Where finance and media intersect with reality


Why financial engineering has gone full circle with Terra

Screenshot 2022-05-12 at 12.12.49

I’m really pressed for time but I wanted to get this out as it’s important. Here as a result is the longish thread-based reply I just posted on Twitter in response to this question:

One question: if I buy tether, my cash was supposedly going into a short term dollar asset (asset side of tethers balance sheet). But if I bought luna or terra ust, where would my cash go? You know that?

As I replied:

This is the question isn’t it. In theory this was structured like a synthetic ETF. So the issuers could use the proceeds however they saw fit to “defend the peg”. The logical thing to do from a risk mgmt perspective was invest in dollar securities.
But the issue with MMFs/Stablecoins is that simply matching things one to one is an incredibly low margin exercise. If not negative margin exercise. It’s also subject to slippage risk. The way MMFs compensate for this is by charging fees.
But if you charge fees in the hyper competitive world of crypto you will lose market share. So the incentive has always been to play the “delta one” game. I.e. take the proceeds and invest them in assets that will OUTPERFORM the dollar.
You then get to pocket any outperformance for yourself while returning the par value dollar return to the customer. The least risky way to do this is to invest in higher yielding “safe assets” that are highly correlated to the dollar.
But in crypto, the preference was to take a lot more risk by investing in almost everything — including your own asset creation via your own loan-making business — and simply mark it to market all the time.
This essentially transformed these stablecoins into shadow investment banks and even in some cases de facto commercial shadow banks.
When a bank’s loan book sours, or its assets need to be written down, this has a direct impact on the equity value of a bank. The same applies to the investments of these stablecoins.
If you think of short-term bank liabilities to depositors as the equivalent of “stablecoins” that banks issue, you can think of bank equity as the supporting investor equity that backs the capacity of banks to defend those liabilities.
This is a similar relationship to what Terra had with Luna. Think of Terra as the short-term depositor liability, and the Luna as the bank equity.
Now imagine if a bank could control for the stability of its “money liabilities” not via a liquidity guarantee from the central bank, but by being able to continuously tap the market for its own equity whenever it needed a top up?
That’s essentially how the “arbitrage mechanism” at the heart of Terra/Luna operated. Like a permanent rights issue to the market. And as long as the “bank equity” was supported by the market, it could fund the liabilities.
A key thing not well understood about stablecoins is that unless their business models’ incorporate a fee structure (and most of them do not), they can’t really make enterprise profits. The only way they can generate profit aside from charging fees is by taking undue risk with the funding they receive from depositors.

But this is basically the business that banks are in: maturity and liquidity transformation.

Banks take short-term unpredictable funding from depositors and invest it in longer term but higher yielding assets, generating a profit. Alternatively — as was the case in the pre Volcker rule days — they would also take the cheap funding from depositors and use it to punt on other securities and assets, hoping that these assets would outperform their depositor liabilities on a “number go up” basis. The differentials were captured as industry profit.

The final model is to charge fees for par management. This, ironically, is both the least risky and most transparent model, but also the least popular. Mostly because the market is entirely rubbish at recognising that if it’s free you’re probably the product.

Of course, what the global financial crisis showed us was that even licensed banks could not be trusted to operate these models in a prudent manner. The vulnerability in the system was always the fact that “networked” banks could create money out of thin air by creating loans funded by their own ledger entries. What this generated was a market practice where banks would lend first (creating balances out of thin air) and fund those balances later.

By this I mean, banks would seek to find funds that could cover the risk that those liabilities they had just created might be redeemed and taken out of their closed systems.

Very much like crypto, the more the equity value of banks went up, the easier it was for banks to take funding risk, as in theory (but rarely in practice) they could always sell more equity to cover any mismatches.

