissertation is based, for valuable input and many interestingdiscussions about social science. Carlo Knotz, the final member of theteam, deserves my gratitude for reading and commenting on everythingI have produced since starting the program. Although my improvementshave been sluggish at best, he never gave up. Tomas Bergström, HannaBäck, Lina-Maria Ellegård, Mats Fred, and Svend-Erik Skaaning readthe entire manuscript and their suggestions were crucial in helping mefinalize the dissertation. Thank you.

Next, I wish to thank several people who generously commented andgave advice on parts of the dissertation, in particular Despina Alexiadou,Lucy Barnes, Jens Bartelson, Deborah Boucoyannis, Björn Bremer,Douglas Brommesson, Rasmus Broms, Johan Bo Davidsson, Carl-HenryGeschwind, Emma Hällqvist, Michael Klitgaard, Vytautas Kuokötis,

Victor Lapuente, Klas Nilsson, Linda Nyberg, Kenneth Shepsle, SvenSteinmo, Kitty Stewart, Mariano Tommasi, Sten Widmalm, and BjörnÖstbring. I apologize to those I might have forgotten to mention.

I also want to thank the people who are in some way responsible forme ending up in a doctoral program. Eva Mörk and Matz Dahlbergdemonstrated to me that social science is not only an important andfruitful endeavor, but is also great fun. Furthermore, I want to expressmy gratitude to Khalid Medani who encouraged me early on to pursuea Ph.D. in Political Science and to Kåre Vernby for pointing me in thedirection of Lund University.

Generous financial support from Stiftelsen Lars Hiertas Minne, KarlStaa�s Stiftelse, Jacob A Letterstedts resestipendiefond (through Kungl.Vetenskapsakademien), and the Swedish Network for European Studiesin Political Science has allowed me to travel to numerous conferences andarchives around the world. I also want to thank Ben Ansell for arrangingfor me to spend a term at Oxford University. Hossein Ordobouadianprovided excellent assistance with text editing and proofreading.

Finally, I thank Emma Hällqvist for her support, endless patience,and for helping me to keep academic pursuits in proper perspective.

Contents

� Introduction �

Taxation and the State: A Short History �

Political Institutions �

Taxation: Causes and E�ects ��Taxation and Modernization, �� – Taxation and Economic Growth, �� –

Taxation and War, �� – Taxation and Democracy, ��

Taxation and Constitutions: How the Papers Fit Together ��Actors and Preferences, �� – Institutions, �� – Summary, ��

A Note on Research Design ��

Taxation in the World ����-����: The Government Revenue Dataset��Data Collection and Coding, �� – Patterns of Taxation: ����-����, ��

Summary of Papers ��

Contributions and Future Research ��

Bibliography ��

� Papers ��

Paper I: Democracy, Urbanization, and Tax Revenue ��

Paper II: Left-wing Tax Strategy Depends on the Electoral System ���

Paper III: Political Institutions and Taxation by the Left: Britain andSweden Compared ���

Paper IV: Urgent Problems, Long-Term Problems, and Political Insti-tutions ���

�

Introduction

T

axation is not only intimately related to the emergence of themodern state, but is also an ideal policy for a study of the impact of

political institutions. This dissertation presents new data and analysisof the emergence of the modern tax state using a novel dataset spanningfrom ���� to ����. Moreover, it uses tax structure to study how democ-racy and democratic institutions shape the struggle between economicclasses over policy. Using a range of methodological approaches, thesefour papers cast light on di�erent aspects of the politics of taxation,from the long-term e�ects of democracy to the impact of ideology andleft-wing tax strategy to a normative analysis of the performance ofpolitical systems.

William H. Riker once claimed that if “[p]olitics [...] is the au-thoritative allocation of value”, then, political science is the study ofdecision-making (����, p. ��f). This dissertation is concerned withdecision-making with respect to tax policy. Taxation is particularlyinteresting for at least two reasons. First, no state can exist withoutrevenue; that is, as long as taxation is used to finance the state, deci-sions about taxation are inevitable. Second, because taxation is by itsvery nature about allocating value, conflicts will arise concerning howtax obligations will be distributed.

This centrality of revenues – and in particular tax revenue – asa reflection of a society’s values was famously stressed by Joseph A.Schumpeter:

The spirit of a people, its cultural level, its social structure,the deeds its policy may prepare – all this and more iswritten in its fiscal history, stripped of all phrases. He who

� introduction

knows how to listen to its message here discerns the thunderof world history more clearly than anywhere else ([����]����p. ���).

Similarly, the Austrian sociologist Rudolf Goldscheid held that “thebudget is the skeleton of the state stripped of all misleading ideologies”,pointing out that taxation is crucial to understanding the rise of themodern state.� If Schumpeter and Goldscheid are right, much can begained from studying taxation during the last two centuries, an era thatsaw dramatic changes not only in the extent of taxation (i.e., taxationgrew from being a relatively minor activity in the nineteenth century torepresenting around half of the economy in modern industrial countries)but also in economic and political organization.

The industrial revolution is one of the largest transformations inhuman organization since the development of agriculture and sedentarycivilization. Apart from profoundly changing the economic structureof societies and leading to an unprecedented increase in human well-being, it also altered the way the state and society interacted. The neweconomy increased the demands from the state, from providing basicinfrastructure to financing ever costlier wars. Industrialization alsoamplified the process whereby increased dependence of the state on taxrevenue was met with increasing demands for representation (Bates andLien ����; North and Weingast ����; Schumpeter ����). The inclusionof a wider part of society in decision making culminated in modernrepresentative democracy with elected assemblies and accountableexecutives.

The first question asked in this dissertation pertains to the rela-tionship between tax systems and democracy. It has been claimedthat democracy, by including the previously disenfranchised masses ingovernment decision-making, a�ects the way states tax as democraciesbegan to rely on a progressive and redistributive taxation structure.However, the empirical evidence for this hypothesis is mixed. Myfirst paper is concerned with this issue and suggests that the reasonfor earlier conflicting results is that the poor have been treated as ahomogeneous group with identical preferences with respect to taxation.By tying tax preferences to sectors of the economy such as agricultureand manufacturing, I derive more specific testable hypotheses.�

Another set of explanations for the variation of taxation around theworld focuses on ideology, in particular, the rise of left-wing political

�Quoted in Schumpeter [����]����, p. ���.�All papers are summarized in section �.�.

�

parties during the late nineteenth and early twentieth century. Inter-estingly, scholars have failed to find a consistent influence of ideologyon economic policy in general and on taxation in particular. Thisunresolved issue is the second puzzle in this dissertation. In Paper �and Paper �, I use insights from the literature on institutions to suggestthat the e�ect of ideology depends on institutional context.

The third part of the dissertation is concerned with what makestaxation di�erent as a policy and what this means for the performanceof constitutions. The idea is that not only is taxation a central policyfor “stateness” in itself but also belongs to a class of policies commonlyreferred to as “political investments”. Investments are policies with short-run costs but long-run gains. As such, taxation presents the politicalsystem with challenges that are very di�erent from other policies. InPaper �, I and Johannes Lindvall contrast political investments withsituations where swift decisive action is needed in a formal modelevaluating the relative performance of constitutions that concentratepower compared to constitutions that share power.

Together, the four papers in this dissertation contribute to ourunderstanding of the relationship between taxation and political insti-tutions. It does so in four distinct ways. First, it presents evidencethat the influence of democratization is not uniform but depends onthe urbanization rate. Second, it argues that the e�ect of ideology ontax policy is di�erent depending on the institutional context. Third, itextends our theoretical understanding of how di�erent constitutionsa�ect the ability of governments to make investments (of which taxa-tion is a subset) and handle urgent crises. Finally, it presents the firstcomprehensive data set over government tax revenues from ���� totoday covering �� countries in Europe, North America, South America,Australia, New Zealand, and Japan.�

I have confined my analysis to taxation as a domestic politicalproblem driven by distributive conflict. This means I am not concernedwith tax compliance (e.g., Levi ����; Timmons ����), the relationshipbetween taxation and economic growth (Auerbach ����; Widmalm����), or international factors such as tax competition (Genschel ����)or interstate warfare (e.g., Hintze ����; Scheve and Stasavage ����;Tilly ����). My work is also restricted to the nineteenth and twentiethcenturies, a time when industrialization and urbanization led to a

“threat of concentrated working-class collective action” (Tilly ����, p.��) as well as a novel political environment.

�The data were collected and compiled together with Thomas Brambor and isdescribed in more detail in section �.�.

� introduction

This introductory chapter presents the general scholarly contextto which the four papers speak and outlines the common theoreticalunderpinnings of the papers. I begin by providing a brief overview oftaxation and institutions – i.e., what they are and what they meanin the context of the papers. This discussion is followed by a sectionon previous research. In the subsequent section, I describe the overallempirical strategy, outlining my methodological starting point and howthe di�erent methods used in the papers complement each other. Thelast part of the chapter presents the dataset as well as an empiricalexploration of the history of taxation and institutions. The chapterends with a summary of the contributions made and suggested avenuesfor future research.

�.�

Taxation and the State: A Short HistoryTaxation is together with law and order one of the “main attributesof sovereignty and the most visible demonstrations of the power ofauthority” (Brélaz ����). As such, it is at the heart of a core conceptin comparative politics, states. Max Weber famously defined a state asa “human community that (successfully) claims the monopoly of thelegitimate use of physical force within a given territory” (����, p. ��),but maintaining the monopoly of violence is not free, and the state isoften (but not exclusively) financed by taxes.

According to Schumpeter, the modern state evolved from a conditionwith no clear distinction between the government budget and the privateeconomy of the monarch. Increasing spending pressure as a result ofwar meant that the private revenue of the monarch could no longersustain the increasingly larger armies needed to defend the state, forcingthe ruler to rely on the aristocracy for additional funds. Since thefunds were needed to finance a public good (i.e., defense), the rulerand the nobility had a common interest. However, the increasingdemands for revenue were met by calls for influence over how it wasspent. Eventually, this agreement developed into what Schumpetertermed a “tax state”, where most citizens contributed a share of theirincome to the state (Schumpeter [����] ����).�

�Extending this narrative, many scholars have linked demand for revenue todemocratization (see section �.�).

taxation and the state: a short history �

Schumpeter’s well-known narrative of the development of the taxstate in Europe from the Middle Ages suggests a rather linear evolutionwith war as the ultimate cause; however, the general mechanism ofwarfare leading to increased government expenditures (and as a conse-quence, a need for more revenue) goes back much farther. In AncientGreece, the costs of maintaining large fleets led to a sharp increase inthe cost of war, which in turn meant heavier taxation. Recent work evenmakes explicit comparisons between Athens and the Schumpeteriantax state (Gabrielsen ����).�

Although pre-industrial states were much smaller compared totoday’s states, the challenges associated with paying for the state weresimilar. There are many ways to finance government activities:

�. The state might control important resources (e.g., diamonds, oil,copper, trading routes) that generate revenue without imposinga direct cost on its citizens.�

�. The government can charge fees for specific services providedby it, for example, the use of a bridge or a highway system or itmight charge a fee for the lawful transfer of property.

�. Most states can also borrow money to finance its activities.Importantly, since the loan generally needs to be repaid, thistype of revenue is short term and must involve some of the othersources in the long term.

�. Finally, the state can impose a tax to fund its activities.�

All these types of revenues di�er in distributive impact (i.e., who orwhich groups pay more or less) and how it regulates the relationshipbetween citizens and the state. For example, a fee implies a verydi�erent relationship than a tax, as the latter involves more discretionon part of the government (which needs to be complemented by truston the part of the tax payer).

�Section �.� provides a brief review of the literature linking taxation to warfare.�However, the use of many of these sources exclusively by the government

introduces a number of indirect costs, for example increased prices associated withmonopolies.

�The state can also choose to privatize a certain activity, by which a service isprovided and paid for without the involvement of the state. Throughout history,many activities that we today associated exclusively with the state – such as defense– has been privatized (at least in part) (for examples from the Byzantine empire seeTreadgold (����), ch. �-�).

� introduction

This dissertation is concerned with the last of these sources ofrevenue – taxation – and how decisions regarding taxation are structuredby political institutions.� The point of departure is that politicalinstitutions shape the outcome of the distributive conflicts stemmingfrom the contentious bargaining over taxation.

Before moving on, we have to establish what a tax really is. Accord-ing to Supreme Court Justice Oliver Wendell Holmes Jr., taxes “arewhat we pay for civilized society”� and Alt (����) claims that “theanswer to the question ‘Why have a tax?’ is the same as the answer tothe question ‘Why have a state?’ ”(p. ���). In other words, taxationis the cost paid by citizens in exchange for a state. A slightly di�erentdefinition is used by the World Bank (����): “[Taxes are] unrequited,compulsory payments collected primarily by central governments”. Thisclearly delineates taxation from fees in exchange for specific servicesand natural resource rents. While succinct and capturing the key ele-ments of what a tax is, this definition hides one of the most importantproblems tax reform seeks to redress: evasion and avoidance. Notingthat including the payment of a tax in its definition e�ectively ignorescompliance, Martin et al. (����) propose the following definition:“Taxation consists of the obligation to contribute money or goods tothe state in exchange for nothing in particular” (p. �, emphasis inoriginal).��

Throughout history, states have taxed a variety of things, frombeards and musical events to oars and champagne. A fairly commonway to categorize di�erent taxes is to begin by distinguishing betweendirect and indirect taxes. Direct taxes are levied on a person’s incomeor wealth paid directly by the taxpayer, whereas indirect taxes arelevied on goods and services. Direct taxes include taxes on income,land, corporations, and profits and indirect taxes include tari�s, excise,and broad-based consumption taxes.��

To continue the discussion of these categories of taxation andhow they are related to politics, I have to introduce the concept ofdistributive impact. To understand the preferences of voters andpolitical parties as well as the challenges associated with investment

�Non-tax revenue is also an important topic in itself. For a recent study of itsrole in early modern Sweden, see Nilsson (����).

�In Compania General de Tabacos de Filipinas v. Collector of Internal Revenue(����) , p. ���

��The empirical definition used when assembling the data is similar in spirit, butmore detailed. See section �.�.

