As discussed in the previous post, the net impact on balance sheets of the various approaches to government spending – (A) overt monetary financing, (B) sale of bonds to the private sector and (C) interest on reserves – is the same. There is a direct and immediate increase in net financial assets held by non-government as well as a creation of income. Both effects are equal in size to the amount of government spending. But in the case of method B, these effects are achieved in an inefficient and convoluted way that obscures the origin of government money, which in reality is created through the act of government spending or lending rather than being “borrowed” from the private sector. Even so, the practice of auctioning bonds to the private sector has often been rationalized on the supposed grounds that methods A and C are more expansionary in their demand effects than method B, and so are claimed to carry a greater inflation risk. In truth, there is no significant difference between the three methods when it comes to the expansionary impact or inflation risk of government spending. If anything, method B might carry marginally more, not less, inflation risk to the extent that bonds attract higher interest payments than reserves (inducing slightly more consumption), but any such difference will be minimal. If, in comparing effects, it is assumed that the same interest rate applies under each method, then the choice of method will affect the composition of net financial assets but make no difference to the expansionary effects of the government spending.
The impact of government spending on net financial assets
In aggregate, the net financial assets held by non-government are always equal, as a matter of accounting, to the sum of currency in circulation, reserve balances and outstanding government bonds:
Net Financial Assets = Currency + Reserves + Government Bonds
The reason for this is that all other financial assets held by non-government have an offsetting financial liability within the non-government sector, and so net to zero.
In the case of overt monetary financing (method A) or interest on reserves (method C), the increase in net financial assets created by government spending registers as additional reserve balances. The change in reserves matches the amount of government spending and so also matches the amount received by spending recipients.
Under method B, the influx of reserves caused by government spending is offset by a sale of government bonds to the private sector. This action drains reserves, because the reserves are required for final settlement of the bond purchases. Consequently, the extra net financial assets held by non-government show up, in aggregate, as bonds rather than reserves. When it is banks purchasing the bonds, this simply involves a debiting of their reserve accounts. When members of the non-bank public purchase bonds, they end up holding a correspondingly smaller amount of other financial assets (e.g. bank deposits). But here, too, the reserve accounts of banks are debited, in final settlement of the transactions. In either case, the additional net financial assets created by government spending register, in the aggregate, as bonds rather than reserves.
When bonds are auctioned off to the private sector to match the government’s net expenditure, the effect is as if (i) the government first spends, then (ii) reserve accounts are credited and then (iii) the reserves are exchanged for bonds. This depiction of reality is actually accurate, but the reality is obscured by the operational procedures of method B, which are quite complicated and roundabout. The government spending that comes first appears to be the final step of a prior round of the method B process. The operations involved in the US are described in detail in a previous post. In the first step of what is presented as a round of government spending, the central bank adds reserves. That is, it creates government money ex nihilo (“out of nothing”) by crediting the reserve accounts of banks. It does this by purchasing already existing bonds from primary dealers in the first leg of what is called a ‘repo’ or ‘repurchase agreement’. This is essentially a short-term government advance to primary dealers, with government bonds required as collateral. It is worth noticing that the bonds serving as collateral in this “first step” can only exist because of past government deficits, underscoring the fact that, in a sovereign currency system, it is indeed government spending that is prior to all else. In any event, the central bank advance of reserves is carried out so that the Treasury can issue new bonds to the private sector without this action causing a reserve shortage that would send the short-term interest rate above the central bank’s policy target. Later, the central bank completes the second leg of the repo, which involves primary dealers repaying the short-term advance (resulting in a debiting of reserve accounts) and receiving back the collateral. The central bank’s purpose in this step is to drain reserves so that when the government spends, which causes an influx of reserves, the short-term interest rate does not fall below the central bank’s policy rate due to an excess of reserves.
The Federal Reserve Bank of New York website provides this explanation of its use of repurchase agreements:
Among the tools used by the Federal Reserve System to achieve its monetary policy objectives is the temporary addition or subtraction of reserve balances via repurchase and reverse repurchase agreements in the open market. These operations have a short-term, self-reversing effect on bank reserves. … The New York Fed makes payment for the securities by crediting the reserve account of the dealer’s triparty agent, a commercial bank. This act of crediting the bank’s account actually creates reserve balances. When the repo matures, the dealer returns the loan plus interest, and the Fed returns the collateral. The return of funds to the Fed extinguishes the reserves that were originally created by the repo. (emphasis added)
This makes clear that there is no actual government “borrowing” involved in any of this. The trick is in the initial ‘reserve add’ by the central bank, which far from being an act of borrowing is in fact more akin to a temporary government advance to non-government. Moreover, the bonds the central bank requires as collateral only exist because of prior government spending. The Treasury auction is then settled with the government money (reserves) that have been created out of nothing for the purpose. The reserves are created by the central bank to enable the private sector’s purchase of newly auctioned government bonds to proceed without this causing instability in the short-term interest rate. But by the time the government spends (an action that will add reserves), the second leg of the repo needs to be complete (so that the temporarily added reserves have been drained), again to maintain stability of the short-term interest rate.
Irrespective of the operational gymnastics, the aggregate effect on balance sheets of government spending is, as already noted, basically the same under bond sales to the private sector as under overt monetary financing or interest on reserves. Namely, non-government possesses additional net financial assets to the extent of the government spending, and new income of the same amount has been created. The only difference, in the case of bond sales to the private sector, is the composition of net financial assets, which are held as bonds rather than reserves.
Inflation risk is the same irrespective of the operational regime
Since the effects on net financial assets and income are the same under all three methods of coordinating government spending and interest-rate management, the impact on demand of government spending will also be the same. Assuming, as has been done, that the relevant interest rate is the same under all methods, government spending is no more nor less expansionary with one method than the others. In all cases, the demand effect is in the spending itself, including subsequent multiplier effects and the impact on the accessibility of private credit. Whether savers have the option of holding their financial wealth in the form of government bonds or must instead opt for other assets will have no bearing on the demand – and potential inflationary – effects of the government spending.
Suggestions to the contrary have usually been based on the deeply flawed “money multiplier” tale traditionally told in textbooks, in which an expansion of reserves is claimed to result in a multiplied increase in bank loans and deposits. This tale has long been rejected by Modern Monetary Theorists and Post Keynesian economists. It has also been explicitly rejected in the publications of leading central bank research departments, including those of the Bank for International Settlements, such as here p. 19, and the Bank of England here. (Some interesting statements on the topic by central bankers and economists from the Bank for International Settlements, the Federal Reserve Banks of New York and Minneapolis, and the European Central Bank can be found at this link in the section entitled “Problems with the Money Multiplier Model”.) In reality, it is the act of lending that creates deposits. Reserves, if necessary, are added after the event by the central bank. The central bank, as lender of last resort, stands ready to do so at a price of its choosing. It does so to control the short-term interest rate (if not paying interest on reserves) and to maintain financial stability. Banks take into account the price of obtaining reserves when making their lending decisions, since it will affect the profitability of lending, but they are not constrained by a lack of reserves.
A 2015 Bank of England research paper puts it this way:
In the deposit multiplier model, the creation of additional broad monetary aggregates requires a prior injection of high-powered money, because private banks can only create such aggregates by repeated re-lending of the initial injection. This view is fundamentally mistaken. First, it ignores the fact that central bank reserves cannot be lent to non-banks (and that cash is never lent directly but only withdrawn against deposits that have first been created through lending). Second, and more importantly, it does not recognise that modern central banks target interest rates, and are committed to supplying as many reserves (and cash) as banks demand at that rate, in order to safeguard financial stability. The quantity of reserves is therefore a consequence, not a cause, of lending and money creation. (emphasis added)
Since lending is not reserve constrained, the question of whether banks hold financial assets in the form of reserves or government bonds has no direct implications for private credit creation. In either case, private credit creation will be driven by the expected profitability of lending, subject to capital constraints:
In the real world, banks provide financing through money creation. That is they create deposits of new money through lending, and in doing so are mainly constrained by profitability and solvency considerations. (emphasis added)
The profitability of lending, in turn, will depend on the demand for loans from creditworthy borrowers. A healthier level of aggregate demand such that firms are more likely to hit their sales targets and employees are less likely to lose their jobs is relatively conducive to profitable lending, because banks can then be more confident that the loans will be repaid.
Notwithstanding these considerations, it might be wondered whether the impact of government spending is dampened in instances where members of the non-bank public, rather than banks, purchase bonds. On a moment’s reflection, though, it should be clear that this is not the case. Nobody in the non-bank public is forced to acquire government bonds. Those who do so are making a portfolio decision to increase the portion of their savings held in the form of bonds rather than other financial or physical assets. The decision to save rather than spend is prior to this portfolio choice. If members of the non-bank public opt not to acquire government bonds, that does not mean they will decide to spend. In the case of savers, it is simply evidence that they prefer to hold their savings in some other form. Whether savers choose to hold their financial wealth in one form or another, the implications for demand and inflation are the same. Saving is simply the act of not spending.
It is certainly true that the non-government’s capacity to spend, including through private borrowing, is bolstered by government spending. But this is true no matter what operational procedures accompany the government’s spending. The non-government’s capacity to spend is strengthened because of the extra income and net financial assets created by the spending. This enhanced capacity to spend is enjoyed by bondholders and other wealthholders alike. There is nothing to prevent a bondholder from converting to cash (by selling the bonds to a saver) in order to spend. And holding government bonds, just like ownership of other assets, enhances an agent’s prospects of securing a loan. Government bonds can be used as (high quality) collateral.
Debate over the expansionary effects of government spending under different operational regimes is a red herring. The expansionary impact – and any inflation risk – of government spending is always in the spending itself. This is so whether the government adopts overt monetary financing, sells bonds to the private sector or pays interest on reserves.