Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I've been trying to figure out what SVB's reserve requirements were and found (https://www.federalreserve.gov/monetarypolicy/reservereq.htm):

"As announced on March 15, 2020, the Board reduced reserve requirement ratios to zero percent effective March 26, 2020. This action eliminated reserve requirements for all depository institutions."

Wat?



Reserve requirements are pretty silly. Many countries never had them.

What you want to keep banks afloat are capital buffers, not reserves.

Expressed differently: reserves are like cash in a vault (or electronic equivalents). What you want instead is a big buffer of equity in the capital structure, so that shareholders can absorb huge losses long before creditors do.

In regulatory terms something like this is called 'minimum capital adequacy ratio'. But it's generally better to set up the rules of the game so that banks naturally want to have more of an equity cushion, instead of giving them strict rules on capital buffers but leave lots of incentives to work around those rules.

As an example of incentives: many tax codes around the world let you pay interest with pre-tax money but dividends have to be paid with post-tax money. (That's simplified, since there's lots of different taxes.)


Slightly ironic that you comment about the silliness of reserves on a post about insolvency of a bank that likely had little reserves


Right, but reserve requirements are in the order of a few percent, certainly not enough to prevent a bank run, or preventing you from being insolvent when the bonds you're holding went down in value by 20% due to interest rate movements. At the end of the day what you actually care about is whether you have more assets (cash or equity) than liabilities.


Reserves are but one tool in the arsenal and it's also for tightly managing the maximum leverage. If all your assets are not cash-equivalent and they fluctuate in value then there is no maximum leverage that can be guaranteed a priori and there is operational risk.


Yes, reserve requirements don't guarantee any maximum leverage either.

Suppose your assets are 50 dollar reserves and 50 dollar investments.

The liability side of your balance sheet is 20 dollars of equity and 80 dollars deposits. For a leverage ratio of 1:4.

If the investments drop 10 dollars in value (to 40 dollars), your leverage ratio goes to 1:8.

If the investments drop 20 dollars in value (to 30 dollars) your leverage ratio goes 1:infinity.

If your investments drop below 30 dollars (say to zero), you are insolvent.

Yes, reserves are a tool that banks can use. But that doesn't mean that legal minimum reserve requirements are a good idea.


Reserve money and insurance are very silly until you need them.


Rubber duckies are very silly, until you need them. (Afterwards, too.)

Just to be clear: the problem is that SVB didn't have enough loss absorbing equity. If they had more, they wouldn't be insolvent.


Why? I don't see any irony here.

The bank also had very few chocolate coins (probably none), but that doesn't mean requiring them to hold more of those in their vault would have improved matters.

The problem is that the bank is insolvent; not so much that the bank is out of liquidity.

If they were solvent, someone would lend them the liquidity they need.


I get that other countries are also doing this but “oh this bag is equities is safe because it’s diverse” proving to be false all the time feels like a good argument against this logic.

Granted if SVB had all these bonds that would pay out “guaranteed” that feels pretty strong


The problem was specifically that they were not diverse, though?

And, absence a run, they were somewhat safe. They effectively concentrated all of their risk in this category. And then got hit there


How are they proving to be false all the time? The FDIC said the last bank they took over was in 2020.


What I was glibly saying was that “we have securities X Y and Z that are in aggregate worth T dollars at current market prices, thus this is like we have T liquid dollars” shortcuts have lead to so many problems when macro economics happen.

I am being glib, though I am very wary of the safety of things that aren’t just like… cash. “We haven’t had to take over a bank for 2 years!” Isn’t as much of a vote of confidence in a system as I’d like.


I’d be curious to hear why such a cadence (let’s say 15 years) is an acceptable price to pay in your book. Do you think this system leads to such substantive increases in American standard of living that it’s worth giant banks toppling do often?


Just treat these leveraged entities like SEC would have brokerages treat margin accounts. There are different leverage ratios counted depending on the risk of the position and if you don't maintain margin you get called and liquidated. Such regulations should be uncontroversial but I guess SVB lobbied for infinite margin and then used it!


