“All banking is a swap of IOUs.”
For notes on a related course see: Macroeconomics: Marginal Revolution University Course.
When looking at a bank, we check assets (like bonds and credit) and liabilities (like, crucially, money from depositors). A bank is solvent if assets > liabilities. A bank has liquidity if it has enough money to ‘clear’ the day, which is a day-to-day duty.
The course thinks of liquidity first, as opposed to starting from solvency. Shadow banks are a solution to the problem of liquidity.
Shadow Bank: an entity that holds short-term assets with as little risk as possible, which they can use as collateral to acquire liquidity (borrowing from the money market -mutual funds, corporate investors-) to then provide to banks. The money running through shadow banking is bigger than through ‘regular’ banking.
Hierarchy of financial instruments
In an idealized Victorian, Gold-standard system with no shadow banking.
If you map out a pyramid where gold is the apex, and securities the base, it gets wider as you go down and it has liabilities on the right and assets on the left, then what you see is it expands and contracts cyclically over seasons (days, weeks, etc.). It looks like it is organically breathing.
The pyramid expands in booming periods, and contracts in recessions. When it expands, credit ‘looks more like money’ and when it contracts the difference becomes starker. I am guessing if you raise interest rates, this pyramid also contracts.
Principle of Elasticity (of derivatives) vs Principle of Scarcity (at the top of the pyramid) or Discipline. You could look at a system and say it is more disciplined or more elastic at a point in time.
Market makers: For each level of the hierarchy there’s a “market maker”: central banking for gold to currency, banking system strapped between currency to deposits, and security dealers between deposits and securities.
When a market stops functioning for lack of elasticity, the central bank, which is the only market-making institution that isn’t solely motivated by short term profit, may step in and provide liquidity in exchange for securities.
Things a central bank does:
As per Bagehot, the central bank needs to “Lend freely at a high interest rate, against good collateral” during crises. This is what the Bank of England was already doing organically, but not explicitly. Bagehot brings it to public consciousness.
Countercyclical policy (raise interests when the economy is too boomy, lower them when it contracts too much) and financial policy (regulating banks to e.g. maintain standards of solvency) comes from this fact (the central bank will be responsible on the ultimate crisis, so let’s prevent it).
Monetary policy, by way of changing interest rates, aims at making the system more/less elastic.
The Fed regulates overnight interest, whereas the market sets the price of long term bonds. You may for instance raise interest of overnight lending above 30-year one, in which case you’ll see a huge contraction, or just raise or lower a little to make small impact.
Before the Fed, you had regional or local banks, but no central one. The big banks in NY (especially JP Morgan) served as de facto central banks in that they acted as intermediaries for international commerce. There was the gold standard, and banks were forced to deposit reserves in reserve banks, which in turn had theirs in NY banks.
In the Civil War, the state withdrew all gold from the banking system (!) to use for international purchases (presumably of war equipment), and started printing legal tenders, which were by law to be accepted as currency, but internationally devalued fast as they were not exchangeable into gold.
Eventually the supply of tenders was too big and the government switched tactics towards giving bonds to banks that could be used as collateral to emit currency. This means independent of the emitting bank, the bank notes could be exchanged (presumably at some government office) for bonds. From this point up until the creation of the fed, the national base money supply was fixed.
Fast forward to 1907. There was a seasonal component to liquidity due to the economy being mostly agricultural, whereby on harvest time farmers would withdraw their deposits to do farm stuff. This forced the long tail of small banks to withdraw reserves from regional banks, that withdrew from NY banks in turn. Thus, once a year there would be a liquidity crunch, asset prices would fall (due to worse access to credit and speculators having to sell to give back loans to farmers) and the effects would ripple to the international (British) financial system.
The Fed was partly created as a response to this, by Woodrow Wilson right before WWI. The hierarchy of banks becomes
An agricultural country has a seasonal, yearly fluctuation of the demand of bank notes and deposits. Let us then seasonally adjust the supply.
The mechanism by which this is made is: member banks loan to Main Street freely, and don’t need to worry about liquidity: they can acquire bank notes from reserve banks, borrowing them with their discounted loans as collateral. The reserve banks at the same time can use the discounted loans as collateral to borrow from the Federal Reserve, from which they may obtain bank notes, which they then provide to member banks completing the process. This way the supply of bank notes can adjust to demand.
The fact that the Fed ended up swamped in Treasury Bills instead of being mostly full of Main Street loans is a historical accident, and not something envisioned by its framers. It’s due to the need for financing WWI and it stuck.
Minsky hypothesis: the financial system is inherently unstable (swinging from too contracted to too elastic), and it is the banker’s job to vie for stability in its swing.
This lecture first proposes a very simple, finite money-only system where the only two individuals can trade using IOUs for goods, to solve for the seasonality of their production, assuming they can trust each other.
The pool of IOUs is fixed, and so in a case of disparity a liquidity problem may emerge. Alternately, we may propose an intermediary that holds the IOUs and emits its own liabilities, even for more than 2 actors. These liabilities in the form of credit are functionally money, and they are backed by the IOUs, but unlike the model before the outstanding deposits may grow indefinitely (up to a borrowing limit per actor). This credit system is how the payment system works, even today, though it now has much more complexity. Credit is necessary in order to facilitate elasticity.
Minsky proposes analyzing the economy by looking at units in terms of cash flows (ignoring goods). Units include households, banks, the State, etc. The emphasis, as expected given the background of the course, is in liquidity, rather than solvency, and especially in the strong liquidity constraints imposed on e.g. households and businesses. ‘Your wealth doesn’t matter if you can’t pay – Liquidity kills you quick’.
Survival Constraint: Cash inflow > Cash outflow.
At the time of an outflow, it is optimal to pay with cash reserves (discipline principle) but we generally want to use the elasticity afforded by credit. Thus we may instead borrow, or sell assets to pay. However money is the only payment we can take without depending on a counterpart.
Several firms may have the same net worth (and all be solvent) and yet face different liquidity constraints. E.g. 1 bank that needs to deposit X overnight every night and knows its cash flow will be 2X/day faces no liquidity constraint. If instead it needed to do just one 3X deposit the first day (and no others), it would need to somehow come up with the money to do it.
Thus we may see equally solvent firms behaving differently due to solvency, and as a response to how aligned their future flows are. This is partially known at any point as they have different bonds, loans, etc. Note that their creditors also know this and will adjust their own behaviors accordingly, adding complexity to the market.
It is a reasonable claim that the money market interest rate is a good measure of how solvent the entire system is at any given point: if too many firms are facing constraints, it will shot up, and the opposite happens if there is enough money in the market.
You could have a payment system purely based on money: each bank has reserves and any transfer between banks will imply a transfer of reserves (whereas intrabank transfers are free). A credit-based system however would allow you to change deposit amounts without affecting reserves directly. (example with the One Big Bank model).
Recommended reading: “Chapters on the Theory and History of Banking”, Charles F. Dunbar.
When we move to a N-bank system, particularly modeling the interaction between two banks, two possibilities emerge (assuming a credit-based system).
Starting point: Bank A keeps a deposit on Bank B, and there is an equivalent deposit from B in A. Assuming customer α is sending money to customer β, A notifies B and B increases β’s deposit (credits β), while A debits α. Then either:
The difference is which of the banks shrinks its balance sheet. Though logically both options would be plausible, in real life the accounting is usually kept by the higher ranked bank (e.g. a regional bank keeps a deposit from a local one). If a deposit reaches 0, the smaller bank will increase it again periodically.
This presents a big convenience advantage over a purely money-based system where banks would have to transfer reserves upon every inter-bank transaction. Note that there is also periodic netting (I guess daily or hourly, though the course doesn’t mention it just yet), so that correspondent balances (the interbank deposits) are only updated on checkpoints (reminiscent of database log flushing -it’s in Spanish, gomen ne-).
