Greg Mankiw betrays his astonishing ignorance of monetary economics and the institutional structure of international finance. He thinks that the capital markets set US interest rates and determine the yield curve, so a shrinking trade deficit would reduce buyers of US Treasuries, driving up interest rates across the yield curve.
Their analysis of trade deficits, starting on page 18, boils down to the following: We know that GDP=C+I+G+NX. NX is negative (the trade deficit). Therefore, if we somehow renegotiate trade deals and make NX rise to zero, GDP goes up! They calculate this will bring in $1.74 trillion in tax revenue over a decade.
But of course you can't model an economy just using the national income accounts identity. Even a freshman at the end of ec 10 knows that trade deficits go hand in hand with capital inflows. So an end to the trade deficit means an end to the capital inflow, which would affect interest rates, which in turn influence consumption and investment.As Professor Mankiw observes, this is a freshman error, and is he the one making it! Apparently he cannot distinguish between a model with simplifying assumptions and the real world. The good professor is describing a the world as he would like it to be, not the way it actually is at present.
The Fed sets the interest rate and the yield curve is a projection of the interest and expectations about future Fed rate policy. There is never a lack of USD existing as settlement balances to purchase Treasury securities because the amount of Treasury securities offered is equal to the reserves injected into the settlement system by government spending. Treasury security issuance simply serves to drain the excess reserves created by government spending from the settlement system as reserve accounts at the Fed into Treasuries, which are transferable time deposits held at the Fed.*
Furthermore, the Fed has the capacity to manage the amount of settlement balances in the settlement system so that all transactions clear. When the Fed is not paying IOR and doesn't choose to set the rate to zero, then it sets its target and lets quantity float, by using open market operations, for example.
There is nothing wrong with Professor Mankiw's model as an economic model. However, it is not representational model of way the real world works. While it might have relevance as a teaching gadget, students would be given the wrong idea if they were lead to conclude that the world works like that.
Greg Mankiw's Blog
Greg Mankiw | Robert M. Beren Professor of Economics at Harvard University
* L. Randall Wray, Modern Money Theory: The Basics, at New Economic Perspectives