How does fed control interest rates?

[muscleman]

I only have limited understanding that the Fed can set the Fed funds rate at 0-0.25% and it seems like they will keep this rate low for the next few years. But I noticed the 30 year mortgage rate, which has very high correlation with the 10 and 30 year T-bond rates, have been steadily falling. I wonder what impacts that rate and what scenarios will lead that rate to go up instead of down?

[muscleman]

Could anyone please share some thoughts?

[JimBowerman]

This is probably one of the more complicated subjects but will give it a short explanation and we can go from there.

 

Interest rates in my mind correlate (over the long term) with nominal GDP growth.  And NGDP is determined, in the long run, by the velocity adjusted monetary base.

 

Increase MB by 100x and you should increase NGDP by a similar amount over the long term (i'm ignoring interest on reserves for now to keep it simple)

 

In practice, the Fed controls the monetary base by setting expectations and adjusting short term rates to meet that expectation.  They have a 2% target now, and will adjust the overnight rate to a level where they feel MB will increase enough to match that 2% inflation target.  If expectations are below 2% they can lower the rate, which should increase MB and increase inflation.  However monetary policy is 90% expectations and 10% actual actions, so this can be confusing.  Many times, the expectations can change without a change in actual Fed action.  Dropping short rates to 0% can be stimulative if the market expects the Fed to keep rates at 0% for a long while.  However if Fed signals they'll raise short term rates in the near future, then the initial drop to 0% will have much less stimulative effect.  This happened in 2009, when the market saw that the Fed would raise rates too early (which they did in 2015).  Many folks saw the 0% rates in 2009-2014 as easy monetary policy, when in fact monetary policy was tight during that time period.  Again, expectations are the key. 

 

As a simple thought experiment: If the Fed pegged short rates at 0% forever, what would happen to longterm nominal bond yields? Answer: They'd go sky high and you'd have hyperinflation almost immediately

 

We should judge monetary policy based on an NGDP level target, and we can clearly see that NGDP was well below the pre 2008 trend ever since the GFC.  Monetary policy was entirely too tight for nearly a decade.  I fear we have a similar problem now.  Despite 0% short term rates, long term rates have remained stubbornly low, implying tight monetary policy and low NGDP growth going forward.  The Market correctly sees that monetary policy is currently too tight.  Truly successful QE and easy monetary policy should *raise* long term interest rates, not lower them.

 

And of course, ever since 2008, the Fed has paid interest on reserves.  This means that an injection of QE is much less stimulative because the commercial banks have less incentive to lend. Pre 2008, during open market operations (OMOs), commercial banks received 0% yielding reserves in exchange for say 2% yielding short term treasuries.  Since these reserves can't be disposed of, the commercial banking sector had to lever up its balance sheet by making loans etc, which increased NGDP and draws the link between OMOs and higher NGDP. Now however, the Fed is paying interest on reserves (IOR).  Commercial banks are receiving 2% yielding reserves in place of 2% yielding treasuries.  Comm. Banks incentive to lend more is reduced drastically, showing how paying too high a IOR rate leads to a neutering of OMOs or QE and keeps NGDP too low (as we saw after 2008 and now)

 

Anyways, that at least an (unorganized) start to the conversation :)

 

 

 

[wabuffo]

In practice, the Fed controls the monetary base by setting expectations and adjusting short term rates to meet that expectation.

 

Jim - you keep saying this but I don't think its correct.  Let me ask you - where do incremental bank deposits come from?  How do they originate?  I'm talking in aggregate for the US commercial banking sector in its entirety -- and not for an individual bank like JPM or WFC? 

 

The US commercial banking sector saw an almost $3t increase in total bank deposits in 2020.  What caused that increase? Where did it come from? 

 

Hint:  ...not the Fed.  Further hint: ..... the US Treasury

 

If the Fed pegged short rates at 0% forever, what would happen to longterm nominal bond yields? Answer: They'd go sky high and you'd have hyperinflation almost immediately

 

Again - agree to disagree. In that scenario, I believe new issuance of long bonds would go off at a zero coupon rate as well (regardless of inflation or not).  There are three forms of risk-free money from the US government that the private sector can choose from: 1) currency in circulation, 2) bank reserves, and 3) US Treasury debt.  The first form has no interest rate.  The second form, if set, permanently to zero by the Fed - would cause no 3 to yield zero as well (in fact, some of the long-term debt would go negative YTM due to the mechanics of yield to maturity and bond prices). 

