Recent Posts

Tuesday, March 24, 2015

Primer: Can We Consolidate The Central Government And Central Bank?

One of the strengths of Modern Monetary Theory (MMT) is that it provides a clean analytical framework for the analysis of "modern" economies (economies with a free-floating currency and which controls its central bank). One of the ways in which it does this is to consolidate the central bank with the fiscal side of the central government. Such a consolidation has extremely important effects for understanding government default risk, and is controversial as a result.

UPDATE: This text was incorporated into the eReport Understanding Government Finance (link).

It should be noted that this is a somewhat abstract issue, and it was generally not the direct topic of debates. Instead, academic debates revolved around the more concrete implications of this issue. However, since the concept is consolidation is used a lot within MMT, this topic provides a natural starting point for addressing those other debates. However, I keep the discussion here relatively short, as I hope to discuss the more substantive issues elsewhere.

What Is Consolidation?

Consolidation is a term from accounting, and it is a merger of the accounting statements of two (or more) entities. Financial analysts use consolidated accounting statements all of the time, possibly without realising it. Pretty well every major multinational "corporation" actually consists of dozens if not thousands of separate legal entities. When analysts look at the financial statements of a public corporation, what they are looking at are the consolidated statements of all of the underlying corporate entities.

What we are interested in here is the consolidation of the central bank and the rest of the central government, which is the "fiscal arm" of government. I will refer to the fiscal component of the government as "the Treasury" herein, although it may be labelled "the Ministry of Finance" in some countries (such as in my home country of Canada). For the purposes of economic analysis, we typically are only concerned with monetary policy and fiscal policy, and so we abstract away from the other components of the central government (such as the judiciary).

If we used corporate accounting principles, my feeling is that the Bank of Canada and the United States Federal Reserve would qualify for consolidation with the rest of their respective Federal Government. The central banks are wholly-owned subsidiaries*, they work within a framework dictated by the rest of the government, and the top officers are political appointees. Their "independence" is roughly the same level of autonomy that other subsidiaries have.

However, governments use a different set of accounting principles, and the Treasury and the central bank are not consolidated. Therefore, the consolidation accounting has to be applied by the analyst.

An example of consolidation is given below, starting from a much simpler starting point than the full set of national accounts. In it, the Treasury and Central Bank behave in a fashion similar to the Canadian Federal Government. The main simplification of this framework is that "bank reserves" no longer exist. In terminology that only applies to the American banking system, the "required reserve ratio is 0%". (In Canada, a deposit at the central bank is referred to as a "settlement balance".) I will discuss further how government finances operate in Canada in later articles.

First is the balance sheet of the Treasury. I assume that the government has non-specified fixed assets with a value of $200, a deposit at the central bank**, as well as 100% ownership of the Central Bank worth $5. The Treasury has $180 in bonds outstanding (which presumably includes Treasury Bills as well). This leaves the government with Capital (equity) of $55.  [Update: I added the value of the central bank equity to the balance sheet; I previously had accounted for the Treasury as a stand-alone entity which is incorrect. Thanks to Joseph Lalibert√© for spotting that problem.]

Deposits at Central Bank $30
Fixed Assets $200
Equity in Central Bank $5
Government Bonds $180
Capital $55

The central bank does not have deposits from private banks ("reserves" or "settlement balances"). Therefore, the only liabilities of the central bank are currency (dollar bills and coins) as well as a deposit from the Treasury. The Central Bank operates with only $5 in Capital. This bank is following standard modern procedure (for the "Anglo" economies - see the note below on "overdraft economies"), and only has Government Bonds (including Treasury Bills) as financial assets, and with $2 of fixed capital (a very small currency museum?).

