The beast is not easily starved

New (2009) finds stronger support for Niskanen’s empirical case against the starve-the-beast hypothesis (and therefore a stronger violation of orthodox price theory) by making some adjustments to Niskanen’s regression analysis and getting a statistically negative relationship between government revenue and spending.

Obviously, voters do not have perfect information, and that fact can explain why they are not influenced by tax rates. But it is important to recognize that voter indifference to tax rates would exist even if voters have perfect information.

See http://​www.​foxnews.​com/​politics/​2012/​08/​04/​obama-brings-back-biden-pitch-suggests-tax-paying-patriotic/​.

Brennan and Buchanan in 1984 acknowlege that he (Buchanan) only hinted at the analysis of what became known as expressive voting. See Brennan and Buchanan (1984, footnote 7).

For a vote to be decisive the election would have to be a tie in its absence.

Neither does Tullock refer to Buchanan’s 1954 article, which is understandable since Buchanan doesn’t explicitly discuss the extremely low probability that one vote would count in the sense of being decisive. Downs (1957) is cited by Tullock, but on a point that is unrelated to expressive voting. Downs does make use of the indecisiveness of a vote to develop the concepts of the “rational ignorance” and “rational apathy” of voters, but does not apply that indecisiveness to explain the importance of expressive satisfaction from voting.

To my knowledge, it was Brennan and Lomasky (1993) who first used the term “expressive voting.”

See Brennan and Lomasky (1993, Table 4.1).

This is surely a higher probability of decisiveness than exists in most federal elections. For example, if there are 200,001 voters and the expected difference in probabilities between the yes and no votes is 1 % point either way, then the probability of a decisive vote is one in 12.3 million. Again, see Brennan and Lomasky (1993, Table 4.1). It should also be pointed out that if a vote is for a political representative, or even the president, who claims to favor a policy, his or her election does not guarantee passage of the policy. So the probability that a person’s vote is decisive in the election for a politician who favors a policy is larger than the probability that the vote will determine decisively that the favored policy is enacted.

One could object to this conclusion by pointing out that poor people are more likely to vote for political candidates who are most vocally in favor of transfers to the poor. But that is consistent with ideology being more important than the transfers. Before the 1930 s it was common for the poor to consider it a threat to their dignity to take transfers, even from private charities. As pointed out by Eberstadt (2012, p. 25) “[o]vercoming America’s historical cultural resistance to government entitlements has been a long and formidable endeavor.” This resistance has obviously been reduced in response to the efforts of government agencies dispensing transfers, and the poor seeing ever larger numbers of others, including the non-poor, taking transfers. And few things are easier to do than adjusting one’s ideology to favor policies that are personally beneficial. If you work for General Motors it is easy to believe that “what’s good for General Motors is good for America.” It should be pointed out, however, that ideological opposition to taking government help is far from extinguished. One measure of this opposition is that the poor vote for Republican candidates in higher percentages than one might think, given that the rhetoric of Republicans is less sympathetic for expanding anti-poverty programs than is the rhetoric of Democrats. See the article http://​www.​theguardian.​com/​world/​2012/​oct/​29/​working-class-voters-america-republican. (accessed April 15, 2015) It is clear from this article that poor voters commonly vote against their financial interests for ideological reasons. One example is from the following quote in the above link: “Many people say they are angry because the government is wasting money and giving money to people who do not deserve it. But more than that, they say they want to reduce the role of government in their own lives. They are frustrated that they need help, feel guilty for taking it and resent the government for providing it.”

Note that voting for programs that are seen to help others at a large personal cost is completely consistent with the self-interest assumed in a utility maximizing model. For a good discussion of this consistency see Brennan (2008).

This is good place to point out that voters are no less self-interested than are members of special interest groups. Consider the fact that in elections with a very small number of voters, one’s vote may be reasonably likely to be decisive, which can easily make the cost of voting against one’s financial interest much higher than the benefits from remaining faithful to one’s ideology. In the case of politically influential special-interest groups, bureaucratic agencies, and politicians, letting ideological expression trump their financial self-interest in how they use (“vote” with) their influence is very costly. So we would expect that the more influence a group has on political outcomes, the more likely they are to use that influence to advance their private interest at the expense of their ideological desire to promote the public interest. The differences in how voters vote and how special interests use (“vote with”) their influence is more a function of the different incentives they face than differences in their self-interests.

As evidenced by widespread believe that the poor, who pay little or no federal income tax, are motivated to vote for anti-poverty programs by financial advantage. If the poor thought that their individual votes were decisive in determining whether or not he or she received the government benefits being proposed, the voter turnout of the poor would be far higher than it is.

This is not saying that the value of the expressive satisfaction from voting is several thousand times greater than the value that is likely to be gained or lost as a result of the election. It is how much the voter weighs those values. In our example in this section, the weight applied to the value of expressive satisfaction is 1 which is 10,000 times greater than the weight applied to the value that will be gained or lost if the proposal receives a majority of the votes.

See http://​www.​cnbc.​com/​id/​100335218 (accessed April 15, 2015).

See http://​www.​reuters.​com/​article/​2013/​05/​30/​us-usa-fiscal-taxbreaks-idUSBRE94S1FS201​30530 (accessed April 15, 2015).

Revenue reductions are commonly defined by politicians as smaller increases in tax revenue than they had planned on.

Most workers are required to pay extra taxes for Social Security, Medicare and Medicaid, so they have prepaid for at least some of the benefits they receive from these programs. But in all three cases the amount paid in has for many years been consistently less than the amount paid out, with many recipients receiving significant positive transfers.

Sixty-five percent of the elderly (65 or older) receive over half their income from Social Security and 33 % receive over 90 % of their income from Social Security. And the median elderly recipient depends on Social Security for 67 % of his or her income. See http://​www.​epi.​org/​publication/​ib206/​. (accessed April 15, 2015)

Yet those American voters aged 66 to 83 gave slightly more support to Democrats at times in the 1990s, but have since strongly supported the Republicans. See http://​www.​people-press.​org/​2011/​11/​03/​the-generation-gap-and-the-2012-election-3/​ (accessed April 15, 2015).

Friedman (1998, p. 123) lamented his contribution in the early 1940 s to withholding federal income taxes by stating “it never occurred to me at the time that I was helping to develop machinery that could make possible a government that I would come to criticize severely as too large, too intrusive, too destructive of freedom. Yet, that was precisely what I was doing.” Tax withholding reduced the perceived financial cost of paying more taxes, which pushed out the limit on how much taxing and spending was considered acceptable by the public. Any reduction in the cost of voting for more spending operated through a change in some voters’ attitudes, or ideologies, which had started shifting noticeably in favor of larger government in the 1930 s and increased the expressive satisfaction many voters received from voting for more spending.

See Higgs (1987, pp. 83-85).

I have focused my discussion on spending and explicit taxation since those are the most relevant to the starving-the-beast debate. But it should be kept in mind that the increased influence and cost of government is not limited to how much government spends or explicitly taxes. Government has greatly increased the regulation of business and nonbusiness activity in recent decades, much of which has had the effect of benefiting organized industries and occupational groups that either lobbied for the regulation, or influenced how it was written and is enforced.

Some studies indicate that 96 % of Americans receive, or have received, government transfers. Most of these people have received these transfers for a relatively short period of time and end up paying more in taxes than their transfers are worth. See http://​www.​washingtonpost.​com/​blogs/​wonkblog/​wp/​2012/​09/​26/​do-96-percent-of-americans-receive-government-benefits/​.

The persistently poor, without the ability to help themselves, would not be a party to the “exchange” and would continue to receive more in benefits than the taxes they pay for the benefits of others. But we could reach a situation in which the vast majority could benefit from the “exchange” among themselves while still paying taxes to help the relatively few who are truly and persistently needy. This possibility is strengthened when we consider the large wealth losses being imposed by dead-weight losses from the taxes and transfers that would be reduced or eliminated.