Direct Democracy and the Effects of Voter Ignorance on Tax Policy Ballot Propositions

Section IV: Analysis and Interpretation of Data

One of the major problems with the polling data for both cases is that one does not know whether those people polled voted for or against the propositions. The responses cannot really be said to convey voter ignorance if they include the responses of some citizens who noted the interconnections between taxes and services.

Sears and Citrin’s data has the advantage of Proposition 13’s fairly large margin of support. Essentially, one can assume that because the proposition passed by a majority, the probability that some of those who voted for the proposition were included in the data showing increased demand is fairly high. The probability of this is much lower in the Oregon case, as the votes were very close.

However, even if we do not know the exact percentage of people who voted for the propositions included in the data, one can still use this data to suggest a logical inconsistency. Data from Oregon and California was collected after the ballot propositions passed. Logically speaking, tax cuts imply less government revenue. Less government revenue implies less government spending on public services. So one should assume, after making these connections, that the government cannot be expected to increase its services or possibly even maintain them. That the polling data reveals citizens’ high expectations for government services after the propositions passed suggests a failure to make these connections.

Whether or not one voted for the propositions is irrelevant in this respect. Indeed, it suggests that even a vote against tax cuts may still imply a logical inconsistency between taxes and services. Admittedly, whether such a vote is a logical inconsistency depends on the situation and an ability to make connections in situ. Whether one voted against tax cuts does not make an increase in demand automatically consistent with that vote. It depends on the situation and the ability to apply knowledge of the general relationship between taxes and services to that situation.

Advertisement

This certainly applies to the Oregon case. Even as cities and counties were losing funding and cutting services, 56% of citizens still demanded more services. This also applies to the California case, albeit in a slightly different way. California’s surplus mitigated the effects of revenue loss caused by Proposition 13. This may explain why people continued to demand increased expenditure for services even after cutting a significant source of funding for those services: they had not felt any negative consequences of Proposition 13 on levels of service.

Yet an ability to make logical connections between taxes and services may have provided a different outcome. First, people must be aware that there is a surplus. Then they must realize that the surplus will run out if the government heeds their demands for increased services. After all, the government’s ability to sustain the surplus depends on taxes (Perry and Sonstelie 1982, 125). Since property taxes are cut, the government’s ability is severely curtailed. For instance, the loss in property tax revenue depleted California’s surplus (Perry and Sonstelie 1982, 125). Moreover, because of the tax cuts, there is not much money outside of the surplus to fund the expenditures for services. Indeed, by 1981, California no longer had a surplus, but a one billion dollar deficit (Perry and Sonstelie 1982, 125).

On a methodological note, I realize that these cases are not perfect examples. For instance, I only used one data source in the Oregon case to convey public opinion. Clearly, one data source cannot unequivocally represent the opinion of all who voted in Oregon. However, this was the only empirical source of public opinion for Oregon I could find. Indeed, lack of such data limited many of my options. I had intended to study similar cases of tax cuts passed by ballot propositions, such as Idaho’s Proposition 1 (1978), Massachusetts’ Proposition 2 1/2 (1979), and Michigan’s Headlee Proposal (1978) but declined due to lack of empirical public opinion data (Sears and Citrin 1982, 227-9). I could have just explained these cases based purely on theory, by citing the number of votes in favor of the propositions and explaining that these votes imply a demand for increased services since, according to Converse, people generally tend not to make the connection between taxes and services. Indeed, the cases of California and Oregon are primarily intended to be theoretically true, but with some sort of empirical basis. That is, the evidence presented is not intended to generate a new theory, but to generally confirm Converse’s established theory of logical inconsistency. I believe that the amount evidence I used suffices in this regard.

Section V: Critique of Voter Ignorance and Ballot Propositions

Bowler and Donovan (1998) argue that voters may not be all that ignorant and that ballot propositions do not convey voter ignorance. Bowler and Donovan impute paradoxical outcomes (such as a demand for tax cuts and an increase in public services) to the attributes of ballot propositions and direct democracy, not to voter ignorance. Indeed, they believe that voters may be aware that cutting taxes means endangering the state’s ability to provide public services (Bowler and Donovan 1998, 257).

Many states have a single-subject rule, in which a given ballot proposition or referenda must encompass only one issue (Clark 1998, 467). This means that the issues on ballot propositions are presented to voters one at a time. This is supposed to clarify issues for voters and mitigate confusion (Clark 1998, 467). According to Bowler and Donovan (1998), this means that “tax limitation initiatives are rarely, if ever, linked to cuts in specific programs or the adoption of alternative revenue sources” (257). As a result, voters can only express their fiscal preferences, such as a tax cut, and not their preferences regarding from where money should be cut to provide for preferred services (Bowler and Donovan 1998, 258). Furthermore, “direct democracy provides no comparable, readily used mechanism for aggregating preferences about numerous decisions and tradeoffs that must be made about spending over hundreds of programs and agencies” (Bowler and Donovan 1998, 258).

