Seeking Robin Hood: How Democracy Shapes Taxation and Redistribution

Meltzer and Richard’s theory of government growth rested on an appealingly simple and robust result: that given an unequal distribution of pre-tax income, under certain assumptions, the typical citizen would be better off if the government raised more money in tax and spent it on social programmes (for the formal exposition, see [6]; for a recent discussion of the large related literature on voting and redistribution, see [7]). The Meltzer-Richard theory drew on a well-known political science model of elections, in which most voters congregate around the middle of a left-right spectrum, and the decision of the ‘median voter’ – the one at the very centre-point of this spectrum – would decide which party would triumph [8]. This median voter would on average prefer more government spending, and with competitive elections, politicians would seek to outbid each other in offering more and more redistribution, inflating government spending and leading to crises of the kind seen during the 1970s. This logic built on a long tradition of fears that democracy, by giving equal votes to the whole adult population, would lead to the expropriation of the wealthy elite.

But if the median voter theorem seemed to do a good job of describing secular trends in the political economy of the western democracies, it was less successful in explaining variations across countries. Indeed, if what drove the political demand for redistribution was the gap between the median pre-tax income and the mean, as Meltzer and Richard implied, then the data in fact seemed to point in the opposite direction. The countries which levied the highest tax rates and spent the most on welfare provision were those that had the lowest income inequality before tax: the so-called the ‘Robin Hood paradox’ [9, 10, 11]. In other words, where more redistribution was of least benefit to median voters, the more redistribution seemed to be offered to them, a pattern that was particularly clear for the social democratic welfare states of Scandinavia. In contrast, high inequality countries, such as the United States and the United Kingdom, have been far less likely to raise taxes to pay for extensive social programmes, despite typical citizens potentially standing to benefit even more from such an approach, as inequality was high and rising.

Although there have been important changes over time in the nature of redistribution, as this article will discuss, these cross-national differences have been remarkably stable. As the role of the state in distributing economic resources expanded in the post-war period, institutions emerged to shape taxation and spending that proved extremely resilient, such that early differences in approach persisted. For example, the failure of the United States, almost uniquely amongst the high-income countries, to establish a federal sales tax made financing the kind of government spending needed to sustain a substantial welfare state almost impossible [12, 13]. The heavy use of tax expenditures as a form of social policy in the United States also locked in a form of welfare which relied less on government transfers and disproportionately benefited higher income groups [14, 15, 16]. As a result, the United States has long been a laggard amongst the wealthy democracies in its use of overt redistribution to reduce inequality, with a lower tax burden and lower social transfers than any comparable country.

At the other end of the scale, the countries of Northern Europe have mostly entrenched relatively high levels of taxation and government spending, with the result that poverty and inequality are amongst the lowest in the world. In 2019, the highest taxed country in the OECD was Denmark, with total government revenues of 47 per cent of GDP, getting on for twice as high as the United States (25 per cent) (see Figure One). The difference in tax take between these two extremes reflects an almost identically sized difference in social expenditure by the government. Remarkably, the same two countries were already at opposite ends of the scale in terms of social policy spending in the late 19^th century [10 p14], but the gap opened up decisively in the post-war period, when welfare states expanded dramatically in Western Europe. In 1965, the United States was already lagging between the European democracies in terms of its fiscal capacity, with only Switzerland (at that time still denying women the vote) and the Southern European dictatorships taxed less. By 2019, only Ireland among the longstanding democratic countries had a smaller state than the US (partly because of the peculiarities of the way GDP accounting plays out in a small country where many multinational companies report their revenues).

These different levels of taxing and spending, not surprisingly, have material consequences for a number of policies. One quite consistent (if not entirely surprising) pattern is that the higher the tax take, the more redistribution governments are able to achieve. Figure Two shows the tight correlation between levels of taxation and the degree to which the government is able to reduce inequality in disposable income (measured as the difference in the Gini coefficient pre– and post-tax).

On one level this is almost a banal finding. It seems obvious that the greater the tax revenues available to government, the more scope there is to make progressive shifts of resources from high to lower income groups. But despite the claims that government spending simply ‘churns’ resources, or even redistributes in favour of the better off, or that higher income groups are able to shield themselves from tax, the overall picture is unambiguously that the more governments tax, the more they shift resources down the income distribution.

But this tight correlation accentuates the puzzle of why some governments play ‘Robin Hood’ more than others. With few exceptions, the OECD countries discussed here have had stable democratic rule and highly developed economies for decades, and yet some of them tax and redistribute much less than others, even controlling for levels of GDP. The median voter theorem should imply that free elections will inevitably lead to pressures for redistribution, yet we can see from the data that there is a good deal of variation that cannot be accounted for by democracy alone. The next section discusses some answers to this puzzle.

The Power of History: Why Some Countries Tax and Redistribute More than Others

Comparative political economists have tended to explain variations in taxing and spending across the democracies in terms of longstanding institutional differences. In Esping-Andersen’s influential Three Worlds of Welfare Capitalism [18], patterns of popular mobilization and political alliances in the early 20^th century entrenched enduringly different types of welfare institutions: liberal, conservative and social democratic. In liberal welfare regimes (the English-speaking countries and Switzerland), the weakness of the left limited demands for government spending and therefore the need for high tax revenues. The lower than average revenue shares of these countries in Figure One, many of which have historically weak left parties, reveals this effect clearly. In contrast, the revenue shares of the other nations show the consequence of a consistently strong left-wing political presence. In Scandinavia, social democratic parties, allied with agrarian parties, dominated government coalitions for most of the 20^th century, entrenching a high tax, high spend equilibrium. The countries of continental Western Europe, where initially authoritarian regimes, then social and Christian democrats, pushed for bigger welfare states, are on average close behind. Finally Southern Europe, mostly denied democratic representation until the 1970s, followed the conservative continental model, but with a lag [19]. Esping-Andersen’s typology, built on data from the 1980s, does a remarkable job of predicting tax revenue shares today.

The ‘Three Worlds’ approach gives primacy to mass mobilization and electoral politics, albeit only in a brief window of mid-twentieth century history [20]. The main alternative approach, ‘Varieties of Capitalism’ [21], focuses instead on the patterns of institutional development which produce two main types of market economies, liberal (LME) and coordinated (CME). The implications for taxation are similar to ‘Three Worlds’. Liberal market economies leave little scope for government activism, and labour markets in particular are left as much as possible in the hands of the price mechanism, which high taxes and generous welfare benefits could potentially distort. In contrast, coordinated market economies involve building institutions to cement cooperation across firms, their investors and their workers. Generous welfare spending is often part of the package, with well-funded unemployment compensations and training provision, and the role of taxation, especially of labour, in paying for these institutions is broadly accepted by all actors. All of the lowest taxed countries in Figure One are LMEs in the Varieties of Capitalism model (except for Japan, where patterns of economic coordination with a lesser impact on government budgets play a similar role [22]).

It did not escape the attention of the Varieties of Capitalism scholars that LMEs and CMEs tended to have distinct electoral systems. The former mostly have majoritarian ‘first past the post’ electoral rules, and the latter forms of proportional representation. The relationship between electoral systems and the growth of government was first recognized by Arend Lijphart, who argued that ‘consensus democracies’ with PR, where multiparty governments predominate, tend to have bigger governments and overall ‘kinder and gentler’ politics than ‘majoritarian democracies’ where a ‘winner-take-all’ logic of single party government prevails [23, 24, 25]. Lijphart’s argument was that the greater inclusiveness of PR meant that more groups would be able to influence policy, leading to a greater dispersion of power, and ultimately, more redistribution of resources. Again, Figure One presents supportive data, with most of the countries with the highest tax take having PR, whereas the majoritarian democracies (US, UK, Australia, Canada) tend to have lower tax revenues, although there are several anomalies (e.g. France at the top, Switzerland and Ireland at the bottom).

