Public Finance and Budgeting

CHAPTER 5 Public Choice and the Political Process

PUBLIC POLICY PERSPECTIVE

Publ1 Ch< ,~ U S C1 1M De Pc rnstatuho 11s Matt.,?

In the United States, municipal government takes two basic forms. One form relies heavily on a pro- fessional city manager who, along with a staff, makes the day-to-day decisions about how to run the c ity and plays an important role in advising elected public officia ls on expansion of public facili- ties. The city-manager forrn of government typically involves nonpartisan elections for mayor and city cou nci l members . In the second form, a mayor, dep- uty mayor, and other pllblic officials are elected directly, usually in partisan elections (in wh ic h candi- dates are members of a political party). In this form of government, the mayor and city council members wield considerably more power than they do under the city-manager form of government.

Does the form of city government affect the pub- lic choices made in mun icipalities? Some scholars have argued that politicians under partisan govern- ments respond to different incentives than city man - agers. For example, one researcher has argued that city managers act as technicians who view city capi - tal and labor simply as inputs for the production of city services. However, partisan mayors view these inputs partly as political assets that can affect their power base and ability to be reelected.'

To test the hypothesis that the form of city gov- ernment makes a difference for public choices, Kevin Duffy-Deno and Douglas R. Dalenberg col - lected data from 26 large U.S. cities chosen at ran - dom . Half were city-manager governments and ha lf were ru n by elected mayors.2 Using estimates of the capital stock for each city, researchers sought to expla in differences in city capital per person and capital-labor ratios between cities. City capital

included police and fire facil ities, parks, recreation centers, highways, water and sewer systems, and health and welfare facilities . Among the variables used to expla in these differences were the number of facilities located in the metropolitan area of each city of the sample, municipal population density, numbe r of manufacturing firms in the city, per cap ita personal income, percentage of owner-occupied homes, median age of city residents, the city's region, population change in the preceding five yea rs, percentage of homes built in or before 1950, and number of services offered by the city.

After accounting for the influ e nce of all variables, the researchers found that the per cap ita stock of cap- ital was 12.5 percent greater in those cities managed by elected mayors than those with the city-manage r form of government. The capital-labor ratio was also higher in the cities run by elected mayors rather than city managers. Although the form of government did not seem to influence total municipa l expend itures, it did seem to influence the means by which services are produced. Public choices in cities managed by elected mayors appear to result in more capital- Intensive production methods and larger stocks of capital relative to population. Therefore, public choice rules do seem to matter when it comes to choices of input mixes of U.S. cities!

•s.., J . S. Zax, "EconomtCS Effects of Municipal Gove mmenl lnstitution1,· wortingPape,No .1 657,National8ureauof Economic RMearch,1985 'S.. Kevin T. Ouffy-Oeno and Douglas R. Dalenberg, " Do ln,nitution, Mi tter? An Empi,i,:., I Note ," National Tu Journa f, 43. 2 (June1990): 207- 215.

POLITICAL PARTIES AND POLITICAL EQUILIBRIUM Individual s with si milar idea s on the role of governmem and other iss ues often group together to form pob1ic1l pa!'lie-i, Politicians seek elective office for a variety of reasons: power, prestige, desire to serve others, and personal financial returns aher leaving office. The y are influential in formulating the alternatives that are pre- se nted to voters and in dispersing information on the relati ve merits of alternative measures and candidates on the agenda for approval. When information is scarce, the behavior of politicians can be influential in determining the actual political equi- librium that emerges. Competition among political parties, particularly under

~- ... ,0 1,c....., u .. All-"""""'-""'1- >< _>< __ u _ _,_ . - .. - ...,_ Do< o,d« . nci,,...- • ..,,_,. , .,__.,,_.., , .... _."' ""-"'' · ..,...,_,, ... _ _,,..,._..,,a - ""' '"''"""'"'""' ~~-u ·,......,.,.. .,,.,,._ ..,. . ..., ·-..,.. -.,.,,, · "· ·

PART ONE The Economic Basis for Government Activity

a system of majority rule, has been analyzed by economists in an attempt to gain a better understanding of the political process. 8

Thus far, littl e has been sa id about the role of political parties in the formula- tion of the alternatives presented to th e electo rate. Clearl y, political parties play an importanl part in defining issues and in attempting to influence the results of elec- tions. For the individual voter, the marginal benefit of any particular budget pro- posal will depend not only on the level of expenditures but also on the mix of types of expenditures within the budget. The willingness of any citizen to vote favorably on any given budget also will depend on the particular tax-sharing plan proposed to finance the expenditures. In some ways, political parties act as brokers to encourage vote trading among voters. Political platforms ofte n include proposals for programs that benefit only a minority of voters. However, by including these benefits a nd spreading the costs over the majority , the parry ca n gain votes.

Economists characterize political parties as vo1e maximizers, because they tend to pu1 together political programs and tax-sharing arrangements designed to maxi- mize the voles that they receive. Under majority rule, the party most successful at maximizing votes wins the election. Therefore, vote maximization is a prerequisite to successfully obtaining political power in a democracy.

The Median Voter, Political Parties, and Political Equilibrium under Majority Rule Suppose the positions of political parries can be ranked according to a scale that measures the quantity of government activity per year. For example, conservatives who argue that government activity should be reduced or curtailed would rank low according to this sca le. Liberal ca ndidates who propose more government activity would rank high on the scale.