What makes you think they had infinite margin? (Or lobbied for it?)

They had assets on their balance sheet that lost in value. They lost enough value to wipe out all the equity, and become insolvent.

Very similar to how your brokerage account can go to zero, if you trade on margin.

The brokerage liquidating your account is pretty similar to the FDIC taking over SVB.

(Yes, in a technical sense SVB went to infinite leverage for a brief moment. Just your brokerage account can go to zero if the market moves faster than your broker pre-emptively liquidates your stuff.)


>Expressed differently: reserves are like cash in a vault (or electronic equivalents). What you want instead is a big buffer of equity in the capital structure, so that shareholders can absorb huge losses long before creditors do.

That sounds like a Ponzi scheme. Wiki:

>A Ponzi scheme (/ˈpɒnzi/, Italian: [ˈpontsi]) is a form of fraud that lures investors and pays profits to earlier investors with funds from more recent investors.


Sure, so when the market goes down and depositors want their money back, they can't.


Pick a bank that only invests in stuff you think won't go down in nominal (!) value.

Like eg a money market fund that sticks to short term government debt. Or a 'narrow' bank.


Kind of happens anyway if there is a bank run, and doesn't happen if there isn't a bank run.


The reserve requirement ratio is the amount of money the bank needs to hold at its account with one of the Federal Reserve Banks per dollar of deposits. A literal pile of cash would not count a single cent towards that requirement.

Modern requirements are based on core capital ratio, which is basically the ratio of a pile of cash that is set aside to deal with losses of assets as a fraction of the assets (weighted by risk, so, e.g., you don't need to set aside any money to protect against a literal pile of cash but you need lots of money to protect against a shitton of shitcoins).


I mostly agree.

> A literal pile of cash would not count a single cent towards that requirement.

Are you sure about that? Do you have a source?

As far as I am aware, vault cash is fine, just way less convenient than an account at the Fed.

But I admit that I don't know the exact rules, and would be happy to be proven wrong.

> Modern requirements are based on core capital ratio [...]

I think you can strike the word 'modern' from that sentence. Capital cushions have been a thing for a long, long time. What's 'modern' is that the US mostly stopped having reserve requirements (though the new rules are written in such a way as to all-but force American banks to hold lots of American government debt as a sort-of reserve in disguise).


You are conflating two things. Reserve requirements are not same as capital. And there is a modern capital requirement, as per Basel III, in effect since 2022. Capital historically was calculated very differently.


> You are conflating two things. Reserve requirements are not same as capital.

I know. I never wanted to imply otherwise. What made you think so?

> And there is a modern capital requirement, as per Basel III, in effect since 2022. Capital historically was calculated very differently.

Yes, the rules change all the time. So there's a modern incarnation of capital requirements. But capital requirements in general aren't new.


I've gotten my information from https://www.federalreserve.gov/monetarypolicy/reserve-mainte..., although I may be misreading it.


Thanks!

Following a link from there, I get to https://www.federalreserve.gov/monetarypolicy/reserve-mainte... and this says:

> During each reserve maintenance period an institution must satisfy its reserve requirement in the form of vault cash or, if vault cash is insufficient to satisfy the requirement, in the form of a balance maintained with a Federal Reserve Bank. The portion of the reserve requirement not satisfied by vault cash is called the reserve balance requirement.

In any case, I'm still agreeing with you!


The thing to understand about the change in policy regime that led to the elimination of the reserve requirement is this: banks were (and are) holding massively more reserves than they needed to. They didn’t need the Fed to tell them to do this, so it was not an effective monetary policy tool any more.

In economic terms: the supply of reserves was so high that it was in an inelastic part of the demand curve for reserves.

The new regime (“ample reserves”) depends on administered interest rates, rather than reserve requirements, to set short-term interest rates.

It is absolutely not the case that the reserve requirements were removed to allow banks to reduce their reserves.


All cars have seat belts. It isn't effective policy to require them.


You need to extend your metaphor.

Imagine instead that the government was paying car manufacturers more than the cost of installation for every seat belt in a car.

You wouldn’t need to have a policy requiring seatbelts: every car would be liberally festooned with them.


But it costs the taxpayer money that way via the interest payments, whereas one more regulation doesn't.


Honestly, if you don't understand the difference between the Treasury and the Federal Reserve (as this comment suggests), I'm sorry to say that you're not going to find this conversation very enlightening.


You must not know that the Federal Reserve returns income to the Treasury.


In 2008, the Fed started paying interest on excess reserves.


They removed reserve requirements (ie. cash you need to keep on hand) but there's still capital requirements (ie. assets you need to keep on hand).


They relaxed the need for a capital conservation buffer, too. These actions were part of the Fed's response to the market panic at the onset of the pandemic, and you can read a bit about their reasoning in their own press release:

https://www.federalreserve.gov/newsevents/pressreleases/mone...

Capital requirements are complicated. There are different types of equity and assets are risk-weighted, but treasuries and the types of bonds SVB bought are generally given the lowest risk-weightings.


Funny that they didn't deem T-Bonds to be high risk given that it was they who crashed the value of those T-Bond by increasing interest rates.


You identified the root cause of this whole thing. Long-term bonds get preferential treatment in the risk-based capital calculation because they have low default risk, which allows banks to run with thinner equity capital, even though they're exposed to high interest rate risk and have a portfolio highly concentrated in one high-risk industry. Capital ratios don't adequately account for actual risk.


Yes. That's (part of) why FDIC is a bad idea: it partially insulates depositors from the need to monitor what their bank is doing.

In banking, we want a flight to quality.


FDIC insulates customers, but it does not at all insulate the banks; if a bank fails, it fails, FDIC or no. So it doesn't at all remove the incentive for banks to avoid failure by not making poor investments in the same way that, say, a bailout might.


It doesn't have to be all or nothing, and the current cap on insured deposits is a nice balance.

Small depositors aren't finance pros, they don't have the training or network to monitor bank management, so we protect them from rogue management.

Large depositors can and should worry about their bank's solvency, so we focus their minds by leaving their deposits uninsured.

Unfortunately, when large depositors catch a whiff of insolvency, they don't help fix the problem, they're first to pull their deposits and leave everyone else to pay the bill.

So we should treat them as we treat creditors in a bankruptcy, and claw back the cash they were able to withdraw in the days leading up to the bank's closure.


And the FAA is a bad idea because it insulates travelers from the need to monitor what their plane is doing?


Maybe? You could replace the FAA with voluntary certification. People could demand the same standards as the FAA enforces.

Let me bite the bullet:

Flying is arguably way too safe. In the sense that flying could compromise a bit on safety, and still be much safer than cars. If that compromise would lead to lower costs and thus prices, perhaps more people would fly and fewer would drive; leading to better safety on average. Despite flying becoming less safe.


Never thought I would hear someone making that argument. You think we should let more people die in plane crashes so that flying is cheaper?

We already know cheap flying with current safety standards is economically viable, with costs approaching the cost of fuel; see low cost airlines like Ryanair. So the most effective way to lower prices is for airlines to downgrade amenities and services onboard.


Let me bite here.

I think flying could be made cheaper and massively more convenient by getting rid of airport security the way it is done right now. Imagine planes being boarded like trains: there might be a security tradeoff but it would be offset by the massive time benefit for everyone involved. I know I would take that risk, I consider my own time to be more worthy than TSA considers it to be


I don't think anyone is arguing about security to board a plane. They're talking about maintenance requirements and safety requires for when a plane is in flight. Things like airplane inspections.


I'm talking about all aspects.

As an example: planes almost never crash. Which is great! So you could drop the requirement to carry life-vests, without compromising safety numbers.

(If you want, you can invest 50% of the cost savings into eg anti-malaria nets, and you'd come out way ahead in terms of lives saved.)

Similarly, airplane seats are massively over-engineered. You could loosen restrictions there, and save weight and thus costs.