Iversen and Soskice showed how different electoral systems cause the median voter to view redistribution differently [11]. In their model, under first past the post rules the median voter faces a simple choice between left and right candidates. Such a voter may fear that a single party government of the left would skew redistribution in favour of the poor, leaving middle income citizens on the hook for higher taxes. PR facilitates a mutually beneficial pro-redistribution alliance between middle and lower-income groups by allowing them to vote for different parties that combine after the election in a coalition government. The checks and balances and broader dispersion of power in most democracies with PR reassure the median voter that redistribution will still leave them better off overall, with the strain taken by higher income groups. This model does a good job of explaining why politics in some of the liberal market economies has often revolved around a battle between parties of higher spending and parties of lower taxes, and why the latter have prevailed more often, leading to lower levels of taxation on average than in CMEs.

The high level of persistence of electoral systems – parliaments elected under a given set of rules tend to be disinclined to change them – is one explanation for why these differences persist. But electoral systems are themselves quite strongly correlated with other country-level characteristics which could also plausibly explain the development of fiscal capacity. For example, the high tax levels in Scandinavia could also be a consequence of their society’s size and homogeneity leading to more sustained solidarity, in contrast to countries like the United States [26, 27]. Yet small homogeneous countries, such as New Zealand, have remained economically liberal, while large, diverse countries like France and Germany are more redistributive. Moreover, other studies have identified a related but distinct correlation between parliamentary rather than presidential government and higher levels of taxation [28].

Other macrosocial characteristics could also explain some of the differences we can observe. For example, religion also seems to play a role in politics which affect the state’s capacity to tax and spend [29]. Different patterns of state development in Europe are closely related to paths taken to the ways in which the critical juncture of the Reformation played out [30]. First, in Catholic countries Christian democracy, which was uncomfortable with the social changes brought by industrial capitalism, encouraged the emergence of the welfare state. Meanwhile, even where the Reformation was successful, different traditions took different views on the welfare state, with reformed Protestantism militating against and Lutheran Protestantism in support [31]. These differences precede the process of democratization and are plausibly independent from the choice of electoral systems, suggesting a much more historically rooted pattern of fiscal development.

To complicate matters more, religious divides are partly mirrored in other forms of institutional variation that could explain different approaches to taxation and redistribution. Institutional economics identified legal origin – the sources of countries’ basic judicial systems – as an important predictor of a range of economic and political outcomes, with the English Common Law tradition being associated with strong institutions for the protection of property rights, and the French, more codified, approach to law undermining governance in a number of ways, including fiscal capacity [32].

Catholic Europe has been more strongly influenced by the French model, spread by the Napoleonic Empire, whereas the English common law has had more influence over regions governed by the British Empire. Meanwhile, Northern Europe, untouched by the British or the French, reflects its own legal traditions. This maps quite closely onto the Varieties of Capitalism and Worlds of Welfare typologies, even though it implies quite different channels of causation. La Porta et al. claim that ‘common law stands for the strategy of social control that seeks to support private market outcomes, whereas civil law seeks to replace such outcomes with state-desired allocations’ [33 p286].

The consequences of this distinction for taxation are far-reaching, but also contradictory. On the one hand, the codified civil law model has great potential to enhance state power to intervene in the distribution of property rights, and thus reallocate economic resources through taxation and regulation. English Common Law’s fundamental preference for market outcomes instead implies greater protection of private property from the ‘grabbing hand’ of government. On the other, overly intrusive political intervention in the market, especially if it takes on an arbitrary or partisan form, can have the effect of undermining compliance, leading among other things to widespread corruption and tax evasion [34, 35]. The large size of the informal economy in former French colonies, and even in civil law parts of Europe, shows how the ‘quality of government’ acts as an important intervening variable in the analysis of taxation [36]. Where the state is unable to enforce compliance with tax law, fundamental inequities arise in the extent to which citizens contribute to public goods, and to redistribution.

