There’s no risk for the Fed . . . eral government
If you are a corporation is it risky to buy back your own debt? Most people would say no. How about if you are the Federal government? Again, it doesn’t seem risky.
QE is not risky for the US government. (Assuming the MBSs that are being purchased are already guaranteed by the Treasury.)
Is QE risky for the Fed? Not really, although I suppose if a person really wanted to believe it was risky, they could think of some sort of convoluted theory to explain why.
The big “risk” that people worry about is that interest rates might rise sharply, and this could cause large capital losses if the Fed’s huge portfolio of bonds were sold. I see several problems with that theory:
1. First of all, the thing that would likely cause higher interest rates—increased NGDP growth, would actually improve the budget situation of the Federal government. Since the Fed is part of the Federal government, that’s all that really matters.
2. But let’s say you were really worried about the Fed’s balance sheet. Some argue the Fed could simply hold the bonds to maturity. But in that case if they wanted to avoid hyperinflation they’d have to sharply increase the interest rate paid on excess reserves, which could cause losses in the short run. So holding bonds to maturity doesn’t avoid losses that would be incurred by selling bonds at a loss if nominal rates rise unexpectedly. There’s no free lunch there. Even so, why should we care if the Fed loses money? It’s not a for-profit institution; it’s part of the Federal Government.
3. One fear is that the Fed might become bankrupt, as its losses on bonds might exceed its capital. It could lose its independence. However their liabilities also include a trillion dollars of currency in circulation. Or should I say “liabilities” as it’s debatable how we should regard currency. If nominal rates rise to 5% then the Fed might have to go a few years without paying a dividend to the Federal government. But is that so bad? In the past few years they’ve been turning over massive profits to the Treasury, several times larger than normal. If they went a few years without paying any money to the Treasury, it would merely offset the recent excess profits. As the older low yielding T-bonds were replaced by newer bonds yielding 5%, the Fed would turn back into the government version of Apple Computer—a money machine churning out profits of $40 or $50 billion per year.
4. Is there some tail risk I’ve missed? Could the losses exceed the stock of currency? Not really, but let’s say they did. Suppose the US shifted to electronic money just as this bear market in bonds was occurring. All currency was removed from circulation. What then? The Treasury could simply issue another trillion in T-bonds and give it to the Fed. Now this might prove embarrassing for the Fed (although I can’t imagine why–it’s not their fault if the US suddenly shifted to all-electronic money.) But is it a problem? Clearly not. And it’s not going to happen–there’ll still be a trillion in cash outstanding 10 years from now, when the Fed has adjusted to the higher interest rate.
Here’s what is risky; continued tight money, which worsens the fiscal situation so much that the Treasury eventually turns the Fed into its lapdog. Tight money leading to fiscal dominance. It happened once before (1930 to 1951), let’s not let it happen again.
BTW, I’d add that it’s very unlikely that 10 year T-bond yields will rise back to 5%. But I didn’t want to rely on this argument. Trust me—if the Fed were privatized and did an IPO, its market cap would greatly exceed Apple’s.
PS. The Japanese case is far more interesting. Suppose they adopt a 2% inflation target, and their NGDP growth rises to 3%. This would be a huge boon to the Japanese government, even if the nominal interest rate on newer bonds rose by the full 3%. That’s because the Japanese government, like all debtors, gains from unexpected inflation. Its existing debt was sold under expectations of no inflation. But even better for the Japanese, it’s quite possible the nominal interest rate on newer debt would not rise by the full 3%, as they are currently stuck at the zero bound. Indeed rates might not rise at all. In that case the Japanese have been walking right past 10,000 yen notes lying on the sidewalk, without bothering to pick them up. I can’t see that situation lasting much longer.
HT: Chris Beseda
Tags:
28. February 2013 at 12:46
This would be a huge boon to the Japanese government, even if the nominal interest rate on newer bonds rose by the full 3%. That’s because the Japanese government, like all debtors, gains from unexpected inflation
By the same token, of course, it’s a huge whammy for the holders of those bonds, which is mainly the Japanese people.
