I just spent the past few days at the annual economics meetings in Atlanta, in some ways a depressing experience. (I arrived home with a cold.) I discovered that I’m not making any headway in convincing people of the importance of monetary offset of fiscal stimulus. Even worse, there seems to be a growing consensus that monetary policy is ineffective at the zero bound. I view this idea as not just wrong, but silly.
Today I’m going to suggest a seemingly inconsequential reform that might improve policy by removing a misconception about monetary stimulus at the zero bound, the false idea that aggressive QE involves vague “costs and risks”.
During the period of quantitative easing, some Fed officials worried that an excessively large Fed balance sheet could expose the Fed to the risk of capital losses, perhaps even bankruptcy. In fact, these so-called “risks” are an illusion, as Fed holdings of Treasury bonds represent an asset for the Fed and an equal liability of the Treasury. Because the Fed’s profits go to the Treasury, any gains or losses on its investments are a wash.
Fed officials know this, but worry about the “optics”. What if they had to go to Congress and ask for a bailout because they’d bought a lot of Treasury bonds, which had subsequently declined sharply in value when interest rate rose? If there was then a sudden upsurge in spending requiring a tight money policy, then the Fed might not be able to sell enough assets to hold down inflation. Hence the need for a bailout. Of course this scenario is extremely unlikely, but fear of embarrassment is a very powerful human motivator.
Because this is a phony problem there is a very simple solution—abolish the Fed’s balance sheet. Have the Treasury take over the responsibility of conducting open market operations to adjust the monetary base. Have banks hold reserve accounts at the Treasury, not the Fed. After all, didn’t the framers of the Constitution intend that the Treasury would have the duty of “coining money”?
Otherwise, nothing would change. The Fed would still make all the monetary policy decisions, just as they do now. The Fed could still be independent. The only difference is that once the FOMC had decided on a course of action, the instructions would be sent to the Treasury, not the NY Fed’s open market desk. The Treasury would be nothing more than an errand boy, with no discretionary power to determine monetary policy
The beauty of this system is that the Fed would no longer have to worry about going bankrupt. Would the Treasury now be assuming this risk? No, because there never was any risk to assume, it was all a cognitive illusion. The markets would now know that the Fed would be free to do whatever it takes to hit their target (whether inflation or NGDP.) And that would allow the Fed to do far less.
It’s conceivable that (at the zero bound) even a policy of buying up the entire national debt would not be enough to hit the Fed’s target. I think that’s very unlikely in the US, and extremely unlikely if a couple reforms are adopted at the next zero bound event:
1. No payment of interest on bank reserves.
2. A switch to level targeting of prices once interest rates fall to zero, as Bernanke has proposed.
In my view, there is now a pretty strong consensus in favor of those two steps, or something similar.
Inevitably, people ask what the Fed should do if even those steps are not enough. In that extremely unlikely case, I’d suggest the Treasury buy as many alternative assets as necessary to hit they target. I.e., the Treasury would essentially be borrowing money to create a sovereign wealth fund. I call that “monetary policy”, but I don’t really object to people calling it “fiscal policy” if they prefer. The point is that we should never, ever, ever react to a depressed economy with tax cuts and/or higher spending—do it all with money creation. Tax cuts and spending increases balloon the deficit, putting a burden on future taxpayers. A sovereign wealth fund merely adds a bit of risk, with no extra expected future tax burden.
But you should probably just ignore the previous paragraph; in reality the Fed won’t need to go that far.
To summarize:
For decades, we’ve wrongly assumed that the Fed is a bank. It’s not. It’s a monetary authority.
Because we’ve wrongly viewed the Fed as a bank, we’ve assumed it has a “balance sheet” like other banks. It does not.
Because we’ve wrongly assumed the Fed had a normal balance sheet, we’ve wrongly assumed it could suffer from a bankruptcy that would hurt taxpayers. That’s false. Its assets are a liability of the Treasury.
Because we’ve wrongly assumed there was balance sheet risk, we’ve wrongly assumed that highly aggressive QE might be “risky”. It isn’t.
Because we’ve wrongly assumed that aggressive QE might be risky (if only the feelings of Fed officials who might have to ask Congress for a bailout), we did too little QE, creating millions of excess man-year of unemployment during 2009-15.