��Section �.� shows how revenue from these di�erent taxes have changed overtime.

taxation and the state: a short history �

dilemmas, it is essential to understand how taxes a�ect di�erent groupsof citizens

Distributive impact describes how heavily the tax falls on citizensdepending on their income or wealth. Progressive taxes are used toreduce the tax incidence of people with lower ability to pay and byextension require higher income individuals to pay a larger percentageof their income; regressive taxes, on the other hand, have the oppositee�ect – the relative tax burden decreases as one’s ability to pay increases.The distributive impact of a tax is also called its “incidence”. It iscommon to distinguish between “statutory incidence”, indicating who islegally obliged to pay the tax, and “economic incidence”, indicating whoreally pays the tax, or more specifically “[t]he change in the distributionof real income induced by the tax” (Rosen and Gayer ����, p. ���).For example, a tax on the sale of a good might be legally paid by thefirm selling it, but if the tax induced the firm to raise the price of thegood, the ultimate payer of the tax is the consumer. Exactly how muchof the tax is absorbed by the firm and how much is passed on to theconsumer is of course an empirical question. By “incidence”, I mean

“economic incidence”.To simplify theoretical analysis, it is common to make assumptions

about the incidence of di�erent types of taxes. Direct taxes on incomeand property are regularly assumed to fall on income earners andproperty holders, respectively, and are classified as “progressive”, whileindirect taxes on consumption are assumed to be passed on to consumersand are considered “regressive”. Although incidence is often assumed,there is empirical evidence in support of consumption taxes beingregressive and income taxes being mainly progressive (Joumard, Pisuand Bloch ����; Prasad and Deng ����). The progressivity/regressivityof taxation has always been politically contentious since the conceptsdescribe who pays for the state and is related to what is considered

“fair” and “just” taxation.With the rise of democracy, the incidence of the tax system became

linked to the distributive basis for ideological di�erences between politi-cal parties. These di�erences are related to changes in social cleavagesas a result of industrialization and are particularly interesting duringthe period this thesis considers (����-����). During this period, thenew urban working class emerged as a political force, represented by left-wing parties and trade unions. These two groups are the main actorsin the second and third paper, where I focus on left-wing governments.I elaborate more on the relationship between economic and political

� introduction

change and how this is related to preferences and representation – andultimately tax policy – in section �.�.

In sum, the distributive influence of di�erent taxes a�ects thepreferences of voters and is a main point of conflict between politicalparties. The outcome of these conflicts is guided by political institutions.

�.�

Political InstitutionsMy focus is on how democracy and di�erent kinds of democracy a�ectthe way distributive conflicts are resolved. The main theme follows theearly insights of Knut Wicksell. Wicksell noted that how the state isfinanced is closely connected to how the state is ruled and that thereare great gains to be had by analyzing them together (I elaborate moreon this approach to the politics of taxation in section �.�). How a stateis ruled is commonly codified in a constitution. These codified lawsare formal political institutions. Electoral institutions are those lawsregulating how voter preferences are aggregated and turned into taxpolicy. Douglass C. North defined institutions as follows:

Institutions are the humanly devised constraints that struc-ture political, economic and social interaction. They consistof both informal constraints (sanctions, taboos, customs,traditions, and codes of conduct), and formal rules (consti-tutions, laws, property rights). (����, p. ��).

Although informal institutions are important (they have beenstudied extensively in the literature on the politics of taxation),�� Ifocus on formal institutions.�� There is of course almost as many typesof democracy as there are democratic states. Moreover, the types ofdemocracies are constantly changing and what is meant by the veryconcept “democracy” constantly changes. The last issue is of lessconcern since I restrict myself to the period between ���� and ����

��For a recent attempt to assess the relative importance of informal institutionsin the context of explaining tax compliance, see Zhang et al. (����).

��Importantly, while I treat institutions as exogenous and concern myself withhow they moderate the impact of ideology, they are also themselves the result ofpolitical, strategic choice. That is, actors deciding whether to democratize, andwhat institutions to employ, take the expected e�ects of these institutions intoaccount (Acemoglu and Robinson ����; Ahmed ����; Boix ����; Leemann andMares ����).

political institutions �

and to modern representative democracies.

Democracy as a method of governing a state is a key concept in all fourpapers. The first paper is concerned with the impact of democratizationand the other three papers focus on variation within democracies. Myargument focuses on how democracy extends influence to previouslydisenfranchised groups. Defining and classifying regime types has beena central part of political science since Plato’s four (and later six) typesof political regimes.�� Plato’s work and later refinements by Aristotleremained the starting point for analysis of political systems well intothe nineteenth century (Almond ����).

What a “democracy” really is has been an issue of continuousdiscussion within the discipline, and the definition used often variesdepending on the question asked. One influential definition is fromRobert A. Dahl, who proposes five criteria for a maximal/ideal democ-racy: e�ective participation, voting equality, enlightened understanding,control of the agenda, and inclusion of adults (����, ch. �). In realworld large-scale democracies, Dahl instead talks about “polyarchy”,which has six minimum requirements: elected o�cials; free, fair, andfrequent elections; freedom of expression; access to alternative sourcesof information; associational autonomy; and inclusive citizenship (ibid.p. ��).

While Dahl’s definition covered a range of factors important for aworking democracy, others have gone in the opposite direction, definingdemocracy with a minimum number of conditions. For example, JosephA. Schumpeter defines democracy as “that institutional arrangement forarriving at political decisions in which individuals acquire the power todecide by means of a competitive struggle for the people’s vote” ([����]���� p. ���). Similarly, democracy, according to Przeworski et al.(����), is “a regime in which those who govern are selected throughcontested elections” (p. ��). In their empirical operationalization, theyfurther emphasize that the executive o�ce and legislature are filledin competitive elections whose results are respected and that they arefree to rule without responding to some other body or person (suchas a monarch). To be classified as a democracy, this process must berepeated and power must be peacefully transferred according to theoutcome of competitive elections.

��The first four were timocracy, oligarchy, democracy, and tyranny, and thelater expansion consisted of monarchy, tyranny, aristocracy, oligarchy, democracy,ochlocracy.

�� introduction

Boix, Miller, and Rosato ���� have a related approach, but theyare more directly influenced by Dahl. They focus on contestationand participation. For a country to be democratic, they believe anexecutive should “directly or indirectly [be] elected in popular elections”and should be “responsible either directly to voters or to a legislature”and the legislature (or executive) should be “chosen in free and fairelections.” As for participation, to be democratic the franchise needsto be expanded to a majority of adult men, a factor not emphasizedby Schumpeter (ibid. p. ����).�� These last concepts are dichotomous,but democracy is also frequently conceptualized and measured on acontinuous scale (e.g., Polity (����) and Freedom House (����)).��

In sum, there are di�erent ways to think about what democracy is,both theoretically and empirically. Since the key aspect of democracyin my argument is the inclusion of previously excluded groups, it isimportant that the definition includes participation. For this reason, Iused the Boix, Miller, and Rosato ���� definition (and data) in thefirst two papers.��

There is considerable institutional variation within the group of demo-cratic countries. Although my focus is on one institution (the electoralsystem), I will briefly present some of the more important di�erencesbetween modern representative democracies.

• The first is whether the head of government is elected in separateelections (like most presidents), or if he/she is appointed by theparliament. The first system is called presidentialism and thelatter parliamentarism.

• Another distinguishing feature is the degree of centralizationwithin a country. Some states consist of a federation of smallerunits, often represented nationally in a separate assembly (e.g.,the United States Senate or the German Bundesrat). Othercountries are centralized, unitary states.

��Excluding female su�rage runs the risk of bias if women have di�erent pref-erences than men (see Lott and Kenny (����)). In the first paper, I use analternative definition of democracy where full male and female su�rage is requiredas a robustness test. For more details, see Paper �.

��For longer reviews of di�erent empirical measures see Boix, Miller and Rosato(����), Hadenius and Teorell (����), and Munck and Verkuilen (����).

��Empirically, I am restricted by data availability. Many indices measuringdemocracy do not cover the time period from ����. See Paper � for more details.

political institutions ��

• Some countries have two assemblies, usually referred to as an“upper” and a “lower” house. This feature is called bicameralism.Examples of upper houses include not only the Senate and theBundesrat (representing territorial units) but also the SeanadÉireann in the Republic of Ireland (a unitary state). The impor-tant factor is that this is an additional body with influence overpolicy.

• Another key institutional feature is whether there is judicialreview performed by an independent constitutional court.

• A final distinguishing feature is the type of electoral systemused. The two main families are majoritarian and proportionalrepresentation systems.

Due to space limitations, I have just mentioned these institutionsvery briefly. For longer theoretical and empirical treatments, see Li-jphart (����), Persson and Tabellini (����), and Powell (����).

The institutional dimensions just mentioned are often used tocategorize democratic states. For example, Lijphart (����) distin-guishes between “consensus” and “majoritarian” democracies, andPowell (����) divides democracies based on how well they fulfill theproportional and majoritarian “visions”. A key element in these andsimilar categorizations is the electoral system, an institution playingan especially important role in my argument in Papers � and � (butalso in Paper � although in somewhat di�erent role).

The term “electoral system” is broad and sometimes refers to slightlydi�erent things. For example, it might include how political partiesinternally select candidates, how the executive is selected, and electionsat di�erent levels of government. I use a rather narrow conceptionroughly equal to what Rae (����) and Cox (����) mean by “electorallaw”: the methods used for translating votes to seats in elections for themain legislative assembly (these include district magnitude, electoralthresholds, and electoral formula). The reason I restrict myself tothis particular election is that assemblies are at the core of politicalpower in democratic societies. Moreover, one of the key functions ofthe parliament is to make decisions pertaining to the budget. Indeed,one of the first tasks for parliaments historically was to approve newtaxes.��

��See Burns and Kamali (����) for an overview over the historical evolution ofparliaments.

�� introduction

A key function of electoral systems is how votes are translated intoseats. The results of this is often interpreted in terms of proportionality– i.e., how the distribution of votes among parties/candidates corre-sponds to the distribution of seats. The electoral formula, the districtmagnitude, and electoral thresholds all a�ect the proportionality of rep-resentation.�� I follow earlier literature in distinguishing between twomain types: majoritary/plurality (winner-take-all) and proportionalrepresentation systems (PR), what has been called “the most funda-mental dividing line in the classification of electoral systems” (Lijphart����, p. ���) (see also Colomer (����), ch. �; Persson and Tabellini(����); Powell (����); and Rae (����)). For example, Lijphart (����)holds that “[m]ost electoral formulas fit the two large categories of PRand plurality-majority, but a few fall in between” (p. ���) (Lijphartfound only a couple of elections, both in Japan, that do not fit in thesetwo main categories.) While there are di�erences between them (e.g.,simple majority versus plurality and varying district magnitude), theimportant di�erences are between the two broad categories rather thanwithin them (Rae ����).

The typical form of majoritarian system is single-member districtplurality. In this system, the nation is divided into a number ofgeographical districts electing one candidate. The candidate witha plurality of the votes wins the seat.�� The votes cast for othercandidates are thus not represented.

The type of electoral system is strongly linked to the number ofpolitical parties. In an early analysis of electoral systems, MauriceDuverger ���� predicted that plurality rule in single-member districtswould lead to two party systems while proportional representation(PR) in multi-member districts would result in multi-party systems.Single-member plurality is associated with two-party competition fortwo reasons. The first is strategy on part of political parties and voters;the latter do not want to waste their votes on candidates with a smallor no chance of winning a seat and the former have no incentive tofield candidates that are not competitive. The second reason is that

��There are ways of summarizing several aspects of electoral systems into onemeasure, for example, Lijphart’s e�ective threshold. However, countries generallyfall into roughly the same three groups as they would if one were to focus only onthe electoral formula: majority/plurality, PR, and mixed systems (see for exampleTable �.� in Powell ����).

��In countries using a “majoritarian formula” instead of a plurality, it is commonto have a run-o� election between the two top candidates if no candidate reachesan absolute majority in the first round such as in French and American primaryelections.

political institutions ��

the formula translating votes to seats limits the number of feasiblecandidates. More specifically, the district magnitude sets an upperlimit on the number of parties/candidates. This limit is the number ofseats in the district plus � (Cox ����). This means that PR does notdeterministically lead to multi-party systems; it just allows it. Anotherimportant result reported in Cox (����) is that the e�ects generallyoperate on the district level. This means, for example, that morethan two parties may be viable nationally in a single-member pluralitysystem if a party has strong regional support.

In PR systems, seats are allocated in proportion to votes received.��

The district magnitude has the opposite e�ect here compared to pluralityrule: the larger the district, the higher the proportionality.

The important aspect in my theoretical framework is how theelectoral system a�ects the concentration of power and oppositioninfluence. The di�erence between the two systems with respect to thisaspect is stark: majority/plurality systems concentrate power in themajority, excluding minorities from influence, whereas PR leads tobroad representation of diverse groups, who often rule in coalitions. Forexample, in Lijphart’s (����) analysis, over ��% of the governments inplurality-majority systems were manufactured�� majorities, comparedto only �.�% in PR systems. Moreover, ��.�% of governments in PRsystems were natural minorities. As Powell (����) (ch. �) points out,even without strategic voting, majoritarian/plurality rules in single-member districts manufactures legislative majorities, often with largedisproportionalities between number of votes and number of seats.This is elaborated upon in Paper � and Paper �. The next sectionreviews earlier literature on taxation and its relation to modernization,economic growth, interstate warfare, and democracy.

��The most common type of PR is a system where voters cast ballots for partylists. But there is also the so-called mixed-member PR and the single transferablevote. Moreover, there are many di�erent mathematical formulas that are used toallocate votes to parties. The interested reader is referred to Lijphart (����) (ch.�) for a brief overview.

��Majorities that are a direct result of how the electoral system transform votesto seats.

�� introduction

�.�

Taxation: Causes and E�ectsTaxation – and public finance more broadly – is a central componentin many of the earliest treatises on statecraft, from Pseudo-Aristotle’sOikonomikos and Kautilya’s Arthashastra in the fourth and secondcentury BCE, respectively, to Ibn Khaldûn’s The Muqaddimah in thefourteenth century CE. Later works include Francesco Guiccardini’sLa Decima Scalata in sixteenth century Florence and Adam Smith’sThe Wealth of Nations in eighteenth century Scotland. The moreimmediate predecessors are the late nineteenth and early twentiethcentury economist Knut Wicksell and his intellectual successors ErikLindahl, Richard Musgrave, and James M. Buchanan, but also thefiscal sociology tradition with key contributions by Rudolf Goldscheidand Joseph A. Schumpeter. Lately, this topic has generated renewedinterest (e.g., Martin, Mehrotra and Prasad (����); Monson andScheidel (����); Yun-Casalilla and O’Brien (����)).

