5 The Basie Logic of Voting Introduction Ty onvER to plan its policies so as to gain votes, the goverment mnt disover some rltionhip between what does and how citizens vote. In our model the relationship is de- rived from the axiom that citizens act rationally in politics. This axiom implies that each citizen casts his vote for the party hie be. lieves wil provide him with more benefits than any other ‘Though this definition seems obvious, it is actually based upon concepts which are both complex and ambiguous. In this chapter eae carefully in order to show what “rational voting” I. UTILITY INCOME FROM GOVERNMENT ACTIVITIES ‘The benefits voters consider in making their decisions are streams of utility derived from government activity. Actually, this definition is circular, because we define utility as a measure of benefits in a citizen's mind which he uses to decide among alternative courses of action. Given several mutually exclusive alternatives, a rational man always takes the one which yields him the highest utility, 36 “THE nASIC LoGé OF VOTING 7 ceteris paribus; i.e, he acts to his own greatest benefit. This fol lows directly from the definition of rationality which is given in ‘Chapter 1. All citizens are constantly receiving streams of benefits from gow- emment activities, Their streets are policed, water purified, roads repaired, shores defended, garbage removed, weather forecast, cte, "These benefits are exactly like the benefits they receive from private economic activity and are identified as government-eaused only by their source. Of course, there are enormous qualitative differences between the benefits received, say, from national defense and from eating mince pie for dessert. But no matter how diverse, all benefits must be reduced to some common denominator for purposes of al- locating searce resources. This is equally true of benefits within the private. sector, The common denominator used in this process we call pity, Te is_possible for a citizen to receive utility from events that are only remotely connected to his own material income. For example, some’ citizens would regard their utility incomes as raised if the government increased taxes upon them in order to distribute free food to starving Chinese. There can be no simple identification of “acting for one's own greatest benefit” with sellishness in the narrow sense because self-denying charity is often a great source of benefits to oneself. Thus our model leaves room for altruism in spite of its basic reliance upon the self-interest axiom. Using this broad concept of utility, we can speak of a utility ineome from govemment activity. This income includes benefits which the recipient does not realize he is receiving. It also includes benefits he nous he is receiving but the exact source of which he does not know, For example, many citizens are probably not aware that the water they drink is inspected by a government agency. If inspection ‘were discontinued, they might not realize their utility incomes had fallen until they received polluted water. Even then, not all of them ‘would know that a cessation of government activity had caused this drop in income. ‘The fact that men can receive utility income from government actions without being aware of receiving it may seem to violate the ‘usual definition of income, Nevertheless, we must insist upon it, be-AN ECONOMIC THEORY OF DEMOCRACY. ‘cause an important political strategy of governments is making vot aware of benebts they are already receiving, However, only bonefts which voters become conscious of by election day can influence their voting decisions; otherwise their behavior would be irrational. I. THE LOGICAL STRUCTURE OF THE VOTING ACT ‘A, TERMINOLOGY OF ‘THE, ANALYSIS By defining income 2s 2 flow of benefits, we have involved out selves in time, since flows can only be measured as rates per unit of time. he wat of tine we use isthe election period. I is defined 2s ie time elapsing between elections, and it forms the prin of judgment in a voter's mind. one At Ieast two clection periods enter into a rational voter's : into s rational voter's calcila- tions: the one following the coming election, and the one ending on election day. We will refer to these periods ¢-+-1 and ¢ respectively. To illustrate the verbal analysis, we also employ bots as follows: . eee U stands foram individual voter's real or hypthetcl ti ol voter’ real or hypothetical utlty income 1 omgorerimen stn dong oe een pod isthe incumbent party, Le, the governing partyin period & Bs th ston py, the pay ot of poor pdt 1 part of te anata, we assume e two party Sytem) stands for utility income actually received eived during a period. It is the yr Billy Sncome provided by the party ia power using that period. stan for the aiity income which a voter believes the highest he could pasbly have received during some period. It is the utlity income which the ideal government would have provided him if it hhad been in power during that period. E stands for expected value. 