Political candidates tend to take a position that represents the median on the sca le. Political parries and candidates who take extreme positions arc likely to lose elections. The candidate who most accurately approximates the median most- preferred outcome will emerge as the victor.

This is illustrated in Figure 5.6. The graph plots the net benefit (after paying taxes) that each voter recei ves from each possible political platform on government activity. Assume that the greater the quantity of government goods and services per year, th e more liberal is the platform.

The graph assumes that th e most-preferred outcome of voters varies consider• ably. Some conservat ive voters' most-preferred outcome occurs at zero government goods and services per yea r. The most-preferred outcome of more liberal voters cor- responds to higher amounts of government goods and services per year. In Figure 5.6, the net benefit curves of nine voters are drawn. The median most- preferred outcome is Q•, which corresponds to the peak of the net benefit function of the median vote r. Assuming that each voter ha s a single-pea ked net benefit func- tion, Q• wi ll emerge as the political eq uilibrium. This is because the net benefits of more than half of the vorers wi ll be high er under Q • than for any alternative quan- tity of government goods and services, gi\•en tax shares.

Figure 5.7 shows that if the most-preferred outcomes of voters arc normally distributed , a political party can maximize the number of votes by raking a center

"Politic•I p.rtin •nd their influence on the political procn, ••e described in tt,,, clnsic study by Anthony Down•, An fcooornic n..ory of o.moc.-.tcy. (New YOfk: Harper & Row, 19S7)

--"" "'''c,...,u · .All ~ .,_.., _ _,-.,.-M<d. · .....,,. - ,..._ o., . · -- . ..,,_., ; ..,__._ "",11oo1.'"""'""-"•l · .._...,._,,...,...,_...,.,,_.,...,,.,,. _....., ,.,.,._ · c_ u_ ,,,_,,,. ..., .. .,.,,... --• ·--... ,,,__,,. · ·

CHAPTER 5 Public Choice and the Political Process

hii&HIS The Median Voter and Political Reforms

Net Benefit for the Median Voter /

a· Output of Government Goods and Services per Year

Given tax shares, the political party proposing Q• units of government goods and ser- E' vices per year will win an election over any other party proposing an alternative quantity. ! This is because 0- is closer to the most-preferred outcome of a simple majority of the J voters.

FiihH!ii Number of Voters and Government Output

a· Output of Government Goods and Services per Year f

The annual output, O", is the political equilibrium under majority rule because it can t obtain more votes than any other alternative. l

position corresponding to Q•. The implication of thi s analysi s is that political par- ti es or candidates who take extreme positions on issues are doomed to lose elec- tions. Political parties that seek to maximize votes will always have an incentive to straddle the median position.

Ample evidence exists to prove that when political candidates in the United States rake extreme positions they do in fact lose. For example, in the 1964 presidential election, Barry Goldwater, a presidential ca ndidate, proposed a platform that was apparently far more conservative than the one preferred by the median voter of thar time. The result was a landslide victory by his Democratic opponent,

- - ... ,01, c... u .. All_,.,......._ M .. .. -..-,_-_,_ . -,.- ,..._ Do< ,od« . --• ..,..-. ... __.,,_._ , .... ....,-"' '°"<'«"'' ·· .,....,_....,..,_,,...,,_..,,a - ""' .,,.,...,._ ~~- u · .....,.,.., -.,..,,._ --• ·.-_.-... ,,,- · ·

m

PART ONE The Economic Basi s for Government Activity

Lyndon Johnson. Similarly, in 1972, the Democrats chose George McGovern as their candidate, but his position was apparently far to the left of the median peak. The result was a landslide victory by the Republican candidate, Richard Nixon. In 1976, in 1he Jimmy Carter versus Gerald Ford race, both candidates proposed pla1- forms close to the median most-preferred position. As a result, the election was very close. Similarly, in the 2000 clC'Ction as both Al Gore and George W. Bush straddled the median on issues, the result was an election so dose that the results were hotly disputed. Recounts and court decisions were required to select the winner.

Over time, the median most-preferred outcome can change. For example, a movement to a more conservative point of view br voters can provide opportunities for candidates who propose conservative platforms to win elections. Perhaps the landslide victory of Ronald Reagan in 1984 can be interpreted as the result of a movement to a more conservative median most-preferred outcome. In 1992, the vic- tory of the more liberal Bill Clinton over a conservative George Bush reflected a movement to a more liberal median most -preferred outcome. The election of Barack Obama over the more conservative John McCain in 2008 can also be interpreted as a result of a movement to a less-conservative median most-preferred outcome than the previous eight years of the George W. Bush presidency. In any event, political issues are seldom single dimensional. In presidential elections in particular, many issues are bundled together in the platforms of the political parties. Often, ranking of candidates along a single dimension is imposs ible. For exa mple , a candidate might be regarded as conservative on domestic policy but li bera l on foreign policy.

The Effect of Nonvoting on Political Equilibrium Politicians seek to offer a political platform in line with the median most-preferred outcome of all voters. If all citizens do not vote, thi s outcome can differ from the median most-preferred outcomes of all citize,is.

