If I buy T-bonds, their price rises. If the Fed buys T-bonds, their price (usually) falls
I notice that lots of commenters insist that bondholders gain when the Fed injects money by buying bonds. Even if this were true, it would have no bearing on my criticism of Richman. That’s because any effect on bond prices would be identical if the Fed injected money by paying government salaries in cash, rather than buying bonds. But there’s a much bigger problem with this fallacy. It’s unlikely that monetary injections would raise bond prices at all.
It’s always a mistake to use microeconomic intuition in macro. People think; “If I buy a few bonds the bond price will rise, so surely the laws of supply and demand tell us that if the Fed buys lots of bonds then bond prices will go up a lot.” Before telling you why this is false, let me provide some evidence:
During the first couple decades after WWII the Fed bought modest amounts of bonds. Then after 1965 bond purchases exploded, and the Fed absorbed vast quantities of T-securities. And yet between 1965 and 1981 we saw an epic collapse of bond prices. One of the worst bear markets in all of American history. Is there a cause an effect relationship?
Yes, there most certainly is. The huge bond purchases caused the monetary base growth to accelerate sharply. At the end of 1945 the base was $33.7 billion. At the end of 1965 it was $49.9 billion. But over the next 16 years the Fed bought bonds at a furious pace, and the base rose to $150 billion by the end of 1981. This explosion in base growth drove NGDP growth rates and inflation much higher, which raised nominal interest rates sharply. That’s why bond prices fell sharply.
Is there another explanation? Could it be big deficits that did the trick? No, debt as a share of GDP fell. The big deficits began after 1981, when Reagan took office. It’s inflation and NGDP growth that determines bond prices, not what the Fed buys.
Why is it different at the individual level? Because when I buy bonds I don’t increase the money supply. When the Fed buys them the money supply rises. Supply and demand still hold true, but we must think about two markets, bonds and money. It’s not like a microeconomic S&D problem. The value of money falls when the supply increases, at least in the long run. Expectations of a decline in money’s value reduce the expected real value of future interest and principal, which is paid out in nominal dollars. That’s why more money means lower bond prices, other things equal.
Do monetary injections always reduce bond prices? No, just most of the time. Obviously there are special cases that relate to how the injection changes expected future policy, and very small effects depending on which maturities are purchased. But the dominant effect is that more money means lower bond prices.
Bob Murphy concedes I was right when I criticized the mechanism by which Richman believed monetary injections would matter—who gets the purchasing power first. But he also argues that it matters what the Fed buys. I’ve always agreed with that. He also argues that one of my blog post titles was misleading:
The problem is, the market monetarists overstepped. Scott Sumner could have said, “I agree with Sheldon Richman, the central bank certainly can make some connected players wealthy at everyone else’s expense, for example by buying up toxic assets or taking over AIG. Clearly it matters very much where the new money is injected into the economy. However, I don’t at all like the mechanism Richman cited. He said it has to do with people getting the money at an earlier date than other people. That’s totally wrong, and here’s why…”
But no, Scott didn’t say that. Instead he wrote a post with the palpably false title, “It makes very little difference how new money is injected.” On a plain English reading of that sentence, it is false; Scott doesn’t believe it for one second. If the new money is injected into the Tomahawk missile sector versus the MBS sector versus the Treasury sector, of course it will make a huge difference. That’s why some of us flipped out and were asking Scott questions that must have struck him as an absurd waste of time.
I plead not guilty. As I made clear in my recent posts, I was holding fiscal policy constant. I still think that if the Congress doesn’t authorize different tax rates or spending levels on the basis of monetary injection, that is if monetary and fiscal policy are handled by different groups, then it matters very little whether the Fed buys T-bonds, or MBSs, or injects money into public employee salaries. And I think that’s a very reasonable thing to assume for two reasons:
1. Over the years my comment section has been loaded with misconceptions on this very point. In the real world the Fed agonizes over whether to spend the new money on MBSs, discount loans, T-bonds, etc. Bernanke doesn’t agonize over whether to spend the new money on Tomahawk missiles. So I am dealing with monetary policy as it is, not some alternative world where the Fed might buy a trillion dollars in bananas, and drive up banana prices. It’s so obvious that fiscal policy can have sectoral impacts that I saw no need to insult readers’ intelligence by stating the obvious.
2. Many readers argue the Fed purchases enable the spendthrift Congress by holding down rates. The problem here is that an even tighter monetary policy would push us toward the Japanese scenario, with even lower long term rates. Any enabling comes via stronger NGDP growth, not lower rates. If the only way to stop Washington from spending is to drive the economy into such a deep depression that Washington has to spend less, then I’m not on board. And in any case, how’d that “tight money saves us from big government” work out for the conservatives who ran monetary policy in the early 1930s?
Perhaps my previous post was too pessimistic about influence. Here’s commenter MikeF:
When I first heard this Cantillon argument I actually though it made sense (I didn’t think about it very hard) but it’s pretty obvious that Scott Sumner is right about this. I wish Austrians would do some self-assessment and realize that the reason they haven’t done anything important in half a century is that they can’t seem to bring themselves to let go of their dogma even when it is essentially proven to be nonsense. If this were a worthwhile economic movement, there would be some adults in the school somewhere who would stand up and say “alright, alright, look, we thought this at one point but it turns out to not be correct, or circumstances changed or whatever, now we need to move on.” No science has ever moved forward without admitting it was wrong in the past. I’m saying this sincerely as someone who wishes there were a decentralized, free-market, school of macroeconomic thought that someone could take seriously.
Time for me to watch It’s a Wonderful Life one more time?
PS. I’m certainly not suggesting all Austrian economists are wrong about Cantillon effects—I am reacting to the Austrians that comment over here, and some media commentary I’ve read. Although Milton Friedman was my favorite economist, I was wrong in initially supporting his money targeting idea (when I was in grad school.) All schools of though need to constantly re-invent themselves. Even MM.
PPS: For instance, here’s an Austrian economist who does understand Cantillon effect.
The Austrian understanding of injection effects cannot be separated from its understanding of the price system and capital theory. It doesn’t matter so much WHO gets the money first. What matters is that SOMEONE does and that the dispersion across the price system is not even. THAT matters because it ripples through the capital structure with real, not just nominal, effects. The Austrian theory of the business cycle is one kind of systematic real effect, but there can be others.
HT: Greg Ransom
Tags:
7. December 2012 at 07:33
Scott,
“When the Fed buys them (UST’s) the money supply rises”
Is the money supply really increasing when banks are just swapping UST’s for reserves at the Fed? QE does not seem to be increasing the money supply, just changing the mix of private sector assets (UST’s vs. reserves). What am I misssing here?
7. December 2012 at 07:35
I’d like a simple explanation of the gains of trade and opportunity cost of Fed transactions. To me it’s a nonsense statement that since the bank and Fed are making a market exchange (eg, $100 value bonds for $100 cash), each side is no better off after the exchange.
Of course, the truth in any market exchange is that both sides are better off.* The exchange has to hold enough gains to overcome our indifference and lead us to trade in the first place.
* Well, I say both sides, but in this case, only one side of the trade is engaging in a market exchange because the Fed is trading a good for which it faces no scarcity and receiving a good for which it has no use, thus, it’s hard to categorize it as really being a market participant.
In any case, yes isn’t there a difference in the folks actually engaging in the trade and the folks just sitting around and having something happen to their asset values? Of course there is, or there wouldn’t be a trade in the first place.
7. December 2012 at 07:48
Sumner writes: “It’s always a mistake to use microeconomic intuition in macro. […] Before telling you why this is false, let me provide some evidence.”
What a perfect summary of the debate. The internet Austrians (including our local troll MF) have made it clear that they only value “logical proofs” in their economics (which in practice means uncritically extending micro to macro, as well as ignoring the fallacy of composition), and they have no interest in real world data.
No wonder there can be no agreement.
7. December 2012 at 07:54
Can I confirm my understanding is correct?
– The whole point of monetary policy is to adjust the amount of money and the value of money to its optimal level
– Almost any method of changing the money supply has some distributional side effects. These affects may be unintentional and/or undesired. In any case they would be categorized as fiscal policy.
– Govts use fiscal policy on a huge scale to carry out distributional effects and this dwarfs any fiscal side- effect of monetary policy. The govt if it chose to could use fiscal policy to back-out any unintended side-effects of monetary policy.
– Once all this is understood then one is left with a clear view that the govt can get the money supply to any level it wants and minimize the distributional effects of doing this.
– Therefore we should really be discussing either what the optimum level of the money supply is to optimize the economy or we should be discussing how fiscal policy can be used to benefit society. We should not confuse these 2 discussions
7. December 2012 at 08:02
Let me first say that I agree with your conclusion that it really doesn’t matter who gets the money. However, I don’t find this argument completely convincing. I hadn’t thought about the mechanism you outline for bond prices. I find it interesting, but I can’t imagine it would dominate in the short run. Fed funds and 3-month t-bills are highly correlated for example. I could conceive an impulse response of t-bills to expansionary fed action starting out positive and asymptoting to something negative, but that doesn’t imply that in the short run bond holders couldn’t benefit. Of course , if you are right about the long run decrease, to keep that gain they would have to move out of bonds eventually to something else.
7. December 2012 at 08:03
dis737, The monetary base increases. M1 may or may not rise.
Ron, You said;
Almost any method of changing the money supply has some distributional side effects. These affects may be unintentional and/or undesired. In any case they would be categorized as fiscal policy.”
Increasing the money supply has distributional effects even if it is done by buying T-bonds. But I consider that monetary policy, not fiscal policy. If the distributional effects of other methods are vastly different from bond buying, that is likely to be due to fiscal aspects of the policy. But in the real world the various methods chosen by actual monetary policy makers have roughly the same distributional effects. So switching from method A to method B has little addition distributional effects, over the basic effect of increasing M and inflation.
Obviously if they did something crazy like buying bananas the distributional effects would suddenly become much different.
7. December 2012 at 08:04
did any austrian suggest that cantillion effects dominate?
obviously when supply increases, prices fall.
7. December 2012 at 08:08
Sheldon Richman has already said directly that what you said he had in mind is *not* what he had in mind.
There is no evidence that you’re account of what Richman hand in mind with his expressive language is what Richman actually had in mind, indeed, there is no reason to believe it, because before using the public consumption language he uses Richman invoke economists who don’t have in mind what account you put in Richman’s mind.
Scott writes,
“the mechanism by which Richman believed monetary injections would matter”
Scott, have no evidence that the mechanism you invented and ascribed to Richman ever existed anywhere but in your own mind.
7. December 2012 at 08:12
Scott, if the government allows everyone to buy long term government bonds as an alternative to whatever taxes they owe, is that monetary policy or fiscal policy?
Would that have any effect on the coordination structure of the economy across time involving alternative production processes and consumption pathways?
7. December 2012 at 08:14
James, If that were true there’d be lots on $100 bills lying on the sidewalk waiting to get picked up.
Greg, I’m not a mind-reader. All I can do is respond to what people write.
7. December 2012 at 08:22
Scott,
Asset purchases are a one-off for the Fed, and carry no further obligation or impact — they buy at market prices. The Fed’s true cash creation obligation is felt on the matching liability side, where the Fed creates/owes excess reserves (to buy the assets) — these are long lived, and this is “where the cash is”. This is why the monetary base is important! The asset purchase side is a red herring.
Banks have so far gotten this cash/capital/reserves — just as they did in the early 1930s. I’d argue that the Fed concentrate purchases in Treasury markets in order to *not* have an impact on secondary market prices — its the biggest and most liquid market out there — they just need some asset to hold. The real cash delivery obligation is on their liability side, and their real concern is the banking system.
