skip to Main Content

Where do we get the funds from to pay our taxes and buy government debt?

I have been (involuntarily) copied into a rather lengthy Twitter exchange in the last week or so where a person who says he is ‘all over MMT’ (meaning I presume, that he understands its basic principles and levels of abstraction and subtlety) has been arguing ad nauseum that Modern Monetary Theory (MMT) proponents are a laughing stock when they claim that taxes and debt-issuance do not fund the spending of a currency-issuing government. He points to the existing institutional structures in the US whereby tax receipts apparently go into a specific account at the central bank and governments are prevented from spending unless the account balance is positive. Also implicated, apparently, is the on-going sham about the ‘debt ceiling’, which according to the argument presented on Twitter is testament to the ‘fact’ that government deficits are funded by borrowings obtained from debt issuance. I received many E-mails about this issue in the last week from readers of my blog wondering what the veracity of these claims were – given they thought (in general) they sounded ‘convincing’. Were the original MMT proponents really overstating the matter and were these accounting arrangements evidence that in reality the government has to raise both tax revenue and funds from borrowing in order to deficit spend? Confusion reigns supreme it seems. Once one understands the underlying nature of the financial flows associated with government spending and taxation, it will become obvious that the argument presented above is superficial at best and fails to come to terms with the basic questions: where do the funds come from that we use to pay our taxes and buy government debt? Once we dig down to that level, the matter resolves quickly.

The allegation raised in this Twitter exchange is that when MMT proponents claim that taxes and debt-issuance do not fund the spending of a currency-issuing government they are not describing reality and thus open themselves up to obvious criticism and rejection from pundits who note the existence of accounting arrangements that tie various bank accounts (such as Treasury Tax and Loan accounts or their equivalent in other nations, etc) and debt ceilings to public spending.

Various accounting conventions follow that suggest that the government cannot spend without tax revenue or borrowing from the non-government sector.

Apparently, the debt ceiling in the US or similar contrivances elsewhere clearly means that the funds obtained from issuing debt are used to fund deficit government spending.

Further, central banks will not allow overdrafts on the Treasury accounts which means that transfers from the TT&L like accounts into general purpose accounts held at the central bank, are directly involved in funding government spending.

That is the contention.

The contenders acknowledge that these institutional arrangements are political in nature (that is, imposed by legislation or regulation), which means they can be altered if there is a political will.

This 2014 article by a senior official at the Federal Reserve Bank of New York, Kenneth Garbade – Direct Purchases of U.S. Treasury Securities by Federal Reserve Banks – documents the history of changes to the Federal Reserve Act in the US with respect to allow or withdraw authority for direct purchases of Treasury debt by the Federal Reserve bank.

I analysed this issue in this blog post – Direct central bank purchases of government debt (October 2, 2014) – just after Kenneth Garbade’s staff note was published.

So, all these accounting arrangements and restrictions are variable according to the whims of the government.

But the argument being made this week on Twitter is that while accounting arrangements stand they mean that the government has to spend tax income and funds from borrowing if it is running a deficit over its tax revenue.

When other Twitter users responded that these accounting smokescreens serve to hide the intrinsic capacity of a sovereign government, which can never revenue constrained because it is the monopoly issuer of the currency, the response was then something along the lines that if that was so then why is there all the angst, in the US context, about the debt ceiling.

The problem with this ‘obvious’ argument is that it fails to consider what is actually going on in a fiat monetary system at the level beyond the accounting arrangements.

The basic contention made that these ‘accounting’ arrangements mean that taxes and borrowing directly fund government spending and that MMT proponents should stop saying otherwise is incorrect.

It all comes down to abstraction layers and temporal causality.

Of course, if there is a rule that says that the Treasury can only spend if it has $1 in some specific account in some specific bank or another and cannot borrow another cent if some arbitrary debt ceiling is breached then as long as the government obeys the rule of law those rules will be binding.

A law to stop government spending can stop it. But that is a separate issue from understanding where the ‘funds’ come from that the government spends.

I have written about the role that these voluntary constraints play in placing political discipline on a currency-issuing government many times before.

