Money Supply: M0, M1, M2 and M3 Compared

M0, M1, M2 and M3 are one stock of money counted at four widths. M0 is what the central bank issued. M1 is what can be paid out this second. M2 adds savings held outside the banks. M3 adds deposits locked for a term. Each width admits something harder to spend than the one before, and that single rule rebuilds all four.

Money counted this way is a stockA quantity sitting in place on a stated day, as against a flow, which is a quantity passing through over a stretch of time. A bank balance is a stock; a salary is a flow.. The stock is an amount standing in a country on a given day, not an amount passing through it over a year. Everything difficult about the money supply is an argument about which holdings belong inside that count and which do not.

Two things already established carry the weight underneath. The first is liquidity. Liquidity is a position in the banking system rather than a mood in it, and stating a position means counting something. The second is the central bank, the body whose rate the whole of this subject works against. The narrowest of the four widths is precisely the part of the money stock that body creates.

There is more than one measure of money because there is more than one honest answer to how fast a holding can be turned into a payment.

Why is there more than one measure of money at all?

The difficulty shows up in a household in four lines and needs no economics at all. Start there. How much money does that household have?

Counting the notes in the purse gives a number. Adding the bank account the salary lands in makes the number bigger. Nobody in that house will open the sealed jar the wedding money goes into before the wedding. Adding the jar makes the number grow again. Adding the five-year deposit the grandfather made in the child's name, breakable but only by giving up something for the privilege, makes it grow once more.

Four counts. Four different numbers. Every one of the four is truthful, and a person who reports the smallest while thinking of the largest has misled the listener without stating a single falsehood.

A country has exactly this problem at scale, and M0, M1, M2 and M3 are its four truthful counts of the same money.

A single question asked at four settings separates the four: how many steps stand between holding this and paying with it? At the tightest setting, the answer has to be none. A note in the hand clears that test and so does a balance that can be transferred while standing in the shop. Loosening the test admits holdings that need a trip, a counter and a form. Loosening it further admits holdings the holder agreed to leave alone for a stated term, reachable early only by paying for the early exit.

The single question makes the four a ladder rather than a list, and the difference matters far more than it sounds. A reader who memorises four definitions has four unconnected things to remember and no way to check any of them. A reader who learns the ladder rebuilds all four from one rule: each width admits something harder to spend than the width before it.

Two things to fix in place before the rungs. The first is that M0 does not sit on that ladder at all. M0 is sorted by a different question: who brought the money into existence, rather than how fast it can be spent. That difference is why the step from M0 to M1 is the only step of the four where something leaves rather than joins. The second is that the four widths are cut from one common set of holdings. The four widths are not boxes nested inside one another, and reading them that way is the commonest error made about money supply.

What is M0, and why is it called the base?

M0 counts what the central bank has itself put into existence, and it counts nothing else. In the Republic of Sankhya, an invented economy carrying invented figures throughout, that comes to two items and no more.

The first item is currency with the public: Rs 3,00,000 crore of notes and coin sitting in purses, tills, cash boxes, envelopes and mattresses across Sankhya. Every one of those notes is legal tenderMoney a creditor is obliged by law to accept in settlement of a debt in that country. The obligation comes from statute, not from the willingness of the person being paid. put out by the central bank and nobody else.

The second item is bank reserves: Rs 60,000 crore that the banks of Sankhya hold in their accounts at the central bank. Same issuer, different holder, and that second half is what makes reserves strange. A bank cannot buy vegetables with a reserve balance, and neither can a household. A reserve balance is a claim on the central bank held by a bank, and it exists so that banks can carry out settlementThe final step in which a payment is discharged and the funds are irrevocably transferred, as against the earlier step in which the two sides merely agree what is owed. with each other and with the central bank.

Add the two together: Rs 3,00,000 crore plus Rs 60,000 crore. M0 for Sankhya is Rs 3,60,000 crore, and the arithmetic ends there.

M0 is the only one of the four widths the central bank creates directly, and that is the reason every other width gets measured against M0 rather than the other way round.

Here is the household version of what is peculiar about M0. A shopkeeper writes paper chits and hands them out, and everybody on that street treats the chits as good. Some of the chits are in customers' pockets. Some are sitting with the shopkeeper's own suppliers, who have not cashed them and are keeping them as a running balance with him. Counting all the chits that shopkeeper has written, wherever they are sitting, counts what he issued. That is M0. The chit count withholds the buying power sitting on that street at lunchtime, and buying power is a separate count altogether.

What drops out on the way from M0 to M1?

Now change the question. Stop asking who issued the money and start asking what could be paid out in Sankhya this second, before anybody fills in a form or waits for anything.

Two things clear that test. A note in a hand is a payment waiting to happen, so currency with the public clears it, all Rs 3,00,000 crore of it. A demand deposit is a balance the holder can move or withdraw without notice, and that is what the word demand is doing in the name. Demand deposits clear the test too, all Rs 2,40,000 crore of them.