Please, let’s abolish the Fed’s balance sheet and give the Treasury the responsibility to adjust the monetary base. I also favor abolishing IOR. But if that policy is to be maintained, then have the Fed set the IOR rate while the Treasury makes the actual payments to banks. (Ditto for discount loans, another facility I’d abolish.)
PS. Here’s a graph from The Economist showing the maturity of the Fed’s bond holdings in 2017:
READER COMMENTS
John Hall
Jan 7 2019 at 4:36pm
“Would the Treasury now be assuming this risk? No, because there never was any risk to assume, it was all a cognitive illusion.”
I agree for Treasury bonds, but what about mortgage backed securities (and similarly for your sovereign wealth fund example)? These are not a liability of the Treasury.
Scott Sumner
Jan 7 2019 at 5:42pm
John, The Treasury has now indicated that they back agency MBSs, which is what the Fed buys. I think that policy is unwise, but these bonds are now effectively the same as Treasury bonds. (These bonds are issued by Fannie, Freddie, etc.)
The sovereign wealth find is merely hypothetical, and almost certainly would not be needed in the US
Matthew Waters
Jan 7 2019 at 6:02pm
The Greenbacks or National Bank Notes should be better known. The tools were not great in 1861-1913, since they were constrained by Congrssional budgets and Treasury issuance.
However, the mechanics are better than a pseudo-private corporation in the Fed banks. An electronic Greenback system can adjust both issuance and payment of interest. For looser policy, either issue Greenbacks instead of Treasury auctions or reduce interest on Greenback accounts. Tighter policy can reduce circulation by overissuing a Treasury auction or increase interest.
Use Treasury auctions instead of the Primary Dealer system. The Primary Dealer and New York Fed desk is very opaque and has a terrible appearance. We have the technology to open it up in a more competitive system.
The electronic greenback system has a fundamental constraint. Once interest rates are zero and greenbacks replace all Treasury auctions (newly issued and rolled over), the policy cannot get looser. Treasury should do an open bid for market Treasuries at 0% rates, with lower durations getting new money first. If this method is exhausted, taxes should be cut until Greenback issuance hits target.
I *strongly* prefer the tax method versus the SWF. It’s very anti-libertarian to have an SWF. Countries that do it either are horribly corrupt or have truly temporary resource revenues. It should absolutely, 100% not be done by non-resource dependent countries.
Matthew Waters
Jan 7 2019 at 6:25pm
I realize this is basically the Chicago Plan and Milton Friedman’s 1950’s plan. But I do not understand why the Chicago Plan isn’t obvious. Why in the world go through a horrible SWF thing (or the discount window generally) as opposed to simply printing money and then cutting taxes if absolutely necessary?
Matthias Goergens
Jan 9 2019 at 12:23am
No need for the tax cuts. Just widen the assets the Fed/treasury is allowed to buy for monetary policy. Scott has written about that.
Eg Singapore uses the foreign exchange rate and never has to deal with any zero bounds on interest. Ie the asset the Monetary Authority of Singapore buys and sells is foreign currency. The US has some ideologically problems with such ‘currency’ manipulation. But they could just buy and sell commodities futures or NGDP futures etc. See also Scott’s remark about calling that part of the fed/treasury balance sheet a sovereign wealth fund, if politically necessary.
Tax cuts are great when they are done as part of sensibly supply side reform. Just don’t mix your monetary and your fiscal policy. See Scott’s writing about the curious case of the missing recession after the fiscal cliff of extreme austerity of 2013.
Scott Sumner
Jan 7 2019 at 7:20pm
Matthew, It’s a big mistake to use tax cuts for stimulus, as they require future tax increases. But I will admit that it’s less of a mistake than spending increases.
In any case it’s a moot point, because neither SWFs or tax increases are needed; ordinary open market operations are more than sufficient.
Benjamin Cole
Jan 7 2019 at 7:25pm
If the Fed is a national central bank ( or monetary authority ) then it should have the power to print money. That is one of the powers of the sovereign.
So what if the Fed buys a lot of bonds and then the bonds decline there after. If a counterfeiter a buys a house, and the house declines in value, is the counterfeiter in a bad way?
There must be some extremely tortured logic or confused accounting system that explains why people are concerned with the Fed’s balance sheet.
I do wonder why there is not more of a discussion of money-financed fiscal programs. It may be that QE in combination with a federal deficit is in fact a money-financed fiscal program, or so contends Michael Woodford.
