David R. Henderson  

Monetary Policy: Giving vs. Buying Redux

PRINT
Self-Recommending could-have-b... Why Take Yoga Classes?...
The Fed doesn't expand the money supply by uniformly dropping cash from helicopters over the hapless masses. Rather, it directs capital transfers to the largest banks (whether by overpaying them for their financial assets or by lending to them on the cheap), minimizes their borrowing costs, and lowers their reserve requirements. All of these actions result in immediate handouts to the financial elite first, with the hope that they will subsequently unleash this fresh capital onto the unsuspecting markets, raising demand and prices wherever they do.
This is from Mark Spitznagel, "How the Fed Favors the 1%," Wall Street Journal, April 19, 2012. This is confused on a few levels. I've written about this before and so you can see it in more detail here. The basic idea behind my critique is that it's important to distinguish between giving and buying.

A couple of things specifically:
1. The term "capital transfers." A transfer is usually thought of as a transfer: giving wealth to someone with nothing in return. Spitznagel understands that and so he tries to justify the idea that the Fed gets less in return. How? By "overpaying." How does he know the Fed is overpaying? He doesn't say. How about "lending to them on the cheap?" He's on potentially firmer ground here. But I would like to see more.
2. The Fed "lowers their reserve requirements." Well, it sometimes does. But that's a reduction of a tax. It's not a handout.


Comments and Sharing


CATEGORIES: Monetary Policy



COMMENTS (16 to date)
James writes:

David asks, "How does he know the Fed is overpaying?"

Because the Fed must come in with a higher price than any other participant in the market. Otherwise, the sellers would sell to some party other than the Fed's open market desk. The difference between what Fed pays for bonds and what anyone else would pay for the same bonds is a gift.

Even if you claim the size of the transfer is small (and I'd believe anyone who claims it is less than a basis point which is still large when multiplied by the size of the recent increase in the money supply) the fact remains that the Fed remits interest on those bonds back to the treasury.

So the Fed's behavior is still a transfer, specifically a transfer from people with less influence over federal spending to people with more influence over federal spending. This transfer may be good or bad for the economy, depending on whether or the people who direct federal spending use the money in ways that are good or bad for the economy. I'll happily defer to you on this one.

Joseph K writes:

I think you make too much of this giving/buying thing. The point is that the Fed is increasing the money supply and that money has to enter somewhere, and one of the big points of entry is the banks, in the form of subtle transfers. If they lend the bank at low interest rates and the banks can relend for a profit, then that's a transfer at the Fed's expense. Also, with lowering the reserve requiements, then they're permitting the banks to take on higher risk, and since the federal govt insures those risks, then when the banks lend out that additional money for profit it's at the expense of those insuring them. It's the same type of thing when the govt subsidized the trains in the 19th century by paying exorbitant freight rates for mail delivery (since the postal service was run by the federal govt).

David R. Henderson writes:

@James,
You say we know that the Fed is overpaying because "the Fed must come in with a higher price than any other participant in the market.” I’m not sure that’s literally true. The Fed might pay the same as all buyers. When I go to Safeway to buy grapes, I pay the same as other grape buyers pay.
But let’s say you’re right. If you are, then does that apply to other markets? Does the fact that I bought grapes mean that I necessarily overpaid for them? If so, then, by that logic, no one should ever buy anything.

James writes:

David,

If the Fed wants to inject $X into the economy via open market activity, the open market desk must be the highest bidder in $X worth of transactions. If the other buyers want to buy $Y worth of bonds, then price of bonds goes up to the point on the supply curve where all orders for $X + $Y get filled. I suppose there may be special cases where the open market desk just crowds out all other buyers, for example if the open market desk bought the entire federal debt. In practice, this is not what happens.

I'd say that my idea only applies when one party has the power to tax or to print money and uses that power to make the wininng bid in an auction on a quantity large enough to move the price of a good along the supply curve. Your grape purchases probably differ. If you did make a habit of printing money to buy mass quantities of grapes in an auction setting, I'm sure we agree that this would lead to a change in the agricultural industry to better satisfy your preferences. Do you think the Fed has no similar effect from an industrial organization perspective?

Bu let's say you're right. If the open market desk is only exchanging like for like, the much larger transfer from the politically disconnected to the politically connected remains.

Mike Sproul writes:

If the fed pays $101 for a bond worth $100 then the fed has just transferred $1 to the lucky bond seller. If the fed lends at 2% in a 3% world, then the fed has just transferred 1% to the lucky borrower. But as long as the fed only trades and lends at market rates, there is no transfer to anyone. Furthermore, such transactions leave the fed's net worth unchanged and therefore cause no inflation.

Bob Murphy writes:

David,

You raise an interesting point; it's almost philosophical. What does it mean to say any buyer "overpaid" for something?

Clearly, if Bernanke wrote you a check for $10 million to put your house on the Fed's balance sheet, we can all agree that it was a handout or gift to you, and that Bernanke overpaid.

(I'm assuming here that you don't in fact live in such a nice house, but maybe I'm wrong...)

Now what if Bernanke held out a standing offer to buy $10 million for all such houses that fit a certain description, such that the resulting new market price were $10 million? Would it still be true that the Fed was overpaying?

How now do we distinguish that from a situation in which there is a large market player who, with its demand, makes the resulting equilibrium price higher than it otherwise would be?

I'm not trapping to trap you, I'm trying to think through these issues. As you can guess, I think the Fed overpays for these assets, but I'm admitting it's a little tricky to define exactly what I mean by that term.

James writes:

Bob writes, "I think the Fed overpays for these assets, but I'm admitting it's a little tricky to define exactly what I mean by that term."

It's simple Bob. If the Fed gives you a dollar in exchange for 99 pennies, then the Fed is overpaying. If the Fed pays a dollar to buy assets which sell in a liquid market for $0.99, this is equivalent to the first case and so this is also a case of overpaying.

Why else are primary dealers willing to do business with Fed at all? If they don't do business with the Fed, they can either sell assets at market prices or (2) not sell those assets. Trading with the Fed must bring some added benefit or no prift seeking organization would do so. What is that added benefit if not the inflated prices that the fed pays for those assets?

Bob Murphy writes:

James wrote:

Why else are primary dealers willing to do business with Fed at all? If they don't do business with the Fed, they can either sell assets at market prices or (2) not sell those assets. Trading with the Fed must bring some added benefit or no prift seeking organization would do so.

James, replace "the Fed" with "organization X." Your logic still holds. And you just proved that all organizations overpay for financial assets.

Jeff writes:

This is just silly. In every voluntary transaction, the buyer values the thing bought more than the seller does, else the transaction wouldn't take place. This is not "overpayment", it's just normal, everyday transacting.

Spitznagel's story about how Fed-initiated changes in interest rates happen is also wrong. If the Fed wants to lower the funds rate from 3 percent to 2 percent, it issues an announcement to that effect and directs the NY Fed to make it so. And, like magic, before the NY Fed does anything at all, the rate drops to the new 2 percent level immediately. It's what Nick Rowe calls the Chuck Norris effect: Everyone in the market knows that the NY Fed can make the new rate stick by loaning unlimited amounts of money it creates on the fly, so they immediately take this into account when calculating what rates they themselves are going to offer and/or accept. There is no interval in which some favored insiders get to borrow from the Fed at 2 percent while everyone else is paying 3 percent.

tom cullis writes:

Jeff,

Spitznagel's story is correct because there are lenders/borrowers in the system without access to the Fed that are not both lenders and borrowers. For example anyone with a savings account with a bank is a lender into the system but without access to the Fed. These people (expect to) see their interest payments drop when the Fed lowers rates. Because they cannot borrow from the Fed directly they must go through one of an artificially limited number of institutions to borrow through the Fed, and an artificially limited number of institutions means an artificially less than perfectly competitive market which means there are rents to be gained by those participants. Perhaps those rents are in the forms of fees, higher rates or lower customer service but those rents are expected.

@ David Henderson re:#2

Altering reserve requirements for a small portion of banks grants them favored status. Unless the Fed is altering requirements on a perfectly even basis across all markets they influence they are granting additional profits to a small segment based on political status, not comparative advantage, which is in effect a redistribution from those without that status.

David R. Henderson writes:

@tom cullis,
Altering reserve requirements for a small portion of banks grants them favored status.
True. Does the Fed do that? I’ve never heard of the Fed lowering reserve requirements for some banks and not others. Do you have a cite to cases where the Fed has done that any time in, say, the last 25 years?

Jeff writes:

@tom cullis,

There are thousands of commercial banks in the US. About a third of them are members of the Federal Reserve System and have their own accounts at a regional Federal Reserve Bank. The others use correspondent accounts. Competition for banks keeps fees for the correspondents very low.

Barriers to entry into the banking business are not very high. If there are so many rents to be had, why aren't you a banker?

James writes:

Bob,

No. I'm just assuming that the bond market is perfectly efficient before the Fed shows up. Any buyer in an efficient market is overpaying if his transaction causes prices to go up, right?

You can say that I believe anyone who buys any set of financial securities other than a scaled version of the market portfolio is very very likely to be overpaying for those assets. I'm also assuming away the possibility that participants in financial markets hane non-financial motives e.g. they enjoy holding stock in firms they admire.

Jeffrey Rogers Hummel writes:

Much of this discussion seems to involve minor semantic quibbles over what constitutes a government transfer. To the extent the substantive issues have been raised, they involve (1) forgetting that the Treasury is also involved in the market for its securities, as well as the Fed, or (2) confusing government seigniorage with Cantillon effects.

Regarding the first point, remember that the U.S. Treasury sells and pays off its own securities, and in theory could also buy them back early to reduce its outstanding debt. So if the Fed is simply expanding the monetary base to cover an increase in the Treasury deficit, you have the Treasury selling its securities, ceteris paribus reducing their price slightly, at the same time that the Fed is buying Treasuries, ceteris paribus increasing their price slightly. The net effect on Treasury prices should be a wash. Admittedly the Fed cannot legally deal directly with the Treasury, so the primary dealers may earn a slight spread, but they earn that anyway when they resell Treasuries to the general public. To the extent that that Fed and Treasury actions in the open market do not match up perfectly, they may either be increasing or decreasing the price of Treasury securities, depending on whether purhcases exceeds sales or vice versa.

Whenever the Fed increases the monetary base, this also generates seigniorage. The seigniorage gain results from the impact of monetary expansion on the price level the therefore the public's real cash balances. This indeed is a transfer, but one that accrues directly to the government at the expense of the general public, whose money holdings have lost purchasing power (relative to what they otherwise would have had) exactly equal to the government gain. Except for the Fed's operating costs, the seigniorage always reverts to the Treasury in the form of rebated interest earnings. Moreover, as I have pointed out in several past articles, seigniorage has become a trivial source of revenue for governments in the developed world. In the U.S., seigniorage was less than half a percent of GDP (covering only 2 percent of government expenditures) during the great inflation of 1970s, and has been still lower ever since.

Price level effects can also cause redistributions among the general public, and the magnitude depends on how accurately inflation is anticipated. Holders of NET nominally denominated liabilities will gain and holders of NET nominally denominated assets will lose. Most financial institutions simultaneously borrow and lend, so the impact is mainly reflected in the relative maturities of these assets and liabilities. Banks and other depositories, by borrowing short and lending long, tend to be hurt by unexpected inflation, as we saw in the 70's inflation and subsequent S & L crisis.

Cantillon effects, in contrast, result from the impact of monetary expansion NOT on the price level but on relative prices. Thus, ALL government expenditures have potential Cantillon effects. A new, tax-financed subsidy of hula hoops would initially decrease their relative price, and this change would persist as long as the subsidy is in effect. The Cantillon effects of monetary expansion depend on where the new money is initially injected. To the extent that Fed monetary expansion flows almost directly to new Treasury borrowing, such effects should appear not in financial markets but wherever the Treasury is spending the new money. This is most notably the case during wartime inflations. But if Fed open-market purchases far exceed Treasury sales, then the initial effect will obviously be in financial markets. And the size and the duration of the effect will depend upon how large the Fed operation is relative to the market and how quickly prices throughout the economy adjust.

Bob Murphy writes:

Jeffrey Rogers Hummel wrote:

Much of this discussion seems to involve minor semantic quibbles over what constitutes a government transfer. To the extent the substantive issues have been raised, they involve (1) forgetting that the Treasury is also involved in the market for its securities, as well as the Fed, or (2) confusing government seigniorage with Cantillon effects.

Wow, everything that we've written until Jeff showed up was bad. I'm left trying to classify my own mediocre thoughts. Were they a minor semantic quibble, did they forget about the Treasury, or did they confuse seigniorage with Cantillon effects? It may take me a day to sort out my mistakes.

Bobby writes:

We also have to consider the implications of a Fed asset purchase.

This creates moral hazard, the banks have it both ways- it does not matter if the Fed is overpaying for the assets (I believe they are), they transfer the risks to the Fed rather than losing individually.

The banks get all of the upside without any downside. Heads they win, tails we lose. The fact that the Fed is paying anything for impaired loans with present value of zero or even negative is a net gain for the banks, and increases the profitability for a bank.

So what we see ultimately is that the Fed creates excessive profits when times are good, and mitigates losses when the banks suffer from declines in markets.

Comments for this entry have been closed
Return to top