I know you’re debating with a lunatic (Murphy, not me) but I really wish you weren’t *also* wrong in some annoyingly subtle ways.
In particular: you conflate two distinct steps in the process of 'taking' a deposit or 'making' a loan in your various writings on this topic. The result is that you confusingly use 'deposit' to mean both the cash/reserves that a customer hands over to a bank AND the liability that the bank subsequently records to that customer. But these are clearly two completely different things, not least because the former is an asset of the bank and the latter is a liability.
Similarly, you seem to use the word 'funding' to mean two different things too! Sometimes you use it to mean 'the actual cash/reserves needed to satisfy outflows' and sometimes you mean ‘the liability that matches any newly created asset on the balance sheet’. Again, one is an asset and the other is a liability. They can’t possibly be both at the same time!
Can I humbly suggest the following:
* When talking about cash or reserves that are in the possession of the bank you say just that: cash, a balance sheet asset.
* When talking about something the bank *owes* a customer, then call this a deposit. A _liability_ of the bank. But don’t use ‘deposit’ in other contexts.
They’re linked, of course: when a customer ‘makes a deposit of cash’, the asset side of the bank’s balance sheet increases (there’s more cash in the vault or reserves at the BoE), and the liability side also increases (a matching liability, the ‘deposit’, is recorded as an increased balance in the customer’s account). But they’re two different concepts.
This is important because it lets us tease apart the constituent parts of your go-to example of a bank making a loan to a customer, to see exactly where Murphy is wrong. We can break it down into 1) the creation of the loan, followed by 2) the (likely) withdrawal of the loan amount.
In part 1, the balance sheet is updated as follows:
* A new asset is created (the loan)
* A new liability is created (the deposit)
That deposit gives the customer the right to demand cash (or a transfer to another bank settled in central bank reserves) on demand. And so the bank needs to be ready to pony up the cash or reserves, but that hasn’t happened at the point the loan is made. It’s just something that could happen later.
This disentanglement is helpful because it means we can then distinguish between three different types of constraint a bank faces: leverage, liquidity and profitability which, collectively, lead to the empirical phenomenon that ‘deposits usually roughly equal loans’. But the key point is: this observation is not an iron law of economics or mathematics. It just works out like that in practice.
These are the constraints I think are in play:
1. The primary constraint on how many loans a bank can make is the degree of leverage it is allowed to have (or considers prudent). Specifically: how big can the asset side of the balance sheet be relative to the bank's equity? The answer to this is informed by thought-experiments such as ‘do we go bust if x% of our customers default and we have to write down the value of our loan book by y%?’ This is the main reason Murphy is so obviously wrong. There is a practical, and regulatory, limit on the prudent ratio of loans (or assets more generally) to equity. (Note: equity, not cash).
2. As a separate question, we can then ask: how much cash or cash-like stuff should a bank have on hand at any given time to satisfy anticipated cash outflows? This is a liquidity question. The bank needs to have a model for what proportion of its deposits will be demanded as cash on any given day, along with assessment of likely inflows and other non-deposit outflows.
3. And then there’s the question of profitability. Each asset on the bank’s balance sheet (loans, cash, whatever) is there because somebody agreed to let them create/buy that asset. What did the bank give them in return? A promise to pay interest? A promise to pay some fixed amount in the future? In any case, what is that corresponding promise worth? The answer to that is the liability that we say ‘funded’ the asset. And a key insight (which you make in your post) is that successful banks pay less for funding than is paid to them on the assets created/paid for by that funding. And deposits can be a cheap form of funding as no/little interest is paid. But they’re just one form of funding. In other words, deposit funding isn’t special because it brings cash in to the bank - there are lots of ways of bringing cash in to a bank. Deposit funding is special because the liability it creates is usually pretty *cheap*.
Now, it so happens that the amount of cash a bank tends to have on hand is roughly equivalent to the amount of deposits it holds. eg check out the balance sheets in the annual reports for pure-ish-play banks (eg nationwide or even Lloyds). But this is mostly an artefact of the interplay of the above… and not because of some iron law of nature that ‘deposits are needed to make loans’. Because that statement just ain’t true.
