Fiscal policy correction

by on March 14, 2013 at 3:29 pm in Economics | Permalink

In my earlier post on fiscal policy, there is a mistake on point #1.  I failed to correctly distinguish between “total return” and “rate of return per annum.”  As the post read, it referred to the latter when it should have referred to the former.  Apologies!  Sometimes when I am traveling I don’t get to give posts enough rereads and am then more prone to errors; that is an explanation but not an excuse.

tomrus March 14, 2013 at 4:07 pm

Dear Prof. Cowen,
Could you asterisk your posts made when traveling so we will know to discount them?

Michael Collins March 14, 2013 at 4:16 pm

Would it have been so painful to say, Brad DeLong was right?

prior_approval March 15, 2013 at 1:17 am

Even when right, DeLong is wrong. He makes a Mercatus Center scholar look like a paragon of disinterested virtue, a mighty accomplishment.

Mike March 14, 2013 at 4:21 pm

It’s been my experience that the primary criterion by which one can differentiate an excuse from an explanation is whether or not either was requested. Explanations are usually given in response to requests; excuses are usually given without request.

Andrew' March 14, 2013 at 4:23 pm

The 20% seemed a bit high. The other thing is that the hurdle rate heuristics are relative. So comparing government projects versus comparing to private projects versus comparing peak-to-trough would all be different. In my prior life at a real company, we used to use a 1-year payback which seemed ridiculous to me. It was to encapsulate all the agency problems, vagaries, and uncertainties. Still ridiculous.

Steven Kopits March 14, 2013 at 4:49 pm

For a project will a

- 20% deadweight loss
- a two year construction period
- a 25 year life
- no residual value
- and a 5% cost of capital,

the hurdle rate is about 12%

Andrew' March 14, 2013 at 4:56 pm

I don’t see how that guts an argument against people claiming 4 or 5%.

Andrew' March 14, 2013 at 4:57 pm

(I get the part about Treasurys being cheap, I just don’t fully buy it.)

Andrew' March 14, 2013 at 5:34 pm

(As in, rates are low because the private sector is hurting, not truly because of a vote of confidence in government spending. So, that suggests a utilization of under-utilized resources only and NOT going gangbusters on projects that extracts more resources from the private sector. To do that would require (don’t we wish) competing with the private sector hurdle rates and THEIR cost of capital.)

Steven Kopits March 14, 2013 at 5:14 pm


Could you elaborate on the 4-5% comment? That’s much, much different than 12%.


Andrew' March 14, 2013 at 5:35 pm

This is what people such as Joe Stiglitz will talk about I suppose due to the Feds’ lower cost of capital.

Steven Kopits March 14, 2013 at 5:08 pm

The capital costs for a given project is the market rate for projects with comparable risk / return profiles, not the cost of borrowing for the government.

If a biotech companies requires a 25% irr to attract venture capital, then a government investment at a 6% irr is not a good investment, but rather a simple transfer of value from the government to the company–it’s a subsidy, not more, not less. If a government offers below market rates, it will take too much risk, as the example of Solydra shows.

Andrew' March 14, 2013 at 5:38 pm

I think you are saying what I’m trying to say but in a much less ignorant, talking out your arse kind of way.

Brian Donohue March 14, 2013 at 6:08 pm

Agreed. I also wish to affix my approval to everything Stephen Kopits says until further notice.

BobC March 14, 2013 at 9:15 pm

Sorry to be dense….I am just not getting this calculation…..can you be more explicit?….thanks!

dbeach March 15, 2013 at 1:20 pm

How do you figure? I get 7.06%.

I also don’t know why you’d use a 5% cost of capital when the government’s real rate is no more than 2% on 30-year treasuries, which would imply a (real) hurdle rate with the 20% deadweight loss of 3.8%.

MYB March 14, 2013 at 4:36 pm

Good, thank you! Unfortunately that guts your argument, because even with generous allowance, all your cowboy and DWL premia only add up to a few percent (at least with current Treasury rates).

The much more serious problem with current government spending is that so little of it is actually intended to create economic or social marginal benefit. Spending on the military and the elderly, the big pieces of the pie, aren’t subjected to that kind of analysis at all.