1B. THE TWO PARTY DIFFERENTIALS Each citizen in our model votes for the the party he believes will pro- vide him with a higher utility income than any other party ding Jieoneeees ‘THE BASIC LOGIC OF VOTING 39 the coming election period To discover which party this is, he compares the utility incomes he believes he would receive were each party in office. In a two-party system, this comparison can be set up as a simple subtraction B(U$,,) — EUR) ‘The difference between these two expected utility incomes is the citizen's expected party differentia. If it is positive, he votes for the incumbents, if it is negative, he votes for the oppositions if it is zero, he abstains? ‘At first glance, rational voting thus appears to be a very simple matter. But its apparent ease is deceiving, for a crucial question re- mains: how should a rational voter calculate the expected utility in- comes from which he derives his expected party differential? Tt is in answering this question that we encounter dificulties. ‘When a man votes, he is helping to select the government which will govern him during the coming election period (ic, period t-/-1); Therefore as we have just shown, he makes his decision by comparing future performances he expects from the competing parties, But if he is rational, he knows that no party will be able to do everything that it says it will do, Ffence he cannot merely com- pare platforms; instead he must estimate in his own mind what the parties would actually do were they in power? ‘Since one of the competing parties is already in power, its per- formance in period t gives him the best possible idea of what it will do in the future, assuming its policies have some continuity.+ But + From sow on, the texm uty income refes specially to ailty income from government activity oles otherise noel ‘gfe acu the deisbu role foe mllptty systems late in this chapter. 2 The govesnng party our model Such broad power that perhaps it conid cary out al fi promis, Neverteles we assume here that it Gnnot for fe secon (1) tthe real world avd incur own ameertainty model, goverment Clmtforme te wake ty enone cy ht Eas esse pon the souctre of votes” thinking, aod (2) ina two pty system, each pasty WBibeetely makes ambiguous promises hence patfonn are poor harbingss ot ehons cren'n oor model The second pet is died in detail in Chapter 8 rhe tendancy of every enioal party to matin continuity ims polices discussed in Chopeer 7.AN ROONOMIC THEORY OF DEMOCRACY it would be irrational t ‘0 compare the current performance of on party with the expected fatute performance of another, For 2 valid omparion, both performances must take place under the same con- ee in the same tine petiod. Therefore the voter must weigh formance that the opposition party would have pr i Peto it had eon in power ns Me Prue in True, this performance is purely hypothetical; so he can only imagine what utility income he would have derived from it. But party B's future is hypothetical, too—as is that of party A. Thus he rust cithes compare (1) two hypothetical future utility incomes or (2) one actual present ulility income and one hypothetical present one. Without question, the latter comparison allows him to make nore direct use of eonerete fats than the former, Not only is one af its terms a real entity, but the other can be calculated in full cm butt calculated in full view of the situation from which it springs. If he compares future utility incomes, he enjoys neither of these advantages. Therefore, we be love it is more rational for him to ground his voting decisi corent evens than puey on fete ones Asa result, the most important 2 isi ‘ important part of a voter's decision is the si: OF his current party differential, te, the diference between the utility income he actually received in period t and the one he would have received if the opposition had been in power.’ i i entity is calculated as follows eee et (Uz) — E(Uz) It's the major determinant of his expected pasty diferentia __ However, this conclusion does not mean that citizens in our model ignore the future when deciding how to vote. Obviously, such an attitude would be irational, since the purpose of vating i to select a future government. Therefore the zational men in our model ap plies two futuzeorienting modifiers to hh 8 to his cureent party di in order to caleulte his expected pay diferente 4-7 avid conan, we opt the followin ty fret app witout the’ adgtive cunt weanetinc tem We dente etal eke see "ME BASIC LoGIe OF vorING 7 (¢. THE, TREND FACTOR AND PERFORMANCE RATINGS ‘The first of these modifies we call simply the trend factor. It is the adjustment each citizen makes in his current’ party differential to account for any relevant trend in events that occurs within the current election period. For example, let ns assume that a voter believes the present government made, many mistakes upon first taking office but has steadily improved and is now governing ex pettly. He may feel that this experiness will prevail throughout the next election period if the incumbents are retlected. Therefore he adjusts his current party differential to eliminate the impact of their initial blunders. Conversely, if he feels the government started out superbly but has continuously degenerated, le may project only its ‘bad performance into his expected party differential “The second modifier comes into play only when the citizen eannot see any difference between the two parties running; ie, when he ‘thinks they have identical platforms and current polices.