I suspect this is one driver of current policy — while I personally would gladly trade my bond yields for future growth, a lot of people wouldn’t, especially older people who won’t see gains in employed income. The fear of inflation is still very strong with the Booomers.
28. February 2013 at 14:16
Scott
Why raise the IOR rate? Why not just raise the required reserve ratio?
28. February 2013 at 14:19
Scott
Did you read Miskins paper. Why do you think he’s making a big deal on this issue? Or do you think he just getting off by re-stating the obvious and trivial in mathematical form.
28. February 2013 at 14:27
I can imagine a model where an NGDP-targeting CB could carry out all its operation via CB-bonds. These bonds would have no other purpose than to be traded by the CB. (If the govt wanted to raise money via bonds this would be a totally separate operation)
If NGDP is above target it sells CB-bonds (which it will create if it doesn’t have any on its books)
If NGDP is below target it buys bonds.
It will always buy and sell at the market rate/price. These bonds would be totally risk-free since the CB will always be in a position to honor them.
In some years the CB may make a paper profit and in some years a paper loss. But this will not matter since its sole function is to stabilize NGDP. Its profits would never be shared with the government.
28. February 2013 at 14:38
Risk-less purchase of your own debt? Sounds fishy. Under what conditions is it not risk-less? Are you sure you’re not leaving some other parties who have to bear the risk out of the equation?
28. February 2013 at 14:58
It’s kind of amusing to think that a corporation that makes a profit for over 100 years straight taking a loss in one year would be a big deal. But with the mindless Fed bashing by the know-nothing Ron Paul/Glenn Beck crowd I guess it’s not surprising.
But if it’s such a big political issue, the Fed does have 250 million ounces of gold on its books at the official price of $42 an ounce, or $11 Billion. It could sell a portion of that to cover its paper loss.
But my favorite solution is dtoh’s, raise the required reserve ratio. I don’t consider that a tax, it’s a user fee for the privilege of being a member of the Fed system – no different than charging royalties for drilling on public land, or charging to enter a national park.
28. February 2013 at 15:13
If you are issuing more currency to buy back debt, that doesn’t seem riskless. If you are raising money some other way (say, using retained earnings if you are a corporation), that seems less risky.
28. February 2013 at 15:21
QE is not a debt buy back. The size of the public debt is unaffected.
The Fed could easily neutralize interest rate risk if it wanted. Certainly interest rate risk has nothing to do with conventional monetary policy.
But I guess if the Fed did that, they would have a harder task convincing people (other than die hard monetarists) that QE “works”.
28. February 2013 at 15:44
TallDave, Remember, it’s not a zero sum game.
dtoh, Higher reserve requirements would put a massive tax on the banking industry. Not saying it’s a bad idea, but it would be highly controversial.
I didn’t get the impression that they were all that worried about the risk, but I just skimmed the paper.
Carl, I don’t see the risk.
Negation, Yes, they should raise the gold price to current market value—mark to market.
Max, I hate to tell you this, but the stock, bond and forex markets also believe QE “works”, if by works you mean impacts expectations of inflation and growth. Only MMTers who pay no attention to the real world deny this. Even Krugman has recognized the market reaction. So it’s not just monetarists.
And I’d add that QE is not the monetary policy proposed by “market monetarists.”
28. February 2013 at 16:02
Scott,
I agree with most of what you are saying, but on a narrow point Jeremy Stein and Miles Kimball are right. Stein fears that the risk premium on long term government bonds is negative, while Kimball has pointed out that there are plenty of assets the Fed is prohibited from buying that have positive risk premia.
Another point – if highway dept loses money on some complicated interest rate gamble, we will get more potholes. The same applies to the Fed – if the Fed loses money, we will get more macro volatility in the future.
28. February 2013 at 16:22
When the Fed buys treasury paper from the Chinese central bank and then acknowledges that the paper is worthless by ripping it up, isn’t this effectively a handout of public funds to the Chinese? Is there really nothing objectionable about this?
28. February 2013 at 17:00
Is there any sense in which the Fed could run out of (anti)ammo – like its balance sheet is insufficient to buy back enough MB, in the face of hyperinflation? Is there any reason higher IOR would be an inadequate substitute for OMO?
28. February 2013 at 17:19
Those who say we should be worried about the Fed putting too many “risky assets” on their balance sheet, most often don’t really grasp the fact that the Fed is able to create money virtually out of thin air, and never have to go bankrupt in the technical sense.
This is not to say that we should not be worried at all about what the Fed is buying or doing though.
28. February 2013 at 17:31
Infaltion does not benfit all debtors.
With the levels of debt that Japan faces increases in Inflation without increases in real GDP will cause debt service to rise faster than NGDP and tax revenue. It is an ustable situation. If government expenses (and pensions) are baselined to inflation as they are in the US, the problem is doubled.
28. February 2013 at 17:56
I’m not sure what point 4 is trying to say. Federal Reserve accounting dictates that losses in excess of what’s necessary to equate retained earnings to capital paid are to be booked as a deferred asset, effectively making it impossible for the Fed to ever be insolvent. So, why does it matter if losses exceed the stock of currency? I dont think it does, except, perhaps for appearance issues, which the Fed would face in any event.
28. February 2013 at 18:13
What would happen if the Fed stopped rolling over the bonds on its balance sheet? Or burned them?
28. February 2013 at 19:41
I am trying to figure out what Quantitative Easing is really all about.
If I buy a $10,000 10 year treasury bond (increasing the federal debt) and some time later the federal reserve gives me $10,000 to buy back my bond (reducing the federal debt), is the fed purchase of my bond quantitative easing?
If this is quantitative easing then it seems to me that quantitative easing is the same as the federal reserve acting to reduce the federal debt. Is this correct?
28. February 2013 at 22:01
Very thoughtful blogging.
We must consider that the Fed will simply hold the QE-securities to maturity. And then transfer the funds to the Treasury. It never “takes a loss.” The Fed prints money anyway. It is federal debt anyway. I get a headache think gin about the metaphysics of this.
Transferring proceeds from maturing securities (that is, from the Fed to the Treasury) should allow tax cuts, btw, and thus be stimulative.
In Japan, 2001-6, they went with QE. John Taylor gushed about the positive results. Obviously, Japan’s QE was not long enough or hard enough, despite John Taylor being so satisfied.
Okay, if five years of QE in Japan was not long enough, maybe 10?
Or, might QE become a permanent part of the Fed arsenal? The new “conventional” policy? The economics profession must start to contemplate this reality.
So the Fed may be building a balance sheet for a long time. That’s okay, it is deleveraging our economy. Perhaps it should only buy Treasuries, to avoid conflicts of interest.
People keep thinking rates will go back up. Maybe so. That has not been the experience of Japan.
And Scott Sumner has pointed out we may well see recession again before we see interest rates go back up. See you in Japan, if the Fed does not get aggressive.
But most importantly, as Sumner says, we cannot cater to people who see boogeymen behind every tree, on the road to an expansionary monetary policy. There are risks with every policy.
There is less risk when the economy is growing, besides being a lot better place to live and work.
1. March 2013 at 04:09
[…] See full story on themoneyillusion.com […]
1. March 2013 at 06:31
123. If the Fed loses money from a tight money policy we will get more macro instability. If it’s from an easy money policy we will get less (unless the policy is too easy–producing excess NGDP growth.)
Rademaker, It’s worthless to the US government, but very valuable to the Chinese.
Josh, Technically possible, but realistically impossible.
Doug, What matters is real debt service costs. Those don’t rise with inflation.
Alex, You may be right (which makes my argument even strongerr.) I don’t know the accounting method they use.
Michael. In realtiy nothing would happen, but it might be politically embarrassing.
Bob, Yes.
Ben, Good points.
1. March 2013 at 06:48
Scott: “If the Fed loses money from a tight money policy we will get more macro instability. If it’s from an easy money policy we will get less (unless the policy is too easy-producing excess NGDP growth.)”
Assume a benevolent central bank. Greater stability of NGDP requires the ability to do temporary expansions and reductions of money supply. This ability is impaired if the Fed loses money.
1. March 2013 at 09:05
Scott: I would think that if there were no risk, the government should perform quantitative easing all the time. Are you saying that there is only risk when the economy is in the particular state of having sub-optimal NGDP growth?
And, why are you able to rule out the risk that by increasing the market for government debt, you are increasing the demand for government at a time when the economy is least able to afford it?
1. March 2013 at 11:04
“Doug, What matters is real debt service costs. Those don’t rise with inflation.”
They do! If a country is more than 100% indebted, a spike in inflation (with RGDP continuting to grow) will result in debt service growing faster than NGDP. Debt Service / GDP grows even though Debt / GDP declines.
The increased debt service costs is long lived, and by the time the high cost debt is repaid — holding all else equal Debt / GDP will stabilize at a higher level than had we not had the inflation spike.
1. March 2013 at 11:39
Doug:
“They do! If a country is more than 100% indebted, a spike in inflation (with RGDP continuting to grow) will result in debt service growing faster than NGDP.”
Agreed.
With high debt loads, there arises a non-linearity between revenues and expenses. Then expenses rise exponentially relative to revenues. Real costs do in fact rise with inflation in this case, because one’s after expense net income, corrected for inflation, falls.
The error in the claim “Real costs do not rise” assumes (among other assumptions) that revenues and expenses always rise at the same rate.
1. March 2013 at 11:45
Doug M & Scott,
I think you’re both right, but I’m going to quibble a bit here. Real debt service costs often rise with inflation because the expectation of future inflation rises with it.
The increased debt service cost can be long lived, but is not always so. In my opinion the bottom line is that it’s the investors that ultimately matter the most. Higher inflation can (counter-intuitively) be supportive of bond prices if investors believe that the inflation is indicative of a reduced risk of default. You see that effect more clearly with junk bonds, where an improved economy and decreased risk of default can offset the upward movement of interest rates brought on by a recovery.
Higher inflation can also (more intuitively) do serious damage to bond prices, thereby raising borrowing costs, if investors believe that the inflation/devaluation is a panacea that is not backed up by the sort of structural changes necessary to make the long term outlook more positive.
1. March 2013 at 18:10
123, I don’t think losing money impairs their ability, as I tried to explain in this post.
Carl, It’s the recession that increases the demand for government, not QE. Ever since NGDP growth plunged government has become a bigger share of GDP.
Doug, You are confusing real and nominal variables. Inflation doesn’t increase real debt service costs. This is just basic economics.
1. March 2013 at 21:47
“Inflation doesn’t increase real debt service costs. This is just basic economics.”
It does, it does!
The country of Ruritania has
GDP 100 RUR (the currency of Ruritania)
Debts of 200 RUR
Real GDP growth is 1%
Inflation is 1%
NGDP growth is 2%
the real interest rate is 1%
the nominal rate is 2%
Debt servie is 4 RUR
The 20% of the country’s debt is refinance each year into new 5 year bonds
Spending is 20% on GDP
taxes are 20% of GDP
The budget defict is 4 RUR
Deficit to GDP = 4%
The numbers were chosen such that the debt stays constant as a fraction of GDP.
In 5 years time
NGDP would be 110.4 RUR
and debt would be 220.8 RUR
or still 2x
Debt service is 4.4 RUR
debt service to GDP is 4%
But if you ran with a 5% inflation rate, with real rates and real growth at 1%
NDGP = 133.8 RUR
Debt = 248.5 RUR
Debt / GDP = 1.86x it has fallen
But your deficit service would be 15 RUR
Debt service / GDP = 11%
2. March 2013 at 01:59
Scott, you are talking about the ex-post impact on macro stability by the central bank that is too hawkish, I am talking about the ex-ante impact on macro stability by the central bank that is doing optimal policy. You are talking about the net benefits of policy, I am concerned with the measuring of all the costs of the policy properly.
Consider a third world central bank that operates a currency peg. Central bank losses will increase the probability that the peg will be broken. This will be reflected in the financial markets.
Consider a first world central bank that operates a NGDP peg. Central bank losses will increase the probability that the peg will be broken. This will be reflected in the financial markets.