Central Bank
Government Bonds $75
Fixed Assets $2
Currency $42
Deposit from Treasury $30
Capital $5

What happens if we consolidate the Central Bank and the Treasury? When we consolidate two entities, we net out claims between the two. The final result is:

Fixed Assets $202
Currency $42
Government Bonds $105
Capital $55

The changes include:

  • The Central Bank's holdings of Government Bonds are netted out. All we are left is the $105 in bonds that are held outside of the Central Bank.
  • The Treasury's deposit at the central bank is both a liability and an asset to the consolidated entity, and is netted out to zero.
  • Fixed Assets represent all fixed assets on both balance sheets, and is the sum of the two values. The consolidated Capital is equal to the original Capital of the Treasury, as it already included the Capital of the Central Bank. (In a real world example, consolidated balance sheet valuations may be different than is the case for the unconsolidated entities. This might result from using historical costs versus market cost to value assets and liabilities. In such a case, Capital would need to be adjusted to bring the balance sheet back into balance.) [Update: corrected.]
From the perspective of entities outside the government, all that matters are their (net) financial assets, which are Currency holdings ($42) and their Government Bond holdings ($105). These amounts are unaffected by consolidation. This is why consolidation makes for a cleaner economic model - it reduces the number of variables to be tracked, but it does not affect the position of the non-government sector.

[Update.] As Neil Wilson helpfully informed me, the United Kingdom publishes a consolidated set of accounts - the Whole of Government Accounts (link). At the time of writing, I am unaware of other developed countries that follow suit.

Aside: Overdraft Economies

The above example is one where the central bank buys central government bonds, which is standard practice in the "Anglo" economies (Canada, United States, United Kingdom, Australia). This framework is fairly standard for MMT analysis. However, not all central banks operate in this fashion. An alternative framework is for private banks to borrow directly from the central bank, possibly in the form of overdrafts (a negative deposit balance). In such an "overdraft economy", the assets of the central bank are loans to private banks.

This is discussed in Section 4.3.8 in Professor Marc Lavoie's textbook Post-Keynesian Economics: New Foundations. This distinction in operating procedures has shown up in arguments between MMT economists and those in the other wings of the post-Keynesian school. I will not discuss those arguments here, but it should be noted that consolidation of the Central Bank and the Treasury in an overdraft economy accomplishes little, as inter-government claims are greatly reduced. That said, it is unclear how much of a difference this makes in practice.

Why Does This Matter?

There is not a whole lot of agreement between various camps of economists around many topics of government finance. However, there appears to agreement upon the following point: in a regime where the currency is non-convertible, a central bank can be insolvent (have negative equity), but it cannot be illiquid (unable to meet payment obligations). (In a regime where there is a legal obligation for the central bank to convert currency into something external, such as gold, the central bank could become illiquid.) This means that central banks in such a regime should be free of default risk.

Therefore, if we consolidate the Treasury with the Central Bank, the Treasury will inherit this property. To use mathematical terminology, being able to consolidate the Central Bank with the Treasury is a necessary and sufficient condition for Treasury bonds being default risk free. I think it would be safe to argue that there is little consensus about floating currency governmental default risk, hence there is little consensus about the validity of consolidation.

The MMT Position

My interpretation of the Modern Monetary Theory position on this topic can be summarised as follows. It is not based upon particular references, rather it is my restatement of strands of thought that I believe is consistent with the existing literature.
  1. The best way to analyse a floating currency government is with the Central Bank and Treasury consolidated. (This is a positive - "value free" - statement about economic theory.***)
  2. MMT economists have delved into the details of monetary operations in the developed economies with free-floating currencies, and used this evidence to argue that consolidation is a valid analytical technique. (This is an empirical statement.)
  3. Governments should organise their monetary operations in such a way that consolidation is never called into question. (This is a normative statement.)
I agree with position (1) in my list above; consolidation is the best way of analysing an economy. The only question is whether it is theoretically justified. Since I do not believe that default risk is significant, I believe that it is justifiable. The justification of that stance is relatively complex. I would need to delve into the operational details of monetary and fiscal operations (following the lines of the research done in statement (2)). I hope to cover these topics in greater detail in later articles.

