Non-eocnomist here. I'm trying to understand how money works, and what "money supply" means.

I understand about how banks take deposits and lend them out, turning short term loans (the deposits) into long-term loans. And I understand this effectively creates money, because the bank lends the deposit out while at the same time it stays in the depositors account, so its in two places at once. And I get that M1 and M2 are the sum totals of these amounts of money, with M2 including accounts that aren't necessarily accessible on demand.

But now suppose instead of depositing money in a bank for them to lend out I directly buy a 5 year bond issued by FooCorp for £1,000. From the banking POV (as I understand it) I now have £1,000 less, so the bond issue hasn't changed either M1 or M2. Its just the same as if I'd bought £1,000 worth of widgets. But it looks to me like I now have a financial instrument which is worth £1,000. The bond may not mature for another 5 years, but I can get that £1,000 back almost immediately by selling the bond (maybe a week for the payment to clear). So from a practical point of view a bond issued by a private non-bank corporation looks very much the same as a short term savings account, even though it is counted as "not money" by the banking system and (AFAICT) economics in general. And the same argument would go for shares, albeit with more volatility.

To put it another way, by directly buying a tradeable bond I've done exactly the same as the bank would have (turn my short term deposit into a long term loan) but without the bank in the middle. But because the bank isn't involved it looks like the impact on M2 is different.

Have I got that right? If so, how does an expansion in non-money liquid wealth, like lots of bonds being sold or a rise in share prices, relate to the rest of the economy? Is it inflationary in the way that an increase in M1 would be?

Edit the thing that made me wonder about this was statistics about how some small number of men have more wealth than the bottom 50% of the world population, and similar. Clearly this wealth isn't in the form of M1 or M2 money, because that would show up as a big savings ratio. But it also mostly doesn't seem to be in personal ownership of real estate and chattels like super-yachts. So that leaves "investments": bonds, shares etc. This is clearly liquid wealth, but doesn't show up in the economic statistics.

  • $\begingroup$ Maybe a linear regression would help here. Independent variables: change in government spending, change in household spending, change in bond issuance or outstanding. However, the central bank will tend to stabilize the change in the price level so the effect on inflation from the independent variables alone will be obscured. $\endgroup$
    – H2ONaCl
    Oct 1, 2023 at 7:02
  • $\begingroup$ "But because the bank isn't involved it looks like the impact on M2 is different." That seems to be attributable to the definition of M2 and nothing more because as you said, bond issuance is like a time deposit but ironically the bond is more liquid. $\endgroup$
    – H2ONaCl
    Oct 1, 2023 at 7:12
  • $\begingroup$ @H2ONaCl regression using those independent variables would not help here because those independent variables are endogenious so linear regression that regresses those variables on inflation, is the worst thing you can do here, you need at least VAR or some other simultaneous equation approach $\endgroup$
    – 1muflon1
    Oct 1, 2023 at 9:25
  • $\begingroup$ @1muflon1 Yes, they're endogenous but isn't that a matter of degree? For example If the linear regression incorporates a time lag in the inflation measure, can we say the bonds outstanding variable has some degree of independence? $\endgroup$
    – H2ONaCl
    Oct 2, 2023 at 7:44
  • $\begingroup$ @H2ONaCl It is not so simple, the bias does not depend just on covariance but also variance, hence even if inflation causes the other variables just a bit your estimates might be severely biased depending on variances of the variables involved. Adding a lag of inflation as an independent variable might improve estimates but it creates also additional issues as once you go with ARDL you need to have correct dynamic structure or again you create additional bias. Also, adding controls never solves the reversed causality issue, it can only control for OVB. Lastly even small amounts of bias $\endgroup$
    – 1muflon1
    Oct 2, 2023 at 8:55

2 Answers 2


I think there is some confusion here. I'll try and explain what's going on as I see it.

Firstly, when you state that

banks take deposits and lend them out

you are referring to the Intermediation of Loanable Funds (ILF) model of credit creation. This is mechanistically and materially false as a model of banking.

In reality, banks simply create deposit balances when they extend credit and remove deposit balances when these loans are repaid. There is no connection or requirement for loans to be funded from previous deposits. See this from the Bank of England: BoE Working Paper No. 761

Now, let's consider the effect of corporate bond purchases on the money supply. If a private sector corporation issues savings bonds to raise cash, you are purchasing a financial instrument (presumably in return for interest payments). It will be instructive to analyse precisely how this transaction occurs.

Let's assume you and this company hold deposit accounts at different commercial banks. When you inform the company that you wish to buy their bond for £1000, your bank and their bank will communicate. Your bank will debit your deposit account, lowering its value digitally by £1000. Your bank will simultaneously instruct the Bank of England to decrease its reserve balance held with it by £1000 and credit the reserve balance of the company's commercial bank by £1000. Once the company's bank's reserve account has been credited by the Bank of England, it can proceed to credit the company's deposit account with them by £1000. The net result is that you have £1000 less in deposits but they have £1000 MORE in deposits. You will then own the issued corporate bond contract issued by the company to you.

Nowhere in the above analysis does the aggregate level of deposits in the economy increase. It is just a re-distributional effect endogenous to the private sector.

