The Inevitable Awkwardness of Money Markets

Most confusing concepts I have encountered during my finance career included plays around the word money: smart money—very high self esteem, real money—then there’s indeed fake money, fast money, sound money, multiple of money—not to be confused with money multiplier, etc. Of those money-related depictions, money markets have been the most opaque concept for me. What is exactly a market that trades money? Isn’t money a payment tool in itself, and therefore why would I buy money…with money? It turns out that those are actually markets where money is fabricated, at least temporarily, rather than properly purchased. It also turns out that my confusion wasn’t a coincidence, since money markets indeed make of opacity one of their core functioning characteristics. Sometimes a bit of opacity is a good thing.

Opacity As a Feature In Money Markets

What are MMs? In short, a money market (or MM) instrument is a short-term / low-risk way for one side to temporarily park liquidity in a safe place and for the other to obtain short-term financing at very effective rates. A lot of money gets created and destroyed like this, by pledging safe assets among financial institutions and central banks. MMs instrument are everywhere, and exactly because of this they work very differently from most others.

I am not the first one to point out how the observable functioning of equity markets has often brought participants to imply all markets should work under similar premises. In BIS Working Paper 479, Bengt Holmstrom wonderfully depicts how the MM, for one, has been structurally running away from price discovery in order to fulfil its key functions. MM instruments are indeed founded on the concept of no questions asked, or NQA-ness—more on NQA-ness can be found here. Those markets are very big, practically ubiquitous, and operate too fast to accomodate too many questions. A well-functioning (meaning liquid and resilient) MM is, in other words, founded on satisfactory opacity, where users ignore the details and idiosyncrasies residing behind the instruments themselves as they continue to go along with their lives. This is totally opposite to what happens for equity markets, where price discovery is king. Using Holmstrom’s words—emphasis is mine:

Stock markets are in the first instance aimed at sharing and allocating aggregate risk. To do that effectively requires a market that is good at price discovery. […] Information will quickly be reflected in prices and, since prices are common knowledge, beliefs will not be biased one way or the other to permit someone with just the knowledge of prices to make money. […]

The purpose of money markets is to provide liquidity for individuals and firms. The cheapest way to do so is by using over-collateralised debt that obviates the need for price discovery. Without the need for price discovery the need for public transparency is much less. Opacity is a natural feature of money markets and can in some instances enhance liquidity, as I will argue later.

(Good) money is (over-collateralised) debt → If money, as we have reminded the reader in innumerable occasions, is debt, then well-functioning money is typically significantly over-collateralised debt. It is through the use of a significant enough buffer that money markets circumvent time-consuming negotiations and therefore enhance liquidity. We could say that in MMs obfuscating information becomes beneficial to both sides—with a large enough collateralisation ratio private information about the true punctual value of such collateral is totally irrelevant and transactions can be facilitated.

In determining this collateral ratio, the purpose of the lender is to instate a painful-enough punishment for the borrower in case of default. Borrowers, on the other hand, have the incentive to minimise such a collateral ratio for purposes of capital efficiency (the good faith borrowers) or fraudulence—the bad faith ones. While this is the case for any debt market, this negotiation is even more important for money markets, where both sides benefit from the liquidity and pervasiveness of the liability claim. All those considerations drive the collateral ratio towards a fungible equilibrium. This is not only true for repo markets (the most atomistic money markets) but also for more complex money printing machines—namely banks, where the boundary condition is more broadly an equity rather than a collateral ratio—from #57 again.

In graphical terms (see below) information becomes relevant for collateralised debt instruments only around the so-called default boundary, i.e. only when the market value of the collateral gets close to the boundary. You don’t care about the business practices of your depositary bank until it starts becoming public that you should. Holmstrom reminds the nerdier ones that in asset pricing jargon a borrower corresponds to some sort of buyer of a put option, i.e. the right but not the obligation to give away the collateral (fixed value) at a fixed maturity date by defaulting. That’s why the payoff curve of the debt instrument is a concave one.

As with other option instruments, the shape of the curve depends on a set of assumptions, and structuring decisions. We can use Black-Scholes to estimate the probability of default of such an instrument.

Where N is the standard normal cumulative distribution function, B is the face value of the collateral, D is the default boundary—i.e. the face value of debt, r is the interest rate, Sigma is the underlying volatility of the collateral, and T the time to maturity. We can summarise the factor impacting the probability of default, and therefore the shape of the red line:

• Debt vs. collateral: the larger the buffer at inception, the closer the market value will be to the face value of the debt—i.e. more insensitive to private information
• Maturity: options increase in value with longer maturities, given that a disparate number of events becomes more plausible—in the context of debt markets longer maturity dates decrease the market value of debt vs. its face value
• Interest rates: changing risk-free rates has an impact on the time value of money and therefore the value of debt, but it is irrelevant for this discussion
• Volatility: the volatility of the underlying collateral has huge implications on the market value of debt—higher-vol collateral reduces the market value of debt and can only be compensated by larger collateral buffers

Structural Implications For Money Market Instruments

The observations above have several implications on the structure of MMs that should be interesting for money users, regulators, as well as designers of future money instruments.

(i) MMs and debt markets have different ambitions → Although MM instruments are a specific type of debt instruments, they serve a different primary purpose: system liquidity. This drives an extreme need for standardisation and frictionlessness that is not necessary relevant in more general debt markets. If debt instruments are some form of put options, MM ones are deeply out of the money (or alternatively hyper-short) put options, which are extremely insensitive to private information.

(ii) Capital efficiency shouldn’t be the primary ambition of MM → The meme of capital efficiency has been a pervasive one in crypto since the beginning of DeFi, and used each time to defend a newly-branded construct: endogenously collateralised stablecoins, under-collateralised lending protocols, heavily algorithmic money markets, etc. The meme states that requiring large amounts of collateral to back a financial claim is very inefficient and we should find (smarter) mechanisms to reduce the collateral locking. It is fair (to me) to say that DR is not against (good faith) innovation in financial infrastructure, but also fair to ask ourselves what is the price-benefit relationship of each innovation. Due to their central place in financial intermediation MMs should with high probability remain the most conservative markets; given their structural pervasiveness a failure at the core of such markets could have devastating ripple effects across the whole economy. Imagine what a default of the US Treasury could do to the global financial markets, or a structural debasement of $USDT/ $USDC could cause within DeFi.

(iii) Volatile (or opaque) collateral is inefficient collateral for MMs → There are, however, bad and good ways to improve capital efficiency of MMs. If reducing collateral ratios alone is a bad improvement, adopting less volatile / opaque collateral is a good one. In option pricing jargon, this would mean reducing the volatility expectation of the underlying reference asset and hence increase the curvature of the red line. Crypto, unfortunately, doesn’t have (much) non-volatile collateral when measured in dollar terms—technically stablecoins are supposed to be those more stable assets. Ironically, with debt being structurally more stable in valuation than equity claims, debt-on-debt is a very good way to structure a MM instrument—which is exactly what happens in the repo market and what makes the debt collateral used even more pervasive and impactful across the economy. The cornerstone collateral of MMs are US short-term treasuries.

(iv) More information doesn’t always lead to a better outcome → In a world of total transparency over the collateral characteristics (imagine continuous mark-to-market or a live accounting feed for a bank’s balance sheet) private information would be even more valuable, and the red curve above way more unstable, with unpredictable and discontinuous effects on MM claims. As Holmstrom affirms, “intentional opacity is a rather ubiquitous phenomenon” and, “purposeful opacity can enhance liquidity in MMs”. Money, the ultimate MM instrument, is absolutely opaque, and that is good enough—until it’s not. We ignore the quality of the US balance sheet, or of JP Morgan’s, and don’t bother to check the CDS curve of the depositary institution holding our savings. Money is quintessential NQA-ness. What matters more is the structuring of the liability claim, and the original positioning of the boundary condition: if you have banks going under one after the other people will start caring with whom they deposit their money.

(v) Non-observable risk means tail risk → There is no free lunch, and creating a system that in normal conditions is insensitive to private information (point iv) and structurally incentivises leverage on leverage (point iii) pushes a lot of risks into the tail. Everything is great, making people oblivious of the underlying market structures, until something goes wrong, then it’s the apocalypse. Panic is, in financial engineering jargon, an information event: expectation on the market value curve abruptly shift causing a domino effect across the market structure. This is what typically happens in MMs, see the Great Financial Crisis on which enough has been written. It is unclear whether the rise of the so-called shadow banking problem has actually had negative implications (due to increased hidden leverage) or actually acted as a better risk allocator throughout the stack. Yes, it has made contagion more sismic, but with the benefit of increasing liquidity and growth (through better capital allocation) for a long period of time. It is arguable that whas is happening in DeFi today, with structures being created on top of other structures to expand leverage and, let’s say, improve capital efficiency, has parallels with the emergence of the shadow banking sector.

(vi) Credible bailouts are good preemptive measures → Contagion dynamics seem therefore inevitable given the structures involved, and panic a reasonable, if rare, price to pay for liquidity and fungibility. The best we can do is to improve structuring making panic less common, and to have safety measures in place when panic occurs. Bailouts might be inevitable for a useful MM. Question is how expensive those bailouts are, and who’s bearing the ultimate costs. But those topics are beyond the point of this DR entry.