Additionally, it should be noted that the value of consolidation in analysis is aimed more at the development of models of the economy - either verbal or mathematical. The idea is that consolidation guides our thinking about how the economy behaves. If one is interested in a detailed analysis of the structure of the national accounts, consolidation may or may not be useful. From the point of view of an analyst, the fact that national accounts are typically given in unconsolidated form is an advantage, as that provides extra information. Diligent equity and credit analysts would be very happy to be able to get unconsolidated accounts for public corporations, so that they could get a much better handle on the fortunes of various subsidiaries. [Update: this paragraph was added in response to comments by Ramanan.] 

As for (3), the standard MMT proscription is for the government to stop issuing bonds; its only liability would be money. Although I do not think there would be dramatic effect (with Quantitative Easing, Japan and the United States is halfway there), the policy environment would be different (nominal risk-free interest rates would be stuck at 0% permanently). However, it would be trivial for the government to keep issuing bonds but change operating procedures and still ensure that the central government would never face the possibility of default for financial reasons. These procedural changes would have no observable effect on the economy. It appears that the only reason these changes have not been made is the result of economic superstitions. I hope to discuss this issue further in later articles. 

Postscript: Mainstream View

I have never seen consolidation come up within mainstream economic analysis. There has been analysis of sovereign default risk, analysis which could be best described as confused. But if we look at standard Dynamic Stochastic General Equilibrium models, consolidation would be justified. (I am referring to models similar to those found in Woodford's Interest and Prices.)

Within these models, fiscal policy is specified as an exogenous set of primary fiscal balances, and monetary policy consists of the central bank setting an interest rate by trading Treasury Bills (versus money). (There are typically no bonds within these models, just 1-period Treasury Bills.) Within the models, the two arms of government are thought of as distinct, but the accounting are effectively consolidated. What matters within the model are the private sector holdings of money and Treasury Bills, and inter-governmental claims are not tracked.

As for default, the word does not even appear in the index of Woodford's 785 page book. The governmental budget constraint implies that Treasury Bills are rolled over every period without incident. Within the context of such models, default is unthinkable. This makes the case for consolidation water-tight.


* There is a certain amount of silliness surrounding the ownership structure of the United States Federal Reserve floating around on the internet. Private banks do own the equivalent of preferred shares, but what matters for the purposes of corporate control are common equity. This class of equity is wholly under the control of the United States Government.

** In Canada, this is the Consolidated Revenue Fund of Canada (Wikipedia link). Canada does not deposit government money at private banks, which is a practice that the United States Treasury uses periodically to smooth cash flows within the American banking system. Avoiding complications like that is one advantage of using the Canadian system as an example.

*** One could reasonably argue that it is very difficult for economic analysis to be "value free". But in this case, I find it hard to see what ideological positions could bias the relative merits of accounting treatments within an economic model.


(c) Brian Romanchuk 2015


  1. Brian,

    About your Bank of Canada, two things:

    Earlier the government used to transfer funds between the government's account at the Bank of Canada and in commercial banks. It then figured, it could "auction" the funds.

    It's possible that this is not perfect and the government is left with some funds to transfer to private banks at the end of the day.

    Second even if all funds are auctioned, the absence of funds at commercial banks doesn't mean that the government's assets are zero.

    So the Treasury has financial assets which is either deposits at commercial banks or funds lent to financial institutions or a combination of the two. Either way, the government has financial assets other than deposits at the central bank.

    (My usage of the phrase "government" is different from yours).

    Second, I do not think the best way is to consolidate. National accountants don't do so. Check flow of funds, SNA etc. I am not sure your position is value free as the guides do offer some reasons to do so in their way.

    1. Thanks for the feedback.

      This article is an excerpt from the first draft of a longer document. There is some wider context missing. I will take a look at your comments and see how I can tighten up my wording here so this works better as a stand-alone document.