Voters might abstain from vot ing because they arc indifferent to the platforms of rwo competing political parties. Barry Goldwater's campaign slogan in 1964 was "A choice, not an echo!" Perhaps he hoped that a significant number of conservative citizens who had not turned out in the past because they were indifferent to the posi- tions of the candidates might turn out in hordes to support his position even though it was to the right of that espoused by the median voter. Apparentlr, he was mistaken.

When vote rs do 1101 vote because the platforms of the candidates appear to he similar, the impact on political equilibrium depends on the number of voters on each end of the political spectrum who do not vote. Assume that candidates can be ranked on a single-dimensional sca le, such as the liberalness of their positions. The median most-preferred outcome of citizens sti ll will be the political equilibrium, provided that as many conservative citi7..ens as liberal citi7..ens do not vote because of indifference. Only in the case in which non voters are predominantly conservative or liberal will the median most-preferred outcome of all citizens diverge from the median most -preferred outcome of all actual voters.

Another reason for abstaining from voting is alienation. Citizens might find the positions of the cand idates too far from their own most-preferred positions to bother to vote at all. The effects of alienation on the outcome of elections are com- plex. Assume once again that the positions of voters can be ranked according to a single scale of liberalness. When voters choose not to vote because ther arc alien- ated, the tendency for political parties is to move toward the mode rather than the median of the most-preferred outcome of citizens. If the distribution of the

c-'"' "'" """""" u......_ .,._,.,.,,, . ..._,...,,_ .,_ .,_,,_,._M<d. '" .....,., .,,..._ o., _,<1«.....,rip,u,._...,_,,__..,, . ..,__._.,,, .... _.., ""-",1 · .._..._...,...,_...,.,_,.,..,.,a - ..., ..,,.,,. .. ,.__, · _c_ u · """""°""'"",._ --•.., ·.-_.. -.,.,.,.,,.,..,.,,. · ·

CHAPTER S Public Choice and !he Political Proce ss

most-preferred outcome of voters is symmetric (as is the case for the familiar bell- shaped normal distribution shown in Figure 5. 7), then the median and the mode will coincide. Under these circumstances, the median most-preferred outcome of all citizens still will dominate, provided that the distribution is unimodal. When the distribution of most-preferred outcomes is either asymmetric or multimodal (one that has two peaks), alienation can result in political equilibriums that differ from the median most-preferred outcome of all citizens. 9

VOTING ON MORE THAN ONE ISSUE AT A TIME : LOGROLLING When more than one issue is voted on simultaneously as a package-as is commonly the case-voters are sometimes confronted with political packages that include both favorable and unfavorable items. If the voter feels more strongly about some issues compared with others, or if he or she is better informed on some issues than on others, then his or her vote may be a function of the extent to which the package supports those issues he or she strongly favors. When intensities of preference differ on issues, there are incentives for groups to trade votes for those issues of great inter- est to them. Such a vote-trading process is called l(Jj(rolllng.

Suppose, for example, the ballot contains two issues, neither of which can pass separately because each issue provides benefits to only a minority of voters. One of the issues significantly benefits oil producers, and the other greatly benefits shoe manufacturers. Suppose that oil producers can gain considerably when their issue passes, and these gains outweigh any losses they might incur by voting for the issue favored by the shoe manufacturers. They then will have incentives to offer to vote for the issue of interest to shoe manufacturers in exchange for the shoe manu- facturers' positive vote on the issue of interest to the oil producers. The shoe produ- cers will agree to such a trade, provided that simultaneous passage of both issues will provide them with net benefits.

Incentives to trade votes exist when an asymmetry of gains on the issues is involved. If the oil producers' gains resulring from passage of their issue were exactly offset by losses from passage of the legislation of interest to shoe manufac- turers, the incentive to trade votes disappears.

Trading votes might not he success ful in accumulating enough votes ro pass an issue. However, opponunities for logrolling clearly result in passage of some issues that otherwise could not command a simple majority.

Incentives to engage in logrolling depend on rhe relative intensities of voters' prefer- ences on issues. If voters felt rhe same about all issues, the gains from passage of any one issue would be exactly offset by the losses expected as a result of passage of any other paired issue. Again, no incentive to trade votes exists under those circumstances. JO

Implicit Logrolling Implicit lr~rnll1nt1, occurs when political interests succeed in pairing two (or more) issues of strong interest to divergent groups on the same ballot or the same bill. This is a common practice in legislatures, where riders are ohen attached to bills.

' Otto A. Davis, Me Mn J. Hinich, and Peter C. Orde .t.ook, - An Expository Development of a Mathe matical Model olth& El&ctoral Proc&ss,- Am&ri.:anPolirica/Sc:i&nce Review64tJune 1970); 426-448. ' °For an 1nalysis of logrolling, Sff J1mes S. Colema.n, -n.e Possibility of e Social Welfare Function." America,, Economic Review57 (t)ecember 1967): 1311 - 1317.

-- "" "''' C<oc>c< U · A.II ~ .,_ ., _ ., __ u_M<d. ,. ......,,,, · ,..._ o., . · rip,u..,._ , .., ,_. ; ..,_...,. ,,_ ..,,11oo1. _.., ""-"'' ·· .._..._...,...,_.....,_,., ...,,. - "" °"'""'" "- ·.,. c_ u .• ....., ..... ...... ,... -- • .., · - ·"'"'"""""'"' " · ·

PART ONE The Economic Basis fo r Government Activity

PUBLIC POLICY PERSPECTIVE

State Governm•nt Spandtng DOH th• Sir• of th• Leg1slotur• Matter?