I genuflect to the QE totem in the Museum of Macroeconomics. But then the Japan situation does cause me some concern, during my chanting.
Matthias Goergens
Jan 9 2019 at 12:29am
In the extreme case, the assets the Fed bought become worthless. That’s equivalent to giving away the money as a literal helicopter drop in the first place.
The only problem for the Fed is if inflation picks up too much: the usual response to that is to decrease the amount of dollars outstanding by selling assets from the balance sheet.
If there’s nothing of worth on the balance sheet, you can’t do that. And that’s it—worries about bankruptcy of the central bank are nonsense.
Japan only strengthens Scott’s argument: Japan needed more QE, but less fiscal deficit spending.
dlr
Jan 7 2019 at 7:47pm
Scott, by the time the we reach your emergency step 3 and start buying risky assets, how is it supposed to be working?
one possible way is as a mere signaling device, i.e. merely establishing credibility for future policy when off the zero bound. but in that particular case when even level targeting didn’t work as a signal, it seems harder to believe asset buying would work as a pure signal.
mechanically, on the other hand, it only seems to beat Wallace neutrality (I.e. the classic woodford objections if crediblity is elusive) if buying assets actually increases the risk on government liability such that holders (including holders of money) require a higher risk adjusted return. but succeeding at raising the riskiness of government debt seems economically equivalent to increasing the deficit by reducing taxes. in order to create the same inflation as a hypothetical deficit you have to lever up a lot.
you seem to be imaging just a tiny increase in risk but that only works if it’s really just a schelling focal point in disguise.
artifex
Jan 7 2019 at 11:14pm
The motives for using fiscal policy instead of quantitative easing are the same as for using fiscal policy instead of lowering rates below zero. The balance sheet is an excuse, as is the worry that, because asset prices are higher when interest rates are low or negative, lowering rates supposedly increases likelihood of bubbles.
Matthew Waters
Jan 7 2019 at 11:16pm
“Matthew, It’s a big mistake to use tax cuts for stimulus, as they require future tax increases. But I will admit that it’s less of a mistake than spending increases.
In any case it’s a moot point, because neither SWFs or tax increases are needed; ordinary open market operations are more than sufficient.”
Some portion of the federal budget will always be monetized. The Fed system is a layer on monetizing deficits through Treasury auction then OMO. The FDIC and implicit banking guarantees also monetize Treasuries, as the Treasuries have no capital charges.
Monetized tax cuts were the concrete proposal in Benanke’s helicopter money paper. The monetized tax cuts may never need corresponding tax increases if the demand for money remains. I realize this sentiment sounds like MMT silliness. But if demand for a currency exists, then the government will always eventually use monetization up to acceptable inflation.
Most importantly, the Bernanke tax cut defuses arguments for banking backstops. These arguments say just not enough Treasuries exist to have 100% reserves. A 50% rate of conversion in MZM to reserves will keep helicopter money from happening. A 70% rate still has substantial tax cuts in 1999-2002.
https://fred.stlouisfed.org/graph/?g=mAoP
Regulators will also need to look maturity mismatch issues with no backstop private banking. For example, securities lending played a large role in the financial crisis. Regular investors in funds ultimately funded Lehman Brothers and Bear Stearns through lending securities and investing collateral in Money Market Funds. This is how the financial crisis could have hurt regular people even though they had FDIC insurance. The regulation of fiduciary scenarios will need to be case by case.
Matthias Goergens
Jan 9 2019 at 12:34am
Yes, and the argument for more private money creation (via more fractional reserve banking) instead of via issuing and then monetising more debt is about allocation of resources to the private vs public sector.
It’s never about the government going bankrupt or otherwise unable to pay.
(Up to the limit of acceptable inflation.)
Basically, using money demand to cover government debt is not free in real terms. Even if it looks free in nominal terms.
Andy Harless
Jan 8 2019 at 12:00am
On the one hand you argue that the possibility of having to bail out the Fed is not a real risk. On the other hand you argue that tax cuts and spending increases raise the burden on future generations. This seems inconsistent to me. A bailout would require either cutting the budget elsewhere (or raising taxes) or accepting a larger deficit, which presumably would burden future generations. (Obviously it’s the same if the Treasury “bails itself out”: disinflationary OMOs would directly take the form of issuing new debt.)