Andrew' March 14, 2013 at 4:45 pm

How does it gut the argument? What if interest rates signal no good projects? So then it is back to ‘unutilized resources’ and those apparently don’t actually get utilized. So, it might be a REALLY bad time to use resources.

Your second paragraph is an example of not doing the analysis, which could be one subset of doing it poorly.

MYB March 14, 2013 at 7:16 pm

The whole debate is the value of public investment. Tyler was astronomically, nearly order-of-magnitude wrong about something at the core of the debate and not even particularly complex. Even if he’s correct on all other points (doubtful), public investment is clearly a much better option than Tyler thought it was.

A lack of private sector investment seems the same thing as “no good projects” to me, and therefore not explanatory. The explanatory variables are crazy complex, just off the top of my head you’ve got risk premia, time preference, global savings behavior, federal reserve purchases, on, on, on. Also, by your very point that the resources utilized by the private and public sectors don’t perfectly overlap, you’re admitting that the lack of good private sector projects can’t be assumed to hold for public projects. Potential reasons include the government’s scale, unusual legal abilities (eg eminent domain), valuation of non-monetary benefits, ability to capture some positive externalities through taxation, and unusual risk and time-horizon characteristics.

You’re making too much of ‘unutilized resources’: resources can be un- or under-utilized, or even just under-valued, and there are resources besides the low-skilled unemployed (who, although they probably won’t become employed due to public investments, do accrue benefits). The recession caused an unusual number of projects to come in under budget, precisely because the private sector wasn’t as interested in competing for those inputs. If we want do something regardless of recession, like fixing our bridges, it makes sense to do it when the inputs are relatively cheap.

Luckily, a public investment analyst wouldn’t have to worry about all this interplay; the amazing thing about market-pricing (of resources, of government debt) is that it’ll help him decide what to do.

That we choose not to do cost-benefit analysis on most government spending is a problem of substantially different nature than Tyler’s incorrect assertion that government spending need meet some crazy IRR. This oversight in your thinking isn’t very promising…

Tyler Cowen March 14, 2013 at 9:13 pm

You can start with 7 percent or so on the risk premium…add on agency costs, taxes, it is still a significant hill to climb and not close to the near-zero rates of return which are cited so often.

Ritwik March 15, 2013 at 7:34 am

What risk premium is 7%? You’re saying beta of marginal government investment is 1? Why does the bond market disagree so violently with you?

Again, if you were to make the more limited argument that the bond market is demanding a bigger deficit, but that should come largely from tax cuts, you’d be right.

Tyler Cowen March 15, 2013 at 7:46 am

Come on, it is about the covariance of value for the *consumption* outputs, not the bond rate.

Ritwik March 14, 2013 at 9:08 pm

Very admirable!

Now, that’s why, point 2 was really the fulcrum of your critique. Governments very often destroy net wealth. So do private organizations. Private organizations have some absurd command and control mechanisms to check that. At my previous firm, cap-ex investments had to break even in 2 years (40%-50% hurdle rate) to be approved by the senior committee. Now admittedly, that’s too risk-averse. But technocratic challenges are the same.

Important to recognize, though, that your point 3 wasn’t that clear either. Is a long govt bond a risky asset or is it (one of the ) numeraire(s)? The implied Jensen premium from the markt movements in actually negative.

John Hatchett March 14, 2013 at 9:22 pm

Nice use of a semicolon.

Rich Berger March 15, 2013 at 9:09 am

I understood that the deadweight loss was a one-time haircut. Using “hurdle rate” was inconsistent but so what? The longer the length of the borrowing, the lower the impact of the deadweight loss. The bigger issue is that politicians do not (and do not want to) subject political “investments” to any sort of rate of return analysis.

Travis Allison March 15, 2013 at 11:06 am

I don’t think DeLong is right that the hurdle rate is changed by 20%/duration of debt.

Think of a perpetuity. If DeLong is right, than the hurdle rate is not affected at all no matter how much you pay.

If you finance the project with a perpetuity and your project earns a 10% return on invested dollar and because of taxes, the total cost of the project to the economy is 20% more than invested dollar, then the true return of the project is 10%/1.2 = 8.33%.

Comments on this entry are closed.

Previous post:

Next post: