Friday, January 29, 2016

A monetary policy sound check

It's healthy to ask others for a sound check every now and then. I'm going to give a short description of how I see the monetary policy transmission process working, then readers can tell me how far off I am. Hopefully this sound check will bring some more rigour to my thought process.

Briefly, the story from start to end it goes like this...

1. A central bank reduces interest rates.

2. After a delay, consumer prices will be higher than they would have been without the rate cut.

Here's some more detail on how I get from 1 to 2.

A) In the first moment after the rate cut, banks find themselves earning a smaller return on balances held at the central bank than on competing short term/safe financial assets (like government bills and commercial paper). Central bank balances are overpriced, government bills and commercial paper are underpriced.

B) To maximize their profits, banks all try to sell their overpriced balances, driving the prices of government bills and commercial paper up and their expected returns down. The relative mispricing has been fixed; returns on central bank balances are once again equal to returns on other short-term/safe financial assets. What about other financial assets?

C) In the next moment the reaction spreads to the rest of the financial universe. Financial market participants (many of whom don't have an account at the central bank) observe that the returns on government bills and commercial paper in their portfolios have been reduced relative to returns on other financial assets. They try to sell their bills & paper and buy underpriced risky assets like stocks, gold, and bitcoin, driving the prices of these instruments higher and returns lower until the arbitrage window is closed.

D) Very quickly, these adjustments brings the expected returns on all financial assets into balance with each other. What about goods markets?

E) In the next moment the reaction spreads beyond financial markets. Investor begin to notice that the returns on the financial assets in their portfolios have suddenly become inferior to the return they can expect on consumer goods and services. Investors try to re-balance by selling their financial assets and buying underpriced consumer goods.

F) Unlike financial prices, goods prices may be slow to adjust. This means that the window for enjoying artificially underpriced consumer goods stays open for a period of time. With people flocking to enjoy free lunches, the quantity of consumer goods and services sold speeds up relative to the pace that would have prevailed without a rate cut. We get a boom.

G) At some point, shops increase prices and close the arbitrage window. We've now arrived at 2 and the story is complete.

You may notice that I didn't include bank lending in my sound check. That's because I'm not convinced that bank loans are vital to the monetary transmission process. That being said, we can introduce an optional step between F and G.

i) To take advantage of underpriced consumer goods, investors may take on bank debt in order to buy more goods than they might otherwise have afforded, so the quantity of debt increases.

But even if people choose not to take on additional debt, or for some reason the banks decide to hold back lending, the arbitrage process ignited by a rate cut will still play itself out with an increase in consumer prices being the final result. The key role banks play in the transmission process is at A & B, the effort to sell reserves for alternative safe assets, not at the i) level. And no matter how sick a bank is, it won't forgo arbitrage at the A & B level.

So the purpose of the Bank of Japan's recently-announced negative interest rate policy is not to make Japanese banks lend more. The point is to set off an arbitrage process out of Bank of Japan deposits and into goods & services through a series of other intervening assets, eventually leading to higher prices.


Previous posts on the transmission process:

Robin Hood Central Banking
Toying with the Monetary Transmission Mechanism

Monday, January 25, 2016

The social function of equity deposits

One more post on equity deposits.

My last post described a dirt cheap (and hypothetical) way for long-term investors to get exposure to equities. Briefly, an investor commits a certain amount of money to a one-year term deposit that promises an equity index-linked return. The manager of this equity deposit (ED) invests that money in an appropriate number of shares in the companies that make up the index, then lends these shares out to borrowers for one year at a fixed rate. At the end of the year the stocks on loan are recalled, sold, and the investor's deposit is repaid. The interest earned on stock loans is shared with the depositor, boosting their returns.

It's worth pointing out that ETFs already lend out shares, but unlike an ED they can only do so on an overnight basis. So an ETF can't harvest the extra term premium on long-term loans.

