We need a policy, not an expedient
Several people sent me an editorial by The Economist, which seemed to endorse NGDP targeting:
Our suggestion is that the bank, backed by the chancellor, George Osborne, should make clear that it will not tighten policy until nominal GDP, currently £1.5 trillion, gets to a level that is at least 10% higher than today.
Let’s be clear about one thing; this isn’t really NGDP targeting, level targeting or otherwise. You need a rate of change, not just a number. And it’s also unclear how it would work. But I can think of all sorts of ways it could fail:
1. Suppose it took 10 years for NGDP to grow 10%—everyone who worries about liquidity traps would say that NGDP targeting has failed.
2. Suppose it took one year for NGDP to grow 10%, and inflation was 7%. People who warned that the policy was inflationary would be vindicated.
3. Suppose the British economy recovered at exactly the same rate it would grow without NGDP targeting. Then the policy would be seen as ineffective.
If the British government plans to switch to NGDP targeting, then they need to stick with it. Otherwise we’ll be in the worst of both worlds—-ineffective at controlling NGDP growth, and confusing to the markets.
PS. Yesterday I gave a talk at the Council on Foreign Relations in NYC. They sure have a nice boardroom! I also had the opportunity to meet Michael Woodford for the first time, and had a nice chat with him.
PPS. A couple days ago I did a post suggesting that BEA data implies NGDP growth was probably about 4% in Q4, not the 0.5% initially reported. Yesterday the government announced that job growth averaged nearly 200,000/month in Q4 (a strong quarter.) I think it’s pretty safe to say that NGDP growth was not 0.5%. Next month we’ll get the NGDI estimate, which is (ironically) a more accurate measure of NGDP than NGDP itself.
PPPS. Svetozar Pejovich tells me that there is interest in market monetarism in Serbia. If there are any Serbian speakers out there, I’d be curious as to what they have to say.
Tags:
2. February 2013 at 07:00
One can hope that there were more details behind the suggested offer than were actually reported, and I wondered whether the proposed level might be more effective than the unemployment target we now have, within a specific time frame. Good to hear that you were able to meet Michael Woodford!
2. February 2013 at 07:42
Becky, Nope, there were no more details.
2. February 2013 at 08:08
“I also had the opportunity to meet Michael Woodford for the first time, and had a nice chat with him.”
TELL US MORE.
2. February 2013 at 08:30
Article also shows the old “promising to keep interest rates low” fallacy. People still thinking in “real expenditure” terms – monetary policy works by making people want to buy more stuff.
2. February 2013 at 09:57
@ Saturos
What’s wrong with “monetary policy [expansionary monetary policy, that is] works by making people want to buy more stuff [including services and capital goods, of course]”? Isn’t that just Scott’s “hot potato effect”?
2. February 2013 at 13:22
Philo, You said;
“What’s wrong with “monetary policy [expansionary monetary policy, that is] works by making people want to buy more stuff [including services and capital goods, of course]”? Isn’t that just Scott’s “hot potato effect”?”
I think there is subtle but very important distinction. Scott’s HPE model says, people buy more real goods and services because they have more money. Functionally this is an accurate model, but I think Scott knows it’s not true and has conceded that in his “Bill Gates doesn’t buy a Ferrari because he has more money” analogy.
What really happens is that economic entities exchange financial assets for money because they want to buy more real goods and services, and the reason they want to do this exchange of financial assets for real goods and services is because of change in the real price (1 / the expected real return) of financial assets relative to the price of real goods and services (1 / expected return on real goods and services so to speak).
Fed action drives this exchange by raising the real price of financial assets (lower returns) and by raising the real return (lowering the real price) on real goods and services through higher expected NGDP.
Once you understand this simple exchange mechanism, MM and monetary policy in general becomes very easy to understand.
2. February 2013 at 13:35
I wish I could have been a fly on the wall during your chat with Woodford and hope you will write more about it in the near future.
2. February 2013 at 13:47
[…] Scott Sumner has a less optimistic comment on The Economicst’s endorsement of NGDP targeting. Share this:Email Pin ItLike this:LikeBe […]
2. February 2013 at 17:24
Scott
You wrote: “A couple days ago I did a post suggesting that BEA data implies NGDP growth was probably about 4% in Q4, not the 0.5% initially reported.”
I think that´s quite unlikely. NGDI will growth will most likely be quite a bit higher than NGDP, but it doesn´t mean NGDP was much above 0.5% (in principle it could even be lower):
http://thefaintofheart.wordpress.com/2013/01/31/ngdp-ngdi-two-sides-of-the-ledger-and-playing-catch-up/
2. February 2013 at 19:29
Saturos, I noticed that too.
dtoh. The HPE explains why people make more nominal expenditures on goods and services when there is more money in the economy.
Sticky wages explain why people make more real expenditures when there is more money.
The HPE does not occur because people are richer, but rather because people have more excess cash balances. I don’t see the HPE as a useful fiction, I think it actually explains why more money makes NGDP rise.
