A Case Study on HUSD

Huobi recently announced a trial upgrade to their stablecoin service, now titled HUSD Solution V2.0. The intent is for Huobi to help meet user demands when it comes to stablecoins by offering support for interchangeability between Paxos Standard (PAX), True USD (TUSD), USD Coin (USDC) and Gemini Dollars (GUSD). Users can use these stablecoins on their exchange, where these four tokens are represented under one quote currency titled HUSD.

Interestingly enough, this rollout was revealed only 3 months after the launch of the first version. What went wrong? The purpose of this article is to highlight HUSD’s design flaws and outline the events which occurred because of it. Detailing these episodes is important because it provides insight on how the ecosystem of the crypto market works and how the players respond accordingly.

Snippet from the announcement of the HUSD solution

The first launch of HUSD allowed for 1:1 interchangeability between PAX, TUSD, USDC, and GUSD. This is a seemingly intuitive rational, as all of these are fiat-backed stablecoins — which means all of them can be exchanged for 1 USD. The intent was to make it easier for people to trade against different, supported stablecoin pairs. What was overlooked in this service is that stablecoins are not interchangeable on a 1:1 basis even though they are equivalent in redeemable value. There are many reasons why they aren’t the same, and all of these contributed to the events that unrolled.

These stablecoins all have different:

• perceived counter party risk
• price fluctuations
• redemption schedules
• KYC/AML procedures
• discounts offered to market makers

What we eventually saw was a situation where various stablecoins issued tokens to people at a certain rebate to improve liquidity. Traders then took advantage of of the situation, using the HUSD solution to redeem and arbitrage for full price. In this situation, the 1:1 ratio between stablecoins offered by HUSD did not hold, and market makers took advantage to earn quick profit. Huobi overlooked this design flaw and it resulted in the proverbial “free lunch” for these traders. If a trading desk were to acquire any sort of stablecoin less than a dollar, they can lock in profit without going directly into the market, which is what they did when these stablecoins were moved around through HUSD on a 1:1 basis. This was discovered and quickly exploited — the longer the HUSD solution was in place, the more this strategy would be utilized. Arbitrage would continue to persist so long as the opportunity is there.

What happened next could have happened to any of the four fiat-backed stablecoin companies, but unfortunately Paxos Standard was the preferred vehicle to convert to fiat for some traders. Paxos Standard does have more lenient standards for redemptions, but they also have regulatory backing from the state of New York. When these traders were looking to move risk off their balance sheet and redeem to fiat through the stablecoin, the company saw huge redemption requests — nearing, at times, almost $20 million in a day. Paxos Standard found this behavior to have serious implications. To manage their requests and uphold compliance standards, they began to inquiry into these activities, which was not looking upon favorably in the crypto trading community. At one point, traders’ Paxos Standard accounts were frozen or closed indefinitely — with some amounts in the millions.

Mildly invasive questions asked by Paxos Standard

Making matters worse was the fact that traders were also circumventing the light KYC/AMC restrictions on Huobi. To work around the $10,000 withdrawal limit imposed by the company, people created many accounts to move more funds than typically allowed.

One of Huobi’s wallets showing PAX outflow under the 10,000 limit. Src: etherscan

In response to all of this activity, Huobi needed to go back to the drawing board to conceptualize a new design for the solution because the original was being taken advantage of. In addition to primarily trading on the Huobi exchange, HUSD was intended to convert between stablecoins. Since Huobi did not account for price differences directly from the market or from the rebates, the exchange itself had to take a loss to maintain their service. Below is an analysis on the total inflows and outflows from Huobi wallets for these stablecoins during the duration of v1.0.

Huobi wallet activities, Src: etherscan

With this view, it’s hard to fully assess what flow activity can be attributed directly towards arbitrage or for HUSD’s intended solution. We can get a clearer depiction of these transfers by breaking down inflows and outflows to reserves owned by Huobi and to wallets not in their control. We can also assume internal transfers are for operational usage while external transfers are derived from activity to arbitrage.