 

Think of it like you/me/we (the private sector) are customers of a bank called the US Treasury Federal Reserve Bank.  We have excess funds so we can hold it in cash (currency in circulation) or in a demand deposit (bank reserves) which pays zero or in a time deposit (US Treasury debt) which can be locked up for 1 month/3 months/10 years/30 years etc.  Since we have excess funds and want the safety of holding them in a bank, we will take any time deposit which gives us even a tiny/teeny interest rate no matter what.  What other choice do we have?  Put our large excess funds under a mattress in cash and banknotes?  So the yield curve asymptotically approaches zero too if the policy of the bank is to never pay an interest rate on demand deposits.

 

Commercial banks are receiving 2% yielding reserves in place of 2% yielding treasuries.  Comm. Banks incentive to lend more is reduced drastically, showing how paying too high a IOR rate leads to a neutering of OMOs or QE and keeps NGDP too low (as we saw after 2008 and now)

 

Again - I do not believe this is how it works.  Reserves have nothing to do with lending.  US commercial banks have $3.2t of reserves and $15.1t of bank credit ($22t of total assets).  As you note banks can't "lend out reserves" since these are inert deposits that are immovable from the Federal Reserve account, so how do you want banks to proceed?  If I understand what you are saying correctly, you are proposing that banks would lend more and balloon their balance sheet size in order to overcome the drag of their zero-interesting earning reserves that in aggregate they are stuck with.  But the ratio required would mean an immense growth in loans to reduce the ratio of reserves to total loans/assets.  Is there enough good credit out there in the private sector for that.  Do banks have enough capital?  Nope and nope. 

 

As an example, let's go back to how it was before the GFC.  Just picking a date in mid-July 2007.  Banks had $5.9b in reserves against $8.9t in bank credit (and $12.1t of total assets).  So if your goal is to force banks to lend so that they maintain the same ratio as then, the US commercial banking sector would have to expand their loan book 531x!  They would have to lend $4.7 quadrillion dollars!  The US economy is only $22t in size.

 

Banks don't lend based on what's happening to their reserves.  They lend based on their capital and their credit evaluation.  Reserves (and what they're paid on them) never enters the equation at all.

 

But that's just my 2-cents.

[DocSnowball]

http://aswathdamodaran.blogspot.com/2015/09/the-fed-interest-rates-and-stock-prices.html

 

 

From 2015, but the Myth or Truth discussion in the post remains evergreen:

1. The Fed sets interest rates

2. Low interest rates are the Fed's doing

3. The reason stock prices are so high is because rates are low

 

[Guest cherzeca]

a short answer: fed doesn't control long term rates. it controls short term rates...because it sets the rate at which it lends short term...mostly to banks...who dont like borrowing from the fed much, looks shabby. lately the fed has had an effect on long term rates by buying treasuries and agency securities.

[wabuffo]

Let's review this all-important Fed interest rate (and the role of the Fed):

 

1) it's an interest rate that is paid on reserves.

2) reserves are stuck at the Fed and can't go anywhere else.  You and I can't get a "reserve" account at the Fed.  No one actually wants the Fed's reserves - they are illiquid. Banks don't create reserves - the Fed does.  Nothing that banks do increases the size and quantity of reserves.  Only the Fed can increase or shrink the size of reserves.

3) reserves don't affect bank lending.  Banks lending is affected by bank capital and regulatory ratios.  It has never been affected by reserves - nor do banks need reserves to lend.  In fact, "required reserves" are now temporarily set at zero.

 

The Fed is like the Wizard of Oz - so long as you ignore that little man behind the curtain.  The real reason the Fed exists is to do one job - run the payments system and make sure no inter-bank payment in the United States ever fails to clear and settle.  It does this by using reserve accounts to clear payments between banks and between banks and the US Treasury.  The Fed also buys and sells assets with the banks and uses reserves to do this.  The banks have no choice in the matter.

 

wabuffo

[muscleman]

Thank you for the discussion guys. It is mind blowing to me how complicated an answer can be and I am unable to understand what Jim is saying.  ::)

Thank you to Chris's short version. Yeah it was my understanding too that the Fed sets the short term rate, which the banks can borrow from it at that rate for as much as they want.

But if they are also buying long term treasuries, then would there be a case where the long term rates would go up? Would that ever happen unless FED stops buying those long term treasuries?

It blows my mind how low the 30 yr mortgage rates are right now. It was 5% in 2009 and it looks low. Then it was 4%, and then 3% looks like a steal, and now it is 2.6%.

[tede02]

Jeffrey Gundlach has expressed a view that he believes the Fed is in-fact manipulating long-term rates via open market purchases. It's an interesting question whether or not long rates could go higher if the Fed doesn't want them to. You would think at some point, if inflation started heating up for example, that market forces would over-power the Fed's actions.

 

One interesting relationship I've spent more time on this year is betweeen the 10 year Treasury and 30 year mortgage. The correlation between the two is nearly 1. My understanding of this is most mortgages are guaranteed via Fannie/Freddie, and therefore ultimately by taxpayers. Thus the market views Fannie/Freddie mortgages (and ultimately mortgage backed securities) as similar credit risk as Treasurys. But if you look at this relationship, the long-term average spread between the 10 year Treasury and 30 year mortgage is about 1.5%. I'm not an expert in this area by any means but one of the main reasons cited for the 1.5% spread over the 10 year is the pre-payment risk that comes with mortgage bonds. Watching the spread seems to be one of the best ways to determine how attractive mortgage rates are. For instance the spread gapped out last spring when the 10 year cratered all the way down to like 0.50%. Even now, the spread over treasurys is higher than average which suggests rates could be lower.

 

Attached is a chart I keep on spreads. Couldn't figure out how to get it into message body.

 

[cameronfen]

Thank you for the discussion guys. It is mind blowing to me how complicated an answer can be and I am unable to understand what Jim is saying.  ::)

 

I’m with you muscleman and I’m a PhD student in economics.  For what it’s worth I think Jim is expressing the conventional agreed upon wisdom.  Is it right? I don’t know as this isn’t my thing. 

[LC]

https://www.investopedia.com/articles/economics/08/monetary-policy-recession.asp

[JimBowerman]

Cameron, I'm largely parroting here the market monetarist school of thought (Scott Sumner and George Selgin both have blogs that I largely agree with).  I'm oversimplying a bit, but the old school "Monetarist" school of thought was close to accurate, but critically assumed the money velocity was constant.  Market monetarists update this to focus on the "Velocity adjusted, monetary base"...Basically they target NGDP (and wish to automate most of the Fed's actions by setting up an NGDP futures market and making open market operations largely automatic whenever these NGDP futures deviate from a 4 or 5% growth path - ie making dollars convertible (in unlimited amounts) into NGDP futures.

 

WaBuffo, always enjoy out discussion here and on twitter...Don't have a ton of time, and sorry if I'm repetitive with some comments...understandable if we want to agree to disagree, but with that said...

 

Whats the point of QE if reserves are simply for clearing payments?  If QE is not stimulative, then why don't we do more QE and collect less taxes? Is it not a free lunch? Again, sorry if i'm repetitive...i think we've had this conversation before but I do enjoy them!  ;D

 

(I'd also note that its MB [ie reserves + cash] that should correlate to NGDP over the long term.  Reserves stayed flat or even dropped for much of the rapid NGDP growth period from 1970s and 1980s...but MB grew much more so.  looking at just reserves can lead to wrong conclusions)

 

Regarding "They would have to lend $4.7 quadrillion dollars!" - You can't do a simple ratio like this because of IOR.  Comparing reserve amounts pre 2008 vs post 2008 isn't possible because IOR changes everything.  That said, if the Fed did inject the amount of reserves that they've done, and did NOT pay IOR, then we'd certainly have hyperinflation and well on our way to that $4.7 quadrillion.  And yes, while the current IOR rate is only 0.1%, the market correctly expects the Fed to raise IOR at the first sign of overheating...which in turn keeps inflation moderate.

 

 

[cameronfen]

Cameron, I'm largely parroting here the market monetarist school of thought (Scott Sumner and George Selgin both have blogs that I largely agree with). 

 

Scott Sumner = NGDP I know that much :).  Although it’s that’s definitely not totally orthodox in economics.  Can you tell I don’t know what I’m talking about...

[JimBowerman]

Cameron, Yes Scott Sumner and market monetarists are far from mainstream.

 

At least by my rough twitter count, It doesn't seem like there are more than a few dozen people in the world who are outspoken supporters of market monetarism. 

Hopefully that changes over time.

 

Market monetarists are some of the few (only?) folks who correctly see that the goal of QE should be to *raise* long term bond yields (via higher NGDP growth).  I wish folks on the FOMC would read more from: Scott Sumner, George Selgin, Lars Christensen, Marcus Nunes, David Beckworth, etc.  This talk of yield curve control, etc will only bake in lower growth expectations (which comes from from incorrectly thinking that the goal of QE should be to lower bond yields - its the exact opposite!)

 

 

[Guest cherzeca]

no need to get theoretical here, just look at practice.  Fed issues money with a keystroke (as opposed to Treasury, which just borrows money). Fed can buy securities with "hit the return key" money. voila! so it controls a lot of the liquidity of the economy, which may be more important than borrowing costs...interest rate setting is only the beginning.

[wabuffo]

Fed issues money with a keystroke (as opposed to Treasury, which just borrows money)

 

nope not how it works.  The US Treasury spends first, which creates a new bank deposit somewhere in the private sector (and a matching reserve deposit at the Fed for the bank receiving the deposit).  US Treasury then borrows to remove the reserve it creates in the banking system with its spending by replacing it with an interest-earning time deposit (ie, US Treasury bond) held at the Fed in a custodial account.

 

The Fed is an asset swapper - its only asset is a reserve balance which can't go anywhere except inside the Fed's payment clearing account system.  So despite the hoopla, the Fed is basically inert and can't do diddly when it comes to creating new financial assets in the private sector.

 

I think you have it backwards, I'm afraid.  Money creation comes from the US Treasury.  That's why everyone wants them $600 cheques.

 

wabuffo

[Guest cherzeca]

"US Treasury then borrows to remove the reserve it creates in the banking system with its spending by replacing it with an interest-earning time deposit (ie, US Treasury bond) held at the Fed in a custodial account."

 

nope, unless Treasury is borrowing from Fed...which it doesn't. that bond its held by the Treasury buyer/extender of credit...which is not Fed

[wabuffo]

Whats the point of QE if reserves are simply for clearing payments?

 

I agree QE is a pointless exercise.  Its stupid in too many ways to count.  We don't need it and it makes things worse.  And here I am refering to the exercise of the Fed forcing the banks to hold excess reserves and then adding to those already unproductive reserves by buying Treasuries and swapping them for even more reserves.

 

Reserves stayed flat or even dropped for much of the rapid NGDP growth period from 1970s and 1980s...but MB grew much more so

 

One has nothing to do with the other.  Reserves have nothing to do with lending.  To the extent the monetary base grows, it is growing from increasing deposits in the US commercial banking sector.  You didn't answer my question - where do incremental bank deposits come from?  I'll answer it - from net US Treasury deficit spending.  So I can rephrase your comment this way:

Reserves dropped because banks never needed them to lend.  Growth in MB didn't come from reserves but instead from the larger deficits of the 1980s, which created a surge of new deposits.  See?  The Fed and its reserves had nothing to do with it.

 

That said, if the Fed did inject the amount of reserves that they've done, and did NOT pay IOR, then we'd certainly have hyperinflation

 

How does the Fed "inject" reserves? It swaps them for Treasuries - ie an asset swap with the private sector.  Reserves can't go anywhere.  They're stuck.  So even if Fed doesn't pay an interest rate (or charges an interest expense) - what can the banks do?  Nothing.  They are stuck with those reserves.  They can't lend them, they can't reduce them.  The only way they can in theory reduce them is to exchange them for currency.  I'd like to see that.  Have JPM, WFC, BAC etc ask for $3.2 trillion in tens, twenties and hundred dollar bills.

 

wabuffo

[wabuffo]

nope, unless Treasury is borrowing from Fed...which it doesn't. that bond its held by the Treasury buyer/extender of credit...which is not Fed

 

Repeat after me - US Treasury "borrowing" is a reserve maintenance function.  It really isn't borrowing despite its appearances.