Yes, you own a liquid financial asset which was created out of thin air by the company issuing it. But it is quite rightly not considered currency and part of the money supply because it can not be used to pay taxes liable to the Government and therefore it cannot be used to purchase goods and services in the economy (You could theoretically try and offer the bond in exchange for a good (say £1000 valued car) but how would you pay the VAT due on that purchase in £ Sterling?). It's possible for these financial assets to be accepted as payment of debt (which has no tax obligation associated with it) but all this would do is cancel out some loan asset-liability balances and leave the created deposit (net financial assets remains unchanged) but this is rare.

Similarly, buying shares in a publically traded company has no effect on the net money supply. You purchase a share from someone who already owns it, or directly from the company issuing it. Your deposit value is transferred (via commercial bank reserve accounts at the Bank of England) to the seller's deposit account and you now own the share. No currency creation has taken place. You may later sell your share for a larger amount than you purchased it. But this does not create new money. YOU end up with a larger deposit balance at the end of the buying and selling transactions, but someone else ends up with a lower deposit balance (in aggegate). It's ALWAYS only distributional.

There are TWO ways for the money supply to increase. Either the demand for net saving in the private sector increases, resulting in the Government's deficit (private sector surplus) increasing, or the demand for credit in the private sector increases, resulting in commercial banks creating more deposit balances out of thin air.

Note that only the former method (Government deficit spending) can increase the NET money supply of the private sector because they are the only entity that sits outside the private sector. When banks create credit financial assets, they must simultaneously create financial liabilities. The net is always zero. But the Government only ever increases net financial assets (in £) when it spends.

Finally, let's consider the issuance of Government bonds (UK Gilts or US Treasuries for instance). In a given year, the Government spends its budget to provision itself and make payments to public sector workers, state pensions, companies in exchange for goods and services, etc. It does this by instructing the Bank of England to credit the reserve accounts of the relevant commercial banks, who then proceed to credit the deposit accounts of the final beneficiaries. Tax is then returned to the Government in the reverse process.

At the end of the year (in this simple example), there is often some £ money left over in both the deposit accounts AND the reserve accounts because less tax was returned than was initially spent. I'm ignoring commercial bank credit for now as it makes no difference to this analysis. To provide a safe store of value and target a particular interest rate, the Bank of England offers Government bonds. The excess reserves held by the commercial banks at the BoE are used to buy these new Government bonds. This DOES have the effect of deleting some of the reserve account balances held at the BoE and so reducing "High powered money" (i.e. liabilities of the BoE). However, it has no impact on the level of deposit account balances held by individuals and firms at the commercial banks. The excess £ money left over in deposit accounts after the Government spending and tax process is complete REMAINS in the economy after Government bonds are issued.

The conclusion of the above paragraph is that the money supply available for spending on goods and services in the real economy has gone up as a result. This could lead to an increase in the price level but not definitely as it depends on multiple other factors (such as production, bank credit, spending behaviour, etc).


The explanation is not entirely correct. The reason why you lending money through 5 year bond does not increase money supply is that 5 year bond is not included (using UK definition - money supply definitions are different in main economies like US, UK or EU) in M2 or M3.

If in fact you would buy debt security (e.g. bond) with maturity up to 2 years, you would create new money. Debt security would count toward M2, and the money you used to buy it are still part of money supply. So if bond with maturity up to 2 years costed you 1000 pounds you increased money supply by 1000 pounds. Bank does not need to be involved here at all.

Purchasing stocks never increases money supply because they do not count in any of the Ms.

Share prices are not directly affecting inflation (in fact there is well known opposite effect from inflation to shares e.g. see the Feldstein's 1983 paper). There can be some second order effects here. For example, there might be income/wealth effects, but richer shareholding households have also lower marginal propensity to consume so their spending does not affect inflation that much. It is an empirical question how big these or other second order effects are if there are any, but unfortunately there is no research on this to my best knowledge.

When it comes to higher borrowing (e.g. more bonds) it depends on situation. For example, economic models generally predict that central bank purchases of government bonds will result in higher inflation, because these are done with new high powered money, but the same would not hold for a private individual purchasing government bond because the increase in aggregate demand from new government spending could be offset by decrease in aggregate demand necessitated by individual using the money to purchase bond as opposed to consumption or investment spending. However, even here exact parameters matter (e.g. what is the current value spending multiplier?).

  • $\begingroup$ Yes, I understand that bond and share purchases are excluded from the M* numbers by definition. My question is about whether that makes sense. Does this make measures like M2 and the savings ratio meaningless because they ignore a big chunk of what they are supposed to be about? $\endgroup$ Oct 1, 2023 at 16:50
  • $\begingroup$ @PaulJohnson no, measures of M* measure money supply, typically going from narrow meaning of money to broad meaning of money. However, technically anything can serve as money as long as it fulfills functions of money. Salt can be money. However, in modern economy virtually nobody uses salt as money. Short term debt is actually not extremely uncommon in settling transactions to pay for some goods and services, although it’s not common (but hence it is part of the broadest money supply measures), shares are rarely used, and hence they don’t fall into these measures $\endgroup$
    – 1muflon1
    Oct 1, 2023 at 22:25
  • $\begingroup$ I don’t see how it in any way invalidates the measures of money supply $\endgroup$
    – 1muflon1
    Oct 1, 2023 at 22:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.