“You create billateral credit systems so that the reserves don’t have to flow”.
Originally there were bank hierarchies: hinterland banks < regional banks < NY banks (Citibank, JPMorgan, etc.).
You first net payments to reduce transaction count, then just shift correspondent ballances in the appropriate (ultimately NY) banks. NY Banks need to clear with each other, and that’s where eventually the central bank (big clearinghouse) comes in.
The clearinghouse was established by all the big NY banks. It had gold reserves and emitted clearinghouse certificates.
Banks would net all their payments to and from other NY banks, add them all up as if they were too and from the clearinghouse, and finally at end of day would either claim certificates if net receivers, or pay if net payers.
There is a credit expansion over the day that creates elasticity, and then a clearing at the end of the day that creates discipline.
The whole system is elastic, because the clearinghouse provides elasticity.
If a bank cannot clear with their clearing certificates (because they ran out) they may borrow from others (especially, in practice, JPMorgan). If no one wants to lend them, usually for insolvency reasons, that bank has to default and is banished from the system, after which the rest of the members will clear the debts using the gold reserves. This typically would not happen, but is an example of an aggregate of agents being more robust than any individual one and providing security and solvency.
In times of crisis, a feature was added whereby the clearinghouse could provide, with good collateral and high interest (e.g. 6%), clearinghouse loan certificates to member banks, which could be used as currency when they were deemed solvent but illiquid.
This sort of ‘lender of last resort’ system was a precursor to the federal reserve, but it being a mutual association that left other banks out it was not seen as democratic enough, one of the reasons or the creation of the Fed after 1907.
Eventually when the Fed was created, it operated on similar principles. Its reserves (assets) were gold, and its liabilities would be FR notes, and bank deposits (reserves).
If a bank cannot meet its requirements (it needs to transfer from bank A to bank B as a net credit, and it has no balance at the Fed), it has three options, analogous to those in the clearinghouse:
Note that for most transactions it’s just a balance update at the Fed.
What can kill a clearinghouse? Draining its gold. There are two kinds of drains: internal drains (people withdraw money from banks), which can be solved by the Fed and the clearinghouse, and external drains (Bank of Englan want their money back) which necessitate withdrawing actual money (gold) and can cause a suspension of payment whereby the bank simply stops giving out gold at all (thereby breaking the peg). This happens in emergencies.
The C5 banks (which command most of the world economy) are the Central Banks of Switzerland, England, Japan, Europe and the Fed. No mention of China, I wonder if that’s still up to date.
Deleveraging: Banks shrink both sides of the balance sheet.
Three relevant markets, each has a different interest rate: Fed Fund market, RP (Repo) market and Eurodollar.
Fed Funds market is an interbank borrowing market, and what they’re borrowing are reserves (liabilities of the Fed). Size of 1T in 2012.
Reserves at the Fed are high-powered money. Fed Funds are interbank promises to pay money. Credit expansion on top of reserves on the Fed, they are assets and liabilities of the banks, not the Fed.
As the rate went to 0, the Fed switched tactics towards QE (expansion of the balance sheet) instead.
Two different intermediaries, sort of clearinghouses: a private one (CHIPS) and the Fed.
The Fed has an overnight requirement that all banks’ balances be non-negative. Over the day a bank may transfer to another by moving their reserve amounts. If the amount becomes negative they may develop a daylight overdraft, whereupon the Fed does still transfer the money (its own liability, true money) to the receiver.
On end of day, the bank with an overdraft must settle. They may borrow from the Fed, which will always lend at Fed Funds rate + 100bp, or more conveniently find a lender in the Fed funds market (at the effective fed funds rate). Note that the Fed’s punishment for not clearing is not the offender’s utter destruction, but just a painful interest rate on the loan, as the Fed is not in the business of making payment harder.
Analogously, members of CHIPS will net their payments over the day (unlike the FED which is real time) and transfer reserves from one bank to another -the assets in CHIPs are collateral of some form-, and if the sender were unable to clear their payment then the other CHIPS members will cover for it (and it’s presumably kicked out).
There is elasticity during the day, and contraction at end of day (the survival constraint).
In both cases, when a bank has a deficit, another has a surplus, as this can be modeled as a closed system. Then it becomes a game of finding your counterpart and borrowing, so you can clear (both in CHIPS and at the Fed). In CHIPS, the money market is the eurodollar. Not every bank can be a member of Fedwire/Fed Funds market.
The basic operation is, bank A needs to increase reserves to clear at end of day, so they borrow in the Fed Funds market. Typically an overdraft with the Fed may be small or last very little, as there is a penalty for average overdraft over the day (as another discipline factor). Bank A may borrow from the same bank B to which it needs to pay, and then pay with those same reserves it just got, ending in no net changes of reserves but now bank A owes bank B a Fed Funds to be paid next morning, with the interest.
Worth pointing out: now that there is a market, you may want to trade Fed funds independently of whether you’re trying to clear an overdraft. You may just hold and resell for a higher price like any other asset. You become a dealer in the Fed Funds market, with a spread between interest rates so that you collect the difference.
In a transaction between banks, you can have, as a facilitator, either a dealer or a broker. A broker, like a real estate one, finds two interested complementary parties and helps them meet. A dealer instead takes an intermediary role: both bank A and B can abstract from who the other party in the transaction is, and just trade with the intermediary (e.g. giving them fed funds for money and back).
When too many banks need to borrow reserves and they are scarce, interests go up. The opposite will happen if the market is flooded with money.
The Fed is neither borrowing nor lending in the Fed Fund market (!). It has a target for the Fed funds, but it influences it by manipulating the quantity of reserves in the economy as a whole. Particularly by entering the repo market and using it to increase or reduce the supply of money. This means it needs to anticipate the scarcity of money correctly.
“They change money in order to influence the rate”, using overnight loans (repurchase agreement loans) that create additional reserves.
The fed emits RP Loans, which a securities dealer may buy to obtain liquid reserves (which they then lend in the fed funds market). By doing this the fed increases the elasticity of the market, creates additional reserves and thus lowers the Fed Funds rates. The rate for Fed Funds market transactions, averaged over the day, is the Effective Fed Fund rate.
The Fed Fund loans are unsecured. This means there is no collateral: if you default, the counterpart gets nothing. This is risky, and each bank sets rates and decides how much to expose itself to each other bank so as to minimize this risk as per their perception. A bank may refuse to lend to another if it is insolvent, or may demand an intermediary broker to be involved, but a broker may also refuse trades. The repurchase agreements market, in contrast, is a secured loans market (the typical collateral being T-bills), so their loans are more trustworthy.
In the Repo market, as said above, a bank will sell securities to the fed, with a promise to repurchase them overnight adding an interest rate. The securities in effect are acting as collateral for a loan. Analogously, the money acts as collateral for the securities, if the Fed were to sell them. Both parts may also be security dealers.
The repo market actually came before the Fed or the Fed Funds market: it’s how banks would trade with each other –if you don’t know the other bank, you’ll want collateral.
Unlike the Fed Funds market, almost any agent may participate in the repo market. You don’t need to be a bank, etc.
When the Fed participates in the repo market (which used to be frowned upon: why should the Federal Reserve need collateral when it borrows?!) it may do a repo operation (loan to the dealer with securities as collateral, increasing the balance sheet) or a reverse repo (shrink the balance sheet absorbing liquidity and releasing securities).
Typically, the rates are Repo < FF < Libor/Eurodollar. This is, roughly according to Mehrling, similar to how in Bagehot’s times, there would be the market rate of interest and the BoE’s, such that all borrowers would try to borrow at the markets first, but as interest rates would go up when liquidity dried, then borrowers would flow to the Bank of England.