Add those two: Rs 3,00,000 crore plus Rs 2,40,000 crore. M1 for Sankhya is Rs 5,40,000 crore.

Look at what just happened to bank reserves. Reserves were Rs 60,000 crore of M0 and they are not in M1 at all. The reserves did not shrink and they were not netted off against anything. The new test asks who can spend it, and for a reserve balance the answer is nobody, at least nobody buying anything in Sankhya. Reserves simply failed it.

Bank reserves are money to a bank and not money to anyone spending, so the step from M0 to M1 is a step where something leaves, and noticing what leaves a measure teaches far more about that measure than noticing what joins it.

The habit that breaks is a strong one. Most people meet these four widths as a sequence of additions and read them the way a shopping bill is read, each line making the total bigger. M1 at Rs 5,40,000 crore is indeed bigger than M0 at Rs 3,60,000 crore, so the habit survives the encounter without being corrected. But the Rs 1,80,000 crore of difference between the two is not an addition of Rs 1,80,000 crore. The difference is Rs 2,40,000 crore joining and Rs 60,000 crore leaving. A reader who never notices the leaving cannot answer the simplest question about the pair, whether every rupee in M0 is also in M1. It is not.

The street of chits makes it concrete again. The shopkeeper wrote chits worth a certain amount, and some of them are sitting with his own suppliers as a running balance rather than circulating on the street. For anyone asking what the street can spend at lunchtime, those particular chits are of no use at all. Those chits are real, they are his, and they are not going to buy anybody a plate of rice today.

Why is M3 not M2 plus something?

Loosen the test again. Stop asking what can be paid out this second and start allowing holdings that can be reached, but not instantly. Two quite different kinds of holding walk through that door, and the whole difficulty of this guide is that they walk through separately.

The first kind is small savings held outside the banks. In Sankhya these come to Rs 60,000 crore, sitting in savings schemes run through the post office network rather than through a bank. A household holding one of these can get its money, and it is not a matter of pressing a button. There is a counter, a passbook and a form. So this holding fails the M1 test and passes a looser one, and the width that admits it is M2. M2 for Sankhya is M1 of Rs 5,40,000 crore plus Rs 60,000 crore. The total is Rs 6,00,000 crore.

The second kind is time deposits. In Sankhya these come to Rs 12,60,000 crore, and they are far and away the largest holding in Sankhya. A time deposit is money a household or a business has agreed to leave with a bank until a stated maturityThe date on which a deposit, a loan or a bond is due to be repaid in full. Reaching it early is a separate arrangement, usually on worse terms than waiting.. A time deposit can be broken early. Breaking it costs something, and that cost is why the holder treats a time deposit as a different kind of money from the balance in the account the salary lands in. So this holding fails the M1 test as well, and the width that admits it is M3. M3 for Sankhya is M1 of Rs 5,40,000 crore plus Rs 12,60,000 crore. The total is Rs 18,00,000 crore.

Read those two builds again and notice what both of them start from. M2 is M1 plus small savings, and M3 is M1 plus time deposits, so M3 is not M2 plus anything and M2 is not inside M3 at all.

Most readers go wrong exactly here, and it is a completely reasonable thing to get wrong. The names run M0, M1, M2, M3 in a sequence, and every sequence a reader has ever met, from school grades to shoe sizes, gets bigger by containing what came before. The four widths do not. They are cut off one common set of holdings, and two of the four happen to be cut from the same starting point. The small savings of Rs 60,000 crore are inside M2 and they are nowhere inside M3. Small savings are, in the plainest possible words, not part of M3.

The household version is a jar and a deposit certificate. A household says it can lay hands on the wedding jar money if it has to, and separately that it can lay hands on the five-year deposit if it really has to. Both statements are about the same starting point, the salary account. Neither statement is about the other. Adding the jar to the deposit would count the salary account twice and produce a number the household has never had.

The four widths built from one set of components

Here is the whole build in one place. Every figure below belongs to the Republic of Sankhya and stands for one illustrative day rather than a period. The base is what the four widths disagree about, so read the base column first.

Two columns in that table are doing more work than the numbers. The base column shows that M2 and M3 share a starting point and that M1 does not start where M0 ended. The added column shows that only one entry describes something absent, the reserves line, and that is the entry most readers skim past.

What does it mean that M3 is five times M0?

Divide the widest by the narrowest. M3 of Rs 18,00,000 crore over M0 of Rs 3,60,000 crore comes to exactly 5.00 times. Turn it around and the same fact reads differently: M0 is one fifth of M3, and the other four fifths, Rs 14,40,000 crore of it, is money that exists inside the wide measure and was never issued by the central bank.

The central bank issued Rs 3,60,000 crore and the widest count of money in Sankhya is Rs 18,00,000 crore, so four fifths of the money in that economy was put there by nobody with the power to print anything.