      With regards to Canada, my next article is discussing the "Canadian model" in more detail. The simplified model I describe follows from what I refer to as the "simplified Canadian model". (Since the article does not yet exist, I could not work in a link.)

      Although I should clarify here, I am more interested in "model accounting" than formal the formal national accounts definitions. For example, should we consolidate the central bank and Treasury within a stock-flow consistent model? Also, as long as nobody defaults, we can abstract away from the whole issue of clearing. Someone who is looking at "real world" data will need to apply a mental transformation from those simplified models to the national accounts data. And if they are worried about the nuts and bolts of clearing, they will not find it within this article.

      The lending from the central government fund is an interesting wrinkle, which will probably get an article on its own. As a spoiler, I believe that it creates an interesting hybrid between the "asset based" central bank model and the "overdraft economy" model.

  2. Ben Bernanke was entirely clear about the role of the Fed in relationship to the US Treasury in a lecture he gave 29th March 2012 at the George Washington University School of Business when he said:-

    "We act as the agent, the fiscal agent for the Treasury."

    See page 10 on PDF.

  3. You can find the original source of all four lectures here in video and PDF format (top of web page):-

  4. "However, it would be trivial for the government to keep issuing bonds but change operating procedures and still ensure that the central government would never face the possibility of default for financial reasons."

    It never can anyway. The central bank will not go against the elected will of the legislature. To do so would cause a constitutional crisis and of course the legislature can eliminate the central bank if it wants to with a legislation change.

    So there is *no* control point at which a central bank can say no to an authorised request from the legislature independently of another legislature request not to (hence the debt ceiling debacle in the US). In other words find the person you think can say no, and you'll find a person looking for a new job if they do.

    It's all smoke and mirrors with back fill procedures in place to present the appropriate end of day picture.

    The whole concept of an entity free of politics or influence is nonsense. It is a pure Solomon Fallacy - as we've seen with the Swiss situation. Institutions are run by people and people are influenced by beliefs and biases. If those people are unelected and have any actual power you end up with a near feudal situation.

    Note that in the UK we do have consolidated 'Whole of Government' accounts which does an FRS 10 consolidation on the public sector including the Treasury and the Bank of England.

    1. As a dour fixed income analyst, it is not in my temperament to rely solely upon political arguments. In any event, I will be extending this body of analysis to look at look at what other people are saying about default risk, and they would not grant the same political assumptions.

      Also - thanks for the heads up on the UK accounts.

    2. You are not dealing with the weather or robots, so when assessing the function of a human system you have to take into account the behaviour of people and the power relations between them.

      The currency system is a power relationship. Specifically a monopoly.

      The singular failure of economics has been to drop the 'political' prefix. Because politics is the human condition. Even more amusingly the economists are constantly playing politics with each other over their analysis.

    3. Neil says “If those people are unelected and have any actual power you end up with a near feudal situation.”

      I have disastrous news for Neil: about 99.99% of decisions taken by governments are taken by unelected bureaucrats or committees of bureaucrats (shock horror) who wield “actual power” (more shock horror).

      Almost the entire world (apart from Neil) is quite happy with democratically elected politicians delegating decisions to bureaucrats, with politicians obviously having the ultimate right to overrule bureaucrats.

      Indeed, the idea that politicians have the time or expertise to take any more than a MINUTE proportion of decisions is pure fantasy.

  5. Joseph Lalibert√©March 25, 2015 at 7:08 PM

    Great piece. You have now become my #1 economic blog destination!

    Specifically on the issue of consolidation. Two things to note... First, at consolidation, the subsidiary assets and liabilities must me evaluated at market value (granted, this would not make much difference in the case of the central bank, except for bonds maybe). Second, the government of Canada 100% ownership of the central bank would have to be recognised before consolidation as an asset on its balance sheet (if the government was to follow "normal" accounting principles, that is). Therefore, after consolidation, this asset item would disapear to be replaced with the market value of assets and liabilities of the subsidiary (any difference between market value and accounting value would be recognised as goodwill). So it is incorrect to sum up the equity of each entity when consolidating (therefore, before consolidation, capital of the GoC would be at 55$ and would remain so after consolidation). These are all minor technical elements obviously in the context of the main points of your analysis.