 

US Treasury sends you a $600 stimulus cheque.  The Fed moves $600 from the US Treasury general account to your bank's reserve account thus performing the Fed's main job (payment clearing).  This is a new financial asset for the private sector (your bank gets both an asset and a liability.  A $600 deposit in its reserve account at the Fed and a $600 deposit liability to you).

 

You then decide that you don't want to hold $600 in cash at the bank cuz it doesnt pay any interest.  So you buy T-bills with the $600.  Your bank now exchanges its $600 reserve balance on your behalf (and your $600 deposit) for $600 of T-bills.

 

The private sector (you) still has an asset ($600 of T-bills).  Your bank has neither $600 of reserves nor deposits.  The issuance of T-bills has removed $600 of bank reserves that the initial stimulus cheque created.  Thus reserve maintenance activity. 

 

Let's review:

US Treasury spend -- new financial asset

US Treasury borrow -- asset swap, financial asset still exists but in different form.

 

wabuffo

[Guest cherzeca]

repeat after me. treasury sells/issues a t bill to pimco. fed is not involved. pimco credits (reduces) cash, debits (increases) securities account. treasury debits cash, credits liabilities.

[JimBowerman]

Whats the point of QE if reserves are simply for clearing payments?

 

I agree QE is a pointless exercise.  Its stupid in too many ways to count.  We don't need it and it makes things worse.  And here I am refering to the exercise of the Fed forcing the banks to hold excess reserves and then adding to those already unproductive reserves by buying Treasuries and swapping them for even more reserves.

 

 

This is our fundamental disagreement right here.  If the Fed does QE to the tune of injecting $1,000,000,000,000,000,000,000,000 worth of new reserves and buys literally every asset on the planet, you're saying it will have no effect on bank lending?

 

 

[Spekulatius]

repeat after me. treasury sells/issues a t bill to pimco. fed is not involved. pimco credits (reduces) cash, debits (increases) securities account. treasury debits cash, credits liabilities.

 

I can see wabuffo‘s point. If the treasury issues a t-bill to Pimco, it would in exchange for cash. Now the treasury would hold cash. Nothing would change (no money creation , just an asset swap)  until they spent the cash.

[Cigarbutt]

During 2020 (‘recession’ period), the Case-Shiller house price index went up by almost 10% (!). Why?

 

 

For many reasons and (very) low interest rates? Isn’t the Fed driving interest rates down?

 

Share of the Fed-held agency MBS vs total agency MBS (%)

Before 2009: 0%

2010: 15%    2012: 13%    2014: 24%    2016: 23%    2018: 20% (attempt at ‘tapering’ and in 2019, withdrawal tantrums)

2019: 16%    end 2020: about 23% (in 2020, there was record MBS issuance, re-fi activity and Fed involvement)

Isn’t the Fed an essential player here even if only a swapper? In early March, their extra involvement was key to restore the volatility in spread that happened in the agency MBS market but the market has become habituated to their long-term ‘involvement’. Opinion: Like the Eagles song, the Fed can check out anytime but can never leave.

 

The 30-yr mortgage rate is influenced by the effective Fed fund rate and is tightly correlated to inflation.

 

 

If you know where inflation is going, you know where the 30-yr mortgage rate is going. Note: crossing zero though is mostly uncharted territory.

 

[wabuffo]

If the treasury issues a t-bill to Pimco, it would in exchange for cash. Now the treasury would hold cash. Nothing would change (no money creation , just an asset swap)  until they spent the cash

 

you can also see it empirically.  In 2020, the US Treasury spent $3 trillion more than it took in via taxes, fees (think CARES Act, etc).  What was the increase in all US commercial banking deposits in 2020?  Answer: $2.94 trillion.  Its a match (well almost - I can never get it exactly to match which makes me think there's leakage into currency in circulation as some people take money out of their ATMs, etc).  This happens every year if you flip through the Fed H.4.1 and H.8 reports.

 

wabuffo

 

 

[Guest cherzeca]

@cigar

 

fed has become a big buyer of treasury and agency mbs, and so is definitely providing cash demand (which the fed key stokes on its magic computer) to meet issuance supply, thereby lowering long term rates. but my view is that fed cant control long term rates...at some point, you will see long term rates trend up. that point will be when the dollar has become so weak that foreign sovereign funds etc find attractive purchases elsewhere...which is Chairman Xi's fondest wish.  yes, this will become a foreign policy issue, not just a finance issue.