* ‘To escape from this deadlock, he alters the basis of his decision to whether or not the incumbents have done as good a job of governing as did their predecessors in office. Our use of this particular tiebreaking device may seem rather arbitrary. Why should a rational man pay attention to the past in selecting a future government? Why should the present similarity Of parties canse him to drag past goveraments into his decisions? The answer to these questions is derived froun the impact of elec tions por se upon party behavior. In effect, every election is a judg: ment passed upon the record of the incumbent party. But the standards used to judge its record are of two types. When the oppo- sition’s policies in period ¢ have differed from those of the incum- Dents, the judgment expresses the voters’ choice between the future projections of these two policy sets. Bat if the opposition’s policies ‘© When perfect information exists, citizens think paris’ policies are identical only when they relly ae identical, But in a-wood Where men ace nok fully i foracd, some actual ferences between parties may escape notice because they Be pot sign enough fo cc vlc perepin tls, For afte explanation of this possbity, see Section LIT ofthis chapter.a AW HcoNOMEG THEORY oF pescoceAcy have been identical with those of the incumbents, Provides the voters with no real choice. In this ease, their judgment ‘expresses whether they sate the incumbents’ record as good or bad according ta some abstract standard. ‘Thus every clection is a signaling device as well as a govemment sclector. However, in a two-party system, itis limited to giving one of two signals. The incumbents always regard retlection #8 aman. date to continue their former policies. Conversely, the opposition party regards its triumph as a command to alter at least some of the incumbents’ policies; otherwise, why would people have voted for 3 In short, the outcome ealls for either “no change” or “change.” Hence it always makes a difference which party is elected, no matter how similar their records in period t, If the opposition wins, it is sure to carry out policies different from those the incumbents would have carried out had they been reelected, However, no one knows in advance just what policy changes the Opposition will make if it is elected. Nor can they be discovered by looking at the epposition’s hypothetical record in period ¢, since (we are here assuming) it is identical with that of the incumbents. But if men do not know what change signifies, how can they rationally «Vote for or against it? © Rational men are not interested in policies per se but in their own utility incomes. If their present utility incomes are very low in theit own eyes, they may believe that almost any change likely to be made will raise their incomes. In this case, it is rational for them to vote against the incumbents, ic, for change in general, On the other hand, men who are benefiting from the incumbents’ Policies may feel that change is likely to harm rather than help them, ‘Truc, the opposition might introduce new policies which would raise their utility incomes. But their incomes are so high already that they fear any break in the continuity of present policics. Hence they tationally vote for the incumbents, ie, against change in general. Clearly, both actions arc rational responses to the fact that elec- tions inevitably signal change or no change. They show that even when the parties running have identical records in period t, many Sitizens may reasonably expect different utility incomes from each mere projection ‘me nasic Locie oF voriNe 8 pasty in period ¢ 4-1. Thecefore abstention rational only a nges that will be made i -n believes that either (1) the policy changes ‘ the opposition s eleted lane no wet fest upon bis wii covne or (2) these changes may affect his income, but the robubiy ‘that they will raise it is exactly equal to the probability that they will lower it; i, the expected change is zero. oe ‘Two things are to be noted bout thls resoning, isl, we have admitted a degree of uncertainty into our certainty moda Hover the pape of this model isto prepare for snags ofthe uncertainty model; hence we feel justified in taking uncertainty int whenever it affects the basic structure of sain ee ae mid, we have argued that the incumbents’ record can be spd x tad cen shen i eal ith the rd oF he = positon. But what standard for judgment exits in this ease? With can the incumbents’ record be compared? : "i there wry men fen compar wat goverment ng with what it shonld be doing without seeing to soy thee pe Instead they are implicitly comparing the utility in he a tetualy reeving with those they would be receiving if the a wernment were in power. Of course, every man does not have ception oF the ideal government to asign a performance rating fo the incumbent patty or any other party.