Marcus, You are looking at a different issue. Yes, it may be true that the two series converge because of rising NGDI. My point is different. We know that actual NGDP and actual NGDI are always exactly identical. The only question is whether measured NGDI or measured NGDP more accurately measures actual NGDP/NGDI. I think the measured NGDI number is more accurate, and in this case the Q$ NGDI is far more consistent with what we know:
1. We know that lots of jobs were created in Q4.
2. We know that although reported military spending fell at a 22% rate, actual military spending didn’t change very much.
Hence I think the NGDP number is erroneous.
2. February 2013 at 21:44
Regarding the Economist at least they are getting onto the same page—now if we can just get them to recognize the errors they have made. Of course, 10 percent higher NGDP is a failure if accomplished over 10 years….
Has anyone ever tried to get Scott Sumner on the Charlie Rose show….great exposure….
.
2. February 2013 at 22:34
Scott,
You have said (paraphrasing)”People hold the amount of money they need for transactions,” and “Bill Gates does not buy a Ferrari because he has money.” I completely agree.
For people to hold more money, they need to first make a voluntary decision to hold more money (exchange financial assets for cash). They don’t do this out of sense of civic responsibility to accommodate the Fed.
(Excluding financial intermediaries) They do this because the real price of financial assets has risen relative to the price for real goods and services, and they intend to exchange financial assets for real goods and services. Money is just the intermediary instrument that they acquire in order to effect this exchange.
More money will be closely correlated with increased purchases of real goods and services (increased AD) but it is merely an intermediate effect…not a cause.
I think you say people spend more because they are richer (the wealth effect) but this is just a subset of the more general case of exchanging financial assets for real good and services because of a change in the relative prices of financial assets and real goods/services.
The HPE model/mechanism which you rely on is functionally equivalent to the real financial asset price model/mechanism and will produce the same results, but I think it is less intuitive, more difficult to understand, and does accurately portray real world behavior.
2. February 2013 at 22:39
Correction
and doesNOT accurately portray real world behavior.
3. February 2013 at 05:54
@everybody
Take alook at this post by Bonnie (DaJeeps):
http://dajeeps.wordpress.com/2013/02/03/i-protest-preventable-spread-of-recession-not-recession-itself/#comment-2369
3. February 2013 at 08:54
dtoh, You said;
“You have said (paraphrasing)”People hold the amount of money they need for transactions,” and “Bill Gates does not buy a Ferrari because he has money.” I completely agree.”
I do believe this, but I also believe that if the Fed injects more money than people want to hold, their nominal expenditures will rise. This is one way of getting rid of the excess balances–at least in the long run. Of course interest rates also fall in the short run.
I agree that my approach is less intuitive than yours.
3. February 2013 at 08:58
Ben, Thanks, but I doubt whether I get asked.
Marcus, thanks for the link.
3. February 2013 at 10:57
Scott, You said,
“I also believe that if the Fed injects more money than people want to hold, their nominal expenditures will rise. This is one way of getting rid of the excess balances-at least in the long run. Of course interest rates also fall in the short run.”
I agree with this too, and I think it is quite consistent with a real financial asset price triggering mechanism/model whereby nominal expenditures rise because the increased price of financial assets causes causes an exchange of financial assets into real goods/services with money acting as the intermediate instrument. IMHO, there are advantages to thinking about it this way because:
1) I think this is actually what happens.
2) It responds to the argument that the Fed can issue money but it can’t make people spend it.
3) It is a little easier for Keynesians with a “lower interest rates spur investment/consumption” mindset to get their heads around your arguments.
4) It makes it easier to understand some of the other implications and issues surrounding monetary policy.
As I said, I think a pure HPE model/mechanism is functionally equivalent to the financial asset price mechanism and predicatively accurate. I also think it comes out of the monetarist “money supply is paramount” way of thinking and thus there is a historical attachment to the HPE model, but as I said I think it makes the MM argument much harder to understand and accept if HPE is described/viewed as the triggering mechanism.
4. February 2013 at 01:52
dtoh, you write “They do this because the real price of financial assets has risen relative to the price for real goods and services, and they intend to exchange financial assets for real goods and services.”
… or perhaps they think bonds are overpriced, so they’ll cash out now “before rates go back up” and watch for a bit so see what happens. Perhaps they’ll invest in some other financial instruments… waiting for a sign to get back in (like higher rates). I don’t see where they necessarily will purchase “real goods and services.” What, do you mean like a Ferrari? Doesn’t it drop 20% in value the second you drive it off the lot? Terrible investment for someone shooting for capital preservation for his golden years.
“real goods and services” are not a good substitute for someone looking for a safe long term investment. It seems to me the private sector will be incentivized to invent a substitute for safe Treasury bonds if the Fed buys them all up. Aren’t people investing in bonds looking for exactly that and not necessarily “real goods and services?”
I know that’s my circumstance. I had everything in equity and bond index funds and stayed the course for 20+ years (even through 2008/2009) until the debt ceiling fight of 2011. Then I sold everything and put it in a “low risk” stable principal fund, which strives to maintain $1 per share value (like a Money Market) but which pays 2 to 2.5% (while Money Markets are 0% essentially). So that’s where I’m PARKED waiting out the storms for the foreseeable future. Low risk capital preservation is what I’m after for the next 20 years. I’ll go back into index bond (and maybe equity) funds if the waters look safe, but I still don’t trust them now. I can’t imagine what would motivate me to trade those funds for “real goods and services” anytime in the next 20 years. Those are my SAVINGS!
