Exercising Currency Sovereignty Under Self-Imposed Constraints

A currency-issuing government is not revenue constrained. It is always able to purchase whatever is available for sale in its own currency. This simple reality is partially concealed by a variety of contrived hoops through which modern day governments require themselves to jump. There are at least two different ways in which we can see past the confusion. The easiest way is to stand back and look at the big picture, both from the standpoint of logic and by considering the monetary and fiscal authorities as two parts of the same entity, the consolidated government sector. For the eagle eyed, this approach may appear to overlook potentially consequential details in the way governments actually spend. In practice, the monetary authority plays one set of roles, the fiscal authority plays another, and many governments have introduced various restrictions on the way in which the two can interact. The present post begins with a bird’s eye view of government spending, and then turns to a more detailed consideration of the way in which self-imposed constraints and convoluted operational procedures complicate but do not undermine the sovereignty of a currency-issuing government. The case of the US government is taken throughout as an example, but much of the discussion is also broadly applicable to other currency-issuing governments.

A Basic Point of Logic

As a simple matter of logic, a currency-issuing government that imposes a tax denominated in its own money cannot receive tax payments until it has first issued that money. Likewise, it cannot borrow its money from the non-government without first having issued money. A currency-issuing government creates money when it spends or lends. Tax payments and purchases of government debt result in the destruction of money. This makes clear, as a matter of first principles, that government spending or lending must occur before tax payments or government borrowing can take place.

The US government, for example, requires taxes to be paid in dollars. To the extent the government spends more than it taxes, the fiscal authority (Treasury) sells new debt in the form of treasuries to the non-government. Both tax payments and purchases of treasuries are settled via Fedwire when the monetary authority (the Fed) debits the appropriate reserve accounts. Reserves are a form of government money. They are a liability of the Fed and asset of the banks. For taxes to be paid, or treasuries to be purchased, there must be sufficient reserves in the system. If there are not, the Fed will need to do a ‘reserve add’ by lending to the non-government.

The Consolidated Government Sector

Often, when we are analyzing the economy as a whole, it is convenient to ignore the distinction between the monetary and fiscal authorities and simply consider them as part of a consolidated government sector in which Congress and the President are at the top of the hierarchy and the Fed and Treasury are their agents. This is fine for many purposes, because internal interactions between different parts of government have no impact on the non-government. What matters to non-government is the combined effect of the consolidated government sector’s operations.

When viewed in this bird’s eye way, the overall effect of government spending and borrowing is easy to see. Government spending, considered in isolation, results in the creation of extra reserves. The recipients of the government spending end up with additional balances in their bank accounts. However, the overall levels of bank deposits and reserves are ultimately unaffected because some members of the non-government purchase treasuries matching the amount of the government spending and this, considered in isolation, draws down bank deposits and drains reserves. In aggregate, total deposits and reserves remain unchanged, but there is an increase in net financial assets, held in the form of treasuries. It is as if the government spends, adding reserves, and these extra reserves are exchanged for treasuries.

Distinguishing Between the Fed and Treasury

The bird’s eye view of government spending is fine for most purposes. But if we want to understand the way in which government spending actually occurs under current practices, and confirm that these operational details do not diminish currency sovereignty, it is necessary to peer inside the consolidated government sector. In particular, we need to separate out the actions of the Treasury and Fed and consider the consequences for their respective balance sheets.

Many governments have put in place various self-imposed constraints on their own operations that require them to engage in a more circuitous (and inefficient) series of actions to achieve the same ultimate effect as could be obtained simply by crediting reserve accounts and exchanging them, if desired, for public debt. Two self-imposed constraints are currently operative in the US:

  1. The Treasury is required to keep an account with the Fed and to draw upon that account when it spends;
  2. The Treasury is forbidden from borrowing directly from the Fed (that is, the Fed is not permitted to buy treasuries directly from the Treasury but must instead obtain them in the open market).

At first glance, these constraints might seem problematic. We know as a matter of logic that government spending or lending is required before taxes can be paid or treasuries can be purchased. But due to the self-imposed constraints, the Treasury needs balances in its account at the Fed before it can spend and cannot get them directly from the Fed. This makes it appear as if the government is not a currency sovereign at all. It seems as if the government must depend on the non-government to obtain the dollars it spends. It creates the illusion that the government could be forced into insolvency, unable to pay its debt, or fall victim to the bond vigilantes. Yet, we know that this is a logical impossibility for a currency-issuing government.

