The Politics of Redistribution Redistribution through Voting

The Politics of Redistribution

Redistribution through Voting • Democracy often leads to redistributive policies. – In principle, any 51% of voters could vote to redistribute income or wealth to themselves from the other 49% • Note that this does not imply any notion of social justice – For example, the richest 51% could vote to enslave the poorest 49%. – A religious majority could vote to tax religious minority groups. – Populations with aging populations might expect to see large intergenerational redistribution

Public Spending on Private Goods • Again, we are returning to the concept of using public finance to supply private goods. • Recall that there is generally not an efficiency case to be made for such spending • This type of spending is justifiable (primarily) on equity (social justice) grounds alone

Voting Coalitions • Question: If we see redistribution through voting, on what scale are we most likely to see it? • Answer: Minimum Winning Coalition— You should expect to see coalitions with 51% of voters. – The more people in the majority, obviously the less they can collect from the minority – The fewer people in the minority, the larger the burden of being in the minority

Minimum Winning Coalitions • The nature of democracy creates incentives for redistributive coalitions to form, and these coalitions will be of the 5149 type. • Imagine a scenario with 100 individuals, and assume 51 of them vote to redistribute income from the other 49. • The 51 will receive $500 each, and the 49 will lose $600 each (DWL).

Minimum Winning Coalitions • There is an incentive for people to leave the majority and join the minority—why? – The 49 in the minority could all chip in (say) $50 each and promise $1225 to the first two defectors. • Suppose this happened—surely the new minority could bribe 2 people from the new majority…

Coalition Stability • Note that this is similar to the cycling result from voting and vote trading before. • What will determine if a coalition is stable or unstable? – Reputation—if you are known as a “switcher, ” nobody wants you in their coalition anymore – Fixed costs—if coalition building is costly, individuals have less incentive to switch – Expectations—Individuals know that this switching is possible and will demand guarantees.

Limits on Redistribution • Income distribution in society is often skewed, so the mean income is above the median income • The median voter would like to increase his income, but there limits on how much he can do so. – If he raises his income too much, he will no longer be the median voter! – The stable coalition will be that of the median voter and everybody poorer than he

Limits on Redistribution • Wealth distribution often has the same shape, but is less likely to be redistributed than income – Wealth taxes are essentially “retroactive” taxes. – Violate principles of Rule of Law • While many people think that wealth redistribution does not have DWL, in fact it does. – Changes incentives for accumulation of wealth, which distorts incentives and decisions on the margin – A one time, unanticipated wealth tax will not have a DWL, but will generate future inefficiencies

Income redistribution • Why will the stable coalition will be that of the median voter and everybody poorer than he? • Assume: – a non-discriminatory “flat tax” of amount t. – n individuals in society – each individual earns an income of yi which varies among individuals. – No DWL

Income redistribution • Revenue=ty 1+ty 2+…+tyn=t(y 1+y 2+…+yn) • Tax revenue is redistributed equally to each of the n taxpayers, so each person receives an amount of money • S=t(y 1+y 2+…+yn)/n=tyavg • Each individual’s after tax income is: • Ii=(1 -t)yi+tyavg=yi+t(yavg-yi) • If we ask people to vote on t, what will happen?

Income Redistribution • Recall that Ii=(1 -t)yi+tyavg=yi+t(yavg-yi) • For the person where yi=yavg, they are indifferent across all tax rates. • Those who earn more than the average want a 0 tax rate • Those who earn less than the average want a 100% tax rate • Since the median is less than the mean, the outcome of voting would be full redistribution • Why don’t we see this in the real world?

Income Redistribution • Excess burden of taxation • Would anybody work if there were a 100% tax rate? • If the median voter knows this, would he want 100% redistribution? – Probably not.

Income Redistribution • Taxation also has a justice component • Horizontal equity – Requires tax laws to treat people with equal incomes equally • Vertical equity – Requires tax laws treat people with unequal incomes equally • Neither are fully evident in practice, however the consideration of both reduces the desire of the median voter to levy fully redistributive taxes.

Income Redistribution • Social Mobility – Would the poor vote for high tax rates if they thought they might have a high income in the future? • The ability to realize social mobility also reduces the desire for income redistribution.

Suffrage • From the preceding slides, it should be obvious that the extent of income redistribution will depend on who votes and who has the right to vote. – Historically, expansion of the suffrage has led to greater degrees of income redistribution. • For example, historically, granting the right to vote to women has led to an increase in government transfer payments.

Intergenerational Income Redistribution • Tax finance v. bond finance of spending has a large intergenerational component. – Bonds are simply obligations to tax future generations. – Future generations are usually not allowed to vote, especially if they haven’t been born yet. • Social Security and Pensions – In the next 20 -40 years, demographers are predicting that, in a number of countries, over half of the population will be retired!

Constitutional Restraint • In many cases, democracy can not be relied upon to limit redistribution, especially when those who will be paying taxes cannot vote. • Various examples of this include: – Balanced budget amendments (limits debt financing) – Restricting the ratio of government debt to GDP

Income Redistribution in Democratic Societies • What sort of redistribution do we actually see? • While the median voter does not benefit much from publicly financed income transfers, they do benefit disproportionately from public spending on goods or services. – Education—the rich send their kids to private schools, while the poor tend to go to school less – Police—the rich often buy security, and the poor have less stuff to defend.

Income Redistribution in Democratic Societies • There are limits with income redistribution that make it very difficult to actually do – As the amount of transfers increase, it becomes more difficult to target those being taxed and those being subsidized. • Fiscal Churning—the practice of taxing and subsidizing the same people – All this does is generate DWL and bureaucratic waste without actually doing anything.

Fiscal Churning

Political Behaviour and Public Policy The principal-agent problem in Representative Democracy

Public Interests and Special Interests • There will be a principal-agent problem between voters and politicians if the personal objectives of political decision makers differ from the policies desired by the voters. • What is the personal interest of a politician? – Winning elections (vote maximization) • This requires both popular support and lots of money

Public Interests and Special Interests • Politicians need money for lots of things: – – Campaigning Advertising Mass mailings Polls • In order to get this money, political candidates will seek campaign contributions, often from individuals, groups, or corporations whose agendas are not aligned with the “public interest”

Public Interests and Special Interests • What is the “public interest? ” – Concept is extraordinarily nebulous because individuals differ considerably within any society • Nobody may agree with the median voter • Consensus through Clarke taxes or Lindahl solutions are impractical • Consensus on a Social Welfare function is extremely unlikely to be found (Arrow’s theorem) • In spite of these difficulties in defining the “public interest, ” it is often discussed as though it is easily identifiable!

Public Interests and Special Interests • One thing we are pretty sure of, however, is that policies that provide discriminatory benefits to a small group in society are often not in the public interest, especially when they impose losses on the society at large. • This is what special interest politics is all about.

Public Interests and Special Interests • A special interest group is simply a collection of individuals who, for whatever reason, are in favour of the passage of certain public policies. • The policies may be narrow or broad, and are not necessarily “bad” – – Environmental groups Pro/anti abortion groups Pro/anti gun groups Pro/anti tax cut groups

Public Interests and Special Interests • Typically, the most successful interest groups will have narrow interests. • This is because special interest groups are engaging in collective action, and thus are subject to a free rider problem. – Groups with narrow interests are often composed of a small number of individuals, and thus can more easily overcome free-riders. • Typically, large groups (with broad policy interests) can only offer votes, whereas small groups (with narrow interests) can offer money.

Public Interests and Special Interests • Political Decisions are often the source of rents. • We refer to the incentives to obtain and protect rents as rent seeking. • Although rents are not always generated through political mechanisms, it is the predominant way in which rents are created.

Public Interests and Special Interests • Special interest groups seek rents in the form of political favours in exchange for campaign contributions. • Politicians who accept such contributions are entering into an implicit contract with their contributors. • When the interests of the special interest groups differ from that of the “public interest, ” accepting this money constitutes a principal-agent problem.

Voters • The principal-agent problem can be fixed if the principal (voter) can provide incentives for the agent (politician) to act in the interests of the principal (public interest) rather than the interests of the agent (special interests). • There a lot of problems that make such monitoring nearly impossible.

Voters • As said before, large groups will have difficulty raising money due to free riding, and can only offer votes. • Small groups, while they offer relatively few votes, can donate large sums of money. • Note that there is a free riding problem within any given group—sure a voting bloc of (say) 5000 voters is powerful, but it is hardly more powerful than a bloc of 4999 voters.

Voters • Group size is also important because of the concept of “dispersed costs and concentrated benefits, ” or “concentrated costs and dispersed benefits. ” • The concentrated group is more likely to form a interest group because the stakes are high for them.

Voters • Example of dispersed costs and concentrated benefits: Automobile Tariffs • If an automobile tariff is enacted, it will create a massive benefit for a very small sector of society —those engaged in the production and sale of automobiles. • The costs will obviously outweigh the benefits (DWL), but those costs are dispersed over the entirety of the car buying public. Each individual in this group has a very small stake in this policy.

Voters • Elections are the primary means by which voters can discipline politicians—if a politician shirks, he/she can be replaced • Two problems with this method of discipline: – Voters typically take a retroactive or backward -looking approach, which leads to incumbency bias (were they good v. were they good enough) – Voting does not happen very often.

Voters • Voters are rationally ignorant. • Because information about politicians is costly to acquire, they will only acquire information up to the point where marginal private benefit=marginal private cost of acquisition. • This typically leads to voters being grossly uninformed about the behaviour of their elected politicians. • Voters will also be uninformed about most of the policy positions espoused by candidates, focusing on just a few issues.

Voters • Monitoring politicians is a public good subject to the free rider problem—I would like the politician to be monitored, but I’d prefer some other taxpayer/voter to do it. • This is worsened by the ability of politicians to “hide” rents given to special interests – Hiding special interest concessions in unrelated legislation – Using strange terminology (e. g. calling tariffs “variable levies” or referring to quotas as “voluntary”) – Giving subsidies indirectly rather than as cash grants (industry tax breaks, exemption from regs, etc)

Voters • Rationality of voting – The likelihood of any one vote influencing the outcome of an election is ridiculously small. • In general, any individual vote does not matter – If people place a positive value on time, why would you engage in an activity that has an anticipated benefit of close to zero? – The disinclination of individuals to vote strongly works against the ability of voters to discipline politicians.

Voters • Let: • • UP=utility if preferred candidate wins UL=utility if other candidate wins P=probability that your vote influences the election C=opportunity cost of voting • In equilibrium, you should vote if: • P(UP-UL)>C • P is extraordinarily small • If MVT is correct, so is UP-UL • Opportunity cost of voting almost certainly outweighs any expected benefit of voting.

Voters • Rationality of voting – Why do people vote? • • • Voting as a consumption good Minimax regret Social norms Peer pressure Penalties for not voting – Changes previous equation to: • B+P(UP-UL)>C

Voters • In order to monitor politicians, voters are reliant upon an objective and informative press. – This can help reduce rational ignorance • This role is undermined when the press is not objective – Often the press is subservient to, or prejudiced against a particular party or viewpoint.

Voters • Even when the press is objective and informative, a free press is unlikely to be that important – Complex ideas are often compressed into sound bites, compromising the information delivered. – Most stories disappear after a few days, giving journalists little incentive to provide in-depth and quality news stories. – Politicians try to create “spin” that is not easily seen through.

Politicians • To win elections, politicians need votes and money. – Usually, politicians are not able to acquire enough money from their own personal wealth to finance an election campaign, and have to rely on special-interest money – Taking special-interest money typically entails taking positions that differ from those that maximize social welfare

Politicians • To get special interest money, a politician will have to offer favours (rents) to the special interest groups. • A rational politician wants to offer short term or easily reversible rents (i. e. tax concessions) rather than long term or permanent rents (i. e. a freeway) – The politician wants to have something to offer during the next election cycle.

Politicians • Do we have evidence of this sort of behaviour? – Zupan (1990) finds that compared to “professional” politicians, retiring politicians have voting patterns that more closely mirror the preferences of their constituents.

Politicians • These sort of “deals” often put the politician in a position where he must espouse contradictory policies. – This puts a politician in a difficult position, as he wants credibility from both the interest group and voters in upcoming elections. • How to resolve this problem? – Hiding special interest benefits – Spin – Asking who is likely to have the shorter memory

Sources of Campaign Financing • In many democratic countries, campaign financing comes not only from private sources but also from the government itself: – Matching grants – Lump sums based on votes received in prior elections

Sources of Campaign Financing • Matching grants make the principal agent problem worse – This effectively subsidizes politicians’ moving away from the median toward a policy preferred by a special interest group. • Grants based on prior performance create barriers to entry – New parties have difficulty getting started and have trouble competing with existing parties. – Still does not obviate the need for pandering to special interests.

Maximizing Political Support • A politician seeks to maximize some political support function: • S=ƒ(U 1, U 2, …, UN) • Note that this looks remarkably similar to a SWF • Let’s say we rewrite this as an expected vote function, where: • V=Σƒ(Ui(PX, PY)) • PX and PY are the policies espoused by candidate X and Y, respectively.

Maximizing Political Support • Politician X will choose a platform PX that maximizes votes, and in equilibrium: • V*=Σ∂ƒ/∂Ui ∂Ui/∂PX at the optimal policy PX • It turns out that this is analogous to maximizing a Bergsonian SWF, where ∂ƒ/∂Ui is the individual lambda weight for each individual. – Here, the lambda weight characterizes the responsiveness of each voter to a change in the policy variable—voters who are “on the bubble” get a higher weight.

Maximizing Political Support • These equations imply that politicians are willing to make tradeoffs—good policy for special interest money. • The rate at which the politician is willing to make tradeoffs (analogous to a marginal rate of substitution) depends on how responsive to policy special interests are relative to voters.