4. February 2013 at 06:31
dtoh, You may be right about the marketability of your mechanism.
4. February 2013 at 11:05
Tom Brown,
A couple of quick clarifications.
1) The Ferrari was not intended as an example of an alternative investment but rather as an example of spending on goods and services. There had been some past history of posts/comments regarding Bill Gates buying a Ferrari as an example of spending on goods and services. I should have been clearer in my comment. (And just to be clear, when I speak spending on goods and services, I refer to any spending whether it be the purchase of groceries, a Ferrari, or a factory).
2) I understand you argument about switching from Treasuries to alternative investments, but I think this is incorrect or at least this portfolio balance effect is not significant from the point of view of the effect on aggregate demand (AD). To oversimplify a bit, Fed purchases of Treasuries (or any other financial asset) will push up the price of all financial assets, and therefore the real price financial assets in general will have risen relative to the price of goods and services. As a result there will be a marginal increase in the exchange of financial assets for goods and services and thus an increase in AD.
4. February 2013 at 13:43
dtoh, I think I see your point about AD and it’s a good one. I’m just questioning how real that effect is if most bond investors are like me and primarily concerned with safety. Because I’m not really interested in investing my nest-egg in “factories” either at this point… or even well diversified equity funds which ultimately invest in factories. But maybe that’s just me!
Also, if the government continues to deficit spend, then there will be more bonds available… and if the Fed continues to purchase them driving up their price… then why not purchase more, since you can always turn around and sell them to the Fed for a quick profit. Most Treas auctions on-sell the bonds to non-PDs, correct? Since I don’t directly invest in bonds (only bond index funds), and I don’t want to gaze at my portfolio more than once a year tops, I’m not interested in that kind of trading.
In fact, I really could use an education as to what I’ve put the bulk of my money in… it’s supposedly low risk, but what are “managed synthetics?” Related to factory building at all? Ha! … probably very very tenuously at best.
What if you knew for a fact (which I confess I don’t!) that most everyone buying T-bonds demanded safety above almost anything else. Would that change your AD argument any?
I guess risk is folded into the calculation, is that what you’d say? There’s a risk inflation will outstrip performance for the “safe” assets. So the professional capital preservation portfolio manager is marginally more inclined to rebalance toward factory building assets? Seeing how well these “professionals” did “managing risk” from 1999-2009 makes me more than a tad skeptical of their analytic/clairvoyant powers though!! Hahaha!
Final question: so the shares of a company building factories are “financial assets” aren’t they? So your argument says there should be a shift away from those because of that? So are shares of a company building factories “financial assets” or “real goods and services?”
4. February 2013 at 16:10
Tom,
Let me try to answer a bit.
I think I see your point about AD and it’s a good one. I’m just questioning how real that effect is if most bond investors are like me and primarily concerned with safety.
It’s important to keep in mind that this all happens at the margin, so the Fed doesn’t have induce all investors to sell… just a few that are wavering at the margin. Also it’s important to note that I’m talking about economic entities (not just investors) so the effect occurs regardless of whether it’s company selling Treasuries (or taking out a bank loan) to buy a new factory, or someone taking out a mortgage to buy a new home, or a consumer running up a little more debt on their credit card.
A second point on this is that modern portfolio theorists make an “irrelevance” argument, i.e. that investors are just looking at the pecuniary returns and the risk is already baked into the pricing.
Also, if the government continues to deficit spend, then there will be more bonds available… and if the Fed continues to purchase them driving up their price… then why not purchase more, since you can always turn around and sell them to the Fed for a quick profit.
Everyone else will think the same way bid up the price at auction and the opportunity to make a profit will go away.
What if you knew for a fact (which I confess I don’t!) that most everyone buying T-bonds demanded safety above almost anything else. Would that change your AD argument any?
I don’t think so. If Treasuries were supply constrained and investors reluctant were to part with them for safety reason, then you still have the exchange into real goods but Fed purchases would cause a bigger move in the price of Treasuries in order to effect the exchange. The economists who have written on this stuff generally assume though that assets are only being held for the pecuniary returns and that there are unlimited assets to buy.
Final question: so the shares of a company building factories are “financial assets” aren’t they? So your argument says there should be a shift away from those because of that? So are shares of a company building factories “financial assets” or “real goods and services?”
Yes shares are financial assets. But if you hold shares in a company and just exchange them with someone else for other financial assets (say Treasuries) so that you now hold Treasuries and the other person holds shares, this has no impact on AD.
If on the other hand, Bill Gates sell Microsoft shares to buy a Ferrari or Microsoft issues new shares to build a factory, then that does have an impact.
Also keep in mind as I have said that the exchange of financial assets for goods and services can take many forms including: sale of shares by an individual investor to buy a new car, increase in credit card debt to go our for dinner, share issuance or new borrowing by a company to build a factory, etc. A way to think of it is any increase in the net financial liabilities of the non-financial sector.
Sorry for being long winded. Hope that clarifies.