For example, please read my blog post – On voluntary constraints that undermine public purpose (December 25, 2009) – for more discussion on this point.

First, consider what constitutes ‘government’.

In my blog post – The consolidated government – treasury and central bank (August 20th, 2010) – I make the point that beyond all the hoopla about who might own the central bank and whether it sets interest rates without political interference, etc, there can be no sense that the two macroeconomic policy arms are not part of government.

There are several links within that blog post that provide further elucidation.

I also recommend you read this paper (Levy Working Paper No. 788) by Éric Tymoigne – Modern Money Theory and Interrelations between the Treasury and the Central Bank: The Case of the United States – which shows how “monetarily sovereign governments have a very flexible policy space that is unconstrained by hard financial limits”.

It also contains several references to further material to help you understand the argument and avoid falling into the trap that the Twitter debate encountered this last week.

The notion of a consolidated government sector is a basic MMT starting point and allows us to demonstrate the essential relationship between the government and non-government sectors whereby net financial assets enter and exit the economy without complicating the analysis unduly.

A simple understanding that net financial assets can only be created and destroyed in the non-government sector through transactions with either the central bank or the treasury should disabuse you of the idea that the central bank is not part of a currency-issuing government.

Further, as I have outlined in some detail in the introductory suite of blogs – Deficit spending 101 – Part 1Deficit spending 101 – Part 2Deficit spending 101 – Part 3 – the way in which fiscal policy decisions impact on bank reserves, which directly influence the way the central bank conducts monetary policy, means that the two arms of macroeconomic policy are intrinsically linked and cannot be independent.

At this simple level – that the currency-issuing ‘government’ comprises the consolidation of the central bank and the treasury functions – makes it obvious that taxes and borrowing do not fund ‘government’ spending.

A central bank is always able to issue bank reserves and currency notes (in some nations notes and coin issuance is separated by department).

If you think about it, central banks, in addition to conducting monetary policy and engaging in asset swaps with the non-government sector, directly purchase goods and services from the non-government sector.

They buy things like office supplies, IT and other equipment including professional services and training, building maintenance services, and other items that are use to equip, train and provide for their operations and staff.

Large purchases usually require competitive tender situations. But that does not abstract from the reality – these purchases are equivalent to the treasury department buying goods and services from the non-government sector.

Where do the funds come from to facilitate these purchases? From the currency-issuing capacity of the ‘government’. The central bank would just credit bank accounts in favour of these orders and create bank reserves ‘out of thin air’.

This is a different operation from quantitative easing where no new financial assets are created in the non-government sector. QE just allows the non-government sector to alter its portfolio of financial assets – less government bonds (for example) and more bank reserves.

In both cases the wherewithal comes out of ‘thin air’ but the impacts on the net financial position of the non-government sector is different.

In the case of direct central bank purchases of goods and services, no tax revenue or debt issuance would be required. In doing so, new net financial assets would be created in the non-government sector and we would not think for one minute that the ‘spending’ was made possible by tax revenue or borrowing from the non-government sector.

When talking about the treasury department, things seem to get more complicated because there is an array of these accounting processes that give the impression to those who do not really understand the abstraction involved that tax revenue and funds raised by selling debt provide the ‘money’ which allows the treasury side of government to spend.

But, the reality is akin to the central bank purchases and the accounting arrangements are just smokescreens designed to present a false causality.

The government deficit (treasury operation) determines the cumulative stock of financial assets in the private sector. Typical decisions taken by the central banks, then determine the composition of this stock in terms of notes and coins (cash), bank reserves (clearing balances) and government bonds with one exception (foreign exchange transactions).

Any payment flows from the government sector to the non-government sector that do not finance the taxation liabilities remain in the non-government sector as cash, reserves or bonds.

Note the causality here – the government finances the taxation revenue.

A very simple example, which captures the essence of the relationship between the government and non-government sectors is an economy with a population of just two; one person being government (which issues the currency) and the other deemed to be the non-government sector (which uses the currency the other person issues).

If the government runs a balanced fiscal position (spends 100 dollars and taxes 100 dollars) then non-government accumulation of fiat currency (money) is zero in that period and the non-government budget position is also balanced.

