Terra/Luna Debacle: From a Financial Engineering and Risk Lens (Part I)

BY LEO TAY AND PHILIP TE

“History teaches us that a crisis often causes problems to correlate in a manner undreamed of in more tranquil times” ???- Warren Buffett

“Assumptions are made and most assumptions are wrong” - Albert Einstein

???????????????????????????????????????????????????????????????????????????

Over the last few days, the Terra (UST/Luna) collapse hs dominated the headlines not only of crypto media but of mainstream business as well.

This seems to lead to a sudden explosion of overnight clairvoyant experts who quickly dismissed Terra/Luna, the darling of the DeFi space, as nothing but a scam and a Ponzi scheme. ?Many predicted that this will the end of DeFi and dismissed the entire crypto world as nothing but a creation of charlatans and for the hundredth time predicted the demise of Bitcoin and crypto as a whole.

?Admittedly, the authors of this article initially adopted the view that Terra/Luna may be just one of those crypto scams. However, we decided to explore this topic deeper and understand what happened.?

We read the Terra whitepaper co-written by Do Kwon (the South Korean co-founder of the Terra ecosystem). We read working papers both supportive and skeptical about this project. We studied precedent cases (IRON TITAN being the most prominent).

Coming from a financial engineering, trading and risk management background, we realized that there are many parallels between the Terra/Luna debacle and past failures in financial engineering (most notably, the 1987 stock market crash partially caused by program trading/ portfolio insurance and the 2008 subprime mortgage crisis). ?

This will be Part I of a two-part series exploring the financial engineering mistakes of Terra/Luna. In Part I, we will explore the problems that Terra/Luna wants to solve, how it solved the problem and why the solution resulted in unintended consequences that caused its problems over the last few days.

The Problem that Terra/Luna Wants to Solve – Price Stability

Paul Krugman, a vocal crypto skeptic (who of course, famously dismissed the power and influence of the internet), posed a very important question to Anderseen Horowitz (a16z) partner Katie Haun – despite all the news about bitcoin, what does it actually do? ?Other than as a store of value, bitcoin seem to have failed to live up to its potential as a medium of exchange.

Part of the reason why Bitcoin has not succeeded as a medium of exchange is because of its extreme price volatility.?In a rising market, who would want to pay in Bitcoin if in the next month or so it will rise sharply??On the other hand, from a merchant perspective, who would want to receive Bitcoin if it will sharply collapse??

?The Terra/Luna Answer: A Two Coin Approach

Terra solves this problem by coming up with a simple and elegant solution: a two-coins system:

• ?Coin 1 to maintain the peg against fiat (UST – shorthand for TerraUSD).?This Coin 1 is also known as the stablecoin part of the Terra ecosystem.
• Coin 2 to absorb market volatility (Luna)

The goal is to keep Coin 1 (UST) equal to 1 USD.?

To achieve this, Coin 2 (Luna) is used as a shock absorber to absorb market volatility in situations where 1 UST is not equal to 1 USD (i.e. if the peg is not as expected).??

What do we exactly mean when we say Coin 2 (Luna) acts as a shock absorber???

The Role of Luna

There are two main types of exchange rate regimes:?floating exchange rate and fixed (or pegged) exchange rate.?In the floating exchange rate regimes, exchange rates are determined by market forces which means that the volatility is absorbed by market participants.?In fixed exchange rate regimes, to maintain the peg, someone has to bear the costs of volatility to maintain the peg.

In traditional finance, central banks intervene in the foreign exchange markets and absorb the volatility of the pegged currencies they choose to protect.?

In Terra, users absorb the volatility through Coin 2 (Luna) – the governance token of Terra. ?

By shifting the volatility away from Coin 1 (UST) to Coin 2 (Luna) – the system effectively solves the problem of price stability and seems to immediately made Terra both a viable store of value and medium of exchange.

How Does Luna Exactly Do It?

Scenario 1:?Price of UST is Greater than 1 USD

If the price of UST is higher than 1 U.S. dollar, the goal is to push the price of UST lower by making more of it. The more supply of UST, the lower the price will be.?

To do this, users and arbitrageurs participating in the Terra ecosystem should exchange their Luna with UST.?The protocol will mint more UST and the theory is this will help make UST at parity with the anchor fiat currency. ?

Scenario 2:?Price of UST is Less than 1 USD

What if the price of UST is less than 1 USD? ?The goal here is to push the price of UST higher.?To do that, we should make UST scarcer.?The lower the supply of UST, the higher the price will be.

To do this, users and arbitrageurs participating in the Terra ecosystem should exchange their UST with Luna. By selling their UST, the protocol earns more Luna. The less the supply of UST, the higher its price will be until UST reaches the desired parity equal to 1 Dollar.

Parallels in Traditional Finance ?

Scenario 1 is functionally equivalent to an expansionary monetary policy where central banks increase money supply when price levels are above a target fixed level. ?In many ways, this is an easy problem to solve.?

Scenario 2 is functionally equivalent to an contractionary monetary policy where central banks decrease money supply when price levels are below a target fixed level. ?This is a more difficult problem to solve than the previous scenario.

In traditional finance, to contract money supply, central banks have to raise interest rate and issue debt instruments (to siphon liquidity and incur interest expense). Central banks and governments, therefore, bear the cost of the volatility in a pegged currency.