4. February 2013 at 16:49
dtoh, yes, thanks much for your reply. Don’t worry about being long winded, I MUCH prefer clarity to pithy undecipherable quips.
So in summary you’re saying the Fed purchasing Treasuries causes all financial assets to raise in price, which causes an increase in the purchase of “goods and services” from those in the non-financial sector wavering on the edge of selling their financial assets. In other words, this causes an increase in AD. Is that it?
How certain are you that this mechanism is true? Is there some link in the chain that you’re uncertain about? Assuming it’s all true, are there any downsides to this policy? Isn’t the central bank creating price distortions in financial assets? I suppose you’ll say “that’s exactly the point!”
Is there a way to implement this policy (NGDPLT) in a limited experimental way to make sure everything works as planned? Has that already been done in some country somewhere, for example?
4. February 2013 at 18:03
Tom Brown,
So in summary you’re saying the Fed purchasing Treasuries causes all financial assets to raise in price, which causes an increase in the purchase of “goods and services” from those in the non-financial sector wavering on the edge of selling their financial assets. In other words, this causes an increase in AD. Is that it?
Yes. To be more specific the real price (including adjustment for expected inflation) of all financial assets rises relative to the price of goods and services so at the margin there is an increased exchange of financial assets into goods and services (which is by definition an increase in NGDP). (I tend to prefer to use the word “exchange” rather than “sell” just to make it clear that transactions like new borrowing by businesses and consumers are included.)
How certain are you that this mechanism is true?
Pretty certain. I haven’t really been able to come up with anything that contradicts this and I’ve spent a lot of time trying. I think Scott and other monetarists would describe the mechanism more in terms of the quantity of money and/or the hot potato effect (HPE). I think that’s a function of their training. But in my mind, Fed Open Market Operations (OMO)are not just the issuance of money but an exchange of money for financial instruments. It’s yin and yang. They can’t exist without the other and they are indistinguishable. Using the quantity of money (HPE) model produces exactly the same result as the financial asset price model I have described. The two models are functionally equivalent, but I think the financial asset price model more accurately describes actual behavior and is easier to understand.
Is there some link in the chain that you’re uncertain about?
A couple of things.
1) Scott will argue that higher financial asset prices (lower real interest rates) will increase people’s willingness to hold money and thus reduce velocity which will tend to slow AD growth. I don’t think this is significant at low rates of interest and inflation or for small changes in these rates, but I don’t think I’ve convinced Scott.
2) It not clear 100% clear to me how NGDP growth breaks down between RGDP growth and inflation. Scott believes inflation is mostly a function of the money supply. I believe inflation (at least in the short term) is largely determined by imbalances between short term supply and demand expectations.
3) It’s also probably appropriate to make some comment about ER and the duality of money at the ZLB, but I’ll save that for later.
(Scott, apologies if I’m misrepresenting your views.)
Assuming it’s all true, are there any downsides to this policy?
I don’t think so.
Isn’t the central bank creating price distortions in financial assets? I suppose you’ll say “that’s exactly the point!”
Yes that is the point. But to be more specific, I would call it a change rather than a distortion, and not a change in the relative price between different financial assets, but rather a change in the real price of financial assets generally relative to goods and services.
One final thing that is important to keep in mind is that there is a critical third variable involved which is the expected return on purchases of (investments in) goods and services. Expected NGDP growth is a good proxy for this. If expected NGDP growth goes up, the return on investment in goods and services goes up (i.e. their real price goes down) so the exchange of financial assets for goods and services becomes doubly attractive. IMHO, if the FED does a good job at communicating, then this impact is even more important than the impact of the change in the real price of financial assets.
Just as footnote, when I talk about real prices, I like to think of it as (1/the expected real risk adjusted after-tax annual return).
4. February 2013 at 18:09
@dtoh, wondering if you’d agree with my assessment of the only ways in which reserves can leave the banks’ consolidated balance sheet:
1. Withdrawn by customers as physical cash, which, aside from lost or destroyed money, is eventually deposited again, and thus returns to being reserves (either as vault cash or exchanged back to the Fed as electronic Fed deposits). A wash.
2. Used to purchase Treasury bonds at auction. Those proceeds being spent by the gov back into the economy, thus again a wash.
3. Fed OMOs. Not happening now, in fact the reverse happening with QE.
4. Taxes. But unless the gov runs permanent surpluses, these again enter the banks as reserves when spent by the gov.
Given that we have a glut of excess reserves now (correct?), then Bank loans only convert an amount of their excess reserves (equal to 10% of the loan value) into required reserves… but they’re still reserves.
Thus I conclude that unless the Fed resumes 3. or unless the gov starts running a permanent surplus, reserves stay on the consolidated banks’ balance sheet. I realize this doesn’t describe movement of funds between banks to clear payments. I’m just trying to see if I’ve forgotten something here.
4. February 2013 at 18:30
dtoh, thanks for your detailed reply. I’ll mull it over.
4. February 2013 at 18:41
Tom,
So if we’re talking about reserves, there are required reserves and excess reserves both of which can either be held as vault cash/coins or deposits with the Fed. If cash and coins leave the banks/Fed, then they are no longer reserves.