* Algebraicaly, itis comput as follows: UE [vs mee ratings are extremely useful for comparing gover coping nile ine ino ee nH et “They ae nec sons because absolute levels o ‘They are necessary for such comparisons if wily come Hom dflent tine pods cannot be compared ose isaac () fo "Ty cause ing lpia nt 30 4) Reena Mote ee ene a aie SESE toe tensa Ga de ae Sei tie yea ims igri pe acu oe a eaAN ECONOMIC THEORY OF DEMOCRACY rectly, a5 we saw earlier. The performance rating of a govemment may change for the following reasons: (1) it changes its actions while other conditions remain the same; (2) it keeps the same at. tions, and they give rise to the same utility as before, but other ein cumstances change so that the ideal utilityincome level alter; or (3) it keeps the same actions, but other eizcumstances change $0 that these actions mo longer produce the same utility incomes. 1 our model, performance ratings enter a voter's decisionmaking whenever he thinks both parties have the same platforms and eue fe oliies, At fist glance, this rule seems to imply discontinuity in the voters thinking, but in fact it does not. Every rational voter knows that if the opposition party is elected, it will alter some of the policies now being followed by the incumbents. But whenever the two parties have different platforms or current policies, he also Knows just what changes will be made. Therefore he ean choose between partes by deciding how he likes these specific changes, However, when he believes the two parties have identical plat. forms and current polices, he no longer knows what specise changes Will occur if the opposition wins. Therefore he is forced to base iis decision upon his attitude towards change in general. There is no shift im his method of deciding how to vote; rather a shift im the svidence available causes him to discard one tool and use anothes ‘The object of both tools is the same—to estimate the gain he will get from voting for one party instead of the other. ‘Thus voters use performance ratings only when their current party differentials ate zero and not always then. A man’s cunent party liferential ray be zero for two reasons: (1) both parties ‘have ‘tential policies and platforms; or (2) though theis policies and platforms are different, they produce identical utility incomes for him. Inthe latter case, penformance ratings are useless to him because be already knows what changes will take place if the opposition Wins. Sinee these changes do not alter his utility ineome, he abstaine, But in the former case he does not know what changes the oppos. towned change in general. We have alteady shown that (1) this attitude depends upon how good a job he thinks the incumbents 5 ‘THE BASIC LOGIC OF VOTING 4 ity income and (2) he can are doing in providing him with utility income a wie the acurdbent’pesformance against an idea pcormane. But by what standard does he evaluate, say, a rating of 40 percent as g Ping an wat es hoi il ces of past governments. In our model, each vot mnents. By computing their performance tinge ces 3 mee I with which he can discover whether the incumbents Sour dongs gard, bet or inferest jb of govern He tes for them if crating is good, aginst them is bad and not a Alf ts idfeent* Thus he may ratoly sign a non 250 vale to his expected party differential even when both parties have i cal records in period t. II. PRELIMINARY DIFFICULTIES CAUSED BY UNCERTAINTY puting their party dif 0 hove glibly spoken of voters computing : ttn nd prenan® ting on ining ew nee is current party differcntial computation is. In order to find his current party ia 2 ote ira top) stem mat do the lowing (1) eatin 2 phases of government action to find ont where the two partis ‘would behave differently, (2) ia Bae ae reuld his utility income, and (3) aggregat ‘ este stam ae whos he ey ol ethan the ott. Thi show rational vote: would bebave i Peau ot complete and cots information he same world in which dwell the rational consumer and the rational producer of tr tional economic theory. : int fe wae world, uncertainty and lack of information prevent tles, a yoter using preference ratings always votes if the suman bev dene good (or tad) but this nok tse bs on Cost inthe later cnt he oes or Benes) he expect for ange in ened must be lnge enogh to oatweigh the cost of voting though incumbents donot have an tniferen ating, For a more aches Rene Sintetin when voig say soe Copter Ut.AN ECONOMIC