    Finally, some of Ramanan's points are incorrect. First, the government is always able to auction 100% of the amount it so choose. The reason is simple: auctions striking price pretty closely match the overnight rate. If participants take a pass on these auctions at 16:15, they may have to borrow at the target plus 0.25% (which would be irrationale). Second, no government deposit is left at commercial banks, those amounts are consolidated in the government deposit account at the central bank at the end of each day.

    1. "Finally, some of Ramanan's points are incorrect. First, the government is always able to auction 100% of the amount it so choose. The reason is simple: auctions striking price pretty closely match the overnight rate. If participants take a pass on these auctions at 16:15, they may have to borrow at the target plus 0.25% (which would be irrationale). Second, no government deposit is left at commercial banks, those amounts are consolidated in the government deposit account at the central bank at the end of each day."

      Yeah I did talk of the possibility of auctioning all funds.

      But the government still has financial assets which is more than central bank deposits. That was my main point. The financial asset is funds lent to the financial system.

    2. The size of those assets depends upon the size of the lending from the consolidated revenue fund. I have not come across data on that front; I have sent a query to the Bank of Canada. If it is just used for small gap filling, it may be relatively trivial. The BoC periodically has repo assets with the private sector as well. Additionally, there are Crown Corporations that have financial assets, and the government has Foreign reserves and gold.

      I will be covering the real world balance sheet in an upcoming article, although if anyone has some info on the lending from the Consolidated Revenue Fund, I would be happy to hear about it.

  6. Thanks! Yes, I should add a line item for the equity in the BoC; I was doing the rest of the government as a stand-alone entity. The valuation of bond holdings would match what is used on the liability side of the balance sheet.

  7. This comment has been removed by a blog administrator.

  8. This comment has been removed by a blog administrator.

  9. Dear Mr. Romanchuk, I found your interesting report.
    My name is Prof. Park Seung-Joon (Kwansei Gakuin Univ. Japan), major in Environmental Economics. I have a small study group "People's Economic Policy (", together with some other professors.
    We are interested in economic policy which liberal political party should adopt. In Japan, liaral (left-wing) political party and media are not aware of economics and, interestingly, they are often more austerian than conservative LDP (Premier Abe).
    Do you mind if I introduce this your report (translated into Japanese by me) on our website?

    1. That would be great. Obviously, the only confpdition is that you leave a reference to me and my site.

      When you are done, could you let me know where the web page is? I can add it to my small list of translations that I cannot read... Thanks.

    2. Dr. Mr. Romanchuk,
      sorry for the delay. It took time to refine the translation.
      Our translation is now on web.
      Please let me know if there is anything to be fixed.
      Park Seung-Joon.

    3. Dear Mr. Romanchuk,
      sorry for delay of my reply. I have tried to reply into this blog more than 5 times in vain. I hope it succeed this time.
      We have uploaded your paper (Japanese version)
      Please check it, and let me know if there is anything wrong.
      Park Seung-Joon.

    4. Hello,

      It looked interesting. Your comments needed to be moderated, and I was not able to do so for a few hours. Thanks.

  10. Dear Mr. Romanchuk,
    finally we could post the translation. I hope this is a helpful translation,
    if not perfectly beautiful one.
    Please let me know if there is anything to be fixed.
    Park Seung-Joon.

  11. The eurozone has specific institutional arrangements, linked to the fact that member states are not monetarily sovereign. Do you think that the consolidation of the treasuries and of the ECB is relevant for it?

    1. Looking over my text, I didn’t call out the euro area exception specifically, but the discussion references currency sovereigns - and the euro area doesn’t count.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.