Of course, the central bank solvency costs of the latest QE are small, QE is a good program, albeit not an optimal one.
However, some people are unduly concerned with accounting measures of central bank solvency, and are ignorant about the economic measures of it. This imposes a de-facto solvency floor on the central bank activities, and as a result central banks are too cautious in expanding money supply temporarily. In September 2008 after Lehman, Bernanke was afraid to expand money supply temporarily because of the risk concerns, and he asked for TARP instead.
2. March 2013 at 09:02
TallDave, Remember, it’s not a zero sum game.
Oh, don’t get me wrong, I agree. Some have more to lose than others, though.
2. March 2013 at 12:30
Dr. Sumner:
“Doug, You are confusing real and nominal variables. Inflation doesn’t increase real debt service costs. This is just basic economics.”
This assumes that revenues rise equally with expenses.
But if inflation raises expenses more than it raises revenues, then it very much does raise real debt service costs, because debt service costs takes up a larger ratio of income.
Doug is I think suggesting that for individuals with high debt levels, inflation, if it does not affect their revenues and expenses equally, does not have a zero effect on real debt costs.
2. March 2013 at 12:32
The troubling thing is that this is not the first time that this non-linearity issue has been raised, and yet there is still accusations that people are conflating real and nominal variables when they bring this up, AFTER it was shown that there is no such confusion.
2. March 2013 at 15:29
Doug, You are confusing stocks and flows. You cannot divide nominal interest payments by NGDP and get a real interest burden. You need to divide real interest payments by RGDP.
123, Central bank losses would not lead to the peg being broken unless it led to fiscal dominace. But the sort of event that causes losses at the central bank actually helps the Treasury much more.
2. March 2013 at 16:02
Dr. Sumner:
“Doug, You are confusing stocks and flows. You cannot divide nominal interest payments by NGDP and get a real interest burden. You need to divide real interest payments by RGDP.”
Real interest rates are what money can buy given a rate of return.
RGDP is just historical NGDP.
Real interest rates divided by RGDP is a real quantity (goods) divided by a nominal quantity (historical NGDP).
The quantity would just be real goods relative to historical spending.
3. March 2013 at 04:16
“Central bank losses would not lead to the peg being broken unless it led to fiscal dominace.”
A rise in probability of fiscal dominance is a cost of QE.
“But the sort of event that causes losses at the central bank actually helps the Treasury much more.”
Two points. First, you can never be sure that Treasury will recapitalize the Fed. Consider the stupid debate in Britain about the cancellation of BoE bond portfolio. Second, you are talking about net benefits and costs, where there is no disagreement between us, while I am focusing on the measurement of each element of cost. Fed’s losses is a cost, as treasury would gain from a QE even more if the Fed held other assets that would appreciate instead of losing value.
Hall and Reis have a paper “Controlling Inflation and maintaining Central Bank Solvency under New-Style Central Banking”, which is excellent.
( http://erevents.frbsf.org/conferences/130301/papers/Hall-Reis.pdf )
QE should be defended the way Hall does, by arguing that the costs are small, and by pointing out that there are less costly ways to do QE: “The ECB has an important potential advantage over the Fed because it holds a substantial
fraction of its assets as collateralized loans to banks rather than as outright ownership of
bonds.”
15. December 2014 at 10:25
Manny U.
Does anybody know what a “balance sheet” effect is, in regards to buying MBS and Treasury securities by the FED? Most people I read about in this blog use terms, like QE, assets. And liabilities rather loosely. Was a success of the FED the implementation of the monetary policy of the QE for the past 6 years in the USA, as measured by the growth of GDP and reduction of both the traditional rate of unemployment and NAIRU? We begun at 5.8% in July 2008, rose to 10.2 in Oct 2009, and down again to 5.8% last month!. BB estimated in Dic 2008 the rate would be 6.5% in Dic 2013; the actual rate was 6.7% then. Was it close, or way off mark? Was BB’s QE program a success or just sheer luck? Is the Fed’s balance sheet dangerously QEed up?