So now we can explain why banks can’t create deposits ad infinitum (and hence why Murphy is of course wrong).
* It’s imprudent to lever the bank too much. This is point 1 above.
* You do need to have some cash on hand in some proportion to the amount of deposits. And cash, like any other balance sheet asset, needs to be funded. And one of the cheapest forms of funding is deposits! (I know this seems a bit circular.) The point is: banks *could* fund the cash on their balance sheet some other way (eg a lot of repo perhaps). It’s just that they’ll likely make less money. This is the interplay of points 2 and 3.
Net-net: Murphy is wrong because (point 1) bank management teams are usually not insane and so don’t lever themselves up beyond a certain point (and prudential regulation exists in any case) and (points 2 and 3) retail deposits are cheap so basic profit-seeking encourages banks to attract as many as possible.
As I’ve said before, I applaud your efforts to fully expose the madness of Mr Murphy. But I can’t help thinking you’ll be more successful in this narrow area if you could just be a bit more precise with your language!!!
"and not because of some iron law of nature that ‘deposits are needed to make loans’. Because that statement just ain’t true."
It's a legal requirement. As far back as Adam Smith we get commentary on bankers issuing more paper than they've deposits to cover (Smith himself thought a little bit of it might be a good enough thing in fact). But it's simply the law, now, that a bank must balance the books once a day - usually at 4.30pm. Which is why I say that it must. It's not, at all, something that is inherent in the nature of banking. It's a law.
"we get commentary on bankers issuing more paper than they've deposits to cover" << this is a perfect example of the language point I'm trying to make :)
"issuing paper" << this is the language of liabilities. As in banks issuing IOUs. OK, fair enough. That's something they do. Indeed, it's a perfectly reasonable form of words to describe the act of issuing paper notes of the form banks used to do. No problem with that part of your statement.
"they've deposits to cover" << hmmm.. deposits are liabilities. I'm now lost. Rewriting, you're saying Adam Smith was writing about banks issuing more liabilities than they had liabilities to cover. Are you sure that's what he meant?
Are you sure he wasn't talking about banks whose liabilities exceeded their (liquid) *assets*? eg banks issuing paper (a liability) that gave customers the right to demand payment (in gold presumably) whilst knowing they didn't have remotely that much gold (an asset) in the vault? And/or no realistic prospect of laying their hands on any pronto?
Indeed, this is basically my entire point... deposits can't simultaneously be liabilities and deposits. Pick one :)
Eh? A deposit is a liability to the bank. It owes money to someone else as a result of the deposit. The loan out is an asset of the bank - someone owes money to the bank.
"The more traditional view is that banks do need deposits. Because while they may invent money when they lend they must finance that by the close of play that day."
Let's decompose it.
Sentence one: talks about "deposits". Fine - context is set and is clear.
First part of sentence two: "invent money when they lend". OK - that's one way of describing a deposit... as in a balance sheet liability that has not yet been demanded by the customer. Fair enough.
Second part of sentence two: "they must finance that by the close of play". Wait. What?! Why would anybody want to finance a *liability*? What are you talking about?
What you mean, I presume, is that they need to lay their hands on physical *cash* - or reserves at the BoE - in order to fulfil customers' requests to *withdraw* said deposits.
That isn't 'financing'. It's about liquidity... it's about getting their hands on said cash. And, yes... if they don't have enough on hand they need to go looking for some (eg in the repo market).
To repeat: I'm agreeing with you that Murphy is wrong on everything. My point is it sometimes possible for *two* people to be wrong about something, and in different ways!
"First part of sentence two: "invent money when they lend". OK - that's one way of describing a deposit.."
No, that is describing the loan, the asset. Here Mr housebuyer, here's the money I just invented, you can go spend it. Which is the thing the bank then needs to finance, by 4,30, with a deposit.
No!!!!! This is the point of my first comment... you're conflating too many things.
First, when the loan is made there are TWO things happening at precisely the same time.
1) The bank records that it can expect a some nice cash flows in the coming years from the customer that just signed the loan agreement, and is hence obliged to pay the principal, with interest. This is an asset.
2) The bank records that it owes the customer the face-value of the loan, payable as cash, or as a transfer to another bank (settled in central bank money) on demand. This is a liability.
And then, at some point in the future, quite possibly the same day, the customer might say: "thanks for the loan. I now demand that you give me the money".
Fine. If so, the bank knows that, by 4:30 if you like, that it will be making two more (linked) updates to its systems:
* It will reduce the customer's deposit balance. So the liability to the customer will go down.
* And it will hand over some cash to the customer or, equivalently, send some BoE reserves to a bank of the customer's choice. The banks's holdings of cash/reserves (assets) will go down by the same amount.
These two events are what 'withdrawing a deposit' means in terms of accounting records.
And your point, I think, is that - by 4:30 if that's the time you care about - the bank better have some cash or reserves on hand in order to make this happen.
Maybe it already has some. Maybe it needs to repo out some bonds. Whatever. *Shrug*.
We're back to the points I was making in my first comment.
Can you point me to that law please? And what, precisely, do you mean by 'balance the books'? Books are balanced at all times. That's the point of double-entry book-keeping. There is nothing to be 'balanced' at the end of the day.
Sounds like you're using 'balance the books' as shorthand for some other concept.
I suspect you mean 'have sufficient cash or reserves on hand by a specific point in time in order to pay out that amount in settlement of obligations becoming due that day' or somesuch. Which is precisely point 2 (liquidity) above. And, assuming so, I would remind you of the existence of the repo market.
What is the point of this repo market of which you speak Richard? If as you say 'books are always balanced', why would any bank even begin to fool around, speculate or try to make a gain in such a market?
Repo is short for 'repurchase'. If a bank needs to lay hands on cash quickly - e.g. to satisfy same-day outflows - then one option it has is simply to *borrow* it from a counterpart who is looking to make a bit of return overnight on excess cash they happen to have hanging around.
And one way to do this is in the repo market. Repo is short for 'repurchase', because the way borrowing in that market is structured is that the bank sells some high quality assets (eg government bonds) in exchange for cash, at the same time as entering into an agreement to _repurchase_ said bonds some time later - eg the next day. In this way a bank that needs cash quickly, and short term, can get hold of it.
This is why I find it so hard to understand why Tim persists on painting a weird picture of bankers running up and down Britain's high streets every afternoon, desperately trying to persuade random people to deposit money with them in order to make the 'books balance' at 4:30. His description of how things work is ludicrous on its face.
Now I guess he might say "no...! That's not what I meant! I was just saying that a bank has to have sufficient cash on hand by the end of each day to satisfy that day's outflows. And I'm using 'attract a deposit' as shorthand for 'borrow it'"
Why oh why Richard would "bank needs to lay hands on cash quickly - e.g. to satisfy same-day outflows"?
I mean if there was a regulation that required that I could understand it. But you are not saying that from what I can establish. So what is the reasoning behind your speculation that they might need to?
It's possible we're talking at cross purposes? Money moves in and out of banks _all day long_.
To take just one example, imagine you'll soon completing on a house purchase and need to transfer a GBP50k deposit, say, to your solicitor, who uses a different bank. When you instruct that transfer it will go over CHAPS, which settles in real-time. This means your bank will need to have at least GBP50k in their Bank of England reserve account - right then (not at "4:30" btw) - to satisfy the request. Similarly, if you try to draw out some cash over the counter, you expect the bank to have some cash on hand there and then. So the bank clearly needs to have some cash (and reserves) on hand at all times.
But the bank also knows that money will be coming *in*. For each house-buyer sending cash to a solicitor at another bank there might be a house-buyer at some other bank sending cash to a solicitor at *this* bank. So history and experience will tell the bank what the likely net inflows and outflows will be as the day progresses. They even have pretty little graphs of this that they update and monitor extremely closely all day long.
And then there are the bulk/batch payments (eg Faster Payments settles three-times-per-day settlements whereas BACS, Visa and various other payment schemes have daily/multi-day cycles), but the principle is the same: cash (well, reserves as all of this is done electronically of course) will be moving in and out of the bank throughout the day, with particular spikes at certain times (end of day being one of them).
A core function at the bank is to manage this process.. to ensure there are enough cash and reserves on hand at any given moment to satisfy outflows, but without having so much on hand that the bank can't make any money. After all, one business model would be simply to keep ALL the bank's assets as cash, but it would be hard to earn enough to pay the staff let alone interest on its liabilities. There clearly has to be some mix, and optimising this is a key competence of a bank.
All of this is obvious from a moment's introspection.
But can you see that the above process is all about the _asset_ side of the balance sheet...? It's about ensuring the right balance between cash/cash-like assets (earning low/no return but available to satisfy outflows) and less liquid assets, which the bank needs to earn a decent return. How these are *funded* (deposits, long-term debt, equity, etc) really has no relevance to the question of ensuring the bank can perform these day-to-day cash management functions.
Yes - how they're funded impacts profitability (deposits are cheaper) and is an input to the above calculations (deposits are more likely to lead to withdrawal requests than 90-day notice accounts, etc), but when it comes to ensuring a bank can satisfy a given outflow on a given day it's all about the composition of the asset side of the balance sheet.
Aside: this is where the repo point I made yesterday comes in: if the bank knows it will need more cash in a few hours than it presently thinks it will have on hand then yes... I guess it could adopt Tim's idea and walk up and down the high street asking passers-by to make a deposit. "Roll up! Roll up! I need to 'balance my books!!'" Or it could temporarily exchange some bonds for cash in the repo market. My money says that the thing they actually do is the better description of reality. But what do I know?
I've actually done this job in a bank. OK, I was a v minor part of it as an intern but......the bank proceeds along, issuing loans, taking deposits, business happens. But at some point it's necessary to check that everything balances. The time and hour of this is 4,30 each day. That's just what banks do. They do not "balance the books" minute by minute. But once a day.
This is why the overnight market used to exist. And that's when it all happened too.
Check *what* balances? What is an example of something that might not 'balance'? Put another way, what is something that might be discovered just before 4:30pm that demands a response, and what would that response be?
I know you’re debating with a lunatic (Murphy, not me) but I really wish you weren’t *also* wrong in some annoyingly subtle ways.
In particular: you conflate two distinct steps in the process of 'taking' a deposit or 'making' a loan in your various writings on this topic. The result is that you confusingly use 'deposit' to mean both the cash/reserves that a customer hands over to a bank AND the liability that the bank subsequently records to that customer. But these are clearly two completely different things, not least because the former is an asset of the bank and the latter is a liability.
Similarly, you seem to use the word 'funding' to mean two different things too! Sometimes you use it to mean 'the actual cash/reserves needed to satisfy outflows' and sometimes you mean ‘the liability that matches any newly created asset on the balance sheet’. Again, one is an asset and the other is a liability. They can’t possibly be both at the same time!
Can I humbly suggest the following:
* When talking about cash or reserves that are in the possession of the bank you say just that: cash, a balance sheet asset.
* When talking about something the bank *owes* a customer, then call this a deposit. A _liability_ of the bank. But don’t use ‘deposit’ in other contexts.
They’re linked, of course: when a customer ‘makes a deposit of cash’, the asset side of the bank’s balance sheet increases (there’s more cash in the vault or reserves at the BoE), and the liability side also increases (a matching liability, the ‘deposit’, is recorded as an increased balance in the customer’s account). But they’re two different concepts.
This is important because it lets us tease apart the constituent parts of your go-to example of a bank making a loan to a customer, to see exactly where Murphy is wrong. We can break it down into 1) the creation of the loan, followed by 2) the (likely) withdrawal of the loan amount.
In part 1, the balance sheet is updated as follows:
* A new asset is created (the loan)
* A new liability is created (the deposit)
That deposit gives the customer the right to demand cash (or a transfer to another bank settled in central bank reserves) on demand. And so the bank needs to be ready to pony up the cash or reserves, but that hasn’t happened at the point the loan is made. It’s just something that could happen later.
This disentanglement is helpful because it means we can then distinguish between three different types of constraint a bank faces: leverage, liquidity and profitability which, collectively, lead to the empirical phenomenon that ‘deposits usually roughly equal loans’. But the key point is: this observation is not an iron law of economics or mathematics. It just works out like that in practice.
These are the constraints I think are in play:
1. The primary constraint on how many loans a bank can make is the degree of leverage it is allowed to have (or considers prudent). Specifically: how big can the asset side of the balance sheet be relative to the bank's equity? The answer to this is informed by thought-experiments such as ‘do we go bust if x% of our customers default and we have to write down the value of our loan book by y%?’ This is the main reason Murphy is so obviously wrong. There is a practical, and regulatory, limit on the prudent ratio of loans (or assets more generally) to equity. (Note: equity, not cash).
2. As a separate question, we can then ask: how much cash or cash-like stuff should a bank have on hand at any given time to satisfy anticipated cash outflows? This is a liquidity question. The bank needs to have a model for what proportion of its deposits will be demanded as cash on any given day, along with assessment of likely inflows and other non-deposit outflows.
3. And then there’s the question of profitability. Each asset on the bank’s balance sheet (loans, cash, whatever) is there because somebody agreed to let them create/buy that asset. What did the bank give them in return? A promise to pay interest? A promise to pay some fixed amount in the future? In any case, what is that corresponding promise worth? The answer to that is the liability that we say ‘funded’ the asset. And a key insight (which you make in your post) is that successful banks pay less for funding than is paid to them on the assets created/paid for by that funding. And deposits can be a cheap form of funding as no/little interest is paid. But they’re just one form of funding. In other words, deposit funding isn’t special because it brings cash in to the bank - there are lots of ways of bringing cash in to a bank. Deposit funding is special because the liability it creates is usually pretty *cheap*.
Now, it so happens that the amount of cash a bank tends to have on hand is roughly equivalent to the amount of deposits it holds. eg check out the balance sheets in the annual reports for pure-ish-play banks (eg nationwide or even Lloyds). But this is mostly an artefact of the interplay of the above… and not because of some iron law of nature that ‘deposits are needed to make loans’. Because that statement just ain’t true.
So now we can explain why banks can’t create deposits ad infinitum (and hence why Murphy is of course wrong).
* It’s imprudent to lever the bank too much. This is point 1 above.
* You do need to have some cash on hand in some proportion to the amount of deposits. And cash, like any other balance sheet asset, needs to be funded. And one of the cheapest forms of funding is deposits! (I know this seems a bit circular.) The point is: banks *could* fund the cash on their balance sheet some other way (eg a lot of repo perhaps). It’s just that they’ll likely make less money. This is the interplay of points 2 and 3.
Net-net: Murphy is wrong because (point 1) bank management teams are usually not insane and so don’t lever themselves up beyond a certain point (and prudential regulation exists in any case) and (points 2 and 3) retail deposits are cheap so basic profit-seeking encourages banks to attract as many as possible.
As I’ve said before, I applaud your efforts to fully expose the madness of Mr Murphy. But I can’t help thinking you’ll be more successful in this narrow area if you could just be a bit more precise with your language!!!
"and not because of some iron law of nature that ‘deposits are needed to make loans’. Because that statement just ain’t true."
It's a legal requirement. As far back as Adam Smith we get commentary on bankers issuing more paper than they've deposits to cover (Smith himself thought a little bit of it might be a good enough thing in fact). But it's simply the law, now, that a bank must balance the books once a day - usually at 4.30pm. Which is why I say that it must. It's not, at all, something that is inherent in the nature of banking. It's a law.
"we get commentary on bankers issuing more paper than they've deposits to cover" << this is a perfect example of the language point I'm trying to make :)
"issuing paper" << this is the language of liabilities. As in banks issuing IOUs. OK, fair enough. That's something they do. Indeed, it's a perfectly reasonable form of words to describe the act of issuing paper notes of the form banks used to do. No problem with that part of your statement.
"they've deposits to cover" << hmmm.. deposits are liabilities. I'm now lost. Rewriting, you're saying Adam Smith was writing about banks issuing more liabilities than they had liabilities to cover. Are you sure that's what he meant?
Are you sure he wasn't talking about banks whose liabilities exceeded their (liquid) *assets*? eg banks issuing paper (a liability) that gave customers the right to demand payment (in gold presumably) whilst knowing they didn't have remotely that much gold (an asset) in the vault? And/or no realistic prospect of laying their hands on any pronto?