The link between taxation and the development of modern stateshas attracted researchers from sociology, economic history, economics,and political science. In this overview, I have divided contemporaryresearch into four main camps, relating taxation to modernization,economic growth, interstate warfare, and democratization. However,many of the topics and authors cross the artificial boundaries set uphere, and the categorization is made more for presentational purposesthan to reflect clearly defined intellectual borders.

taxation and modernization

How taxation and institutions are related have long been studied bysociologists, economists, and political scientists. One of the earliestempirical investigations of the determinants of tax structure was Hin-richs (����). Building on social mobilization theory (e.g., Deutsch����), Hinrichs proposed that countries go through di�erent stages oftaxation during the course of economic development. Direct taxes onland dominate at very low levels, but with modernization countriesshift to taxes on trade up to a certain level where the most modernstates go back to a revenue system that relies mainly on direct taxes(in particular on income). Importantly, economic development doesnot lead deterministically to one particular tax system; moderniza-tion makes more tax tools available for the state, leading to greater

taxation: causes and effects ��

variability of tax systems among developed states. Richard Musgravedescribes this change in tax systems as follows: “If the availability oftax handles places a constraint on total expenditures in low incomecountries, it may be expected to do so even more with regard to thetax structure mix. This constraint again loosens as per capita incomerises, and in high income countries the composition of the revenuestructure becomes a free policy choice” (���� p. ���). Hinrichs arguesthat the observed variability is a function of the rather opaque conceptof “cultural-political” preferences (���� p. ��.).

Modernization is also related to administrative capacity, anotherimportant factor in determining tax policy choice. Aidt and Jensen(����) argue that an e�ective income tax can only be implemented if acertain level of tax collection technology is reached, and this technologyin turn is a function of the level of education, urbanization, and theexistence of an updated census. Similarly, Co�gel and Miceli (����)posit that the evolution of tax collection methods can be traced tochanges in bureaucratic capacity, costs of measuring the tax base, costsof monitoring, and the variance in the tax base. Others have tiedeconomic development and tax systems together by highlighting lowfiscal capacity as a key impediment for economic growth. Investing intax capacity is a crucial challenge for developing countries (Besley andPersson ����, especially chapter �). Importantly, solving the investmentdilemma depends on how political institutions shape the distributionof power in the future (ibid.).��

taxation and economic growth

Another dimension of the politics of taxation is concerned with howtaxation is related to incentives, economic growth, and compliance. Iftaxes are too high, people will choose to produce less or try harder toavoid/evade paying the taxes they owe. This observation is one of theoldest – and most recurrent – in the historical literature on taxation.

Writing at the dawn of the industrial revolution, Adam Smithstressed that “[e]very tax ought to be so contrived as both to takeout and to keep out of the pockets of the people as little as possible,over and above what it brings into the public treasury of the state”([����] ����, p. ���). Raising tax rates above a certain point reducestax revenues since it also decreases productive activity and encourages

��The fourth paper of the dissertation concerns a related problem when consid-ering di�erent democratic institutions, but finds that there is a trade-o� betweeninvestments and crisis management.

�� introduction

avoidance and evasion.�� And a couple of centuries later, Schumpeterargued that “the tax state must not demand from the people so muchthat they lose financial interest in production or at any rate cease to usetheir best energies for it” and “there exists a level beyond which furthertax increases mean not an increase but a decrease of yield” ([����]����,p. ���-���). In short, it has long been recognized that “just the rightamount” is key to a sound tax policy. Importantly, Schumpeter linkedthis to constitutions when he argued that a broader political base allowsfor a broader tax base (ibid.).

The point that taxation a�ects overall output has spurred an enor-mous wealth of research in optimal taxation (for early contributions seeRamsey ���� and Mirrlees ����). This tradition usually analyzes taxa-tion with the goal of maximizing social welfare. A common omissionin optimal taxation theory is administrative costs, an important issuein real world tax policy reform. This means, for example, that whilein (some) theories di�erent goods should be taxed at di�erentiatedrates according to their demand elasticities (e.g., Ramsey ����), a realworld optimal consumption tax involves fewer rates since administrativecosts increase with the number of di�erent rates (Slemrod ����). Evenmore problematic is that optimal taxation theory in general ignores thepolitical process behind tax policy and does not consider the possibilitythat political actors may maximize something else than social welfare(Slemrod ����; Wicksell [����] ����). Since political actors not onlyare interested in taxing e�ciently but also are concerned with the dis-tributive impact of taxes and how the revenue is spent, it is unrealisticto expect these actors to implement optimal taxes from the point ofview of aggregate social welfare. This is especially important in democ-racies where tax policy is usually the result of struggle and compromisebetween di�erent groups typically represented by political parties ratherthan the enlightened decision of a benevolent social planner. In short,taxation is just another tool politicians use to win elections. Thereis evidence of this view in recent work on consumption taxation. Forexample, James (����) argues that although the modern value-addedtax (VAT) was introduced mainly with reference to e�ciency, the realVATs actually introduced were far from the theoretical constructs inthe optimal taxation literature and more the result of political conflictand compromise. For example, although the (modern) theoretical idealprescribes one (or a few) uniform rate over all goods and services, real

��This relationship is known to most social scientists as “the La�er curve”, namedafter the American economist Arthur La�er (Wanniski ����), but was pointed outalready by Ibn Khaldun in the ��th century.

taxation: causes and effects ��

world VATs are characterized by di�erentiated rates and exemptions(ibid.).

taxation and war

While only indirectly related to institutions, this overview would beincomplete without mentioning the impact of interstate warfare. Theargument is that war provides a way out of the vicious circle of low taxrevenue and low state capacity through the extreme demands war putson the treasury. As mentioned before, Schumpeter saw war as the maindriver behind the evolution from the domain state of medieval Europeto the modern tax state. A version of this argument is that war wasan urgent factor forcing rulers to make representative concessions inexchange for taxation, setting in motion the “taxation for representation”dynamic that ended in democracy (Tilly ����).��

Important contributions stress the importance of war or war pressureas factors leading to increased levels of taxation as well as inducinggovernments to invest in new types of taxes (Besley and Persson ����;Herbst ���� ch.�; Hintze ����; Karaman and Pamuk ����; Scheve andStasavage ����; Tilly ����).��

However, it is not necessarily warfare itself that provides the shockto which the state responds with tax reform. Based on a compara-tive historical case study of England, Japan, and China, He (����)argues that severe fiscal crises, which could be but does not have to beconnected to warfare, led to the centralization and modernization offiscal systems. Which taxes were centralized depended on the economicstructure at that time and place. Moreover, centralization made itpossible for the state to leverage long-term loans to finance furtherreforms of the revenue system.��

However, war (or a crisis) is by no means a su�cient or evennecessary condition. Schumpeter (rather dramatically) pointed outwith respect to the New Deal in the United States that “irrespective ofthe war, a tremendous transfer of wealth has actually been e�ected, atransfer that quantitatively is comparable with that e�ected by Lenin”

��The link between taxation and democratization is discussed in more detail inthe next section.

��The a�nity between warfare and government revenue was even emphasized byThucydides in History of the Peloponnesian War. For an excellent analysis of therelationship between war and state finances in Thucydides, see Kallet-Marx (����)and Kallet (����).

��It has also been claimed that economic crises have been a catalyst for reformin more recent times (see for example Drazen and Grilli (����)).

�� introduction

(Schumpeter ���� p. ���). Many countries also introduced the incometax when there was no war. For example, when Sweden introduced anincome tax in ����, it was the influence of peasants that was crucial(Åkerman ����).�� Another powerful critique of the link betweentaxation and warfare is that this view has mainly been informed bythe European experience. Centeno (����), for example, has questionedthe validity of the argument in Latin America. A further issue is pathdependency. Morgan and Prasad (����) argue that warfare tends toamplify existing patterns of taxation rather than changing the systemcompletely. That is, war leads to an expansion of existing taxes, not theimplementation of new ones. Finally, there are other ways to financewar. For example, tying tax payments directly to the war e�ort throughallotment systems (or similar methods) turns the tax into a fee andresistance to taxation can be reduced.

taxation and democracy

The slogan “no taxation without representation” is commonly associatedwith the complaints of American colonists in the mid eighteenth century,leading up to the American Revolution.�� The notion that expansionof taxation led to democratization is widespread in explanations of therise of the modern state in Europe (see North and Weingast (����)and Schumpeter [����] (����)). This literature points to the fact thatincreased levels of taxation led to rising demands for representationthat in turn led to democracy (important contributions include Batesand Lien (����), Herb (����), Moore (����), and Ross (����), butsee also Boucoyannis (����) for a di�erent view).

A related idea is that of the “resource curse”: an abundance ofnatural resources impedes democratization since it allows the state tooperate without having to resort to taxation (Ross (����,����)). A ver-sion of this argument is that democracy in itself reflects a commitment

��This was possible because the fourth estate – the peasants – needed to approvethe new constitution. And without a new constitution Sweden could not negotiatea peace with Russia (since the king had recently been deposed in a coup d’etat).That is, the peasants had a unique opportunity where they actually had influence,and they took it. The tax was removed only a couple of years later, although it didyield significant revenue (Åkerman ����). Importantly, the tax was not introducedbecause of revenue needs associated with war, but because of peasant influence andtheir tax preferences.

��Importantly, it was not taxation itself that sparked the protest known asthe Boston Tea Party, but unjust taxation. Specifically, the British Parliamenthad recently passed a tax exemption for the East India Company, an exemptionunavailable for the American colonists (Thorndike ����, p. �).

taxation: causes and effects ��

to redistribution. The existing political elite in a nondemocratic societywants to prevent a possible revolution from the lower classes but cannotuse redistribution as appeasement since the promise of redistributionis not credible over time. An extension of political rights changes theequilibrium distribution of power and therefore serves as a commitmentdevice leading to more redistribution and a lower risk of unrest (Ace-moglu and Robinson ����; Boix ����). However, democracy mightnot be stable if the level of inequality (and thus redistribution) is sohigh that the former elite mounts a coup (Acemoglu and Robinson����, ����). This rendering of the argument has been criticized byscholars who claim that intra-elite competition is the main determinantof democratization (Ansell and Samuels ����).

Some of the models of democratization mentioned above, in particu-lar Acemoglu and Robinson (����) and Boix (����), are based on theanticipated impact of democracy on taxation. That is, the behaviorof the elite in non-democracies is contingent on what they expect thee�ect of democracy on taxation will be. This expectation is basedon the notion that demand for redistribution is a main determinantof tax policy (Alt ����). One of the most influential theories linkinginstitutions to tax policy posits that democracy leads to redistribu-tion. The argument is based on the median voter theorem (Black ����;Downs ����; Hotelling ����) and the distribution of income amongvoters. In a democracy, the median voter sets the tax rate, and if theincome of the median voter is lower than the mean income, the ratewill be redistributive. This means that for a given inequality of income,extending su�rage to the poorer segments of society will lower the in-come of the median voter and thus increase redistribution (Meltzer andRichard ����; Romer ����). While theoretically straightforward, theempirical evidence is mixed, with some reporting results in support ofthe theory and others rejecting it (Ansell and Samuels ����; Lee ����;Lindert ����; Mares and Queralt ����; Mueller and Stratmann ����;Scheve and Stasavage ����). In my first paper, I build on this literature,arguing that the e�ect of democratization on redistribution dependson the tax preferences of the previously disenfranchised, preferencesthat vary depending on urbanization.

Within democracies, scholars have highlighted several importantinstitutional features for tax policy outcomes. For example, propor-tional representation (PR) electoral systems have been linked to largergovernments and more redistribution (Austen-Smith ����; Iversen andSoskice ����; Persson and Tabellini ����; Steinmo and Tolbert ����).However, recent evidence suggests that this is a modern development,

�� introduction

reporting a negative e�ect of PR in the time before the Second WorldWar (Aidt and Jensen ����). The impact of electoral systems is alsoclosely connected to geography since district size and formula havedi�erent e�ects depending on the spatial distribution of voters (Rod-den ����, ����). Earlier research emphasizes the distinction betweenpresidential and parliamentarian systems (Hettich and Winer ����),where presidential countries are associated with smaller governmentsand less redistribution than parliamentarian countries (Persson andTabellini ����).

While political institutions certainly matter, the input side of politics(e.g., the ideology of political parties) also plays a role. Scholarshave explored a range of topics such as the impact of ideology onthe size of government, unemployment, and welfare spending (e.g.,Cameron ����; Hibbs ����; Hicks and Swank ����). Among the findingsare that left parties are associated with larger governments (Tavits����), more redistribution (Bradley et al. ����; Iversen and Soskice����), and (under some conditions) higher taxes on labor (Beramendiand Rueda ����; Cusack and Beramendi ����). While this suggeststhat partisanship matters, others find that government ideology hasno impact (see Imbeau, Pétry and Lamari ���� for a meta study).There is also recent historical research questioning the relationshipbetween the left and redistribution. For example, Scheve and Stasavage(����) found that mass warfare was an important determinant of taxprogressivity during the First World War in Europe, but they foundno e�ect of left party influence. Others have shown that, contrary toexpectation, the first income taxes were adopted by conservatives inpolitical systems with very limited su�rage (Mares and Queralt ����).Since tax systems are strongly path dependent (Morgan and Prasad����), this exemplifies the potential of historical research to contributeto these debates.

Focusing mainly on the post Second World War era, several scholarsemphasize the co-development of taxation and the welfare state (Bera-mendi and Rueda ����; Bradley et al. ����; Cusack and Beramendi����; Ganghof ����; Kato ����; Kemmerling ����; Steinmo ����).However, the nature of this relationship is contested. Some argue thatan e�cient tax system (in terms of generating revenue) facilitateda large state (e.g., Kato ����), whereas others claim that spendingpressure caused the state to increase revenue capacity (e.g., Ganghof����).

Somewhat ignored is the fact that the e�ect of democracy, espe-cially considering franchise, depends on the underlying distribution

taxation and constitutions: how the papers fit

together ��

of preferences. These preferences vary both between countries andwithin countries over time, making it problematic to assume a uniforme�ect of democracy on tax structure. The structure of the economyalso matters for how we should think about political parties (Lipsetand Rokkan ����). For example, farmers were an important force innineteenth century Scandinavia, and this group does not necessarilyshare the tax preferences of the urban working class (Baldwin ����).Moreover, how these cleavages a�ect policy is di�erent in democraciesand non-democracies (Olper ����; Thomson ����). As I will arguein the next section, it is important to be clear about the distributiveimpact of di�erent taxes and then link these to the preferences of thepreviously disenfranchised. In Paper �, I conclude that the influence ofdemocracy on tax structure depends on the geographic distribution ofvoters.��

The impact of ideology is not uniform either. How a party achievesits goals depends on the actions of other parties and how the rules ofthe game (i.e., political institutions) shape their interaction. Earlierresearch on the influence of ideology is either undertheorized or based onseveral assumptions not made explicit. In this dissertation, I clarify theconnection between preferences for equality and di�erent approaches toredistribution, and present quantitative (Paper � and �) and qualitativeevidence (Paper �) to support my argument.

�.�

Taxation and Constitutions: How the Papers FitTogether

My theoretical starting point is that rule by the many or rule by thefew matters. Moreover, how the state is ruled a�ects its goals. Ormore specifically, the rules of government (i.e., its political institutions)determine how those in power secure their power and how they interactwith other groups, which by extension relates to how ruling groupsdesign and implement policies.