EDs have a broader social purpose than just saving a few bucks. Here's a quick list:

1. Equity deposits would reduce the dead weight loss currently being incurred by long-term investors.

The investing world currently discriminates against long-term investors by requiring them to invest in securities that are tailor-made for short-term traders. Stocks, ETFs, and mutual funds enjoy a permanent trading window--the ability to cash out of the stock market in a millisecond. Long-term investors who have precommitted to the stock market for ten or twenty years simply don't need this feature. Unfortunately, not only do they not have a choice (all securities have these windows), but they must pay the fees involved in the maintenance of said window. This is an an efficient allocation of resources, or what economists call a dead weight loss.

Equity deposits are tailor-made for the long-term investor. By removing the trading window, long-term investors no longer have to pay for a feature that caters to traders, thus lowering investors' costs and improving their returns. The world is made more efficient.

2. Borrowers of stock are missing a market. Equity deposits would fill this gap.

Anyone who wants to borrow dollars from a bank can do so overnight or on a long-term basis. It's not the same when it comes to other financial instruments. Anyone who wants to borrow stock can only do so overnight. An equity deposit provides the missing market.

The reason for this gap is that institutional owners of stock like ETFs and mutual funds face the possibility that they might be besieged at any moment by redemption requests. This means that they can only lend out stock on a short term basis to borrowers, usually overnight. Because owners of equity deposits have committed to a fixed holding period, the manager of an ED is free to lend underlying stock out on a long-term fixed rate basis. Borrowers should be willing to pay an ED manager a premium rate of interest for the certainty its fixed products provide.

Which leads into points 3, 4, and 5...

3. Equity deposits would improve market liquidity and reduce price volatility.

The job of a market maker is to facilitate trading in a security by maintaining tight spreads between the bid and ask price. Market makers need an inventory of securities to do their job, and they will often borrow to ensure that supply. Because most shares are lent out on an overnight basis, this source of liquidity is flighty. Stock can be recalled without warning and lending rates can get ratcheted up suddenly. A manager of an ED can offer market makers a guaranteed supply at a fixed price, thus reducing the uncertainty involved in market making. Hopefully this will help make for a thicker and tighter market.

4. Equity deposits would improve price discovery.

Arbitrageurs need to borrow stock to put on the short leg of their strategies. Because most equity loans can be recalled at any moment and lending rates can change daily, it can be difficult to know ahead of time the return of a certain strategy. Equity deposits provide a stable long-term supply of stock for arbitrageurs at a fixed lending rate, thus helping to ensure that the arbitrage process keeps prices in line.

5. Equity deposits provide a useful risk management tool.

Hedgers may need to borrow stock and sell it to hedge some other position they hold, thus offsetting risk. Long-term fixed interest rate loans may offer the hedger more peace of mind than a series of overnight loans that might be reset at higher interest rates without warning.

In closing, ETFs and index mutual funds have become more than just vehicles for retail investors to get passive market exposure. They have also become intermediaries between overnight lenders and borrowers of stock, even if most retail investors in ETFs do not actually realize that they have become lenders. An equity deposit mimics the passive market exposure provided by an ETF while extending the lending business from an overnight basis to what should be a more profitable long-term basis.

It is generally accepted that stock lending brings stability to the marketplace. But as we know from the financial crisis, overnight markets are run-prone. Long-term stock lending via an ED solves the run problem; it seems like an incremental way to create a more robust system.

Monday, January 11, 2016

Even cheaper than an ETF

John Bogle, father of passive investing

With fees as low as 0.10%, passively managed ETFs are one of the cheapest ways to get exposure to equities. Not bad, but here's a financial product that would be even cheaper for investors: an equity deposit. I figure that equity deposits would be so cost efficient that rather than charging a management fee, investors would be paid to own them.

To understand how equity deposits would work, I want to make an analogy to bank deposits. Think of an equity ETF as a chequing account and an equity deposit, or ED, as a term deposit. In the same way that chequing deposits can be offloaded on demand, an ETF can be sold whenever the owner wants, say on the New York Stock Exchange or NASDAQ. Equity deposits, like term deposits, would be locked in until their term was up up. Issued in 1-month, 3-month, 1-year, 3-year, and 5-year terms, EDs would replicate a popular equity index like the S&P 500.

