Thursday, November 18, 2010

Quantitative Easing and Fed Insolvency

I was reading Robert Murphy's criticism of quantitative easing.

One of the issues regarding quantitative easing is that it is risky. By purchasing longer term bonds, the Federal Reserve is subject to interest rate risk. If interest rates rise, say, due to higher expected inflation, then the value of the Fed's asset portfolio will fall. If it falls below the value of its liabilities--the currency and reserves it has issued--then it will become insolvent.

Now, Murphy doesn't focus on the possible insolvency of the Fed directly, but rather notes that if the Fed needs to reduce the quantity of base money through open market sales, it is the market value of each asset that determines how much base money is destroyed by the sale. If the total market value of all the Fed's assets is less than needed decrease in the quantity of base money, then the Fed cannot use open market sales to reduce the quantity of base money enough.

For example, suppose the Fed undertakes quantitative easing and base money rises to $2.6 trillion. Money expenditures begin to rise, resulting in higher prices (and, hopefully, higher production.) The rising prices lead people to expect higher inflation. Long term interest rates rise. Some of the assets the Fed currently holds, which includes mortgage backed securities and longer term bonds, fall in current market value because of the increase in interest rates.. If the Fed wants to get base money all the way back to $800 billion, then it will need to sell approximately $1.8 trillion worth of assets. If the market value of its asset portfolio falls more than 31 percent, then the Fed will be out of assets to sell, and base money will still be greater than $800 billion.

As we imagine the Fed following this policy and reducing base money, the expected inflation should decrease, raising the market value of the bonds. While, normally, selling off these assets would tend to lower their prices because of the liquidity effect, remember, that the reason base money needs to drop is that banks are strongly expanding lending, perhaps by making commercial loans, but also by purchasing bonds.

Presumably, the sensible strategy for the Fed would be to sell off all of the bonds it holds with short terms to maturity, and hold off on selling the long term bonds. But it is possible that it would need to sell off long term bonds and mortgage backed securities at a loss. And if the losses are great enough, the Fed might become insolvent, and further, the insolvency could become so great that it would not have enough assets to reduce the quantity of base money enough to keep money expenditures from rising too much, resulting in above target inflation.

Murphy does make an odd error when discussing the possible strategy of the Fed paying higher interest on reserves. He says that investors will figure out that an exponential increase in reserves will hardly allow the Fed to control inflation. Now, suppose the expected inflation rate does rise to 10 percent and that a 12 percent interest rate is needed so that the demand for reserves will be high enough so that the demand for base money remains $2.6 trillion. Murphy seems to imagine that since each reserve balance would increase by 12 percent each year, base money would rise at a 12 percent annual rate. No. The Fed would simply have to sell between $200 billion and $300 billion in assets each year to leave base money the same. Rather than having to sell off $1.8 trillion in assets post haste, they could sell much fewer assets and prevent any excess inflation. Of course, controlling inflation would require the Fed sell more assets than the amount needed to keep the quantity of base money constant while paying sufficiently high interest rates to maintain the demand for reserves. While insolvency might still be the eventual result, there is time to get inflation and inflation expectations under control.

Speaking of the interest rates paid on reserves, these future risks are a reason why the interest rate on reserves should be cut from its current low level to zero, or perhaps negative. This at least partially substitutes for the need for quantitative easing and so for the amount of longer term securities the Fed needs to buy now. The fundamental problem that quantitative easing should solve is an excess demand for base money. Quantitative easing involves the Fed bearing increased risk for banks and their depositors. The banks get to hold low risk reserves, and the Fed holds the long term bonds with interest rate risk.

The possibility that the Fed might become insolvent, and even so insolvent that it cannot undertake the needed contraction in base money should not be a deterrent to quantitative easing. If necessary, the government should bail out the Federal Reserve. The simplest approach would be for the Treasury to swap the Fed's long term bonds for short term bonds at par. The Fed can then sell the short term bonds to contract the quantity of base money. The Treasury, of course, will have to sell new short term bonds to pay off the ones the Fed sold when they come due. The interest rates the treasury will have to pay will be higher, and so this will increase the government's interest expense. And the government should pay that expense and reduce other sorts of expenditures. Of course, the point of quantitative easing is to expand money expenditures on output, and raise production and employment. This will tend to reduce government social expenditures and raise tax revenue, helping with the deficit.

If the Fed becomes insolvent, there is little doubt that it would become subject to greater political scrutiny. Good. Fundamental monetary reform is needed. Most importantly, Congress should impose on the Fed a rule for a slow, steady growth path of money expenditures. Since such a growth path is inconsistent with persistent high inflation, a large run up in interest rates due to expectations of out of control inflation would be less likely. Further, the amount of quantitative easing needed to expand money expenditures an appropriate amount would likely to be less with a clear commitment to expanding money expenditures.


  1. "If the Fed becomes insolvent, there is little doubt that the result would be subject to greater political scrutiny. Good. Fundamental monetary reform is needed." This seems too hasty. Political scrutiny often leads not to *reform* (i.e., improvement) but to *making things worse*. Maybe we should cling to the devil we know.

  2. There are many ways to reduce inflation beyond selling bonds. If you believe that fiscal stimulus is expansionary, the government could pay down the debt to lower inflation. Since policy expectations are as powerful as current policy, the Fed could commit to not buying bonds for an extended period. Additionally, there are many games they could play with reserves, such as raising required reserves or implimenting a one-time lump sum tax on bank reserves. On the more extreme end of the scale, the government could tax and burn money, or partially default on debt.

  3. The Fed will simply hold all of the bonds to maturity. There is no solvency risk. They can still control inflation via the IOER. They'll simply raise the rate when the time comes. Big deal. There is no solvency risk because of the Fed's balance sheet.

  4. Holding the bonds to maturity only works if the yield on the bonds is greater than what the Fed pays on reserve balances. It is true, however, that it is only the difference between the Fed's interest income and the interest rate it pays that results in an expansion of reserves. This reduces the amount of assets the Fed must sell to prevent base money from rising.

  5. What does the Fed care if interest rates rise? They are buying largely 1.5% paper with 5 year maturity. The treasury will continue to pay the interest on the bonds they hold and one day they will mature and all will be cleared. It doesn't matter if rates rise to 2% in 5 yaears. The bonds will still mature at par. There is no risk of insolvency here.

  6. I don't undersand. The Fed owns 1.5% paper. Suppose it must pay 5% on reserve balances? It is losing money at a rate of 3.5% a year. After 5 years, that adds up to enough to wipe out a significant capital cushion. Of course, currency is a significant part of Fed liabilities, and the interest rate paid is zero, and so that provides a profit at a rate of 1.5%.

  7. This comment has been removed by a blog administrator.

  8. Can the treasury bail out an insolvent fed if it, too, is insolvent?

  9. We to be very careful with Fed decisions, as it at times can be seriously unpredictable, so we need to make sure we are fully prepared for everything and I hardly worry about all these things with OctaFX broker. They are awesome with having lovely conditions from lowest possible spreads at 0.1 pips to high leverage up to 1.500 while they also have mighty 50% bonus on deposit which is use useable and really helps us out in great way and ease.