Recent Posts

Thursday, March 26, 2015

Minsky On Bank Position-Making Instruments

Although the financial system appears extremely complex, the underlying principles tend to be stable over time. One of the fundamental needs of a banking system is for position-making instruments, which are financial instruments that can be bought and sold to manage liquidity. The economist Hyman Minsky often highlighted the importance of these instruments. This article summarises his discussion of this topic, which is useful for understanding some of the operations underlying government finance, in particular the repo market.

The quotes here are taken from his book, Stabilizing An Unstable Economy, which was first published in 1986, and republished after the global financial crisis. Page numbers refer to the second edition (2008). The end of this article turns into a longer discussion of some position-making instruments, which is my expansion upon Minsky's discussions.


Minsky analysed the real economy from a financial ("Wall Street") perspective, which differs from the conventional economic view which starts from a barter economy.

For any economic unit (corporation, household or even government) that finances itself with short-term debt, it faces risks associated with cash outflows. The core assets of most economic units are illiquid, such as bank loans, or the physical plant of a manufacturing corporation. These assets are termed the positions of the unit.
The act of acquiring cash to finance the assets essential to a unit's business is called, following banking terminology, making a position, and the instrument used for such purposes is the position-making asset or debt. An asset or debt is a good position-making instrument if it has a broad and active market. Furthermore, the market for a position-making asset should be resilient in that there will be a flood of orders to buy this asset if the price falls a bit; its price, then, will not change much under normal sales pressure. (Page 80.)

The Evolution Of Bank Position-Making Instruments

Chart: U.S. Bank Treasury Holdings As A % Of Assets

In Chapter 4 of Stabilizing an Unstable Economy, Minsky traced out the history of position-making instruments within the banking system. The chart above gives a very good summary of this evolution.
At the end of World War II the commercial banks were replete with government securities. The government-security market was the primary position-making market, and the Treasury bill was the primary position-making instrument. Banks that had excess cash would buy Treasury bills, and banks that had cash (reserve deposit) deficiencies would sell Treasury bills. These sales would go through either independent dealers or dealer department in large banks. (Page 81.)
However, the reduction in government debt-to-income ratios meant that those large Treasury weightings could not be maintained. Additionally, expectations that the United States economy would re-enter a depression faded as the automatic stabilisers and Keynesian policies held out the promise of taming the business cycle. This rekindled the animal spirits of bankers, and risk was once again added to bank balance sheets over time.

This also led to complexity.
We now have a banking system in which normal functioning depends upon a wide variety of money-market instruments being available for position-making. Since the end of World War II, the banking system has evolved from the simplicity of the Treasury bill's monopoly as the position-making instrument to a complex situation in which a representative bank juggles its government-security account or its federal-funds position, has large denomination certificates of deposit [CDs], repurchase agreements, Eurodollar borrowings (or sales), and borrowings at the Federal Reserve. (Page 86.)
This in turn leads to an increasingly fragile financial system, and hence economy.
When banks sell CDs or enter into repurchase agreements, a substitution of bank time deposits, or promises to pay, for demand deposits takes place. Such transactions increase the ability of the banking system to finance activity. But the financing banks provide tends to be short-term; thus, the measures that allow bank financing to grow at a rapid rate leads to an increase in the short-term financing of nonbank activity. Rapid growth of short-term financing tends to make the financial system increasingly fragile. (Page 87.)

Financial Repression, Or "Help, Help, I'm Being Repressed!"

Some economists having been trying to sell the theory that there is a sinister move underfoot by governments to put in place "financial repression" as a result of rising government debt burdens. There are two planks to this "repression": low policy interest rates, and moves to raise the regulatory liquidity buffers (expressed in terms of government security holdings) at banks. These claims are dubious.

Firstly, policy rates are low due to the mental contortions of New Keynesian economic models that are in place at central banks. Are we to believe that Woodford et al were raised in prominence by some conspiracy at the Treasury decades earlier?

Secondly, raised liquidity buffers are exactly what a sensible analysis (such as by Minsky) would suggest to do. Bankers were left in a lightly regulated state, and they proceeded to burn down the global economy in a very efficient fashion. Regulators will have to rein in their animal spirits, as it appears that the private debt markets and shareholders will not do so. The time series plotted above is going to have to be pushed back to a more sensible level.

