Steve Landsburg and I can have a perfectly civil discussion on theology. He agrees not to mock me (to my face) for thinking a guy can walk on water, and I don’t make fun of him for worshipping irrational numbers. But when it comes to government debt, I think Steve and I might just have to agree never to talk about the subject. Today he writes:
How high should taxes be? High enough to cover expected outlays going forward — but no higher.
That’s because any additional revenue would be used to pay down the federal debt, which is a bad idea. It was almost surely a mistake to run up this much debt in the first place, but now that we’ve got it, the best thing to do is to keep it forever.
Every $100 in outstanding debt commits the government to making payments with a present value of $100, and hence to collecting tax revenues with a present value of $100. In a world where the interest rate is 3%, the options include collecting (and paying off) $100 immediately, or $50 this year and $51.50 next year, or $11.38 a year for ten years running, or $3 a year forever. Because deadweight loss (i.e. the economic damage due to the disincentive effects of taxes) is roughly proportional to the square of the tax rate, it turns out that the latter — the policy of paying interest forever without ever making a principal payment — is (at least roughly) the policy that minimizes the present value of deadweight loss.
In other words, as far as the existing debt goes, we’re hosed no matter what we do — it’s painful to pay it off, and painful to keep paying interest forever. But at least if we pay interest forever, we’re minimizing the deadweight losses associated with raising taxes.
As usual with Steve, I think he is probably correct as far as the strict mathematics of the argument. However, I have the opposite intuition. Now the ground rules of this game (even though Steve didn’t say so) are presumably that we rule out government default on its debt as cheating, and we don’t worry about the ethics of taxation per se. In that framework, I would certainly recommend that the US government start paying down its debt, both in terms of outstanding Treasury securities but also in terms of paying people lump sums to renounce their future Social Security and Medicare claims.
Now the tricky thing is, I suspect if we spelled out our assumptions a little more, Steve and I would end up agreeing. I could even imagine after 5 emails, Steve saying, “Yeah, we’re basically saying the same thing.” And yet, it sure doesn’t seem like that, does it?
To see what I mean–that I think Steve and I are actually closer than it first appears–consider that of course I would NOT [added!--RPM] recommend raising taxes, to start paying down the debt. Rather, I would recommend slashing spending tremendously. Then the other subtlety is that you would want to allow for emergency situations where deficits occurred temporarily in the future. So by paying the debt down today (and as a general rule), you were actually just giving yourself breathing room so that the debt down the road didn’t end up higher than where you’re starting today.
In other words, Steve is saying that if the debt is 85% of GDP today, then in a century he wants it to still be 85% of GDP. (For sure that’s what he’s saying if GDP is constant; in the comments of his post it wasn’t clear if he actually wants to maintain constant dollar amount of debt, or constant debt/GDP ratio.)
Suppose I too want the debt/GDP ratio to be the same in 100 years. But, I predict that there will be periods between now and then when the deficits will be so large that the debt/GDP ratio will grow during those years. In order then to reach my target, there must be periods when debt/GDP shrinks. And unless I happen to think the present is a time that requires growing debt/GDP, why wouldn’t I start paying it down right now?
(This is a little awkward because presumably right now–coming out of the financial crisis–is an “emergency” time when tax revenues are depressed and you could expect the government to run a budget deficit. But, that’s not really important for Steve’s argument. He would have written the same post in 1984 I believe.)
Anyway, the above is just to get your wheels turning. The real problem with Steve’s analysis–by which I mean, the assumption that makes his conclusion right, but which is a bad assumption to make–is this:
Note: The above assumes that spending policies and tax policies can be set separately (subject to the constraint that tax revenues are at least high enough to cover interest payments). Things of course get more complicated if you believe that debt levels and/or tax rates constrain future spending through some political mechanism. I’ve often heard this alleged, but I’m unaware of any strong evidence for this assertion.
How about this Steve?
(The footage is from a riot in Greece earlier this month.)