Given that both ETFs and EDs track the same index, and both provide the same dividends, the sole difference between the ETF and the ED is their liquidity. A commitment by an investor to an ED is irrevocable (at least until the term is up) whereas an ETF allows one to change one's mind.  ETFs represent liquid equity exposure; EDs are illiquid equity exposure.

An investor might buy an ED rather than an ETF for the same reason that they might prefer term deposits to a chequing account; they are willing to sacrifice liquidity for a yield. For a trader with a holding period of a few minutes or hours, an ED would be an atrocious instrument. On the other end of the spectrum, long-term buy-and-hold investor would be perfect candidates for substitution from ETFs into EDs. Come hell or high water, investors following a buy and hold strategy have pre-committed themselves to owning equities till they retire. As such, they don't need the permanent liquidity window that ETFs provide. Now if that window were provided free of charge, then investors may as well buy ETFs. But liquidity doesn't come without a cost, as I'll show below. Which means that long-term investors who own ETFs are paying for a worthless feature.

Better for a buy and hold investor to slide a portion of the portfolio that has already been dedicated to ETFs into higher yielding 5-year EDs, rolling these over four or five times until they retire. In doing so, investors get a higher return while forfeiting liquidity, a property they put no value on anyways.

How is it that an ED can provide a higher return than an ETF? Here's how. Once investors' funds have been irrevocably deposited into a 5-year vehicle, the manager buys the stocks underlying the S&P 500. Next, the manager offers to lend this stock to various market participants, either short sellers looking to make a quick buck or market makers who want to replenish inventories. These loans, which are quite safe due to the fact that the borrower provides collateral, earn a recurring stream of interest income which the ED manager shares with the ED investor. This return should be high enough to more-than counterbalance management expenses such that on net, ED investors end up earning an extra 0.25% or so each year rather than paying 0.10-0.50%.

But wait a minute, why don't ETFs do the same thing? Why don't they lend stock and share the income with ETF investors? Actually, they already do. And in some cases, ETF managers are already providing investors with more in lending income than they are docking them to manage the ETF. See the screenshot below from an iShares quarterly report:

The iShares Russell 2000 ETF, which has a net asset value of around $25 billion, provided investors with $34.58 million in stock lending income in the six months ended September 30, 2015, well in excess of advisory fees of $27.85 million.

So yes, ETFs can and do earn stock lending income, but my claim is that an ED manager following an equivalent index would be able to earn even more from lending out stock. To understand why, we need to think about how stock lending works. Stock loans are usually callable, meaning that the lender, in this case the ETF, can ask for a return of lent stock whenever they want. Callability is terribly disadvantageous to the borrower, especially a short seller, as they may have to buy back and return  said stock when they least want to, say during a short squeeze. In order to protect themselves, a short seller will always prefer a non-callable stock loan, say for 1-year, then a callable one, and will be willing to pay a higher interest rate to enjoy that protection.

ETFs and mutual funds are not in the position to provide non-callable stock loans because ETF and mutual fund units can be redeemed on demand by investors. For instance, if performance lags a mutual fund manager may start to experience large redemption requests. To meet those demands, the manager needs the flexibility to recall lent stock and quickly sell it. As for ETFs, units can be redeemed when authorized participants submit them to the ETF manager in return for underlying stock. So an ETF manager is limited in their ability to lend out stock on a long term basis lest they are unable to fulfill requests from authorized participants. Because ETFs and mutual funds can only lend on a callable/short-term basis they must content themselves with a correspondingly low return on lent stock.

As one of the only actors in the equity ecosystem with a long-term pool of pre-committed stock-denominated capital, an ED manager is in the unique position of being able to make non-callable term stock loans. Put differently, redemption of EDs is distant and certain, so only an ED structure allows for the perfect matching of long term assets with long-term liabilities. This means EDs should enjoy superior stock lending revenues, more than offsetting the costs of running the ED.