Interbank Versus Shadow Bank Position-Making Instruments

As Minsky noted, there are a lot of position-making instruments. The complexity partially arises out of the alleged need to arbitrage regulations. But I would divide them into two broad categories: interbank instruments, and "shadow banking" instruments.* I will take a look at two examples: the fed funds market, and the repo market.

The market in federal funds ("fed funds") is an interbank market. In it, banks trade balances (reserves) at the Federal Reserve banks. It is the instrument that is targeted by the Federal Reserve (there used to be a target interest rate; it is now a target range). It is often used as a proxy for default risk free rates in the U.S. dollar fixed income complex. Since I was never involved in bank treasury operations, I will not attempt to go into details here.

However, a key issue with this market is that it is in fact largely open only to banks, and there are large financial institutions which do not take part. This creates the effect that the fed funds rate can deviate from other risk free rates, possibly on the order of 20 basis points (0.20%). This can cause hand wringing amongst monetary policy wonks, as "the Fed has lost control of interest rates!"

A more realistic response is: deal with it. The fixed income market continuously has to deal with small spreads between various financing rates that are theoretically supposed to be equivalent (generically known as "basis risk"). Unless the basis blows out to hundreds of basis points, it does not matter. And if various bases have been driven to zero, that is probably the time to get nervous. This means that leveraged speculators have squeezed every last basis point out of convergence/reversion trades, and things can only go one way thereafter.

The other position-making instrument that I discuss here, the repurchase agreement ("repo") might be more important in practice. This is because it is used throughout the financial system - banks, nonbanks, and even the Central Bank (although they typically only trade with primary dealers). However, I am unsure whether repos should be considered an "instrument" or not, as they are just collateralised lending against other financial assets, most notably government bonds and bills.

An Introduction To Repos

A repurchase agreement is an agreement between two parties to:

  1. one party sells a security to the other at a fixed price now;
  2. it undertakes to buy the security back at a fixed future price at a later date. (Quite often overnight; term repos are agreements with longer holding periods, such as three months).
The difference between the buying and selling price determines an implicit interest rate. For example, if a bond was purchased now for $100 (full invoice price, or "dirty price") and sold back at $100.10 at the end of the repo, the $0.10 difference is a form of interest payment (10 basis points for the period). As is the case for all other fixed income instruments, the implied interest rate is calculated on a annualised basis using some arcane day count convention.**

Although there are legal and accounting differences that I am unqualified to discuss (nor do I have an interest in them), a repo is operationally equivalent to collateralised lending - the borrower is the side that sells the bond. It is just that the lender takes possession of the collateral outright.

General Collateral Repo Rate

For most government bonds, the market repo rate is uniform for a given term, which is known as the "general collateral repo rate". Some bonds trade "special", with a lower repo rate. (Owners can borrow against them at a lower rate of interest.) This occurs when there is demand to borrow a particular issue so that it can be sold short (for example, by dealers looking to hedge their interest rate risk). Typically, only recently issued bonds go "on special", although it can happen to bonds that are "cheapest to deliver" into futures contract. The fact that lower financing is available explains how benchmark bonds can trade at a premium (lower yield) than nearby non-benchmark bonds.

This general collateral repo rate is actually the most important risk free rate within an economy, as it determines the pricing of the bulk of the government bond curve. However, this rate is somewhat more opaque and less well known, which is why economists often focus on the fed funds rate.

The Joy Of Repos

Repo transactions involving government bonds allow it to function in its "efficient" modern format. For example, they allow investors to buy and sell bonds forward. Additionally, they allow the oversized money market fund complex to integrate with the formal banking system.

A government bond repo is a very secure investment for a money market investor. There are two legal entities promising to repay the "loan".
  1. The counterparty of the repo transaction is legally obligated to buy back the instrument at the agreed price, regardless of its future market value.
  2. The issuer of the bond (in this case, the central government) has a legal obligation to make payments on the underlying collateral until its maturity date. To protect against adverse market value moves, the loan advanced is less than the original market value of the collateral, by an amount of the so-called "haircut". For United States Treasury securities, I believe that the standard haircut is 2%; that is, you borrow 98% of collateral market value.
This makes it possible to synthetically convert long-term bonds into money market instruments (for a short period of time). This is needed, as there is a structural imbalance between the allocation towards money market instruments and the demand for short-term borrowing. In order to provide money market funds with "safe" assets, government debt issuance might have to be almost exclusively short term.

