An Unsatisfactory Platonic Dialogue on the Impacts of Quantitative Easing...
Thrasymakhos: So Jan Hatzius and company at Goldman Sachs think that the Federal Reserve should take the Federal Reserve's balance sheet up from $3T to $5T, and announce that it will take the balance sheet higher if necessary unless and until nominal GDP growth rises to 7% per year--and maintain that higher nominal GDP growth rate until total nominal spending gets back to its pre-2007 trend.
Adeimantos: But why should the nominal GDP growth rate rise? The Federal Reserve buys $2T of bonds and creates $2T of cash, but--as Robert Barro says--this is simply swapping one zero-yield government asset for another, and will have no effect on anything.
Glaukon: But Jan wants the Federal Reserve to buy not short-term debt but long-term debt--agency debt and long-term Treasuries and related securities with an average interest rate not of 0% but of 3%/year.
Kephalos: It seems to me that that is relatively little. Say half of that interest rate is an expectational effect--a belief that short-term interest rates will rise in the future and thus that long-term bonds should incorporate that expected rise in their interest rates. That leaves 1.5%/year as compensation for bearing risk. That means that $2T of quantitative easing means that the Federal Reserve takes onto its books and bears risk that the private market currently values as worth $30B a year to bear. That does not seem to me to be very much. Total GDP originating in Finance and Insurance is, after all, $1.2T. $30 B is only 2.5% of that "GDP originating" total...
Thrasymakhos: Is that small? The $1.2T of GDP originating in Finance and Insurance is a slippery beast. Real insurance, for one, should not be in there--that's $400B a year that insures indiviuals against fire and flood and accident and has nothing to do with providing the risk capital to finance economic activity. And of the remaining $800B, my guess is that 1/4 is the casino: Wall Street as Las Vegas charging people for buying and selling as they watch their stocks go up and down. 1/2 is the scarcity of capital: the fact that access to finance so that you can buy and build capital is valuable whether it is really risky or not. That leaves only $200B a year that is compensation for bearing risks. $30B is 15% of that.
Glaukon: So if $200B a year of risk-bearing capacity drives $1.2T a year of investment spending, then adding $30B/year to that risk-bearing capacity has the potential to drive an extra $180B a year of investment?
Adeimantos: Add a multiplier of 2 to that and you have an extra $360B a year of GDP--enough to push the unemployment rate down by maybe 1.5 percentage points. That gets us down from 9.1% to 7.6%.
Kephalos I don't think that this argument is well-grounded in either theory or the data. I know that I have never seen anybody seriously think about the relationship between economy-wide risk tolerance and aggregate investment spending, and that is what we would need to make these kinds of arguments...