Bank equity (and the banker bonuses that went with it) were the crypto “number go up” of its day. Here as an indicator is a chart of the long-term share performance of Goldman Sachs to illustrate the point. The period up until 2008 was one in which bank equity was helping to support the liabilities of the bank:

This was even more the case with commercial banks like Bank of America:

As Anat Admati always liked to point out, banks didn’t like using equity to finance their risk taking because of its relative expense. The equity return ratio was a sacred cow in the banking industry for a long time.
Why? Because the whole point of investing in bank equity was the opportunity to ride the “number go up” confidence wave and benefit from attractive dividends. If you actually had to be on the hook for the liabilities that the banks were creating, that equity buffer would be impossible to price due to its susceptibility to market panic and liquidity events. It would turn the rationale of the investment upside down.

That made equity funding of liabilities extremely expensive, and incentivised incredibly low tier one equity ratios in the system as a whole.

What’s more, as a network, banks didn’t need to rely on their collective equity to defend their liabilities. In a crisis they always had the ability to tap the lender of last resort for liquidity — allowing the proceeds of equity gains to be spent however they saw fit (usually in dividends and bonuses). Admati’s key criticism was that banks preferred debt financing over equity because it lowered the amount of taxes they paid but also because it allowed them to benefit from implicit guarantees by the government, rather than running prudential businesses directly. The incentives, in other words, were all warped.

As she noted in her paper at the time:

Government guarantees that allow banks to en joy cheap debt financing create numerous distortions and encourage excessive leverage and excessive risk taking. Because of the distorted incentives as well as the difficulty for governments to commit never to bail out banks, it is challenging to neutralize this eff ect by charging banks for the true cost of the guarantees on an ongoing basis. In this context, equity cushions are particularly valuable, as they reduce the likelihood and cost of the guarantees

These are important considerations for the Terra debacle.
What the Terra founders de facto created was a system wherein the par value of Terra liabilities was supported by an ETF-style arbitrage mechanism linked to the value of its own equity — as represented by the price of Luna cryptocoins. The crypto community viewed this as genius financial engineering equivalent to alchemy itself.
What it really was, however, was an explicit commitment to the market that Luna’s stock-market value would always be used as a cushion to defend the price stability of the system’s “money like” liabilities.As Bloomberg’s Matt Levine describes the arbitrage at the heart of the arrangement (TBS’ emphasis):

One UST is supposed to be worth one US dollar, and one UST can always be exchanged for a floating quantity of Luna with a market value of $1. If a UST is trading at $0.99, you can buy it for $0.99 and then exchange it for $1 worth of Luna, making an instant profit. If it is trading at $1.01, you can buy $1 worth of Luna (for $1) and use it to buy a UST worth $1.01, making an instant profit. Because of this arbitrage relationship, while the price of Luna can fluctuate, the price of Terra should always be $1: If it trades above or below $1, people will exchange Terra for Luna or Luna for Terra until the price of Terra gets back to $1.

But another way to phrase the above is as a promise from Luna Foundation that it will always be prepared to use its own “money like” liabilities to buyback its own shares from the market when it feels they are undervalued in relative terms. This undervaluation occurs when its capacity to “debt finance” is uninhibited (and thus super cheap) because it is being inundated with inflows into Terra — possibly to the point that it risks breaking par value to the upside.Rather than break par in this way (and return the imbalance to its own customers) the system always has an incentive to return the positive slippage to its primary insiders and investors. On the flip side, the commitment extends to a promise to always issue more stock to the market at a discount if it feels its stock has become overvalued relative to its capacity to manage redemptions. (Which is to say, in the event it can no longer raise debt financing as cheaply as it can raise equity financing).

The two commitments, however, are just another manifestation of something the old school banking sector has been up to forever.

On the share buyback side of the equation it’s the same phenomenon as this:

And as Euromoney goes on:

Barron’s also describes the phenomenon here — a trend that clearly coincides with the new run up in bank equity and the influx of “dollar liquidity” into the banking system as a result of Covid stimulus.
On the flipside, Luna selling its stock at a discount relative to the value of Terra, is really no different to this:

Which links back to my original Tweet thread observing that the crypto market was always going to be vulnerable to the return of a positive yielding fiat environment. This is because its gains were always likely the function of an excessively cheap debt financing world.