If we’re talking about the monetary base (MB), then it’s Fed deposits, bank vault cash/coins, and all other cash/coin held outside of banks and the Feds.
Ignoring tax evaders and drug dealers, the important point to remember is that your point 1) is not a wash. As NGDP increases, the cash required for transactions increases proportionately (assuming constant V), and therefore the cash returns more slowly to banks so the amount of money (MB) held outside of banks and the Fed increases.
I may be incorrectly anticipating your question, but I assume you are wondering what will happen to all of the excess reserves. I think the answer is twofold. First as increased NGDP requires additional money for transactional purposes, bank excess reserves are reduced as they are converted into cash and held by the public. The other part of the answer is that if interest rates rise, then the banks will more rapidly liquidate their excess reserves (converting them into cash which gets distributed to the public) as they find investment opportunities that are better than holding excess reserves. This could cause a glut of cash which would require the Fed to start selling assets on their balance sheet rather than buying them. One of the benefits of NGDPLT is that it makes operations simple for the Fed. They don’t have to think about anything…just start selling assets when NGDP gets above the level target.
4. February 2013 at 20:14
dtoh, thanks again for responding. I’m talking reserves: vault cash (paper bills and coins that I referred to as “physical cash”) plus electronic bank Fed reserve deposits. And I’m not making a distinction between excess and required reserves (except in my side point about loans, which don’t affect the overall level of reserves if the consolidated bank balance sheet still has excess reserves).
My point #1 was really just about the paper bills and coins component of reserves (vault cash), and yes I agree it stops being reserves as soon as a customer withdraws it from his bank deposit (which is why I listed it as a way reserves leave the system). I can see that it might be possible for more paper bills and coins to leave their reserve status (be withdrawn from the bank) than return to it (be deposited at the bank), but I guess I’d be surprised if that were significant. I’m reasoning from personal experience (I know… dangerous!) in surmising that it wouldn’t be significant: I do 99% of my transactions electronically, as does pretty much everyone I know. I don’t see why the quantity of paper bills or coins I CARRY would increase no matter how high the rate of NGDP growth got. I’d still just take $40 out of the ATM and leave it in my wallet indefinitely (just in case) like I do now… unless after a decade or two of inflation I’d feel more comfortable upping my standard $40 to $80.
This brings up an interesting side issue: I can’t believe that going all electronic (getting rid of paper bills and coins) would significantly affect the economy. Do you agree or disagree?
This part of your answer confused me:
“…as they find investment opportunities that are better than holding excess reserves. This could cause a glut of cash which would require the Fed to start selling assets on their balance sheet rather than buying them. One of the benefits of NGDPLT is that it makes operations simple for the Fed. They don’t have to think about anything…just start selling assets when NGDP gets above the level target.”
Up until this point I was interpreting your “cash” as paper bills and coins. In the first sentence, who is “they?” The banks? If a bank or a bank customer wants to invest in better opportunities then I have a hard time seeing them using paper bills and coins to do it with. So I’m going to assume you are including electronic excess reserves here. In which case they just get transferred from bank to bank to clear the transaction, and that only happens if buyer and seller (of the investment opportunity) have accounts at different banks.
Same goes for buying assets from the Fed: I can’t imagine using paper bills and coins to do that. If instead electronic payments are used, then reserves do leave the banking system when this happens, which is my #3: OMOs in the direction of the Fed exchanging assets for reserves.
Just for laughs, let’s assume we did get rid of paper bills and coins. Then I’d just strike my #1 from the list, and the other three stand as the only ways in which reserves leave the system, correct? Sure they might get passed around between banks to clear payments and for inter-bank loans to meet reserve requirements, but they really don’t leave on a permanent basis except for #3 and #4. #3 only if the Fed wants to reduce the NGDP growth rate, and #4 only if, for some odd reason, the gov decides to take in a permanent surplus. I guess #2 also, if for some even odder reason the gov didn’t spend all of its proceeds from a bond auction. So really I just see #3, and I don’t see that happening until the NGDP growth rate gets too high (not a problem now!)
I don’t have any other follow on questions aside from that. I just wanted to see if you agree.
4. February 2013 at 21:11
dtoh, I recently saw Nick Rowe’s response to a similar question about whether or not he assigns any special significance to paper bill and coins:
“To me, the physical form of money is irrelevant. (OK, that’s too strong, because there are lots of practical reasons why physical form matters, but I normally abstract from all that when I’m doing monetary economics.)
What matters is who promises what, and who uses what, and who does what. Physical form matters only insofar as it affects those things.” — Nick Rowe
4. February 2013 at 21:20
Not to be too self-promotional but Nick’s comments on this can be read in full where he left them intitially
http://diaryofarepublicanhater.blogspot.com/2013/02/do-we-need-banks-and-phystical-cash.html?showComment=1360040647753#c1702875419412675064
4. February 2013 at 21:52
Mike! Sorry… I should have provided a link!
4. February 2013 at 23:10
Tom,
I can see that it might be possible for more paper bills and coins to leave their reserve status (be withdrawn from the bank) than return to it (be deposited at the bank), but I guess I’d be surprised if that were significant.
