Some new thoughts on the “can you make losses printing money” theme, this time with a Hans-Werner Sinn angle.
My thoughts on the ECB’s QE announcement. Perhaps a bit long, perhaps a bit disorganised ….
While the Federal Reserve’s quantitative easing program has not yet caused an increase in inflation, many are concerned that the money creation due to QE will provoke serious problems at some point. So when leading economist Robert E. Hall of Stanford releases a paper (co-authored with Columbia‘s Ricardo Reis) focusing on the “risks to solvency” facing the Fed then many people are going to pay attention. This is particularly so when the paper includes scenarios in which the Fed loses control of the price level and “the central bank would likely be dissolved and a new currency introduced.”
So is the Fed going to go bust and be dissolved? Will the dollar have to be replaced? Let’s start with a spoiler alert. In the end, Hall and Reis conclude the possibility of a doomsday scenario for the Fed is “remote”. Still, given the nature of controversies about the Fed, the extreme scenarios cited in the paper are likely to become part of public debate on the Fed’s future. In this post, I argue that even the “remote possibility” scenarios put forward by Hall and Reis rely on a range of counter-factual assumptions.
A Potential Problem?
Hall and Reis’s concerns can be summarised as follows. Because the Fed now holds very large amounts of long-dated low-interest-rate bonds, future increases in interest rates may reduce the value of these assets. They describe a scenario in which the Fed reaches a point where its assets don’t generate sufficient income to cover its costs.
In particular, Hall and Reis are concerned about the Fed’s ability to pay interest on the reserve accounts that banks hold with it. Payments made to Fed ultimately arrive via deductions from various reserve accounts that institutions such as private banks or the Treasury hold with it while interest paid on reserves to banks takes the form of additions to the relevant bank reserve accounts.
Given this structure, if the interest payments made to the Fed don’t cover the interest payments it makes on reserves, then the total stock of reserves will increase. Hall and Reis express a concern that this process of additional money creation stemming from such an increase in the stock of reserves would result in higher inflation. (Exactly how this gets so bad that the currency ends up being dissolved is a little unclear to me.)
So it turns out that Hall and Reis are not actually concerned about the Fed going bust in the usual sense of being unable to meet its obligations. Rather they express a concern the Fed might only be able to meet its obligations by creating a quantity of money that generates high inflation. Alternatively, they discuss a scenario in which the Fed keeps the interest rate on reserves low to restrict the growth in total reserves but this low interest rate results in a loss of control of the price level.
After crunching some numbers, Hall and Reis finally assure worried readers that their scenario in which the Fed loses control of inflation isn’t a likely event. However, this still leaves some room for doubt that there may be some more extreme set of losses than those contemplated in the paper that could result in a doomsday scenario.
A Problem that Simply Isn’t a Problem
In fact, the doomsday scenario outlined by Hall and Reis simply doesn’t make any sense and it doesn’t require crunching of numbers to dismiss it.
The scenario revolves around the problems caused by payment of interest on reserves. However, it’s worth noting that, far from being a time-honored tool for controlling interest rates, the Fed only began paying interest on reserves in 2008 and had little difficulty controlling interest rates prior to the introduction of this policy.
As explained in this paper, the principle argument for paying interest on reserves is that it acts as a floor on money market interest rates and so reduces day-to-day volatility. However, when central banks pay interest on reserves or a deposit facility, this interest rate is rarely the one used to set the target for short-term private sector borrowing rates.
For example, the ECB has a deposit facility that pays interest to banks, an emergency marginal lending facility for overnight borrowing and a weekly refinancing operation in which it lends money to banks. It is the refinancing rate that usually dictates average money market interest rates with the deposit facility rate acting as a lower floor. So, in general, there are few grounds for the idea that setting a low interest rate for remuneration of reserves (or indeed a zero rate) will lead to the Fed losing control of inflation.
Still, with banks currently flooded with reserves, one could argue that it might be difficult over the next few years to raise the federal funds rate without hiking the interest rate paid on reserves. So one could still ask whether concerns about income shortfalls could lead to the Fed having to abandon paying interest on reserves and the operational efficiency gains that come with it. This is extremely unlikely.
Firstly, the Fed holds large amounts of assets that are set against currency “liabilities” which in practice pay no interest. So it could incur losses on the rest of its operations and still make a profit. Moreover, there is no operational need for the Fed’s stock of assets to exceed its stock of notional liabilities.
Secondly, the Fed holds long-term assets and short-term liabilities and yield curves generally slope upwards. So, on average, the Fed is going to make profits even without the assets backing currency notes. It could thus go through a period of making losses and simply offset these losses with later profits that are partly retained rather than handed over to the US Treasury.
Hall and Reis note in their paper that this latter option of smoothing out profits and losses by varying dividends to the Treasury may not be possible if the Treasury decided to prevent the Fed from accumulating too much profit income without handing it over to the rest of the US government. Here I can only say that any future world in which the Treasury is allowed to bully the Fed into providing it with more income would indeed be one in which price stability is endangered but only because the Fed will have lost its independence, not because of QE-related losses on assets.
The US economy has many real problems for people to focus on. Economists can help by focusing on those problems rather than fairy stories about the Fed going bust and the dollar being abandoned.