The riddle in this lecture (taking place around the Crisis) is: why is the Fed Funds rate now lower than the Repo market’s, contrary to historic trends and common wisdom? The answer is active involvement by the Fed: They want to bring the Repo rate down, as it is a proxy for a lack of liquidity, and so they are providing liquidity in abundance. To do this they peg the Fed Funds rate much lower than the Repo one, effectively allowing for arbitrage so that liquidity is injected into the market.
Keynes used to say the government is better at slowing down the economy than accelerating it: It is like pushing on a string.
Volcker Rule: A bank may not speculate: if they serve a derivative they must also find a buyer for its counterpart and only profit from the spread.
Fencing: Separetely from reserves corresponding to liabilities related to trading, a bank needs to have reserves that cover retail deposits.
The Eurodollar market rate (or LIBOR) is usually slightly above Fed Funds rate (typically 2bp).
The Eurodollar market is much bigger than the fed funds one, but it operates with no intervention from the Fed: It is the market for dollars in foreign banks, which need to have a corresponding deposit in a correspondent, american bank (even for branches of the same bank). In effect the ‘real’ money (deposits at the Fed) never leaves America.
This is the market for funding outside of the USA, and it trades virtually par with Fed Funds most of the time, except when pressure is exerted. Most interbank trade happens between banks with a surplus of deposits and banks with a surplus of loan capacity.
Foreign banks in the Eurodollar market don’t have reserves at the Fed. They merely hold correspondent reserves in banks in NY. That’s why the discipline factor is a lot stronger in this market. To compensate for this and limit their exposure to foreign currency, banks will try to keep a match book. Not just in amounts, but in time: they will make deposits have a term, and align these terms with loan terms.
To keep a match book and not be exposed to currency fluctuations, as well as being compliant with regulations, banks will emit liabilities or procure assets so as to cancel them out over time. E.g. if bank X knows in 3 months it will receive a 6-month deposit for N dollars, it will ensure it can provide a loan for the same amount at the same moment for the same period, and so on.
Suppose bank X has a 2-month loan and a 5-month deposit with bank Y as counterpart. A 3 month loan emerges after 2 months for it to provide. It can borrow that money at LIBOR rate.
Another possibility, always with the goal of procuring liquidity in mind, is Bank X and Bank Y make a FRA (Forward Rate Agreement) at rate F: This means Bank X effectively borrows from Y at dates t0 to t1, with interest rate (F - LIBOR).
This way, back to the previous example, X may borrow from 2-months from now to 5-months from now to fund the loan, at (F - LIBOR). There is an incentive for it to set F as low as possible, and for Y to set it as high as possible. On the day of, Y funds itself with the IBOR market, lends at F - LIBOR (hopefully positive).
Since bank X already set interests for 5-month and 2-month deposits, it follows that the rate F is already set, as the only number (1 + R(0,5))/(1 + R(0,2)) - 1 such that there is no emerging arbitrage. The same happens with foreign exchange, but scaled by the future exchange rate estimation. This is called covered interest parity.
Spot rate: The ratio between currency values (e.g. 720 for USD:ARS).
Market makers create liquidity by offering buy and sell options.
Suppose there are two businesses: one is retail (A), the other (B) is wholesale (e.g. they sell apples). In practice if A buys apples from B, it will give B a bill of exchange, stating that it will pay the corresponding amount in e.g. 90 days. B can then sell that bill at a discount of its fase value to a bank, getting the money instantly. The bank then absorbs the risk of A becoming insolvent, and loses liquidity, but gets an interest rate in exchange.
In this manner the bank is sacrificing liquidity for profit, which is not a trivial tradeoff.
Other options for the bank instead of providing notes are providing a deposit (thereby only expanding its balance sheet without sacrificing liquidity, sometimes providing an interest if it is a time deposit) or instead of buying the bill providing an acceptance: a contingent liability that pays in case A does not, but charges an interest rate in exchange (this is a precursor to credit default swaps).
In most cases the bank knows business A and is not concerned about outright default as much as insolvency and delays in payment.
The rates of discount can go up to discourage operations when the bank is running back on liquidity, generating a tightening of liquidity downstream. Once the bank has liquidity again as older bills mature, they can lower the rates to get back more operations. Ideally they want to not have any cash sitting idle (buy as many bills as possible) but also never run out of liquidity (books have to match up every night).
A bank can also rediscount at another bank, effectively running a spread. Banks are money dealers.
In the Victorian times (Bagehot’s times) the Bank of England acted as a central bank for the world economy, and did the profit-maximizing operation other banks do: use liquidity completely, buy discounted bills. But what it also did was quoting a discount rate that was above the market’s rate. This way in normal times no one would borrow from the BoE, but in times of crisis as interests rose until reaching the BoE’s, suddenly all borrowing would take place at its doorstep. This made the BoE the lender of last resort, which was not a profit-driven mission but a public duty. Crucially at this point the BoE would emit deposits, not notes, which would be bound to be treated as the same thing even though they markedly aren’t. By regulating this interest rate, the Bank could send signals about the liquidity situation for the whole economy: impose discipline or inject flexibility depending on e.g. how much gold they had left. This also required coordination with other central banks.
Starts with a small reading of the Global Financial Stability Report from the IMF.
Markets can be liquid or illiquid. In a liquid market, a given transaction does not affect prices, as there is enough to go around. Prices in a liquid market for a given good will be continuous, whereas for illiquid markets discontinuities are present.
Supermarkets are in a way liquid: in the morning shelves are full and no one is buying, after work it’s rush hour, they’re full of buyers and the shelves empty out, but prices do not go up. The key for why this is so lies on inventories. The supermarket has stocks of goods and therefore can set a single price and maintain it.
Dealers of securities and assets work like this as well. They have capital, allocated between securities and just cash reserves, and this allocation fluctuates with demand, but doesn’t necessarily change prices. They could merely profit from the spread. An example of an illiquid market where you depend on a broker instead of a dealer is real estate, where you may want to buy a house that is simply not in the market.
As a model of two-way dealers (we will use Trainer’s model from 1987), we see three factors: in the x-axis, inventory, liquidity constraints: there is a maximum long position (when you burn all your capital, or your maximum theoretical leverage) and a maximum short (when you sell all of your inventory of a security + short it to the max). On the y-axis, prices fluctuate between two constraints: value bid and value ask. The value bid is the lowest price an asset can reach, as anything below that will be scooped up instantly for arbitrage (ask Warren Buffet). The value ask is the point where an asset is so priced that it becomes obvious it deserves a short. Finally, a dealer will sell their inventory on a bid-ask spread. This spread will be wider or narrower given how volatile the market is, and how much assymetry of information the dealer perceives (arguably these are also correlated? Seem like variance and bias to me).
In a way, dealers are buying funding liquidity and selling market liquidity.
This is also where the Volcker rule comes in: it establishes that a bank must have matched book (an inventory of 0) for each security. This in theory means for every long position the bank must have an equal opposite short one, and therefore price risk must be 0. In practice, ‘there is no such thing as a perfect hedge’ and liquidity risk also emerges.
In practice, a dealer will not be dealing with its own capital (or its own capital will be a small part of the total assets) –That would be a trader. Instead, it will be leveraged. What’s more, the dealer will commonly hold not securities and cash per se, but reverse and repo: some hedge fund or similar institution will be the ultimate holder of securities, whereas a bank (very typically the dealer itself will belong to a bank) will provide funding in the form of repos.
