The Fed’s balance sheet used to be fairly small. For those of you who quite sensibly do not think much about central bank’s balance sheets, its size is determined entirely by its liabilities (there is only a small sliver of equity, as profits are pushed up to its owner, the Treasury). The three main components of these are currency in circulation (“dollar bills”), and deposits by the Treasury and private banks (the latter of which are often called “reserves”, although reserve requirements no longer exist). Currency (dollar bill) demand is relatively stable and small (although there can be seasonal effects), and is largely outside the control of the central bank (although severe negative rates might create demand for currency, which has an implicit interest rate of 0%). (Textbooks argue that there ought to be a portfolio demand effect based on interest rates, but the reality is that since institutional investors do not stick dollar bills into vaults, there is very little observable effect on holdings.) Government deposits are driven by whatever the fiscal arm of the government is doing. This just leaves bank deposits as a free variable.
Back in Ye Olde Days (pre-2008 in the United States), central bankers had not gone nuts with QE and so reserve balances were held at the minimum required (with very small excesses). (Note that reserve requirements still existed in the pre-QE era.) This allowed the central bank to operate with a small balance sheet. (Canada did not have reserve requirements, so the balance sheet was even smaller relative to the size of the economy.) However, the act of purchasing bonds by the central bank creates deposits (“money”) at the central bank that cannot removed by (sensible) private sector transactions, so the balance sheet increases (the added bonds show up on the asset side, the private bank deposits on the liability side).
An Immediate Unwind is Easy, Harder to Do Later
The issue raised by the Beckworth article is that the larger balance sheet of the central bank is matched by balance sheet changes in the banking sector. They necessarily are stuck with more deposits at the central bank. Although these deposits are safe, they also are not lucrative — nobody in their right mind wants to invest equity capital in a private bank so that it can park funds at the central bank and earn the overnight rate. The investor can cut ought the middle-person and just buy a Treasury bill fund and achieve the same outcome.
If the central bank “almost immediately” unwinds the balance sheet expansion, not much would change. There might be a bit of thrashing around of positions in the money markets, but the aggregate system returns to its previous allocations and so nothing much has changed. The issue is what happens when the central bank distorts private sector banks for an extended period?
The argument is — and something that I previously largely ignored — is that the large reserve balances get baked into what is the “expected” bank balance sheet. Both internal and regulator stress tests start using the current levels of reserve balances as “normal” — and so so there is an expectation that banks will preserve those levels.
What happens if we get a Fed Chair who insists that the Federal Reserve shrink its balance sheet to match its previous size relative to the size of the economy? Since such a step mathematically forces private banks in aggregate to shrink their reserve holdings, we can see that if all banks are targeting keeping their reserves at current levels, something has to give.
Pressure on Weak Links
If we just look at the aggregate bank balance sheets, QE/QT seem manageable. Transactions are two-way flows, and so the effects on bank balances should appear to be balanced. The reality that not all banks have identical asset/liability mixes is what causes problems. The clients of banks who ultimately buy bonds are exchanging their bank deposits for bonds, which means that their banks face a liquidity drain. (They lose a deposit liability, but have to make a payment into the payment system.) Unless the banks receiving liquidity push their excess into the interbank market, we see that problems arise.
An alternative way of looking at this: if the macro brain trust at the Fed decides that they want to shrink aggregate bank “reserves” by 20%, there has to be a mechanism to ensure that banks desired reserve balances shrink by 20%, allowing them to flow from banks not losing 20% of their reserves to the ones losing more than 20% of their reserves.
This is not impossible to achieve, but it would require central bankers and bank regulators to have a grasp of how the macro accounting lines up with bank-level accounting. One way to smooth the process is cancel out bond selling via the central bank getting involved in term deposit transactions with banks (a point discusses in the Beckworth article).
Term Deposits
The current operating procedures of the Federal Reserve are complex, and this complexity obfuscates the true economic purpose of its financial transactions. The complexity is usually justified on the necessity to follow various laws defining what the Fed can and cannot do, which is a rather precious and archaic attachment to the Rule of Law in the current political environment.
But if we were able to strip away the complexity, the Fed could just offer term deposits to private banks. They just publish a grid of bid/offer interest rates on loans on a curve over a number of maturity dates. Banks can either borrow or lend at those rates on a guaranteed basis with the central bank, without having to run off to various wacky wholesale funding markets. (The Swedish central bank — the Riksbank — already does this.)
(Note that a term deposit is economically equivalent to a non-marketable bond. So if the central bank takes in loans via term deposits, this is equivalent to issuing non-marketable central bank bonds. Going the other way, the central bank is buying non-marketable private bank bonds.)
This allows a way to avoid the problems with QT. Let us imagine that a not very large bank has a client that buys $10 million in bonds from the Fed, which hits the bank with a $10 million dollar unexpected outflow. Instead of scrambling elsewhere, the bank just sidles up to the Federal Reserve and takes a $10 million term deposit loan for a week to cover the outflow, giving it a week to figure out how to patch its liquidity hole.
However, observant readers will have noted that this sequence of events does have an important side effect — the central bank’s balance sheet is unchanged. It lost a $10 million bond from its assets, but replaced it with a $10 million term deposit at the bank. In theory, the term deposit is self-liquidating, but the liquidity drain associated with paying off the term deposit might force the release of another term deposit.
Lessons Learned
There are two lessons to be learned from this experience.
Although it was easy to mash on the “Buy Bonds” button at the central bank, the “Sell Bonds” button has to be pushed in a careful fashion, and the entire bank regulatory apparatus needs to take into account the effects on bank balance sheets. This means that the mashing needs to be done at a deliberate pace.
Quantitative Easing was a stupid policy that was based on cargo cult monetary economics. If banks needed liquidity, just do lender-of-last-resort operations. If the objective is to shape the yield curve, do not attempt to rely on quantitative “supply and demand” effects on yields. Instead, be honest, and use price signals: just offer yield targets that which the bank will transact.
No comments:
Post a Comment
Note: Posts are manually moderated, with a varying delay. Some disappear.
The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.
Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.