Conjunctural Observations For So-Called Stablecoin Instruments

Most of the implications described above can be applied to the current vintage of crypto-facilitated (so-called) stablecoin instruments. Let me clarify before continuing: the spirit of this section is not to criticise the status quo, but to refine a personal mental map and guide researching and building activities.

CDP is the way → So-called crypto-native stablecoins became very interesting to economists given their impact on the money creation process itself. Maker’s Collateralised Debt Positions (or CDPs) allowed anyone with eligible collateral to access the protocol discount window and mint a fully-fungible, soft-pegged, dollar-denominated liability. For the borrowers, such a liability would have improved the liquidity of their collateral without foregoing ownership—and upside; for the lenders, i.e. the $DAI holders, the comfort of transacting with a digitally-native stable liquidity instrument. Maker’s original construct was what attracted most minds into DeFi. We can clearly see the parallelisms between the Maker-printed and over-collateralised $DAI of the early days and the more traditional MM instruments described above.

$BTC-backing + $USD-denomination is problematic → The decision to denominate crypto MM liabilities (like $DAI) in dollar terms had to do more with marketing than with financial engineering, and yet exposed the structure to clear weaknesses due to very high volatility of the underlying collateral (mainly $BTC and $ETH) relative to the denomination of the liability. That mismatch was solved in the only way possible, i.e. with severe collateral haircuts—causing capital inefficiency. This type of capital inefficiency was unavoidable: while $BTC and $ETH were the most stable crypto-assets available, they weren’t stable enough. Almost every other design followed through along the same lines. Analysts often point also to the lack of fixed yield for those assets, but to me this is a minor problem when compared to the volatility costs—but I haven’t done the math.

Who’s going to bail crypto out → We have admitted the necessity of the existence, even ex ante, of credible safe valves when MM markets go havoc. The dollar MM has such credible bailout infra, but what about crypto? Today, our best bet is to piggyback on real-world central banks. To my knowledge, nobody has worked on an ultimate, blockchain-wide, safety insurance for DeFi MM, and it might well be one of the transformational innovations we need to go full digital in money creation.

The pegging obsession → One of my main problems with the term stablecoin has been the overemphasis over short-term stability at the expense of other factors. While in the long term stability is a good proxy for solvency, in the short one it is better correlated with liquidity. This obsession has driven projects to over-invest on secondary liquidity at the expense of stability (see Tether) or to pollute their original designs introducing dangerous stability mechanisms—see Maker and the PSM.

Crypto is a mind-bending beast, where the boundaries between assets and the infrastructure where those assets flow are blurred. Many of our conceptual mistakes can be traced back to this. Stablecoins sit at the junction of two worlds trying to satisfy both, and hence they are convoluted by design. What would then be alternative ways to apply useful blockchain-enabled constructs more appropriately?

1. Crypto repo rails + USD instruments: blockchain protocols continue to be used as very effective repo channel for dollar-stable, i.e. non-crypto native, collateral such as US Treasury bonds in order to issue dollar-denominated MM instruments on the blockchain; to me this is the most powerful use case today
2. Crypto repo rails + crypto-native instruments: blockchain protocols are effective repo channels where pledgers provide crypto-native collateral to obtain novel crypto-native liabilities that have nothing to do with the USD. This system has a lot of advantages, last but not least elegance, and structural soundness given that e.g. $BTC issuance is disjointed from monetary policy consideration within the ecosystem. The disadvantage is that this construct has very little application in today’s USD-dominated world
3. Crypto repo rails v2 for hybrid structures: constructs that produce USD-denominated liabilities dealing only with crypto-native collateral might not be applicable today but might well be in a future where the digital collateral pool expands or the dollar valuation of digital assets, stabilises. This is a powerful construct but is not today’s reality

The term stablecoin is a marketing gimmick that intends to communicate a blanket promise of stability that cannot be verified—stablecoins tend to be very stable until they are not stable anymore. Under the blanket there are too many different business models, too many different constructs, and too many divergent set of risks. Ultimately, the term stablecoin diminishes money to a mere payment tool while what is more interesting about money (to me) is how money is created in the first place. Regulators and central bankers know very well that what matters is who’s controlling the printing machine, and are trying to perpetuate the meme money = payment (hence the term stablecoin) because it helps cornering the phenomenon simply within a sub-category of digital payments, staying away from more fundamental questions. Perpetuating the term stablecoin is the best way to protect the banking sector.

Disclaimer：

1. This article is reprinted from [Dirt Roads. All copyrights belong to the original author [LUCA PROSPERI]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.