Bryan Caplan  

Name that Blip Redux

PRINT
Two New Books... Means-Testing is Awesome...
Tom Ault makes an interesting point in the comments:
I'd just like to point out that some commenters are looking at a graph of total government spending (the sum of federal, state and local spending) and drawing correlations based on which party controls one or two branches of the federal government. A quick look at graphs of federal, state and local spending shows federal spending has hovered around 20% of GDP since the 50s, while state and local expenditures have risen steadily over the same period.
At first, Tom's "quick look" at total federal spending looks dead-on:

govspend2.jpg

Note, however, that there's pretty big Korean War spike in the early fifties.  From the end of the Korean War to the early Reagan years, the upward trend isn't as steep as it is for total spending, but it's still there.  But it's the second part of Tom's comment that activates my spidey sense:
I don't think it is fair to hold either party at the federal level responsible for what the states do -- at least not without a more sophisticated analysis of the impact of federal legislation on state expenditures, but certainly not on a naive correlation drawn from a simple graph.
The amazing thing about state and local spending in the post-war era is how heavily it relies on federal grants.  At the end of WWII, it was about .5% of GDP.  Now it's about 3.5% of GDP.

govspend3.jpg

Most economists probably doubt that these intergovernmental transfers really increase the size of government much.  After all, can't states and localities respond to federal grants by spending less of their own money and cutting taxes?  But the naive analysis is probably closer to the truth.  The literature on the so-called flypaper effect confirms that federal money really does stick where it hits.  When I worked with this data, in fact, I generally found that a dollar of grants increased state spending by more than a dollar.  (Unfortunately, my data lumped together matching and non-matching grants, so it's not clear quite what the mechanism is).

Tom raises some important caveats, but focusing on the feds ultimately makes a lot of sense.  Even though I'm more skeptical of Tiebout competition than almost any economist I know, I still think that without federal subsidies, tax competition between states and localities would have kept their governments a lot smaller than the ones we see today.  And I still can't figure out whether the current spending spike is a blip, or just a return to long-run trend. 

Any further thoughts?


Comments and Sharing





COMMENTS (8 to date)
Chris writes:

There has also been a significant trend of tying regulation to federal funding as well as matching funding for some programs.

Think Medicaid, NCLB and 55MPH speed limits. All have state and federal spending tied together.

Wasn't there also provisions in the recent stimulus package forcing state spending and prohibiting states from rolling back the new spending?

Justin Ross writes:

...I'm more skeptical of Tiebout competition than almost any economist I know...

Your skepticism of Tiebout moving-as-revealed-preference mechanism seems to be that the ability to move is mitigated by capitalization, forcing us to rely once again on voting.

It is worth noting, however, that this same mechanism makes voter consumption of irrationality very costly. In your book, you make the point that you believe people consume irrational preferences when it is cheap, and voting makes it basically free. Since most households have a significant percentage of their wealth tied up in their homes, hits to their housing value would represent fairly costly consumption of irrational policies. I think this would be a fairly satisfying "Caplanian" answer to why rent control is popular in polls yet rare in practice, especially rare in municipalities where homeowners dominate the politics.

Agree or no? I think if you agree, you have to be somewhat pleased that this Tiebout mechanism makes rent control more scarce than it otherwise would be.

kebko writes:

I propose that Amendment 17 is the worst amendment of the constitution. By changing the Senate from state-elected to voter-elected, the amendment basically removed the states as a balance of power in Washington. The amendment was ratified in 1913, and the second graph here looks like it starts moving within a few elections of that change.

Now, instead of a Senate that votes down any attempts by Washington to control the state houses, we have state houses who can vote to waste $100 million local dollars so that a $1 billion federal grant brings money to local contractors to build something useless.

Vacslav Glukhov writes:

What strikes me in this image http://www.life.com/image/91241743 is first - their happiness, good mood, confidence that - yes, they can build it, and that no, there's nothing we, the disorganized class of private industry employees, can do about it.

And second: the fraction of overweight people among public employees. One of them even eats in the course of the rally - for more public funds.

Snorri Godhi writes:

Good point about the Korean War spike, and I am happy that removing that spike confirms Truman as the most fiscally conservative President ever. (At least in terms of repaying the national debt.)

Taking out that spike, the federal spending increase from end of ww2 to 1980 is about 8% of GDP, about the same as the increase in state spending over the same period. State spending started from a lower base, though, so its increase is relatively bigger (and looks bigger on the charts).

The increase in federal-to-state transfers, from ww2 to 1980, is only 2.5% of GDP: the increase in state spending can be blamed only partially on such transfers.

I realize that all of this does nothing to answer the question about the current spike, but I think that such a question can only be answered with hindsight.

Chris Koresko writes:

Bryan Caplan writes: "When I worked with this data, in fact, I generally found that a dollar of grants increased state spending by more than a dollar."

This suggests that Snorri Godhi's conclusion that "the increase in state spending can be blamed only partially on such transfers" needs to be treated with caution.

I suspect the mechanism is the one that Chris described above, i.e., that federal funding to the states comes with attached strings that require the states increase their own spending. I saw a Wall Street Journal article a few weeks ago which said that the stimulus law is so full of those strings that a lot of states are wishing they'd joined the handful which refused to take the money.

blink writes:

After all, can't states and localities respond to federal grants by spending less of their own money and cutting taxes?

Probably less often than it seems at first glance The grants may be tied to specific projects/objectives that (almost by definition) the State was not already funding. It might be better to think of some "State" programs as really federal programs that are administered locally.

Jim Glass writes:

When I worked with this data, in fact, I generally found that a dollar of grants increased state spending by more than a dollar.

Look no further than the horror show that is Medicaid.

In NYS Medicaid financing is split three ways -- federal, state, local.

Thus every local-level politician figures "Every mere $1 that I finance gives me $3 of spending I can take credit for -- great deal for me!!" And the state-level politicians figure the same.

Conversely, when confronted with the idea of cutting waste they recoil with: "Yuch, I'd have to take $3 away from my voters to get credit for saving only $1 of taxes, no!, no!, no!".

Not surprisingly, NYS is one of the worst of all the Medicaid disaster states.

E.g. regarding fraud prevention...

New York cut the number of health-department staffers combating Medicaid scams from 200 in the late eighties to 50 by 2005, even though the state’s program (including federal, state, and local expenditures) has grown by $30 billion since then. [City Journal]

Medicaid fraud in NYS is estimated as high as 40% of all expenditures. Not waste, fraud. (40% is an estimate for California too.)

Comments for this entry have been closed
Return to top