It is key to remember the dealer is trying to only profit from the spread, and have as much balance sheet size as possible, while avoiding exposure to prices. So ideally they will try to long and short securities to cancel out price risk. To short a security, a dealer may lend to a hedge fund in the repo market (obtaining a reverse instrument) and then sell the security instantly. This is pretty awesome, a mechanism to obtain a short position in a security. The long position is straightforward.
Additionally to the spread, most dealers benefit from the breakdown of the expectations hypothesis (lend long, borrow short).
Markets have different qualities: they present different slopes in prices (steeper or plainer), and different bid-ask spreads. Since each dealer is motivated to provide liquidity to the least liquid markets as a way to increase their spread, in aggregate they end up providing liquidity to most markets.
Your ability to trade in liquid markets is, in fact, brought about by the fact that the price you are getting is different from the fundamental value.
Fischer Black said that asset prices are 90% of the time within a factor of 2 from their fundamental value. This was his efficient market hypothesis.
Trainer says dealers are concerned about their net position, as it exposes them to price risk. Their positions could be split into matched book positions and exposed ones.
Empirically in normal dealers, an order of magnitude bigger share of their position is matched book, while only the rest would be e.g. borrowing to long bonds (thus incurring price risk), therefore speculative book.
The dealer may be borrowing short and lending long (using securities as collateral in the repo/reverse market). This exposes them to liquidity risk, as they are functionally operating as a bank. They reborrow every day from the repo market, but if the bank they are borrowing from becomes illiquid, they would face liquidity issues as well.
The dealer chooses to expose itself to liquidity risk by borrowing in the repo overnight market and lending in the term market, with a corresponding (positive) spread in yields.
That spread in yields presents a curve, not unlike that in Trainer’s model of dealers but flipped (so it goes up as X increases) where Y axis is yields (with a spread between overnight and term) and X axis is liquidity exposure. The dealer may expose to more liquidity risk by increasing both sides of the balance sheet in the corresponding repo/reverse operations, or less by shrinking it. A negative position would be holding cash? Note that this curve mostly describes the term rate, as the overnight rate will be strongly affected by the Fed intervention.
The dealer is dealing in securities and dealing in money. This same story, by the way, would apply for banks: you can think as an overnight deposit versus a term loan to a business. In this case the bank is a money dealer instead of a securities dealer.
One thing the Fed does is create an outside spread for the Fed Funds market. Another is to Keep the Fed Funds rates stable.
After 2008, the Fed became very explicit about the Fed Funds rate outside spread: it set a minimum (by paying .25% interest, the interest on excess reserves -IOER-) and a maximum (by lending at .75%, the discount rate). At this point, the effective rate can only move between these values. Before the crisis these values were 0 and 1%+target.
Temporary open market operation: The Fed can lower interest rates in the Fed Funds market by expanding its balance sheet and dumping them in the repo market (for a temporary period of time). They are temporary as opposed to normal (permanent) OMO where the Fed buys the treasuries. The Fed buys the repos from a dealer through a bank, injecting liquidity into the markets. The dealer can then use those reserves for e.g. paying a loan, investing, keep them as deposits, etc.
Note: for a bank, the payments part of the system may be a loser, and their profit making part would be the lending in the money market. They are able to keep par in the deposits/payment system thanks to the money market, but operating in the money markets provides them with the profits. As always, a trader / dealer could be thought of as a dealer of securities or one of money, depending on how it’s looked at.
What monetary theory cares about are the transmission mechanisms between what tools the Fed / Central Banks have (interest rate setting, monetary emission) and the market prices of money. These can be understood either as the supply of money (for any definition of supply and money, picturing the expanding pyramid of liquidity) or the interest rates of money (picturing the curve of yields, that starts low with the overnight rate of the Fed and starts growing as the terms lengthen).
input equation here
This equation describes and fits the interest rates set by central banks outside of anomalous (crisis) times. It was developed mostly as a descriptive tool, but it became normative in the ’80s. The last two terms are sort of ‘error correction’ terms, in that they represent the Fed’s targets.
The first part is the Fischer Effect: the idea that the nominal interest rate is the real interest rate (ro) plus expected inflation (pie). The markets adjust interest rates to match their expectations of inflation.
The other parts are the way the Fed has to correct against what the market will do on its own. To lean against the wind so to speak. These two terms, with weight coefficients, are one corresponding to how much higher inflation is than the target, and one to how much lower output is than full employment.
This is what is called inflation targetting.
At this point, we’ve seen three models.
First a dealer in the bond market, using Trainer’s model. The dealer sets prices based on the size of their inventory, exposing itself to higher or lower price risk.
Then, a lender (in the repo market or others), who in practice is a dealer in money. This lender will set their yield/interest rate higher the bigger their exposure to liquidity risk is (that is, to how expanded their sheet is), so the curve slopes to the other side.
Finally, the Fed Funds market, where interest rates are bounded (by the Fed’s intervention) and yet fluctuate, increasing as banks expose themselves to higher settlement risk.
The Fed is in practice only setting a target Fed Funds rate. It may also trade in temporary OMO. This fixes a certain settlement risk, which is the discipline in the market.
The transmision mechanism goes from the Fed’s target, which dictates the overnight Fed Funds rates, to the term rates for dealers (in e.g. the repo market) to finally bond yields and loan interest rates in the private sector. So understanding this, one can see how e.g. raising interest rates will reduce lending and funding in the private sector.
Before the crisis, Citibank had Citi SIVs (Investment Vehicles) which would own RMBSs (mortgage-backed securities) as assets and finance their operations through borrowing in the commercial paper market (so liabilities: ABCP, with these RMBSs as collateral). On the other side of operations would be e.g. MMMFs that have these ABCP notes as assets and their own shares as liabilities.
As trust in RMBSs went down, MMMFs wanted to be rid of them and did not roll their commitment to the SIVs after term. Citibank misjudged, and decided to keep the RMBSs in the SIV, and loan the money to the SIVs themselves (!) while financing those operations in the repo market first. The buyers of these repo would be e.g. the same MMMFs. Note that these are more secured: they are secured loans backed by the bank itself. Eventually as trust kept going down, the securities used as collateral (the same RMBSs) did not look safe enough and MMMFs would prefer safer securities as assets, so that Citibank stopped borrowing from the RP and had to start obtaining funding from the bond market (emitting commercial paper), bringing the entire bank as the support for their loans but in so doing putting the RMBSs completely into their balance sheets.
Finally, when the MMMFs did not want to keep holding the commercial paper and sold it, liquidity was needed in the eurodollar market (these MMMFs were in big part foreign) and so action from the Fed and the Treasury were needed. In the end the MMMFs ended up holding T-bills, whose emission the Fed funded with an expansion of its balance sheet. The reserves from that went to the ECB as liquidity swaps, and finally the ECB loaned that money to e.g. Citibank. At this point the Fed had brought the RMBSs funding into its own balance sheet, while later it would start having to buy RMBSs themselves in the wholesale market.
End of part 1
There is a hierarchy of currencies, implied by trading volumes. USD is the world currency, then Yen, Pound and Euro make the big 4, and the CAD, AUD and Swiss Franc make the most traded ones. Other currencies trade not between each other but by switching to USD and back.
Their markets are liquid, you can do huge trades without moving the price.
It is not surprising that there were two traditions in monetary theory, given that there were always two kinds of money. These are fiat currency and gold, one corresponding to Chartalism, the other to Metallism. These represent respectively retail and wholesale trading, public and private spheres of influence. Historically there has always been an exchange rate between these: you may use domestic money for day to day living and paying taxes, and international one for trade, banking, and as a store of value.
The distinction is the King’s money versus international money. A King would have to move back and forth between the two, particularly notoriously in a foreign war.
It can be argued that the current system is a hybrid of the two, both institutionally and historically.
The idea that the state desires elasticity while the private sector provides discipline lets you to think of the relationship between the treasury and the central bank as one where the treasury issues debt at an interest rate, and the government’s bank (the central bank) buys it and emits currency as its non-interest-bearing liability. This is why we see the state funding itself through the central bank and why in most developed nations it is the central bank that buys most treasury bills.
In the quantity of money theory (MV = PT), where T is aggregate transactions, T has to do with the real economy, and P is more about price levels.
Another model says that the rate of exchange is (1 + R)/(1 + R*) where R and R* are the interest rates. This is a chartalist model, whereas the metallist model would say money has value merely proportional to its metal content (the “mint par”: how much gold do I get for this?), give or take some noise from the actual transaction costs of doing so (a different price would otherwise allow for arbitrage).
In the foreign exchange market, when country A wants to pay country B (and we want to allow for credit, so payment is possible with no liquidity -otherwise payment is trivial-), country A pays X*E to a forex dealer, who pays X/E’ to the country (typically in the form of its own liability in turn, not cash unless requested which would pose severe liquidity constraints). Then it takes matching, opposite positions in the term market to offset exchange rate risk. He is this way still exposed to a different (as of yet undisclosed in class) risk, but at least avoids problems due to exchange rate fluctuations. On the other side of this, there will be speculators lending or borrowing at term to allow for this transaction. This would incentivize countries to move their interest rates to allow liquidity to hit the market (obvious examples come to mind).
What is the exchange rate?: Relative price of goods (purchase power parity), relative price of assets (forward interest parity), relative price of money (in terms of money), which takes into account liquidity.
In the past (think early modern) the government would have a government bank, and use it for funding (especially for war finance). They would print bonds or treasury bills and the central bank would buy them and keep them as assets, while printing money as a liability. In parallel, the private world would have banker’s banks, that would emit money backed by gold as its asset. There was credit and elasticity albeit much less than now.
What happened in the late modern period (around the time of the American Civil War) was the government took the gold from the private banks and gave them fiat money deposits instead. But then they gave those deposits their blessing (which only government bank money used to have) in exchange for keeping the gold. This made bonds much more desirable as an asset and allowed for much bigger opportunities for government (and eventually military) funding. This is the hybridity of banking. Modern monetary systems are always hybrid, one way or another.
Before WWI, there was a strong discipline factor due to the gold standard. Much as central banks and states would like to print infinite money, they needed to maintain their pegs to gold.
This brought integration and stability to the international monetary system. Failures were not infrequent, but they were correlated. It is possible to remember it fondly compared to the chaos that followed.
As said before, central banks wish to avoid discipline. To do this, there were several proposals, but they required coordination (e.g. if all of us print money at the same time then gold just gains value). After the war, the dollar standard emerged (with everyone pegging to the dollar), but this had its own problems as the dollar suffers its own stabilities due to domestic issues.
Reading: R. A. Mundell’s article (15 pages, looks good🌱).
Confrontation of the Fed and the Gold Standard: After WWI many countries (all the Europeans) go off the standard, as all the gold was flowing to the US to pay for the war. After this several of the main powers (UK, FR, DE) try to go back to the gold standard, but this leads to several issues like deflation. The US Fed makes a mistake by raising interest rates, so all the gold flows back to them and European powers need to drop off the standard again. At this point (in the Great Depression) the most senior founders of the Fed had died, and the chairmen at that point were inexperienced in comparison and did a conservative measure, which backfired. (Mehrling mentions a biography of Benjamin Strong, which he says is really good 🌱).
According to the Fed’s website, they raised interests at that point to prevnt speculation in the stock markets (it obviously backfired) and also failed to be lenders of last resort for collapsing banks facing liquidity issues.
Confrontation of National Macroeconomic Management and Fixed Exchange Rates: After the war, in Bretton Woods, Keynes (who was British) proposed the Bancor plan: make an international bank that lends to deficit countries and gets deposits from surplus ones. But it would give out negative interest rates to the lenders (!) and force them to lend, so that there was an incentive for them to consume. This did not go over well with the clear surplus country (the USA).
Instead of Bancor, what we got was the IMF. The IMF works under the principle of discipline (some could say too much of it) instead of elasticity: the balance sheets do not expand, rather there is a fixed supply of reserves, but instead of gold these are reserves of paper money, so that they can be expanded arbitrarily but not randomly.
Each member country deposits gold and their own currency in the IMF, obtaining (through a complicated mechanism of conversions) SDR -Special Drawing Rights-. These SDR can be used by deficit countries to pay surplus ones. However if SDRs reach 0, they must then borrow from the IMF (typically at a high interest rate).
Bretton Woods: In the Bretton Woods system, after WWII, there were international reserves (in the form of gold + SDRs), the USD was pegged to them (35$/oz) and major currencies were pegged to the USD. USA provided funding for the European reconstruction using dollars, not reserves, which were accepted as money. In a way the new global reserves became USD. As credit kept expanding, gold became undervalued (that is, reserves were ever smaller with regards to USD in circulation) to the point where there was a run on the dollar in 1971, after which the convertibility to gold ceased.
In a way, one could model the relationship between USA and Europe/the rest of the world as that of a bank and a depositor: they would sell their bonds to the US, which would pay in dollars. Then countries would get (non-interest-bearing) dollars in exchange for interest-bearing bonds. In effect lending long and borrowing short.
1972-1999: Flexible exchange: At this point the central banks began targeting inflation. Not being pegged to anything, purely fiat currencies tended to produce high inflation, and this is where monetary policy came along. Central banks and speculators, without being completely integrated into an international system, bred volatility.
Wars happen when people can’t figure out what to do next.
Imagine again the world of Bagehot, where businesses are using bills of exchange to trade. But now picture Businesses A and B such that neither is in England, even though both conduct their banking in London. Even if you’re not in England, you can take a bill to the City of London and discount it.
Now business in London must be conducted in the pound sterling, which at this point is the international currency. However both firms need their local currencies to pay their workers, and also are paid in their local currencies by their respective markets. In order to transact in London they will go to a foreign exchange dealer.
Dealer model for foreign exchange: We can use Trainer’s model for exchange rates. The exchange rate will be the ratio between the mint pars, give or take a delta that make an enclosing range of prices. E.g. if USD can be exchanged for 1 oz of gold and GBP for 1.2 oz, then the ideal exchange rate would be 1.2 USD per GBP, but in real life there is a non-negligible transaction cost to selling and moving gold, so that there is noise around that value.
The gold points act as an outside spread, and put a cap on how much risk the dealer may take. If they are long a currency and that currency depreciates below the exchange ratio - delta, the dealer may always just exchange back into gold and use that to potentially acquire more of the international currency (GBP).
Suppose now that the price of the currency reaches below the minimum point in the outside spread. At this point the dealer takes the USD to the ‘central bank’ for gold. The central bank can choose between paying, in which case its gold reserves go down but the money supply shrinks (literally contractionary policy: the money that entered the bank was already its own liability, so they cancel out and the balance sheet shrinks). Or it may sell assets (especially bonds) for gold. Selling t-bills for reserves can however put pressure on their price, effectively raising yields. The third and last option is to borrow reserves.
The central bank will protect the currency value mostly not by keeping big reserves, but by using these mechanisms to keep reserves stable.
There cannot be liquidity without dealers, and there won’t be dealers without a price curve that generates profits. This is the essence of monetary economics.
When there is a liquidity crisis in the central bank whose liability is the international currency (in practice, the BoE), it has three options: it may acquire more gold (typically impractical), raise interest rates (at this point in time, the 90d term interest discount for exchange bills) to bring discipline, or suspend payments. Even after suspending payment in species, currency prices would not collapse, which is interesting.