I think there is a real (though IMO very small) risk if one insists on inflation targeting. The original QE works because the CB makes it large enough that it would be too costly to take it back (making the “promise to be irresponsible”, as Krugman would say, credible). So if the CB really does end up insisting on taking it back (which could happen if it insists on not allowing the inflation rate to rise above target to compensate for an earlier undershoot), there is a real price to pay. It’s fine if you just target the price level, as in the proposal you cite, but I think the anxiety about QE is predicated on the assumption that the target is always the inflation rate.
Andy Harless
Jan 8 2019 at 12:18am
Just occurs to me, in reference to my above comment: inflation targeting at the ZLB is always costly. We’ve made an artificial distinction between cases where seigniorage revenue is sufficient to cover the cost and where it’s not. There is no qualitative difference between a bankrupt Fed and a solvent one, but the former indicates that costs have been particularly large. There is a risk to QE, but it’s somewhat arbitrary to describe that risk as insolvency. The risk is simply that the cost may be larger than expected, perhaps much larger than expected, of which technical insolvency would be one indication.
bill
Jan 8 2019 at 8:16am
Under this system, when the Treasury issues currency, it’s essentially issuing small denomination perpetual bearer bonds that pay no interest, right?
Historical question. Who played the greatest role in starting the payment of IOR in October of 2008? I recall that there was already legislation in place to permit the Fed to pay IOR in 2011 or so but then that permission got moved forward? October of 2008 seems like the worst time in the last 75+ years to increase the rate of interest paid on reserves.
Matthias Goergens
Jan 9 2019 at 12:40am
George Selgin wrote a book about the floor system of IOR and its history. And has a lot of blog posts.
Vaidas Urba
Jan 8 2019 at 12:41pm
Good points by Andy Harless on the cost of monetary policy.
The business of central banking is a natural monopoly, and as with other natural monopolies, it would be wrong to restrict output and minimize the cost of services supplied.
Matthias Goergens
Jan 9 2019 at 12:43am
Emitting money is far from a natural monopoly. If left to their own devices, like in Scotland and Canada and Australia and some other countries during their free banking episodes, you get a a healthy competition of multiple issuers.
What is a natural monopoly is for the unit of account. Banks solved that in practice by standardising on a specific amount of gold per unit they offered to freely concert into and out of their notes. (But gold is far from the only possibility.)
Scott Sumner
Jan 8 2019 at 1:33pm
dlr, It would be really nice if it didn’t work. Then the Japanese could buy up the entire world, retire, and live lives of leisure based on the sweat of foreign workers.
Seriously, where did this idea that it would not “work” come from? If countries were not able to debase their currencies then it would mean the end of scarcity, indeed the end of economics.
Matthew, Sorry, I didn’t follow all that. Exactly what argument are you making?
Andy, You said:
“A bailout would require either cutting the budget elsewhere (or raising taxes) or accepting a larger deficit, which presumably would burden future generations.”
I don’t see this. Imagine that a bond market crash reduces the value of the Fed balance sheet by $1 trillion. It needs a $1 trillion bailout. That event also reduces the Treasury liability for Fed bond holdings by $1 trillion, thereby freely up exactly the resources needed for the bailout. They could print up $1 trillion in debt and give it to the central bank. AFAIK, higher taxes are not needed.
In years when the Fed makes a capital gain, they can give the profit back to the Treasury.
Now imagine the Treasury were doing monetary policy. Then there’d be no reason to worry if the price of T-bonds changed in that case.
Matthew Waters
Jan 8 2019 at 2:07pm
I’m sorry. My general point was that the monetary base would increase substantially without support for banking. By support for banking, I include deposit insurance, discount window and implicit bank guarantees.
I used the MZM money stock for the highest possible monetary base in this system. MZM was higher than federal debt outstanding from 1997-2010. Therefore, in the worst case, the US government needed substantial monetization to match MZM.
Some portion of MZM would still use private liabilities for banking. No government guarantee at all would exist. 50% conversion of MZM to gov reserves would not have needed new monetization for 1997-2010. 70% conversion would have needed some monetization above debt outstanding.
Kurt Schuler
Jan 8 2019 at 10:12pm
Scott, you write, “The Treasury would be nothing more than an errand boy, with no discretionary power to determine monetary policy.” That is not quite correct. The Treasury has control over exchange rate policy. It could in effect decide to abandon inflation targeting, or nominal GDP targeting, by setting some type of exchange rate target, although in current circumstances that would be unlikely. Also, the Treasury has the Exchange Stabilization Fund, which has in the past been much more significant in monetary policy than it is now.
