Bryan Caplan  

Easy Answer All My Students Should Know

PRINT
Significant Links... Native American Economics...
Last week, I presented a question almost all my Public Choice students have trouble answering:

Suppose voters were rational [in the Rational Expectations sense] and the SIVH [Self-Interested Voter Hypothesis] were true.

T, F, and Explain: Democracies would spend a higher share of their budgets on genuine public goods.
Students usually claim that self-interested voting will lead to more redistribution and less public good production: Selfish people want free stuff, but don't care about society.  But that's pretty silly.  Key point: Redistribution is zero-sum, but genuine public goods are positive-sum.  As a result, there are generally far more selves who selfishly benefit from public goods than redistribution! 

And since we've assumed rationality, the electorate will, on average, recognize this difference.  Whatever you think about rational, selfish voters' desired level of government spending, their relative support for public goods over redistribution raises the budgetary share devoted to genuine public goods.  This result is even clearer if you realize that many alleged "public goods" aren't public goods at all - and that rational, selfish voters would, on average, realize this fact and support them less, further boosting the budget share of genuine public goods.

P.S. I'm on vacation for the next four weeks.  Expect light posting during that time.  Happy holidays to all!


Comments and Sharing






COMMENTS (22 to date)
ColoComment writes:

In support:

When I moved to Colorado in the late 1970s, I was surprised to learn of the concept of "open space."

Residents of Jefferson County, Colorado, voted to tax themselves for the purpose of buying and preserving open land from development. Apparently, they were the first in the country to initiate and fund such a "public good."

http://jeffco.us/open-space/about/history-jeffco-open-space/

Dylan writes:

"Key point: Redistribution is zero-sum, but genuine public goods are positive-sum."

Sure, but what's the distributional effects of that public good spending? And do the public goods have decreasing marginal return? Seems like you could end up spending more on an absolute basis on public goods, but less on a relative basis since I'm pretty sure tax collection would go way up if both were true.

Pajser writes:

Caplan: "Key point: Redistribution is zero-sum, but genuine public goods are positive-sum. As a result, there are generally far more selves who selfishly benefit from public goods than redistribution! "

If you claim that redistribution is zero-sum, it means you calculated taxes as costs. Then, why do you claim that genuine public goods are positive sum? Is it definition of "genuine public good?" Because public goods, generally, are not necessarily positive sum.

For instance, expensive program for search of cure for rare disease is negative sum; it will cure few people. Yet it is public good, because knowledge obtained on that way and published in public domain (as typically required for research programs supported by state) is non-excludable and non-rivalrous. Furthermore, selfish people wouldn't purchase cure for rare disease. Empathetic people maybe would.

Is it that I do not know something about "genuine public goods" that you and your students know? (I am pretty sure that Caplan will not answer, but maybe someone else would.)

Denver writes:

In order to know what a "genuine public good" is, I would have to spend a lot of time learning economics, and then wasting that hour or so going to the polls to vote.

If I'm rational and self-interested, I would only due this if my expected value for voting exceeds the costs. Which is unlikely in any modern public election.

So while I think you're right to the extent that public goods could be positive-sum, I don't think it follows that they are necessarily positive sum given the premise of your question.

Which is probably why this confuses so many people.

MikeP writes:

The economic definition of public good requires that it be positive sum. Otherwise it would be a public bad.

Pollution and smallpox are also non-excludable and non-rivalrous, but they are certainly not public goods.

Presumably the question uses the term "genuine public good" in order to reinforce that it intends the economic definition and not any definition like "a good provided publicly".

Ted Sanders writes:

This post was disappointingly uninformative. I wish there was more explanation and less assertion. I remain wholly unconvinced, personally.

Andrew_FL writes:

@MikeP-

The economic definition of public good requires that it be positive sum. Otherwise it would be a public bad.

"Good" in Public Good is a noun, specifically in sense 3. You are incorrectly interpreting its use in this context.

Philo writes:

"[T]here are generally far more selves who selfishly benefit from public goods than redistribution!" But so what? The rational, self-interested voters will get all the public goods that pass a cost-benefit test for the median voter, but is this less or more than voters in our system are getting? I do not see that economic theory provides an answer. And there will also be redistribution in the rational self-interested system to the extent that majority coalitions can be formed to exploit the minority. What extent is that; is it more or less than we have in our present system? Again, I do not see that economic theory provides an answer. In short, I found this post unconvincing.

MikeP writes:

"Good" in Public Good is a noun, specifically in sense 3. You are incorrectly interpreting its use in this context.

?

I am using 'good' as a noun. I am also using 'bad' as a noun.

How else would my sentences even parse?

James writes:

Caplan's reasoning seems wrong.

All economists agree that rational actors will spend more money on private goods than on public goods so I won't bother to spell out the case for that.

Now consider a modified scenario where everything is funded by the government and the costs people bear for things are not money, but time spent waiting in lines. For all of the same reasons that pertain to the case above, economists should agree that rational people will spend more waiting time on private goods than public goods.

Now consider a further modification in which everyone waits in the same line and when they reach the end of the queue, they mark a ballot expressing what they want. Are we to believe that the same people who prefer to spend a majority of their money on private goods also prefer to "spend" their votes so differently?

DMXRoid writes:

[Comment removed pending confirmation of email address. This is your final notice. We have attempted to contact you several times previously. Email the webmaster@econlib.org to request restoring this comment and your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Andrew_FL writes:

@MikeP-You are not using in in OED noun sense 3, that's for sure. You are interpreting it as something which has an antonym "bad" which OED noun sense 3 does not have.

Jeff writes:

I don't have any idea what "rational in the sense of Rational Expectations" means. Rational expectations is a modeling tool.

Suppose you have a model that can make some predictions of a set of variables Z given expectations of a subset X of Z. Given some expected X, you feed them in, crank through the model, and it predicts Z, which includes X and possibly some other stuff. You can then take the predicted X and feed them back into the model as expected X again. Crank, rinse and repeat. Eventually, you might stumble across a fixed point, wherein the X outcomes from one iteration turn out to be the same as the outcomes of the previous iteration. This is a fixed point, and it's also the definition of Rational Expectations.

That's all it means. Rational Expectations are the fixed point of a particular model. They have nothing to do with rationality in the usual sense of the quality or state of being reasonable. You might say that some other set of expectations is unreasonable because they aren't consistent with your model, but someone else could counter by saying that model-consistent expectations demand an unreasonable amount of information processing on the part of agents.

Or someone with a more philosophical bent could wonder if it's reasonable at all to postulate an economic model that "works" only if everyone being modeled believes the model's predictions. Why should they? If they don't, does the model become intractable? Maybe that's reason enough for the modeler to believe, but it's no reason at all why the agents in the model should do the same. But that's another discussion altogether.

BC writes:

Since this is a question from a Public Choice class, I would have thought that the answer would consider concentrated benefits and dispersed costs. Rational voters choose to be rationally ignorant on matters that only have small dispersed effects on them individually even if there are large effects in the aggregate. Rational voters prefer instead to focus their political energies on matters that have large concentrated effects on them individually, such as rent seeking and special benefits. Some public goods such as defense and police may have large effects even on individuals. However, other public goods may be positive sum in the aggregate but may have only small dispersed benefits for any given individual. Thus, it's not obvious to me that the share of budget spent on genuine public goods would increase if all voters were rational. We might instead see a proliferation of special-interest lobbying, where almost all voters ignore matters of general interest and focus exclusively on rent-seeking and obtaining special benefits.

On redistribution, we would expect to see a decline in 1%-type redistribution, i.e., taxing the very "wealthy" to redistribute wealth to everyone else. Those costs are concentrated since only 1% of the population would pay those taxes. The benefits would be spread across 99% so no individual would really get that much. The 1%, however, would have a lot of incentive to fight those taxes, as they would be quite large for each individual payer. The more likely type of redistribution would be the kind where we impose a small amount of tax on a lot of people to give individually large benefits to a small number of people, e.g., old-age benefits to seniors, education benefits to college students, subsidies to farmers, etc.

Hmm, this sounds a lot like our current state. Maybe, voters are not so irrational.

MikeP writes:

You are not using in in OED noun sense 3, that's for sure. You are interpreting it as something which has an antonym "bad" which OED noun sense 3 does not have.

I am using "public bad" as an antonym in one dimension to "public good" -- viz., a good in OED noun sense 3 that is bad.

That's how the Encyclopedia Britannica uses "public bad" too.

T Boyle writes:

"Democracies would spend a higher share of their budgets on public goods"...

...as compared to what?

baconbacon writes:

There is not enough information to answer the question.

Key point: Redistribution is zero-sum, but genuine public goods are positive-sum. As a result, there are generally far more selves who selfishly benefit from public goods than redistribution!

This is an assumption without proof. What happens if 99% of people decide to tax and redistribute the wealth of the 1%. This could be a net negative for the economy and still benefit the 99% at the expense of the 1%. As long as whatever threshold is needed in terms of % of votes is cleared then redistribution would occur. In such a world you would have constant referendum where people would try to identify groups that they were apart of and >50% of the population to receive taxes from out of the group.

For public goods the distribution of benefits is non linear. A project may heap large benefits on a small portion of the population and impose small costs on a large portion of the population (and be neutral of others).

J Mann writes:

T Boyle, Brian means as compared to the current situation. According to his class notes linked in the last post on this topic, Bryan believes that voting is primarily expressive rather than based on voter's interests.

In other words, since no individual vote is likely to matter, voters show up because they enjoy voting or feel guilty about not voting, and they vote to express something to themselves or others rather than based on a cold hearted analysis of what policies would benefit themselves personally. Of course, this is still rational and self-interested, but I take it that the Self-Interested-Voter-Hypothesis is that voters vote based on their interest in the policies at stake, not that they vote for whichever candidate they find most emotionally pleasant to vote for.

There were some good caveats in the last thread, in particular:

- Caplan's hypothetical rational self-interest voters might prefer policies that benefit VOTERS at the expense of non-voters, which wouldn't be true public goods.

- Caplan's hypothetical voters might be systematically mistaken.

Still, I imagine Caplan would still argue that voters voting for policies that met their self-interest would prefer more true public goods than voters who treat voting as an expressive act.

Andrew_FL writes:

That Encyclopedia Britannica makes the same error as you does not make it right.

Pajser writes:

MikeP -- definitions of "good" I saw do not take into account costs of production or purchasing of "good." If I buy expensive table, and experience show me that benefits were not worth the costs, the table is still "good." It is "bad" only if I do not benefit from it any more and cannot sell it, and it takes space in my room. Surely, individuals try not to do such purchases, and if they are "rational," they do not.

But democratically organized collectives are different. Both collectives of selfish individuals and collectives of empathetic individuals have their own, different reasons to purchase "negative sum" public good, such good that sum of individual costs are greater than sum of individual benefits.

MikeP writes:

That Encyclopedia Britannica makes the same error as you does not make it right.

It is not hard to find others using the concept of public bad. Indeed, you can find it in Byran Caplan's own Advanced Public Choice lecture notes:

A. What about national defense? Isn’t that clearly a public good?

B. Answer: It depends. “National defense” is not a public good for the world because if no country had “national defense,” no country would need it!

C. Implication: Countries’ “national defense” programs are often a public bad – and the losers typically include their own citizens.

D. Simplest reply to the national defense objection: Our country’s national defense is a public bad, and both we and the rest of the world would be safer without it.

1. This argument became far more convincing after the fall of the Soviet Union.

As evidenced by Pajser's examples, public bad is an extremely useful concept. Consider that, by construction, the universe of public bads outnumbers the universe of genuine public goods by an unmeasurable amount.

thaomas writes:

Is not living in a republic without huge differences in consumption levels a public good? It would require a good bit of distribution, but not self-interested distribution per se.

Comments for this entry have been closed
Return to top