There are three different theories as to what predicts foreign exchange rates. Covered Interest Parity (the ratio of interest rates equation), Uncovered Interest Parity (F = E(S)) and Expectations Hypothesis of the Term Structure (the interest rate for a term T should be the product of interests for shorter disjunct terms that form it).
Empirically only the first one holds up, and this lecture explains why.
We see again the Trainer model applied to the dealers in FX under the gold standard (and remember that as a lender of last resort, the BoE would adjust its term discount rate for bills of exchange, and had a last resort of suspending payment in specie).
We go back to the dealer in FX who facilitates payments between two nations (A to B) that don’t have the international currency by allowing A to pay in local currency to the trader, the trader pays in international to B, then makes two opposite trades to hedge against exchange risk by buying forward trades (effectively borrowing the local currency and lending the international one), while making interests R and R* on them, plus some spread that makes it profitable. This is not a perfect hedge, but it’s profitable enough given the risk, from the perspective of the dealer.
On the other end of this, there will be a speculator: a different dealer taking the opposite positions on the forward trades, by borrowing and lending the opposite currencies at these same rates.
To absorb the price risk from this dealer, a speculative dealer will lock in the forward position, provided that it expects a profit. This means the speculative dealer will have a value E(S) of expected spot rate for the currency at term, and buy future positions at a lower value than that -presumably, assuming R is positive, the s.d. assumes the future value is lower than current?-. Assuming they are right, they can profit from that difference.
As the matched book dealer takes on more liquidity risk by being more exposed to fluctuations in the national currency, he will raise the interest rate he expects from lending to term. F/S = (1 + R*)/(1 + R), where R is basically a constant, so that the other country will end up adjusting R* to increase or decrease the elasticity of the FX market. This means as the dealer is raising his exposure and how much of the local currency he has, he bids up the term rates in the deficit country. He wants to buy spot more cheaply than he sells forward.
In a world without liquidity premium, you expect the term rates are the expectations of the roll over of the overnight rates. This is the expectations of the terms hypothesis.
“China is a surplus country. If everyone needs to make a payment to you, your liabilities are money. It’s no longer true that everyone needs to make payments to the USA (as it was after WWI or WWII), but the eurodollar market is already formed, without having much to do with the balance of trade.”
The survival constraint is a real thing, and this is how you solve it. You cannot abstract liquidity away.
“A shadow bank is money market funding of capital market lending”. Shadow banks fund themselves in the Repo market or by emitting ABCP, and use this to buy securities.
The prices the shadow bank is dealing with on both sides of the balance sheet are market prices.
Historically, money markets and capital markets were treated like separate things, to the point where they are taught about by different departments. This is no longer quite true, as shadow banking now plays an enormous role in both.
In Bagehot’s day, the BoE was principally focused on discounting bills of exchange, which are relatively short term, and didn’t involve itself with bonds and other long term lending, which worked on a separate market. The American banks were much more prone to take part in loans and bond trading, as they also used bonds in a proto-repo market, especially before the creation of the Fed, in order to settle interbank payments.
In order to do repo payments, banks would know how safe a bond was by using rating agencies, like Moody’s. These were here before the Fed.
In 1929 the Fed allowed the shadow banking system to collapse, without supporting mortgage credit or bonds, as they were deemed ‘improper’ for banks. In contrast, the Fed did backstop the shadow banking system in 2008.
The liquidity concept in the UK was formed around having reserve to liquidate payments on time while discounting bills of exchange: short-term lending. In contrast, in America liquidity was more about market liquidity, which at the time was explained as ‘shiftability’ of bonds. That is why in the 1920s it was argued that the Fed had to step in and provide liquidity buy performing OMO.
[Some recommended reading 🌱 goes here.]
Schumpeter said that in order for a country to develop it is necessary for there to be development banks, which borrow short term in the money markets and lend long in the capital ones. They fund development. He also insisted on the importance of monetary expansion ex nihilo.
Direct vs Indirect finance: in indirect finance, the bank lends a loan to a business. To fund this loan, it will use deposits, ideally from the final recipient of the money it lent, but otherwise from money markets and so on. In direct finance, the business will sell bonds to lenders, side-stepping the bank altogether.
Indirect Financing:
| Business | Bank | Society | ||
|---|---|---|---|---|
| + deposits | + loan | + loan | + deposits | |
| - deposits | - machine | |||
| + machine | + deposits |
Direct Financing
| Business | Society | |
|---|---|---|
| + funds | + bonds | - funds |
| + bonds | ||
| - funds | - machine | |
| + machine | + funds |
Recommended Reading: Gurley & Shaw: Money in a Theory of Finance 🌱.
Moulton said, and Gurney & Shaw agree, that banks are not only intermediaries in the money market, but in the capital market: long-term lending, which is necessary for development.
Two important intermediaries: pension funds and insurance companies.
Pension Funds give out fixed incomes as a liability (typically a percent of salary before retirement), and hold mostly equity as assets. Insurance companies usually hold bonds and give out contingent payments -insurance policy- as a liability. In this way, both are intermediaries beetween corporations and households.
The liquidity of insurance companies depends on the liquidity of bond markets.
Indirect finance ‘solves development’ because it stands between households and companies and gives each one what they want. Mutual funds remove the intermediary by giving households ‘shares’ directly into a portfolio of equity/bonds, passing the risk through to the holders. The rise of finance is the replacement of indirect finance with direct finance.
The government is in the middle of banks providing stores of money for households, and funding for businesses and even the Fed. In that position it acts as a securer of liquidity (with the Fed backstop) and a securer of solvency (FDIC). In exchange for the taxpayers’ intervention, this created regulation. Because of this an opportunity for arbitrage emerged, and this created the shadow banking system.
Financial evolution: indirect finance to direct finance. Banking evolution: loan-based credit to market-based credit.
The shadow banking system was created with no lender of last resort. The thought being that if you get rid of solvency risk (using an insurance company that sells swaps) then this is a risk-free, liquid asset. That turns out not to be true: market liquidity is created by dealers, and even if your asset is risk-free you still need someone who wants to buy it. When assets lost their shiftability, the whole thing came down.
Two prices made this market: international dollars for funding, and RMBSs for holding assets. Once risk became apparent and dealers stopped making the RMBS market, the prices were no longer well-defined, lending with them as collateral became untenable, and then the liquidity crunch set in.
Class action clause: a term in a
Advance Clearing: ways in which changes in expectations into the future (and thus prices in markets) cause cash flows today and may trigger liquidity constraints.
These may be positive cash flows and you’re freer, or negative ones, and you’re more constrained, and may be forced to sell something.
Survival constraints not only change asset prices (stocks, bonds) but also money market interest rates.
A future can be modelled as a) a swap of IOUs b) between a firm and a bank where c) the bank receives a 3-month deposit and a 6-month loan (with the bank receiving the correspondent counterparts). This is effectively equivalent to a 3-monht loan, 3 months from now. However, the interest rate is locked-in. This would be called f(3,6). Given that the bank has defined the interest rates R(0,3) and R(0,6) for 3- and 6-month periods, R(3,6) is implied as the ratio between the two.
In practice these are forward contracts, listed as forward loans or deposits.
The bank could hedge this by finding a firm B that does the reverse agreement, so that: A deposits for 3 months in bank, bank deposits for 3 months in B. B gives a loan for 6 months to bank, bank gives a loan for 6 months to A.
This is match-book assuming everyone pays, and there’s only a bid-ask spread for profit. It’s a way for a firm to secure future funding if they fear liquidity crunches.
In this system, there is the client who wants a forward -e.g. F(3,6)-, the bank who does the loan (FRA), the speculators at the futures exchange who sell futures -the hedge- and finally there will be a real spot rate -R(3,6)- in 3 months.