State legislatures have grov.,-, in size during the 20th century. Only a minority of U.S. state governments (about20 percent)have the same numberofseats in their legislarures that they did in 1902. In more than half the states, the size of the upper house has inaeased since the beginning of the 19COs, while 23 states have eKpanded their lower house as well. For state programs, a larger number of d istricts represented in the legislature suggests that the tax costs per district for state public goods ard services if distributed equally among districts will be lower. If constitl..lents of each legislator's district pay one nth of the cost of a statewide program, where n is the number of districts, the cost in each district falls 'Mlen the number of districts goes up. The theory of logrolling suggests that as more d istricts are available to distribute the costs of public sperding, there will be more incentives for irdividual legislators to engage in vote trading to expard state government spending.

To test th is theory, Thomas Gilligan and John Matsusaka collected data on spending by the 48 con- tiguous states from 1902 to 1942. 1 AccOl"ding to their research, the number of seats in state legislatures had a sign ificantly positive impact on state and local gov- ernment expenditure in the first half of the 20th cen- tury. However, it is the number of seats in the upper house (the Senate) that accounts for the upward effect of government spending. The impact on the number of seats in the lower house did not have a statistically significant impact on st.ate and local government spending during this period.

Gilligan and Matsusaka also examined the impact of political parties on spending. They could not find any significant effects from control of the legislature of the governor's office by varying politi- cal parties on spend ing over the period. The incre- ment in spending attributable to increased legislature size was most concentrated in govern- ment spending on highways and educat ion . Since these are two aspects of government spending that tend to result in geographically restricted ben- efits, this suggests that increased legislature size does contribute to more logrolling. The researchers also conclude that much of this increased spending was by local governments that often benefit from state aid.

As with most economic studies of this type, the researchers used statistical methods to account for other possible influences on state and local govern- ment spend ing including pop ulation , income, fed- eral aid, population growth, percent of population in rural areas, and other demographic and political characteristics. The results indicate that after con- trolling for these influences, growth in leg islature size in states during the period studied, in fact, con- tributed to higher government spending.

15- Thomas W. Gilligan and John G. Mabus.aka, " Fiocal Policy,LegislatureSize,andPoliticalParties: Evidencefrom State and Local G011emment in the First Half of the 20th Century.· National Tu.Journal LIV, 1 (March 2001): 57- 82

For example, two unre lated issues, s uch as import quotas for textiles and the fund- ing of a new bomber, might be included on th e same ballot. By doing so, legislators will have to vote for each of the issues together even if they would gain from the passage of only one issue. Each issue, if voted on separately, would be defeated because each alone provides benefits only to a minority of voters. However, the combined package might succeed in passing by a simple majority if each minority special -interest group votes for it to get its favored program. In effect, each special-interest group is induced to support the program of another special-interest group in order to receive benefits from its most-favored program.

Again, the willingness of each special-interest group to vote for the combined package is a function of the relative intensity of preference on the two issues. If the gains on th e most-favored issue are balanced equa lly by the losses resu lting from passage of the issue mos1 favored by the other group, no incentive exists to muster support for the combined package. Citizens will have incentives to engage

c_.., ,., ,.c..c..,u-. AU R;""'"""'«L "-',- ""_"_,,_,._-4. ., _., .,,..._ ""' "'"".....,"""'-,._....,_,,_...,, . .,__.,,_ "", .... _."' ""-"'l ·· .._..,,_...., _ __ do<,a - "' ""'""'""- . . C- U . ......,.,.. ·1,., ,._ --• ·.-...,..-.,.,,. . " ..

CHAPTER 5 Public Choice and the Political Process

in implicit logrolling only to the extent to which it provid es them with positive net benefits.

Many argue that logro lling is a positive safety va lve in a democratic society because it allows citi zens an opportunity to ex press their intensity of preference for particular outcomes in terms of their willingness to trade votes. A problem often overlooked in democratic societies is that allowing one vote per voter on each issue provides no direct basis for individuals to express their intensity of preference on iss ues. A vote is a vote. It says nothing about the extenr to which a ci tizen is made bener or worse off by a given political chan ge. Under simple majori ty rule, a coalition of individua ls ca n succeed in defeating issues about which certain minority groups have extremely strong feelings. If the minorities have no outlet for venting these intense preferences, the result might very well be socia l instability and eventua ll y a revolution to change political institutions. Thus, logrolling can be thought of as a safety valve.

Although logro llin g has the potenrial to account for intensity of preference, the danger remain s that it will be used by skillful politicians as a mea ns of gain in g approval for programs with purely redistribu ti ve benefits. In fact, the common judgment is unfavorable toward logrolling because of its repu tation as a mechanism for members of Congress to use the political system to gain benefits that accrue only to th eir constitue nts. Those who argue against logro llin g beli eve th at it exte nd s the size of the public sector over and above what wou ld be the case in its absence. Many believe that this extension is primaril y in programs th at redistr ibute in come to certain gro ups, rat her than providing positive net benefit.