In other words, voters are aware of tradeoffs involved in tax cuts and the provision of public services, and would like to somehow balance the two. Unfortunately, they are constrained by ballot propositions and direct democracy in general. The ballot propositions do not allow them to specify precisely from which programs funding should be cut in order to increase or maintain funding for other services. Direct democracy does not provide them with a means to express this preference (Bowler and Donovan 1998, 257). The paradox of tax cuts and increased services are due to systemic oddities, not logical contradictions.

Section VI: Rebuttal

Both California and Oregon employ the single subject rule, which enables us to apply the theory to the cases of Proposition 13 and Measure Five. Yet this theory cannot explain why the public opinion data shows an increase in demand for services after the tax cuts passed, given their effects. It seems odd that such “logically consistent” voters would make these demands when the state is suffering from a significant loss of revenue (as in Oregon) or when the state will suffer a significant loss of revenue after the surplus runs out (as in California). It implies that voters are not aware that state revenue (largely consisting of tax dollars) provides for these services.

One could argue that California citizens attempted to do what this theory posits: cut funding for specific services in order to continue state expenditures for other services. According to Figure 1, those polled supported decreasing expenditures for public housing and welfare after Proposition 13 passed.

However, this argument is not very strong. As people supported decreased expenditures for public housing and welfare, they supported increased expenditures for nearly every other program. The demanded decreased spending for public housing and welfare was not enough to offset the increased spending for the other programs. Moreover, by September 1980, people actually supported increasing expenditures for public housing and welfare, along with increasing expenditures for every other program (Sears and Citrin 1982, 48).

This theory is especially weak in light of the Oregon data. 6% decreased their expectations for the level of public services. One could assume that they decreased their expectations for certain public services, in order to redirect the expenditures for those services to other services they wished to see continued. Even if this assumption happened to be true, 56% still increased their expectations for public services. So although some citizens may have tried to negotiate the tradeoffs between tax cuts and continued services, they were a minority.

Section VII: Concluding Remarks

Advertisement

What good are ballot propositions that pass tax cuts, if this outcome is logically inconsistent? The outcome may be logically inconsistent, but it is also predictive. Citizen demand for increased services will most likely rise as state revenue goes down. The ability to predict citizen opinion in this regard could be useful for public officials or scholars interested in public opinion or voting behavior.

Should one even allow ballot propositions of this type, if the outcome is “ill-informed?” I believe one should. Even if the outcome proves to be logically inconsistent with the opinions of those who engendered it, this is not due to the ballot propositions themselves. Ballot propositions can be used to imply logical inconsistencies or, when combined with public opinion data, explicitly reveal them. But ballot propositions do not cause such inconsistencies. Eliminating one mechanism through which voters express their ill-informed preferences will not remedy voter ignorance, whether or not voter ignorance can even be remedied at all.

---

References

Bowler, Shaun and Todd Donovan. 1998. “An Overview of Direct Democracy in The American States,” “Responsive or Responsible Government?” in Citizens as Legislators. Ed. Shaun Bowler, Todd Donovan, and Caroline Tolbert (Columbus: Ohio State University Press).

Campbell, Angus, Philip Converse, Warren Miller, and Donald Stokes. 1980. The American Voter (Chicago: University of Chicago Press).

Caplan, Bryan. 2007. The Myth of the Rational Voter (Princeton University Press).

Clark, Sherman. 1998. “A Populist Critique of Direct Democracy,” Harvard Law Review 112 (2): 434-82.

Converse, Philip. 1964. “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent. 1964. Ed. David E. Apter (London: Macmillan Limited).

Cree, Nathan. 1892. Direct Legislation by the People (Chicago: A.C. McClurg and Company).

Oberholtzer, Ellis. 1900. The Referendum in America (New York: Charles Scribner’s Sons).

O’Toole, Daniel and Brian Stipak. 1998. “Coping with State Tax and Expenditure Limitation: The Oregon Experience,” State and Local Government Review 30 (1): 9-16.

Sears, David and Jack Citrin. 1982. Tax Revolt: Something For Nothing In California (Cambridge: Harvard University Press).

Shapiro, Perry and Jon Sonstelie. 1982. “Representative Voter or Bureaucratic Manipulation: An Examination of Public Finances in California Before and After Proposition 13,” Public Choice 39 (1): 113-42.

Somin, Ilya. 1998. “Voter Ignorance and the Democratic State,” Critical Review 12 (4): 413-47.