THEORY OF DEMOCRACY {he most intelligent and well-informed voter from behaving in pre cisely the fashion we have described. Since he cannot be certain what his present utility income from goverment is, or what it would be if an opposition party were in power, he ean only make estimates of both. He will base them upon those few areas of government ae, tivity where the difference between parties is great enough to im. press him. When the total difference in utility ows is large enough so that he is no longer indifferent about which party is in office, his party differential threshold has been czossed, Until then, he remains jnaiferent about which party is in power, even if one would give him a higher utility income than the other. The existence of threshe olds raises the probability that the expected patty differential will be 2600, ity that abstention will occur. It also makes it possible to change a voter's mind by providing him with better information about what is already happening to him, At this point, we encounter two major problems, open the door of our model to uncertainty, undesirables as errors, false information, this chapter we deal only with the basic opie of voting, we will post. ene consideration of these factors until Iter except for one proviso, Throughout this thesis, we assume that no false (ie, factually in- correct) information exists, though incomplete information can exist ‘Ths we exclude deliberate lies from our model, though envors and misleading data may remain, The second problem is rooted in the very concept of a voler’s changing his mind about how to vote, As we have shown, every voter Imukes bis voting decisions by comparing various rel and hypotheti cal sircams of utility income. To decide what impact each govern nent act has upon his income, he apprases it as good ot bad ia the ght of his own view of “the good society.” This procedure i me. tional because every citizen in our model views goverment as a ‘means to the achievement of the good society as he sees it Jans 4 mat’s ewaluation of each party depends ultimately upon (1) the information he has about its policies and (2) the tlauon between those of its policies he knows about and his conception of the good socicty. Once a voter has even provisionally decided hon ta st, when we wwe must also admit such ‘and ignorance, Because in 7 ‘THE aASIC Loic OF vormNG 4 mind only if one of these vote, he can be persuaded to change his two ‘aut are To simply the analyse asume that every eliza as a fixed conception of the god selety and as aad 10 his uowledge of pty pois fn a consent manne oct nly now infomation en posuae him change nen re assuming that citizens’ political tastes are fixed. in essence, we are assumain usta a Even though these tastes often change rally aie mg We eve out assumption is plansible inthe shor em, basing ware or praia upheavals. Infact Sed polis tests se a fad plausible to us than fixed consumption tastes, whic sumed in demand studics. IV. VARIATIONS IN MULTIPARTY SYSTEMS i of a two-party system, but its 1 snalyss as so far been ia terms of a two perty tis ae can easily be extended to 2 multiparty system, a 7 lates, veer follows the sme eles sin he formes, but compares the incumbent party with whichever’of the onposton ares hs the hit pecent pafrmasce ating i wold yield Bin st utility income if it were now in office. He poe Seer ae ein a twoparty sytem: a tational voler may a fora pty other hn the one he mot pt. For ange, wen rogresive Paty rn a candidate in the American President Uiatntaf 1918, some votes nko peered the Present tate to al oer eres voted for the Dente candidate Thy ido ese the ete vote cand no ehance il, and Je voted for him, the fever would Sa pean lest deste goup fom te Poste pit of Yiow-wonk win, Ths 2 woe fr thet forte condidate tenia frase the probability thatthe one they favored a od To avoid the later outcome they voted forthe et in the middle oftheir preference ordering a lea, thi i rational behavior, But it contradits our simple48 AN ECONOMIC THEORY OF DEMOCRACY Tule for how voters should act. This discrepancy demands an ex- planation. First we must point out that in our model, elections are evices for the selection of govemments, though they actually serve ‘many purposes besides this one. They can also be (1) means of creat. ing social solidarity, a5 they are in modern communist countries, (2) expressions of political preference, (3) devices for releasing personel ‘gression in legitimate channels (e.g, in political campaigns), and (4) incentives for citizens to inform themselves about current events, Nevertheless, we ate interested in elections solely as means of selecting governments, and we define rational behavior with that end in mind. A ational voter first decides what party he believes will benefit him ‘most; then he trics to estimate whether this party has any chance of Winning. He does this because his vote should be expended as part of a selection process, not as an