7. December 2012 at 08:30
Scott, I just want to get you to understand why I was asking you such apparently sophomoric questions in the original threads. Here you wrote:
But there’s a much bigger problem with this fallacy. It’s unlikely that monetary injections would raise bond prices at all.
OK, so this is why I was asking you stuff like, “Can the Fed, even during normal times when we’re not up against the ZLB, ‘cut interest rates’ like the intro textbooks and the press and central bankers all seem to think?”
It sounds to me like you’re saying no, the Fed can’t reduce short-term interest rates, not even in the short run, as a general rule.
OK, that’s a coherent position to hold, I just disagree strongly with you, and I think so would most other economists.
If that’s *not* what you’re saying, then the statement I quoted above isn’t what you really believe, right?
(Again, not trying to do a “gotcha” here, I really keep updating my mental model of what your views are.)
7. December 2012 at 08:43
Bob M —
I’d say its because the Fed’s real power is base money. They have taken this power and used it to traditionally dominate the short-term money market via regular intervention, and later worked through expectations. In any interest rate market, however, all the Fed can do is try to match the natural rate of interest that their monetary base policies create — which has been tight money/low yield.
The true money policy creation takes place in the monetary base — the liability of the Fed — in currency- and reserves-. They may go out and buy Treasury bonds, but if the parallel growth in the monetary base side is sufficient, then yields will rise. If they go out and sell Treasuries and reduce the monetary base in parallel, then yields might very well fall! As Scott says, buying the bonds has a trivial impact — the impact is felt in the monetary base (which is the mirror image of Fed assets).
7. December 2012 at 08:52
Bob, that is exactly what I am most interest to here about. Treasuries don’t rally in expectation of Fed easing? Really?
7. December 2012 at 09:00
James — think of it in terms of 1/NGDP is set by monetary base policies. If money is tight, long end yields anticipate lower NGDP, for instance, the curve inverts, and then the Fed brings the short end down to the natural rate of interest. This natural rate is set by how much base money they have produced — and what the long end market has priced in.
Interest rates are the tail that is wagged by the monetary base dog via NGDP. Setting Fed Funds is not the Fed’s main power.
7. December 2012 at 09:04
“That’s because any effect on bond prices would be identical if the Fed injected money by paying government salaries in cash, rather than buying bonds.”
Really? Printed money, with no asset, even if it is a loan from the goverment, behind it has the same impact as money created through asset purchses? When I asked that question in college 20 years ago, the grad student advisor said that only the poorest African countries print money when they need it.
” As I made clear in my recent posts, I was holding fiscal policy constant.”
What do you mean here? It sounds like you are saying that as pretend central banker, that you are independent from the pretend fiscal autorities. But, in the previous thread you suggested that, in senario 1, all stimulus will be monetary and in scenario 2 all stimulus will be fiscal, but the total level of stimulus will be the same.
7. December 2012 at 09:10
We really, really, need to understand the structure of the Fed balance sheet. People seem to think that the Fed creates money by buying assets. They don’t. They create money by creating liabilities — currency- and reserves-.
7. December 2012 at 09:12
Look what’s back;
http://www.washingtonpost.com/business/economy/could-the-platinum-coin-option-solve-the-us-debt-crisis/2012/12/06/d6dc7956-3fe5-11e2-ae43-cf491b837f7b_story.html?tid=pm_pop
‘If President Obama wants to avoid an economic calamity next year, he could always show up at a news conference bearing two shiny platinum coins, each worth .”‰.”‰. $1 trillion.
‘Okay, that sounds utterly insane. But some economists and legal scholars have suggested that the “platinum coin option” is one way to defuse a crisis if Congress cannot or will not lift the debt ceiling soon. At least in theory.’
7. December 2012 at 09:14
I’m glad you mentioned that Horowitz post. I thought it was far clearer than anything else that had been written on this subject.
7. December 2012 at 09:21
Basic background on Cantillon Effects and trade cycle science from David Glasner:
http://uneasymoney.com/2012/12/06/those-dreaded-cantillon-effects/
Glanser points out that the simple minded views that Scott Sumner has uncharitably put in the mind of Sheldon Richman are trivialities that have nothing to do with the processes and effects invoked by Richman, ie that are trivialities that have nothing to do with what Richman had in mind, and attempted to communicate to a popular audience with compelling but non-technical metaphorical language.
David Glasner,
“What accounts for the difference between the empirically rich theory of systematic Cantillon Effects articulated by Hayek over 80 years ago and the empirically trivial version on which so much energy was expended over the past few days on the blogosphere?”
More Glasner,
“t seems to me that the reason Cantillon effects were thought to be of import by the early Austrian theorists like Hayek was that they had a systematic theory of the distribution or the incidence of those effects. Merely to point out that such effects exist and redound to the benefits of some lucky individuals would have been considered a rather trivial and pointless exercise by Hayek.”
Glasner takes the time to elaborate a number of central points I’ve made here before in very quick language.
7. December 2012 at 09:26
Traditionally the Fed targets one interest rate, and the market sets the rest. When the Fed is buying bills, more Fed purchases could drive bond prices higher or it could drive them lower.
The 30 day rate is approximately equal to the expectation of the average overnight rate over the next 30 days. The 60 day rate is the average between the 30 day rate and the expected overnight rate in the 30 day period beginning 30 days from now. Using this method we can “boot strap” the forward curves.
If the market is efficient, an steep yield curve suggests that the Fed will be raising short-term rates some time in the intermediate term. If the fed is too easy now, it will just have to raise rates that much more in the future. So, yes easier policy now may cause longer term rates to rise. I have certainly seen the long bond go both the direction of the short-rate move and the opposite direction of the short-rate move on day when the Fed has changed policies.
Now the Fed is buying longer-dated Treasuries, MBS, etc. The Fed has sufficient resources that it can push any rate it wants to to any level it wants to. If the Fed wants to drive long-rates lower, even if the market sees that as inflationary, long-rates are going to be lower.
Is this good policy to drive long-rates lower? What is the rationale behind “Operation Twist”? The rationale is that the Fed is trying to crowd out investors from these relatively low risk intermediate Treasuries into higher risk holdings, which will ultimately drive down risk premia across the board. Is that what we want?
If we want to expand the money supply and grow GDP, we want banks to underwrite new loans. But if they get paid by the Fed to sit on excess reserves, and are not going to be paid to take risk, then OT has a negative effect.
7. December 2012 at 09:31
“It’s unlikely that monetary injections would raise bond prices at all”.
This again? Monetary expansion does not equal inflation. Krugman addressed this in 2009. Inflation affects bond prices, monetary expansion does not.
“I still think that if the Congress doesn’t authorize different tax rates or spending levels on the basis of monetary injection, that is if monetary and fiscal policy are handled by different groups, then it matters very little whether the Fed buys T-bonds, or MBSs, or injects money into public employee salaries”.
I don’t understand why you insist that your arguments are based on the real world one paragraph after saying this. Burns, Volcker, Greenspan… were these guys independent of the government or politics? Hardly. Is Bernanke? You decide.
7. December 2012 at 09:31
“People seem to think that the Fed creates money by buying assets. They don’t. They create money by creating liabilities.”
One person’s liability is another person’s asset. That is double book entry accounting.
7. December 2012 at 09:34
If the economy was in equilibrium there would an interest rate that would reflect that equilibrium and allow investment and consumption spending to be in the correct ratio.balance.
If money is over-valued and prices are sticky then spending will be depressed. The fed would buy T-bills which in the short term may well push interest rates down (indeed in normal times the fed policy may well be to target a lower interest rate). As the feds t-bill purchases address the over-valuation of money spending will return to normal and interest rates actually may rise back to the equilibrium despite the feds policy to lower them.
If at that point the fed continued to target the low rate and pursued that policy by buying more t-bills they would increase the money supply sufficiently to generate inflation concerns. This would push the price of t-bills down despite the fed buying them and yields up.
At the zero-bound the fed cannot meaningfully target an interest-rate. It can however continue to buy t-bills and increase the money supply until money is correctly valued. At that point interest rates would again be at their equilibrium level.
How will the fed know when money is correctly valued and it can stop increasing the money supply ? Seems like an NGDP target would provide a good rule of thumb.
Whats more: Controlling the money supply by buying interest bearing assets is self-regulating. When the money supply is at the optimum level the price and yield of these assets will be set independently of any fed buying or selling. This will minimize any “Cantillon effects”.
7. December 2012 at 09:35
If you can’t read Horwitz, Glasner, Boettke et al and not see that you are boxing your own shadow, Scott, and you are not addressing or engaging with a real position held by anyone, why shouldn’t we all conclude, that you are *very happy* to be misrepresenting & inventing bogus accounts of the causal conceptions and mechanisms of a scientific explanatory rival, your whole desire is to be boxing your own shadow, making marginalizing rhetorical gains with red herrings, straw men, and other informal argumentative fallacies.
7. December 2012 at 09:35
Doug M — forward rate curves tend to be an artifact of discounting money, and have no relationship to the realized future path of interest rates.
Try the exercise of cutting reserves out of the monetary base picture entirely — all currency. When the Fed buys Treasurys, it pays you in cash. As an individual, you have simply gotten market value for your asset. In aggregate, there is a ton more cash floating around. Printing this currency is what made Zimbabwe’s NGDP skyrocket. Tweaking the Treasury yield — buying your 10y rather than Joe’s 2y — is a trivial effect in comparison.
7. December 2012 at 09:37
Greg,
David Glasner most emphatically did not imply that Scott Sumner was simple minded; he was placing the dialogue into a historical context, which is a big part of what he does so well with these discussions. You provide no one any charitable help or otherwise by your implication.
7. December 2012 at 09:40
Doug M —
The Fed’s liability is your asset (currency). The Fed’s asset is your liability (treasurys).
Which constitutes money printing for you and your operating in the economy?
7. December 2012 at 09:47
jknarr, That’s right.
Bob, The Fed can obviously cut T-bill yields, but easy money usually raises T-bond prices. As a whole, easy money hurts owners of Treasury debt.
Doug, The impact on debt held by the public is the same either way–think about it.
Tyler, Surely you don’t believe that Congress will set fiscal policy differently if the Fed buys T-bonds, as compared to the Fed buying MBSs. Obviously monetary policy per se can influence fiscal policy, as I pointed out in my Great Depression example. But that’s not what’s being debated he. It’s a question of whether it matters which asset the Fed buys. For plausible assets the answer is “hardly at all.” When Bernanke starts DIRECTLY buying Tomahawk missiles, please let me know.
7. December 2012 at 09:48
Conservatives have a natural affinity for NGDPLT for many reasons, one of them is:
Given $17T in debt, the best way to shrink government spending in the near term is to get interest rates increasing.
So if the target will be 4.5%, with a couple points of make-up… the faster we get the make-up done, the faster we can see rates rise.
Until interest rates rise, the stretched rubber band can’t snap back.
7. December 2012 at 09:49
Becky, I could see Greg was playing tricks again. He loves to misquote–did so just yesterday. The “lucky individual” argument is the same point I made, so even without reading Glasner it was obvious that Greg was once again being deceptive.
7. December 2012 at 09:54
Scott — I don’t play the deception game.
7. December 2012 at 09:55
“If the market is efficient, an steep yield curve suggests that the Fed will be raising short-term rates some time in the intermediate term. If the fed is too easy now, it will just have to raise rates that much more in the future. So, yes easier policy now may cause longer term rates to rise. I have certainly seen the long bond go both the direction of the short-rate move and the opposite direction of the short-rate move on day when the Fed has changed policies.”