Analyzing the state as an agent is common when focusing on compli-ance where the state and citizens engage in contracting and bargaining.For example, Levi (����) assumes that the ruler or the state has

��The general point that geography matters is not novel – see, e.g., Baldwin(����), Lipset and Rokkan (����), Rodden (����), and Thomson (����).

�� introduction

its own utility function with one argument, revenue maximization.This perspective is also common in the optimal taxation and welfareeconomics literatures. However, I follow the approach advocatedby Knut Wicksell by treating the state as an arena where di�erentgroups compete for power, groups that do not see the welfare of thesociety in general as their goal: “[T]he members of the representativebody are [...] precisely as interested in the general welfare as aretheir constituents, neither more nor less” (Wicksell [����] (����), p.��). From this viewpoint, political institutions define the contextin which the competing preferences of these groups are translatedinto policy. Moreover, the relevant bargaining is not between theruler and the people, but between di�erent representatives of thepeople. This approach sees politics and political conflict as the maininput variable, and political institutions as the process whereby theseconflicts are resolved. In the end, this conflict produces observabletax policy. In this section, I outline some intellectual forerunners anddiscuss important definitions and assumptions. In particular, I fleshout who the relevant actors are, what their preferences are, and whatinstitutional contexts regulate their interaction.

Following Wicksell, the starting point tying the papers together is thattaxation and political institutions are inextricably linked. Wicksellis famous for two ideas in the second essay of his FinanztheoretischeUntersuchungen: taxation and expenditure should not be separatedand the near unanimity rule. The first claim has been interpretedmainly in normative terms and influenced James Buchanan’s work onfiscal constitutions.�� The second claim – the near unanimity rule – canbe interpreted both normatively and positively. Wicksell, writing ata time when su�rage was severely limited in Sweden, states explicitlythat he expects an expanded franchise to lead to heavier taxes on therich because of increased influence of the poor, and uses this as anargument that the rich should reform the public finance process so thattaxation is tied to expenditure.

The modern research in this tradition was sparked by the seminalpapers by Romer (����) and Meltzer and Richard (����). Their modelsare in a sense refinements of the intuitive notion that democracy leadsto redistributive taxation. However, their iteration of the argument is

��The two main intellectual heirs to Wicksell are James M. Buchanan andRichard A. Musgrave, a�liated with the approaches public choice and public financerespectively. For a concise summary of these two approaches, see Buchanan andMusgrave (����).

taxation and constitutions: how the papers fit

together ��

constrained in that it is wedded to the median voter theorem as themain political mechanism translating preferences into policies, therebyignoring real world institutions. But, as Powell (����) has shown, realworld democratic constitutions vary considerably in their ability torepresent the median voter.

The papers in this dissertation are part of the same tradition,acknowledging the central importance of formal political institutionswith respect to taxation. In terms of theoretical structure, they aremore specific than Wicksell, but they are not wedded to a specific modeltranslating preferences to policy (such as the Meltzer and Richard (����)model). Instead, the papers emphasize the diversity of institutionsin the real world and how these institutions have real e�ects on taxsystems. The papers are also concerned with a specific, althoughfairly long, period (����-����), where factors such as industrialization,urbanization, and mass party democracy driven by modern ideologiesare central.�� As mentioned above, this is critical when discussing termssuch as “ideology”, “political parties”, and “democracy”. Moreover,as was pointed out by Tilly (����, ch. �), the threat of collectiveaction by the new industrial working class presented new challengesand opportunities for rulers.

On a general level, all papers in this dissertation use the sameunderlying model of politics. This model assumes that voters preferpolicies that align with their economic self-interests and act accordingly.Voters are represented by political parties. How elections are won – andby extension how voters’ preferences are represented – and how policyis made is determined by formal political institutions. It is importantto note that there is a distinction between parties’ preferences withrespect to taxation and their ultimate goals. It is common to identifythree key goals motivating parties: policy, o�ce, and votes (Müller andStrøm ����). My theoretical argument is on a lower level of abstraction,making it di�cult to distinguish between these three motivations. Anybehavior seemingly in line with a party being policy motivated canequally be the result of a party seeking o�ce or votes. Moreover, partiesmight seek votes to gain o�ce in order to a�ect policy. For example, if aparty has a preference for income taxation it can be consistent with theparty being motivated by policy but it can also be consistent with thenotion that policy is just an instrument to achieve the goal of attainingvotes and/or o�ce. For my argument it is not necessary to make anyassumptions regarding the ultimate goal of parties, the important factor

��Paper � is concerned with the entire period from ���� to ����, Paper � focuseson ����-����, and Paper � is concerned with the post Second World War era.

�� introduction

is that parties have divergent preferences over taxation and that thesepreferences are linked to those of the voters (for a discussion aboutdi�erent goals behind party behavior in general, see Müller and Strøm(����)).��

All four papers assume that actors behave rationally. Rationalbehavior can be seen as “behavior that shows [...] goal-directednessto a particularly high degree and in a particularly consistent manner”(Harsanyi ����, p. ��). More specifically, an agent is rational if he canrank actions based on the utility they yield, that he chooses the actionyielding the highest utility, and that if he prefers A over B and B over C,then he prefers A over C. Rationality as internal consistency of choiceis used by classics such as Arrow (����) but also by contemporarytextbooks like McCarty and Meirowitz (����), usually referred to ascomplete and transitive preferences.��

While the rational choice approach has not performed well whentrying to explain individual human behavior such as turnout, it hasbeen more successful in explaining party competition and economice�ects of constitutions and ideology (Bendor et al. ����, ch. �). As-suming rationality brings order to actors’ behaviors that are neededfor theoretical analysis. This assumption is especially important inPaper � (which uses game theory) since “even in very simple gamesituations (involving only two players and having a rather simple logicalstructure in other respects) it may be virtually impossible to suggestreasonable hypotheses about the players’ likely behavior without havinga clear systematic idea of what it means to behave rationally in therelevant class of game situations” (Harsanyi (����), p. ��). If we do notmake any assumptions regarding the behavior of agents (i.e., if theiractions are random with respect to their goals), then an analysis of howinstitutions a�ect outcomes of distributive conflict becomes di�cult, ifnot impossible. In the words of Harsanyi:

[W]e cannot really understand and explain a person’s behav-ior (or indeed the behavior of another intelligent organismor even of an intelligent robot) unless we can interpret iteither as rational behavior in this particular situation or asan understandable deviation from rational behavior. For ex-ample, we may be able to interpret it as an understandable

��Assumptions with respect to the basic motivation of parties are more importantwhen analyzing coalition formation and minority governments (e.g., Strøm ����).

��McCarty and Meirowitz (����) distinguishes between “thick” and “thin” ratio-nality, where the former includes specifying the goals of actors while the latter isjust about behavior, not the goals as such.

taxation and constitutions: how the papers fit

together ��

mistake, as an understandable emotional reaction, or as anunderstandable intentionally suboptimal response. (����,p.��, emphasis in original).��

In sum, political agents are assumed to act rationally to attaintheir own goals, not societal welfare. These rational actors often findthemselves in what I refer to as a “strategic situation”. A strategicsituation is “when anyone premises her instrumental actions on what sheexpects some relevant others might do” (Johnson ����, p. ���).�� Thepolitics of taxation reflect strategic interdependence where institutionsregulate how players interact.

actors and preferences

When considering the preferences of political actors and who theyrepresent, it is important to be precise. In fact, one of the main causesof confusion and debate in the literature on tax policy is a lack ofclarity on this point. A recurrent theme in my dissertation (especiallyin Paper � and Paper �, but also in Paper �) is that preferences relatedto taxation, in particularly with regards to redistribution, are oftenignored or treated too simplistically by earlier research.

In his ���� Nobel Lecture, James M. Buchanan (����) called forbehavioral consistency, arguing that analyzing actors in the market-place should be no di�erent than in politics: “[T]he choice behaviorof the individual is equally subject to the application of analysis inall choice environments. Comparative analysis should allow for predic-tions of possible di�erences in the characteristics of the results thatemerge from market and political structures of interaction” (p. ���).��

But, as mentioned in the previous section, this idea is much older asWicksell argued that in a democratic context government policy isnot the outcome of a benevolent social planner but of self-interestedparties. Parties need to bargain in order to reach their goals, and thecompromises as the result of political negotiations in no way guaranteesthat the best interest of the population is prioritized.

��An example of an “understandable intentionally suboptimal response” is achess move, since humans lack the cognitive abilities to calculate the optimal movein a complicated game as chess every move is intentionally suboptimal (Harsanyi(����), p. ���).

��Note that this is di�erent from a decision-theoretic situation in which theprobabilities used in expected utility calculations are exogenous (i.e., set by “nature”)(Tsebelis ����).

��This is also the approach taken by Anthony Downs in his An Economic Theoryof Democracy (����).

�� introduction

I treat political parties as representatives of economic classes. Ana-lyzing tax policy in terms of class conflict is not new, and this conflictis commonly analyzed in terms of demand for redistribution (Alt ����).Regardless of whether one is interested in the e�ect of democratizationor the impact of ideology, one needs to make assumptions about how taxpreferences are formed and how they are translated into policy. In hisseminal Essays on Taxation, Edwin Seligman points out the importanceof conflict between economic classes such as farmers, urban workers,landed elites, and capitalists when explaining modern tax systems. Infact, Seligman holds that “[t]he history of modern taxation is largelythe history of these class antagonisms” (����, p. ��, cf. Hinrichs ����).

A common starting point in the literature is to consider only twogroups: the rich and the poor. The poor wants to redistribute resourcesfrom the rich to themselves and since the poor are more numerous thanthe rich, democracy will lead to the poor controlling government andimplementing redistributive policy (see Acemoglu and Robinson ����;Meltzer and Richard ����). Implicit in these models are assumptions re-garding the tax and spend systems, their impact on economic inequality,as well as assumptions about the identity of relevant groups and theirpreferences. Consequently, these models perform best in a frameworkin which government policy is also very simple (pro-rich or pro-poor).However, tax systems are much more complex, and so are economicinterests in society. First, “the rich” is not a homogeneous group withidentical preferences over taxation; for example, landed elites havedi�erent economic preferences than urban elites (Ansell and Samuels����; Mares and Queralt ����). Moreover, I present evidence in thefirst paper suggesting that the urban poor have di�erent preferencesthan the rural poor and that this di�erence a�ects how democracy func-tion. Similarly, ethnic heterogeneity influences how states, specificallydemocracies, prioritize social spending, suggesting that ethnic identityamong the poor matters (Jensen and Skaaning ����). That is, theactors in the first paper are voters, and their preferences are influencedby sector (urban or rural).

In Paper � and Paper �, the actors are political parties representingeconomic classes. To discuss parties’ tax preferences, tax incidence – i.e.,what groups carry the burden of a specific tax – is key. As mentionedabove, it is common to distinguish between “progressive taxes” – fallingrelatively heavier on the rich – and “regressive taxes” – falling relativelyheavier on the poor. Traditionally, income and property taxes areviewed as progressive taxes, whereas consumption taxes are viewedas regressive taxes. A simple model assumes that parties to the left,

taxation and constitutions: how the papers fit

together ��

which typically presume to represent the poor, prefer to shift taxationfrom regressive to progressive taxes. But this is too simplistic. Wealso need to consider the impact on total tax revenue and by extensionthe spending programs in place. That is, more tax revenues meanmore funds available for government programs, and the e�ect of theseprograms can more than make up for the distributive e�ect of taxation(Engel, Galetovic and Raddatz ����). As Alt (����, p. ���) notes that“[a]ny political account of redistribution must deal with the fact thatthose seeking redistribution may choose to pursue it through benefitsor taxation” (my emphasis). This general argument – that one cannotfocus solely on taxes or expenditures when considering the impactof a tax – was articulated by Wicksell in ����, and was later pickedup and developed by Buchanan in ����. Buchanan argues that ifbenefits from government spending are equal for all, then we can focussolely on taxation and if they are not, we cannot study taxation inisolation: “It is impossible to speak of the ‘burden of taxation’ withoutconsidering, at the same time, the benefits from expenditure madeout of such taxation” (Buchanan ����, p. ���). When examiningthe preferences of political actors over taxation, we need to considerwhat assumptions are made regarding spending, because, as Buchananputs it, “[t]he statement that progressive taxation will redistributeincomes but that proportional taxation will not implies that benefitsare returned to individuals in proportion to incomes and wealth [...][t]he same amount of redistribution may be as well accomplished bythe levy of a sales tax to provide expanded social services as by anincrease in the higher-bracket income-tax rates to finance additionaldefense expenditure” (ibid. ���). Moreover, it has been shown thatproportional (or even regressive) taxes under certain conditions yield ahigher overall revenue, freeing up more resources for spending (Engel,Galetovic and Raddatz ����). This means that an actor wishing toredistribute resources in society might prefer less progressive taxes,if the taxes provide better opportunities to redistribute wealth andincome using government expenditures. This reasoning involving bothtaxing and spending was prevalent in British and Swedish tax policyin the post-war era (see Paper �).

In sum, expenditures should be considered when analyzing the taxpreferences of di�erent groups. Moreover, a proper understanding ofthe politics of taxation requires a more nuanced approach than merelyrelying on the rich-poor and progressive-regressive dichotomies.

�� introduction

institutions

Preferences over taxation are key to understanding tax policy changesince they determine the goals actors pursue in the political arena.However, equally important is the fact that the outcome of this conflictis shaped by the rules of the game – i.e., political institutions. Inthe first paper, the independent variable is democratization, a majorinstitutional change. Paper � and Paper � examine how the electoralsystem shapes left-wing tax strategy. The fourth paper contrastsinstitutions that concentrate power with those that share power.

When we have a minimal definition of the behavior of the actors,we can consider the role institutions play in the outcome of politicalconflicts over taxation. Writing in a time before universal su�rage,Wicksell was not only concerned with how the enfranchisement ofthe working class would a�ect tax policy but also with how di�erentconstitutional rules would shape this e�ect. In Wicksell’s ideal politicalsystem, government projects generating a net benefit would alwaysbe implemented since taxes and spending can spread the gains.�� Inpractice, the so-called unanimity rule would hardly be feasible, buthe considered the proportional representation electoral system to bea close approximation since it shared power broadly (as comparedwith plurality systems) (Wicksell [����]����). While Wicksell did notignore the problem of credible commitments, he did not present a deeperanalysis of the timing of costs and benefits or the future distribution ofpolitical power (for a discussion of this key issue see Powell ����).