Say that an ED can beat an ETF by around 0.5% a year thanks to its superior stock lending returns. That doesn't sound like much, but compounded over a long period of time it grows into a large chunk. For instance,  If you invest $2000 each year for 25 years in an ETF that earns 8%, you end up with $157,900. Place those funds in an ED that returns 8.5% and you end up with $170,700. That's a pretty big difference.

EDs don't exist. But if they did I'd probably sell a significant number of my ETFs and buy EDs. I could imagine putting 20% of my savings in a 5-year S&P 500 ED, for instance. What about you?

PS: Feel free to torture test this idea in the comments
PPS: It is very possible that this product already exists.
PPPS: The devil is in the legal details.

Related posts:

An ode to illiquid stocks for the retail investor 
A description of the moneyness market 
If your favorite holding period is forever 
Beyond Buffett: Liquidity-adjusted equity valuation 
Liquidity as static 
No eureka moment when it comes to measuring liquidity

Sunday, January 3, 2016

What makes money special, the lawyer's edition (with a guest appearance by bitcoin)

Juan Galt recently introduced me to one of bitcoin's biggest problems. Bitcoin is not money, at least not according to the law.

Economists like to say that money is unique because it is a medium of exchange, store of value, and unit of account. Lawyers and judges have a different story to tell about money's uniqueness. Unlike goods, money can't be 'followed.' When a good is exchanged, its entire history goes with it. This history may be checkered. Say that a car has been stolen at some point in its past and then sold, and the police discover this fact. The current owner—though having purchased the car innocently—is required to return it to its rightful owner. The law 'follows' goods.

With money things are different. Each time a monetary instrument is transferred, its history is wiped clean. As long as the recipient accepts the money in good faith, the original owner of stolen dollars cannot make a claim for those dollars.

This peculiar legal treatment of money, dubbed money's liability limitation by Steve Randy Waldman, ensures fungibility. When all members of a population can be perfectly substituted for each other, than we say that they are fungible. If each monetary unit's unique history becomes a datum that merchants must take into account before selling a good, then fungibility no longer prevails. One unit may be worth more than another because its history is more pristine.

Fungibility is important because it promotes the smooth functioning of a monetary system. If merchants have to analyze each piece of money they are offered to ascertain its legitimacy, long lineups will develop. Exchange grinds to a halt.

So why not extend the status enjoyed by current forms of money to bitcoin? What follows is a quick tour through the history of how jurists have rationalized the legal treatment of other forms of money, including coins, banknotes, and bills of exchange. This should provide us with enough grist to analyze bitcoin's current legal status.


Let's start with coins. The basic principle of nemo dat quod non habet governs property; no one can give away that which they do not have. According to early common law jurists, coins were exempt from nemo dat because they couldn't be followed. The inability to follow coins arose from the fact that they were homogeneous. In the words of the jurists of the day, 'money has no earmark.' Whereas one pig could be differentiated from another thanks to the practice of earmarking—cutting out a distinct piece from a pig's ear—coins could not be earmarked, and therefore could not be differentiated.

Thus there was no way for a victim to lay claim to lost or stolen coins. With no way to prove that the coins in the accused's pockets had not already been there, mixed coins could not be sufficiently distinguished to establish title. James Fox, for instance, cites a 1614 case in which a gambler, Warde, "thrusts" his coins into the stack of another gambler, Aeyre, perhaps hoping to get a tell from of his opponent. Aeyre refuses to give the coins back. The judge upholds Aeyre's rights to the entire stash since money has no earmark, and therefore nemo dat does not apply. Once mixed, who ever possesses the pile of coins has the best title.

Interestingly, the only way to preserve ownership of coins in the medieval era was to keep them in a bag. Since they could now be identified by the distinctiveness of their container, like any other good they were subject to nemo dat. Had Aeyre's coins been bagged, he could have easily mixed them with Warde's without losing title to them.

The fact that coins had no earmark meant that each piece's distinct past was irrelevant. While this was awkward for poor Aeyre, society was made better off by this decision. Coins became much more fungible than they otherwise would have been, and this would have dramatically promoted their use in trade, greasing the wheels of commerce in general.


