>would only raise interest rates by 1.7 percentage points

Presumably, such estimates are necessarily contingent on the supply elasticity of foreign capital inflows into the American capital market from abroad. If so, what is the supply elasticity - and what likely consequences will there be for foreign capital markets from capital exports to America?

On what evidence is it linear? There is none.

Do expanding deficits though scare off foreign capital? I thought long-term rates, like mortgages, fell in the 90s because foreign capital liked the fiscal responsibility under Clinton. Will exploding deficits stem this capital flow?

I agree with Daoud at least up to a point on this. My thought is that its got to be non-linear: very small deficits don't have any impact on interest rates at all. In fact, one could argue there are very important kink points (threshhold effects) in the relationship between deficits and long-term rates.

But what I DO find interesting is that Hubbard seems so much surer than other economists about where the non-linearities are. He seems to firmly believe either that the kink point is far enough out that we don't have to worry about it or that there is no kink point at all. Instead of presenting his informed opinion about where the kinks are, he is pretty confident in assuring as that the relationship is linear and small as 'far as the eye can see' in Mitch Daniels' words. It seems that the more accurate description of the state of knowledge among other economists is that there are 'all else held equal' difficulties in empirically finding out what the relationship between deficits and interest rates really is.

What scares me most is the seeming patchwork of ideologies that hold the plan together. For instance, I would like a supporter of Bush to help me find answers to the following questions.

1. Deficits in the hundreds of billions of don't matter but we need to hold down the cost of those school lunch programs. Why? Just borrow the damn money it is a piss in the ocean.

2. Deficits don't matter but we need to ensure that government does not grow faster than the rate of growth of personal income. Why? Under Rubinomics, spending grew way slower than consumer's incomes but that was apparently not a well-founded plan. So why do we need to control spending, if the money can be borrowed at a ridiculously low growth cost of 0.1 percentage points of increased interest rates per $200 billion?

3. We need to have policies that stimulate private savings because more savings presumably stimulates growth. But according to the powers that be the level of government dissavings is almost irrelevant for growth. So does this mean that to stimulate growth by a full percentage point, we need to get the U.S. public to increase savings by $1 trillion? Or is the calculus somehow different on the positive impact of savings on growth as compared to the negative impact of deficits on growth?

4. And is the calculus that
nat'l savings = private savings + public savings
just wrong?
Is it really
nat'l savings= priv. savings + 0.1(pub. savings)
Otherwise how do we explain the weight given to increasing private savings but not to gov't savings?

The result is that when I try to reconcile these inconsistencies, I come away with the belief that Hubbard like the rest of us believes that savings matter, and that long-term structural deficits have important costs. Why he then pursues in acts that are clearly designed to put the long-term budget health in jeopardy makes me fear about intellectual bankruptcy.

Surely the deficit is not $3.5 trillion, it's about $300bn isn't it? Do you mean the stock of debt?

The Orszag and Gale article is very enlightening"

"The magnitude of the effect on interest rates from the studies that incorporate projected
deficits is generally consistent with the results from the structural macro-econometric models. For
example, Elmendorf (1993) finds that an increase in the projected deficit of one percent of GNP
raises five-year bond yields by 43 basis points; Canzoneri, Cumby, and Diba (2002) find that an
increase in the projected CBO deficit averaging one percent of current GDP is associated with an
increase in the long-term interest rate relative to the short-term interest rate of 53 to 60 basis
points; and Cohen and Garnier (1991) find that an increase in the projected OMB deficit of one
percent of GNP raises the 10-year interest rate by 53 to 56 basis points. Note that these effects
represent the short-term impact on long-term interest rates, not the ultimate impact on long-term interest rates. The range – an increase of roughly 40 to 60 basis points in long-term interest rates from an expansion in the projected deficit of one percent of national income – should thus be compared to the results from the structural macro-econometric models for the effect on long-term
interest rates after one year. Both approaches suggest effects in the same broad range"

Therefore, a jump of 0.8% is more likely for a deficit of 200B. Deficit apologists argue that the effect on short term interest rates is compensated for by growth. But consider: What's a reasonable annual return on a typical investment? Is it not 5% or 10%? Therefore (in an excessively simplified form), the deficit can be thought of as a Net Present Value series, something like this:

200 = p/(1+n) + p/(1+n)**2 + ...

where p = the return on investment and n = inflation rate (this is government, so all we want is breakeven). If the return on investment is as little as 4-6 billion per year, it pays itself off.

But of course, this is oversimplified. Let's assume that interest rates rise by 0.8%. The additional interest cost to the US Treasury *alone* is on the order of $50B. Therefore, the real return on investment should be $4 + $50 B. And then there's the effect on growth, which further depresses revenues. A SWAG therefore suggests that deficit spending of $200B should be guaranteed a return of no less than $54 B and probably more like $75B. Annually. Forever.

Not many investments meet these criteria.

The Keynesian approach would involve much smaller sums. For example, paying 1 year of unemployment insurance to each of 2 million unemployed would add less than $20B to the deficit. That corresponds to an 0.08% increase in interest rates, or an additional $5B charge on the Treasury. This is doable. And there is a positive effect on growth as people spend their benefits and generate jobs and tax revenue. The same cannot be said of tax cuts going primarily to the wealthy, who are just as likely to invest them in China, where they may serve to decrease American unemployment.

From even these pathetically oversimplified back of the envelope calculations, it's clear that the current deficit spending is irresponsible.

James asked: "Surely the deficit is not $3.5 trillion, it's about $300bn isn't it? Do you mean the stock of debt?"

No, Prof DeLong meant 3.5 billion. Hubbard said that 200 billion imposes only a rise in interest rates bu 0.1 percentage points. Therefore, seventeen and a half times bigger (17.5 * 200 billion is 3.5 trillion) should cause an increase of 1.75 percentage points (17.5 * 0.1), if the relationship is linear. But that would seem to be nonsense, so the relationship is either very nonlinear or 0.1 is way too low an increase arising from a 200 billion deficit.