If you are not already familiar with it, see Maxwell Fry’s book Money, Interest and Banking in Economic Development, which has a chapter on central bank bankruptcy. It is, however, oriented toward central banks that incur substantial foreign exchange liabilities, either from their own operations or from being saddled with liabilities when the government involves them in bank rescues.
Finally, fun fact that some of your readers may not know: The Fed had its main offices in the Treasury building until 1937, when its headquarters building was completed.
Greg Jaxon
Jan 8 2019 at 10:19pm
Hardly. Congress (not the Treasury) had the obligation to [establish the US Mint in 1792 or so] charged with ‘regulating’ (i.e. standardizing) the weights of gold and silver best suited to monetary purposes. The government itself was pretty much dead broke at the time. Money that was coined then came from (and could easily be melted back to) the bullion held by citizens who (perhaps through their banks) brought their monetary metals to the mint to be coined. If we ask who the Constitution empowers to “control the money supply” (not an 18th century concept) the answer must be that, through an open 0-seigniorage mint, that “power” was reserved to the people themselves.
Of course much later Legal Tender cases, brought before an intentionally-packed Supreme Court, disagree with me. But the dissents in those cases are mighty persuasive.
Matthew Waters
Jan 8 2019 at 11:28pm
The majorities of the Legal Tender cases rightly say Congress does a lot that isn’t specifically outlined. Greenbacks were very much necesary and proper for waging war and other enumerated powers.
The Federal Reserve is technically 12 federally chartered corporations, with very strict charters. States give corporate charters at will now, for unlimited duration or purpose. Before, corporations were chartered at the discretion of states. The Federal Reserve banks are in that mold of charter, like the first two central banks.
Thefirst central bank was federally chartered in 1791. So I believe most framers considered a federal chartered corporation constitutional. I don’t think the Fed *should* exist in its current form, but I think it is constitutional.
Greg Jaxon
Jan 9 2019 at 1:43pm
Matthew,
During warfare, extraordinary sources of credit and extra coercion to obtain it are understandable. U.S. Grant was clearly spoiled by this wartime dispensation, whereas Salmon Chase (who’d lobbied for the 1862 Legal Tender Act) recognized it as an emergency excess-use of monetary power. On the same day Chase’s court announced that (by 3 to 4) it would not require Mr. Griswold to accept Mrs. Hepburn’s bank notes at face value in lieu of the gold she owed, Grant appointed two new judges and within a year the L.T. Act was back in force by his new 5-4 advantage. This politicized view–that the government runs the money–held through several retests during the 1870s. It was the dissents in those cases that I found persuasive.
These cases construct a new identification between legal tender notes and a fixed quantity of monetary metal that is legally (rather than economically) secured. Doing this socializes the cost of securing the redeemability of the government’s bank notes. It improves the circulation-efficiency of the notes. But it really is a brand new (to the U.S.) legal construction, not anticipated by the founders. There being no free lunch, however, we should not ignore the duty to maintain parity with the prevailing gold standard which was a new burden the government tokk upon itself then. A well-regulated militia of banks was clearly even more required henceforth.
I don’t dispute that the federally chartered reserve banks were instituted constitutionally. The original (1913) Act isn’t even particularly objectionable to me. Its clear intent (in the form of the F.R. Agent’s collateral-posting to back FRN issuance, and in the original open market operations–not the retroactively(1933)-approved Treasury+Fed check-kiting scheme used today) was to “back” the monetary value of issued notes with what were then known as real bills. In 1913, the government’s bond was not considered a proper reserve. Indeed, that is almost certainly the reason to separate the Treasury’s balance sheet from the Fed’s, despite its non-profit charter. The ban on private innurement from operations of the Fed and the socializing of its economic costs seems, however, to have altered the incentives of its members. They are really inseparable from the political forces in play and are not beholden to either the people or to the commerce which the issued currency supports.
I know “Modern” monetary theorists conjoin the income tax amendment (1913) with the FR Act(same year) to say that this revolution enshrined a third source of very-low-entropy values (the power to tax economic flows) which might be suitable as a base money (the first two being the monetary metals and the social circulating credits found in the commodity goods markets). But where they (and I think Scott Sumner) see this as evolutionary substitution, I think it would be far more stabilizing to view it as incremental addition of entropy-lowering principles. Surely the income tax does not repeal the value or stability of the gold money supply, and yet that is the recurring theme of legal tenderers ever since 1934.