Let's move on to paper credit, namely bills of exchange and banknotes. While bills of exchange developed in the 12th or 13th century, the first notes would not have appeared in England until the 17th century. Though English common law was useful for land disputes, it had not yet developed the expertise to deal with commercial disputes. Indeed, common lawyers' expertise with commercial matters was so limited that Josiah Child, an English trader, complained that he could only make his lawyers understand "one half of our case, we being amongst them as in a Foreign Country."

Rather than resorting to common law, problems arising from the usage of negotiable instruments like bills were governed by lex mercatoria, or merchant's law, a private form of commercial law or custom that had been developed by European merchants over the preceding centuries. Market courts, operated by the merchants themselves, guaranteed a decision the day after a complaint, a necessity given the mobile nature of commercial life.

According to Lowry, the close-knitted nature of the merchant class began to unravel by the end of the seventeenth century, making merchant law less enforceable. As commercial cases were increasingly brought to common law courts, jurists had to decide how to treat these new financial innovations.

Lex mercatoria had always accepted the principle that, as in the case of coins, bills of exchange could not be followed. Since those who accepted bills of exchange didn't have worry about whether they had been stolen or not, this would have made trade in bills of exchange extremely fluid. However, the stance taken by lex mercatoria was an anathema to common law logic. Unlike coins, which couldn't be followed due to their lack of earmark, both bills of exchange and banknotes did have earmarks. Whereas coins were issued in uniform denominations, bills of exchange were usually made out in non-standard ones, say $101.50, making for easy identification. Bills were also signed by a unique debtor and a range of consignees. As for banknotes, these had serial numbers on them. Without the homogeneity of coins, there seemed to be no way to save the these relatively new financial instruments from the harsh strictures of common law nemo dat. Goods they were to be, not money.

It was Lord Mansfield, an English jurist, who took on the task of incorporating lex mercatoria into English common law (Adam Back notes a similar case in Scotland). Take Miller v Race, Mansfield's definitive ruling on banknotes in 1758. The note in question had been issued by the Bank of England "to William Finney or bearer on demand" and subsequently mailed to a third party by Finney. Along the way it was stolen and used to buy room and board at an inn, the innkeeper Miller innocently accepting the note. Finney, upon learning of the robbery, asked Race, an employee at the Bank of England, to stop payment of the note, upon which Miller the innkeeper sued Race. If the bill was treated as a regular good, then Finney would have prevailed. However, Mansfield ruled that despite the note having been stolen, Miller had the best title and was allowed to keep it.

In justifying his ruling, Mansfield dismissed as "quaint" the old earmark principle for not following money. Instead, he appealed to the common mercantile practice of the day. Banknotes, wrote Mansfield, are:
not goods, nor securities, nor documents for debts, nor are so esteemed; but are treated as money, as cash, in the ordinary course and transactions of business, by the general consent of mankind, which gives them the credit and currency of money to all intents and purposes. The are as much money as guineas themselves are, or any other current coins that is used in common payment as money or cash. 
The true reason that money cannot be followed, said Mansfield, is upon "the currency of it; it can not be recovered after it has passed in currency." Thus had Finney sued the robber before he had spent the stolen note, he would have succeeded in claiming title since the note had not yet passed into currency. But once Miller accepted it, the note was "in currency" and thus out of nemo dat's reach. In subsequent rulings, Mansfield extended this same protection to bills of exchange, cheques, bonds, and exchequer bills. Any contrary decision would "incommode" trade and commerce, wrote Mansfield. Thus the customs of merchants were transcribed into common law.


So both lex mercatoria and the common law tradition that superseded it accepted the principle that in order to protect commerce, highly liquid instruments should not be subject to nemo dat.  Given this precedent, why not extend this same broad amnesty to modern monetary innovations like bitcoin, Fedcoin, or other digital bearer tokens?

One reason could be that bitcoin hasn't proven itself yet. Whereas bills of exchange and banknotes had been widely accepted for decades, even centuries prior to Mansfield's ruling, bitcoin is less than a decade old. It fails the my-grandmother-uses-it-test or, in Mansfield's words, lacks the "general consent of mankind." People seem more intent on hoarding the stuff than trading it around in the "ordinary course of business." Unfortunately there is a chicken-and-egg dynamic at play here; how can bitcoin gain enough consent to be granted amnesty by the law if it needs amnesty to gain consent in the first place?

Lacking common law amnesty from nemo dat, an alternative would be to modify bitcoin so that it is completely anonymous. Although it is true that the real world identity of a bitcoin owner remains unknown, the blockchain itself is a publicly-distributed ledger that reveals the history of every single bitcoin. Removing the ability to see the ledger's history would restore true anonymity. In the same way that coins were originally exempt from nemo dat because they were physically impossible to follow, modern law would not be able to trace any given bitcoin because there would be no means to do so. Anonymity in turn guarantees fungibility, without which mass market adoption might never happen. My understanding is that extensions such as Zerocoin or Zerocash would be able to achieve this sort of true anonymity.

The third route is to roll with the punches and accept non-fungibility. If merchants must search each bitcoin's past, they will innovate solutions to cope. One innovation would be to set up a system for grading bitcoin so as to save on transaction time. Tokens that pass a test of authenticity would be accepted at par whereas low grade bitcoin, that which has a soiled history, would pass at a large discount to pure bitcoin. I believe that a few bitcoin grading services have emerged, including Mint Exchange, which sells freshly-mined bitcoin (which are unburdened by a history) at a premium to regular bitcoin.


Let's explore the third route a bit more. There is precedent for non-fungible monetary systems. During the so-called Wildcat banking era in the early to mid 1800s, U.S. privately-issued banknotes of the same denomination (say $1) were often  accepted at varying discounts to par. A $10 note from a the Bank of Talahassee might only be worth 98% that of a $10 note from the Bank of Fargo.

While banking regulations prevented note-issuing banks from establishing branches beyond state borders, nothing kept their notes from circulating outside of their home state. However, for notes to be settled in gold, they had to be returned to the issuing bank. Given the large distances involved and lack of transportation infrastructure, this could be an expensive process. To recoup this cost a merchant would typically accept local notes at par while applying discounts to non-local notes. The discount acted as a fee that covered the merchant's transportation costs. And since each bank's brand of notes involved different transportation costs, there were a bewildering number of discounts.

To solve the non-fungibility problem, a new profession emerged, that of a banknote analyst. In addition to providing merchants with information on how to spot counterfeit bank notes, an analyst would publish a weekly banknote reporter that advertised the market price of each banknote that circulated in a particular city, say Philadelphia. Gary Gorton provides a visual feel for what one looks like. Philadelphian merchants who subscribed would, upon being proffered a particular note by a customer, consult their reporter and apply the proper discount. I've explained in more depth how this process worked here and here.

While a Wildcat-era sorting mechanism for bitcoins would help merchants cope with the fungibility problem, any sort of grading process would also impose an extra set of costs on the bitcoin system, making it less competitive with banknotes and deposits. The lack of uniformity of U.S. banknotes was recognized to be enough of a problem that the 1864 National Banking Acts required all banks to accept notes at par (it would have been better to allow banks to establish branches across state lines, of course. See George Selgin here).

Uniformity would certainly be the most efficient solution for bitcoin, but lacking a central authority that can enforce par acceptance, bitcoin may have to endure a period of non-fungibility before the law deems the cryptocurrency popular enough to earn amnesty from nemo dat. That's a low bar to set, but if bitcoin is as good as its proponents say, it should be a bar that can be limbo-ed.


S. Todd Lowry: "Lord Mansfield and the Law Merchant: Law and Economics in the Eighteenth Century" (1973) [link]
Benjamin Geva: "The Payment Order of Antiquity and the Middle Ages" (2011)
Kenneth Reid: "Banknotes and their Vindication in Eighteenth-Century Scotland" (2013) [link]
David Fox: "Banks v Whetston" in Landmark Cases in Property Law (2015)
Tim Swanson: Unable to dynamically match supply with demand (2015)
Nick Szabo: From Contracts to Money (2006)