Category Archives: Redistribution II

LT9 redistribution readings….



Follow this link for full essay: INFORMALITY AND DISTRIBUTIVE POLITICS(final)

Rory Creedon London School of Economics (MPA ID)


An individual’s preferences regarding taxation may be derived from a number of sources such as the distance between his income and the average income[1] or notions of social justice. A further particularly salient source is that presented by Alesina and Rodrik [2]. Individuals, they argue, are endowed with labour, capital, or most likely a mixture of the two, whilst governments make productive investments financed by a tax on capital. The basic result of their model is that an individual who derives all of his income from capital will prefer the tax rate that maximizes the economy’s growth rate, whereas anyone who earns even part of his income by selling his labour will prefer a lower tax rate and a correspondingly lower growth rate.[3] To this insight I would like to add another: that the presence of a large informal economy will affect the preferences for taxation of both capitalists and wage earners.

The directional influence that a large informal economy will have on preferences for taxation is not discernable a priori. This is because the true nature of the key mechanism by which the informal economy affects preferences for taxation, namely the interaction between the informal and formal economy, is disputed. This essay analyses two major lines of thought on how the informal economy interacts with the formal economy. Dualists argue that the informal economy is a separate marginal sector not linked to the formal sector in any significant way. Structuralists on the other hand maintain that both the informal and formal economy are part of the same capitalist spectrum.[4] This essay does not assert the primacy of either of these views. Rather, within the stylized model presented by Alesina and Rodrik with the additional assumption of a large informal economy, I seek to emphasize that preferences over taxation will vary according to whether the true nature of the informal sector is closest to the dualist or structuralist tradition, thus reaffirming the importance of the debate. Arguments and examples are drawn largely from the literature surrounding the informal economy of Latin America as the extent of the informal economy in that region is such that it is impossible to ignore in terms of policy making and preferences over policy[5]. Additionally a particularly rich vein of scholarship has emerged in relation to the Latin American informal economy.

[1] A.H. Meltzer & S.F. Richard, A Rational Theory of the Size of Government The Journal of Political Economy, Vol. 89, No. 5, (Oct., 1981)  pp.916

[2] A. Alesina & D. Rodrik, Distributive Politics and Economic Growth The Quarterly Journal of Economics, Vol. 109, No. 2, (May 1994) pp. 465-90.

[3] Ibid. pg. 466

[4] M. Carr & M.A. Chen, Globalization and the Informal Economy: How Global Trade and Investment Impact on the Working Poor, ILO Employment Sector Working Paper on the Informal Economy, No. 1, Geneva, ILO pg. 5

[5] J.R. Franks Macroeconomic Policy and the Informal Sector in C. Rakowski (ed.) Contrapunto: The Informal Sector Debate in Latin America, New York: State University of New York Press (1994)




From Unequal Democracy: The Political Economy of the New Gilded Age

L. Bartels (2008)

In a Nutshell

Escalating income inequality is neither an “economic reality” nor a natural consequence of the US capitalism; it is the result of policy choices. Different patterns of inequality growth reflect consistent differences between the Republican and Democratic policy actions. Under Republican administrations real income growth for the middle and lower classes has consistently lagged behind income growth for the rich. Under Democratic regimes the lower and middle classes see much greater real income growth. Indeed under Democratic regimes inequality reduces slightly whereas it escalates under Republican governments. The data do not support the notion that these cycles are part of some cycle of calibration whereby Democrats expand the economy unsustainably and then the Republican government much see it contract. Nor can oil price shocks be blamed, as the results indicate that oil shocks affect households more or less equally and so have little effect on inequality.

The reason for the difference between the effects on inequality that the parties have is largely due to different macroeconomic policy. Democrats are more likely to run the risk of higher inflation in order to pursue expansive policies designed to yield lower unemployment and extra growth. Republicans fight inflation with tight fiscal policy etc. The result is that unemployment under Democrats is on average 2 % lower and real output about 6% higher. These elements of the economy have the most salient effects on lower income households and very little effect on rich households.

In terms of taxes the parties have very different philosophies. For example Kennedy instituted tax cuts at the lower end of the income distribution in order to stimulate demand to boost the economy. Reagan gave tax cuts to the rich as they have a larger role in saving and investment. Democrats come at the problem from the demand side, and the Republicans from the supply side. Reagan presided over a dramatic redistribution of the tax burden from corporations and the wealthy toward the middle and lower classes. A similar cycle was seen in the Clinton/Bush years.