I think Japan might already have different domestic regulations there. I remember being on a domestic flight between Kobe and Tokyo, and the seats in economy class used a lot of mesh over aluminium frame or so. They looked a lot lighter (and also more breathable) than your typical airplane seats. See https://photos.app.goo.gl/ZDWQTNMB93mQs5Yz5 for a picture that I took.

I also agree with vlack-vingaard, but they already gave some good examples.


This is true, we could definitely do with less "compulsory participatory theatre" at airports.

That said, the FAA != TSA. We like the FAA. They solve actual problems.


My guess is that you have two states of equilibrium. One is that accidents happen all the time. The other is that accidents are extremely rare. Both states can exist, but the middle ground is rare and impossible to engineer. The error in safety margins of overlapping complex systems is just too great.


That's a very abstract argument, and doesn't take into account any special features of planes.

So I can say: I just want planes to be as safe as, say, trains. (Or whatever other form of transportation is safer than individual cars, but not as over-burdened as planes. Perhaps busses?)


> could replace the FAA with voluntary certification

How does that work when a plane lands on my house?


That one post seems to say, too bad, you die but it's for a good cause because two other people didn't die in a car crash.


Essentially, yes.

It's very similar to why we allow cars to crash so much in the first place. To paraphrase JumpCrisscross:

> How does that work when a [car runs me over]?


Idk I think retail depositors should be comfortable leaving their money anywhere that’s fdic insured and it’s the bank shareholders that should have to monitor what they’re doing.


We want more monitoring in the financial system, not less.

Otherwise you could make the same argument you just made, and expand it to: shareholders should be comfortable, it should be regulators (or someone else) that should be monitoring, etc.


It's a good idea: it does not protect shareholders at all.


We want more people in the financial system to be aware of risks and pro-actively deal with them. That includes depositors.

(And really skittish depositors could switch to banks that only invest their deposits in eg government bonds. Which are probably about as safe as FDIC insurance.)


Well, it was their interest risk, not their credit risk, that caused the problem here.

But it's an interesting thought.


Yes. Though as a small nitpick, I wouldn't describe capital requirements as 'assets you need to keep on hand'.

No matter your capital structure, all your liabilities will be matched by assets. What matters is that after your accounting for your fixed liabilities, like deposits, you still have plenty of total assets left over to have a thick equity cushion to absorb losses.

In accounting terms, equity is also a liability. But it's a very benign one, as your shareholders can't demand their money back.

There are other forms of liabilities that act like equity in their ability to absorb losses. But equity is the simplest and generally the most import one.


Wow, I never thought of / heard anyone explain equity as a liability to pay to shareholders. It’s brilliant!


You might like to read up on the very basics of accounting. There's lots of other interesting concepts there.

See https://martin.kleppmann.com/2011/03/07/accounting-for-compu... for an intro.

Accounting might sound rather boring, but at its core its about understanding businesses (and economies) with numbers. It can be as varied and interesting as companies are.

Of course, in practice there's lots of cruft build on top of relatively simple concepts. But the simple underlying concepts are still fascinating. The link above explains double entry book keeping in terms of graph theory and network flows.

The basics of deprecation are also quite interesting (to me, at least).


Just a small comment on this:

To me, accountants have a better grip on reality than economists. It's an accountant who taught me real economics (I was originally pursuing a math degree with economy as a 'minor' (not really how it's working in my country but close enough)). I had to unlearn some of what I learned in my first year, but I had a way better grip on how money work after that (and decided to create value and changed course).


Yes. Well, carpenters have an even better grip on reality.

Accountants and economists are doing different things. Both fields are useful, and there's some small overlap between the two.

Many people could benefit from learning some 'rationalised' accounting, ie accounting without the accumulated historical accidents and tax dodges. (Those are also interesting. But less as a description of a reality, and more in the same vein that the Talmud is interesting.)


But orthodox economic theory goes against most thing i learned. To be honest, i really thought it didn't matter, that macroeconomics was on its own, and that economists are valid expert to listen to. We were still in eurozone crisis, and while i thought "This plan doesn't make much sense" when the Troika laid out what Greece should do, GDP and socioeconomic markers were not stuff i learned or cared about.