Indeed, this is basically my entire point... deposits can't simultaneously be liabilities and deposits. Pick one :)
Eh? A deposit is a liability to the bank. It owes money to someone else as a result of the deposit. The loan out is an asset of the bank - someone owes money to the bank.
Exactly!!!
OK - so I just did a quick google search for articles where you talk about 'deposits'.. and this one came up: https://www.timworstall.com/2022/10/this-years-nobel-how-absolutely-gorgeous/. And it's a pretty good example.
Key line:
"The more traditional view is that banks do need deposits. Because while they may invent money when they lend they must finance that by the close of play that day."
Let's decompose it.
Sentence one: talks about "deposits". Fine - context is set and is clear.
First part of sentence two: "invent money when they lend". OK - that's one way of describing a deposit... as in a balance sheet liability that has not yet been demanded by the customer. Fair enough.
Second part of sentence two: "they must finance that by the close of play". Wait. What?! Why would anybody want to finance a *liability*? What are you talking about?
What you mean, I presume, is that they need to lay their hands on physical *cash* - or reserves at the BoE - in order to fulfil customers' requests to *withdraw* said deposits.
That isn't 'financing'. It's about liquidity... it's about getting their hands on said cash. And, yes... if they don't have enough on hand they need to go looking for some (eg in the repo market).
To repeat: I'm agreeing with you that Murphy is wrong on everything. My point is it sometimes possible for *two* people to be wrong about something, and in different ways!
"First part of sentence two: "invent money when they lend". OK - that's one way of describing a deposit.."
No, that is describing the loan, the asset. Here Mr housebuyer, here's the money I just invented, you can go spend it. Which is the thing the bank then needs to finance, by 4,30, with a deposit.
No!!!!! This is the point of my first comment... you're conflating too many things.
First, when the loan is made there are TWO things happening at precisely the same time.
1) The bank records that it can expect a some nice cash flows in the coming years from the customer that just signed the loan agreement, and is hence obliged to pay the principal, with interest. This is an asset.
2) The bank records that it owes the customer the face-value of the loan, payable as cash, or as a transfer to another bank (settled in central bank money) on demand. This is a liability.
And then, at some point in the future, quite possibly the same day, the customer might say: "thanks for the loan. I now demand that you give me the money".
Fine. If so, the bank knows that, by 4:30 if you like, that it will be making two more (linked) updates to its systems:
* It will reduce the customer's deposit balance. So the liability to the customer will go down.
* And it will hand over some cash to the customer or, equivalently, send some BoE reserves to a bank of the customer's choice. The banks's holdings of cash/reserves (assets) will go down by the same amount.
These two events are what 'withdrawing a deposit' means in terms of accounting records.
And your point, I think, is that - by 4:30 if that's the time you care about - the bank better have some cash or reserves on hand in order to make this happen.
Maybe it already has some. Maybe it needs to repo out some bonds. Whatever. *Shrug*.
We're back to the points I was making in my first comment.
Can you point me to that law please? And what, precisely, do you mean by 'balance the books'? Books are balanced at all times. That's the point of double-entry book-keeping. There is nothing to be 'balanced' at the end of the day.
Sounds like you're using 'balance the books' as shorthand for some other concept.
I suspect you mean 'have sufficient cash or reserves on hand by a specific point in time in order to pay out that amount in settlement of obligations becoming due that day' or somesuch. Which is precisely point 2 (liquidity) above. And, assuming so, I would remind you of the existence of the repo market.
What is the point of this repo market of which you speak Richard? If as you say 'books are always balanced', why would any bank even begin to fool around, speculate or try to make a gain in such a market?
Repo is short for 'repurchase'. If a bank needs to lay hands on cash quickly - e.g. to satisfy same-day outflows - then one option it has is simply to *borrow* it from a counterpart who is looking to make a bit of return overnight on excess cash they happen to have hanging around.
And one way to do this is in the repo market. Repo is short for 'repurchase', because the way borrowing in that market is structured is that the bank sells some high quality assets (eg government bonds) in exchange for cash, at the same time as entering into an agreement to _repurchase_ said bonds some time later - eg the next day. In this way a bank that needs cash quickly, and short term, can get hold of it.
This is why I find it so hard to understand why Tim persists on painting a weird picture of bankers running up and down Britain's high streets every afternoon, desperately trying to persuade random people to deposit money with them in order to make the 'books balance' at 4:30. His description of how things work is ludicrous on its face.
Now I guess he might say "no...! That's not what I meant! I was just saying that a bank has to have sufficient cash on hand by the end of each day to satisfy that day's outflows. And I'm using 'attract a deposit' as shorthand for 'borrow it'"
OK. But, in which case, why not say so?
Why oh why Richard would "bank needs to lay hands on cash quickly - e.g. to satisfy same-day outflows"?
I mean if there was a regulation that required that I could understand it. But you are not saying that from what I can establish. So what is the reasoning behind your speculation that they might need to?
It's possible we're talking at cross purposes? Money moves in and out of banks _all day long_.
To take just one example, imagine you'll soon completing on a house purchase and need to transfer a GBP50k deposit, say, to your solicitor, who uses a different bank. When you instruct that transfer it will go over CHAPS, which settles in real-time. This means your bank will need to have at least GBP50k in their Bank of England reserve account - right then (not at "4:30" btw) - to satisfy the request. Similarly, if you try to draw out some cash over the counter, you expect the bank to have some cash on hand there and then. So the bank clearly needs to have some cash (and reserves) on hand at all times.
But the bank also knows that money will be coming *in*. For each house-buyer sending cash to a solicitor at another bank there might be a house-buyer at some other bank sending cash to a solicitor at *this* bank. So history and experience will tell the bank what the likely net inflows and outflows will be as the day progresses. They even have pretty little graphs of this that they update and monitor extremely closely all day long.
And then there are the bulk/batch payments (eg Faster Payments settles three-times-per-day settlements whereas BACS, Visa and various other payment schemes have daily/multi-day cycles), but the principle is the same: cash (well, reserves as all of this is done electronically of course) will be moving in and out of the bank throughout the day, with particular spikes at certain times (end of day being one of them).
A core function at the bank is to manage this process.. to ensure there are enough cash and reserves on hand at any given moment to satisfy outflows, but without having so much on hand that the bank can't make any money. After all, one business model would be simply to keep ALL the bank's assets as cash, but it would be hard to earn enough to pay the staff let alone interest on its liabilities. There clearly has to be some mix, and optimising this is a key competence of a bank.
All of this is obvious from a moment's introspection.
But can you see that the above process is all about the _asset_ side of the balance sheet...? It's about ensuring the right balance between cash/cash-like assets (earning low/no return but available to satisfy outflows) and less liquid assets, which the bank needs to earn a decent return. How these are *funded* (deposits, long-term debt, equity, etc) really has no relevance to the question of ensuring the bank can perform these day-to-day cash management functions.
Yes - how they're funded impacts profitability (deposits are cheaper) and is an input to the above calculations (deposits are more likely to lead to withdrawal requests than 90-day notice accounts, etc), but when it comes to ensuring a bank can satisfy a given outflow on a given day it's all about the composition of the asset side of the balance sheet.
Aside: this is where the repo point I made yesterday comes in: if the bank knows it will need more cash in a few hours than it presently thinks it will have on hand then yes... I guess it could adopt Tim's idea and walk up and down the high street asking passers-by to make a deposit. "Roll up! Roll up! I need to 'balance my books!!'" Or it could temporarily exchange some bonds for cash in the repo market. My money says that the thing they actually do is the better description of reality. But what do I know?
I've actually done this job in a bank. OK, I was a v minor part of it as an intern but......the bank proceeds along, issuing loans, taking deposits, business happens. But at some point it's necessary to check that everything balances. The time and hour of this is 4,30 each day. That's just what banks do. They do not "balance the books" minute by minute. But once a day.
This is why the overnight market used to exist. And that's when it all happened too.
Check *what* balances? What is an example of something that might not 'balance'? Put another way, what is something that might be discovered just before 4:30pm that demands a response, and what would that response be?
Retail branches will start doing cash reconciliations around 15:30~ish.
If tuppence has fallen behind a desk, the recon. won't balance.
Nobody goes home until it's found.