I think you need to look at this in the aggregate. I agree most people do most of their transaction with credit cards, checks, etc., but they still do some transaction in cash.If NGDP goes up by 2.5%, then the amount of people’s cash transactions will go up by 2.5%, and on average the typical person will be taking out $41 instead of $40 because they’re buying 2.5% more stuff with cash.
This brings up an interesting side issue: I can’t believe that going all electronic (getting rid of paper bills and coins) would significantly affect the economy. Do you agree or disagree?
I think it would have a few effects including a) allowing the Fed an additional policy tool of setting the interest rate (positive or negative) on cash, b) accelerating the loss of all person privacy, c) driving up the price of gold (or maybe Swiss Francs) as drug dealers stopped using the U.S. currency as a store of value, d) probably causing a significant drop in economic activity as all sorts of previously “tax free” income in the gray economy became taxable, and e) screwing up the economy in many developing countries that use the dollar as their main form of currency.
Regarding bank excess reserves…typically there are very few excess reserves. But when interbank lending rates drop to close to zero, there is an inflection point where the banks begin to hold excess reserves rather than holding other assets. If in this way, OMP simply result in an increase in ER, there is no AD effect from the OMP. The OMP are effectively “sterilized”.
Conversely as interest rates rise, banks will chose to reduce their ER and purchase other assets instead (e.g. new housing loans, increased credit to business, etc.) The counter-party to these loans are economic entities (businesses and consumers) who are in effect exchanging financial assets (e.g. taking out a loan) in order to buy goods and services (new factories, cars, etc.)
It is important to remember that these purchases of goods and services have a multiplier effect… the supplier of factory equipment pays their employees, the employees buy new cars, the car dealers goes out to dinner, etc., etc. So while cash only serves as a medium of exchange for some limited percentage of all the transactions, you eventually get an expansion of NGDP sufficient to require additional cash equal in the amount to the ER being drained (less any incremental increase in required reserves resulting from new deposits arising from the increased lending activity). If not, the banks who still have ER will continue to acquire additonal assets with rates of return higher than the rate paid on ER until they have eliminated all of their excess reserves.
The Fed can either slow or stop this process by selling assets and draining the ER that way… or by raising the interest rate paid on reserves (IOR) until the banks no longer have an incentive to replace ER with other assets. The Fed doesn’t have to worry about interest rates, the amount of ER, or the amount of other monetary aggregates. All they need to do is worry about the NGDP and either sell assets when NGDP goes above the target level or buy assets when it goes below.
I’ll make a couple more comments in a subsequent post.
4. February 2013 at 23:35
Tom,
Let me comments on your specific points 1-4.
1. Withdrawn by customers as physical cash, which, aside from lost or destroyed money, is eventually deposited again, and thus returns to being reserves (either as vault cash or exchanged back to the Fed as electronic Fed deposits). A wash.
Again as I have said, not a wash because as NGDP increases, there are increased transactional needs to hold cash.
2. Used to purchase Treasury bonds at auction. Those proceeds being spent by the gov back into the economy, thus again a wash.
Ditto for my comment on (1).
3. Fed OMOs. Not happening now, in fact the reverse happening with QE.
I think you mean OMS (Open Market Sales). OMO (Open Market Operations) refer generically to both OMP AND OMS. I agree with you. If there is a sufficient reduction in excess reserves to cause NGDP to go above target, then the Fed would need to engage in OMS…..or raise IOR….or raise the required reserve ratio in order to slow NGDP growth.
4. Taxes. But unless the gov runs permanent surpluses, these again enter the banks as reserves when spent by the gov.
I don’t think taxes are too significant except to the extent that a change in the tax rate would reduce the return on investments in goods and services and thus tend to reduce the NGDP growth rate. Otherwise, it doesn’t really matter whether the government borrows more money through Treasury auctions (2) or takes in higher tax revenue (4). In either case, the comment with respect to (1) still applies.
5. February 2013 at 00:39
dtoh, I really appreciate you taking the time to answer. I follow and agree w/ pretty much all you write so far.
re: “Regarding bank excess reserves…typically there are very few excess reserves.”
I view that as the “demand for credit is high” scenario (e.g. pre-2007) when banks are creating endogenous money and their problem is finding reserves to meet reserve requirements. I think we’re seeing the same picture, just from slightly different perspectives. Fortunately for them, on aggregate, they only have to come up enough reserves to match 10% of the endogenous money they create… and as I understand it, Fed OMOs regularly supply those reserves (again in times of high demand for credit). Again, as I understand it, loan-assets on the banks’ balance sheets can be used as collateral (for example) to borrow those from the Fed.