Money arriving without being printed is the strangest fact about a money stock, and it is entirely ordinary. There is no missing money, no error and nothing hidden. Every rupee of the Rs 14,40,000 crore is real, is somebody's deposit and would be paid out if asked for. The money simply did not arrive by being printed.

How it did arrive is a mechanism with a name of its own, and it is taken up in full under how banks create money through lending.

The shape of the question matters more than its answer. The ratio of 5.00 is not a constant of nature, a target or a rule. The ratio is an arithmetic consequence of what the households and businesses of Sankhya were holding on the day of the count. Had they held more currency and fewer deposits, the same economy would have shown a smaller ratio with nothing real changed about it. The ratio is a symptom to be explained, never a fact to be quoted.

Which width to watch, and for what?

Choosing a width is not a matter of preference, and there is no best measure. The width is decided by the question, and using the wrong one is how a careful reader ends up seeing either nothing at all or everything at once.

The narrow widths move with what the central bank does. M0 is the part of the money stock that body issues, so an operation that puts reserves into the system or takes them out shows up in M0 first and shows up there plainly. When the question is about what the central bank has been doing, M0 is the place to look, and the wide measures are too slow and too noisy to answer it.

The wide widths move with what borrowing and lending do. M3 is dominated by time deposits. In Sankhya they are Rs 12,60,000 crore out of Rs 18,00,000 crore, or seven ninths of the whole width. When the question is about whether the economy is taking on more debt and holding more savings, M3 is the place to look, and M0 will barely twitch while it happens.

Watch what that does to a growth number. Take Sankhya over one illustrative stretch, with the components moving by the amounts set out here and nothing else changing: currency with the public and bank reserves each up 4.00 per cent, demand deposits up 8.50 per cent, small savings up 2.00 per cent, and time deposits up 12.00 per cent. Rebuild the four widths at the end of that stretch and M0 has grown 4.00 per cent, M1 6.00 per cent, M2 5.60 per cent and M3 10.20 per cent.

The same economy over the same stretch grew its money stock by 4.00 per cent and by 10.20 per cent at the same time, and both figures are correct, so a growth rate quoted without its width is not a fact anyone can use.

The pair in the middle kills the chain reading better than any argument. Look at it. M1 grew 6.00 per cent and M2 grew 5.60 per cent, so the wider of the two grew more slowly. Nothing is wrong. M2 is M1 with a slow-growing holding added to it, and adding something that grew 2.00 per cent to something that grew 6.00 per cent pulls the combined growth rate down. A reader who believes M2 contains M1 and then extra growth has just met a number their model cannot produce.

One more use of the ratio. At the start of that stretch, M3 stood at 5.00 times M0. The wide width grew 10.20 per cent and the narrow one grew 4.00 per cent, so at the end the ratio stands at 5.30 times. The change in the ratio is a fact about the composition of what people were holding.

What does an analyst check first in a money growth number?

Watch somebody competent handle a money supply figure and the first thing they do is unglamorous. A competent reader does not ask whether the growth is high. The first question is which width the number sits on, and if the answer is not printed beside the number, that reader treats the number as unusable until it is.

The demand sounds like pedantry until the Sankhya stretch is set beside it. The same components over the same stretch produced 4.00 per cent, 6.00 per cent, 5.60 per cent and 10.20 per cent. Somebody writes that money supply grew by about ten per cent. Somebody else writes that it grew by about four. Neither is lying, both are quoting a real figure, and the two sentences describe completely different situations. One is about deposits building up across the banking system. The other is about what the central bank issued.

The width is almost never stated, and it changes the number by more than a factor of two on the same data, so recovering the width is the first act of reading a money growth figure and not an optional refinement of it.

The second check follows from the first. Once the width is known, the next question is which component inside it did the moving. A wide measure that grew because time deposits grew is a different story from a wide measure that grew because currency in circulation grew, even though both show up as one number. The same discipline applies to any ratio: a figure that moved always has parts, and the parts are where the meaning is.

A lender uses the narrow end of this for something quite practical. A bank watching system liquidityWhether the banking system as a whole is holding more funds than it needs to settle its obligations, or fewer. System liquidity is a position in the system rather than a description of any single bank. cares about reserves, which sit in M0 and nowhere else, because reserves are what banks actually settle with. The same bank setting its deposit pricing for the year cares about time deposits, and time deposits sit in M3 and nowhere else. One institution, two questions, two widths, and confusing them would have that bank managing the wrong quantity.

A household version exists too, and it is worth having. Somebody reads that money supply is growing fast and worries about what that does to prices. Before that worry can go anywhere, two things have to be settled: which width grew, and how it compares with nominal outputThe value of everything a country produced over a period, measured at the prices actually charged rather than adjusted for price change. Nominal output is the usual denominator when a money or credit quantity is set against the size of an economy. over the same stretch. A money stock growing alongside an economy that is also growing is a different fact from a money stock outrunning it. The comparison with output runs well beyond the widths, and it cannot begin until the width is named.

The Republic of Sankhya is invented.
Educational material. Not advice on any investment, tax, budget or market position.