How QE Could Backfire
This post reviews how quantitative easing is supposed to work. Then it shows how it might go wrong, and why the dangers associated with it going wrong are extreme.
The Fed's quantitative easing program is meant stimulate economic activity through several monetary policy channels. (See this important 1996 paper by economist Frederic Mishkin -- or this highly approachable lecture-notes version here -- and I am going to summarize even further.)
First, by expanding the balance sheet and purchasing assets, the Fed increases demand for assets and decreases their supply, with the net effect being that real interest rates fall. Low real interest rates increases the level of investment and thereby aggregate demand. That's the conventional textbook Keynesian interest-rates channel.
If that is the only channel you knew of...well, you're going to learn a few things today in this blog post. There exist a whole class of channels by which monetary policy affects asset prices and thereby increases economic activity. Let's briefly consider the three: (1) exchange rates, (2) equity prices, (3) real estate.
A reduction in real interest rates in dollar-denominated assets renders them unattractive relative to foreign assets whose yields have not been reduced, leading to capital outflow and an increase in net exports.
The reduction in real interest rates means that bond values have risen. Investors substitute the now relatively cheaper equities -- the Fed often talks about this as portfolio balance effects. As households have significant savings in bonds debt and equity assets, the increase in these two asset prices generates positive wealth effects on aggregate demand.
The third asset-price channel for monetary policy is in real estate. Low real interest rates decrease the opportunity cost of purchasing a home, place of business, or other real estate, and so demand for real estate rises, pushing up real estate values. Home equity represents a substantial fraction of household wealth, making this another important source of wealth effects which act upon AD.
There are also two more channels for monetary policy which work through credit. The first is that monetary policy's reduction in interest rates and expansion of deposits generates credit easing, increasing the supply of loanable funds -- and as the financed projects get underway, aggregate demand increases. The second occurs through balance sheets: monetary accommodation has a number of short-run positive effects on balance sheets, including a reduction of debt-finance costs, an increase in asset values and revenue, etc. Healthier corporate balance sheets make banks more comfortable lending to firms, and health in household balance sheets leads to more spending on consumer durables like cars and washing machines.
But in an environment of extremely low supply of safe assets and extremely high demand for safe assets, I see substantial risks to discretionary quantitative easing. (This post from the FT's "Alphaville" blog explains why safe asset supply has fallen -- a large fraction of mortgage-backed securities and sovereign debts have been kicked out of the "safe asset" category.) These risks offset on a net basis some of the stimulative effects of monetary policy, and they also directly dilute the potency of channels which operate through the financial system. Low interest rates will do no good if there are no willing lenders, nor will easing will buttress asset values if they are collapsing as they did amid the 2008 financial crisis.
My version of this story is that quantitative easing exacerbates both forces pushing up the value of safe assets, crowding out private demand for safe assets. When the Fed is holding roughly 11 percent of the federal debt in Treasuries and 9 percent of all mortgage debt, and when alternative sources of safe asset supply have greatly contracted, further contraction of safe asset supply or its diversion to the central bank may have material negative consequences on the stability of the financial system.
This view builds on an expanding literature which posits a vital role for safe assets as collateral and as a quasi-money, because safe assets are stores of value. (For starters, check out this paper by Ricardo Caballero, which lays out the basic theory.)
These risks are "tail risks," as a financial crisis is an extreme and non-normally-distributed event if there ever was one. If quantitative easing will not be conducted according to a rule -- my strong preference, obviously -- then I would strongly encourage the Fed to consider a few modifications to what appears to be the increasing likelihood of a "QE3" in the fourth quarter of this year.
(1) Stop buying safe assets. I do not know well the legal limitations on the Fed's ability to buy such assets -- the "Worthwhile Canadian Initiative" blog has a post about them -- but apparently the Fed does have the tools to purchase short-maturity state and municipal debt, foreign government debt, and (if it threads the needle on a particular statute) basically anything it wants if it is written as a loan contract in which the Fed is swapping money for the asset it wants.
(2) Buy safe assets, but then issue more. The Fed could further increase the money supply, given that money is acting as a safe asset in this macroeconomic environment. More unconventionally, the Fed or Treasury could initiate a new deposit program which creates a senior liability for the government -- selling a huge amount of this paper would instantly create a highly liquid source of safe assets.