The Fed Already Has a Blank Check

In the interest of continuing the conversation, I want to argue a contrary position and push back a bit on Steve Waldman's post about allowing the Fed to pay interest on reserves, and his worry that this change will allow the Fed to put excessive amounts of public money at risk:

Let's not write the Fed a blank check, by Steve Waldman: Last week, the Fed decided to ask Congress for the right to pay interest on bank reserves. (Hat tip Barry Ritholtz, see also William Polley, Mark Thoma, Brad DeLong) This is a very big deal.

Don't be misled into thinking that the Fed's proposal is just some arcane, technocratic change. The Federal Reserve is asking taxpayers for a big pile of signed, blank checks. That's far too much power to put in the hands of a quasipublic organization with little democratic accountability. This authority should not be granted without some strong strings attached. ...

First, some background. There is a trend among central banks to move from old-fashioned, fractional-reserve banking to a system whereby interest rates are managed via a "channel" or "corridor", and under which fixed reserve requirements might be dispensed with entirely. The basic idea is simple. The Fed ... choose two interest rates, a "floor rate" at which the Fed would stand ready to borrow funds, and a "ceiling rate" at which the Fed would stand ready to lend. As long as there is no stigma attached to transacting with the Fed, banks would never lend for less than the floor rate or borrow for more than the ceiling rate. The interbank interest rate would necessarily lie within a "corridor" defined by these two interest rates. ...

A corridor system would represent a meaty change to how central banking is done in the US, but the approach seems to work okay in other countries. ...

As long as the Fed is conducting ordinary monetary policy, switching to a channel system offers modest benefits at a modest cost to taxpayers. But the Fed's monetary policy has not been ordinary at all lately. In fact, it's been quite extraordinary. It is in the context of this extraordinary policy that the Fed has asked Congress to accelerate its authority to implement a channel system...

The core of the Fed's new exuberance is a willingness to enter into asset swaps with banks. The Fed lends safe Treasury securities to banks, and accepts as collateral assets that private markets consider dodgy or difficult to value. (This is the direct effect of the Fed's TSLF program, and the net effect of TAF and other lending arrangements that the Fed sterilizes in order to hold its interest rate target.) In doing so, the Fed puts taxpayer funds at risk. If a bank that has borrowed from the Fed runs into trouble, the Fed would face an unappetizing choice: Orchestrate a bail-out, or permit a failure and accept collateral of questionable value instead of repayment. Either way, taxpayers are left holding the bag. ...

In December, the Fed had $775 worth of Treasury securities. That stock will soon have dwindled to $300B, give or take. The difference, about $475B, represents an investment by the central bank in risky assets of the US financial sector.

$475B is an extraordinary sum of money. It is as if the Fed borrowed more than $1500 from every man, woman, and child in the United States, and invested that money on our behalf in Wall Street banks that private financiers were afraid to touch. For bearing all this risk, if things work out well, taxpayers will earn about what they would have earned investing in safe government bonds. If things don't work out well, the scale of the losses is hard to predict. ...

If the Fed were to blow through the rest of its current stock of Treasuries, it would have invested more than $2500 for every man, woman, and child in America. Public investment in the financial sector would have exceeded the direct costs to date of the Iraq War by a wide margin. Would that that be enough? If not, how much more? Just how large a risk should taxpayers endure on behalf of companies that arguably deserve to fail, to prevent "collateral damage"? Have we considered other approaches to containing damage, approaches that shift costs and risks towards those who benefited from bad practices, rather onto the shoulders of taxpayers and nominal-dollar wage earners? Does this sort of policy choice belong within the purview of an independent central bank?

Now I don't actually mean to be too harsh. Putting aside the years of preventable foolishness that got us here, ... a crisis emerged that had to be managed and the Fed was the only organization capable of stepping up to the plate. I don't love the decisions that were made, but decisions did have to be made, and there weren't very good options. But now we have a moment to reflect. If the credit crisis flares hot and bright again, how much more citizen wealth should be put at risk before other policy options are considered? That's not a rhetorical question: We need to choose a number, a figure in dollars. My answer would be something north of zero, but not more than the roughly $300B stock of Treasuries that remains on the Fed's balance sheet. But this is a decision that Congress needs to make.

And what does all this have to do with the question that will soon be put before the Congress, whether the Fed should be permitted to pay interest on deposits?

Everything, as it turns out. Suppose the Fed decides it wants to swap more than the $300B in Treasury securities it currently has available in order to support the financial system. Given its current tools and practices, the Fed would have to print money in order to buy more Treasuries to swap. But if it did that, the extra cash would drive interest rates below the Fed's target level, quite likely provoking inflation. The Fed cannot simultaneously swap away more than its existing stock of Treasuries and satisfy its legal mandate to promote price stability, unless it resorts to something weird.

But suppose Congress gives the Fed the authority to pay interest on reserves. Suddenly the Fed can print cash to buy all the Treasuries it wants to swap for troubled assets. When banks find they have more cash than they need, they lend the money back to the Fed, collecting the "floor" interest rate and removing the currency from circulation. Since interest rates can be held to any level by adjusting the "corridor", the Fed would retain the flexibility to respond to inflation. At the same time, it would be able print cash in any amount that it pleases — "to infinity and beyond!" — in order to fund asset swaps (or outright purchases) at taxpayers' risk. This strikes me as a delegation of Congressional authority that would not only be undesirable, but arguably unconstitutional.

So, should we simply refuse the Fed's request? Probably not. Brad DeLong makes an excellent point:

The Fed may also want to raise the general level of interest rates in order to fight inflation--which requires that it sell its Treasuries for safe bank reserves rather than temporarily swap them for risky MBSs.

The Fed is already rubbing pretty close to its "balance sheet constraint". If, after exposure to gamma radiation from televised images of food riots, Ben Bernanke were suddenly transformed into The Incredible Volcker, he might lack the tools he'd need to jack rates up into the muscular high teens, unless he's given this new authority. So what should we do? James Hamilton has an answer:

Congress has a quite proper role in determining the magnitude of the fiscal risk that the Fed opts to assume... [A] statutory limit on the non-Treasury assets that the Fed is allowed to hold might make sense. Perhaps the outcome of a public debate on this issue would be a decision that the Fed needs the power to lend to private borrowers even more than the $800 billion or so limit that it would run into from completely swapping out its entire portfolio... Or perhaps after deliberations, Congress would decide that the business of swapping Treasury debt for private sector loans is one that is better run by the Treasury rather than the Federal Reserve.

I agree. I think that Congress should grant the Fed's request, but it should simultaneously impose constraints on the composition of the Fed's balance sheet that cannot be violated without express legislative consent. This will be a complicated exercise, unfortunately. Besides government debt, central banks quite ordinarily hold precious metals and foreign exchange, and limitations on non-Treasury assets will have to take this into account. Plus, restrictions would have to be written carefully to apply to off-balance sheet arrangements such as TSLF, and contingent liabilities like the insidious reverse MBS swap proposal. Finally, Congress must consider restrictions on the Fed's ability to enter into derivative positions, whether directly or indirectly via special purpose entities, including how the bank's existing derivative book should be managed and whether the bank should or should not guarantee the liabilities of current Fed-affiliated SPEs.

Congress might also limit the quantity of reserves on which the Fed will be permitted to pay interest.

The Fed can retain full independence for the purpose of conducting ordinary monetary policy, exchanging government debt for cash and vice-versa. But if the central bank wants to put ever greater quantities of public money at risk, it will have to accept a lot more public supervision. If the prospect of intrusive oversight is too much for the Fed, then, as James Hamilton hints, perhaps the roles of central bank and macroeconomic superhero should be moved to separate boxes on the organizational chart. If we are not careful, the next bank requiring a taxpayer bailout may be the Federal Reserve system itself.

This is an old problem - how much authority should be centralized thereby allowing quick and immediate response during a crisis, and how much should be retained in slower, deliberative bodies like the House and Senate? The War Powers Act reflects this compromise - we want the ability to respond quickly to an attack or other military developments, but we worry about the concentration of power in the hands of a single individual. Centralization has the benefit of allowing a quick response to a crisis, but it risks being out of step with the democratic process. In the case of financial market emergencies, however, I have more faith in the Fed than in congress to act quickly and correctly. That's partly because I have little faith in the ability of congress to quickly comprehend what the problem is and attack it directly and effectively - many of them admit to not having a clue about economics, and more worrisome are the ones who think they have a clue but don't - but congress should not give up its oversight role.

Some "off the cuff" thoughts:

1. Public money is always at risk when the Fed does open market operations. The amounts may not be as large, but the risk is always there. For example, suppose the Fed prints $100 and purchases a T-Bill for $100, and that while it is holding that T-Bill, the price falls to $75. The Fed has just taken a 25% loss - if it tries to sell the T-Bill back to the public, it will get less for it than it paid originally. The big difference is that with MBS and other risky securities there is default risk - the value could fall to zero - but this is about the magnitude of the loss, not the fact that the Fed is able to put taxpayer money at risk. In addition, every action the Fed takes can affect the public treasury. If monetary policy changes GDP, tax collections change and this affects the deficit. Similarly, when the Fed changes the interest rate, the amount of interest paid on the accumulated federal debt changes - and this can be quite a bit of money. So everything the Fed does has the potential to hit you in the pocketbook, and if it makes a big mistake, the hit could be fairly large (a 3% decline in GDP is more than 400 billion dollars). The fact that the Fed can impact your pocketbook is nothing new. [Update: the first comment makes me realize I should have been more careful here, this is only the Fed's part of the government balance sheet, and I should have noted this is not the main argument - that is in 3 and 7 - i.e. that the Fed already has the authority to put money at risk so that is not a reason to object to paying interest on reserves.].

2. The Fed's ability to purchase risky securities comes under emergency authority. When the emergency ends, the ability ends, so this criticism is only about extraordinary times in financial markets. There haven't been many of those.

3. The channel of corridor system does not alter the Fed's ability to put public money at risk, it can do that now. For example, the worry above is that the Fed will want to take risky securities off the market by trading them for safer government bonds, but it might run out of government bonds to trade. If it does, it can print money to purchase more risky assets, but this could be inflationary. To avoid the inflation, it buys the money back by offering to pay interest on reserves (notice what happens if the value of the risky assets the Fed is holding falls to zero - it gave out money for free, then buys it back by paying interest). But the Fed can already do this under existing authority - it can already purchase risky securities and avoid inflation even when its inventory of T-Bills is zero. Just do as above, print money to purchase the risky securities, then increase the required reserve ratio and make banks hold the extra reserves in the vault or at the Fed (at zero interest). This avoids the potential for inflation, and it does not require that the Fed pay interest on the extra reserves banks are forced to hold. The difference between the current procedure and a corridor system is that under the corridor system banks are compensated with interest for being forced to hold reserves idle, but they can be forced to hold the reserves at zero interest under existing authority.

4. The worry seems to be that the Fed will lose all $475 billion. If there are losses, it is likely to be a fraction of the total, not the full amount. If it is the full amount, we're in big trouble anyway, and the Fed was probably correct to try to prevent such a disaster. But we should probably value the loss as a probability times $475 billion, not as the whole amount.

5. Are there losses? If the Fed prints money and buys an asset, it has created an asset out of thin air - newly printed money - and purchased an asset from the private sector. Thus, by creating as asset out of thin air, the Fed was able to get something of value. If the asset the Fed bought - say an MBS - has its value fall to zero, then the effect is just as if the Fed printed money and gave it to the public - dropped it out of a car on the street, dropped it from a helicopter, something like that. But where is the loss? (That's a question, I need to think about the "seigniorage", etc. aspects a bit more.)

6. Steve says:

It is as if the Fed borrowed more than $1500 from every man, woman, and child in the United States, and invested that money on our behalf in Wall Street banks that private financiers were afraid to touch. For bearing all this risk, if things work out well, taxpayers will earn about what they would have earned investing in safe government bonds. If things don't work out well, the scale of the losses is hard to predict.

But isn't the return is larger than this? Shouldn't we include the (probabilistic) benefit from not having slipped into a major recession? If, say, $250 of the $1500 is lost, but by taking the action the Fed prevented a fall in GDP valued at, say, $500 per person, aren't we better off? As noted above, a 3% fall in GDP for one year is over $4oo billion.

7. The main thing to note is that the Fed already has a "blank check" through manipulation of reserve requirements as described above. If you are worried about the Fed putting public money at risk in an emergency, that risk already exists and the corridor or channel system does not change that. So I don't see this as a valid objection to the channel or corridor system.