Below is a short example.?Suppose the Hong Kong Monetary Authority (HKMA) would like to set a hard peg at HKD 7.80 per 1 USD. ??

In the left hand side scenario, the market equilibrium is at HKD 7.90 per 1 USD because there is more demand for US dollar than for Hong Kong dollar.?To bring it down to the desired level of HKD 7.80 per 1 USD, HKMA should decrease the supply of HKD relative to USD by buying HKD.?This will effectively bring down the exchange rate to HKD 7.80.?Buying HKD against USD to strengthen it up to a certain level will require HKMA to act as a shock absorber.?

A Critical Loophole in the Solution – The Assumption of Demand

So far, we have discussed the ingenious solution of Terra and compared it to mainstream finance situation.?Let’s focus our attention on the more pressing problem – supporting the price of Coin 1 (UST) in the scenario where it is lower than US dollar.??

The solution:

·??????Reduce the supply of UST

·??????How? By getting people to swap their UST to Luna

The solution seems elegant until we ask the most important question – what would get people to swap their UST to Luna??To understand this, we have to understand that Luna is a governance token.?This represents mining power in the Terra network.?

The theory behind this is that in exchange for protecting the system by absorbing short-term market volatility, miners are rewarded with more power and control in the Terra network.

This presupposes that there are enough participants who will be incentivized to protect the system in exchange for more control over the system through Luna.?

In short, that there will be enough demand from incentivized actors who will bear short-term losses in exchange for longer term reward through power and control.

We highly recommend that you read the paper “Built to Fail: The Inherent Fragility of Algorithmic Stablecoins” written by Dr. Ryan Clements of University of Calgary:

“Reliance in a base level of support is perhaps the biggest problem with an uncollateralized algorithmic stablecoin two-coin structure.?The volatility absorbing coin must maintain certain support level of demand- and not fall below a price threshold- otherwise the entire ecosystem fails”.

The reliance on a minimum level of baseline demand for Luna if the UST drops below the desired peg is a crucial loophole.?The second order consequence is if people no longer trust and believe that other people will continue to believe in the Terra ecosystem, people will rush to convert the UST back to Fiat (and hence, the collapse of the peg). ?

To support baseline demand, the Terra ecosystem developed an entire ecosystem of useful protocols and applications:?

The Problem with Assumptions, Risk Interdependencies and Analogical Reasoning

There were so many things that was said about this debacle but we think a large part of the problem is due to the inherent fragility of the system.?And as what we discussed in the previous section, the fragility is largely caused by the need to always have a baseline demand.

Why did Luna’s value drop to zero??Simple – because the ecosystem relies on Luna to act as a shock absorber. That is an assumption. ?However, the crucial aspect that was missed is that Luna relies on the Terra ecosystem for people to accept it as a shock absorber.

One of the key issues that the Basel Committee on Banking Supervision (BCBS) and the Financial Crisis Inquiry Commission identified in the 2008 financial crisis is the interdependencies of risk particularly their mutually reinforcing and amplifying characteristics.?

Charlie Munger said that a lot of problems in business and in life can be solved if one bothers to ask the question – “and then what?”.?

Luna is the shock absorber that will maintain price stability according to the protocol.?But if an overall meltdown in confidence in the system, asking the question “and then what?” will make us realize that sound assumptions normally work well all the time in theory, work well most of the time in practice and in extreme situations, they don’t.?

We highly encourage everyone to read the Terra Whitepaper.?It is concise, well written and well argued. It tries to achieve something ambitious – to replicate the entire financial system using blockchain technology and smart contracts.?However, it has one major shortcoming – its heavy reliance on analogical reasoning.?

Reasoning by analogies (or synthetic replication as how we want to call it) is one of the favorite thinking mode of financial engineers and arbitrageurs.?If you look at history (which, ironically is also mostly reasoning by analogy), financial engineering disasters are caused by incomplete analogical reasoning.?In its quest to create a financial system from the ground up, it relied too much on similarities with how central banks defend the peg (stability) and governments stimulates through fiscal policy (encourage adoption).

Nassim Taleb, the great social philosopher, said it best: ?“Regular minds find similarities in stories and situations), finer minds detect differences”.

While unfortunate, we take heart that despite all this fragilities, like the traditional financial system that it sought to emulate, it takes a lot of time and humility ?to understand the similarities and differences.?

From the collapse of money/ hyperinflation during the Weimar Republic in ?1923 to the Great Depression in the 1930s, the hyperinflation/stagflation of the 1970s, the 1997/1998 emerging market crises in Latin America, Russia and Asia, the 2001 technology bubble/ Enron scandals and the 2008 financial crisis – despite its relatively longer history, our understanding of the financial system - it's logic, dynamics and interdependencies, is still evolving.

In Part II, we will look at history from the 1987 stock market crash to the 2008 financial crisis to the recent failures of stablecoins not only to understand the similarities but also the differences in the current situation.

Recommended Readings:

1.?????Terra Whitepaper:?Terra Money – Stability and Adoption

2.?????Built to Fail:?The Inherent Fragility of Algorithmic Stablecoins by Dr. Ryan Clements

3.?????Stablecoins: Growth Potential and Impact on Banking by Gordon Liao and John Caramichael (Federal Reserve International Finance Discussion Papers)