Thus there is no non-government saving in the currency.

Also note, that there is no capacity in this economy for the non-government person to pay the government person the $100 in taxes until at least that much is spent into existence by the government person.

This is the same causality noted above: spending funds taxation revenue, rather than the other way around.

Say the government spends $120 and taxes remain at $100, then the non-government surplus is 20 dollars, which can accumulate as financial (monetary) assets and represents an increase in the non-government sector’s net worth or wealth.

The non-government saving of $20 initially takes the form of non-interest bearing money holdings.

The government may decide to issue an interest-bearing bond to encourage saving but operationally it does not have to do this to finance its deficit.

An interest-bearing bond is just a piece of paper that says the government will repay a certain amount (the face value) at some specified time (maturity date) and will pay an interest premium in addition (the yield or coupon rate).

The non-government sector exchanges cash for the bond if it wants to earn interest and has no use for the liquid funds, which may be held as deposits in private banks.

The government deficit of $20 is exactly the non-government savings of $20.

Now if government continued in this vein, accumulated non-government savings would always equal the cumulative fiscal deficits.

However, should government decide to run a surplus (say spend $80 and tax $100) then the non-government sector would owe the government a net tax payment of 20 dollars and would need to run down its prior savings, liquidate real assets, or sell interest-bearing bonds back to the government to get the needed funds.

The result is the government generally buys back some bonds it had previously sold.

Either way accumulated non-government saving (financial wealth) is reduced dollar-for-dollar when there is a government surplus.

The government surplus thus has two negative effects for the non-government sector: (a) the stock of financial assets (money or bonds) held by the non-government sector, which represents its wealth, falls; and (b) non-government disposable income also falls in line with the net taxation impost.

Some may retort that government bond purchases provide the non-government wealth-holder with cash. That is true but the fiscal surplus forces the non-government sector to liquidate its wealth to resolve its shortage of cash that arises from the tax demands exceeding current income.

The cash from the bond sales pays the government’s net tax bill. The result is exactly the same when expanding this example by allowing for non-government income generation, private firms and production, and a banking sector.

Now consider this economy with some new accounting rules. The government person might announce that it has decided that it cannot spend further funds into the economy until a ledger shows positive tax revenue.

Does that alter anything? Not really.

It just alters the timing of the government spending but not its intrinsic capacity.

The question you have to ask is this: where did the funds come from, which were paid by the non-government person to the government person to allow the government person’s ledger to show a positive tax receipt balance?

The answer is that the funds came from past government deficits.

Far from imposing limits on government spending, taxation places limits on non-government spending and creates the fiscal space in which governments can take command over the use of productive resources.

As the former Chairman of the Federal Reserve Bank of New York Beardsley Ruml noted in his January 1946 American Affairs article – Tazes for Revenue Are Obsolete – taxes do not “provide the revenue which the government needs in order to pay its bills”.

The same argument applies if another accounting rule was introduced – that government person deficits had to be matched by an equal amount of debt issuance to the non-government person.

Does that alter anything? Not really.

Where did the funds come from to purchase the debt and hold this portion of the non-government person in the form of bonds rather than cash?

The government person is just borrowing the net financial assets that were created when they previously had spent the funds into existence.

Its a wash!

It is true that the private sector might wish to spread these financial assets across different portfolios. But then the implication is that the private spending component of total demand will rise and there will be a reduced need for net public spending.

So we might have an accounting trail that looks something like this:

1. Government has a financial account with the central bank called Treasury General Account (using the US terminology). When the government spends it might account for the flow of funds using its TGA balances.

2. The Treasury department might also keep accounts with selective commercial banks called Treasury Tax and Loan Accounts (TT&Ls), which collect tax revenue and funds raised from conducting debt issuance auctions. They help even out the impact of fiscal policy operations on the level of bank reserves, which helps the central bank in its liquidity management operations.

Alternatively, the funds from taxes and borrowing could be kept within accounts at the central bank called Tax and Loans, which could be a consolidated account or separate accounts. It doesn’t really matter. In this case, accounting for taxes, for example, will immediately drain bank reserves whereas having an intermediately step – using TT&L accounts – means that the reserves are not drained upon the tax payments being made (see next point).

3. Government spending is accounted for out of the TGA at the central bank although accounting entries might see numbers transferred from the TT&L accounts into the TGA balance. The latter transfers serve to drain bank reserves – that is the tax revenue disappears from the banking system having remained within that system while the TT&L balances were positive.

4. The transfers to the TGA account thus come from funds in the TT&L accounts or directly from the Tax and Loan accounts held at the central bank. It doesn’t really matter which for our purposes here.

5. Does this mean that the tax revenue and/or borrowing fund the debt entries in the Treasury General Account? Answer: No.

6. The question that has to be asked (following our simple example above) is where do the funds come from that show up as positive balances in the TT&L accounts or in the Tax and Loan accounts held at the central bank?

The answer to that question – one level of abstraction below the obvious – gives a clue to what is going on and provides for an understanding of the one of the essential contributions made by MMT that the existing macroeconomic theories failed to elucidate.

While the accounting processes make it look as though the tax receipts or the funds from borrowing are the source of these positive balances, the underlying monetary reality is that the funds have come from prior government spending and associated reserve operations conducted by the central bank.

It cannot be any other way.

As Éric Tymoigne notes (in the paper cited above):

This must be the case because, leaving aside TT&Ls, taxes and bond offerings drain reserves, so the Federal Reserve had to provide the funds. The logical conclusion then, is that reserve injection has to come before taxes and bond offerings. More broadly, the theoretical insight that MMT draws is that government spending (by the Treasury or the central bank) must come first, i.e. it must come before taxes or bond offerings. Spending is done through monetary creation ex-nihilo in the same way a bank spends by crediting bank accounts; taxes and bond offerings lead to monetary destruction.

As the following shows, in practice, injections of reserves related to the Treasury have come in several forms: monetary creation by the Treasury, funding of the Treasury by the central bank, funding of primary dealers by the central bank, and maturation of treasuries. The injection of reserves allows banks to buy treasuries or to complete tax payments.

He also notes that when the government runs a fiscal deficit it is required by US law (and there are similar arrangements across most nations) to match the deficit with debt issuance – assuming the TGA and TT&L account balances are insufficient to match the net spending injection.

But he also notes that the “Treasury has at least four ways to bypass this budgetary procedure”. I won’t go into detail but suffice to say these bypass procedures make it clear that “if tomorrow nobody is willing to take treasuries, the Treasury, with or without the help of the Federal Reserve, has the means to bypass that problem if it chooses to use them; it becomes a political issue rather than an economic one.”

In simple terms, once you drill down below the obvious accounting relationships, the causality that is established shows that central banks, in one way or another, fund treasury operations.

What about the ‘debt ceiling’?

Debt ceilings or debt brakes are commonly imposed voluntary constraints found in many nations. In the US context, the ‘debt ceiling’ requires its Congress to approve some arbitrary limit and once reached the Congress must vote to increase the allowable public debt level.

Doesn’t the fact that politicians spit chips over the ‘debt ceiling’ in the US and similar arrangements in other nations mean that there is a financial constraint on government spending?

The rather blunt legislative constraint implied by the US debt ceiling, for example, can stop government spending if the rule states that spending has to be matched by debt issuance and the currently agreed ‘debt ceiling’ is binding.

But that is a separate thing to establishing that the funds received from the debt issuance cause the government spending.

It is like saying that a national leader has to stand on their head for 3 minutes each morning before government can spend a cent. If they cannot achieve that physical feat then no spending can occur.

All the ‘debt ceiling’ saga tells us, other than it is never really binding and just becomes a political stunt, is that under certain circumstances, law makers in nations can use parliament to stop the government from spending.

The inference that this legislative capacity tells us that the borrowed funds the deficit spending, however, does not follow.

We also know that that in the US context, the Treasury department can always issue what are called ‘zero-interest instantaneous maturity securities’ – a note or coin of any value which could be deposited with the central bank as a ‘credit’ against the government’s TGA balance.