Logrol ling and Efficiency Suppose citizens in a comm unity vote whether to support both security protection and comm unity entertainment. Both these goods are pure public goods for the citi- ze ns. Sec uri ty protection is measured by th e number of secu rity guards hired per week. Enrertainmenr is measured by the number of fireworks displays per week. Assume three voters agree to share equall y in the unit costs of these two goods. Assume each fireworks display costs SJ00 per week and the cost of a security guard is also S300 per week. The marginal and average cost of each of these public goods is $300. At first, these issues arc voted on separa tel y, and no logro lling agree- ments are mad e. Each voter is assigned a tax share of S I00 for eac h unit of the good. The collective decisions are made according to majority rule.

In Figure 5.8, the marginal benefit, tax per unit, and margina l cost of each of the goods are shown. Figure 5.8A assumes that the margin a l benefit received by voters A and B is zero at one fireworks display per week. The MBA and MBB curves intersect the horizontal axis at one display per yea r. However, voter C recei ves a marginal ben- efit of $250 from the first fireworks display. C is the only person whose tax docs not exceed the marginal benefit of the first display. She will cast the only favorable vote. Under majori ty rule, no fireworks displays will be provided because A and B vote against the fi rst unit, given their $100 tax per di spla y.

Figure 5.8B shows the marginal benefit, marginal cost, and tax per unit of secu- rity guards. In thi s case, voter A is the only one whose marginal benefit docs not fall short of the $100 tax per guard. The marginal benefit that A receives from the first security gu ard is $250. Voters B and C are pres umed to have zero marginal benefit at a weekly level of protection corresponding to one security guard per week. It follows that, given their $100 tax shares, the y vote no. No security protec- tion will be provided under majority rule. Gi ven the assumptions, neither secu ri ty

c_,.. ,.,,. c... u......_ ... _ .,.,,,.._...,,.,.,., __ ,.__,_.,_,..,,..._ ""' "'"".....,....,__,._...,_,.__,_, . ..,__.,,_ ._.,,..,._."'~'l · ... ......, ... .., ..,.,,.,..,., _ ...., a ..., .,,. ..... .,_ · .c,_u ·,.....,.,., ., ,..,,... -- • ._.,_ . .,., ,._ · ·

PART ONE The Economic Basis fo r Governm ent Activity

11@11!4 Logrolling

~30 MC i250 ··· ···~ M~ ii &10or-- ---,--- 1! MS, i !

Fireworks Displays per Week

8 jGo

50

i2so -

MC

MB, r-- ---,--- ,

Security Guards per Week

MC = MSC

2oo1---__c;;,-~ ,

M", MBA Mft

Security Guards and Fireworks Displays per Week • 1

~g::i:~~t:~y~ result in the passage of two issues together that could not pass if voted ! protection nor fireworks displays will be provided under majority rule. The median most -preferred outcome of the three voters is uro in both cases.

Now suppose rhat A and C collude to get firework s displays and security guards paired together as an issue. The election now consists of approva l of a budget that includes one fireworks display per wee k for each security guard hired per week. In Figure 5.8C, the rota! output of security and fireworks is plotted against the marginal cost, benefit, and tax shares. A budget of $600 is proposed. All voters understand that this budget consists of one fireworks display and one security guard per week. The ta x sha re for each voter is $200 fo r each combi ned secu rity guard and fireworks display per week. Voter 8, who evaluates the marginal benefit of both of these goods as zero, votes against the proposed budget. However, both voters A and C will vote

c_.., ,.,,. .,,.... u_ .... _,.,,,,, . ..._...,,.,. .,_ .,_,,_,._-4. ., _ ,..,,..._ ""' '" "°'.....,....,_ __ _, _..,, . .,__. _ .., ..... _.,..a. ••.• ,,,.1 ·· ... ..._...., .., __,,_ ..,.,a _ ..., ..,,.,. .. ,._ · ,,c_ u · ......,. .., .,.,,.,... __ • . _-..,,.,_ ,. . .

CHAPTER S Public Choice and the Political Process

in favor of this budget because the sum of the marginal benefits of fireworks and security protection exceeds their taxes. For voter A, the marginal benefit of the bud- get is $250, which consists entirely of 1he benefit received from security protec1ion. For voter C, the marginal benefi1 of the budget is also $250 and consists entirely of the marginal benefil received from fireworks displa ys. When they agree to pair the issues, both A and C arc better off than they otherwise would have been if the i~ucs had been voted on separately. Thus, pairing of the i~ues results in passage of both, while neither would pass if voted on alone on its own merits.

The same result would emerge if voters A and C agree to trade votes if each of the is sues were voted on separa1ely. That is, C wou ld agree to vote in favor of secu- rity protection, even though tha1 wou ld make her worse o ff, prov ided th at A ag reed to vote in favor of firework s displays. The net weekl y gain to A from agreeing to vote for security protection would be $50. This is the difference between the weekly margina l benefil of $250 from one security gua rd and the weekl y ta x bill o f $200 that wou ld finance both fireworks and sec urity. Simila rl y, voter C also would get a net weekly gain of $50 by agreeing to trade votes. The big loser in thi s process is voter B, who ends up paying $200 per week in taxes to pa y for both a security guard and a fireworks display , neither of which provides positive marginal benefit to him!