To resolve the apparent contradictions, we need to look more closely at the operations involved. It will help if we have a specific scenario in mind. Suppose the Treasury intends to spend but has a zero balance in its account at the Fed. Suppose, also, to make the situation as tough as possible for the government, that there are initially no funds available in the Treasury’s ‘tax and loan’ accounts, which are held at various banks and can be called in to the Treasury’s account at the Fed.

This gives us the following situation. The Treasury is only allowed to spend by drawing down its account, but the account is empty. Its tax and loan accounts are also empty. The Treasury is not permitted to borrow directly from the Fed. Evidently, the Treasury will have to auction off debt to primary dealers or other members of the non-government. And it will have to do this before it can spend.

We might wonder, then, where the reserves will come from that are required to settle the treasury auction? The answer is that they will come from the Fed. The Fed will have to do a reserve add, by lending to the non-government, specifically to primary dealers.

Not only is it possible for the Fed to take this action, but if it is targeting a positive short-term interest rate, it has no choice under normal circumstances but to take this action. To see why, we need to keep in mind the method by which the Fed sets and maintains control of the short-term interest rate. It does this by ensuring that the level of reserves is consistent with settlement needs of the banking system. An excess of reserves would cause the interest rate to fall towards zero. A shortfall of reserves would cause the interest rate to rise above the Fed’s target rate as banks competed for inadequate balances to meet settlement requirements. In practice, the Fed must stand ready to accommodate demand for reserves and also to drain away any excess.

The requirements of interest-rate management therefore dictate that before the Treasury auction can take place, the Fed has to anticipate the effect of the auction on reserve balances and take steps to ensure that any depletion of reserve accounts as a result of non-government purchases of newly issued treasuries will not cause the short-term interest rate to rise dramatically above its target level.

This is why the first step, before the Treasury issues treasuries, and before primary dealers purchase the treasuries, involves the Fed doing a reserve add, which as currency issuer it can do without limit. In short, the Fed lends to primary dealers, who are required to participate in the process. The Fed does this by purchasing already-existing treasuries from the primary dealers on condition that the primary dealers undertake to buy them back later (a so-called repurchase agreement or ‘repo’). The Fed pays for the treasuries by adding reserves to the reserve accounts of the primary dealers’ banks and directing the banks to credit the primary dealers’ accounts.

So, as logic had already dictated, government spending or lending (under current arrangements, it is lending) must occur before taxes can be paid or treasuries issued.

This also makes clear why there is no danger of a lack of demand for Treasury debt. The Fed has created additional reserves, which earn zero or little interest. Agents would prefer to receive more interest than is paid on reserves, and so form a ready market for treasuries. It is equally apparent that the non-government has little capacity to pressure the Treasury into offering an interest rate that is much higher than it wishes (for policy reasons) to pay. The choice for non-government agents is between reserves that pay little or no interest and treasuries that pay somewhat higher interest.

Notice, also, that the initial step in which the Fed purchases already-existing treasuries from primary dealers is only possible if treasuries have been issued previously by the Treasury. If they had never been issued before, primary dealers would be in no position to sell treasuries to the Fed. If the Treasury were still prohibited (pointlessly) from borrowing directly from the Fed, then the Fed would have to add reserves to the reserve accounts of the primary dealers’ banks as an advance by taking something else owned by primary dealers as collateral.

Monetary and Fiscal Operations When the US Government Spends

It is possible to trace through the sequence of operations that are involved when the US government spends. In normal times, when there is no quantitative easing and the Fed targets a positive short-term interest rate, there are six basic steps in the process. The following description is based on a post by Randall Wray and a freely downloadable academic paper by Scott Fullwiler, both of which (it goes without saying) are well worth reading. The six steps are as follows:

1. Repurchase agreement. The Fed purchases already-existing treasuries from primary dealers with primary dealers promising to buy back at a later date. The effect of this step is to add reserves in exchange for treasuries in order to facilitate the non-government’s subsequent purchase of newly issued treasuries. This step, like later ones, can be represented in T-accounts. (A more sophisticated approach is taken by Scott Fullwiler in the paper linked to above.) It is arbitrarily assumed in the T-accounts that the amount of government spending involved is 10 (million, or perhaps billion) dollars. The repurchase agreement has effects on the T-accounts of the Fed, banks and primary dealers, but the overall impact on net financial assets is zero, since the step just involves an asset swap.

US Govt Spending Operations 1

2. Treasury auction. New treasuries are offered to the non-government. For simplicity, we can assume they are purchased entirely by the primary dealers, but in general that need not be the case. For present purposes, this simplifying assumption makes no meaningful difference to the analysis. The auction will involve a depletion of reserves as primary dealers draw down bank accounts to pay for the treasuries, and will add balances to the Treasury’s account at the Fed. The action in this step affects the T-accounts of the Fed, the Treasury, banks and primary dealers. As with the first step, overall there is no impact on net financial assets.

US Govt Spending Operations 2

3. Deposits in tax and loan accounts. The Treasury adds the balances received as a result of the treasury auction to its tax and loan (T&L) accounts held at various banks. This action therefore adds both to deposits held at banks and reserve account balances. There will be effects on the T-accounts of the Fed, the Treasury and banks. Here, too, the overall effect on net financial assets is zero.

US Govt Spending Operations 3

(NB. Steps 4 and 5 can occur in reverse order without consequence.)

4. Second leg of the repurchase agreement. This reverses step 1. Primary dealers buy back the treasuries that were sold to the Fed in step 1. The T-accounts of the Fed, banks and primary dealers will be affected. Since this is just the reverse of the prior asset swap, the overall impact on net financial assets is zero.

US Govt Spending Operations 4

5.The Treasury calls in tax and loan balances ready to spend. This reverses step 3. There is a depletion of deposits held at banks and in reserve accounts, and effects on the T-accounts of the Fed, the Treasury and banks. Once again, the overall impact on net financial assets is zero.

US Govt Spending Operations 5

6. The Treasury spends. The Treasury spends by drawing down its account at the Fed, resulting in a crediting of reserve accounts and a crediting of the bank accounts of spending recipients. The T-accounts of the Fed, the Treasury, banks and spending recipients will all be affected.

US Govt Spending Operations 6 Latest

As mentioned earlier, the aggregate levels of bank deposits and reserves remain unchanged as a result of steps 1 through 6. The increase in deposits for spending recipients is offset by a depletion of deposits held by the purchasers of treasuries. Even so, the non-government as a whole is better off. This is made clear by the fact that total deposits are unchanged but holdings of treasuries have increased. The effect of the government spending, as expected, is to increase net financial assets by the amount of the spending. A summary of the entire 6-step procedure is shown below.

US Govt Spending Operations 7

The vertical sums in the bottom row of the table show that: (i) the Fed’s assets and liabilities remain unchanged as a result of the operations; (ii) the Treasury has unchanged assets but an additional liability equal to the new issue of treasuries in step 2; (iii) the consolidated government sector as a whole has created an additional liability equal to the new issue of treasuries; (iv) the assets and liabilities of the banks and primary dealers remain unchanged as a result of the operations; (v) the assets of the spending recipients have increased by the amount of the extra balances in their accounts; and (vi) the increase in net financial assets of the non-government equals the change in debt of the consolidated government sector, as it must by definition.

The horizontal sums in columns 6 and 13 of the table show that only the final step in the 6-stage procedure has any net impact on the financial positions of the consolidated government and non-government sectors. The net impact is equal to the reserve add in step 6, which is of the same magnitude as the issue of new treasuries in step 2. As we had already discerned on the basis of a bird’s eye view, it is as if the government spends and the resulting reserve add is exchanged for treasuries. However, the first part of the swap is done earlier, in step 2, enabled by the Fed’s reserve add in step 1.


7 thoughts on “Exercising Currency Sovereignty Under Self-Imposed Constraints

  1. I read some of the comments @ MNE too. My 2/- worth:

    Architects draw designs of any complexity in graphics, with minimal annotation. Imagine trying to ‘explain’ a building to someone, especially a layperson – or visualise it from a written description!

    The quality is in the spec. where conversely there is no graphics.

    As a matter of possible interest or relevance (?) design starts from the inside and moves out. One simple idea after another is added together and harmonised, to make the complex. There is an organising principle (parti) around which the design elements arrange themselves. The design is worked upon (written) like a piece of music; sound after sound, colour after colour, texture after texture, form after form, function after function are resolved – slowly over time; all represented in pictures. Its ‘meaning’ becomes clear; its ‘significance’ is kept faithfully in focus. You can walk around in it and ‘see’ where you are. Zoom in and out from the whole to detail at any scale. Nowadays you do not even have to be able to visualise 3D space, colour and texture etc.

    I have never been able to walk around so freely in an economist’s explanation! There is something about conceptual representation that is really binding.