The temporal nature of a policy and the possibility of crediblecommitments are crucial when considering tax reform. As Besley andPersson (����, ch. �) point out, a main problem in politics is thatchange of power leads to short-sightedness: Why take on costs today ifone cannot be certain to reap the benefits tomorrow? This is especiallyproblematic in tax policy, as taxes are more persistent and harder tochange than expenditures (Alt ����). Moreover, many social programshave insurance-like properties (Iversen ����) that will take time beforethey a�ect inequality. Therefore, introducing a tax today that hurtsvoters in the short term but leads to a higher fiscal capacity, whichcan be used to compensate those voters through social programs, isrisky if you lose power. Since di�erent political institutions allowfor di�erent degrees of opposition influence (Powell ����), politicalparties adapt their redistributive strategies to cope with political risk.

��A similar argument based on constitutional veto points is presented in (Lindvall����).

taxation and constitutions: how the papers fit

together ��

For example, even though left-wing parties have similar preferencesregarding redistribution, they will use di�erent tax strategies dependingon the institutional environment (see Paper � and Paper �).

The question of time not only a�ects decisions to invest in a newtax (or reforming an old one) but also a�ects how governments re-spond to short-term crises. While several scholars have analyzed theinfluence of institutions on the trade-o� between stability and flexibil-ity/adaptability (e.g., MacIntyre ����; Tommasi, Scartascini and Stein����; Tsebelis ����), we identify a trade-o� between being able to dealwith short-term crises and overcoming investment dilemmas. Paper �shows that institutions encouraging investments (e.g., fiscal capacity)are less apt at responding to short-term crises.

summary

The main point in this section – as well as underlying all four papers –is that we need to think carefully about who the actors are and whatthey really want. When we have a clear conceptual understanding ofthe actors and their motivations, we can formulate precise hypothesesregarding the influence of political institutions on policy outcome (Paper�). Institutions a�ect how actors choose to pursue their interest (Paper� and Paper �) as well as their ability to overcome investment dilemmasand respond to sudden crises (Paper �).

Paper � describes how democratization allows previously disenfran-chised groups to participate politically. Urbanization is a proxy fortheir diverse tax preferences. Although the independent variable isdemocratization, the actor is a voter. In practice, in democracies votersare represented by political parties. Therefore, although the properactors are voters, their role is implicit since the independent variableis an institutional change. Urban and rural voters have di�erent taxpreferences: rural voters often want lower property taxes and higherincome taxes, whereas urban voters often want to shift taxation fromconsumption to property and income.

In Paper �, the actors are left-wing political parties and they wantlower economic inequality (compared with the preferences of right-wingparties). The optimal strategy is conditioned by electoral institutions.

In Paper �, the actors are left-wing parties and trade unions. Theseactors can also be viewed as a labor movement coalition with two vetoplayers: left-wing parties and peak trade union organizations. Thepreference of this coalition is lower inequality and how they seek toattain this goal is a�ected by how institutions structure temporal risk.

�� introduction

In Paper �, the actors are once again political parties, but thistime in a more abstract sense. That is, we do not analyze them interms of “right” or “left”. The parties have di�erent preferences overspending on two goods provided by the government. The degree towhich their preferences diverge is allowed to vary and is important forthe likelihood of crises being averted and political investments made.

Table � lists the actors and their preferences in all four papers. Thereader will find a longer discussion about the actors, their preferences,and the impact of institutions in the individual papers.

Table �: Actors and Preferences

Actor PreferencePaper � Rural voters

Urban voters

Lower property tax andhigher income tax

Shift from consumptiontaxes to income and

property taxes

Paper � Left-wing parties Redistribution

Paper � Left-wing partiesand unions

Redistribution

Paper � Parties Levels of governmentspending on two goods

�.�

A Note on Research DesignWhen investigating claims regarding tax structure change over time,previous research has been severely constrained by a lack of compara-tive public finance data with broad geographic and temporal coverage.Comparative information about historical tax revenues for a largecross-section of countries is crucial for a comprehensive investigation ofthe root causes behind the evolution of tax systems. As Lord Kelvinonce wrote: “When you cannot measure [...] your knowledge is [...]meagre [...] and [...] unsatisfactory”.�� The dataset presented in this

��This is the version of the quotation carved into stone in University of Chicago’sSocial Science Research Building at ���� East ��th Street in Chicago. For a

a note on research design ��

dissertation will for the first time allow a measure of the developmentof taxation in as many as �� countries.��

Studies using a small number of countries – e.g., a handful ofEuropean states – run the risk of drawing conclusions based on veryspecific circumstances, such as the impact of regional wars. On the otherhand, analyzing a large cross-section of countries but only for a shorttime ignores important historical variations in many of today’s advanceddemocracies. Moreover, tax systems are strongly path dependent, whichmeans that understanding a tax system’s history is crucial if we wantto explain the variation in tax systems today (Morgan and Prasad����). To learn how today’s rich nations overcame the problem offiscal capacity or how democracy a�ected early tax decisions, we needto cover a much longer period than previous research.

The first two papers use statistical analysis to test theories regardinglong-term patterns in taxation and political institutions.�� While usefulfor testing hypotheses concerned with broad patterns and generalassociations, macro data is less useful when evaluating the theory ona micro level. Often, especially when there is no random variation,several explanations fit the observed patterns, implying that eventhough associations exist, it is not possible to verify which explanationis most valid.

Qualitative case studies are generally praised for their ability toshed light on the causal process at the micro level as well as causalordering. Cases can be chosen in a way to facilitate this exploration,but cases can also be chosen with the purpose of estimating a casuale�ect in mind (e.g., Geddes ����, and King, Keohane and Verba ����).In fact, much of the criticism of studies with small sample sizes (i.e.,small-n studies) starts from the assumption that the ultimate goal isthe estimation of causal e�ects. If that is the case, then there are someissues that cannot be avoided; after all, observations and variation arecrucial when estimating the impact of a variable, and small-n studiesper definition have fewer observations. Often the advice seems to be“increase the number of observations”. This advice has two problems:first, it is not very helpful as case studies are often very time consuming

thorough treatment of the quotation, its history, and Jacob Viner’s famous response“When you can measure it, when you can express it in numbers, your knowledge isstill of a meagre and unsatisfactory kind”, see Merton, Sills and Stigler (����).

��But, as Koopmans (����) reminds us, pure empiricism is bound to be inferiorto clever use of theory together with observation and measurement.

��Since the statistical challenges associated with this particular data structureare quite technical, I refer the reader to the methods sections in Paper � and Paper�.

�� introduction

and it is rarely possible to research more than one or a couple of casesand second, it is a bit confused since what is meant by “observations”di�ers between quantitative and qualitative approaches.

One suggestion for why the concept of “observation” changes whenmoving from quantitative to qualitative work is that the former exam-ines data set observations, whereas the latter examines causal processobservations (Collier, Brady and Seawright ����).�� In e�ect, thismeans that one case is not one observation, but consists of several ob-servations in sequence, an approach that is developed in process tracingmethodology (see George and Bennett (����) for an introduction).

Another source of confusion concerns generalizability. A key elementin a quantitative design, apart from estimating the causal e�ect, is togeneralize. But the goal of a case study is often somewhat di�erent. AsMahoney (����) argues: “Case studies seek to tell us why particularoutcomes happened in specific cases; statistical studies try to estimatethe average e�ects of variables of interest” (p. ���).

In my view, the purpose of case studies is not to primarily estimatecasual e�ects and to generalize. Instead, I think the usefulness ofqualitative small-n studies is more in line with Levi’s approach wherethe aim is “to demonstrate that the model is consistent with the facts”(����, p. �). That is, the purpose is to illustrate the causal process.However, this way of implementing case studies needs to be combinedwith quantitative analysis to realize its full potential, since any theoryis more valuable if it can be shown to be operating in several situations.

Quantitative and qualitative approaches complement each othersince the disadvantages of quantitative studies are exactly what quali-tative studies excel at, and vice versa. Here I use both methods notonly to show robust associations across time and space, but also toinvestigate the plausibility of the proposed causal process at the microlevel. For example, in Paper � I present data suggesting that the impactof ideology depends on the electoral system, and in Paper �, I refinethe theoretical argument and provide qualitative evidence in support ofthe proposed causal process – i.e., strategic considerations of left-wingpolitical parties.��

Finally, Paper � uses formal analysis, an excellent method forexploring the logic of an argument as well as for uncovering new

��Mahoney (����) further categorizes causal process observations into “indepen-dent variable”, “mechanism”, and “auxiliary outcome” sub-types.

��There are also more structured ways of integrating qualitative and quantitativemethods, for example, nested analysis (Lieberman ����) and the Bayesian approachsuggested by Humphreys and Jacobs (����).

taxation in the world ����-���� : the government

revenue dataset ��

theoretical and empirical implications. This type of exercise alsoforces one to clarify the assumptions used and lays bare the logic ofthe argument (or lack thereof). Paper � provides novel insights intohow institutions shape the behavior of political parties and exposes atrade-o� associated with short-term crisis management and long-terminvestment policies. The next section presents the Government RevenueDataset and provides a brief description of the evolution of taxationand institutions from ���� to ����.

�.�

Taxation in the World ����-����: TheGovernment Revenue Dataset

Alexander Hamilton once claimed that “[t]here is no part of the ad-ministration of government that requires extensive information and athorough knowledge of the principles of political economy, so muchas the business of taxation” ([����] ����), and as the first Secretaryof the Treasury, he was a key figure in the early fiscal history of theUnited States. For example, in this position he was behind a controver-sial tax on whiskey in order to repay the debt that had accumulatedduring the revolutionary war (Stockwell ����, p. ���).�� While theyoung federation relied mainly on indirect taxes (in particular customsrevenue), the structure of American taxation today is very di�erent,with individual income tax being the main source of revenue.

The change in the tax structure of the United States is part of amore general pattern. In the �� countries included in the dataset, thetype of government revenue has been transformed radically in the ���years – from a high reliance on tari�s, to the income tax, and in morerecent times, to broad-based consumption taxes.�� Not only has thenature and composition of government revenue changed, but its relativesize in relation to the economy has also increased sharply.

As pointed out above, the period under investigation is character-ized by revolutionary change not only in economic and demographicstructure but also in political institutions. In the early nineteenthcentury, democracy as a political system was rare. At the same time,

��This tax later sparked the so-called “whiskey rebellion” (Stockwell ����, p.���).

��This last step has curiously enough not been taken by the United States, oneof the few developed countries without a VAT (on the federal level).

�� introduction

industrialization had only begun to transform the socioeconomic anddemographic structures. By the end of the twentieth century, however,democracy was the norm and most of the countries in the dataset hadreached an advanced stage of industrialization.

The first part of this section begins with a brief description of thedataset and how it was collected and coded. The second part provides ashort overview of broad international patterns in the history of taxationand institutions from ���� to ����.

data collection and coding

The Government Revenue Dataset is a joint project with ThomasBrambor. The first version (on which the papers of this dissertationis based) covers thirty-one countries from ���� (or independence) to����.�� As far as we know, there exists no comprehensive historicaldataset on states’ revenues. Even among wealthy countries, such asOECD member states, there is no cross-national database providingdata from the nineteenth century up to today.

Our dataset is based on secondary sources providing partial tempo-ral or geographic coverage.�� In many cases, di�erent sources relied onthe same underlying data but reported conflicting estimates of revenueyields and the size of the economy. To complement and adjudicatebetween existing databases, we combined information from these ex-isting datasets with information from country-specific sources. Theoverall aim of the coding process has been to create time series thatare internally consistent within a country over time and connects tocontemporary datasets (such as the OECD for European and NorthAmerican countries and CEPAL for South America) in order to al-low easy updates of the dataset.�� We also aimed at minimizing thenumber of sources for each country while keeping high coverage overtime. When we needed to decide between using one source to obtaincross-country consistency or using di�erent sources to obtain within

��The countries included are: Argentina, Australia, Austria, Belgium, Bolivia,Brazil, Canada, Chile, Colombia, Denmark, Ecuador, Finland, France, Germany,Ireland, Italy, Japan, Mexico, the Netherlands, New Zealand, Norway, Paraguay,Peru, Portugal, Spain, Sweden, Switzerland, The United Kingdom, The UnitedStates, Uruguay, and Venezuela

��For example, Astorga et al. (����); Dincecco and Prado (����); Flora, Krausand Pfenning (����); Mitchell (����b); (����a); (����c); Mauro et al. (����);and OECD (����). In the codebook we list all the country-specific primary andsecondary sources that we used in the final version of the dataset.

��This approach is particularly suitable for fixed-e�ects models employed inPaper � and Paper �.

taxation in the world ����-���� : the government

revenue dataset ��

country consistency, we chose the latter. The reason for this is thatour main interest in assembling the dataset was to explain long-runtrends within countries.

Total central tax revenue is divided into direct and indirect taxes.Furthermore, two subcategories of direct taxes, property taxes andincome taxes, are measured separately. Indirect taxes are dividedinto consumption taxes, excises, and customs (which includes taxeson import, exports, exchange profits, and profits from export/importmonopolies).�� Property taxes include taxes on real estate, wealth,inheritance, and land. Income taxes include taxes on income, profits,and capital gains by individuals and corporations as well as taxes onpayrolls and workforces. Ideally, these categories should be measuredseparately, but because of the limits of the historical data sourcesthis was not possible. Tax on consumption consists mainly of salesand turnover taxes prior to the spread of value-added taxes in the����s. Excise taxes are taxes on specific goods, for example tobaccoor alcoholic beverages. The di�erence between consumption taxes andexcise taxes is that the former are broad-based and the latter onlyconcern specific goods. The important aspect of both consumption andexcise taxes is their regressive impact.

The dataset only provides data on the central level because of poordata availability at the local level. By restricting data collection to thecentral level, we could provide data for a much larger sample, both intime and across countries. However, it is important to note that this isproblematic in cases where there is significant subnational authority inthe area of taxing and spending, as this revenue will not show up in ourdata. We chose to express di�erent categories of the budget as shares oftotal central tax revenue so we could make cross-country comparisons.Another important aspect of the data is that the sample varies overtime. Countries are only included once they are independent, whichmeans that the sample is smaller in the beginning of the nineteenthcentury. This in combination with data limitations also means thatEurope is overrepresented for earlier years.

We have collected information on public finance and economicdevelopment from a number of existing data sources. Several of thesedatasets relied partly on the same underlying data. Nonetheless, manyestimates of our variables of interest di�ered substantially. This is

��This disaggregation is based on the guidelines in the Government FinanceStatistics Manual ���� of the International Monetary Fund (����), but we combinedsome of the categories because of the scarcity of historical information and thespecific focus of our project.

�� introduction

especially the case further back in time. Since many of the sourcesoverlapped, one method is to average values, but this approach isproblematic several reasons. First, many sources rely on the sameunderlying data, so averaging would hide potentially duplicated sources.Second, coverage over time varies substantially between countries, soaveraging would mean that some sources are over-weighted. Finally,since the quality of secondary data di�ers considerably, averaging mightincrease rather than decrease measurement error.