After Mario Draghi’s hints that the ECB is getting ready to purchase large quantities of Spanish and Italian bonds, expect to read lots of commentary claiming “the ECB is taking huge risks with its balance sheet”, that “the ECB risks becoming insolvent, endangering the future of the euro” or that “Eurozone states may have to recapitalize the ECB at huge cost to taxpayers”.
Much of this commentary is based on misunderstanding the facts about the Eurosystem’s balance sheet and the meaning of central bank balance sheets. In this post, I want to briefly explain what a central bank balance sheet is, then go through the basic facts about the capital position of the Eurosystem of central banks and then argue that the common focus on central bank solvency is misplaced.
Central Bank Balance Sheets
Like all businesses, central banks publish balance sheets that show their assets and liabilities. However, central banks are not just any business. They have the power to print money that is accepted as legal tender. So their assets have been acquired via printing of money. Central banks list the money they have created as liabilities and the gap between the current value of their assets and liabilities is labelled “capital”.
So a stylized central bank balance sheet looks as follows.
If a central bank purchases assets that then decline in value, it could end up having negative capital. When a commercial bank has negative capital, it is termed insolvent and either re-capitalised or shut down. The “insolvent” terminology is sometimes applied to a central bank in this situation but, as I discuss below, central banks are unique organizations and this phrase isn’t particularly appropriate.
Bond-Buying Risk: Eurosystem not ECB
A common confusion that arises when people discuss potential losses stemming from sovereign bond purchases or loans to banks comes from the assumption that the risk of these operations lies with the European Central Bank. For example, it is common to see commentary that examines the ECB’s balance sheet and concludes that the ECB faces a significant risk of becoming insolvent. This is because the ECB lists “Capital and Reserves” of only €6.5 billion, a tiny amount relative to the likely scale of bond purchases likely to be executed under the Securities Market Programme (SMP).
In fact, the ECB also has revaluation accounts and provisions, both of which can cover losses, worth €30 billion. More importantly, the risks associated with the bond-buying program are in fact shared across the ECB and all of the national central banks in the Eurosystem. So what matters when thinking about the threat to solvency of the SMP is the size of the Eurosystem’s capital resources.
The ECB releases a Eurosystem financial statement every week. Currently, it shows capital and reserves of €86 billion and revaluation reserves of €409 billion. These figures show the Eurosystem would have to incur much larger losses on its bond purchases than is often believed before the question of insolvency would become an issue.
Would Eurosystem Insolvency Matter?
This still raises an interesting question. What would happen if the ECB published its weekly balance sheet one day and it showed negative capital? For many commentators, the answer is apocalyptic. The euro will have lost all credibility as currency, hyper-inflation will ensue, that kind of thing.
In truth, the answer is that not much at all would happen. The euro is a fiat currency. In other words, unlike the days of the gold standard, the Eurosystem does not promise to swap the euro at some specified rate for another asset such as gold. The currency is not “backed” by the central bank’s assets.
Since central bank money costs next to nothing to create, the “Liabilities” on its balance sheet are essentially notional. Indeed, a central bank’s asset holdings could fall below the value of the money it has issued – the balance sheet could show it to be “insolvent” – without impacting on the value of the currency in circulation. A fiat currency’s value, its real purchasing power, is determined by how much money has been supplied and the various factors influencing money demand, not by the stock of central bank assets.
Taken to an extreme, one could argue that a central bank with sufficiently negative capital could face problems implementing monetary policy because its stock of assets may be so low as to limit its ability to sell assets to reduce the supply of privately circulating money. However, detailed analytical research by ECB economists Ulrich Bindseil, Andres Manzanares and Benedict Weller concludes that
central bank capital still does not seem to matter for monetary policy implementation, in essence because negative levels of capital do not represent any threat to the central bank being able to pay for whatever costs it has. Although losses may easily accumulate over a long period of time and lead to a huge negative capital, no reason emerges why this could affect the central bank’s ability to control interest rates.
Despite the absence of reasons to be concerned about negative central bank capital, it is still unlikely that such an outcome would be allowed to continue in the Eurozone. The actual rules relating to central bank solvency seem a bit murky but it is generally understood that any element of the Eurosystem that had negative capital would need to be recapitalized by governments providing them with assets.
Even then, these “recapitalizing” transactions wouldn’t actually have any effect on the net asset position of Eurozone governments because central banks hand over their profits to the government. For example, suppose a government hands its central bank a bond that pays €500 million in interest each year. That raises the net income of the central bank by €500 million and thus raises the amount that is handed back to the government by €500 million. The bond has no net cost to the government.
By these comments, I’m not suggesting that bond purchase programs have no cost. To the extent that they involve increasing the money supply, they can contribute to inflation so there is an implicit cost for all citizens. But this is the reason central banks need to be careful with their purchases; concerns about central bank solvency are beside the point.
Unfortunately, despite its lack of substantive importance, the idea that the Eurosystem needs to be protected from balance sheet insolvency has featured heavily in discussions of policy options. Indeed, it is well known that senior ECB officials are extremely concerned with protecting their balance sheet and have employed their “risk control framework” to protect themselves from losses at many of the key moments in the euro crisis.
One can only hope that the policy debate about the upcoming bond purchases focuses on things that matter, like the future of the euro, and not on things that don’t, like the ECB’s capital levels.