The forward rate is not an unbiased estimator for the future spot rate.
This is a negation of the ‘expectations hypothesis of the term structure’, which is not empirically true. The forward rate is always greater than the expected spot rate, as there is the extra cost added by the hedging -to serve as an incentive for speculators to take the opposite position-. This predicts a difference between the forward rate, the futures rate -in the middle- and the spot rate (always the lowest).
A forward agreement causes no cash flow except at maturity. Futures agreements cause cash flows day to day.
In general, the bank always wins. They will set a forward rate that is higher than the expected spot rate. Whoever is on the other side is paying a premium as a way to lower risks.
We could see that over time periods i, 1/(1+ Ri(a, b)) = 1 + R(i, a)/(1 + R(i, b) where R(k,k) = 0.
In the end, as the forward rate is always ft=St - Ke-r(T-t), at the maturity point (the only moment where actual cash flow takes place) the short pays the long ST - K (K == S0erT).
In forwards, the only cash that changes hands in at the end. You subtract the interest rates. In futures, you move from the losing to the winning side everyday, by comparing the futures’ rate to the K. If the futures rate goes up it’s a win for the long, and the reverse if it goes down. In practice this means earnings are proportional, every day (with daily cash flows), to the derivative! There is also an extra liquidity risk to a future that is not there in a forward.
One way to produce knowledge: find something that doesn’t seem right and worry about it for a long time.
Cash and carry arbitrage, which makes futures contracts profitable over a forward contract, is possible only if the ‘implied repo rate’ ρ, is higher than the overnight repo rate r. This way Ft = Ster(T-t) is profitable.
In a forward/futures contract, you pay F now, to get the asset that’s currently worth S0 at time T. This is why you pay an extra interest for the ‘carry cost’, while being bullish that the underlying asset’s value will rise higher than the risk free interest rate (you’re effectively taking on risk).
You take a 6-month bond by shorting a 3-month term repo position (you borrow S0 to pay the bond). At month 3, you sell a future for this bond at 3 months for price F. You’re hoping the bond’s F at month 3 is bigger than the S0 price minus the interest you paid on the repo market over the first 3 months. F - S0 - r3.
The banks are using the futures to hedge forward positions. So they in general have long forward positions that they’re looking to hedge. Price fluctuations are equal and opposite to some fluctuation in the forward market.
A bank has a comparative advantage in taking on liquidity risk.
In 1952 the Fed got th right to set short term interest rate targets. Before that, short term lending was at 1% yields, long term at 2% and the Fed would buy bonds trading below that until yields would go back to intended levels. In effect it was the lender of first resort.
During the war, the Fed was pegging the yield curve for all terms. By 1952 dealers were prominent and they were doing arbitrage, so the Fed could intervene less, only influencing the ‘edges’ of the market, provided that it remained liquid thanks to the private, profit-driven dealers.
It understands dealers fund their holdings in long term bonds by borrowing in the money markets.
They emphasize that OMOs are not equivalent to changing the interest rates. They talk of the tone of the market. They talk of moving the rates for treasury bills (where term < 1 year) and how that will directly move the term rates thanks to the yield curve and the expectations hypothesis of the term structure.
The swap yield curve has historically been a spread over the treasuries’ curve. This could be because companies are less creditworthy than the government or due to liquidity risk. During the 2008 financial crisis and few years later, this fact of the world reversed, with a negative spread, allowing for the possibility of arbitrage.
In a swap operation, a firm BBB that wants to lock in a safe rate of interest even if it means expected loss can trade with an (e.g.) AA one that has access to fixed interest long term borrowing.
AA will borrow fixed, long-term at a rate by issuing bonds, and lend to BBB at a fixed rate called the swap rate. These are quoted in the FT and such. BBB will lend, notionally, at the rate of LIBOR or some other such money market to AA. This way it will only pay the spread on the interest rates. Usually the money market moving rate used is renewed over a short period like 6 months, while the bond is a 10-year one or similar.
| AA | BBB | ||
|---|---|---|---|
| fixed rate borrowing bond | flexible rate borrowing (MM) | ||
| [ Fixed rate | LIBOR ] | [ LIBOR | fixed rate ] |
In real life there is not actually a parallel loan. They are just netting payments. No balance sheet is expanded. There is no principal either.
AA is called the seller of swap, short the swap, the payor of floating.
BBB is called Buyer of the swap, long the swap, paying fixed.
Interestingly, if you’re long a swap you are long LIBOR rates, and viceversa. You are hedging against future liquidity risk / discipline.
Suppose AA can borrow from the bank at a fixed rate x with 10-year bonds, and BBB at rate LIBOR + y on (a shorter) term. AA could borrow at LIBOR + y’ (< y) on the same terms.
What will happen is AA could borrow from the bank at rate x, lend to BBB at (effectively) x + d - LIBOR, and this way BBB is effectively borrowing at LIBOR + y and paying x + d - LIBOR, so in effect it’s paying a total interest of x + d + y (still hopefully below what it could secure for long term borrowing). AA is paying LIBOR + x and making x + d, so in effect it gets funding at (LIBOR - d) rate, and absorbs the risk that BB may default.
This is not exactly credit risk, as there is no principal and all that will be lost is the interest payments on the terms.
In reality what happens is BBB borrows from the bank at rate LIBOR + y’ and just pays the principal and rate to the bank, receives that LIBOR part from AA and pays x + d in exchange.
This is the derivatives market: the loans have vanished, there is only spreads and interests here.
This apparatus was developed to create liquid markets in corporate securities. They then moved it into mortgage space with RMBSs, creating a global market.
The place where prices are made is the dealer markets in derivatives. When a bank is offering you a mortgage, they quote the interest rate by looking at the SWAP yield curve.
If the curve for swaps is below the curve for treasury yields (which is an anomalous situation as the state is supposed to represent less credit risk than the private sector), the way you would expect arbitrage to take place is
| Arbitrager | |
|---|---|
| 30-y Treasury Bond | Repo |
| Repo | fixed 30-y loan |
The line below is shorting a swap.
Shadow Banking: money market funding of capital market lending.
None of the regulations, according to Mehrling, are right for this.
Investing in CD Swaps exposes one to private debt that is high risk and high yield.
In CD Swaps, we peel off some of the credit risk from a corporate loan to a not too credit-worthy company (BB, CCC, etc.) and sell it separately.
Unlike Interest Rate Swaps where we take away part of the interest rate risk (lecture 19), in CD swaps we hedge against part of the credit risk.
Riskless security = risky security + interest insurance (hedge) + credit hedge (guarantee)
Here the interest insurance is a interest rate swap, and the credit hedge is the Credit Default Swap. This was predicted by Fischer Black in the ’70s and became true.
There is a buyer of insurance who is long a swap or short credit risk, with their counterpart a seller of insurance, short a swap, long credit risk.
| Long a swap: | |
| Corporate Bond | |
| [ Treasury Bond | Corporate bond ] |
| [ Treasury bill | Treasury bond ] |
The first instrument is a CDS, the second one an IR Swap.
A bond pays a fixed-amount coupon every e.g. 6 months, and at termination pays the principal.
The price of a bond with coupons Ct and final payoff (face value) FT is
P(0) = ∑ ∂tCt + ∂TFT
Where T is the last term and each successive t is another term, with ∂ the discount rate (1/1 + R for R > risk-free interest rate as this is private lending -e.g. the swap rate-).
Fluctuations in the bond prices sort of come down to fluctuations in the risky interest rate, which is constantly being estimated by the market and has spreads that grow bigger as a function of time and also risk grade.