Notice how the res ult depends o n the intensity of preferences of voters A and C. Suppose that the margin al benefit of the first fireworks di splay was on ly S 125 for voter C. She would still vote in favor of it if it were proposed on its own merits, because her margin a l be nefit wou ld exceed her $100 ta x share. Howeve r, under these circumsta nces, she would have no incentive to engage in logrolling, because the su m of her marginal benefits for the first fireworks display and th e first sec ur ity guard would be only $125, assuming again that she receives 1.ero marginal benefit from one sec urity guard. Because thi s falls short of the $200 ta x share necessa ry to support both programs together, she wou ld vote against the combined budget of one fireworks display and one security guard per wee k.

logrolling also can cause lo~ in efficiency. Ar the political equilibrium budget of $600 per week, the marginal benefit received by A and C from both security protection and fireworks is $500. Because the marginal benefit from both of these public goods is zero for voter B, the sum of the marginal benefits for all three is onl y $500 in equilib- rium. This is less than the margi nal cost of $600 necessary to provide both goods. Because the marginal social cost of the budget exceeds its marginal social benefit by $100, approval of the budget mea ns that more than the efficient amount of public spen ding occurs.

Howeve r, logrolling does not a lways ca use losses in efficiency. For example, if B and A were eac h to get $25 in positive benefits from firework s di splays, then it would be efficient to have one displa y per year. But one displa y per yea r would not pass when voted on alone because both B and A still would vote against the pro- gram that would cost each of them S 100 per week in taxes for $25 per week in benefits. However, the sum of the marginal benefits in thi s case wou ld equal the $300 margin a l cost. Similarly, if both B and C were 10 get $25 per week in ma r- gina l benefits from security guards, it would be efficient to ha ve one guard per week beca use the sum of the marginal benefit of the guards would be $300, which equals the marginal cost. Pairing the two issues as before still wou ld resu lt in B vot- ing against the two goods, because hi s $200 tax share wou ld fall short of the $50 benefit he got fro m 1he two goods. Ho weve r, A and C would eac h get $275 in benefit s from passage and wou ld both vore in favo r o f it, given a $200 lax

c_.., ,.,,. C..C... U ....... All-• """'"'- W.,·- k--u--4.•-.. .,,..._ ""'.,"".....,""""--_,,_...,, . ..,__._..,,Bool_."' ""-"'' ·· ... ..._. .... _ _...,,_..,.. ___ .... ,....._ · _c_ u · ,.__,. .. .......... -- • ·-,.....-.,., ,. · "· ·

PART ONE The Economic Basis for Government Activity

share for each. The sum of the ma rginal benefits of the combined programs would be S275 + S275 + SS0 = $600, which equals the marginal cost of $600.

h is nol possible to reac h an unequivoca l co nclu si on about the impact of log- rolling on the efficient use of resources. In some cases, logrolling can allow improved efficiency by contributing to passage of programs with a marginal social benefit that equals or exceeds their marginal social costs. ln other cases, logrolling results in an overallocation of resources to government use by allowing passage of government programs with marginal social costs that exceed their marginal social benefits.

SPECIAL-INTEREST GROUPS AND THEIR IMPACT ON POLITICAL EQUILIBRIUM \ped;d-mtn~I ,:rou("- are organizations that seek to increase government expendi- tures that benefit their constituents. They differ ft-om political pa rties in th at their leaders do not actuall y run for politica l office. The y do, however, see k to put p res- sure on political candidates, bureaucrats, and ultimatel y on voters to support issues that benefit the members of their groups. Special-interest groups can apply pressure to politicians by threatening to tell their constituents to vote against them. They also ca n make campaign contributions to politicians who support their positions and finance advertisements against candidates who do not support their interests.

Specia l-interest groups ex ist to promote policies favorah le to work ers , particu- lar industries, regi ons, racia l min oriti es, et hnic groups, env ironm enta l preservation, and taxpaye rs in general. The Sierra Club, for example, ofte n acts as a specia l- interest group in trying to persuade environmental protection agencies to prese rve wild ern ess area s. In recent years, such groups as the Sierra Club and AARP, which is a special-interest group representing retirees, have become potent politica l forces. Man y of these spec ial -in terest groups ha ve used their power to influence the votes of those who are not members or direct beneficiaries of their efforts. In effect, special-interest groups in modern democracies often campa ign with as much zeal and actual expenditure of money and resources as do the political ca ndidates them- selves. The technical efficiency of the group in influencing the votes of nonmembers ca n make small groups, such as those representing farmers, or environmenta li sts, formidab le and powerful influ ences on voters a nd therefore on the poli tica l equilib- rium . Resea rch on the effectiveness of seve ra l special-interest groups in Swi tze rl a nd in achieving their objectives mea sures their influence and suggests that they have influ enced the political equilibrium in that nation. 11

Ga ry Becker has ana lyzed how specia l-interest groups affect politica l equilib- rium.12 Successful pressure groups succeed in manipulating ta xes, government subsi- dies, and government regulations to rai se the well-being of their members. Becker points out that pressure groups that succeed in obtaining increased benefits from gov- ernment must make the members of competing pressure groups worse off. This is because an increase in government spending for one group increases taxes, or im poses other costs, on other groups. Increased influence of an y one particular