Instead of averaging across di�erent sources, we followed a decisiontree to decide which sources to use as the basis for our estimates. Thefollowing rules were used to guide the coding:

�. Minimize the number of sources. If several sources coverthe same information, we preferred to use a single source acrosscategories for the same time period.

�. Prefer high-quality sources. We prioritized primary andcountry-specific secondary sources. Since these sources oftenprovide more detailed data, this meant that we needed to dosome of the categorization ourselves. However, many of the cross-national datasets were of such a high quality that we confidentlyrelied on them for parts of the dataset.

�. Check the consistency of sources. When relying on two ormore sources to construct a long-run series, we made certain thatthe information is comparable when covering the same overlappingperiod within a country. In cases of a significant jump at theintersection of two series, this is indicated by coding the last valueof the ending series as missing.

�. Time series consistency trumps cross-sectional consis-tency. As mentioned above, our main interest is long-run trendswithin countries. When deciding whether to use the same sourceto obtain cross-country consistency or using di�erent sources toreach consistent within-country estimates, we prefer the latter.

Since the tax categories are expressed as ratios (e.g., the shareof the revenue coming from income taxes), we strongly preferred thenumerator and denominator coming from the same source. For moreinformation about the dataset and coding, see the individual papers(in particular the appendix to Paper �) and the codebook.

taxation in the world ����-���� : the government

revenue dataset ��

Coverage per Country. Table � reports the coverage in time per countryand tax category. Sometimes the information does not extend to ����.There are three main reasons for this. First, the country did notexist or was not independent at the time (e.g., neither Germany norItaly were unified until ����). Second, although countries may havebeen established, they may not have established that particular tax.For example, income tax was relatively rare in the early nineteenthcentury and broad-based consumption taxes only became commonafter the spread of the VAT in the latter part of the twentieth century.Third, although a country and a tax existed, there was no availableinformation.

Since we did not always know if data were missing because the taxwas not in place for a specific year, if it was but not collected, if theinformation was simply missing, or if it was missing because of waror some other major event, the dates in the table begin with the yearwe first have data. I have marked with asterisks (ú) cases where theinterval contains several large gaps. There are di�erent reasons forthese gaps. For example, a tax can be removed or data can just bemissing because of war or occupation.

I have also included two figures showing the development of the dif-ferent taxes over time for each country. Figure � provides information ondirect taxes (income and property) and Figure � shows the developmentof indirect taxation (customs, excises, and consumption).��

��To make it easier to interpret visually, these graphs are based on linearlyinterpolated data.

Table �: Coverage per Country

(�) (�) (�) (�) (�) (�) (�) (�)Tax/GDP Direct

TaxesIncome

TaxProperty

TaxIndirectTaxes

ConsumptionTax

ExciseTaxes

CustomsRevenue

Argentina ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Australia ����-���� ����-���� ����-���� ����-����ú ����-���� ����-���� ����-���� ����-����Austria ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Belgium ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Bolivia ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Brazil ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Canada ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Chile ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Colombia ����-����ú ����-���� ����-���� ����-���� ����-���� ����-���� ����-����ú

Denmark ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Ecuador ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Finland ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����France ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Germany ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Ireland ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Italy ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Japan ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Mexico ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����ú

Netherlands ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����New Zealand ����-���� ����-���� ����-���� ����-����ú ����-���� ����-���� ����-����ú ����-����Norway ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Paraguay ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Peru ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Portugal ����-���� ����-���� ����-����ú ����-���� ����-���� ����-���� ����-���� ����-����Spain ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Sweden ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Switzerland ����-����ú ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����United Kingdom ����-���� ����-���� ����-����ú ����-���� ����-���� ����-���� ����-���� ����-����United States ����-���� ����-���� ����-���� ����-���� ����-���� ����-����ú ����-����Uruguay ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����Venezuela ����-���� ����-���� ����-���� ����-���� ����-���� ����-���� ����-����

050

100

050

100

050

100

050

100

050

100

050

100

1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000

1800 1850 1900 1950 2000

Argentina Australia Austria Belgium Bolivia Brazil

Canada Chile Colombia Denmark Ecuador Finland

France Germany Ireland Italy Japan Mexico

Netherlands New Zealand Norway Paraguay Peru Portugal

Spain Sweden Switzerland United Kingdom United States Uruguay

Venezuela

Income Tax Property Tax

Shar

e of

Tot

al C

entra

l Tax

Rev

enue

Figure �: Direct Taxation per Country

050

100

050

100

050

100

050

100

050

100

050

100

1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000 1800 1850 1900 1950 2000

1800 1850 1900 1950 2000

Argentina Australia Austria Belgium Bolivia Brazil

Canada Chile Colombia Denmark Ecuador Finland

France Germany Ireland Italy Japan Mexico

Netherlands New Zealand Norway Paraguay Peru Portugal

Spain Sweden Switzerland United Kingdom United States Uruguay

Venezuela

International Trade Consumption TaxesExcise Taxes

Shar

e of

Tot

al C

entra

l Tax

Rev

enue

Figure �: Indirect Taxation per Country

taxation in the world ����-���� : the government

revenue dataset ��

05

1015

20Sh

are

of G

DP

1800 1850 1900 1950 2000Year

Figure �: Total Tax Revenue as a Share of GDP

patterns of taxation: ����-����

Perhaps the most striking development during the last two centuries isthe enormous increase in total taxation. Figure � depicts the evolutionof total central tax revenue as a share of Gross Domestic Product(GDP) from ���� to ����. After the Napoleonic wars, the level oftaxation decreases and stays at a fairly stable level until the First WorldWar. The average tax level decreases somewhat in the interwar periodbut never reaches the old levels, and after the Second World War, theoverall level of taxation continues to increase.��

Figure � shows the development of direct and indirect taxes asshares of total central tax revenue from ���� to ���� in the �� countriescovered by our data. While there was considerable variation in thenineteenth century, it is not until the First and Second World War thatdirect taxation reaches the same level as the share from indirect taxes.Between ���� and today, there has been less volatility and the sharesof revenue from indirect and direct taxes are almost identical. Whilemost empirical research into the political economy of taxation focus

��For an analysis of the so-called “ratchet e�ect” of war in the United Kingdom,see Peacock and Wiseman (����).

�� introduction

020

4060

8010

0Sh

are

of T

otal

Cen

tral T

ax R

even

ue

1800 1850 1900 1950 2000Year

Direct Taxes Indirect Taxes

Figure �: Direct and Indirect Taxation

on the period from ���� or later, this figure shows that major changesoccurred long before that.

Analyzing taxation in terms of indirect and direct categories iscommon, especially in more historical research. However, there areimportant developments within these two groups that are not capturedby aggregate measures. Figure � shows the development of propertyand income tax (direct taxes) and customs, consumption, and excisetaxes (indirect) separately. It is important to note that these categoriescan have di�erent distributional consequences. For example, the socialgroups being hurt by raised tari�s are not necessarily the same asthose who are hurt by a broad based consumption tax. Thus, it isan oversimplification to assume that direct taxes are progressive andindirect taxes are regressive.

The graphs show considerable variation within these two categoriesof taxes. Although the aggregate direct taxation has become moreimportant as a source of revenue, Figure � shows that this is drivenby increased income tax. With regards to indirect taxation, bothcustoms revenues and excises have decreased in importance while broad-based consumption taxes have become more important over time. This

Figure �: Subsets of Direct and Indirect Taxes0

2040

6080

100

Shar

e of

Tot

al C

entra

l Tax

Rev

enue

1800 1850 1900 1950 2000Year

Income Tax Property Tax

020

4060

8010

0Sh

are

of T

otal

Cen

tral T

ax R

even

ue

1800 1850 1900 1950 2000Year

International Trade Excise TaxesConsumption Taxes

�� introduction

Figure �: Taxation in the United Kingdom and France0

2550

7510

0Sh

are

of T

otal

Tax

Rev

enue

1800 1830 1860 1890 1920 1950 1980 2010World War I

World War II

Year

Income Consumption

Income and Consumption Taxation in the UK

025

5075

100

Shar

e of

Tot

al T

ax R

even

ue

1800 1830 1860 1890 1920 1950 1980 2010World War I

World War II

Year

Income Consumption

Income and Consumption Taxation in France

development accelerated with the spread of the VAT in the ����s and����s (James ����; Keen and Lockwood ����).

Although there are common patterns in taxation across countries,large di�erences can be observed between countries. Figure � shows thedevelopment of income and consumption taxes in the United Kingdomand France from ���� to ����. It shows that the United Kingdomhistorically relied more on income tax than France, a pattern that wasstrengthened after the First World War and persists today. After theSecond World War, the share of income tax was almost twice as highin the United Kingdom as in France. Instead, France has relied moreheavily on excise and consumption taxes both before and after thetwo World Wars. Hence, despite general patterns, there are importantdi�erences in taxation between countries that cannot be explained bydemocracy or war alone.

In all, these descriptive graphs show that there have been consid-erable changes in taxation during the last two centuries. Income andconsumption taxes have increased in importance while taxes on tradeand property are no longer major sources of revenue. These changeshave not only been driven by a strive for greater revenue capacity ande�ciency, but also by distributional struggles between social groups.Even if a tax reform increases e�ciency, it will not be implemented ifthe group in power will be economically worse o� (see Paper �). Fur-thermore, a change in the tax code is bound to create conflict betweendi�erent groups in society (Alt ����). How these conflicts play outand are solved (or not solved) is linked to the political institutionsstructuring the interaction between actors (Paper � and Paper �).

taxation in the world ����-���� : the government

revenue dataset ��

0.2

.4.6

.81

Shar

e of

Cou

ntrie

s

1800 1850 1900 1950 2000Year

Democracy Proportional representationSources: Boix, Miller and Rosato 2013 (democracy) and Colomer 2004 (proportional representation)

Figure �: Democracy and Proportional Representation

When considering the development of tax reform over two hundredyears, we should keep in mind that the political institutions changeddramatically. Countries not only democratize (and in some casesrevert back to authoritarianism) but also adopt di�erent democraticinstitutions. Figure � describes the evolution of two major changestaking place in this period: the spread of democracy and of proportionalrepresentation electoral systems.�� As the graph shows, in the latenineteenth century most countries in the dataset were non-democraciesand none used a proportional electoral system. After the Second WorldWar, the majority were democratic and used a proportional electoralsystem.��

The papers follow the changes in countries: from the impact ofdemocracy to specific political institutions within democracies to thechallenges of long-term investments and crisis management.

��Using data from Boix, Miller and Rosato (����) and Colomer (����).��As earlier theorists have pointed out, extending political rights to broader

segments of society should have profound e�ects on the tax system (e.g., Meltzerand Richard ����; Romer ����). These descriptive data lend some tentative supportto this idea in that both the number of democracies and the share of income taxare increasing over time.

�� introduction

�.�

Summary of PapersPaper � is concerned with the impact of democracy on taxation. Thecanonical models (Meltzer and Richard ����; Romer ����) predict thatdemocracy, by including the previously disenfranchised poor, leads tomore redistribution and by extension more progressive taxes (such asincome tax). However, existing research presents conflicting results.Although some indeed find this e�ect (e.g., Lindert ����), others claimthere is evidence that democratization is unrelated to redistribution(e.g., Ansell and Samuels ����). In Paper �, I argue that these in-conclusive findings are due to insu�cient attention to the di�erentialtax preferences of low-income voters. That is, urban and rural poorhave di�erent views on taxation. Urbanites want to shift taxationaway from consumption and onto land and income, whereas the ruralpoor are mainly concerned with lowering taxes on land. Using thenovel dataset over government tax revenue as described in section �.�,I present evidence suggesting that the e�ect of democracy does indeeddepend on the urbanization rate. More specifically, democratizationin highly urbanized states is associated with greater shares of revenuecoming from taxes on income and property and lower shares from exciseand consumption taxation. In contrast, democratization in a ruralcontext is associated with less tax revenue from property, excise, andconsumption taxes.

Paper � focuses on tax policy within democracies. I ask why andwhen left-wing governments tax the poor. Intuitively, governments tothe left should tax the rich relatively harder than the poor, but thehistorical record shows that more often than not the pattern is theopposite. In fact, left-wing governments are frequently associated withheavier taxation of consumption relative to taxation of income andwealth. However, there is no consensus regarding if, when, and whythis is. Moreover, previous research is mainly concerned with the lastdecades of the twentieth century, a temporal limitation overcome by thegovernment revenue dataset I use. My argument is that to understandthe e�ect of ideology on taxation we need to consider how taxationand spending interacts with the institutional environment. First, thefact that left-wing governments prefer a more equal distribution ofincome and wealth does not necessarily predict a certain tax structure.Rather, it crucially depends on government expenditures. This impliesthat di�erent tax structures can be compatible with redistribution.

summary of papers ��

Second, what type of tax/spend mix is employed depends on howpolitical institutions shape electoral risk. In a system concentratingpower to one actor, the impact of losing power is greater than in asystem where power is more widely shared. Redistributive taxation andredistributive spending have di�erent temporal distributions as taxationstarts today while the e�ect of spending takes time (barring lump sumtransfers). Electoral systems create di�erent strategic contexts in whichthe left operates. PR systems grant the opposition greater influence,thus extending the time horizon of an incumbent government. In thesesystems, the left shifts taxation from income to consumption, generatinga higher total tax yield, and redistributes mainly on the expenditure side.In majoritarian systems, on the other hand, because the left will haveno influence if out of power, the left will opt for a safer redistributivestrategy that mainly relies on progressive taxation of income. Inthe paper, I present data indicating that left-wing governments areassociated with heavier taxation of consumption in countries usingproportional representation (PR) electoral systems, while the oppositeis the case in countries using majoritarian electoral systems.

Paper � continues the logic in Paper �, but refines the theory byspelling out the causal process on a micro level. To explore the proposedcausal process, Paper � studies reforms of consumption taxation inBritain and Sweden in the decades following the Second World War.In both countries, policymakers acknowledged the advantages of broad-based consumption taxation in terms of revenues and the possibilityof counter-acting the regressive e�ects with compensatory spending.In Sweden, the Social Democratic Party (and initially also the tradeunions) was convinced that they could use a national sales tax toforward a redistributive spending agenda, so the party implementedthe tax. In Britain, massive resistance from within the labor movement,and especially the trade unions, made Labour reluctant. The oppositionfrom the Trades Union Congress was driven by the risk associated witha future Tory government, which might keep the tax but not implementthe promised compensatory spending. In sum, the decision whether tointroduce a new broad-based consumption tax was strongly influencedby risks shaped by the di�erent political institutions in place.