No debt issuance would be involved. Debt ceiling thus become irrelevant even as a blunt instrument to stop public spending.

See the paper released by the US Congressional Research Service (November 2, 2015) – The Debt Limit: History and Recent Increases – for more information on the debt ceiling.


I received many E-mails about this issue in the last week from MMTers who expressed confusion.

I hope this detailed response helps those who have fallen in to the trap of conflating voluntary accounting arrangements as causality.

The causality is always that the government must spend its currency into existence to facilitate tax payments and bond purchases.

Logic 101 tells us what that means.

That is enough for today!

(c) Copyright 2018 William Mitchell. All Rights Reserved.

Spread the word ...
    This Post Has 108 Comments
    1. Hole Mole! I think I just figured this out!!!!! Is this about the Coin? Everybody here is in on this except for me??? Bernie Sanders is fully on board MMT but he’s keeping it to himself? And once he wins the White House in 2020, he becomes the bad-ass, mthfkn, Ninja, Commie-Socialist-Anarchist and whips out the $20 Trillion and deposits it at the Fed! And I’m the only one in the dark about this? OK. Forget everything I said.

    2. wallflower,

      Minting debased Polish / Saxon coins was a case of economic warfare against nominally neutral Poland (which was in personal union with Saxony under king Augustus III) and a “solution” to the problem of financing the war which at some point was nearly-disastrous to Prussia (Russians took Konigsberg and got to Berlin). I don’t think it can be qualified as a clear case of “seigniorage” which would be if king Augustus himself had minted debased coins.

      The distinction between “ex-post” and “ex-ante” meaning of a social accounting identity allows us to highlight the causality and is critically important from the macroeconomic point of view. This is the actual difference between setting up neoclassical / New-Keynesian and post-Keynesian dynamic models. The MMT perspective offers clear understanding of the processes involved.

      Personally I have no problem stating that under the current arrangements taxation and net issue of government debt has to “ex-post” fund or offset government expenditure. This is just re-stating of the so-called “government budget constraint” which is the equation I wrote in the previous comment (to be strict we may need to disaggregate Treasury and the central bank).

      I would never say that the government has to fund its expenditure “ex-ante” by levying taxes and borrowing but this is what is assumed in neoclassical /New Keynesian models.

      This may look too technical but DSGE models based on the assumption of “ex-ante” funding generally underestimate spending multipliers – in the long run (when the frictions resolve) these models assume / predict that there should be no net gain in terms of real GDP from increasing budget deficits. In contrast, Stock Flow Consistent models (such as one presented in “Monetary Economics” written by Godley and Lavoie) would show a greater-than-one spending multiplier effect even in the “long run” – at the asymptotic level as the trajectory shifted upwards – as long as spare productive capacities have not been exhausted.

      This difference in the behaviour of the models and the reason why we need to look at the causation “ex-ante” from spending to saving / taxation is caused by the shape of the consumption function of the private sector, not accounted for in models based on rational expectations / intertemporal utility maximisation and on the investment function depending mainly on the forecast of capacity utilisation and prospective profits not just past profits.

      Econometric research clearly demonstrates that spending multipliers are greater than one (about 1.5) as long as central banks do not perform monetary tightening (Blanchard, O. J., & Leigh, D. (2014). Learning about fiscal multipliers from growth forecast errors. IMF Economic Review, 62(2), 179-212.). We can then ask a question why we recently have had less than full utilisation of productive capacities (so-called “secular stagnation”). A reasonable answer has been provided by Bhaduri and Marglin in their growth model where the distributional shift from workers to capitalists leads to stunted economic growth (shifts the growth trajectory downwards, the long-term growth rate is still determined by productivity growth). The rich have higher saving propensity. The bias point of the whole US economy has shifted since the mid-1970s. There is simply no such a “degree of freedom” in models based on rational expectations and intertemporal utility maximisation.

      If you are interested I can provide further references but I think that we are getting into too much detail.

    3. Dear Francisco J Flores (at 2018/03/18 at 10:19 am)

      It has nothing to do with the ‘coin’.

      It is just about basic principles of a fiat monetary system.