Re: “Conversely as interest rates rise, banks will chose to reduce their ER and purchase other assets instead”
This paragraph is where I have a little bit of a problem. I did mention in a previous post that if banks have excess reserves they can convert those to required reserves (in the aggregate) by loaning out funds. The amount that’s converted is 10% of each loan (or whatever the reserve requirements are). The way you wrote it makes it sound like they loan out the excess reserves themselves which doesn’t make sense to me since on the consolidated bank balance sheet, whoever is the recipient of the loan either keeps the deposit (rare) or spends the money and it’s thus redeposited in a bank again. If payer and payee have the same bank, nothing changes… except that 10% being re-classified. If different banks, still nothing changes on the consolidated balance sheet: reserves are transferred from one bank to another, but if the transferring bank needs to borrow back some to meet reserve requirements or cover an overdraft, it does…. either from the receiving bank, or any other bank that still has excess reserves. The effect on the consolidated bank balance sheet is the same: just 10% of the amount of the loan in excess reserves is converted to required reserves. No reserves actually leave the system. Of course, I’m ignoring paper bills and coins here as insignificant.
Which brings me to your paragraph starting with: “It is important to remember…”
In particular this bit:
“So while cash only serves as a medium of exchange for some limited percentage of all the transactions, you eventually get an expansion of NGDP sufficient to require additional cash equal in the amount to the ER being drained (less any incremental increase in required reserves resulting from new deposits arising from the increased lending activity)”
So I guess you are directly contradicting my assumption here that we could ignore paper bills and coins (which you are calling “cash” here, correct?).
My problem with that argument is that it makes it sound like the paper bills and coins are what’s driving NGDP growth. Are you sure that’s necessary to your argument? Or perhaps I’m misreading you? While your list of changes that would come with getting rid of paper money was significant, it didn’t include this (major) effect. It seems to me that the “physical form” of the money here should be “irrelevant” (to quote Rowe) in driving NGDP growth.
So humor me for a minute and suppose we did get rid of paper bills and coins; would that significantly hamper or otherwise change the mechanism you describe above regarding NGDP growth? I hope the answer is “no.”
I hope I’m clear about my problem with this. In our hypothetical paperless money system, reserves don’t leave the system until the Fed does OMOs to suck them up. They can be (slowly) converted from excess to required by the loan process but that’s it. It shouldn’t matter what the IOR is… there’s no other escape mechanism. I hope that doesn’t ruin your argument, because that would imply that it relies on paper money to work, which seems very odd to me. And of course I’m assuming here the gov spends every dime it gets (i.e. I’m assuming Treasury isn’t a net sink for reserves)
OK, thanks again for going through all this with me! I’ve learned a lot so far.
5. February 2013 at 09:52
“Again as I have said, not a wash because as NGDP increases, there are increased transactional needs to hold cash.”
dtoh, are you saying that holding a higher level of physical cash increaes NGDP?
5. February 2013 at 09:57
dtoh, I reread your piece and your four responses to my four points. I retract my statement above that it sounds like NGDP in this manner depends on the existence of paper money.
However, your responses to my 2 and 4 above really comes after the gov spends tax or bond sale proceeds, true? So 2 and 4 in and of themselves (from the minute traders fire up the Treas Direct website till the gov buys a new drone) are really not the means by which excess reserves leave the system… it’s my item 1 that you’re assuming follows them. True? (When the Northrop employees head out to the bar after). Although I might claim that at any one instant in time, the balance in Treasury’s Fed deposit represents excess reserves drained from the system…. so you could calculate a time average of how many potential excess reserves are instead sitting in that account.
OK, thanks so much for the great conversation. Really enjoyed it! I’ll look for you on this site in the future because I think you have a clear understanding and a great means of communicating to others. Thanks!
5. February 2013 at 10:17
Tom Brown,
I view that as the “demand for credit is high” scenario (e.g. pre-2007) when banks are creating endogenous money and their problem is finding reserves to meet reserve requirements.
IMHO, ER are really a function of whether the IOR rate is higher or lower than the risk adjusted return on other assets.
This paragraph is where I have a little bit of a problem. I did mention in a previous post that if banks have excess reserves they can convert those to required reserves (in the aggregate) by loaning out funds. The amount that’s converted is 10% of each loan (or whatever the reserve requirements are).
Basically right, but you have to add in the “currency drain,” the portion of loans which are retained by the public as currency in order to meet transactional needs. New loans spur new spending (i.e. higher AD) which creates a need for more currency. I don’t know the exact number, but I’d guess it’s comparable to the reserve ratio so not insignificant.
My problem with that argument is that it makes it sound like the paper bills and coins are what’s driving NGDP growth. Are you sure that’s necessary to your argument? Or perhaps I’m misreading you?
I had a hard time with this as well. I think monetarists would claim currency (or maybe MB) drives NGDP growth. In my opinion it’s very closely correlated but not causal. I think the actual mechanism is that real financial asset prices combined with NGDP growth expectations drive NGDP growth. That growth absorbs the increase in currency… thus the very tight correlation. From a modeling point of view, you could assume that the currency drives the growth and you would get the same result.
So humor me for a minute and suppose we did get rid of paper bills and coins; would that significantly hamper or otherwise change the mechanism you describe above regarding NGDP growth?
Not at all.
I hope I’m clear about my problem with this. In our hypothetical paperless money system, reserves don’t leave the system until the Fed does OMOs to suck them up. They can be (slowly) converted from excess to required by the loan process but that’s it. It shouldn’t matter what the IOR is… there’s no other escape mechanism.
The question is why is this an issue? IMHO ER is only an issue when they rise (dulling Fed policy) or drop (excessively stimulating the economy). When ER drop, there are a few things that can happen.