The price of the bond fluctuates every time the risk-free interest rate fluctuates. In the same way the price of the CDS also fluctuates, inversely, on it.
Since CDSs are derivatives and part of a market, you can buy them and sell them to make money, regardless of whether you actually want them as insurance.
Some may say it is not insurance at all, due to how correlated all risks are (whereas in real insurance risks are supposed tobe independent, uncorrelated and therefore actuarial).
Given a corporate bond you want to ‘insure’, your balance sheet will look like
(Buying a CDS) | Bond | | [Libor | Fixed ] | [Libor | Libor + u ]
This means on every term where the bond doesn’t default, you will pay F*u where F is the bond’s face value. However, in case the bond defaults on period N, then you will give it up and receive a treasury bond instead.
In the end you are holding a bond and funding its purchase at LIBOR + s + u where s is the spread on the interest rate swap market, and u the interest premium on the CDS, which takes credit risk away.
That is the position of a buyer of insurance, long the swap. The seller of insurance will sell CDS to several bond holders, then aggregate them into a diversified portfolio, a CDX (an index) and sell those to people who want to be exposed to this risk, allowing the dealer to hedge.
Once you have a dealer in the middle, the order flow will push these prices around.
This pushes the CDS prices around, not necessarily the bond prices until someone does an arbitrage, which creates opportunity for it sometimes.
AIG the seller of CDSs was providing them for i.a. Goldman Sachs. GS added a clause whereby if they went down in rating below AAA, they would have to foot up the amount of the CDSs, in liquid form. That is what happened when the MBSs went down, against AIG’s predictions, forcing them to pay them even though no defaults had taken place. Ultimately the state stepped in to bail AIG, but couldn’t take back the money from GS as it was just a hedge against an offsetting liability.
What caused the problem wasn’t defaults, it was their depreciation, even before.
This lecture wraps up and combines the knowledge from the previous 4~5 lectures into one with policy suggestions and future projections.
It contrasts the shadow banking system (or market based banking system) with the traditional one, and shows that the ‘deposits’ in MMMFs tend to be held by institutional investors, like pension funds and insurance companies.
This makes the shadow banking system wholesale as opposed to retail, where a small amount may be in the 8 digits.
The shadow banking system presents market-based credit. However it has no direct government backstop (no Fed, no FDIC intervention).
Shadow banks would use their own primitive backstops for liquidity crunches in the form of liquidity puts: the backing bank would say they would act as buyers of last resort if e.g. repo agreements were not rolled over. Liquidity put is a libaility of the backing bank and an asset of the MMMF or dealer. A large private entity would act as backstop.
The private backstop only works if the parent is healthy enough to backstop the crisis, or else the government will need to step in as it is now on the hook.
An example shadow bank operation would be funding yourself in the money market (e.g. repo) and buying hi-tranch RMBSs plus CDSs.
There was a mature global funding/financing system (eurodollar) but not a mature way to transfer risk: CDSs were being invented on the fly. As transferring risk was not so easy, when the dollars ran out the crisis of the shadow banking system became a world crisis as it affected the global funding system.
Modern Asset Management is a result of financial securitization and subsequent globalization, and the emergence of capital markets and money/derivative markets.
A shadow bank would fund itself in the money markets, and hold MBSs, ‘stripped away’ of their risk through derivatives (CDS, IR swaps, etc.).
Fischer Black: One person funds the loan, another bears the credit risk, another the price risk.
This way the RMBSs are treated as ‘risk-free’ bonds, like treasuries with better interest, and traded globally.
Shadow Banking: Money market funding of Capital market lending. It can be on the balance sheet of a bank, an SIV or any other firm. You have global funding of local lending.
The Shadow Banking system is upheld by the market pricing of money (very different from traditional banking) and capital, and it puts at the center market making institutions (to understand situations one needs to focus on dealers and such instead of banks), which have a key role.
The reason this exists is not regulatory arbitrage and fraud. Get rid of the fraud, you won’t be rid of shadow banking.
We have a mature money market funding system, but a very immature risk transfer system. Dealers (in derivatives) are key. They are the ones who make market liquidity and set prices.
Mehrling proposes something akin to a clearinghouse for derivatives dealers as a backstop for the global market, but insists policy ideas are still not fully-baked.
We’re living in a Bagehot moment were we have to rethink the role of the central bank.
The balance sheet of the Fed in late 2011 can be seen as them holding t-bills with currency/deposit liabilities, then t-bonds with t-bill as liability (equivalent to an IR swap) and finally risky bonds as asset for t-bonds as liabilities (where the risk exposure is equivalent to CDS).
Several problems emerge from the shadow banking system. One big one happens if a fund has bought certain assets financed with the global money market dealers (and hedged with a derivatives dealer). Were the assets to temporarily lose value in a market fluctuation, the institutional investors depositing in the money market dealers may want out, causing a liquidity crunch. Alternately the assets may actually, permanently depreciate, necessitating payments all the way up the chain of borrowers. This also requires liquidity, which without a proper backstop may not be available. All it takes is for a single payer to refuse or even delay payment for the entire chain to break and the system to be halted in a deadlock.
A possible solution is for the Fed to provide a liquidity backstop for dealers (both MM and DD) who are the key players in the markets. However, this is not highly desirable as the Fed should be a last resort lender, its intervention must not be a common accurrence. An alternative is to regulate dealers to guarantee capital or reserves, guaranteeing the presence of liquidity. Liquidity for matched book dealers, capital for speculative ones.
Market functioning rests not only in liquidity flow but in collateral flow.
You’re going to need capital requirements for dealers to backstop bad positions.
Money view: the present determines the present. All flows downstream of banks trying to make payments with liquidity which they may borrow from the future, at fluctuating interest rates (elasticity from cash inflows, discipline from outflows).
Economics view: the past determines the present. Y = F(K, L) (parsimony and profligacy, think Adam Smith)
Economics view is more appealing to most and more mainstream.
Finance view: ‘the future determines the present’, or rather expectations about the future. P(K) = sum(expected values of coupons in the future).
The biggest intellectual battle in the last thirty years is Economics view vs finance view. Mehrling’s mission is to make the money view dialogue with the finance view, as the 19th century was a dialogue between the money view and the economics view.
MV = poqo + paqa = PQ
Banking view: PQ necessitates a certain M, V pair. Economics view: MV predicts PQ.
There is Fischer’s view: MV = PQ + (1/1+R)F.
In Fischer’s model, there is a universe of possible portfolios, which has a pareto frontier of R (expected return) and σ2 (risk). At risk=0, you have the risk-free investment (think t-bills) and there is a linear interpolation between a maximally risk-averse investor (all i on t-bonds) and a risk-tolerant one (who may go 100% market or even 150% by borrowing).
There is then a balance sheet-based model of this world where two investors, risk-tolerant and risk-averse, with the same NW, one holds 150 market-portfolio and 50 in loans (liability) and the other has 50 on market and 50 on a risk-free deposit, where the bank is an intermediary and gets the deposit from r-a and lends it to r-t.
On a price increase, in order for markets to clear, money needs to be endogenous. This is because if asset prices rise, but both the risk tolerant and risk averse agents want to maintain their balance of portfolios, the bank in the middle will need to expand their balance sheet. Equilibrium in the asset market requires that the money supply is passive.
\1) Market liquidity (buying and selling without moving a price in volume, fast) depends on the dealer system. It’s not free, and every high-volume trade will move prices away from their fundamentals.
\2) Dealer abilities to provide market liquidity depend on their own capital, their own funding liquidity. When dealers lose capital, prices can stay away from their fundamental.
\3) The ultimate source of funding liquidity is the central bank. [This is what Fischer would’ve opposed most strongly].
Course complete! 🎉🎉🎉