"F. Schneide1 ind J. Naumann, "Inter.st Groups in O.mou.-ci .. -How Influential Are They7 Ari EmP,riul E~ami• nM.ionf0<SwiUe rl¥1d,"PublicChoice36(1982): 281 -303 11G..-y S. Becke,, " A Theory of Competition ....ong Pr-•e G, o,;ps {Of Political lnft.....,ce." Ouarterly Journal of E"conomics98(AugU5t 1983): 371-400.

c_.., ,.,,. c..c... u ....... ..u_,.,_,..._...,._..,_.,_,,_ .. _ -4. .,_.,.,,..._ ""' .,""....., __ ...,_,,_...,, . ..,__.,,_..,, .... _...,""-",1 · ... ..._....,..,_....,_,_...,a - ..,.,,.,...,._ -~-u · """"''"""I"",... --• ·-· ... ,,. · "· ·

CHAPTER 5 Public Choice and t he Political Process

group is also assumed to decrease the influence of competing groups. The pressure groups themselves compete fo r political influ ence by spending time , effort, and some of their income on the production of politica l pressure. Th e number of members in a grou p and the resources spent per member in su p!X)rting the gro up's pressu re activi- ties determine the amount of political pressure that a group can deliver. To the extent to which those who benefit from the special-interest group's efforts try to be free riders, the effectiveness in produ cing pressure declines. By spending money to reduce free ridin g, the lea ders of the groups can increa se press ure on po litica l agen ts.

Becker's model ca n be used to ga in some in sights into the success o r fa ilure of competing spec ia l-interesl groups. Successful special-interest groups are likely to be small relative to the portion of the population that pays taxes to support their sub• sidies. This result might see m surprising, but it is really very logical. After all, the greater the number of citizens who pay taxes to support even a rather large subsidy to a group wirh on ly a few members, the lower is th e la x per citi zen relati ve to the su hsid y per be neficiary. Becker points out that this resu lt is consiste nt with emp iri - cal obse rvation s. For example, agriculture tend s to be heavi ly subsi d ized in nations where it is a sma ll secto r, as is the case in the United States a nd in Japa n. Ag ricul• ture tend s to be heavil y taxed in nations where it is a large secto r, as in Po land and in many African natio ns.

People are ofte n members of more than one special-interest gro up. For exam- ple, a pe rso n co uld be a member o f an occ upational pressure group and of a group that supports regiona l growt h. This pe rson cou ld spend mo ney and effo rt as a member of hi s or her occu patio na l pressu re group in ways that result in increases in th e costs of o btaining the benefits desired fo r his o r her special intere st in regional growt h. In effect, the activities of various specia l-interest groups often resu lt in both taxes and subsid ies for their co nstituents. Ma ny peopl e co uld be equa lly as well o ff if both thei r taxes and subs idies fro m eac h pressure gro up were reduced in eq ua l a mou nts. If the effect of these taxes and subs idies is to ca ncel ca.c h other, members of the var ious special-interest groups wi ll not be harmed. But, because ta xes and subsidies to panicular ac tiviti es res ult in efficiency losses, the resu lt will be net gains from reducing taxes and subsidies.

BUREAUCRACY AND THE SUPPLY OF PUBUC OUTPUT Th e establi shment of a gove rnment a lso implies th e development of a b11rc-a1u:rac,.· that is in cha rge of imp lementing public choices made throu gh political in stitutions. The burea ucra cy itself influences the actual delive ry of services and the efficiency with which such se rvices ca n be produced. 13 In sofar as bureaucrats influence the cost of producing public goods, their behavio r is a n important dete rminant of the actual term s o n which such goods ca n be produced and offered to citizens. In other words, bureaucra ts influen ce 1he terms of suppl y o f pub lic goods and thus influ ence the resulting po lirica.l eq uilibrium.

A basic problem exists in measuring the efficiency of production by burea u- crats . In most cases, the output produced is neither easi ly quantifiable into units nor easily sold for profit in markets. Therefore, it is difficult to determine whether government bureaus produ ce thei r o utput at minimum possible cost. For a pri va te firm competing with other firms producing a similar output, such information is rapidly mad e available to owners through the firm's profit and loss statement. A business firm wi th costs of production that are higher than those of competing

"'See Thomas E. Borcherding, ed., Budgets Mid &,,.....,cr,.i,.. ([)um.om, N.C.: Oulte Unive<sity Press, 1977).

c_.., ,.,,.c... u......_ ...,_.,.,,,..._...., . .,. ,,. __ .. __,_.,_.,.,,..._ ""' '"""....., __ ....,_,,___,..,__.,,_._ ,...,._."'~'' · ... ......,..,. _ _,,_..,.,a - ...,.,,.,,.....,_ -~-u · ....,......., -,.,K>,._ -- • ·.-.,...-.,_,,.__,, · ·

PART ONE The Economic Basis for Government Activity

firms will quickly discover that it will be difficult to make a profit unless costs are lowered. Bureaucrats do not directly own the inputs they use for production. Fund- ing comes from an external source (s uch as Congress), and any net financial gains to bureaucrats who produce efficientl y are rare.

