Four years ago (and again two years ago) I argued it might be good policy for central banks to (conditionally) trash their own balance sheets. Now this idea is in the news. See Ralph Musgrave for links.
The recent blogosphere debate over whether money is or is not a bubble must have sounded like angels on pins to people who aren't monetary economists. But as JP Koning notes, it's what's at the root of those very policy-relevant questions. Does it matter if central banks trash their balance sheets by buying worthless junk? Might it even be a good policy, in some cases, for central banks to burn the bonds they own?
My views on this question have evolved a bit over the years. From reading other bloggers' posts, from comments here, and from thinking about it.
Here's a parable. Just in case it's not obvious, imagine I am a central bank, issuing currency, and targeting 2% inflation, so my currency pays a real rate of interest of minus 2% on average.
I could borrow M, invest the proceeds at a real rate of interest r, consume (r+m)M every year, and keep my stock of debt constant at M forever. Or I could borrow M, consume the whole of M immediately, and then consume an additional mM every year as my lenders pay me interest, and keep my stock of debt constant at M forever.
Would it matter to my lenders which I did? In the first case I have a stock of assets equal to M, which I keep in my basement. My IOUs trade at their fundamental value, because I can redeem them at any time. In the second case there are no assets in my basement. Would my lenders care? Why would any of them ever want to look in my basement? Or are the two cases observationally equivalent, as far as my lenders are concerned?There are two reasons why it might matter whether I keep a stock of assets in my basement:
1. Maybe the demand to hold my IOUs at a negative real interest rate varies over time. I can borrow M(t) at a negative real interest rate. If there is a risk that M(t) would drop by (say) 30%, would I need to keep 30% of M in my basement to be able to repay my creditors?
No. Because I can always borrow at positive interest rates against my future income stream mM(t). If M(t) drops by 30%, and stays there, the present value of my future income stream will still be 0.7mM/r. So I will only need to keep 0.3M-0.7mM/r in my basement.
If M(t) fluctuates, but on average grows at rate g, the expected present value of my future income stream mM(t) will be mM/(r-g). Even if (r-g) is (say) 1%, if m is (say) 2%, that will be 2M. The expected present value of my future income from being able to borrow at negative real rates will be twice as big as M. That means that I wouldn't need to keep any assets in my basement unless there was a risk that M(t) would drop by two thirds.
2. Maybe people are only willing to own my IOUs at a negative interest rate provided my IOUs are actually worth something. So there's a second equilibrium in which my IOUs are worthless, and so nobody wants to hold any, and so they are worthless. My total IOUs are either worth M, or zero. But by keeping a very small stock of assets in my basement (say 1% of M) I can knock out that second equilibrium. If my creditors all rush to redeem their IOUs at the same time, I just say to them: "OK, I will redeem them all at 1 cent on the dollar", which means they are worth something, which means thay are worth 100 cents on the dollar, because people want to hold M(t) and not 0.01M(t) of my IOUs. For any value above 0 cents and less than 100 cents on the dollar, there will be an excess demand for my IOUs (provided I don't issue more than M(t)), so their value must rise back up to 100 cents on the dollar.