The final paper, co-authored with Johannes Lindvall, uses a formalmodel of political conflict to analyze government performance withrespect to crisis management (e.g., financial or banking crises) andinvestment policies. An example of an investment is tax capacity (seeBesley and Persson ����, ch. �), since adding, reforming, or in otherways changing the tax system can all be used to increase fiscal capacity,

�� introduction

but all are costly. Introducing a new tax such as an income or avalue-added tax requires considerable investments in administrativecapacity. This cost means that less government funds are availablefor spending today. Importantly, the decision to add or change taxesdepends on what governments spend on in the future. Our analysisshows that countries using institutions that share power widely (e.g.,proportional representation electoral systems and/or with multipleconstitutional veto points) are more likely to engage in investmentpolicies than countries employing power-concentration institutions (e.g.,systems with majority governments with no veto points). However,there is a trade-o�: Power-sharing institutions are worse at handlingsudden crises when swift and decisive action is needed.

�.�

Contributions and Future ResearchThis dissertation consists of four independent papers, each emphasizinga di�erent aspect of the politics of taxation. Empirically, the maincontribution is the novel dataset over government revenues from ����to ���� that is used in two papers to shed light on how democracy andideology influence tax structure. Qualitative and analytical approachescomplement the statistical evidence and provides new insights intoleft-wing tax strategy and how constitutions di�er in their ability tohandle crises and investment dilemmas. The main contributions canbe summarized in four points:

�. The dissertation is partly based on a novel dataset presentinginformation on government tax revenue from ���� to ���� for�� countries. This is a significant contribution since earlier workhas been based on either a small number of European countriesover a long period or a larger number of states for a very shortperiod (most often the decades following the Second World War).By providing data covering Europe, both Americas, the Englishspeaking o�-shoots, and Japan, it is possible to trace the evolutionof the modern tax state without being constrained by a smallsample restricted to Europe.

�. The second main contribution is a new take on the e�ects ofdemocratization. I argue in Paper � that the impact of democrati-zation on taxation is not as straightforward as commonly thought.

contributions and future research ��

This impact depends on the tax preferences of the previously dis-enfranchised citizens. This understanding is important because itmeans if one is analyzing the impact of democracy in a sample ofpredominately rural countries (such as early nineteenth centuryEurope), one finds very di�erent e�ects compared with a samplewith more urbanized states.

�. This dissertation provides a new analysis of left-wing party taxstrategy. The link between ideology and economic policy has longpuzzled scholars and I argue that by taking into considerationhow electoral institutions shape the strategic environment, someof the previously surprising patterns make more sense. Using themost comprehensive database of tax revenues, I present evidenceof an interaction e�ect where the impact of left-wing governmentis di�erent depending on the electoral system. Moreover, Paper �provides additional evidence supporting the theoretical argumenton the micro level.

�. Paper � and Paper � shows that how a government makes taxpolicy decisions depends on the political institutions in place,which a�ects the inter-temporal risks associated with tax policy.Paper �, a collaboration with Johannes Lindvall, develops ageneral theoretical framework where we compare how di�erentinstitutions a�ect the ability of governments to respond to suddencrises and to make political investments.

There are a number of fruitful avenues for future research in the areaof tax policy, both empirically and theoretically. Among the mostimportant are listed below:

�. In this dissertation, I present data on the revenue side of thebudget. A natural next step is to collect information also onthe expenditure side. As mentioned above, our understanding oftax policy is incomplete without considering what the revenue isspent on.

�. In Paper �, I make the point that urbanization matters sinceurbanization is a proxy for voters’ tax preferences. But the geo-graphic distribution of voters also matters for how well di�erentelectoral systems represent di�erent socioeconomic groups (e.g.,Rodden ����). Thus, how the choice of electoral institutions

�� introduction

a�ect politics is likely to depend on the spatial distribution ofvoters. It is likely that the e�ect of democratization depends onurbanization as well as the electoral system.

�. In Paper � and Paper �, the focus is on left-party tax strategy. Anextension should develop an argument for other political partiesas well.

�. The model developed in Paper � is very simple and excludesseveral potentially important aspects of the politics of investmentand crisis management. Two of the more important extensionswould be to incorporate elections and the economy. Electionscould possibly introduce a tension between policy-oriented ando�ce-oriented goals. And including private economic decisionsadds an important constraint on taxation.

For a more comprehensive discussion about possible future researchin the context of the specific contributions of this dissertation, thereader is referred to the individual papers.

bibliography ��

�.�

BibliographyAcemoglu, Daron and James A. Robinson. ����. “Why Did the West Ex-

tend the Franchise? Democracy, Inequality, and Growth in HistoricalPerspective.” The Quarterly Journal of Economics ���(�):����–����.

Acemoglu, Daron and James A. Robinson. ����. “A Theory of PoliticalTransitions.” American Economic Review ��(�):���–���.

Acemoglu, Daron and James A. Robinson. ����. Economic Originsof Dictatorship and Democracy. New York: Cambridge UniversityPress.

Ahmed, Amel. ����. “The Existential Threat: Varieties of Socialismand the Origins of Electoral Systems in Early Democracies.” Studiesin Comparative International Development ��(�):���–���.

Aidt, Toke S. and Peter S. Jensen. ����. “Tax structure, size ofgovernment, and the extension of the voting franchise in WesternEurope, ����-����.” International Tax and Public Finance ��:���–���.

Almond, Gabriel A. ����. Political Science: The History of the Disci-pline. In A New Handbook of Political Science, ed. Robert E. Goodinand Hans-Dieter Klingemann. New York: Oxford University Press.

Alt, James E. ����. “The evolution of tax structures.” Public Choice��(�):���–���.

Ansell, Ben and David Samuels. ����. Democracy and Redistribution,����-����: Reassessing the Evidence. Washington DC: Paper pre-sented to the annual conference of the American Political ScienceAssociation.

Ansell, Ben W. and David J. Samuels. ����. Inequality and Democrati-zation: An Elite-Competition Approach. New York, NY: CambridgeUniversity Press.

Arrow, Kenneth J. ����. Social Choice and Individual Values. Num-ber �� in “Cowles Commission for Research in Economics” New York:John Wiley & Sons, Inc.

�� introduction

Astorga, Pab, Ame R. Berges, E. Valpy K. Fitzgerald and ThorpRosemary. ����. “The Montevideo-Oxford Latin American EconomicHistory Database (MOxLAD).”.URL: http://oxlad.qeh.ox.ac.uk/

Auerbach, Alan J. ����. “The Choice between Income and Consump-tion Taxes: A Primer.” NBER Working Paper ����� .

Austen-Smith, David. ����. “Redistributing Income under Propor-tional Representation.” Journal of Political Economy ���(�):����–����.

Baldwin, Peter. ����. The Politics of Social Solidarity: Class Bases ofthe European Welfare State ����-����. Cambridge, UK: CambridgeUniversity Press.

Bank, The World. ����. Reforming Tax Systems. In World DevelopmentReport ����. World Development Report The World Bank pp. ��–���.

Bates, Robert H. and Da-Hsiang Donald Lien. ����. “A Note onTaxation, Development, and Representative Government.” Politics &Society ��(�):��–��.

Bendor, Jonathan, Daniel Diermeier, David A. Siegel and Michael M.Ting. ����. A Behavioral Theory of Elections. Princeton and Oxford:Princeton University Press.

Beramendi, Pablo and David Rueda. ����. “Social Democracy Con-strained: Indirect Taxation in Industrialized Democracies.” BritishJournal of Political Science ��(�):���–���.

Besley, Timothy and Torsten Persson. ����. “The Origins of State Ca-pacity: Property Rights, Taxation and Politics.” American EconomicReview ��(�):����–����.

Besley, Timothy and Torsten Persson. ����. Pillars of Prosperity.Princeton, New Jersey: Princeton University Press.

Black, Duncan. ����. “On the Rationale of Group Decision-Making.”Journal of Political Economy ��(�):��–��.

Boix, Carles. ����. “Setting the Rules of the Game: The Choice ofElectoral Systems in Advanced Democracies.” The American PoliticalScience Review ��(�):���–���.

bibliography ��

Boix, Carles. ����. Democracy and Redistribution. New York, NY:Cambridge University Press.

Boix, Carles, Michael K Miller and Sebastian Rosato. ����. “A Com-plete Data Set of Political Regimes, ����-����.” Comparative Politi-cal Studies ��(��):����–����.

Boucoyannis, Deborah. ����. “No Taxation of Elites, No Representation:State Capacity and the Origins of Representation.” Politics & Society��(�):���–���.

Bradley, David, Evelyne Huber, Stephanie Moller, François Nielsenand John D. Stephens. ����. “Distribution and Redistribution inPostindustrial Democracies.” World Politics ��(�):���–���.

Brélaz, Cédric. ����. Maintaining Order and Exercising Justice inthe Roman Provinces of Asia Minor. In The Province Strikes Back -Imperial Dynamics in the Eastern Mediterranean, ed. Björn Forsénand Giovanni Salmeri. Vol. XIII of Papers and Monographs of theFinnish Institute at Athens Helsinki: Foundation of the FinnishInstitute at Athens.

Buchanan, James M. ����. “The Pure Theory of Government Finance:A Suggested Approach.” Journal of Political Economy ��(�):���–���.

Buchanan, James M. ����. “The Constitution of Economic Policy.”The American Economic Review ��(�):���–���.

Buchanan, James M. and Richard A. Musgrave. ����. Public Financeand Public Choice: Two Contrasting Visions of the State. Cambridge:MIT Press.

Burns, Tom R. and Masoud Kamali. ����. The Evolution of Par-liaments: A Comparative-historical Perspective on Assemblies andPolitical Decision-making. In Handbook of Historial Sociology, ed.Gerard Delanty and Engin F. Isin. London, Thousand Oakes, andNew Delhi: Sage.

Cameron, David R. ����. “The Expansion of the Public Economy:A Comparative Analysis.” The American Political Science Review��(�):����–����.

Centeno, Miguel Angel. ����. “Blood and Debt: War and Taxation inNineteenth-Century Latin America.” American Journal of Sociology���(�):����–����.

�� introduction

Collier, David, Henry E. Brady and Jason Seawright. ����. Sources ofLeverage in Causal Inference. In Rethinking Social Inquiry: DiverseTools, Shared Standards, ed. Henry E. Brady and David Collier.Lanham: Rowman and Littlefield.

Colomer, Josep M. ����. The Strategy and History of Electoral SystemChoice. In Handbook of Electoral System Choice. Hampshire andNew York: Palgrave Macmillan.

Compania General de Tabacos de Filipinas v. Collector of InternalRevenue. ����.URL: https://supreme.justia.com/cases/federal/us/���/��/case.html

Cox, Gary W. ����. Making Votes Count - Strategic Coordination inthe World’s Electoral Systems. Political Economy of Institutions andDecisions Cambridge, United Kingdom: Cambridge University Press.

Co�gel, Metin M. and Thomas J. Miceli. ����. “Tax Collection inHistory.” Public Finance Review ��(�):���–���.

Cusack, Thomas R. and Pablo Beramendi. ����. “Taxing work.” Euro-pean Journal of Political Research ��:��–��.

Dahl, Robert A. ����. On Democracy. New Haven & London: YaleUniversity Press.

Deutsch, Karl W. ����. “Social Mobilization and Political Development.”American Political Science Review ��(��):���–���.

Dincecco, Mark and Mauricio Prado. ����. “Nominal GDP Series,����-����.”.URL: http://gpih.ucdavis.edu/GDP.htm

Downs, Anthony. ����. An Economic Theory of Democracy. New York:Harper & Row.

Drazen, Allan and Vittorio Grilli. ����. The Benefits of Crises for Eco-nomic Reforms. Working Paper ���� National Bureau of EconomicResearch.

Duverger, Maurice. ����. Political Parties: Their Organization andActivity in the Modern State. New York: John Wiley.

bibliography ��

Engel, Eduardo M.R.A., Alexander Galetovic and Claudio E. Rad-datz. ����. “Taxes and income distribution in Chile: some unpleas-ant redistributive arithmetic.” Journal of Development Economics��(�):���–���.

Flora, Peter, Franz Kraus and Winfried Pfenning. ����. State, Economy,and Society in Western Europe ����-����: The growth of industrialsocieties and capitalist economies. Campus Verlag.

Freedom, House. ����. Freedom of the World. Washington: FreedomHouse.URL: http://www.freedomhouse.org

Fund, International Monetary. ����. Government Finance StatisticsManual ����. IMF.

Gabrielsen, Vincent. ����. Finance and Taxes. In A Companion toAncient Greek Government, ed. Hans Beck. Blackwell Companionsto the Ancient World John Wiley & Sons Ltd.

Ganghof, Ste�en. ����. “Tax mixes and the size of the welfare state:causal mechanisms and policy implications.” Journal of EuropeanSocial Policy ��(�):���–���.

Geddes, Barbara. ����. “How the Cases You Choose A�ect the AnswersYou Get: Selection Bias in Comparative Politics.” Political Analysis�:���–���.

Genschel, Philipp. ����. “Globalization, Tax Competition, and theWelfare State.” Politics & Society ��(�):���–���.

George, Alexander L. and Andrew Bennett. ����. Case Studiesand Theory Development in the Social Sciences. Cambridge, Mas-sachusetts: MIT Press.

Hadenius, Axel and Jan Teorell. ����. Assessing Alternative Indices ofDemocracy. Committee on Concepts and Methods Working PaperSeries, August.

Hamilton, Alexander. ����. Federalist No. ��. In The Federalist Papers.New Haven: An electronic publication of the Avalon Project Atthe Yale Law School : Documents in Law, History and Diplomacy.Originally published ����.

�� introduction

Harsanyi, John C. ����. Rational behavior and bargaining equilibrium ingames and social situations. New York and Cambridge: CambridgeUniversity Press.

He, Wenkai. ����. Paths toward the Modern Fiscal State: England,Japan, and China. Cambridge, Massachusetts and London, England:Harvard University Press.

Herb, Michael. ����. “No Representation without Taxation? Rents,Development, and Democracy.” Comparative Politics ��(�):���–���.

Herbst, Je�rey. ����. States and Power in Africa - ComparativeLessons in Authority and Control. Princeton, New Jersey: PrincetonUniversity Press.

Hettich, Walter and Stanley L. Winer. ����. Democratic Choice andTaxation: A Theoretical and Empirical Analysis. New York: Cam-bridge University Press.

Hibbs, Jr, Douglas A. ����. “Political Parties and MacroeconomicPolicy.” The American Political Science Review ��(�):����–����.

Hicks, Alexander M. and Duane H. Swank. ����. “Politics, Institutions,and Welfare Spending in Industrialized Democracies, ����-��.” TheAmerican Political Science Review ��(�):���–���.

Hinrichs, Harley H. ����. A general theory of tax structure changeduring economic development. Cambridge, Massachusetts: LawSchool of Harvard University.

Hintze, Otto. ����. Military Organization and the Organization ofthe State. In The Historical Essays of Otto Hintze, ed. Felix Gilbert.New York: Oxford University Press.

Hotelling, Harold. ����. “Stability in Competition.” The EconomicJournal ��(���):��–��.

Humphreys, Macartan and Alan M. Jacobs. ����. “Mixing Methods: ABayesian Approach.” American Political Science Review ���(�):���–���.

Imbeau, Louis M., François Pétry and Moktar Lamari. ����. “Left-rightparty ideology and government policies: A meta-analysis.” EuropeanJournal of Political Research ��:�–��.

bibliography ��

Iversen, Torben. ����. Capitalism, Democracy, and Welfare. Cam-bridge Studies in Comparative Politics New York: Cambridge Uni-versity Press.

Iversen, Torben and David Soskice. ����. “Electoral Institutions andthe Politics of Coalitions: Why Some Democracies Redistribute Morethan Others.” American Political Science Review ���(�):���–���.

Jaggers, Keith and Ted Robert Gurr. ����. “Tracking Democracy’sThird Wave with the Polity III Data.” Journal of Peace Research��(�):���–���.

James, Kathryn. ����. The Rise of the Value-Added Tax. New York:Cambridge University Press.

Jensen, Carsten and Svend-Erik Skaaning. ����. “Democracy, ethnicfractionalisation, and the politics of social spending: Disentanglinga conditional relationship.” International Political Science Review��(�).

Johnson, James. ����. “What Rationality Assumption? or, How‘Positive Political Theory’ Rests on a Mistake.” Political Studies��(�):���–���.

Joumard, Isabelle, Mauro Pisu and Debra Bloch. ����. Less IncomeInequality and More Growth – Are They Compatible? Part �. IncomeRedistribution via Taxes and Transfers Across OECD Countries.OECD Economics Department Working Papers Organisation forEconomic Co-operation and Development Paris: .

Kallet, Lisa. ����. Money and the Corrosion of Power in Thucydides:The Sicilian Expedition and Its Aftermath. University of CaliforniaPress.

Kallet-Marx, Lisa. ����. Money, Expense, and Naval Power in Thucy-dides’ History �-�.��. University of California Press.

Karaman, K. Kivanc and Sevket Pamuk. ����. “Di�erent Paths tothe Modern State in Europe: The Interaction Between Warfare, Eco-nomic Structure, and Political Regime.” American Political ScienceReview ���(�):���–���.

Kato, Junko. ����. Regressive Taxation and the Welfare State. NewYork: Cambridge University Press.

�� introduction

Keen, Michael and Ben Lockwood. ����. “The value added tax:Its causes and consequences.” Journal of Development Economics��(�):���–���.

Kemmerling, Achim. ����. Taxing the Working Poor - The PoliticalOrigins and Economic Consequences of Taxing Low Wages. Chel-tenham, UK: Edward Elgar.

King, Gary, Robert O. Keohane and Sidney Verba. ����. DesigningSocial Inquiry. Princeton: Princeton University Press.

Koopmans, Tjalling C. ����. “Measurement Without Theory.” TheReview of Economics and Statistics ��(�):���–���.

Lee, Cheol-Sung. ����. “Income Inequality, Democracy, and PublicSector Size.” American Sociological Review ��(�):���–���.

Leemann, Lucas and Isabela Mares. ����. “The Adoption of Propor-tional Representation.” The Journal of Politics ��(�):���–���.

Levi, Margaret. ����. Of Rule and Revenue. Number �� in “CaliforniaSeries on Social Choice and Political Economy” Berkeley: Universityof California Press.

Lieberman, Evan S. ����. “Nested Analysis as a Mixed-MethodStrategy for Comparative Research.” American Political ScienceReview ��(�):���–���.

Lijphart, Arend. ����. Patterns of Democracy: Government Forms andPerformance in Thirty-Six Democracies. New Haven and London:Yale University Press.

Lijphart, Arend. ����. Patterns of Democracy: Government Formsand Performance in Thirty-Six Countries. Second ed. New Haven &London: Yale University Press.

Lindert, Peter H. ����. Growing Public - Social Spending and Eco-nomic Growth Since the Eighteenth Century. Volume I: The Story.Cambridge, UK: Cambridge University Press.

Lindvall, Johannes. ����. “Power Sharing and Reform Capacity.” Jour-nal of Theoretical Politics ��(�):���–���.

Lipset, Seymour M. and Stein Rokkan, eds. ����. Party Systems andVoter Alignments: Cross-National Perspectives. New York: The FreePress.

bibliography ��

Lott, Jr, John R. and Lawrence W. Kenny. ����. “Did Women’sSu�rage Change the Size and Scope of Government?” Journal ofPolitical Economy ���(�):����–����.

MacIntyre, Andrew. ����. “Institutions and Investors: The Politics ofthe Economic Crisis in Southeast Asia.” International Organization��(��):��–���.

Mahoney, James. ����. “Review Article: After KKV. The New Method-ology of Qualitative Research.” World Politics ��(�):���–���.

Mares, Isabela and Didac Queralt. ����. “The Non-Democratic Originsof Income Taxation.” Comparative Political Studies ��(��):����–����.

Martin, Isaac William, Ajay K. Mehrotra and Monica Prasad, eds.����. The New Fiscal Sociology: Taxation in Comparative and His-torical Perspective. � edition ed. Cambridge ; New York: CambridgeUniversity Press.

Mauro, Paolo, Rafael Romeu, Ariel Binder and Asad Zaman. ����. “AModern History of Fiscal Prudence and Profligacy.” IMF WorkingPaper WP/��/�.

McCarty, Nolan and Adam Meirowitz. ����. Political Game Theory:An Introduction. � edition ed. New York: Cambridge UniversityPress.

Meltzer, Allan H. and Scott F. Richard. ����. “A Rational Theory of theSize of Government.” Journal of Political Economy ��(�):���–���.

Merton, Robert K., David L. Sills and Stephen M. Stigler. ����. “TheKelvin dictum and social science: An excursion into the history ofan idea.” Journal of the History of the Behavioral Sciences ��(�):���–���.

Mirrlees, James A. ����. “An Exploration in the Theory of OptimumIncome Taxation.” The Review of Economic Studies ��(�):���–���.

Mitchell, B. R. ����a. International historical statistics: Africa, Asia& Oceania, ����-����. � ed. New York: Palgrave Macmillan.

Mitchell, B. R. ����b. International Historical Statistics: Europe,����-����. New York: Palgrave Macmillan.

Mitchell, B. R. ����c. International historical statistics: the Americas,����-����. � ed. New York: Palgrave Macmillan.

�� introduction

Monson, Andrew and Walter Scheidel, eds. ����. Fiscal Regimes andthe Political Economy of Premodern States. Cambridge, UnitedKingdom ; New York, NY: Cambridge University Press.

Moore, Mick. ����. “Revenues, State Formation , and the Quality ofGovernance in Developing Countries.” International Political ScienceReview �:���–���.

Morgan, Kimberly J. and Monica Prasad. ����. “The Origins of TaxSystems: A French-American Comparison.” American Journal ofSociology ���(�):����–����.

Mueller, Dennis C. and Thomas Stratmann. ����. “The economice�ects of democratic participation.” Journal of Public Economics��(�-��):����–����.

Munck, Gerardo L. and Jay Verkuilen. ����. “Conceptualizing andMeasuring Democracy: Evaluating Alternative Indices.” ComparativePolitical Studies ��(�):�–��.

Musgrave, Richard Abel. ����. Fiscal Systems. New Haven: YaleUniversity Press.

Müller, Wolfgang C. and Kaare Strøm. ����. Policy, O�ce, Or Votes?:How Political Parties in Western Europe Make Hard Decisions. Cam-bridge, U.K.: Cambridge University Press.

Nilsson, Klas. ����. The Money of Monarchs: The Importance ofNon-Tax Revenue for Autocratic Rule in Early Modern Sweden Ph.D.Thesis Lund University Lund: .

North, Douglass C. ����. “Institutions.” The Journal of EconomicPerspectives �(�):��–���.

North, Douglass C. and Barry R. Weingast. ����. “Constitutionsand Commitment: The Evolution of Institutional Governing PublicChoice in Seventeenth-Century England.” The Journal of EconomicHistory ��(�):���–���.

OECD. ����. “Organisation for Economic Co-operation and Develop-ment - Statistical Division.”.URL: http://stats.oecd.org/WBOS/Index.aspx

Olper, Alessandro. ����. “Determinants of Agricultural Protection:The Role of Democracy and Institutional Setting Alessandro Olper.”Journal of Agricultural Economics ��(�):��–��.

bibliography ��

Peacock, Alan T. and Jack Wiseman. ����. The Growth of PublicExpenditure in the United Kingdom. Princeton: Princeton UniversityPress.

Persson, Torsten and Guido Tabellini. ����. The Economic E�ects ofConstitutions. Cambridge, MA: MIT Press.

Powell, Bingham G. ����. Elections as Instruments of Democracy.New Haven: Yale University Press.

Powell, Robert. ����. “The Ine�cient Use of Power: Costly Con-flict with Complete Information.” American Political Science Reviewnull(��):���–���.

Prasad, Monica and Yingying Deng. ����. “Taxation and the worldsof welfare.” Socio-Economic Review ����(�):���–���.

Przeworski, Adam, Michael E. Alvarez, José Antonio Cheibub andFernando Limongi. ����. Democracy and Development: PoliticalInstitutions and Well-Being in the World, ����-����. CambridgeStudies in the Theory of Democracy New York: Cambridge UniversityPress.

Rae, Douglas W. ����. The Political Consequences of Electoral Laws.Vol. ���� New Haven and London: Yale University Press.

Ramsey, F. P. ����. “A Contribution to the Theory of Taxation.” TheEconomic Journal ��(���):��–��.

Riker, William H. ����. The Theory of Political Coalitions. New Havenand London: Yale University Press.

Rodden, Jonathan. ����. “The Geographic Distribution of PoliticalPreferences.” Annual Review of Political Science ��:���–���.

Rodden, Jonathan. ����. “The Long Shadow of the Industrial Rev-olution: Geography and the Representation of the Left.” BookManuscript.

Romer, Thomas. ����. “Individual welfare, majority voting, and theproperties of a linear income tax.” Journal of Public Economics�(�):���–���.

Rosen, Harvey S. and Ted Gayer. ����. Public Finance. Eighth ed.New York: McGraw-Hill.

�� introduction

Ross, Michael L. ����. “The Political Economy of the Resource Curse.”World Politics ��(��):���–���.

Ross, Michael L. ����. “Does Taxation Lead to Democratization?”British Journal of Political Science ��:���–���.

Ross, Michael Lewin. ����. “Does Oil Hinder Democracy?” WorldPolitics ��(�):���–���.

Scheve, Kenneth and David Stasavage. ����. “The Conscription ofWealth: Mass Warfare and the Demand for Progressive Taxation.”International Organization ��(�):���–���.

Schumpeter, Joseph. ����. Capitalism, Socialism, and Democracy.Revised second edition ed. London: George Allen & Unwin Ltd.

Schumpeter, Joseph. ����. The Crisis of the Tax State. In JosephA. Schumpeter: The Economics and Sociology of Capitalism, ed.Richard Swedberg. Princeton: Princeton University Press. Firstpublished in ����.

Schumpeter, Joseph A. ����. Capitalism, Socialism and Democracy.London and New York: Routledge. First published ����.

Seligman, Edwin Robert Anderson. ����. Essays in Taxation. �th ed.New York: The Macmillan Company.

Slemrod, Joel. ����. “Optimal Taxation and Optimal Tax Systems.”The Journal of Economic Perspectives �(�):���–���.

Smith, Adam. ����. An Inquiry into the Nature and Causes of theWealth of Nations. Volume II. London: Meuthen & Co. Firstpublished in ����.

Steinmo, Sven. ����. Taxation and Democracy. New Haven and London:Yale University Press.

Steinmo, Sven and Caroline J. Tolbert. ����. “Do Institutions Re-ally Matter? Taxation in Industrialized Democracies.” ComparativePolitical Studies ��(�):���–���.

Stockwell, Mary. ����. The American Economy: A Historical Ency-clopedia. Vol. Two: Essays and Primary Source Documents revisededition ed. Santa Barbara, California: ABC-CLIO.

bibliography ��

Strøm, Kaare. ����. Minority Government and Majority Rule. Cam-bridge, U.K.: Cambridge University Press.

Tavits, Margit. ����. “The Size of Government in Majoritarian andConsensus Democracies.” Comparative Political Studies ��(�):���–���.

Thomson, Henry. ����. “Food and Power: Agricultural Policy UnderDemocracy and Dictatorship.” Comparative Politics (forthcoming) .

Thorndike, Joseph J. ����. Their Fair Share: Taxing the Rich in theAge of FDR. Washington, D.C: The Urban Institute Press.

Tilly, Charles. ����. Coercion, Capital and European States: A.D.���-����. Cambridge, Massachusetts: Basil Blackwell.

Timmons, Je�rey F. ����. “Taxation and Credible Commitment: Left,Right, and Partisan Turnover.” Comparative Politics ��(�):���–���.

Tommasi, Mariano, Carlos Scartascini and Ernesto Stein. ����. “VetoPlayers and Policy Adaptability: An intertemporal perspective.” Jour-nal of Theoretical Politics ��(�):���–���.

Treadgold, Warren. ����. Byzantium and Its Army – ���-����. Stanford,California: Stanford University Press.

Tsebelis, George. ����. “The Abuse of Probability In Political Analysis:The Robinson Crusoe Fallacy.” American Political Science Review��(�):��–��.

Tsebelis, George. ����. Veto Players: How Political Institutions Work.Princeton: Princeton University Press.

Wanniski, Jude. ����. “Taxes, Revenues, and The La�er Curve.” ThePublic Interest .

Weber, Max. ����. Politics as Vocation. In From Max Weber: Essaysin Sociology. New York: Oxford University Press.

Wicksell, Knut. ����. A New Principle of Just Taxation. In Classics inthe Theory of Public Finance, ed. Richard A. Musgrave and Alan T.Peacock. New York: The Macmillan Company. First published in����.

Widmalm, Frida. ����. “Tax structure and growth: Are some taxesbetter than others?” Public Choice ���:���–���.

�� introduction

Yun-Casalilla, Bartolomé and Professor Patrick K. O’Brien, eds. ����.The Rise of Fiscal States: A Global History, ����-����. � edition ed.New York: Cambridge University Press.

Zhang, Nan, Giulia Andrighetto, Stefania Ottone, Ferruccio Ponzanoand Sven Steinmo. ����. “"Willing to Pay?" Tax Compliance inBritain and Italy: An Experimental Analysis.” PLOS ONE ��(�).

Åkerman, Sune. ����. Skattereformen ���� - Ett experiment med progres-siv inkomstskatt. Studia Historica Upsaliensia Västerås: Departmentof History, Uppsala University through Sven A. Nilsson and StenCarlsson.