Thanks Doug, this what I essentially what I had in mind above when talking about the impulse response. I need to be careful about specifying maturity length. Above I intended short run only.
7. December 2012 at 09:58
Scott, I don’t like how you repeatedly say, “This is Joe Blows position”, when it’s not Joe Blows position, its your uncharitable invention of what Joe Blow is talking about, made possible by your refusal to be conversant in the background economics that Joe Blow is talking about.
You can call that sort of behavior anything you like, but it only leads to days of talk at cross purposes about trivialities, as Glasner, Boettke and others have pointed out.
7. December 2012 at 10:04
Imagine a neuroscience blogger with a rival explanatory model to Gerald Edelman repeatedly saying on his blog, “This is what Gerald Edelman believes” and what is then said gets Edelman all wrong and really doesn’t have anything to do with Edelman’s explanatory mechanism. And then that blogger goes on to “refute” “Edelman’s arguments” when the thing refuted has nothing to do with Edelman’s actual explanatory mechanism?
And what if this went on for years?
It would be a damn embarrassment.
7. December 2012 at 10:13
“Tyler, Surely you don’t believe that Congress will set fiscal policy differently if the Fed buys T-bonds, as compared to the Fed buying MBSs. Obviously monetary policy per se can influence fiscal policy, as I pointed out in my Great Depression example. But that’s not what’s being debated he. It’s a question of whether it matters which asset the Fed buys. For plausible assets the answer is “hardly at all.” When Bernanke starts DIRECTLY buying Tomahawk missiles, please let me know”.
Well, no I don’t think that Congress will necessarily set fiscal policy differently is the Fed buy’s MBSs versus T-Bonds, although I wouldn’t totally rule it out.
Does it matter which (plausible) asset the Fed buys? The Fed apparently thinks so. Also, and I’m going out on a limb here somewhat, I do think it matters whether the Fed purchases Treasuries versus MBSs, because it affects who the bondholder ends up being, even holding fiscal policy constant.
If the government is going to issue X number of bonds, I think it benefits them to have the Fed holding those bonds rather than, say, China. The Fed, independent or not, is working from a similar set of priorities as our government. China has its own interests to think about. Although it would hurt them terribly to do so, China could dump its trillion in US bonds and do untold damage to our economy. If they decide to pursue their interests more aggressively in the South China Sea, I believe there is not a whole lot we could do about it.
7. December 2012 at 10:13
So . . . if the Fed wants to increase interest rates it should buy bonds and if it wants to decrease interest rates it should sell them?
Really?
7. December 2012 at 10:15
I’ve said again and again that I don’t it as a charitable reading to ascribe the “lucky individual” assumption to Richman, not if we are to credit in wave of the hand to Austrian non-neutral monetary economics.
Scott writes,
“”The “lucky individual” argument is the same point I made, so even without reading Glasner it was obvious that Greg was once again being deceptive.”
I have no idea what I am suppose to have been deceptive about.
7. December 2012 at 10:16
The people at the fed disagre with you, Scott. They buy t-bonds explicitly to lower the interest rates (raise the prices).
7. December 2012 at 10:16
I’ve said again and again that I don’t *see* it as a charitable reading to ascribe the “lucky individual” assumption to Richman, not if we are to credit in wave of the hand to Austrian non-neutral monetary eco
7. December 2012 at 10:19
Didn’t Glasner come out and say that he agreed with Scott against Greg? Why are we still having this argument? Even MoA-MoE didn’t last that long (and Scott was actually wrong about that one).
7. December 2012 at 10:22
bmcburney, no time to go through the whole history of macro embedded in your fallacy there, but the Fed lowers real rates through the liquidity effect from expansionary policy, but also raises them by improving monetary policy so that real growth expectations rise. And nominal rates rise with the Fisher effect when more inflation is expected. In the long run the Fed only affects nominal rates, unless it screws up.
7. December 2012 at 10:23
Greg — I’m part Austrian school myself, I have a deep and abiding love for Hayek. Debt extinguishes and deflates, malinvestment abides and free markets are the best. We clearly need to have an understanding about how the Fed’s monetary plumbing operates before we can have a productive discussion of Austrian debt cycles. In fairness, I also sense some charicaturization of Austrian school here (maybe a bad experience with the Sound of Music). But can we call it a day on the back-and-forth? Agree to disagree and all that and get our focus back on understanding the plumbing?
7. December 2012 at 10:23
Doc, see above. Surely you know better than to make such dumb “criticisms”. Scott must be tired by now of responding to people who are bent on misinterpreting him.
7. December 2012 at 10:24
I do think that raises an interesting discussion though, DocMerlin. Why does monetary injection not equal to inflation? RDGP growth, IOR policies, foreign holdings of US dollars, regulatory restrictions on lending, etc. etc.
So, if I understand it correctly, the Fed’s purchases with the intent to lower interest rates, in the end, are successful because of IOR, regulatory uncertainty, etc. stop that money from entering the system and increasing inflation (thereby lowering bond prices and raising rates back up). And yet the whole point of the lower interest rates is to spur spending by making the cost of capital cheaper and the cost of saving higher.
Am I missing something?
7. December 2012 at 10:24
[…] example, in his latest salvo Scott implicitly says I am wrong on this and a child (but he quotes someone else so the sting […]
7. December 2012 at 10:28
For those who care, here is my latest (and last!) post on this topic.
7. December 2012 at 10:29
Folks, again:
Try the exercise of cutting reserves out of the monetary base picture entirely “” all currency. When the Fed buys Treasurys, it pays you in cash.
As an individual, you have simply gotten market value for your asset. In aggregate, there is a ton more cash floating around.
Printing this currency is what made Zimbabwe’s NGDP skyrocket.
Tweaking the Treasury yield “” buying your 10y* rather than Joe’s 2y “” is a trivial effect in comparison. *Or buying mortgages instead of Treasurys.
bmcburney — yes, buying Treasury bonds will produce higher yields if it is truly stimulative.
DocMerlin — no, they buy Treasuries to expand the monetary base, not lower rates.
7. December 2012 at 10:33
Scott wrote:
Bob, The Fed can obviously cut T-bill yields, but easy money usually raises T-bond prices.
That’s a typo, right Scott? You meant to say “…but easy money usually REDUCES T-bond prices”?
And if so, I am still stumped. Can someone else explain what Scott is trying to say? How can the Fed “obviously” cut T-bill yields, if easy money usually makes T-bond prices go down? I don’t understand what the claim is.
7. December 2012 at 10:33
Ohhh it’s just a T-bill versus T-bond distinction? Short- versus long-term? I am fine with that, if that’s what you’re saying Scott.
7. December 2012 at 10:40
Bob Murphy —
The monetary base drives NGDP levels, interest rates simply discount the NGDP path — this includes the short end.
Low interest rates signify tight money, high interest rates signify loose money.
7. December 2012 at 10:40
Boy, this article looks quite prescient in December of 2012.
http://traderscrucible.com/2011/08/08/godzillas-bond-rally-begins-qe-ii-as-cash-for-clunkers-update/
7. December 2012 at 10:45
Bob,
It took me a while to come to that conclusion as well. What confused me was that there was no mention of maturity, so I assumed he meant short as well.
7. December 2012 at 10:46
Greg, this is a real position held by some people. The whole “The Ben Bernank” video was about this.
7. December 2012 at 10:53
Doc Merlin,
I agree. QE causes a “cash for clunkers” effect in the Treasury markets, pulling forward sales of bonds while there is a known buyer.
This is why we’ve seen absolutely massive Treasury rallies starting almost to the day when QE I and II ended.
I talked about it in detail over at the traderscrucible.com in real time, as it was happening, in 2011.
My first post is from March of 2011, where I predict 2% long bond yields in June 2012 based on this model. We almost got there, down to 2.4%.
The fed buying massive amounts of bonds lowers bond yields.
http://traderscrucible.com/2011/03/11/qe-ii-the-feds-cash-for-clunkers-program-and-why-the-end-of-qeii-means-the-godzilla-of-all-bond-rallies/
7. December 2012 at 10:54
Everyone, Not one of you has given an alternative explanation for why T-bond yields soared between 1965-81. You can’t beat something with nothing. Please offer at least one alternative explanation (to easy money) so I can have fun mercilessly ridiculing it.
Tyler, No, it would not hurt the US if the Chinese dumped a trillion in bonds on to the market. If anthing it would stimulate AD, or maybe have no effect if the Fed neutralized it.
bmcburney, Yes, really.
Doc, They buy them to lower real rates. Nominal long term bond yields rose immediately after QE3. Even medium term bond yields often rise after monetary stimulus (not always). I’ve done many blog posts on this subject.
Saturos, I was tired of responding after one month of blogging.
Bob, Yes, that was a typo, I meant T-bond yields rise with easy money.
Glad to hear I’ll have the last word. 🙂
7. December 2012 at 10:57
Mike, When rates fall to zero there is more demand for ERs. Don’t confuse cause and effect.
7. December 2012 at 11:23
[…] example, in his latest salvo Scott implicitly says I am wrong on this and a child (but he quotes someone else so the sting […]
7. December 2012 at 11:23
Tyler – I might add that China has abetted the US in tight money policies – they receive USD for goods, and instead of demanding currency or produced goods back (and sending the USD back), they pushed it into Treasury bonds. Our NGDP is lower as a consequence, and therefore our yields are lower.
7. December 2012 at 11:34
“Everyone, Not one of you has given an alternative explanation for why T-bond yields soared between 1965-81. You can’t beat something with nothing. Please offer at least one alternative explanation (to easy money) so I can have fun mercilessly ridiculing it”.
You did provide the correct explanation, you just incorrectly placed two separate price pressures on the same continuum. Inflation causes nominal interest rates to rise, which in turn causes bond prices to fall, and yields to rise. Fed purchases (or any purchases) of bonds causes prices to rise, and yields to fall. In the 70s, the inflationary pressure outweighed the Fed’s purchases. Today, there is no inflation, and the Fed’s purchases have indeed caused yields to fall. Repeatedly and reliably.
You continue to incorrectly equate monetary expansion with inflation, which explains your incorrect reasoning. Today’s monetary expansion has not led to inflation, for a variety of reasons, and inflation (not monetary injections) causes interest rates to rise.
“Tyler, No, it would not hurt the US if the Chinese dumped a trillion in bonds on to the market. If anthing it would stimulate AD, or maybe have no effect if the Fed neutralized it.”
This is very basic finance. A huge amount of selling would lower treasury prices and raise yields, which would make it ruinously expensive to fund our deficits and would negatively affect our creditworthiness. Even if the Fed neutralized it, which would require nearly doubling their current holding of US bonds, the associated market movements would make the crash of 1929 look like a minor correction. You have this weird thing about dismissing supply and demand, which I don’t really understand.
Even using your own logic, why would it stimulate AD if a trillion in cash left the system and was replaced with bonds?
7. December 2012 at 11:40
jknarr: “Tyler – I might add that China has abetted the US in tight money policies – they receive USD for goods, and instead of demanding currency or produced goods back (and sending the USD back), they pushed it into Treasury bonds. Our NGDP is lower as a consequence, and therefore our yields are lower.”
I agree, at least if I’m reading that correctly. China bought dollars and sold yuan to keep their exchange rates artificially low, and had to do something with those dollars. I agree that it had the effect of lowering our yields, and in that sense they have willingly been something of a debt drug dealer. That’s why I said it would hurt them terribly to dump our debt – they’ve been profiting from our profligate spending, and presumably they want to continue doing so. And the value of their trillion dollar investment would plummet long before they could unload most of it.