In many societies the very wealthy seem disproportionately likely to be beneficiaries of this kind of state failure. This is particularly the case in countries with weak state capacity and undemocratic rule: what Acemoglu and Robinson call ‘extractive’ (rather than ‘inclusive’) institutions [37]. However, we also know that institutions evolve over time, and that under certain conditions fiscal arrangements can shift quite dramatically. The power of the very wealthy to protect themselves from pressures to share their wealth has transformed over the long run in very visible ways across the high-income countries. The next section examines the underlying reasons for these changes over time.

Change over time: The Rise and Fall of the State?

The debates around types of welfare capitalism and their associated fiscal institutions have given us valuable insights into how countries differ in their approaches to taxation and economic justice. However, the apparent stability of regimes of redistribution in the high-income countries can obscure the extent to which the politics of taxation has changed over the long run, as well as distracting from intriguing cases where countries have undergone fundamental shifts in their broad approach to taxation and redistribution. We therefore need to pay attention to the politics of how taxation and redistribution change over time, and the relative importance of general trends and country-specific phenomena.

One of the general trends noted earlier was that the 20^th century saw both the entrenching of democratic institutions in the high-income countries of the West and the development of complex market economies which generated high levels of prosperity, a prosperity widely shared by the redistributive institutions of the welfare state. This relationship of democracy to prosperity has underpinned an array of grand theories of political development. Unfortunately, these theories point confusingly in conflicting directions in terms of how they relate to understanding of tax politics. First, democracy is seen by as an almost inevitable outcome of economic modernization [38], while others see it as a necessary precursor to democracy [39]. Democracy is a contradictory actor in theories of the development of a market economy, because an autocratic state can’t be trusted to respect property rights, while a democratic state runs the risk of empowering the impoverished masses to confiscate the assets of the wealthy. As Weingast puts it, ‘a government strong enough to protect property rights and enforce contracts is also strong enough to confiscate the wealth of its citizens’ [40].

For this reason, some theories of democratization have looked at how different levels of inequality facilitate relations between the property-owning classes and the rest, albeit producing quite different predictions. Boix argues that inequality makes democratization less likely, because asset holders will be afraid to concede power to those who might tax their wealth away [41]. Acemoglu and Robinson see a curvilinear relationship between inequality and democratization, with middling levels of inequality more likely to produce a transition to a democratic regime [37]. Ansell and Samuels and Ziblatt show how high inequality can, under certain circumstances, facilitate democratization [42, 43]. All these approaches see the elites’ fear of expropriation as the decisive factor in provoking transitions to democracy.

The development of fiscal institutions in democratic countries also reveals an important role for war (and other shocks) in generating the kind of redistribution which wealthy elites have historically resisted [44, 45]. The availability of better longitudinal data over the past decade has confirmed that top income shares in western democratic countries followed a clear declining pattern for most of the 20th century, with the top percentile’s pre-tax share of national income dropping from close to half in the 1920s to more like a third by the 1970s [46], a trend shared by countries as diverse as the United States, France, Britain and Germany. The Great Depression and the Second World War appeared to play an important role in forcing a dramatic downwards redistribution of income, which resulted from a combination of high inflation which eroded capital returns, political changes in the ownership and regulation of capital, and taxation [46 p147–50]. War required governments to efficiently expropriate wealth to finance the survival of the state, but it also had a more profound effect on the politics of taxation by bolstering the moral case for redistribution: Scheve and Stasavage show that arguments that the wealthy should pay more tax to compensate for the sacrifices made by the working class to fight the war were a powerful driver of progressive taxation in the mid-20^th century [44].