expression of preference. Hence even if he prefers party A, he is “wasting” his vote on A if it has no hance of winning because very few other voters prefer it to B or C. ‘The relevant choice in this case is between B and C. Since a vote for A ' not useful in the actual process of selection, casting itis ima tional Thus an important part of the voting decision is predicting how other citizens will vote by estimating their preferences, Each etizen ‘ses hip forecast to determine whether the party he most prefers is really a part of the relevant range of choice. If he believes itis not, then rationality commands him to vote for some other party. In the absence of any information whatever about what other vot: es are likely to do, the rational voter always votes for the party he Prefers. He also docs so whenever the information he hus leads hina to believe his favorite party has a reasonable chance of winning, The Precise stochastic meaning of “reasonable” cannot be defined 4 Prion it depends upon the temperament of each voter. However, the less chance of winning he feels his favorite party has, the more Ukely he is to switch his vote to a party that has a good chance ‘The exact probability level at which he switches will partly de endl upon how important he thinks it is to keep the worst party from winning, For example, let us assuine that there are three par 4 IME BASIC LOGIC OF vorING # tes: Right, Cente, and Let. Voter X prefers Right to Center and Center to Lt, ut he Blivs hat Right bs the let chawcr of winning, If he greatly prefers Right to Center and is alm f ferent hettoon Centr end Left, ne is les likely to switch i sete from Right to Center than if he slightly prefers Right to Center but abhors Left. : lex when we consider is situation becomes even more compl : futsreionted voting, A voter tay sppott a. pert that toy “ hopeless in the Del that is support will emable 2 to goow aud someday become 4 lilly winser—tins giving him # wider age of section i the fur. Also, he may tenors support a Ropes ty as a warning to some other party to change its platform if betier choicealternatives in the future to present participation the selection of a government” Vv. SUMMARY less informa: a world where he is furnished with compete, cos tin, the mtonal einen mas his voting dcion in the folowing way: / y income from government ac- 1. By comparing the stream of utility income a thiy he has teeived ander the present government (adjusted or tends) with those streams he Bere e wold have eer the varions opposition prt had been n ofc the votes ins huis curzent party differentials. They establish his preference among, he competing partis. i 2. Ine tweparty tem, the voter then votes forthe pat hep ” Fecs. In a multiparty system, he estimates what he believes are the preferences of other voters; then he acts as follows: a. Ifhis favorite party seems to have a reasonable chance of wine ning, he votes fori ' b. If his favorite party seams to have almost no chance of win Kor ¢ more detailed dscusion of voting in multiparty systems, sce Chapters Band 9AN’ ECONOMIC THEORY OF DEMOCRACY ning, he votes for some other party that has a reasonable chance in order to keep the party he least favors from winning ©. If he is 2 futureoriented voter, he may vote for his favorite arly even if it seems to have almost no chance of winning in order to improve the altematives open to him in future élections. If the voter cannot establish a preference among parties because at Teast one opposition party is tied with the incumbents for first place in his preference ordering, he then acts as follows:!2 a If the parties are deadlocked even though they have differing platforms or current policies or both, he abstains b. If the parties are deadlocked because they have identical plat- forms and current policies, he compares the performance rat ing of the incumbent party with those of its predecessors in office. If the incumbents have done a godd job, he votes for them; if they have done a bad job, he votes against them; if their performance is neither good nor bad, he abstains, got cone by ny der a on Feet it Et it Toiets for whichever of these top For athlete topzanking pats he thinks has the bet chance oF wate ing. Far other considerations which might have a bearing upon hs decay, see Chapter 9 is sent and 4 The Basic Logic of Government Decision-Making Introduction Traprrional LLY economic theory assumes 7 the social function and private motive of government both a of ition of social utility or social welfare. Ss paieaal i 1) in our model, govern- from this view in Uuee ways: (I) i fe ee social fmetion is not identical with its private a we: ‘ify only the latter, which is the maximization of . Mead owt ower and (3) the green py cn ctng with ol cal of the governing apparatus. In ing with other parties for control of 7 us chapter we use the last two of these axioms to deser i ed Dasic principles of goverment decisionmaking in our model rmoeracy. st