Matthew Waters
Jan 9 2019 at 4:17pm
Greg,
The question is sort of moot now since Greenbacks or National Bank Notes haven’t been issued for nearly a century. It’s now a question of the 12 Fed bank charters. The oversight from either Board of Governors or Treasury is arguably just a strict limitation on their charter.
Even if the 12 banks are de facto public entities, they are still de jure private entities which kick back most profits to the Treasury and have extremely strict regulation from Board of Governors and Treasury.
The focus on only using coin can also get really silly. If only coins can be minted, the Treasury can simply issue large denomination coin. Then allow private depositories and bank notes, with very strict regulation for 100% backing. Congress can then authorize the Treasury to issue or buy back Treasuries like New York Fed does, using transfers on the books of these private vaults.
Vaidas Urba
Jan 9 2019 at 2:01am
Scott, imagine that a bond market crash reduces the value of the Fed balance sheet by $1 trillion because future tax receipts are expected to be lower. Safer Fed’s balance sheet with a diversified portfolio would be a better choice in such a scenario.
Thaomas
Jan 9 2019 at 5:07am
Why isn’t this good idea right now what ever the state of inflation or interest rates?
Scott Sumner
Jan 9 2019 at 3:28pm
Kurt, That would be true if the Fed adhered to the exchange rate target. But without Congressional authorization to do so, I think the Fed would ignore the Treasury’s wishes. And there is little chance that Congress would authorize the Fed to hit an exchange rate target.
Since the Fed has far more power than the Treasury, it would win any battle. All the Treasury can do is try to indirectly impact the exchange rate via changes in national saving rates.
Greg, I never said Congress intended that the Treasury make monetary policy decisions, as should be clear from the fact that it does not make monetary policy decisions under my proposal.
I was inaccurate, however, as the Mint did not become a part of the Treasury Department until 1873. It was originally in the State Department.
Vaidas, Do you really think other assets would do well in your scenario? I don’t.
Thaomas, It is a good idea.
Vaidas Urba
Jan 10 2019 at 1:57am
Scott, foreign assets and commodities could do quite well in such scenario.
Tony S
Jan 10 2019 at 11:11am
I have two questions.
1. What is the point of using the Treasury to conduct operations for the FED. Is it just that, that way the Fed would never have to potentially have to ask for a bailout or that it would make clear that the FED and Treasury are a wash on the federal level?
2. I agree that the FED is not a bank in any normal way because they can essentially create money and control the money supply which no real bank can do. Plus they have the backing of the full faith and credit and whatnot.
I also understand that when the FED buys Treasuries that that asset is offset by a liability for the Treasury so it is, from the federal governments perspective a wash. However, printing all this money is not cost-less.
By increasing the total amount of currency in existence, isn’t the FED causing those outstanding dollars to be worth relatively less(each less scarce dollar buys less goods and services relative to the past which is the real value of a dollar) and fueling inflation? Since, that is how inflation works, too many dollars being created chasing to few goods, that balance sheet matters but maybe we should just publish M1, M2, M3 and velocity metrics instead of a BS misleading balance sheet.
I personally favor a gold backed 100% reserve currency and would abolish the FED. But that will never happen because it funds wars, greedy politicians, rich bankers, and Paul Krugman articles.
Matthew Waters
Jan 10 2019 at 7:11pm
Since the Fed does not redeem liabilities for gold anymore, its liabilities are meaningless. The Fed printing unlimited money on bad spending would have severe consequences. But it cannot go bankrupt in any meaningful sense.
The 12 Fed banks are privately chartered Federal corporations, as much as Apple is a Delaware chartered corporation. This took me awhile to understand.
A federally guaranteed private bank is not new either. National Banks got notes to issue from Bureau of Engraving and Printing like Fed does today. The OCC issued 90% of market value it received in Treasuries (later 100%). For gold convertibility of notes, the US government nearly ran out of gold in 1895 when JP Morgan organized a note issue. Then Fed gold convertibility was stopped altogether in 1933, as you know.
The Cross of Gold has a big appeal of certainty. Money has “true” value. But the gold standard is filled with half measures, suspension of redeemability, bimetalism, etc. Prices and demand had far greater fluctuations. Even between Second Bank of US and Greenbacks, Treasury often used private state banks instead of specie. There was never a pure time for the gold standard. Why use it at all compared to fiat currency with direct price or demand targeting?