      Your most recent responses exude a certain desperation (private banks creating some ‘currency’ which a government without a currency suddenly accepts; and then the Bernie Sanders nonsense).

      I think we have probably exhausted this topic. I thought by providing a more detailed account of how the system works rather than how it appears to work to the layperson, that I could help advance understanding for those who were embroiled in the Twitter debate on the topic.

      It seems that all I have done is import the unproductive (and error-ridden) Twitter exchange into my blog and the positions haven’t changed one bit. As they say, one can only ‘lead a horse to the water’!

      So for all readers, please do not post any more comments on this topic.

      best wishes

    4. Professor:
      • The Coin/Bernie Sanders comment was a joke.
      • No Desperation. It was a mind experiment: I didn’t say, “no currency”. I said “no money”. In other words the new economy has a sovereign currency, say the Mitchell Dollar, but the sovereign hasn’t spent any of it into the economy yet. To demonstrate that banks don’t need money from the sovereign in order to create and lend money out of thin air. And since they’re able to do so, it can be taxed to create reserve deposits – all without relying on the govt to push it into the non-govt sector.
      Since you don’t want to continue the discussion here, please allow those who want further clarification to the following blog:

      Thank you,

    5. Francisco …

      “Now again: with new country, new bank, no reserves, no money. Bank creates $10 million in loans and deposits. Govt feels like taxing $1 million deposit holders just for the hell of it. (It doesn’t need to of course, they just feel like it because the holders were not obsequious enough.) Credit: Treasury Account at Central bank, Debit :Bank Account at Central Bank. That’s Federal Reserve Deposit creation out of bank deposits.”

      I have no idea what you’re trying to say here, but your imaginary bank in an imaginary country that doesn’t exist has NOT created any government currency or government issued reserve balances, just as real bank are unable to create government issued currency or government issued reserve balances in the real world.

    6. We have the coexistence of taxes, government spending and government bond
      issuance .A historical analysis of their evolution does not prove causality.
      We have to look where the power lies .Who does the spending the taxing and the
      bond issues .Does the private sector use fiat currency, money by government decree.
      The debt ceiling quirk of the US government is something a government which wishes
      to raise fiscal stimulus has to deal with by raising it ,abolishing it ,minting a large
      denomination coin whatever .It does not change the nature of the currency.
      Who seeks to gain by bond issues debt ceilings all the smoke and mirrors which obscure
      monetary sovereignty ? It is the very wealthy who wish to have first pick of the worlds
      resources and the labours of their fellow humans.
      The power of the state IS a threat to the wealthy and their hangers on.
      Mainstream economics is a fig leaf for their interests.
      They wish to fool everyone else that their control of resources and labour
      is best for everyone.

    7. I see Bill has asked us to stop commenting, but I’ll crave his indulgence to make this last comment:

      Adam Sawyer says:
      Sunday, March 18, 2018 at 8:44

      Thanks for that great comment there. If you want to co-operate on your “MMT in the UK for Dummies” document, and would like to get in touch, could you possibly email me at
      MWE [at sign]
      I’ve found some MMT sympathisers not too far from where I live, and there is the germ of a plan to organise some sort of talk/lecture/teach-in in our area. A document such as the one you propose would be a superb handout, or at least could serve as input for our own handouts. Thank you.
      [Bill, if you prefer that my email address should not appear in your blog, I hereby give you permission to pass my email address on to Adam Sawyer, if you would be so kind. This is part of “spreading the word”! :-)

    8. Dear Mr. Mitchell,
      You are saying that digital reserves facilitate the paying of tax and bond purchases. Are you saying that the government is financing itself? That the taxes and bond values that are lowered in our personal accounts count for nothing? That they do not add into the TGA liability account balance and are decremented when there is a FED payout with reserves? If so, as you say, then the personal funds do go quiet while they are in the TGA account (they have no reserve effects) but they are reborn when the FED spends into some other place in the economy where the funds facilitate the creation of personal account balances (inside money). Or are you saying that taxes and bonds don’t fund the government – that our personal funds are useless and the government is self-financing as clearly as if it created a trillion-dollar coin?

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Back To Top