1) OMO as you note.
2) New banks loans as you note (together with the concomitant currency drain).
3) Raise the IOR rate to keep the ER from draining.
4) Raise the RR ratio and you have immediately converted
ER into RR.
5. February 2013 at 11:20
dtoh, Great! I was prepared to leave it along at this point, so I’m happy you did address my previous post after all. To answer your question:
“The question is why is this an issue?”
It’s only an issue onto itself!… I’m not using that as an argument against you. It’s just that every time I hear the paper money thing brought in it makes me suspicious, and makes the arguments messier. So in a system w/o it, I just wanted to bounce what that would look like off of you regarding what happens to reserves.
So your four points at the bottom… They are a list of reasons ER could drop? The only one I’m having a hard time with (the one that’s not like the others) is:
“3) Raise the IOR rate to keep the ER from draining”
1st that sounds like it’s going the wrong direction if this is a list of things which lower ER.
2nd, does that imply paper money exists or not, or is it not important for this point? In your 2) you mention currency drain again, so I’m not sure.
3rd, it sounds like the mechanism here is really through 2) because in and of itself (say in my simplified world with no paper money) I can see how raising the IOR might disincentivize the banks from making certain loans and thus reduce 2), but I don’t see how raising the IOR w/o 2) does anything. Follow me? Likewise, w/o 2) lowering IOR wouldn’t do anything. If no OMOs are happening, and there’s no net change in outstanding loans, and RR is fixed, and there’s no paper money, you could lower or raise the IOR as much as you want but the banks would be helpless to do anything about their ER. Again, not trying to argue against your overall point, I agree with your “ER is only an issue when..” comment, I just want to clarify and make sure I’m not missing something.
5. February 2013 at 11:31
I should have said “the banks would be helpless to do anything about their ER in aggregate”… of course any individual bank could reduce its ER simply buy buying donuts from a donut shop that held its deposit account at another bank.
5. February 2013 at 11:55
Tom Brown
So your four points at the bottom… They are a list of reasons ER could drop? The only one I’m having a hard time with (the one that’s not like the others) is: 3) Raise the IOR rate to keep the ER from draining
Sorry I wasn’t clear. It was kind of mixed list of ways ER are reduced, and/or what might be done to absorb the ER, or how to stop the fall in ER.
Point 3) was only to suggest that if the drop in ER was considered undesirable, the Fed could slow or reduce it by raising the IOR rate.
5. February 2013 at 12:48
In fact, couldn’t that be a “hot potato effect?” Again, say a situation with
1. No OMOs
2. RR is fixed
3. No paper money
If ER exists on the consolidated bank BS, and IOR is lowered to 0, then in the aggregate, banks are incentivized to grow their balance sheets with loans/deposits to put more interest bearing spreads on their balance sheets. Converting ER to RR in and of itself, however, doesn’t seem like the goal of banks making loans because neither bear interest. However an individual bank making a loan could possibly replace non-interest bearing ER from its BS in a 1:1 ratio with the principal of a new interest bearing loan asset if the borrower uses the deposit to purchase something from a seller having an account at another bank. If no purchase is made by the borrower (rare) or if the seller has its account at the buyer’s bank, the bank still makes money from the spread between what the borrower pays and what it pays the depositor. The bank only converts 10% of the loan principal amount of ERs to RRs in the process. The remaining 90% can then be used to back other customer loans, etc., or used to invest in other assets.
Regarding those other assets (other than creating new loans), individual banks will compete to use their ERs to purchase assets with “risk adjusted returns” greater than 0. But in aggregate, the banks are not getting rid of their ER through this channel. Perhaps “velocity” of money is increasing here though?
So in the aggregate, in this situation (my three points, 1-3 at the top), banks don’t get rid of reserves (ER+RR), but individual banks are incentivized to improve their BSs due to the loss of IOR. Correct?
5. February 2013 at 13:16
Tom,
So assuming no currency (paper money or coins), then you would not get rid of reserves, but instead through the process of lending, redeposit, and re-lending, you would just be converting ER into RR. I agree with this.
To make the analysis even simpler, assume just one bank. In this case (with no currency) either the lender redeposits the funds with the bank, or the lender buys something and the seller deposits the proceeds of the sale with the bank. In this case, any increase in loan balances will be exactly offset by an increase in deposits of the same amount. The only constraint is the RR ratio. Assuming 10%, the bank could lend out up to 1/0.1 (10X) the amount of ER. All of the ER would be converted to RR, but total reserves would not change.
Not sure though where you’re going with this line of thinking.
5. February 2013 at 13:33
dtoh, yes I agree. Just wanted to clarify.
Hopefully final question on this: So let’s take the one bank, no paper money, etc. scenario:
Does this change your argument/mechanism regarding Fed OMPs leading to greater AD?
I’d guess “no,” correct?
5. February 2013 at 13:57
No. Argument model still holds.
6. February 2013 at 02:54
[…] un régimen de polÃtica monetaria por objetivos de PIB nominal (y está feliz con ello), pero Scott Summer argumenta que en realidad no se trata de eso de eso y que la propuesta de The Economist es […]