Andy Harless
Jan 12 2019 at 9:01am
Scott, I would argue that a decline in bond prices doesn’t really reduce the Treasury’s liability. It doesn’t do so in accounting terms, and I don’t think it does so in practical terms either, unless the Treasury happens to have a surplus at the time, in which case it could buy back its own debt at a discount. Otherwise, the Treasury needs to continue floating debt. If it were to take the profit by buying back old debt, it would have to issue new debt at a higher yield, and it would have to continue paying those high coupons thereafter. Since the Treasury is usually in deficit, an increase in market yields is an adverse event for it. Both the Treasury and the Fed are worse off. If you consolidate them, part of the change would wash out, but the net effect of an increase in market yields would be costly because of the Treasury’s need to borrow and roll over existing debt: the difference is just that the adverse event doesn’t show up on anyone’s balance sheet, because the costs are in the future. And if the Fed had no balance sheet, an increase in market yields would still make it harder to control inflation, because the Treasury would need more money for debt service (on new and rolled over debt), effectively committing it to increasing the money supply in the absence of a fiscal tightening.
Andy Harless
Jan 12 2019 at 9:29am
Here’s another way to look at it. Ignore the Fed and assume the Treasury issues money directly. (It makes the standard promise to maintain the value of money, so it is limited in how much it can issue under any given circumstance, but it can issue as many bonds as it can manage to sell.) The Treasury’s normal short position in bonds constitutes a speculative bet on high inflation. This is good because the bet pays off exactly when the Treasury needs more resources to control inflation. So in a sense, the Treasury’s short position in bonds is a hedge against events that adversely affect its commitment to maintain the value of money. The smaller the short position, the smaller the hedge, and the more net exposure the Treasury has. When it does QE, it is taking off the hedge and increasing its exposure.
Scott Sumner
Jan 12 2019 at 1:58pm
Andy, This is an interesting discussion, I’d be interested in what others think. Here’s one way to think about the issue:
1. If we assume the EMH is roughly true, then the Treasury will not be good at market timing. Whenever bonds and cash are exchanged (either way), it won’t be clear which investment is better.
2. But point #1 doesn’t rule out the possibility that changes in ex ante risk may occur. So we need to consider two scenarios:
a. The Treasury does enough QE to hit its 4% NGDP target.
b. The Treasury does too little QE, NGDP growth falls to 1%, and it has a smaller balance sheet than in case a.
There’s actually lots to consider here. The slower NGDP growth hurts the government net debt position by reducing tax revenues. That makes government debt riskier, if we assume the Fed is independent and will refuse to inflate it away. Also, in the long run, the slower NGDP growth leads to a larger monetary base than the 4% NGDP growth option.
In my view, if they always do whatever it takes to hit the NGDP target, then in the long run QE might reduce risk in two ways. There’d be a smaller balance sheet, on average. And there’d be less NGDP volatility.
But even if I’m wrong and you are right, the optics work in my favor. A Treasury balance sheet would not look risky in the way that the perceived risk of the Fed’s balance sheet currently inhibits policy (or did in 2013). You rarely hear people talk about changes in the government net debt position due to changes in interest rates and inflation. It doesn’t seem to be a concern that inhibits policy.
Thus imagine the Treasury injects $2 trillion in cash, and extinguishes $2 trillion in bonds. Five years later they create $2 trillion in new bonds and buy back the cash. Even if the interest rate on the new bonds is different from the rate on the old bonds, that’s not going to be much more controversial that the Treasury’s current debt maturity management, which is certainly not very controversial.
Scott Sumner
Jan 12 2019 at 2:11pm
Andy, Another thought. Could the Treasury create enough of a hedge by shifting to longer maturity bonds so that on average they get the risk reduction you suggest, and yet still have a countercyclical policy of using QE where needed?
Scott Sumner
Jan 12 2019 at 2:14pm
Tony, Yes, the goal is to reduce the current perception that QE is risky.
The point is not to create inflation, it’s to hit whatever target the central bank is trying to hit.
Suppose you had a gold standard. Then after a big gold discovery they’d have to print money to keep gold prices stable. Of course fiat money is often more inflationary, but that’s a policy choice. I’m not advocating high inflation.
Comments are closed.