Bureaucratic Behavior Among the most significant contributions by economists who have analyzed bureaucra cy is the work of William Niskanen. 14 He ha s argued that bureaucrats see k to maximize the power associated with holding public office. Such power is likely to be correlated with the resources that the bureaucrat has under command. This, in turn, is related to the size of the burea u's budget. N iskanen therefore assumes that the burea ucrat seeks to maximize the size of the bureau's budget. Th e implication of Niskanen's analysis is that attempts by bureaucrats to ma ximize their budgets lead to a ge neral overex tension of the government secto r in excess of the efficient level of o utput.

figure 5.9 shows the bureaucratic incentive to supply more than the efficient amount of output. The marginal social benefit and marginal social cost of the burea u's output per yea r are shown in Figure 5.9A. The output co uld be the number o f new mi ssiles deployed per ye ar for a military burea u such as the air force. For a civili an agency, such as a bureau of public ro.1ds, the output could be me asured as the mil es of new road supplied per yea r. The efficient output, Q,. uni ts per yea r, cor- responds to point £, where the marginal soc ial benefit o f output just equals its mar- ginal social cost. Bureaucrats, however, seek to ma ximize the size of thei r budgets. Th ey see k to obtain as much funding as possible for their output. If they reason that th ey can obtain additional funds as long as th e total socia l benefit (TSR) of the output exceeds its rotal social cost (TSC) , they will try to increase output beyond the efficient level of Q• units per year. This is illustrated in Figure 5 .9 B, where the TSB and TSC of the bureau's output arc planed. The efficient output corresponds to the point where the slope of the TSC curve equals the slope of the TSB curve. At that output, MSB = MSC in Figure 5.9A. The output that the burea u will try to get approved is Q8 , which corresponds to the point where TSC = TSR. Th e burea u's desired level of o utput therefore exceeds the efficient amount.

Figure 5.9A shows the loss in well-being that result s if the bureaucra ts succeed in getting their desired level of output approved. The loss in net benefits is the tri- angular area EAB.

An additional problem with managing bureaucracies to achieve efficiency is that bureaucrats often ha ve monopoli stic power. Single agencies provide such ser- vices as environmenta l protection, defen se, and social in surance. In many cases, the bureaucra ts th emselves have specialized information not available elsewhere. The burea ucrats could seek to attain the budget-maximizing output by trying to make politicians believe that the social benefit of their output is greater than it is in fact. This wou ld shift the TSB curve up to the TSB ' curve in the eyes of funding agencies, as shown in Figure 5.98. The maximum output that the agency could fund would increa se to Q ' ll•

A funding agenr often ha s difficu lty monitoring the activities of its bureaus because of high transactions costs. Budgeting and manageri al improvements that

'"fo, mo,e d etails Ofl Willi.,m A. Nis.b,_,, Jr.'1 pioneering,,;.,_ on bureaucratic behavio, , 1ee hi, worb, 8<1reau- cr«y ..,,d Representative Govemmftll (Chicago: AJd ine-AthertOfl, 197 1) and -s.,reaucral• and Politicians," J.,.......,/ of L,w and Economics 18 (December 1975): 617--<>43.

~-"" "''' """'"" 11 ..... _ ,.,..,. ... .... , ... .. __ .. _ ...._ . - .. --- ""' "'""' . -- . ..,..-., .. __.,,_ ... ,Bool --.. ~·' · .,..,._,, .,. _ __,_..,.,a - "' °"'""'" """' ~~-11 · .....,.,. .., -,.,.,,._ --•-, .__, . .,_,,._ · ·

CHAPTER 5 Public Choice and the Politica l Process

hi!§!f ff- Bureaucracy and Efficiency

Q• QB 0utP,Ut per Year;

Q' QBQi, Output per Year

The efficient output is o• per year. A budget-max imizing bureau tries to get its sponsor to approve 0 8 units per year. This amount would result in a decrease in well-being equal to the area EAB. A bureau could try to get 0'8 approved if it convinced its sponsor that its benefits at any given output level were given by the curve TSB' instead of TSB.

lower th ese transac1ions costs wou ld contribute IO bcner moni1oring of bureaucrat behavior and would help achieve efficient output level s.

Bureaucrats ca n increase the size of their budgets in two ways. They can seek to convince governing authorities that their output needs to be increased, as in the pre• ceding analysis. Alternatively, they can increase the amounts of input necessary to produce any given amount of o utput by using inefficient production techniques. In this latter case, the loss in effic iency results from misuse of input rather than exces- sive production of the bureau's serv ice. Bureaucrats do not achieve efficiency because they maximize a utility function that depends not only on net benefit to their " sponsors" (fund ing authorities) but also on the growth of their budget, fringe benefits, job security, and reduced wo rkload.

The behavior of bureaucrats depends on the constraints that they face. Most of the models of bureaucracy presume that sponsors are at the mercy of the bureau- crats. In fact, however, adequate budgeting procedures can estab lish a set of

~- "" "''' "'-u . ·""-•"""'"'-" .. .,_,, __ .. _ ....._ . - .. -,..._ ""' '""" --- . .., .-. ... __.,,_._, .... __ "''°"'"""'' · .,....,_....,..,_,,...,,_..,, a - ..., .,,.,...,._ ~~- u · .....,.,..., -.,,.,... --• ·-,.....-..,,,,"""""" " · ·

m