    I do not know why economists too, do not create visual representations of their designs in their explanations??? Even engineers do it although they are rumoured to spend most of their time staring at their feet (the extroverted ones stare at your feet)! Engineers and architects always joke about each other (creativity and practicality on both sides) – with the poor builder at the nexus, responsible to everybody. But economists! Are Tables and Graphs it???

  2. Peter, you were kind and helpful to comment on my recent post at NEP, suggesting I read this one of yours (which I already had, by the way.) Maybe you can shed some more light on one aspect of this that’s still not crystal clear: in the reverse half of the repo operation, the Primary Dealers “buy back” the Treasuries, and this is described as simply the “reverse of the initial asset swap.” What’s not completely clear, however, is WHICH dollars are the Primary Dealers using to make the repurchase with–because the new fiat dollars they received in step 1 have now been passed on the the Treasury. Do they get the dollars for the repurchase by selling the new Treasuries they now have in the open market?
    Thanks for any clarification!

  3. I’ll take a stab, JD. If anyone knows better, they will hopefully step in and correct if necessary, or otherwise tighten the description. I followed the explanations of Scott Fullwiler and Randall Wray very closely and my understanding of the operational details is limited basically to what is included in their writings on the matter. Like you, I was trying to provide a somewhat accessible version, this post perhaps being at a similar level of difficulty to Wray’s explanation (which is also included in the draft of chapter 9, section 9.3, of the forthcoming textbook by Bill Mitchell and Randall Wray). The post grew out of a comment I’d written in response to somebody at billy blog.

    Having expressed these reservations, my interpretation is that, yes, the primary dealers will on-sell the new treasuries and will then be in a position to complete the second leg of the repurchase agreement.

    At the end of step 3 of the 6-step process, banks are +10 in reserves and will presumably be looking to exchange excess reserves for treasuries. They’ll be happy to lend to the primary dealers to obtain higher interest.

    Steps 4 and 5 can occur in either order and both involve a depletion of reserves. This means that prior to the second leg of the repo, banks will either be +10 in reserves (if leg 2 occurs in step 4) or 0 in reserves (if leg 2 occurs in step 5). If they are +10 in reserves, they will presumably be looking to exchange for treasuries. If they are 0 in reserves, they will not be constrained by that (as we know) and in any case will realize that the government spending is scheduled to occur in step 6, which will require the Fed to conduct the necessary reserve add.

  4. Here is what I see as the rationale underlying the 6-step procedure, stated very briefly. (This comment is not in response to anyone but might help to clarify the rationale for step 1 in particular.)

    The effect of the Fed’s action in step 1 is to create a situation of excess reserves. The purpose of this is to guarantee a ready demand for the new treasuries about to be auctioned by the Treasury in step 2. Without a prior reserve add by the Fed, there might be insufficient takers at the auction in step 2. Since the primary dealers are required to participate in step 1, there is no danger of the auction in step 2 failing. As a result, the government’s capacity to spend is not restricted by the 6-step procedure but, as always, only by the appropriate governmental budgeting process.

    After that, everything is a formality from the consolidated government’s perspective because banks and other members of the non-government prefer higher interest financial assets when they are available.

    Of course, it would be more efficient if either: (i) the Treasury was simply allowed to spend without borrowing and the Fed paid its target interest rate on reserves (this would not preclude the issuance of treasuries if the consolidated government wished to cater to a non-government desire for a risk-free saving alternative to reserves); or (ii) the Fed was allowed to lend directly to the Treasury, allowing the Treasury to spend prior to (or contemporaneously with) treasury issuance.

    In either (i) or (ii), the Treasury’s spending would result in a reserve add by the Fed. In (i) the reserves would simply mount, earning interest, except to the extent that the Treasury issued new treasuries to satisfy demand for a risk-free saving alternative to reserves. In (ii) the reserves could be swapped for treasuries, for which there is a guaranteed demand due to the reserve add associated with the Treasury’s spending. Or, alternatively, interest could be paid on reserves in case (ii) as in case (i).

    Doing away with the current inefficient arrangements that are made necessary by the government’s self-imposed constraints would not make excessive government spending any more of a risk than it is now. Treasury can only spend what has been decided upon in the governmental budgeting process. Once the spending measures have been determined, spending can proceed irrespective of whether the Treasury and Fed are required to jump through hoops 1-6 or instead were allowed to function according to (i) or (ii) above.

Comments are closed.