V. Lacinese, J.Snyder & C. Testa

Centro Studi Luca d’Angliano Development Studies Working Paper No. 278 (Nov., 2009)


A Summary


In a Nutshell

This is an empirical test of three theories of distribution:

1.       The “swing voter” hypothesis which predicts that a larger share of distributive goods will go to groups or areas that contain larger numbers of swing voters (indifferent between parties on ideological grounds).

o   If voters  can trade off their ideological stances in exchange for public funds and projects then it is cheaper for politicians to “buy” the votes of the indifferent voters  so this will lead to a disproportionate allocation of funds to areas that contain many such voters.

2.       The “electoral battleground” hypothesis whereby distributive goods are allocated to areas where the share of supporters for each party is close to 50%. This is particularly relevant for FPTP systems with two major parties and well defined constituencies.

3.       The “partisan supporters” hypothesis in which politicians favour those areas that contain a larger percentage of their core supporters.

o   This could be a rational strategy in low turnout elections if spending primarily mobilizes voters. Thus it would make sense to target areas with core supporters.

o   It could also reflect the fact that politicians are to some extent policy oriented.


In all three it is assumed that all politicians care about is winning elections.

They use exit poll data on self-identification rather than voting results as voting is endogenous to the model of distribution so this tends to bias estimates in studies that use such data. By using the exit poll data they can construct a measure of “independent voter” that should be relatively clear of measurement error.



·         US data

·         Dependent variables: total federal spending pc.; total spending other than direct transfers to individuals (the most manipulable items of the budget); federal grants pc. (the most targetable).

·         Poll data measuring the share of independents based on self-identification. They show a clear positive correlation between self-identification and voting behaviours meaning that their measure can be thought of as a reliable indicator of partisanship.   



·         There is no evidence that states with more swing voters receive more funds. The swing voter hypothesis is not supported by this data set.

·         There is no support for the battle hypothesis either.

·         The is limited support for the partisan supporters hypothesis, although significance is extremely sensitive to the specification, and as voting is endogenous to the model, the use of vote data means that the estimate is probably largely overestimated. Thus it appears there is some evidence that states with a higher proportion of supporters for the party of the president receive more votes.

·         It also appears that states that have received more funds do not reward the incumbent in the polls. On the contrary, partisanship and ideology have large significant effects which is consistent with with Bartels who finds a strong impact of partisanship on voting behaviour both at presidential and congressional level.



·         If candidates typically fulfill their promises to voters regarding spending then there would be little correlation in the data between funds received and votes for incumbents as there would be no out-of-equilibrium behaviour.

·         It could be that institutional factors that are particularly strong in the US meaning that the behaviour these hypotheses predict are not feasible. For example checks and balances may prevent a tailoring of budgets to suit a politician’s preferences. Committees and agenda setting are also relevant. In other words perhaps although the politicians would like to engage in such behaviour they are unable to, and thus in states with fewer institutional checks, the data may return stronger support for one or more of the hypotheses that found little support in the US data.  



A. Alesina and D. Rodrik

The Quarterly Journal of Economics May 1994

 A Summary

“Economics is concerned with expanding the pie while politics is about distirbuting it”

In a Nutshell

The key finding of the theoretical model and subsequent empirical testing is that inequality retards growth. The government taxes capital in order to provide government services which are themselves essential for growth (think infrastructure, education etc.) and there is thus a growth maximizing level of tax and spend. However as individuals are endowed with different levels of capital and labour, the ideal tax rate is different for different individuals. An individual who derives all of his income from capital prefers the growth maximizing tax rate whereas everyone else would prefer a higher tax rate with a lower level of growth in the economy. The lower an individual’s share of capital income in his total income, the greater he want the tax rate to be, and the lower his ideal growth rate. The more equitable the distribution of capital within a society the more capital is owned by the median voter and thus the lower the tax rate will be. In other words inequality is inversely proportional to economic growth as distributional struggles harmful to growth are more likely to occur when resources are unevenly distributed.

The Model


  • Labour and capital are the primary means of production.
  • Private production requires the provision of public services i.e. the government has a constructive role to play. To finance spending the government has access to a tax on capital income. Capital can be thought of as all kinds of accumulated capital i.e. including human capital.
  • The tax on capital affects the incentive to accumulate which slows growth, whilst it increasing the wage rate as the government can spend to increase productivity.
  • Each individual can earn income from both capital and labour, and income depends not only upon personal capital ownership, but on the aggregate stock of capital as well. The higher the after-tax return on capital the higher the growth rate. At small levels of taxation the productivity effect dominates and returns to capital are increasing with the tax rate. At higher levels the incentive effect dominates, and returns to capital diminish with the tax rate. Thus it is an inverted U shape. It follows that the choice of tax policy affects both consumption and growth.


  • For a pure capitalist, the ideal tax rate is positive (as public services play a role in the aggregate production function) but it is precisely that which maximizes growth.
  • For any individual that earns some money from selling his labour, the ideal rate is above the growth maximizing rate. Thus a government seeking to maximize the welfare of a representative individual would not choose the growth maximizing rate. This indicates that in this context growth and welfare are not the same thing.
  • Additional taxation has a positive effect on those earning labour income, as government spending increases the wage level due to increased productivity. Thus there is an instant increase in consumption so it is rational for someone selling his labour to want a tax rate above the growth maximizing rate.
  • The greater the inequality in capital distribution the larger is the difference between the average capital holder’s, and the median voter’s stock of capital. Assuming that politicians are most sensitive to the median voter’s ideal tax position, this means that the tax rate will be above the growth maximizing rate. The further the median voter is away from the average capital holder in terms of the captia endowment, the further above the growth maximizing rate the tax rate I liable to be.

Empirical Evidence

  • Test using data from OECD and reliable developing countries.
  • Land distribution is the independent variable.
  • The results are that income inequality is negatively correlated with subsequent growth.
  • The estimate coefficients show that if the land Gini increased by one standard deviation then growth in subsequent periods would be 0.8% lower per year.
  • They include a “democracy” dummy which is not statistically significant indicating that this relationship holds in democracies and more authoritarian regimes. It seems that democracies do not grow faster than or more slowly than dictatorships. This either means that dictators face similar pressures to be sensitive to calls for redistribution from the median citizen as are elected officials, or it means that inequality operates on growth through a channel other than redistribution [such as through lack of demand for manufactures – Shleifer et. al].  
  • The results imply that countries that experienced land reform after WWII and hence reduced land inequality subsequently had higher growth than those who did not reform. This is often mentioned in the appraisal of growth in Taiwan, Japan and other east Asian countries that have performed very well since land reforms occurred as opposed to the Latin American economies that have not fared so well and have never been able to instituted wide spread land reform which means that land still tend to be concentrated in t eh hands of a small minority.



D. Acemoglu, J. Robinson

American Political Science Review Vol 95, No. 3, (Setp., 2001)

A summary

In a Nutshell

Redistribution often takes inefficient forms. For example it would be much more efficient to redistribute to farmers on the basis of lump sum transfers rather than price mechanisms which distort the market. Similarly it would be better to finance protected industries directly rather than distorting world prices and domestic resource allocation by the adoption of tariff measures. It is argued that such redistribution occurs because governments cannot credibly commit to the honouring of income redistribution by either itself, or by future governments. Thus, the ability of certain sectors to continue to be able to extract redistribution will depend in some way on the strength of the group in question. This in turn indicates that policies that maintain or even increase group size could be preferred, as larger groups in general have larger political clout supposing that they can overcome the collective action problems associated with larger groups. These policies will typically be inefficient. For example, for farm subsidies to be tied to production means that farmers are encouraged to stay in the sector, and new entrants are encouraged meaning that group size, and thus influence, is maintained. This implies that inefficient redistribution is more likely when the group is under threat from political marginalization i.e when its voting power is diminishing. It seems generally to be the case that declingin industries receive the most distorting transfers. Additionally where skills are relatively non-specific (meaning a worker/producer/member could transfer easily through the labour market) then inefficient redistribution is more likely as it will be harder for the group to maintain its size if it is easy to leave the group.

Alternative Explanations of Inefficient Redistribution

Inefficient methods are harder to reverse and so they serve in effect as a commitment device. [This seems somewhat implausible, there is no real logical reason why it should be harder to repeal a direct transfer law as opposed to a production subsidy law. However it does touch on another point being that subsidies etc. were initially intended to promote output growth, especially in the EU, and they are particularly hard to deal with due to entrenched interests. The economic history of subsidies in the EU offers a challenge to this model, as farmers did not choose the subsidies as they were in decline, in fact they were a very strong and large part of society. Rather output needed to be boosted after the war.]

Where politicians care about one subset of voters they use inefficient redistribution to confound the rest of the electorate. Whereas the giving of direct transfers would reveal his preference for the subset over the majority, a subsidy etc. can be made to look as though it addresses market failure whilst still achieving income distribution.

Application of the Model

Agricultural Policy

Farm policy cannot be explained by market failure and direct subsidies could save considerable resources although it does not appear a viable option. The model states that the current policy mix is designed to maintain a critical mass of farmers.

Labour Market Policy

Severance pay, although it is a good form of insurance, is inefficient. There would be increased efficiency from other means that do not distort the market leading to unemployment etc.

The reason such labour laws exist is to increase “insider” power. i.e. unions to remain powerful in order to continue to bargain for increasing wages, thus it is important that workers cannot be fired easily which would reduce the critical numbers of union members.



K. Murphy, A. Shleifer, & R. Vishny

Quarterly Journal of Economics 1989

A summary

In a Nutshell

In order for industrialization to occur within an economy domestic demand must be sufficient to allow for manufacturing activities that display increasing returns to scale to be undertaken profitably. This will occur more readily in countries which have a large population, relatively homogenous tastes, and concentrated population centres. Critically however, this cannot occur in societies with a radically unequal income distribution, as the natural consumers of manufactured goods are the middle classes. Thus an extreme concentration of wealth in a small minority will manifest itself in demand for handmade and imported luxuries rather than for domestic manufactures. This was seen for example in Colombia in the 19th century when there was a boom in tobacco exports, and subsequently another boom in coffee exports. As tobacco is conducive to cultivation on large plantations, the export earnings were concentrated in the hands of a few wealthy landowners who spent it on luxury imports. Coffee on the other hand could be grown on small family farms meaning that resulting wealth was spread much more evenly throughout the agricultural sector to a large number of people that demanded domestic manufactures thus spurring on industrialization. In general, expansion of domestic demand will involve

Domestic income distribution would not be relevant to industrialization if world trade were costless and free of barriers. In practice however penetrating foreign markets is hard given different tastes, protectionism, transport costs etc. Thus size of local market continues to be of great importance for many goods in many countries. For example at the time of writing 80% of goods manufactured in the developing world were consumed domestically.

The authors do not attach any normative value to industrialization, although they do note that it is often correlated with welfare improvements, growth etc. Technological spillover promotes further growth and capital accumulation etc.

Agriculture as the Leasing Sector

A boost in agricultural productivity can substantially increase the size of the markets for manufactures. This is because the increased productivity means agricultural employment falls as output rises, meaning that wages increase for those left in the sector. Their earnings may be such that the goods demanded will now include manufactures rather than solely food. Additionally, the labour that is freed from agriculture is absorbed by industry as output rises in the already industrialized sectors. This implies increased living standards and further increasing demand for manufactures.

Technological change in agriculture thus fosters industrialization.

[This would seem to imply that the Latin American economies that focused ISI on manufacturing sector at the expense of the agricultural sector were missing a more sustainable way of increasing domestic demand. Additionally, it would appear that development strategies should incorporate agricultural development as much as direct industrialization promotion.]

Failure to Industrialize

If all consumers are just buying food, there is no space for industrialization. In a poor society this failure to industrialize could happen in two ways: total equality, and high inequality. Total equality in a poor country could mean than no consumer has sufficient disposable income to spend on manufactures. In a highly unequal economy there are not enough people to make profitable a conversion to increasing returns to scale processes. There are not enough people to cover the fixed costs of industry so it will not arise. This result suggests that industrialization will occur more readily in countries with a larger population. [This was in fact the case in Latin America. By the 1960s it became clear that if ISI was to succeed throughout the region, domestic markets had to be expanded, as the only countries that experienced large scale competitive industry formation were Brazil, Mexico and Argentina which all had sufficiently large internal demand structures. The smaller countries of the region did not have sufficient domestic demand and so various projects of regional integration were started to try expand demand and hence increase specialization and industrialization.]

The higher is the fraction of profits in the hands of the middle classes the greater the demand for manufactures.

Thus in a poor country income inequality could be a driver behind the failure to industrialize even where there is sufficient investible capital within the state to get industrialization off the ground.


If there is a redistribution from the upper to the middle classes, but not so much that the identities of the members of those groups is changed then demand for domestically produced manufactured goods increases at the expense of luxury import items. Thus the extent of industrialization increases as does employment in those sectors.



A. Meltzer & S. Richard

Journal of Political Economy, Vol 89, No. 5 (Oct., 1981)

A Summary

In a Nutshell

Tax rates have been rising in the UK/US for a century, but why exactly? In the framework set out for analyzing tax policy, the size of government is chosen by rational utility-maximizing individuals who are fully informed about the state of the economy and about the consequences of taxation and redistribution. The key is that the size of government will be determined by the relation between the mean income, and the income of the decisive voter, which for democracies, means the median voter. If income distribution is skewed to the right, the mean income is above the income of the median voter and thus the median voter has an incentive to choose redistribution by taxation on incomes higher than his own. Therefore, any voting rule that means that the income of the median voter is pushed farther away from the mean income implies that the size of government will increase. As the last 120 odd years of British history has seen the franchise extended to include more people below the mean income, there have been increased votes for redistribution.

The logical extension of this argument is that the median voter will prefer expropriation and total equalization of incomes, however this ignores the incentive effects of taxation, meaning that as taxes and redistribution reduce the incentive to work, this provides a natural limit to the potential size of government. This is because a tax rate increase has two effects: each $ of earned income raises more revenue but earned income declines; everyone chooses more leisure and more people choose to subsist on redistribution. The median voter has to take this into account when he seeks to maximize his own utility.