On the other side of the trade, repos are a good position-making instrument for banks. A bank treasury may expect to only need funding for a short period, such as two weeks. They do not want to have to first sell and then buy back Treasury securities outright, as that may expose them to capital gains and losses, which they would prefer to avoid. Even small price changes on Treasury bills are unwelcome if they are on large positions.

The Pain Of Repos

However, not all repos are against central government debt securities. Any paper issued by the private sector can be converted into a money market paper via a repo transaction against it. And up until 2008, pretty much anything was. The problem is that once confidence in private debt unravels, the willingness of money market investors to lend against these private sector liabilities disappears. (The "haircuts" increase, possibly to the point where the market is closed for such collateral.)

Any entity that was funding positions in those private sector bonds ran into illiquidity problems as the position-making instruments it was using cease to trade. This had the effect that we saw during the financial crisis.

Money market funds are an innovation to create something that looks like the liquidity and funding management features of formal banks, but bypassing pesky things like bank regulations and capital. But they are really only focussed on funding, and their cut-price operating model means that they lack the capacity to do thorough credit analysis. Correspondingly, it is no surprise that financial crisis often emanate from the money markets.

Central Bank Repo Operations

Central banks often use repos to adjust the money supply. Within "overdraft economies", where the central bank does not buy government bonds outright (for example, the ECB until recently), they are essentially a necessity, As discussed earlier, this is because it allows them to interact with the wider financial market (indirectly intermediated by primary dealers), and because repos are a more convenient instrument.

The pre-2008 operating procedure for the Federal Reserve was to have a large "permanent" portfolio of Treasury bonds owned outright, with a large repo book. The size of the repo book could rise and fall with the demand for reserves within the banking system.  Since the Fed's bond portfolio is essentially held on a cost basis***, this allows it to never book losses on bond holdings, allowing it to report steady profits. This means that it does not need to run to Congress to authorise a increase in capital. Such a capital increase is largely an accounting fiction, but as we have seen with respect to the Treasury debt limit, avoiding Congressional theatrics is often a good political policy.

Turning to my series of articles on government financial operations, I am going to discuss a simplified framework which is an amalgam of current Canadian operating procedures (no "bank reserves") and the early post-World War II system described by Minsky, where Treasury bills are the position-making instrument. Technically, no government financial system follows the simplified practices I describe, but it is very close to those existing frameworks. Once it is understood, it is straightforward to add "chrome". 


* I use the term "shadow banking" in a very wide sense; for me, shadow banking is all non-bank finance. That includes the bond and money markets, which are not particularly "shadowy", since CUSIPs etc. are in the public domain. However, "shadow banking" sounds cooler than "nonbank finance".

**  The interest rate also needs to take into account whether coupon payments are made during the period. Additionally, bond prices are typically quoted on a "clean" basis, with accrued interest added.

*** I am unaware of the exact accounting conventions used by the Federal Reserve. But the fact that they report steady profits, despite the huge swings in market value of government bonds over the years, indicates that the bulk of their portfolio is effectively accounted for on a cost basis.

(c) Brian Romanchuk 2015


  1. Great summary of a great book, one that every serious student of finance and economics should read. Minsky's work is truly exceptional.

    1. Thanks. At some point I should do a "review" of the book (more of recommendation, since it is old).

  2. Repo are bankruptcy-remote and may introduce systemic risk due to re-hypothecation.

    1. The re-hypothecation issue seems to be a much greater issue for private sector collateral. There's concerns about Treasury repo fails, but I found them to be fairly arcane. Things get messy for the firms involved, but I doubt that things can escalate that much.

  3. This comment has been removed by a blog administrator.

  4. This comment has been removed by a blog administrator.

  5. This comment has been removed by a blog administrator.

  6. This comment has been removed by a blog administrator.

  7. This comment has been removed by a blog administrator.

  8. This comment has been removed by a blog administrator.

  9. This comment has been removed by a blog administrator.

  10. This comment has been removed by a blog administrator.

  11. This comment has been removed by a blog administrator.


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.