7. December 2012 at 11:45
Tyler — how about this:
Let the Chinese sell all their bonds at lower and lower prices — buy high sell low. Classic!
The Fed can simply fund the Treasury as-needed issuance.
I’d love to buy bonds at Chinese firesale prices, won’t make any difference to funding the US Treasury.
In fact, why should the Fed take the Chinese out of their bad positions? Are you asking for a Maiden Hell III bailout vehicle?
7. December 2012 at 11:49
Tyler — actually, the Chinese yield vector is not: buying Treasury paper and lowering yields — it is exporting to the US and not buying back (running a trade surplus). Check the GDP stats — net deficit exports subtract. This drives NGDP lower, and thereby reduces yields.
7. December 2012 at 11:58
OK, I’m a micro guy and I agree with your overall sentiment about applying micro principles at a macro level, which I am finding very interesting. I think I also agree with your point about buying bonds lowering the price, at least in the long run. And I think I get your general theory about tight money and low interest rates (it seems to me, this bond price example is basically that in a nutshell). But I have to issues I’m trying to resolve.
1. While it makes perfect sense to say that a consistent policy of “tight money” over a decade would lead to higher bond prices by lowerin inflation expectations and thereby nominal interest rates, it still seems like I commonly hear headlines like “the FED announced a new wave of asset purchases today. Yields on one year notes fell two basis points on the news.” Am I imagining this (quite possible, I don’t take careful notes on these things)? Or is it possible that the immediate effect is contrary to the long-term effect? Seems like this would open the door to arbitrage but I haven’t thought it through that far yet.
2. One summer they got me to teach intermediate macro. In the course of covering IS/LM and AS/AD it was necessary to explain the mechanism by which the FED affects interest rates by changing the money supply and the explanation I gave was essentially the one you are debunking here. It seems like that’s what the textbook said, though maybe I made it up, I don’t know. At any rate, Keynesian economics seems to hinge on this relationship between the money supply and the interest rate. I know you’re not a Keynesian (nor am I) but does your interpretation of the relationship between OMOs and interest rates throw all of that IS/LM stuff out the window or is there a way to square the two that I’m not noticing?
[Adressed to Scott Sumner but obviously there might be a lot in there to reply to and he’s a busy man so have at it anyone who thinks they know.]
7. December 2012 at 12:06
Well, if we could get away with buying back all our debt at firesale prices, I sure wouldn’t complain! My concern is that the mere act of dumping their holdings would cause the bond markets as a whole to question their previous assessment of our debt as relatively riskless.
Anyways, that’s not really an immediate concern, and I apologize if I made it sound as if it should be. I was just pointing out that it does matter who our creditor is, and if people always did what was in their best interest, we would have a lot fewer wars.
To your second point, I think that’s somewhat semantic. A favorable exchange rate makes their products more attractive and ours less, thus setting the stage for the net deficit import-export relationship we have with them. Chicken and egg. I didn’t say they cared about lowering our yields, just that our demand for loans nicely meets their need for favorable exchange rates.
7. December 2012 at 12:10
Mike, I’m here so I will gratuitously respond —
1) “It is well enough that people of the nation do not understand our banking and monetary system, for if they did, I believe there would be a revolution before tomorrow morning.”
Most commentators have no idea how it really works, and so constantly confuse causes and effects.
Advice: don’t listen to any “because” causal statements in the media — they typically don’t have any idea and are parroting the line, or have an agenda. Log the fact and connect occurrences yourself.
2) I’m not sure what you are asking.
7. December 2012 at 12:15
If Scott had used the simple word “if” as in “If Richmand had in mind X, Y, Z ” then the conversation could have been brought to a conclusion with “No, Scott, that’s not what he had in mind, and if your are interested in what Richman had in mind, go read x, y & z, not inerested, and, by the way, this might bring an end to your shadow boxing with yourself sessions.”
7. December 2012 at 12:16
Mike:
1# – Headlines don’t mean much, when it comes to that. Market movements based on Fed announcements are a reflection of market expectations versus reality. So when Apple is predicted to have sold a gajillion Ipods for the year, and it turns out they only sold a bazillion, the stock falls on the news. Same thing happens to bond prices.
2# – Scott might not agree with my characterization, but in my opinion he is not disputing that rising nominal interest rates lowers bond prices (he says as much above). What he is doing is equating monetary expansion with inflation, which is quite understandable as more money in the system often leads to more dollars chasing the same amount of goods, hence inflation.
In a theoretical environment, OMOs in which the Fed buys treasuries increases inflation and the expectations of same, offsetting the S&D pressure produced by the purchases.
But! As we can see in the current environment, sometimes things interfere with the injected money actually entering into the system, such as the policy of paying interest on reserves, in which case you can have monetary expansion without inflation.
7. December 2012 at 12:16
Tyler: the Chinese would also sell bonds in exchange for currency — strengthening the USD. And what would they do with this cash? They are right back in the export surplus boat where they first started. Maybe they should just burn it so that we can print more.
If they really wanted to hurt the US, they would take the cash to buy gold and then suddenly peg the yuan to an ounce, redeemable anytime, anywhere, no questions. That would make USD an inferior FX asset and cause the funding problems you describe.
7. December 2012 at 12:25
jknarr: Presumably they would turn around and sell the USD they received from selling the bonds, in return for a variety of other currencies, in which case the dollar’s value would plummet and our yields would still be sky high.
Again, that would leave them with a big loss on the trade and probably would still leave them with a more expensive yuan, neither of which they want. But, in my opinion, the damage they suffer would be much less than they inflicted on us.
Also, we’re derailing the thread big time!
7. December 2012 at 12:29
Prof. Sumner,
I have one question relating to the Fed buying bonds. If the Fed buys bonds, the Fed is giving someone new cash for a bond they held, right? Now, if I used to hold a bond and I now have cash, wouldn’t I be much more likely to use the new money to buy other assets. For example, if the Fed gave cash to a bunch of government workers directly; wouldn’t they be much more likely to use the money on goods and services than giving it to a former bondholder. This question has nothing to do with the price of bonds, it has everything to do with what path the money will go once it’s in the hands of a certain person. The question itself might be unrelated to the point that you’re getting at, but I’m just wondering about your opinion on this question.
7. December 2012 at 12:31
“Everyone, Not one of you has given an alternative explanation for why T-bond yields soared between 1965-81. You can’t beat something with nothing. Please offer at least one alternative explanation (to easy money) so I can have fun mercilessly ridiculing it.”
I was not an active Fed watcher in the 1960s as I had not yet been born.
I have very little data aviable to me that predates the 1980s.
But, what is taught in school…. Lyndon Johnson fights two wars — The war in Vietnam and the War on Poverty.
The OPEC cartel drives up the price of oil.
I would say that without easy money an oil price shock should be a temporary increase in inflation, rather than the systemic inflation that it became.
Looking at a shorter term timeframe, long-bond yeilds are more likely to fall (prices rise) in the day and week following a Fed move, regardless of whether the Fed raises or lowers the Fed Funds rate. Largely this is because the bond price has already “priced in” Fed action.
The times when the fed has seemed to catch the maket by suprise – big reactions on the long end, the long end moves the opposite direction from the short end.
7. December 2012 at 12:36
Meet FRED for pre-1980 data:
http://research.stlouisfed.org/fred2/graph/?chart_type=line&s%5B1%5D%5Bid%5D=CURRCIR&s%5B1%5D%5Btransformation%5D=pc1
7. December 2012 at 12:38
http://research.stlouisfed.org/fred2/graph/?g=dzN
7. December 2012 at 12:46
Tyler, Jknarr,
On the China thing…China sell us stuff, we sell them a note. The accounts are ballanced and the currencies do not fluctate relative to one another.
What is this mean for monitary policy, demand, and NGDP. Demand stays the same, but as we are demanding a large quantity of Chinese goods, so that demand isn’t flowing into US GDP.
If China decided to, all of a sudden, dump their US Treasury holdings, what does that mean for F/X, interest rates, balance of trade, demand, inflation, and GDP? Impossible to say without knowing what they do with the proceeds of the sale.
They take their money home? Sell the dollars and buy Yuan. Dollar depreciates, inflation rises, interest rates rise, real GDP falls, NGDP stays about where it is.
7. December 2012 at 12:48
Thanks Tyler, here’s my response.
1. Just to be clear, I’m well aware of the role of expectations in markets and the ineptitude of the media trying to explain economic relationships. My point is just that it seems like the FED announcing more purchases often does drive the price up. Or at least, announcing more purchases than expected drives the price up and less than expected drives it down, rather than vice-versa which seems to be what Sumner is arguing. So news media aside, am I mistaken about this short-term observation? Again, I’m quite open to the possibility that I am.
2. I get the relationship between bond prices and (nominal) interest rates. The issue is the relationship between monetary expansion and nominal interest rates. It seems what you are saying is that there are two effects, a supply/demand effect and an inflation effect where Sumner is only talking about the inflation effect (perhaps assuming it always dominates). If this is true then it opens the door to saying either one could dominate and possibly even saying that the former dominates in the short-run but the latter in the long-run which would explain my above observation, if it is correct. Again this seems supect due to the potential for arbitrage. So if we think of it this way, is there a way to prove that the inflation effect always dominates the S/D effect?
7. December 2012 at 12:50
Oh, and also, my original question: If it’s true that expanding the money supply raises nominal interest rates, does that mean we have to rewrite the intermediate macro textbooks or is there something I’m missing that makes this a less monumental assertion?
7. December 2012 at 13:04
This is incorrect. The effects on bond prices are not identical. The inflation caused by money printing raises the prices of consumer goods, and that puts an inflation premium on t-bonds and bond prices or otherwise lower.
However, this does not mean that it doesn’t matter if bond sellers get the money from the Fed or if consumer goods sellers get the money from the Fed. If inflation reduces the value of bonds, then it is one thing if that is accompanied by bond sellers getting the new money first, it is another thing if that is accompanies by consumer goods sellers getting the new money first.
The argument isn’t over bond prices. Richman’s argument, which you continue to misrepresent and misunderstand, is that IF the bond sellers get the new money first rather than, say, gold sellers or consumer goods sellers, then the bond sellers are relatively benefited as compared to everyone else, because while everyone owns depreciated money, and while the real value of bonds goes down, the bond sellers receive new money, whereas everyone else does not.
The fallacy in your thinking is that you keep comparing the wrong two things. You are comparing inflation with no inflation, when you should be comparing inflation sent to bond sellers rather than inflation sent to gold sellers or consumer goods sellers.
Yes, nobody is denying that bond prices tend to fall with inflation, but if inflation is going to take place, then bonds are screwed regardless. But if the bond sellers get the new money first, then they are relatively benefited as compared to others. Hence, it certainly does matter who gets the new money first.
During the first couple decades after WWII the Fed bought modest amounts of bonds. Then after 1965 bond purchases exploded, and the Fed absorbed vast quantities of T-securities. And yet between 1965 and 1981 we saw an epic collapse of bond prices. One of the worst bear markets in all of American history. Is there a cause an effect relationship?
It is always a mistake to use macro intuition in macro, because all macro is really micro.
If the Fed used the same inflation over that time period and bought consumer goods instead of bonds, then while bond prices would have still fallen, they would have fallen by more precisely because the new money is being used to buy bonds rather than consumer goods. If the Fed used the same inflation to buy consumer goods, then consumer goods prices would have still risen, but they would have risen by more precisely because the new money is being used to buy consumer goods.
Nobody is denying that purchasing bonds tends to make bond prices fall. The actual argument that you continue to not address, is the difference between inflation to buy bonds versus inflation to buy something else. The resulting effect on bond prices is NOT the same in the two cases, precisely because of the supply and demand “micro” intuition that you are claiming doesn’t apply.
You are again comparing the wrong two things. The comparison is not between inflation to buy bonds, and no inflation. It is inflation to buy bonds versus inflation to buy X. The reason this is the comparison is because the whole Richman argument is that it matters who gets the new money first, not that it matters there is new money per se (although that also matters).
7. December 2012 at 13:04
ssumner:
Let me put this another way.
Do you think it matters to bond sellers, that if inflation is to take place (which reduces the prices of the bonds they own) that they receive the new money first, as opposed to taco sellers receiving the new money first? Do you honestly think bond sellers would be indifferent between them getting the new money first, as opposed to the taco sellers, given that inflation will take place, and given that inflation will reduce the prices of bonds?
7. December 2012 at 13:26
Mike
#1 – I think you’re exactly right. Fed purchases do drive up the price of treasuries, the same way that when company B announces it’s intention to buy company A’s shares for $16/share, the price of company A’s stock immediately rises to ($16 – the estimated odds that the deal won’t go through).
2. An interesting question. I would say that the disconnect between monetary expansion and inflation seen in the past few years disproves the idea that inflation always dominates the S/D effect. How about your next thought: does the inflation effect always dominate S/D over the long run? Japan’s monetary base has increased quite a bit over the last 15 years or so, and inflation has yet to catch up.
I think Scott is accurate in stating that fiscal policy can (and often does) strangle monetary policy. In our own case, the interest on reserves policy and a whole lot of regulatory uncertainty make banks unwilling to lend any more than they absolutely have to. Consumers are also, as a whole, making an effort to reduce their debt. Those things are what cause the disconnect between textbook economics and reality, in my understanding of the situation.
7. December 2012 at 14:29
ssumner – So the Fed is currently buying approximately 61% of all debt issued by the US Treasury (and buying more in MBS) to drive interest rates on the 10 year “up” to 1.62%? In 2008, however, it was buying negligible amounts to drive rates “down” to 5.0%?
Could this be the counter-example to 1965-81 you were looking for earlier?
7. December 2012 at 18:02
ssumner:
The Fed can obviously cut T-bill yields, but easy money usually raises T-bond [yields].
Not if the Fed buys T-bonds, as they did with Operation Twist, which consisted of the Fed buying long term t-bonds and selling short term t-bills.
http://research.stlouisfed.org/fredgraph.png?g=dAc
Not that correlation proves any theory correct, but long term rates have fallen to a new low trend as Operation Twist was put into effect.
I hope you’re not going to claim that because the yield on 10 year t-bonds is around 1.6%, that those bond investors are communicating their 10 year inflation expectations as being around 1.6% (that’s not even including the other determinants of interest that add a component to the rate!), rather than the obvious explanation, which is that the Fed is buying those bonds.
7. December 2012 at 18:33
Major —
The technical impact of Fed asset purchases on yields can be measured in basis points. It is trivial. The production of Fed liabilities (currency) dilutes every dollar in your pocket and therefore increases NGDP — sometimes massively in the case of ZIM dollars in Zimbabwe. Currency production — the liability side of the Fed asset purchases — is like napalm, as opposed to the Treasury purchase lower yield flyswatter.
If the Fed is succeeding at what it states its goals are — economic stimulus and growth — then yields ought to rise even as the Fed buys Treasuries — not because of the Treasury purchases, but because of currency formation. The dilutive currency effect is huge, twisting the yield a bit here or there is tiny.
What the market has been saying with lower yields is that base money growth remains insufficient (it has even shrunk of late) and tweaking and twisting the curve is not stimulative in the slightest. Thus yields have fallen.
Stimulative Treasury purchases would boost base money in a form usable by the economy (reserves are not usable by you or I)– i.e. expand the Fed balance sheet. Treasury purchases would be correlated with — but not cause — higher yields. The liability side of currency production produces higher yields, as it dilutes every dollar in existence.
7. December 2012 at 19:01
Mike F. You asked:
“Oh, and also, my original question: If it’s true that expanding the money supply raises nominal interest rates, does that mean we have to rewrite the intermediate macro textbooks or is there something I’m missing that makes this a less monumental assertion?”
Macro texts state very clearly that expanding the money supply rapidly creates high inflation (except obviously at the zero bound.)
Macro texts state quite clearly that high inflation raises nominal interest rates.
Hence easy money produces high interest rates. Milton Friedman made this point over and over again–I’m shocked that anyone is surprised. Actually I’m not longer shocked–I was shocked when I started blogging and discovered that the rest of the world ignored the message of mainstream money textbooks.
bmcburney. The share of Treasury debt owned by the Fed is very small, and is no larger than 10 years ago. Rates are certainly not falling because of a lack of debt available to the public–the amounts have exploded. David Beckworth has posts showing this.
Rates are falling because NGDP growth since 2008 has been the slowest since Herbert Hoover was president. Does that sound like easy money to you?
Obviously the demand for base money soars at the zero bound, especially if you start paying interest on it. Indeed reserves basically are a part of the public debt today, so there has actually been a huge increase in interest bearing public debt in circulation.
7. December 2012 at 19:02
OK, I thought about this a little more and the way I am conceiving of it now is to picture a downward sloping demand for bonds which is a function of inflation expectations and a perfectly inelastic supply representing the quantity of bonds in the market. Instead of treating the FED’s purchase as increasing demand, I will treat it as taking some of the bonds out of the market and therefore reducing the supply. At the same time, if it increases expected inflation, it will increase the demand at any given price. Since both curves are moving the effect on price depends on the relative magnitudes of the changes and the elasticity of demand. Essentially by imagining a downward sloping demand curve for given inflation expectations, I am assuming that the real rate people are willing to accept on the margin is decreasing in the number of bonds they hold. This means that what I called the “Supply and Demand” effect above is actually reducing the real interest rate (by moving decreasing the supply curve in this case) at the same time that the inflation rate is increasing so the effect on real rates (and the price of bonds) depends on the relative magnitudes of these changes. This means, I think, that if the interest rate in the IS/LM model is considered to be the real rate and not the nominal rate (which frankly I’ve never been very clear on…) then this view put forth by Sumner is, at least to some extent, compatible with that model. (I know he’s not advocating that model, I’m just trying to figure out how far he is deviating from it.)
7. December 2012 at 19:05
Everey once and a while I stumble on MF by mistake. This time he’s citing Operation Twist. Great way of showing Cantillon effects—a program that failed to inject money.
I see he hasn’t lost his ability to say something utterly off topic.
7. December 2012 at 19:06
In this quote “expanding the money supply rapidly creates high inflation (except obviously at the zero bound.)” how imporant is the word “rapidly?”
7. December 2012 at 19:08
Let me word that differently: Does more rapidly always mean higher interest rats than less rapidly in the short run or is the reverse possible at all?
7. December 2012 at 19:20
ssumner:
Everey once and a while I stumble on MF by mistake. This time he’s citing Operation Twist. Great way of showing Cantillon effects””a program that failed to inject money.
I thought that was clear. I did say selling t-bills and buying t-bonds, didn’t I? I didn’t use this example as a Cantillon effect.
My point was only to show that the Fed buying t-bonds can be associated with rising t-bond prices. Of course, if the Fed inflated to buy those bonds, then there probably would have been less of a price increase, but I think that considering the fact that the entire yield curve has been falling while the Fed is buying all along the yield curve, it is safe to say that t-bond prices would have still risen, just not as much.
The 1.6% yield on the 10 year is pretty glaring evidence that inflation to buy bonds is having the effect of raising the prices, rather than lowering them.
7. December 2012 at 19:26
I think that once all the money printing the Fed has engaged in the last few years begins to raise consumer prices substantially (there will have to be another extensive boom, which may take some time), and investors begin to observe 5%, 10% price inflation, then they will probably begin to price the 10 years a lot lower.
Honestly, I don’t think many investors today are even assured of 5 years out, let alone 10 years out, when it comes to mental time horizons of what the Fed will do next. The Fed is buying 10 year bonds, and that is enough for investors to buy them and flip them to those who can sell them to the Fed.
7. December 2012 at 19:31
jknarr: That contradicts what the Fed has to say about it’s own operations. They pledge to keep interest rates low for X amount of time, not to inject money into the economy until the inflationary pressure causes interest rates to rise.
What the market has been saying with lower yields is that people flee to safety in times of stock market volatility. You combine a greater demand for less risky assets with a smaller supply of bonds because business are less inclined to borrow money and invest in growth in a poor economic environment, and you have high priced bonds. Investors don’t like the low yields, and some of them start to shift into lower credit quality bonds, dividend paying stocks, etc. That’s why spreads on longer duration and higher risk bonds have been compressing.
The Fed wants to 1) lower the cost of capital for business and individuals, and 2) push people out of low yielding savings and money markets and get them investing again. And as long as our creditworthiness remains more or less sound, interest rates are not going anywhere until the Fed says so.
7. December 2012 at 21:17
Tyler —
Yes it does contradict. The Fed’s explanations seem a bit disingenuous to me — charitably, I suppose that the statement that they are keeping rates low cannot hurt per se, and could be complementary to the expansion of the balance sheet effort. But it is not the main thrust of monetary policy — their balance sheet is.
We should watch what they do, not what they say. They are expanding the monetary base with every Treasury purchase. The vector for boosting NGDP is the monetary base, but they have ensured that excess reserves remain locked up — and at zero rates, few are borrowing by definition at 4% NGDP.
So their effort seems perverse to me — it would appear that they are not interested in effective stimulus (which would be more like 1939 style currency printing expansion of the balance sheet). They are repeating the policy failures of the early 1930s. And if I know this, then they certainly know. They are setting up for ongoing flattening of the yield curve toward a Japan-like scenario — because of their monetary efforts are routed into dead reserves, not alive currency. Again, Treasury purchases tweak yields a bit in basis points — the real action is in the monetary base.
Tight policy (below trend currency growth) leads to slow NGDP. Slow NGDP boosts the value of income-bearing assets. A higher price to income bearing assets boosts P/E ratios and lowers yields. Higher prices for bonds means more volume of bonds shift onto the market. Hence tight money creates increased debt. The Fed tightened in 1980 and remained tight since — yields and NGDP fell during this entire period, and debt and leverage exploded higher.
At near zero, the cost of debt cannot get any lower. There are simply not enough marginal healthy balance sheet borrowers *at this level of NGDP*. I maintain that interest rates are simply the revealed discounting of future NGDP — rates are simply prices that gravitate toward the natural rate of interest (that is set by the pace of the base expansion — i.e. NGDP).
Companies and individuals do not want to borrow debt at present — they want higher NGDP based incomes. Zero rates are a symptom of this phenomenon — the ability and willingness to borrow is not an everlasting oil well (at a given level of NGDP). Zero rates marks the end of borrowing. (there is still space on the long end for large corporates and the treasury).
People are buying fixed income hand over fist, and selling equity — what does this say about their inflation expectations? What do their inflation expectations say about the Fed’s monetary stance? In actuality, they are holding many more money market assets than Fed Funds would indicate — check out demand deposits sometime — skyrocketing.
Low yields are pricing in the fact that the Fed is not stimulating effectively — whether by choice or mistake.
7. December 2012 at 22:09
Mike F, MV = PY. That’s mainstream, although most people wrongly think it’s a useless statement. There is nothing unconventional about Sumner’s views in this post, except for the fact that he takes the theory of Rational Expectations more seriously than most economists do. This means that instead of believing that changes in the money supply have long and variable lags in affecting future nominal spending, he assumes that changes in expected future nominal income affect current nominal income with no lag at all. 30 years of research plus a Chicago PhD have convinced him of the rightness of this approach. Plus the best New Keynesian models are strictly supposed to reason this way as well…
7. December 2012 at 23:24
I get the relationship between inflation and nominal interest rates and I know that’s mainstream. And I’m not arguing with Sumner’s view I’m just saying that it seems like this implication regarding the effect of OMOs on bond prices *in the short run* (would italicize that if I could) is at odds with the (old) Keynesian model I remember from intermediate macro (actually we learned it in first year grad macro as well). It also seems like this doesn’t fit with my own (very) casual observation. And I’m pretty sure I’m not just remembering it wrong, I went and grabbed an intermediate macro text (Bade and Parkin) from the grad computer lounge and looked through it. Here are a few passages. (Note, this is going to be a bit long, I understand if Sumner doesn’t want to read/reply to it. It is not an argument that Sumner is wrong or that people shouldn’t have been aware of what he is saying since Friedman apparently said it, I’m just saying, it seems like I’m not the only one who doesn’t.)
I couldn’t find a section where they explicitly talk about OMOs but here is from the section called “Bond prices and interest rates.”
“[I]f people are holding more money than they want, they will attempt to get rid of some of it by buying financial assets.
This process increases the price of financial assets and lowers the interest rates. Thus when people want to hold less money than is available, interest rates decrease; when they want to hold more money than is available, interest rates increase. But when interest rates change, so does the quantity of money demanded: as interest rates increase, the quantity of money demanded decreases; and when interest rates decrease, the quantity of money demanded increases.”
This is essentially explaining why there is a downward sloping demand for money, then when they increase the money supply (holding prices constant) they move along this demand curve which shifts the LM curve to the right causing lower interest rates and higher output in the short run. Then, of course, in the long run prices increase and it goes back. It seems to me that if expanding the money base actually raises interest rates in the short run, the whole mechanism that they are explaining here is nonsense.
It’s true that textbooks state clearly that expanding the money supply causes iflation and higher inflation causes higher interest rates but they don’t put these together the same way when they are talking about monetary policy. Here is their explanation of monetary policy in the 1980s.
“At the start of the decade, the United States was experiencing a serious burst of inflation triggered by a series of large increases in world oil prices. In 1980-1981, the Federal Reserve decided to apply a severe dose of monetary discipline: the money supply growth rate was not permitted to keep up with rapidly rising prices. As a result, interest rates increased sharply in 1981.
In 1982, money suply groth was held steady but in 1983 the Fed stepped up the growth rate of the money supply, permitting interest rates to decline. For the next few years through 1986, the money supply growth rate was held steady at bout 8 percent a year. But with inflation under control, interest rates gradually declined through this period. Short-term interest rates declined more quickly than long-term rates. There is a general tendency for the swings in short-term interest rates to be larger than those in long-term rates. This was a period of rapid and sustained economic expansion.
In 1987, the Fed attempted to keep inflation under control by slowing the growth rate of the money supply. During the remaining years of the decade, the money supply growth rate was held steady at about 5 percent a year. Long-term interest rates held steady through this period but short-term interest rates increased through 1989, raising questions about whether the Fed had slowed the money supply growth rate too quickly, pushing the economy to the edge of recession.”
Then in a later chapter they talk about the equation of exchange, Fisher equation etc. but they essentially treat changes in inflation as exogenous not as the result of Fed policy (much like they treat them in the above passages).
Also, Bernanke does seem to talk about policy in similar terms as mentioned by others above. So I still feel like this notion deviates pretty significantly from the IS/LM model of monetary policy they teach in my neck of the woods. I guess I should have studied harder for my GRE and gone to Chicago. (=
8. December 2012 at 00:02
Hey Scott, Free Exchange just stole your post title!
http://www.economist.com/blogs/freeexchange/2012/12/americas-latest-employment-report
http://www.themoneyillusion.com/?p=14452
Though apparently Bush said it first?
8. December 2012 at 01:13
Scott,
I think the easiest way to explain this is that Fed action causes a change in expectations regarding NGDP growth. This causes businesses and individuals to sell financial assets (e.g. issue commercial paper, borrow money from the banks, take out mortgages, etc.). This increase in the supply of financial assets pushes prices down (i.e rates up) and the effect of this additional supply (or expectation thereof) is greater than the effect of the increased demand from Fed buying.
IMHO, this is both trivial and obvious. I don’t see why people have such a hard time understanding this especially since it is so well borne out by the empirical data.
8. December 2012 at 02:22
Mike F, yes we agree with all that. Scott had some recent posts (google my handle and “themoneyillusion” (not in quotes), click on the second one) where he explained how he sees the supply and demand for money balances. The IS-LM model is not incompatible with Market Monetarism; I use it myself. You just need to use it correctly, which few do. For example here: http://monetaryfreedom-billwoolsey.blogspot.com.au/2012/05/negative-real-interest-rates.html. Also see Nick Rowe’s blog at Worthwhile Canadian Initiative (in the sidebar) for some truly insightful theoretical exposition.
In the conventional analysis, an increase rate of growth of nominal money balances leads to an increase in the supply of real money balances (remember to put “r” not “i” on the vertical axis in ISLM!), which pushes down the interest rate until people are willing to hold all the money. The lower interest rates stimulates spending. As real income rises (when the economy is not at maximal capacity), not only does the demand for real money balances increase at every interest rate, but also since the marginal propensity to consume from transitory income is <1, the equilibrium loanable funds rate also falls. The result is a somewhat lower interest rate and higher level of real income. Straight out of intermediate macro.
What happens next depends on whether the economy has now exceeded its natural level of output. If it has, then since money is neutral in the medium run, prices rise faster and the supply of real money balances shrinks. This shifts the LM curve backwards until real interest rates return to normal. At the same time, however, people begin to expect higher inflation. This leads to a lower quantity of money demanded at each real interest rate (a lower quantity of money demanded in response to a higher nominal interest rate). This is known as the Cagan effect, and can be found in Mankiw's intermediate textbook, amongst others. This time however the supply of real money balances shrinks further to meet the reduced demand, through even faster rising prices. Hence, "expected inflation causes current inflation". If we put nominal rates in the ISLM diagram, then both IS and LM shift upwards to a higher rate at the same (natural) level of income.
Now the MM analysis is somewhat different. (This is also consistent with what sophisticated New Keynesians would predict using an expectations-augmented IS curve, btw.) We believe that rates move now as a result of rational market expectations of future changes in nominal income. Real rates may go down for a while from the excess liquidity – but the Fisher effect will occur immediately. Also, even if the economy was below the natural level of output to begin with, monetary expansion will shift the IS curve right, and the real and nominal interest rates may both rise. Or they may not. It is hard to say. As Bernanke said in 2003, interest rates are a very unreliable indicator of the stance of monetary policy. Take a look at what happened to the three month T-bill yield in December 07, when Fed policy became unexpectedly tight (commencing the extended Great Recession) and stocks and long bond yields fell. In fact Scott has a post on that (which I won't link to because I want to save my second link for the next line so that this comment gets through the autoblock).
If Scott were to respond to you though, he'd probably point you to this post: http://www.themoneyillusion.com/?p=915
dtoh, isn't that what Scott and David Beckworth have been saying all along?
8. December 2012 at 02:23
Still praying for the day when Scott switches to a blog template which doesn’t waste half the space on the page, forcing our comments into unreadable chunks…
8. December 2012 at 02:28
Mike, I don’t know about your textbooks, but good textbooks such as Mankiw or Bernanke always make it clear that the Fed controls inflation. MMs just make the next step – it controls inflation merely as a consequence of controlling NGDP (aggregate spending). But nominal rigidities mean that falling NGDP can drag down RGDP instead of prices. So to examine the Fed’s performance look at NGDP not prices, and subtract out predictable impacts of deviations of NGDP from trend when evaluating recessions. But underlying RGDP performance depends on the underlying production function which is (mostly) not affected by money, so you shouldn’t target real variables.
8. December 2012 at 04:10
Daniel Kuehn just left an interesting comment under his own blog post:
http://factsandotherstubbornthings.blogspot.com.au/2012/12/pick-up-book-kids.html
Scott, to what extent would you agree with this?
8. December 2012 at 04:39
Saturos,
You said,
dtoh, isn’t that what Scott and David Beckworth have been saying all along?
Yes, but I think his explanation can be clarified and his statement, “It’s always a mistake to use microeconomic intuition in macro” is not applicable in this case as there is a simple micro explanation which explains the macro effect.
8. December 2012 at 06:03
SS: I notice that lots of commenters insist that bondholders gain when the Fed injects money by buying bonds. Even if this were true, it would have no bearing on my criticism of Richman. That’s because any effect on bond prices would be identical if the Fed injected money by paying government salaries in cash, rather than buying bonds.
MF: This is incorrect. The effects on bond prices are not identical. The inflation caused by money printing raises the prices of consumer goods, and that puts an inflation premium on t-bonds and bond prices or otherwise lower.
MF, why is this incorrect? Here is a simple explanation of what happens.
Parties: Government, Central Bank, Government Salary Earner, Bond Seller, Bond Buyer
Situation where Central Bank does not intervene:
The Government has to pay a $1 pre-contracted salary. Bond buyer wants to buy $2 of bonds. Bond seller wants to sell $1 of bonds.
Gov issues a $1 bond to finance salary. Bond buyer uses $1 cash to buy newly issued bond and $1 cash to buy sellers $1 bond. Gov pays salary with $1 cash raised.
Net effect:
Gov debt increases by $1
Publicly held debt increase by $1
Bond seller gains $1 cash, gives up $1 bond
Bond buyer gains $2 bond, gives up $2 cash
Salary earner gets $1 cash.
Money supply remains constant.
If the Central Bank wants to increase the money supply by $1 they can:
a. Print a new $1 note and buy the newly issued $1 bond. Buyer will then buy seller’s bond with $1 cash.
b. Print a new $1 note and buy the seller’s $1 bond. Buyer will then buy newly issued bond with $1 cash.
c. Print a new $1 note and pay the salary directly, thereby extinguishing the Gov’s liability and eliminate the need for the Gov to issue the $1 bond (monetize the debt). Buyer will then buy seller’s bond with $1 cash.
In all 3 cases buyer will not be able to buy the other $1 bond from the specified parties – so in a. and b. demand for bonds goes up by $1, in case c. supply of bonds goes down by $1.
Demand going up and supply going down are equivalent.
The effect in a,b,c is identical:
Gov debt increases $1 in a. and b., stays contant in c.
Publicly held debt stays constant in all cases.
Bond seller gains $1 cash, gives up $1 bond in all cases.
Bond buyer gains $1 bond, gives up $1 cash in all cases.
Salary earner gets $1 cash in all cases.
CB prints $1 cash in all cases and gains $1 bond in a. and b.
Money supply increases by $1.
Gov. debt minus Central Bank held debt remains constant in all cases.
In a. the newly printed $1 goes to the salary earner, but money supply increases because the bond buyer is unable to exchange his 2nd $1 cash for bonds.
In b. the buyer’s $1 goes to the salary earner, but money supply increases because the bond buyer is unable to exchange his 2nd $1 for bonds. Seller gets the newly printed $1.
In c. the new $1 goes to the salary earner, but money supply increases because the bond buyer is unable to exchange his 2nd $1 for bonds.
So in all cases the Central Bank’s intervention increases demand for bonds by $1 (by increasing demand in a. and b., and cutting supply in c. It also increases the money supply by $1.
The onus is on Austrians to explain why there is a contradiction in their reaction to the Central Bank’s actions – a. and b. vs c.
When the CB intervenes through action a. and b. Austrians complain that bond prices are artificially increased resulting in lower yields and an unfair profit for bond holders.
When the CB intervenes through action c. Austrians complain that it is inflationary, which results in higher yields and lower bond prices.
8. December 2012 at 07:28
“T-bond yields soared between 1965-81” because roc’s in M*Vt soared (aggregate monetary purchasing power). My prediction for corporate AAA bond yields in 1981 was 15.48% (actual yield was 15.49%).
The sharp increase in DD velocity beginning in 1964 was the consequence of a variety of factors which included 1) the daily compounding of interest on savings accounts in commercial banks & “thrift” institutions, 2) the increasing use of electronics to transfer funds, 3) the introduction of negotiable commercial bank certificates of deposits, & 4) the rapid growth of ATS (automatic transfers of savings to DDs) & NOW (negotiable orders of withdrawal) accounts.
But the most important single factor contributing to the increased rate of money turnover probably was those structural changes which made virtually all time deposits the equivalent of low velocity demand deposits. These changes included the virtual elimination of Regulation Q (interest ceilings on time deposits) & the introduction of interest bearing checking accounts known as money market demand accounts. High interest rates and expectations of higher prices have been both cause and effect of rising rates of Vt
The importance of Vt in formulating – or appraising monetary policy derives from the obvious fact that it is not the volume of money which determines prices & inflation rates, but rather the volume of monetary flows (MVt) relative to the volume of goods & services offered in exchange.
Unfortunately, financial innovations aren’t one-time events. Greenspan’s housing boom-bust (the misallocation & mal-distribution of available credit within the residential & commercial mortgage markets), was primarily fueled by financial engineering: CMOs, CDOs, CDSs, ABSs CLOs, CFOs (securitization & credit enhancements), via (SIVs, SPEs, & other non-banks).
The housing sector was continuously stimulated by various structural alterations to assets which served as loan collateral (for example, re-hypothecation in “2007, accounted for half the activity in the shadow banking system).
The housing bubble was characterized by its colossal money flows (money X velocity). Note: these money flows went undetected as the Fed discontinued the G.6 in 1996.
8. December 2012 at 08:16
MikeF, Try to search for my post on the myth of Volcker. The tight money wasn’t adopted until mid-1981.
8. December 2012 at 08:18
MikeF, Here is is:
http://www.themoneyillusion.com/?p=11083
8. December 2012 at 10:47
Saturos – Thanks for detailed response. I think I understand most of what you guys are saying, I will check out these posts.
A few notes on the text just to be fair to it (although I don’t feel like it’s that great. I used their principles text when I was visiting at another university one semester and didn’t think it was great either…)
I didn’t mean to imply that nowhere in the text do they make the connection between monetary policy and inflation, of course they do and of course everyone including me has always been aware that monetary expansion causes inflation. It’s just that they don’t, and I didn’t, and it seem silke Bernanke, and most other people I have heard speak about it, don’t go from there to monetary expansion raises interest rates. As you point out, and I actually mentioned above, if the interest rate in IS/LM is the real interest rate then these views are not incompatible but I actually went to a great deal of trouble to determine which interest rate they were talking about and they never specify. The only way I could tell was from a one-paragraph section about real vs. nominal interest rates where they implied that they had actually been talking about the nominal rate the whole time but if they used the real rate it would be the same (talking about investment demand). This is, in my opinion, a flaw in the textbook, but this isn’t the textbook I had used previously (Blanchard) and I distinctly remember always having trouble determining which interest rate they were talking about when dealing with Keynesian economics. (I had the same issue when reading the GT, I was actually convinced, though I might have been mistaken, it was a while ago, that Keynes was sometimes talking about one and other times the other without specifying.
Anyway, if I’m understanding this right now, when the Fed buys bonds, they lower the real rate and (at least potentially) increase the expected inflation rate so the effect on the nominal rate *could* go either way (as I suggested above) and it is then an empirical observation that Sumner is making by saying that it usually goes up. Correct? And Bernanke is confused when he seems to imply that by promising to hold rates down for the foreseeable future, is an expansionary monetary policy, or is that not what he’s doing?
8. December 2012 at 11:45
“Do monetary injections always reduce bond prices?”
The widely held mis-conception (prevailing hubris on the Fed’s technical staff stemming from their Keynesian training), that a tight money policy results in high interest rates (& vice versa) derives from the premise that interest rates are determined by the demand for & the supply of money, rather than the demand for & the supply of loan-funds.
Interest is the price of loan-funds; the price of money is the reciprocal of the price level. It is true that an expansion of commercial bank credit (loan-funds) produces a concomitant increase in the volume of money & that the initial effect is to depress interest rates, other things being equal. If, however, the increase in the volume of money flows exceeds the changes in the volume of goods & services offered in the markets, prices will rise. And if the price increases broadly based & chronic, we have inflation.
Long-term interest rates (& bond prices) are determined not only by the various supply & demand factors that affect short-term rates, but also by a unique factor; namely, inflation expectations. The expectation that price levels will chronically increase injects an “inflation premium” into long-term rates. The tendency of short-term rates to rise concomitantly with long-term rates is largely the consequence of the substantial substitutability of both short & long-term financing.
Rising inflation expectations operate on both sides of the supply-demand equation. Long-term lenders will demand, in terms of group behavior, an interest rate that at least exceeds the “inflation premium” (real-rate). Thus the higher the expected rate of inflation, the higher the long-term rates.
Note: the collapse of the Euro-dollar banking system beginning in July 2008 & the payment of interest on excess reserve interbank balances on Oct 9, 2008 constituted 2 major shifts in the demand for loan-funds (gov’ts). The growth in IBDDs also induced dis-intermediation within the non-banks (velocity), where the size of the non-banks shrink, but the size of the CB system is unaffected (like 1966).
8. December 2012 at 13:45
ssumner – During 2011 the face amount of Treasury debt the Fed holds increased from less than $800 billion to more than $1.6 trillion. During the same year, the yield on the Ten Year went from 3.36% to 1.89%. By my calculation, nominal GDP growth in the US that year was 3.9% and ten years before (2001) it was 3.3%. I may have done the ngdp calculation incorrectly and I am sure there are issues I may have missed. It seems to me, however, that the fall in Treasury yields during 2011 is somewhat more easily explained by Fed bond buying than by historically low ngdp growth rates.
Please explain again why bond purchases by the Fed do not reduce interest rates.
8. December 2012 at 14:27
bmcburney, Since 2008 NGDP growth has been slower than any 4 year period since 1929-33. That’s the reason rates are so low. You need to look at more than one year data, 2001 was coming off a huge boom. Long rates are highly correlated with both the level of NGDP relative to trend, and the expected growth rate.
In 2011 they were swapping MBSs for Treasury debt–the balance sheet didn’t go up quite that much, if I’m not mistaken. In any case, temporary monetary injections at the zero bound have little effect on anything, including long term rates. Long rates rose on the QE3 announcement. My arguments have focused on the much more interesting case of what happens when nominal rates are positive.
MikeF, Easy money is more likely to lower long term real rates than long term nominal rates, but even real rates can go up as a result of easy money.
It depends on lots of things.
8. December 2012 at 14:32
Saturos, How does New Keynesianism “microfounds” the IS curve? (Not that I think microfoundations are needed.)
8. December 2012 at 19:30
The argument that one does not have to actually receive new money from the Fed in order to relatively benefit from inflation, for example, that one can buy t-bills that will later go up in price, and thus one can acquire gains equal to the gains acquired by those t-bond sellers who actually sell to the Fed, while seemingly a valid argument in itself, doesn’t actually prove what it ostensibly sets put to prove.
In order to actually gain by such speculation, it is not enough that the t-bond prices merely rise. The t-bond owners must go out and sell the t-bonds, to other investors. This of course implies the existence of a population of investors who do not speculate in t-bonds, but rather buy them at higher prices from those investors who do speculate.
In other words, not everyone can gain from inflation via speculation! For if everyone did try to gain from inflation via speculating on t-bonds, then there would not be anyone to later sell them to. If there is nobody available to subsequently sell the t-bonds to, at hier prices, then the t-bond prices would actually collapse to zero. This is because prices require actual exchanges to be made. Without exchanges, there are no prices.
Hence, the Sumnerian “solution” for benefitting from inflation the way the primary dealers benefit from inflation, is not actually a universal solution. The “solution” requires a population of people who relatively lose from inflation (i.e. those who initially don’t speculate, but then buy the same t-bonds at higher prices from the primary dealers and the speculators).
Thus, the Richman argument that it matters who gets the new money first, stands. The only alteration needed is to say “It matters who gets the new money first, because those who receive the new money first, and those who own the same assets as those who receive the new money first, will relatively benefit, and those who this group needs in order to relatively benefit, namely, those who do not currently own the asset but will pay higher prices later on, they will relatively lose.
9. December 2012 at 04:00
http://www.bankofengland.co.uk/publications/documents/events/secure/qeconference/document60.pdf
Conclusion:
“In this paper, QE is examined as a monetary policy tool that impacted financial markets during 2009 and 2010, and the impact on long gilt yields is measured. A linear dynamic regression model and an ECM model were adopted using explanatory variables based on well established economic theory and long standing relationships. According to the ECM model, the impact of QE on long gilt yields was a fall of around 95bps, while it was between a fall of 33bps and 63bps in the OLS model (on the short and long run respectively). Hence the overall size of the reduction in the long gilt term premium overtime appears to be significant. In addition to reducing the term premium, liquidity and confidence also improved, which is a prerequisite to a well functioning financial system and a healthy flow of credit in the economy. Based on this evidence, this analysis concludes that QE was successful at reducing gilt yields, and although not tested here, it is reasonable to assume that lower gilt yields helped or will help lowering corporate borrowing rates and stimulating economic activity.”
Is this consistent with your claims/arguments? I’m finding it hard to unpick your explanation of why QE doesn’t affect bond prices.
9. December 2012 at 06:02
Ben, No, I disagree with that stiudy. Most central banks think they are conducting an easy money policy, whereas it is actually a tight money policy. The recent decision to pay interest on reserves has completely changed the demand for base money, so monetary injections are no longer “easy money.” Instead, one form of interest-bearing government debt is being replaced with another. As I pointed out there may be Cantillon effects from IOR, but since the Fed has been doing that only since 2008, it seems unlikely that IOR is what they had in mind.
9. December 2012 at 11:57
ssumner – As I am sure you are aware, the Fed was not injecting large amounts of money into the system throughout the four or five year period of the crisis. (In fact, it seems to me that I may have read that observation on a blog not so very many clicks away from this one.)
Look, if you saying rates fell from just under 5% to 3.4% and stayed there solely because of the dramatic fall in ngdp, I will believe you. In fact, I do believe you. If you go on to insist, however, that the fall from 3.36% to 1.89% in 2011 had nothing to the Fed doubling the size of the balance sheet during the same period, I have to ask that you explain this idea very carefully. Correlation is not causation, but it seems to me that an argument that Catillion effects do not exist ought to address what seems to have happened in the market for Treasury Bonds in 2011.
9. December 2012 at 12:14
I agree with Bob Murphy that the whole thing is a bit misleading…
While it arguably doesn’t matter that individual bondholders get the money “first” in case of treasuries purchases by the Fed at market prices (even though we could also argue that, if all the beneficiaries of the purchases had tried to sell their treasuries at the same time to raise cash, it would have depressed their market price, making the bondholders worse off), we can’t say the same of all monetary injections (in particular, cheap refinancing operations, LTRO or FLS-type – see my comment on previous post), where bankers (and some of their first clients) clearly benefit from getting the money first.
But treasuries purchases by the Fed still has an indirect effect:
1. After selling their treasuries, managers/bankers will surely want to rebuild their portfolio with the same assets (ie other treasuries), using the newly injected money.
2. There is then increased demand for treasuries, allowing the federal government to borrow more cheaply.
3. Then the government will spend this new money. It is actually THERE that the “first” matters.
Granted, point 3 would be fiscal policy. But in this case, we clearly can’t separate the action of the Fed from the resulting implications for fiscal policies, can we?
(sorry if the point has been raised before)
9. December 2012 at 14:27
Scott writes in the post above: “I plead not guilty. (…)
In the real world the Fed agonizes over whether to spend the new money on MBSs, discount loans, T-bonds, etc. Bernanke doesn’t agonize over whether to spend the new money on Tomahawk missiles. So I am dealing with monetary policy as it is, not some alternative world where the Fed might buy a trillion dollars in bananas, and drive up banana prices.”
This is really disingenuous: The point of the original Richman article, as I understood it, was in fact to criticize the current setup of the monetary system. The claim was that it disproportionally benefit the federal government and “particular interests” that get to deal with the Fed directly.
Now, you might disagree with the existence of such effects etc. etc., but you cannot say that “you are dealing with the monetary policy as it is” and that you therefore do not consider alternatives to currently existing policy in response to an article that explicitly criticizes monetary policy “as it is”.
It is not an answer to say “well, I don’t care if there are distortionary effects, this is just the way it is done now.” This might be true, it is just not clear at all if it follows that this is the most equitable or lear distorting way to achieve the goals of monetary policy.
So clearly, you are talking past the intent of the original article.
I actually disagree with quite a bit in the original article, but I really cannot see how one can say that there are not obviously people and institutions who greatly benefit from the way monetary policy is conducted.
9. December 2012 at 17:03
In order to actually acquire a bond with its newly-created money, the Fed must either outbid the marginal bond purchaser, or entice the marginal bond seller to sell – either of which is accomplished with an increased bid. If the Fed fails to entice the marginal seller or outbid the marginal buyer, it will fail to actually purchase a bond, and fail to inject new money into the money supply. Therefore, Fed purchases of bonds must increase bond prices above the price that the market would have set otherwise.
If market-determined bond prices have adjusted lower due to future money purchasing power expectations, then the Fed purchases are increasing a lower bond price, but it is increasing it none the less.
“any effect on bond prices would be identical if the Fed injected money by paying government salaries in cash, rather than buying bonds.” -sumner
Since the federal purchase of bonds must increase its price above the market price, it also must decrease the bond yield. The yield is modified by the fed purchase to be lower than its market-determined value. If the fed gives the new money to the government directly instead of purchasing bonds, then the bond prices will reflect the unmodified market-determined yield, right? Won’t those bond prices be lower, to represent this unmodified yield? Wouldn’t this be true even if the government doesn’t issue new bonds?
9. December 2012 at 17:05
bmcburney, I think if you look at the days the T-bond rates plummented it was when there was bearish news for NGDP growth, not news about QE2 (which was announced in 2010, and hence already priced in.)
Long term bond yields reflect expected future rates. There’s no way that a highly expansionary monetary policy that dramatically increased expected NGDP growth would lower expected future interest rates. That’s why rates rose so sharply in 1965-81, the easy money drove expected NGDP growth much higher. Now we have tight money.
JN, The money will end up in the banks. It doesn’t matter if they get it “first,” or 10 minutes later.
If they get a sweetheart deal that’s different, as I’ve said 10 times.
The rest of your comment is off topic, as everyone agrees that monetary and fiscal policy matter–you don’t have to convince me of that. The question is why.
Shining Raven, It is others that brought up crazy stories of the Fed buying weird assets, to criticize me. I was responding to them. I think everyone now agrees that it doesn’t help the first person to receive the money for “purchasing power” reasons.
The main institution that gains is the Federal government, which earns seignorage, as I said and as everyone understands. The first person to get he money gains almost nothing, as the commissions on OMOs are tiny.
9. December 2012 at 17:09
higgins. If you pay off your mortgage are you “distorting” the mortgage market?
As far as salaries, as I said what matters is net debt held by the public. It’s the same whether the Treasury issues a bond that the Fed buys, or whether the Treasury never issues it because gov. salaries are paid with cash. It’s basic accounting.
10. December 2012 at 01:29
ssumner:
1) The original article by Richman claimed that the specific way in which the Fed conducts monetary policy benefits particular interests.
2) You are of course free to argue that the claimed effect does not exist or that there is no mechanism(“who gets the money first” – I take this to be about liquidity, and believe there is a point there).
3) But of course it is very much relevant to compare to alternative ways of conducting monetary policy if you want to address the question if other people, and it is neither crazy nor off-topic. You are of course free to not engage in this discussion, but in the context (criticism by some right-libertarian guy of the way the Fed does conduct policy) a counter-facutal is perfectly reasonable.
10. December 2012 at 01:30
… if other people could benefit as well…
10. December 2012 at 10:19
ssumner – I have not checked but, assuming your assertion regarding economic news and interest rate changes to be true, doesn’t this evidene cut the other way? I think everyone agrees both that interest rates are influenced by business conditions and that rates will fall when bad news concerning those conditions appears in the press. I also think everyone should agree that if something happens as a result of news it is really the result of investors reacting to news. My guess for the mechanism which produces the corellation between news and interest rate movement is that investors are attempting to “front run” the Fed, buying bonds on speculation that the Fed will soon be buying bonds or otherwise act to reduce interest rates. This would not refute the existence of Cantillion effects. To the contrary, it proves that investors, at least, think they exist.
On the other hand, if causation only worked the other way it is hard to think of a mechanism which would explain a close temporal correlation between economic news and interest rate changes. It seems to me that your analysis requires that “tight money” causes both lower interest rates and lower ngdp at the same time. News that ngdp was dissapointing should not result in any change in interest rates because both things are controlled by another variable (and you reject the idea that causation works in both directions). Do investors believe that bad enconomic news will cause the Fed to sell bonds and drive interest rates down?
10. December 2012 at 10:30
ssumner:
higgins. If you pay off your mortgage are you “distorting” the mortgage market?
Is he counterfeiting money in his basement?
10. December 2012 at 18:08
bmcburney;
“My guess for the mechanism which produces the corellation between news and interest rate movement is that investors are attempting to “front run” the Fed, buying bonds on speculation that the Fed will soon be buying bonds or otherwise act to reduce interest rates.”
I have to work very hard to get my students to stop thinking this way. One approach I us is to point out that interest rates fell sharply during recessions even before the Fed was created.
11. December 2012 at 10:24
ssumner – Yes, my guess for the mechanism which produces a general corellation between recessions and reductions in interest rates (especially prior to creation of the Fed) is that during recessions credit worthy borrowers are in short supply relative to lending standards which are adjusted higher by by lenders in response to the additional risk. More money chasing fewer qualified borrower reduces rates. I believe this mechanism is still in place today and, as I explained previously, can very plausibly account for the initial fall in interest rates at the beginning of the rescesison. During actual recessions, I firmly believe that this mechanism is dominant.
But we were discussing the subsequent change in interest rates from 3.3% to 1.89% which just happened to coincide with massive bond buying by the Fed. On a percentage basis, the fall in interest rates which took place during the calendar year 2011 alone was actually larger than the initial fall in rates from mid-2008 to mid-2009. Meanwhile, ngdp did not fall during 2011 (but I believe growth in ngdp did slow). Why should a relatively modest slow down in ngdp growth have a bigger effect on interest rates than an actual contraction (and a very severe one at that)?
And, for what its worth, I stand by my comment regarding a mechanism whereby short term moves in interest rates may be corellated to news regarding macro conditions. As I percieve it, your comment does not really address the specific issue of short time horizon moves in rates.
12. December 2012 at 05:25
bmcburney, I can’t explain every up and down in 10 year yields, but neither can you. The effects of QE2 were fully priced in by the fall of 2010. Markets move on new information. There was no new QE information in 2011
12. December 2012 at 10:07
ssumner – I certainly agree that neither of us can explain every up and down in 10 year yields. But if that is the case, how can we be sure that the effects of QE2 were fully priced in by the fall of 2010?
I also generally agree markets move on new information and also agree that there waw no new QE information 2011. But the question is whether there were QE2 effects which were not experienced until the actual buying was underway. Also, isn’t it inconsistent to say that there are no Cantillion effects from actual bond purchases but also assert that Fed announcments concerning bond purchases do move markets? Why is it that Fed announcements move interest rates down but actual purchases move interest rates up?
For what its worth, I enjoy your blog and have enjoyed our discussion (even with the snarky comment about your students).
16. December 2012 at 09:35
Sumner’s right. QE2 was fully priced in by 2010 year-end. If you trade interest rate futures you ought to be able to explain why yields fluctuate.
All POMO & SOMA changes were virtually complete by July…but the major move in rates hadn’t yet happened (Jan-July: 3.36% to 3.03%). Rates fell in the 2nd half of 2011 (Aug-Dec: 3.03% to 1.89%) largely because of (1) the EURO-zone “flight-to-safety” (see corresponding re-alignment of the U.S. trade-weighted exchange rate which impacts inflation expectations), plus (2) the Dodd-Frank Wall Street Reform & Consumer Protection Act that provided unlimited insurance coverage on non-interest-bearing transaction accounts…
16. December 2012 at 09:36
Sumner’s right. QE2 was fully priced in by 2010 year-end. If you trade interest rate futures you ought to be able to explain why yields fluctuate.
All POMO & SOMA changes were virtually complete by July ($460b)…but the major move in rates hadn’t yet happened (Jan-July: 3.36% to 3.03%). Rates fell in the 2nd half of 2011 (Aug-Dec: 3.03% to 1.89%) largely because of (1) the EURO-zone “flight-to-safety” (see corresponding re-alignment of the U.S. trade-weighted exchange rate which impacts inflation expectations), plus (2) the Dodd-Frank Wall Street Reform & Consumer Protection Act that provided unlimited insurance coverage on non-interest-bearing transaction accounts…
16. December 2012 at 10:08
“Meanwhile, ngdp did not fall during 2011 (but I believe growth in ngdp did slow)”
Real-output moved pari passu with that of money flows-M*Vt (our means-of-payment money times its transactions rate-of-turnover) during 2011. As the roc in MVt doubled, so did real-gDp (from 83.5b in the first half of 2011 to 176.3b in the second half of 2011).
http://research.stlouisfed.org/fred2/data/GDPC1.txt
18. December 2012 at 06:56
http://seekingalpha.com/article/1069671-will-long-term-yields-rise-in-2013?source=email_macro_view&ifp=0
“The 10-year inflation adjusted Treasury of constant maturity first dropped below zero at the start of November 2010. It bounced back above zero for a short time but then dropped below again, where it has remained. The fall in this yield below zero was tied relatively closely to the financial problems that were being experienced in Europe”
5. March 2016 at 22:13
[…] If I buy T-bonds, their price rises. If the Fed buys T … – 131 Responses to “If I buy T-bonds, their price rises. If the Fed buys T-bonds, their price (usually) falls” dis737 7. December 2012 at 07:33 […]