The expansion of western governments’ fiscal capacity and the increasingly progressive nature of tax policies continued in the decades after the Second World War. The wartime taxes on top incomes were maintained and even ratcheted up, reaching remarkable marginal rates of over 90 per cent in both the United States and the United Kingdom in the 1970s (albeit effective rates remained lower) [44 p63]. This all the more extraordinary in the light of the direction those two countries took shortly afterwards, although it reflects the broad shift in power away from capital owners towards what could be loosely described as the labouring classes in the post-war period. This shift was reflected in the growing strength of the organizations representing the labour movement, most notably trade unions in the workplace and left-wing political parties in the electoral arena [47]. This mobilization and organization of labour – in the context of the Cold War and the communist challenge from the East – conditioned policymakers, and governments responded by expanding social spending, enhancing employment rights and curbing the discretion of owners and managers in a variety of ways. The rise of tax revenues and government spending across western democracies in this period matched the mobilization of the power resources of the labour movement [48].

Labour power peaked in the 1970s, and the push to grow the state’s fiscal capacity for redistributive purposes slowed or even reversed, starting with the UK and the US, where the Thatcher and Reagan neoliberal revolutions took aim at the ‘leaky bucket’ of tax and spend [49]. Income taxes, especially at the top, were reduced, and taxes on capital were also cut in many countries, as policymakers bought into the idea of an ‘overloaded’ state, one which had been forced to spend profligately to meet the ‘insatiable’ demands of voters [50]. The emergence of a new policy orthodoxy of limited government, with key economic policy instruments taken out of the hands of elected politicians, put the brakes on welfare state expansion. Slower economic growth meant that tax rises would be felt more keenly by voters, whilst the prospects of taxing the wealthy ran up against the increasing mobility of capital, newly unshackled from the export controls of the Bretton Woods era. This much more unforgiving environment not only made politicians loathe to promote higher taxes but also made tax reductions an increasingly popular electoral offering. Margaret Thatcher famously blamed Britain’s economic problems in the 1970s on the fact that Labour governments ‘always ran out of other people’s money’.

But although the inexorable government growth that worried Meltzer and Richard did seem to be arrested by the political changes of the 1980s and 1990s, there is actually little evidence that this neoliberal turn led to the ‘rolling back of the state’ that its proponents aspired to. Taking the OECD countries that have data back to the mid-1960s, peak tax revenue as a share of GDP was in 2016; although the pace of revenue growth slowed in the 1980s and 1990s, and flatlined in the 2000s and 2010s, taxes raised as a share of the economy have not fallen despite decades of political effort (see Figure Three). However, this average obscures some interesting variations across welfare regime types. Figure Three also shows that the liberal regimes have indeed tended to reduce the share of tax revenue in GDP, as have the Nordic countries (albeit from much higher levels). On the other-hand the continental welfare regimes have tended to keep revenue shares fairly stable since the 1980s, and the Southern European countries have a continued their upward trend (although since the 2010 Euro crisis this has been driven by a shrinking economy as well as a concerted effort at raising taxes to close their large budget deficits).

These patterns suggest that proponents of the median voter theory of redistribution were not entirely mistaken to point to strong pressures from voters to maintain patterns of taxation and spending. Even in the cases where political impetus behind shrinking the state was strongest, the tax take could only be pared back to the levels of the 1970s or 1980s, the period which had provided the evidence for their claims of an overbearing state that was stifling initiative. In Southern Europe, economic growth was accompanied by the process of increasing government spending and redistribution observed elsewhere, which has survived even the austerity measures imposed in return for financial assistance during the debt crisis. Despite the headwinds of the recent period, public support for high taxes and redistributive social spending remains strong, especially in the generous welfare states of Northern Europe [51]. However, in some places, notably the UK, public enthusiasm for redistribution has declined [52], and in general there is (not surprisingly) greater support for the spending side than for the taxation necessary to fund it.

Death and Taxes: The Decline of Labour and the Exit of the Wealthy

The survival of ‘big government’ in the face of the neoliberal onslaught confirms that it is a lot easier to promise a slimmer state than to actually achieve it. As Paul Pierson puts it, welfare states are an ‘immovable object’ due to the institutional inertia and intense electoral resistance to cuts in social rights [53]. As a result, the ambition of significantly reducing the tax ‘burden’ appears to have run out of political momentum, with even conservative politicians (outside the US) increasingly hesitant to include commitments to general tax reductions in their electoral programmes. But within a panorama of relatively stable overall levels of taxation and spending, we can still observe major transformations with distributional consequences.

The ‘neoliberal turn’ of the 1980s and 1990s may not have led to large reductions in overall tax levels, but it did spur a decline in the progressivity of taxes [54]. Ideas such as the Laffer curve, which claimed that the incentive effects of tax cuts would generate more revenue than they cost, and ‘trickle down’ theory, which argued that lower taxes on high earners and investors would benefit the wider population by raising economic growth, inspired large reductions in taxes for the wealthiest. Top marginal tax rates, which averaged around 60 per cent across the high-income democracies in the decades following World War Two, fell to under 40 per cent by the early 21st century. These policies managed not only to increase post-tax income for the wealthiest, but also likely stimulated high earners to pursue higher pre-tax incomes with even greater determination, contributing to an increasingly top-heavy income distribution [46, 55]. A recent review found that there was no evidence of any wider benefits beyond the direct beneficiaries [56].

Growth in income inequality driven largely by the concentration of income at the top combined with lower taxes on top incomes can only mean one thing: that the burden of taxation is shifted downwards if tax revenues are to be maintained. However, with high inequality, many citizens have little income to tax. As a result, income groups outside the wealthy elite are called upon to contribute more. This has taken the form of a rise in two less progressive forms of taxation: social security contributions and indirect taxes on goods and services, as Figure Four shows. In contrast, personal income taxes, after growing up until the 1980s, have declined as a share of national income. Corporate tax revenues have risen, but raise much smaller amounts (around three per cent of GDP on average across the OECD), whilst property taxes have barely moved, despite dramatic rises in real estate prices in recent decades.

So, while redistributive spending has been maintained across most OECD countries, the tax revenues that pay for it are increasingly raised in ways that place a higher burden on the middle-income voters whose behaviour underpins the median voter model. Capital owners – very wealthy individuals and corporations – have enjoyed a growing share of income [46], at the same time as enjoying rapidly multiplying opportunities to shield themselves from fiscal liabilities. Fears of capital flight in a world in which capital controls have mostly been removed have led governments to reduce corporate taxation rates, whilst the availability of tax havens provide large corporations and wealthy elites with many options to avoid even these [57, 58]. In tandem with this, the rise in social security contributions and consumption taxes, and the parallel fall in personal income tax means the burden for financing the welfare state falls heavily on its direct beneficiaries, the wage earning population, whose share of national income has been shrinking relative to capital since the 1980s. Meanwhile, the discomfort of policymakers towards the persistent budget deficits of the last 40 years, which exploded after the Global Financial Crisis, has provided further pressure to raise taxes (as well as cut spending) [59].

This trend towards a more regressive tax system is not as much of a novelty as might be assumed [60]. The social democratic expansion of the welfare state in Scandinavia did not in fact rest on a levying of corporate profits to pay for social transfers, but instead incentivised capital owners to reinvest, and expanded indirect taxes instead [12]. However the rise of the regressively financed welfare state took place in an age of higher economic growth, lower inequality, and lower capital/income ratios. The politics of sustaining this model in an era of economic stagnation and higher inequality, where the wealth of the top one per cent is visibly growing much faster than for the rest, is likely to be increasingly tense. Voters are torn between paying for government spending they mostly approve of and maintaining disposable incomes which are under pressure. Yet, the kinds of political outcomes that would be predicted by workhorse political economy theories – labour mobilisation and median voter demands for more redistribution – seem to be absent. The political forces that dominated the era of welfare state construction – social democratic and Christian democratic parties – are in steep decline [61].

Political scientists have offered two broad sets of explanations for the paradox of the decline of ‘Robin Hood’ policies in an era of rising inequality. Perhaps the dominant view takes voter preferences as the drivers of change, arguing that electoral support for the kinds of solidaristic fiscal values that underpinned the post-war growth of the welfare state has been compromised. For example, survey data from Great Britain shows an increasing rejection of redistribution by voters over the last three decades, although this does not appear to be a general pattern [62 p26–28]. A more common argument is that the increasing salience of a ‘second dimension’ of political competition along cultural lines, which has replaced social and occupational class as the chief driver of voter preferences [63], undermines support for redistribution and the centre-left parties who have favoured it.

Instead, increasing numbers of lower income voters are attracted to parties of the ‘populist radical right’ who focus more on the need to protect borders than on the redistributive opportunities of the tax system (indeed many have historically argued for lower taxes) [64]. A variant on this argument is that changing occupational structures have transformed the class basis of redistribution preferences, undermining the typical left-right battles of tax and spend that characterized the politics of welfare expansion [65]. These developments dovetailed with the theoretical refinements of the standard median voter model with multiple dimensions of electoral competition, which mostly imply greater difficulty in assembling pro-redistribution voter coalitions by shifting the median voter’s position further up the income scale [7].

This more ‘electoralist’ take on changing policy priorities, however, runs the risk both of understating the impact of economic change on patterns of voter behaviour, and ignoring the ‘supply side’ of politics [66]. First, focusing on the shifts in occupational groups, and especially the ‘upskilling’ of large shares of the workforce as the services economy has developed, can obscure the ways in which the distribution of economic risk has become more unequal over time. This manifests itself in the increasingly skewed income distribution in many countries, but also in the growing role of wealth in structuring economic interests and political preferences. Financialization has ushered in a new politics of wealth, ownership and debt that has shaped political preferences for redistribution in ways quite different to standard accounts of class conflict based on occupation [67, 68, 69]. In particular, wealth inequality is much higher than income inequality, with most assets concentrated at the very top of the income distribution [46]. In this kind of economy, ownership of real estate can matter quite a lot more than occupation or earnings in shaping political allegiances. In many countries age, which is strongly correlated with asset ownership, is an important predictor of voter choice [70]. A large ‘grey power’ bloc of older property-owning voters acts as a brake on attempts to redistribute wealth.

The rising importance of wealth inequality can also be seen in the behaviour of political parties themselves. The decline of organized labour and of political parties as membership organizations has made politics increasingly elite-dominated [71, 72]. A growing literature has documented the ways in which the overwhelming influence of the wealthy shapes political decision-making at the expense of middle- and lower-income voters, especially in the United States [73, 74, 75]. This can occur directly, through campaign contributions and lobbying, or indirectly, through the growing blackmail power of mobile global capital. The veto power of great wealth has a predictable effect: political parties are reluctant to commit to the kinds of tax policies that would redistribute resources from the super-rich to the broader population, and increasingly converge around an anodyne set of broadly market-friendly policies that will preserve this unequal set of power relations [76, 77]. The post-financial crisis backlash over the past decade has shaken this party ‘cartel’ [78], but has yet to generate the kind of political coalition that could challenge the policy orthodoxy of the past 30 years.

Conclusions

Debates around redistribution and inequality in political science tend to focus on the ways governments spend money rather than how the revenue is raised to fund this spending. However recent research has shifted the focus back on tax, and particularly the taxation of the super-wealthy. This paper has looked at how democracy and the upheavals of the 20th century built up sophisticated arrangements for redistributing economic resources through taxation and government spending, how these vary across countries in enduring ways, and how the pressures of the era of globalisation and crisis have conditioned the survival of welfare capitalism. Tax revenues continue to account for the best part of half of national income in most OECD countries, which has shielded lower income groups from the full effects of rising inequality of market income, but political constraints on the ability to tax top incomes and great wealth have placed this redistribution settlement under strain. It is now a priority for researchers to understand better how wealthy elites interact with politicians and political parties, and how the structural power of capital, and the instrumental power of lobbying and political finance, conspire to obstruct attempts to redistribute through taxation of wealth. This has left welfare states underfunded and most voters denied their preferred policies.