If that’s true, it’s entirely possible that the last 10 years or so of sky-rocketing crypto gains were largely a transfer of “cheap money” financial profiteering, which — were not for the post GFC regulatory environment — would otherwise have been captured by the banking industry in the form of sky-high bank equity valuations.

The real question at hand is whether diverting this speculative craze into a largely unregulated “caveat emptor” market that the lender of resort is not responsible for, will have helped to ring fence the stability of the core financial system and in this way protect those who didn’t fancy taking a punt with their life savings on something devoid of intrinsic value?
I think it could well have done. How far any crypto contagion rips is going to be key.

The Daily Blind Spot newsletter

Latest posts

If viewing on a mobile simply tap the QR code

11 Responses

  1. Highly interesting question at the end. Sewer systems and automotive radiators have overflow tanks to hold excess literal liquidity. The excess evaporates if not reabsorbed. It would make sense for the real banks to create crypto nonsense as an evaporative overflow tank.

  2. Excellent analysis. A lot of crypto people would be well served if they spent more time studying the history of banking as opposed to just denouncing it. Supporters of algo stablecoins have always compared their design to central banking, but that’s a false analogy.

    As for the fiat backed variety, I think the model is safer than most people think. Unlike a traditional bank (narrow or not) the liability side of any issuer (including Tether) is verifiable in trustless fashion. People debate about USDT’s reserves, but there’s no debate about it’s liabilities since you can look them up on your favorite blockchain explorer.

    While true that there’s an inherent conflict of interest for the issuer to want to take more risks with the reserves than their users would like, issuers can (and do) have other sources of revenues, like APIs and SaaS services for those interacting with the coin. I think this is an underrated business model given the TAM of traditional fiat payments which are still shockingly expensive.

  3. Crypto is too small to fail?

    “Stable coins are just another manifestation of something the old school banking sector has been up to forever.”

    Which is closed-loop financial engineering with the feedback model always pointed upwards for growth. This works well until the market points down due to the cheap money being withdrawn from the system. Big banks get away with contagion mistakes, because the Fed will backstop big banks errors as they are “too big to fail”. Crypto is not networked for contagion, it is too small to fail… a couple of stable coins failing is due to the feedback model not being corrected when the market was pointed down. The rest could hold-up …

  4. Unstable Coin should have been the proper name for this scam which used non-crypto–dollars–to suck real money into inherently unreliable crypto currencies. The explosion of NFTs and Coins of every sort should have been a warning that we go from actual Tulips to plastic one, to images of Tulips that sell for more than the Tulips themselves! I do not see governments stepping in to save anything related to stable coins. Warren Buffett will not offer a deal as he did to Goldman Sachs in 2008

  5. Isabella. Great post but one question/observation. You compare a reserve based stablecoin balance sheet to a bank taking a deposit. That works for a narrow bank model but most banks obviously do not operate on that model. The “safety” of the deposit is not so much about the assets it is invested in and the equity supporting those liabilities but also by the super senior claim the deposits have. Deposits are a deeply over collateralised claim that extends beyond the equity whereas stablecoins do not typically have this over collateralisation

    1. yes, this is why stabelcoins have even worse run risk. As depositors have no confidence they will even be first in line.

    2. Thanks … and yet we continue to see stablecoin advocates assert that stablecoins are better than bank deposits because they have 100% reserve backing whereas deposits have fractional reserve backing. This comparison is I think we agree just plain wrong and yet I continue to see it made without any pushback. I can see a role for fully collateralised stablecoins (and even plain vanilla narrow banks) but they need a lot more capital and lower risk reserves assets than the current industry standard before they approach the risk status of a bank deposit.

      I had a go at the question of what makes a bank deposit safe in this post

      And of course Matt Levine makes the same point about seniority with much greater brevity and wit


Leave a Reply

Your email address will not be published. Required fields are marked *