Complications and private preferences. There are often two objections to the claim that government spending effectively takes place by simultaneously crediting the recipient’s bank account as well as the bank’s reserves: a) it must be more complicated than this; and b) what if the private sector’s spending and portfolio preferences do not match the government’s budget outcome?
The first of these objections has been carefully dealt with in a long series of published articles and working papers (by Bell (a.k.a. Kelton), Bell and Wray, Wray, Fullwiler, and Rezende who look at actual operating procedures in the US, Canada, and Brazil; I’ll provide references later as well as more details). In practice, the treasury cannot directly credit bank accounts when it wants to spend.
Rather, a complex series of steps is required that involve the treasury, the central bank and private banks each time the treasury spends or taxes. The central bank and the treasury develop such procedures to ensure that government is able to spend, that taxpayer payments to treasury do not lead to bounced checks, and—most importantly—that undesired effects on banking system reserves do not occur. While the end result is exactly as described above (treasury spending leads to bank credits, taxes lead to debits, and budget deficits mean net credits to both demand deposits and bank reserves), it is more complicated.
This often generates another question: what if the central bank refused to cooperate with the treasury? The answer is that the central bank would miss its overnight interest rate target (and eventually would endanger the payments system because checks would start bouncing). Readers are referred to the substantial literature surrounding the coordination (more details for the wonky coming up in a later blog). Nonspecialists can be assured that the simple explanation above is sufficient: the conclusion from close analysis is that government deficits do lead to net credits to reserves, and if undesired excess reserves are created they are drained through bond sales to maintain the central bank’s target interest rate.
The operational impact of bond sales is to substitute government bonds for reserves—it is like providing banks with a savings account at the central bank (government bonds) instead of a checking account (central bank reserves). This is done to relieve downward pressure on the overnight interest rate.
With regard to the second objection we first must notice that if the government’s fiscal stance is not consistent with the desired saving of the nongovernment sector, then spending and income adjust until the fiscal outcome and the nongovernment sector’s balance are consistent. For example, if the government tried to run a deficit larger than the desired surplus of the nongovernment sector, then some combination of higher spending by the nongovernment sector (lower nongovernment saving and lower budget deficit), greater tax receipts (thus lower budget deficit and lower saving), or higher nongovernment sector income (so greater desired saving equal to the higher deficit) is produced.
Since tax revenues (and some government spending) are endogenously determined by the performance of the economy, the fiscal stance is at least partially determined endogenously; by the same token, the actual balance achieved by the nongovernment sector is endogenously determined by income and saving propensities. By accounting identity (presented above) it is not possible for the nongovernment’s balance to differ from the government’s balance (with the sign reversed—one has a deficit and the other a surplus); this also means it is impossible for the aggregate saving of the nongovernment sector to be less than (or greater than) the budget deficit.
So, those are the general responses to those objections. I will do a wonky blog later with more details. But next week we look in more detail at the private saving decision.