Reserves, Governement Bond Sales, and Savings

By L. Randall Wray

Last week we showed that government deficits lead to an equivalent amount of nongovernment savings. The nongovernment savings created will be held in claims on government. Normally, the nongovernment sector prefers to hold that much of that savings in government IOUs that promise interest, rather than in nonearning IOUs like cash. This week we will look at this in more detail.

Bond sales provide an interest-earning alternative to reserves. We can say that short term treasury bonds are an interest earning alternative to bank reserves (as discussed earlier reserves at the central bank often do not pay any interest; if they do pay interest, then government bonds are a higher-earning substitute). When they are sold either by the central bank (open market operations) or by the treasury (new issues market), the effect is the same: reserves are exchanged for treasuries. This is to allow the central bank to hit its overnight interest rate target, thus, whether the bond sales are by the central bank or the treasury this should be thought of as a monetary policy operation.

Reserves are nondiscretionary from the point of view of the government. (In the literature, this is called the “accommodationist” or “horizontalist” position.) If the banking system has excess reserves, the overnight interbank lending rate falls below the target (so long as that is above any support rate paid on reserves), triggering bond sales; if the banking system is short, the market rate rises above target, triggering bond purchases.  The only thing that is added here is the recognition that no distinction should be made between the central bank and the treasury on this score: the effect of bond sales/purchases is the same.

There is a surprising result, however. Since a government budget deficit leads to net credits to bank deposits and to bank reserves, it will likely generate an excess reserve position for banks. If nothing is done, banks will bid down the overnight rate. In other words, the initial impact of a budget deficit is to lower (not raise) interest rates. Bonds are then sold by the central bank and the treasury to offer an interest-earning alternative to excess reserves. This is to prevent the interest rate from falling below target. If the central bank pays a support rate on reserves (pays interest on reserve deposits held by banks), then budget deficits tend to lead banks gaining reserves to bid up prices on treasuries (as they try to substitute into higher interest bonds instead of reserves)—lowering their interest rates. This is precisely the opposite of what many believe: budget deficits push interest rates down (not up), all else equal.

Central bank accommodates demand for reserves. Also following from this perspective is the recognition that the central bank cannot “pump liquidity” into the system if that is defined as providing reserves in excess of banking system desires. The central bank cannot encourage/discourage bank lending by providing/denying reserves. Rather, it accommodates the banking system, providing the amount of reserves desired. Only the interest rate target is discretionary, not the quantity of reserves.

If the central bank “pumps” excess reserves into the banking system and leaves them there, the overnight interest rate will fall toward zero (or toward the central bank’s support rate if it pays interest on reserves). This is what happened in Japan for more than a decade after its financial crisis; and what happened in the US when the Fed adopted “quantitative easing” in the aftermath of the financial crisis that began in 2007. In the US, so long as the Fed pays a small positive interest rate on reserves (for example, 25 basis points), then the “market” (fed funds rate) will remain close to that rate if there are excess reserves.

Central banks now operate with an explicit interest rate target, although many of them allow the overnight rate to deviate within a band—and intervene when the market rate deviates from the target by more than the central bank is willing to tolerate. In other words, modern central banks operate with a price rule (target interest rate), not a quantity rule (reserves or monetary aggregates).

In the financial crisis, bank demand for excess reserves grew considerably, and the US Fed learned to accommodate that demand. While some commentators were perplexed that Fed “pumping” of “liquidity” (the creation of massive excess reserves through quantitative easing) has not encouraged bank lending, it has always been true that bank lending decisions are not restrained by (or even linked to) the quantity of reserves held.

Banks lend to credit-worthy borrowers, creating deposits and holding the IOUs of the borrowers. If banks then need (or want) reserves, they go to the overnight interbank market or the central bank’s discount window to obtain them. If the system as a whole is short, upward pressure on the overnight rate signals to the central bank that it needs to supply reserves.

Government deficits and global savings. Many analysts worry that financing of national government deficits requires a continual flow of global savings (in the case of the US, especially Chinese savings to finance the persistent US government deficit); presumably, if these prove insufficient, it is believed, government would have to “print money” to finance its deficits—which is supposed to cause inflation. Worse, at some point in the future, government will find that it cannot service all the debt it has issued so that it will be forced to default.

For the moment, let us separate the issue of foreign savings from domestic savings. The question is whether national government deficits can exceed nongovernment savings in the domestic currency (domestic plus rest of world savings). From our analysis above, we see that this is not possible. First, a government deficit by accounting identity equals the nongovernment’s surplus (or savings). Second, government spending in the domestic currency results in an equal credit to a bank account. Taxes then lead to bank account debits, so that the government deficit exactly equals net credits to bank accounts. As discussed, portfolio balance preferences then determine whether the government (central bank or treasury) will sell bonds to drain reserves. These net credits (equal to the increase of cash, reserves, and bonds) are identically equal to net accumulation of financial assets denominated in the domestic currency and held in the nongovernment sector.

We conclude: since government deficits create an equivalent amount of nongovernment savings it is

impossible for the government to face an insufficient supply of savings.

This post was first published at New Economic Perspectives.