And then like 5 years ago, i learn about MMT, read about it, disagree on some points, but it overall make much, much more sense and si way closer to reality than Friedman theories to me. It seems like macroeconomics do follow the stuff i learned when i wanted to become a quantitative analyst or whatever (i only wanted to do math tbh, and didn't follow finance classes that much).

And then during Covid we have all those "expert" economists who start to talk everywhere. But now, i am sure they are talking out of their own asses. They had now idea of what production is. The simple idea that production is linked with energy is novel for them. They probably are useful, like sociologists are useful, but i'd like to hear them on medias as much as i hear sociologists. Or even less, since i do think sociologists have real-world application to their thesis, for harm reduction during stampede. Let's say as much as medievalists historians.


MMT is both novel and correct. The problem is that the novel parts aren't correct, and the correct parts aren't novel.

Have a look at https://www.econlib.org/library/Columns/y2021/Sumnermodernmo... for MMT.

I suggest having a look at market monetarism. See eg https://marketmonetarist.com/2015/07/14/the-euro-a-monetary-...

What kind of orthodox economy theory have you had a look at?


The experts are the ones who engineered decades of low interest rates so their buddies can get cheap loans and play financial games with the economy. They've never been trustworthy.


My little brain sees this as a reason for failure here.

SVB had liquid assets (bonds) that it could sell to meet the demands of the depositors, but those assets fell in value creating losses. The bank was forced to crystalise those losses because it had no cash buffer to fall back on.

If some substantial part of the liquidity had been held as cash this wouldn't have happened.


No. This week, the treasuries and other assets in capital are effectively cash equivalents.


The worst part of it, is that once interest rates started raising and they started seeing the obvious (but not critical!) impact on their long-maturity bonds, they simply could have switched to shorter-maturity bonds and be totally fine.

But they preferred to gamble.


What I don’t understand is why they weren’t restored as part of the effort to tamp down inflation, and instead the fed just raised rates. Seems like that should’ve been the first move.


Isn't cash kept on hand the relevant attribute here? Whether they're solvent or not, they ran out of cash to meet withdrawals.


The textbook story about banks taking in deposits and lending some of them back out hasn't been true in modern economies for a long time. Banks don't need deposits to make loans.

It's a long story, so I always recommend going down the Modern Monetary Theory rabbit hole.


Old news ;)


If I get a loan from Bank A, then I use that loan to pay a person who deposits the IOU into Bank B. Bank B will go to Bank A and demand the money in cash because it's a competitor bank. If bank A has zero cash on hand they immediately hit a bank run, so basically bank A wants to keep a certain ratio at all times.

Thus through the existence of competitor banks, banks are NATURALLY incentivized to keep a reserve ratio. A reserve ratio enforced by law is not necessary in a capitalist economy with healthy competition. Competition prevents banks from going crazy with creating money out of thin air via loans. The removal of the reserve ratio by the government is relatively inconsequential.

However this natural regulation through competition is negated by the existence of an entity without competition. The central bank. The central bank functions as an entity that loans money to banks with interest. It is this interest rate that is used to regulate the money supply in the US. Low interest rates are what caused inflation and high interest rates from the central bank are what are now being used to stop inflation.

So in this case Bank A can now borrow a bunch of money from the Central Bank thereby increasing it's reserve ratio allowing it to lend more money out. In a sense, the central bank is essentially the entity where the fractional reserve ratio actually matters.

The central bank is unregulated so they can print money to loan to other banks however much they like. Thus a bank run on the central bank is impossible. The ratio in this case matters more as a metric that correlates with inflation.


I have great sympathy for your argument, and agree with the gist of it.

What you are describing is pretty close to the free banking eras of eg Scotland and Canada.

> The central bank is unregulated [...]

That's not true. Many central banks have lots of regulations on them. However, they are not regulated by the kind of competition you outlined above.

> [...] The central bank functions as an entity that loans money to banks with interest. It is this interest rate that is used to regulate the money supply in the US. [...]

It's probably more productive to think in terms of the total money supply, and less in terms of interest rates.

For one, loaning money to banks is only one part of what the Fed does. They also outright buy and sell assets (eg in open market transactions). In many instances, the banks (technically) lend money to the Fed by having positive account balances at the Fed.

For a contrasting example on how interest rates don't need to be the focus of monetary policy, have a look at the Monetary Authority of Singapore. Instead of using interest rates as a channel to communicate and effect their monetary policy, they use the exchange rate of the Singapore dollar to a basket of foreign currencies. Crudely, instead of 'setting' the interest rate, they 'set' the exchange rate.

Simplified a bit, they 'set' the exchange rate by standing by to buy and sell Singapore dollar to any comer. They have a printing press, so they can push down the exchange rate as much as they want to, and they also have enough assets to prop it up.

Crucially, this framework doesn't need to worry about any zero bound on interest rates. It works as long as Singapore dollars are worth anything more than zero.


> I have great sympathy for your argument, and agree with the gist of it.

I'm not making an argument. I'm stating the current status quo of the US. No argument was ever made here about whether I think it's right or wrong.

>That's not true. Many central banks have lots of regulations on them. However, they are not regulated by the kind of competition you outlined above.

It is true. The central bank is overall unregulated because the central bank IS the regulator. In the same way a government is unregulated so is the central bank. In the US the central bank is more or less the fourth branch of the government.

You're talking about "many central banks." while I'm simply talking about the Federal reserve in the US. I think you're mistaken, I'm not making a general statement about how central banks across the world works.

>For one, loaning money to banks is only one part of what the Fed does. They also outright buy and sell assets (eg in open market transactions). In many instances, the banks (technically) lend money to the Fed by having positive account balances at the Fed.

This is true. However one of the primary ways they influence the money supply is through interest rates. Interest rates are also one of the triggers of the SVB bank run.

>For a contrasting example on how interest rates don't need to be the focus of monetary policy, have a look at the Monetary Authority of Singapore. Instead of using interest rates as a channel to communicate and effect their monetary policy, they use the exchange rate of the Singapore dollar to a basket of foreign currencies. Crudely, instead of 'setting' the interest rate, they 'set' the exchange rate.

They don't need to be, but they ARE quite central in the US. Additionally given how the US dollar is sort of the central peg of all other currencies, the US would rather the Dollar remain the Rate at which all other currencies are set against. That way the US in a way indirectly and collectively controls the worlds monetary value.

I didn't offer any opinions in my initial reply. I'm simply stating what's going on in the US about the nature of the reserve ratio and how it doesn't matter when applied to SVB. It seems you're trying to make an argument here against one I never made?


> The central bank is overall unregulated because the central bank IS the regulator.

Even regulators are regulated. There are laws that prescribe what the Fed can and can not do, and how.

> Additionally given how the US dollar is sort of the central peg of all other currencies, the US would rather the Dollar remain the Rate at which all other currencies are set against. That way the US in a way indirectly and collectively controls the worlds monetary value.

Yes, if you wanted to do a similar system for the USD, you would probably want to peg a basket of commodities instead of the exchange rate.

Or you could have the Fed target the TIPS spread directly: https://fred.stlouisfed.org/series/T10YIE


> Even regulators are regulated

Ron Paul campaigned for "auditing the Fed" perhaps more than he campaigned for president. Was he exaggerating, or does Congress not actually audit and otherwise oversee the Fed?


The Fed gets regular audits as far as I can tell. But perhaps Ron Paul was campaigning for more thorough ones? (Or it was just a good sound bite?)

https://www.econlib.org/archives/2009/07/audit_the_fed_o.htm... and https://www.csmonitor.com/Commentary/Opinion/2009/0803/p09s0... might be interesting.

You might also like https://www.alt-m.org/2020/03/30/when-the-fed-tried-to-save-...


>Even regulators are regulated. There are laws that prescribe what the Fed can and can not do, and how.

I mean sure, you can say that. The US government is regulated too. But in general the government IS the regulator of the people just as the central bank IS the regulator of monetary policy.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: