There appears to be a wide misunderstanding on the subject of money creation and government debt. This misunderstanding is all the more revealing by the number of media articles written about austerity and the coming over supply of Treasury's debt. Interestingly, some authors can be categorized by either conspiracy theorists (often difficult to differentiate from the usual gold bugs) or libertarians (sometimes but not always indistinguishable from plain government haters).
The Himalayan task here is not so much about explaining the reality itself but more about escaping the shortcuts our rationality imposes on us. Just as with the art of investing, it is the process or path we take to understand an issue that is the key.
Private money creation in our fractional reserve system is like an almost ever growing inverted pyramid. Banks create money by extending credit. As banks receive deposits, they keep only a fraction of these in reserves and lend out the remainder. The more they extend credit the larger the money supply. The reason our brain is not screaming with outrage as we grasp this process is that in such a system, every liability (the loan held by the customer) has a corresponding asset (this same loan owed to the bank). These net to zero. Every new amount of money is an asset for one and a liability for another, which is why money is said to be loaned into existence. Our mind is bothered by the vague sense of size and lack of limitations but not repulsed by the concept. An image of an inverted pyramid comes to mind where the larger the amount of credit the taller the pyramid. At the very top you have derivatives with their gigantic notional amount and at the very bottom you have gold, just below hard cash, which is no one’s liability. This is better known as John Exter’s liquidity pyramid, central banker at the Federal Reserve in the 1940s and 1950s. Should we all decide to reimburse our debts, the pyramid would shrink until it reaches base money (vault cash, circulating currency and bank reserves).
Government money creation on the other hand is utterly counter intuitive. The process clashes with every cell of common sense we may have and is rejected as an inconceivable reality. It shakes the very foundation of the relationship between value and scarcity. So how on earth is this medium of exchange that more than ever rules our world really created? Understanding the process will bring us a long way in understanding government solvency and monetary inflation.
In our fiat monetary system, money is issued by the government. The government, through the Federal Reserve, may increase the amount of money in circulation. But there is a crucial difference between money creation by the private sector (banks) and by the public sector (government). Although debt must be the basis for the creation of money by banks the same does not hold true for the government. Let me explain.
The Federal Reserve is an independent entity within the government having both public purposes and private aspects. As the Federal Reserve is an entity within the government neither can be indebted to the other. This means that whilst all money creation by banks always follows the rules of double entry book keeping, with every debit having an equivalent credit somewhere, double entry book keeping between accounts of the Treasury and the Federal Reserve should be considered as consolidated accounts. If so, a debit in the Treasury accounting books and a credit in the Federal Reserve accounting books cancel each other out as neither can be indebted to the other.
The Treasury may either spend (disbursements) or earn (taxes) money. When the Treasury receives tax payments these are deposited on so-called “Treasury tax and loan” (TT&L) accounts with commercial banks. When the Treasury needs to make disbursements it may call on the accumulated deposits of its TT&L accounts. To do so, the Treasury will transfer the funds from the TT&L accounts to the Treasury accounts at the Reserve Banks. The Federal Reserve Banks are authorized to transfer the amount of the Treasury call from the bank reserve account held at the Federal Reserve directly to the account of the US Treasury at the Federal Reserve. As the deposits in the TT&L accounts fall, this loss in deposits for the commercial banks translates in an equal amount of loss in reserves; therefore calling on deposits of TT&L accounts reduces the amount of reserves by an equal amount.
Suppose the Treasury now decides to spend the amount it has just called from the TT&L accounts. As the Treasury makes expenditures by drawing on its balances at the Reserve Banks, the funds find their way back to the banks in the form of deposits and the amount of reserves has been increased by the amount of the Treasury’s expenditures.
Putting all the above together, in a balanced budget world there is no accounting change in the level of reserves, only a transfer of wealth from the private sector to the public sector (tax collections) and then back again to the private sector (government spending).
Now let’s take this a step further and suppose that the Treasury must disburse more money that it has earned through taxes on its TT&L accounts or put plainly, the government is running a budget deficit.
In this case, the act of spending (adding reserves) is larger than the act of collecting taxes (withdrawing reserves). Therefore, government deficits should add reserves and as we know from simple accounting identities, a public sector deficit is a private sector net saving (as detailed in my previous letter).
The Treasury has theoretically two ways of proceeding. The first, and not one day passes without being front and center news, is by issuing debt (Treasury securities). If the Treasury issues bonds these will be purchased either by the banks or by the non-banks private sector (in both cases it will be purchased by the private sector of course). Whether the non-banks or banks purchase the Treasury bonds, these will be settled by a drain from the reserve accounts by the same amount. In the case where it is the non-banks it will simply impact deposits before reserves. At this point there is a reserve drain from the banking sector. However, the act of spending the money received from the bond sale by transferring it back into the private sector will reinstate the previous level of reserves by adding that amount of deposits in the banking system.
Notice that in all cases described above, the fact of withdrawing deposits from the TT&L accounts or issuing government bonds, the result is a drain in reserves by that same amount or put simply, a reduction in the money supply. The fact of the government then spending the money by transferring it back to the banks deposit accounts reinstates the level of reserves.
But there is another important element of this process that is often overlooked. This whole process of spending and issuing debt takes place in parallel. Said differently, deficit spending grows in parallel with private sector savings with deposits and excess reserves created equally in the banking system. Why care about the famous Bid-to-Cover ratio constantly monitored by the doomsayers in a process that creates the demand in parallel with supply? The idea that a government bond auction can fail is a clear misunderstanding of our monetary system. Demand is loaned into existence. If it’s not the non bank sector that purchases the issued bonds it will be the banking sector, eager to transform their low yielding excess reserves into high yielding government debt.
The second spending option, is by not issuing any government debt at all. Remember that one should consider the Treasury and the Federal Reserve as a consolidated entity as they are both government entities but with different mandates. In this case the Fed would simply credit the reserve balances held at the Fed while simultaneously instructing the banks to transfer that amount onto the account of the beneficiaries of government spending. The difference however, is that the amount of reserves in the banking system will have increased by the amount of spending by the Treasury without the ensuing reserve drain of bond issuance. The other important difference is that in this option, the government is not paying interest on its new expenditures. The choice of the Treasury between these two options comes down to whether they wish or not to conduct a reserve drain. At this point you might be thinking that this option is theoretically feasible; however this interest free money creation process is called monetization, and is highly inflationary.
This process is deemed inflationary for one and only reason: If these excess reserves are used by banks to provide loans, the money supply will increase exponentially as the fractional base of reserves grows in tandem with deficit spending. So goes the theory.
This is a fallacy of composition. Contrary to popular opinion banks do not seek reserve balances before they decide to offer loans. And if banks do not seek reserve balances prior to making loans, money supply and credit creation are not influenced by their level of reserve balances but by the amount of opportunities to lend at a profit. An increase in loans will increase the need for reserves, not the other way around. Banks generally increase their level of reserves in order to meet reserve requirements, or even in excess of reserve requirements when it is profitable to do so, by obtaining funds in the Fed funds market which are basically reserves traded amongst financial institutions (borrowing deposits from each other). These transactions do not alter the overall amount of reserves in the banking system. If the overall amount is insufficient, the Federal Reserve will always provide additional funds.
There are two ways the Federal Reserve may inject additional reserves into the banking system. It can either provide loans through the discount window or it may purchase government securities (open market operations).
Banks provide acceptable collateral to borrow at the discount window and will be charged the discount rate which is generally higher than the fed funds rate. Purchasing government securities from a bank will increase reserves at that bank and reduce its holding of securities. Repos (repurchase agreements) work in the same way although they are of a less permanent nature as the bank agrees to repurchase the security at the same price plus interest at a given date. This role of central banks providing reserves to the banking system leads us to believe that there is a causal role between money and reserves and that the central bank controls the money supply by adjusting the quantity of reserves. Makes sense, but it’s absolutely wrong. The supply of reserves through open market operations is not determined by a desired level of lending, it is the result of aiming to control the fed funds rate. If the Fed does not provide sufficient reserves through open market operations, banks will borrow reserves from each other which will bid up the fed funds rate. If the fed funds rate rises above the discount rate, banks will start borrowing from the Fed instead of borrowing from each other in the fed funds market. This mechanism does not limit in any way the access to reserves. Reserves are available to any bank that needs them, a bank never checks its reserve balance before providing a loan.
Open market operations do not seek to increase or reduce overall lending. The Fed must always provide enough reserves to meet known reserve requirements, either through open market operations or through the discount window. The only purpose of open market operations is to control the fed funds rate by providing additional reserves when it rises above the level targeted by the Federal Reserve and by draining reserves when it falls below the target. Excess reserves will ultimately tend to push the fed funds rate to zero, whilst insufficient reserves will simply switch demand from the fed funds market to the discount window. As for controlling the amount of lending, policy makers target a specific fed funds rate that will determine whether it is profitable or not for banks to borrow. A high fed funds rate will simply reduce the amount of profitable lending opportunities for banks. This is the tool to manage the amount of lending, not reserve levels.
Having showed that the amount of reserves provided by the Fed does not determine the amount of lending, let’s go back to the two available options of the Treasury for financing expenditures in excess of receipts. As we have seen, in the first option, the issuance of Treasury debt acts as a drain on reserves of the banking system, draining the same reserves it created by spending. If the Treasury chooses option two, it fails to drain the reserves it created by spending, thereby creating excess reserves. These excess reserves will push the fed funds rate lower as banks will seek to lend the additional reserves in the fed funds market. Excess reserves don’t entice lending but a lower fed funds rate does.
Fast forward to the emergency Economic Stabilization Act of 2008 that enabled the Fed to start paying interest on required and excess reserves. The interest rate paid on reserves acts as a floor on the fed funds rate, keeping the rate from falling below the rate paid on reserves as banks become indifferent on holding the reserves or offering them on the fed funds market. As both rates are now equal, the Fed no longer needs to issue government debt to support the fed funds rate. As a result, even if central banks generate a large quantity of excess reserves, these no longer need to be sterilized. The excess reserves do not change the target rate (lending policy) and have no effect whatsoever on bank’s incentives to lend to firms and households. When the central bank pays interest on reserves at its target interest rate, the textbook money multiplier model ceases to exist. There is no opportunity cost in holding reserves and the money multiplier becomes irrelevant. Taking this a step further, paying interest on reserves has simply given the Fed a new and efficient policy tool to adjust lending incentives without the need of removing excess reserves. I would not be surprised that this new tool remains part of the Fed’s normal operations in the future. The only remaining reason for issuing government debt is to offer banks an opportunity to exchange the interest paid on reserves for the interest paid on government securities.
So we know that reserve levels don’t influence lending and with the Fed now controlling the interest rate paid on reserves these can in no way be a trigger for inflation. Only a policy mistake, keeping rates too low for an extended period, could lead to inflation by stimulating excessive lending. In this world, monetization is not inflationary.
You may be thinking by now that I have written an ode to profligate spending. That sovereign issuers of currency in a non-convertible floating exchange rate system have no real spending constraints and that these have no economic consequences. Absolutely not. What I have tried to demonstrate is that these consequences are not those most evident at first sight. Government deficits do not have to lead to government debt issuance and if the alternative path of monetization is chosen, there is no direct link with future inflation. These are misguided shortcuts that lead to false assumptions.
Interest rates will not rise under the weight of government debt issuance. Excess reserves will not lead to a new wave of lending. There may be correlations at times but there are no causations. In our fiat monetary system, large increases in government debt levels will not translate into higher yields unless it is the government’s wish that they do. The system is designed in such a way so that they have complete control over the yield curve. In effect, it could be argued that their policy mistake was failing to intervene on the abnormally low long term rates between 2002 and 2007. Had they significantly modified the supply of government debt in favor of issuance of 30 year paper, they might have been able to prick the real estate bubble years before.
The next time you read an article warning of the impending rise in long term rates because the government is unable to sell bonds, or that the government may reach insolvency, or that Beijing is Washington’s banker, remember that the government has no technical obligation to finance itself through the issuance of debt obligations. Prior to 2008 the technical reason for government debt issuance was to support the targeted fed funds rate. That reason no longer exists. Countries that enjoy currency sovereignty have no technical limits on the amount of deficits they can run. Interest rates are entirely determined by the Federal Reserve which has complete control on the entire yield curve by controlling the supply and demand of government debt issuance.
So what are the consequences of deficits and elevated debt levels? Remember rule number one in economics: there is never ever such a thing as a free lunch. Somewhere, something will have to give and act as an adjustment factor and in the case of a country with double sovereignty, that factor is the currency. The exchange rate will bear all the pain needed for the economic adjustment to take place. If the monetization is excessive or the spending perceived as irresponsible the currency will gradually be sold driving it to lose value relative to others. The fall in the value of the currency, the adjustment factor, is the process that could lead to inflation, not the amount of excess reserves. The cause of inflation would be the result of a loss of purchasing power. This is the adjustment factor. A weakening of the currency will have to continue until it reaches a level that would gradually lead to an increase in exports which would ultimately reduce the deficit by creating a large current account surplus.
Disclosure: Long TLT