Maximizing Political Support • Note that we have changed our assumptions about voting. – Deterministic voting—voters vote with 100% certainty for whichever candidate delivers them higher expected utility. – Probabilistic voting—voters are more likely to vote for the candidate that delivers higher expected utility—also called stochastic voting.

Probabilistic Voting • Viewing the world as being characterized by probabilistic voting rather than deterministic voting has a number of advantages: – Allows us to take account of rational ignorance – Politicians don’t have perfect knowledge of constituents

Probabilistic Voting • Viewing the world as being characterized by probabilistic voting leads to a few different implications. – No chance for electoral cycles (although every election will be predicted to be a tie) – Politicians seek to maximize a Bergsonian SWF – Politicians want to implement the policies preferred by the average voter (mean voter rather than the median voter).

Special Interest Politics and Corruption • Typically, special interest politics is not viewed as corruption. – Corruption is viewed as using political clout to illegally trade policy for personal gain, through bribes or kickbacks, for example. • In many countries, there is the expectation that the ruler will be corrupt. – These are typically non-democratic countries. – Bribery bears little risk (for either the briber or the person receiving the bribe)

Summary • Representative democracy creates a principal-agent problem between voters and politicians. – Voters have little incentive to monitor agents and few ways of doing so. – Consequently, politicians face few limits in their ability to pursue personal gain through legal means, although corruption is more closely monitored in most modern democracies.

Public Policy and Rent-Seeking

Rents and Rent Seeking • What is a rent? – A rent is a personal benefit that a person receives beyond that necessary to provide incentives to perform particular tasks. • Rent seeking is the using up of real resources in an effort to secure the rights to economic rents that arise from government policies • Why is this important?

Rents and Rent Seeking • Think of the economy as a pie. Real resources can be used for one of two purposes. – Making the pie bigger – Making your slice of the existing pie bigger • Rent seeking is the latter of these activities. – A side effect is that often, when making your slice bigger you are also making the pie smaller! • Even though rent-seeking is a form of competition, the use of scarce resources (e. g. time, effort, money) in rent seeking is usually considered to be socially wasteful.

Rents and Rent Seeking • Rents come in many forms – Monopoly privileges – Tariffs and import quotas – Regulation – Direct transfers – Public goods (through distributional effects) • Just about everything a government does creates winners and losers—these are almost all rents.

Rents and Rent Seeking • In general, the value of the rents are going to be dissipated—that is, any value the rent imparts on it’s winner will be spent in the process of rent seeking. • Example—if government is willing to distribute a monopoly privilege worth $1 M, how much are you willing to pay for it? – Obviously, up to $1 M • The social losses of rent seeking are the expenditures of real resources to attain that rent.

Rents and Rent Seeking P Transfer from consumer to producer (rents) P* DWL MC Peff D Q* Qeff MR Q We all know that the DWL is inefficient, but traditionally the transfer is not considered to be inefficient. However, firms are willing to expend real resources with value up to the entire value of the rent, implying that the transfer may be a measure of inefficiency as well!

Rents and Rent Seeking • If the rent-seekers were merely transferring money (i. e. bribes), would it still be inefficient? Probably • Rent seeking creates inefficiencies in 3 ways: – The efforts and expenditures of the potential recipients of the rents (less the value to those they are lobbying) – The efforts of government officials to obtain the expenditures of the rent seekers – Third party distortions induced by the creation of a rent.

Rents and Rent Seeking • Example: Wool growers want a wool tariff and bribe and/or give “gifts” to politicians. – Straight bribes don’t generate the first type of inefficiencies, but any in kind transfers generate inefficiencies. – Politicians and wannabe politicians know that wool growers are giving out bribes and stuff, so they exert effort to become the person who gets bribed and stuff in the future. – The tariff generates revenues, which may be fought over by different groups within government. It also demonstrates to other industries that there is “protection for hire, ” who may attempt to do the same thing now.

Rents and Rent Seeking • How big are rent seeking losses? • Rent seeking losses to an economy are surprisingly large. • This makes more sense when you take a larger view of rent seeking—any government project or policy with distributional consequences has the capacity to generate rents.

Rents and Rent Seeking Study Economy Year Welfare Loss Krueger (1974) India 1964 7% GNP Krueger (1974) Turkey 1968 15% GNP (trade sector) Posner (1975) U. S. Cowling and Mueller (1978) U. S. 1963 -6 13% GCP (monopoly) Cowling and Mueller (1978) U. K. 1968 -9 7% GCP (monopoly) Ross (1984) Kenya 1980 38% GDP (trade sector) Mohammad and Whalley (1984) India 1980 -1 25 -40% GNP Laband Sophocleus (1988) U. S. 1985 50% GNP Lopez and Pagoulatos (1994) U. S. 1987 12. 5% Dom. Consumption 3% GNP (regulation)