The inflow and outflow numbers for in-between transfers are shown as a sanity check and to avoid double counting

Above is a chart of Huobi’s stablecoin reserves during this time period. As you can see, a majority of the activity occurred with PAX and GUSD. Gemini was bootstrapping their stablecoin and offering discounts to traders to provide liquidity on exchanges. However, these traders simply bypassed the market and swapped GUSD for PAX to lock in quick profits. So instead of having GUSD properly distributed, Huobi actually owns more than 70% of the tokens that are minted. Traders seized the presented opportunity and left Gemini, as well as Huobi, in a less ideal situation to clean up.

After examining the nature of the transactions, we determined that the HUSD solution was abused beyond its intended purpose. The total amount of stablecoins that were sent between Huobi wallets was very minor in comparison to the total amount that was actually moved out. The amount transferred out of Huobi was also larger than the current reserves for each stablecoin, indicating that the primary use case was for arbitrage rather than to keep user funds on the exchange to trade Huobi’s stablecoin markets.

We can estimate that from these breakdowns that, within this short time frame, roughly ~$234 million worth of stablecoins were arbitraged without considering price deviations (possibly up to 3 cents difference based on price discrepancies and rebates), so it’s not a stretch to say that Huobi was on track to lose several million dollars. In addition, from looking at the first table, we see that the Huobi reserves heavily lean towards certain stablecoins as an aftermath which could result in a liquidity imbalance concern.

HUSD is now launching a trial version 2.0 to remove the fixed 1:1 exchange rate, adjusting values between stablecoins based on pricing and various other factors in hopes to provide a meaningful solution. A few key observations arose from the announcement:

• HUSD price is bound based on fiat-crypto markets
• The methodology to price underlying stablecoins is determined by data from mainstream exchanges
• Users have to designate time and amount independently to interchange stablecoins, going from an automatic exchange to a manual one
• The quotes provided by HUSD also account for an extra premium based on the available withdrawal amount on stablecoin positions, and is floored/capped to Huobi’s advantage to prevent possible losses
• Version 2.0 is set to run for one month, where possible improvements can be iterated on

It will be interesting to see how the new version affects trading behavior, activity, and the resulting stablecoin exchange rates on the platform during the trial period. Though we can still make a few comments about v2.0’s design:

• HUSD is not a product that can be transferred out of Huobi, as it only serves as a quote currency specifically for this exchange. Therefore, the product itself has limited utility and benefits when extended to broader markets
• PAX/GUSD/TUSD/USDC are still the only viable options to redeem out of your account. Because there are two offsets (to and from HUSD), one has to compensate when wanting to go from different underlyings. This, coupled with the fact that HUSD’s proposed exchange rate is floored/capped, means that interchanging between two stablecoins with largely different offsets in HUSD may not provide favorable rates on the platform
• The HUSD price itself is explicitly bound to fiat-crypto markets while the underlying stablecoin exchange rates are not; this can be viewed as a way for the company to hedge itself in case of a credit-risk from the supported stablecoins. This can be a negative for users, as there is a premium baked into the use of the platform when depositing or withdrawing stablecoins
• One purpose of HUSD is to access Huobi-specific stablecoin markets, which is not a unique advantage as there are other exchanges with similar pairs and more liquidity. This creates unnecessary time and cost for a trader when other options exist
• Another purpose of the HUSD solution is to interchange stablecoins, though benefits may not be advantageous because the rates are tied to the centralized risk, collateral, and liquidity of Huobi’s reserves. If another alternative is available in the market which can provide better quotes and ease of use, HUSD lacks a competitive use case and could fail in managing both collateral and liquidity without the activity to support it.
• The actual collateral targets for HUSD are unknown to a user, so it’s unclear how Huobi would want to service users in regards to this aspect
• During periods of large volatility, aggregate products have to prove their worth. Managing liquidity risk may present an issue during turbulent markets due to the way pricing is bound under the new design

We can only conjecture how HUSD’s trial version will fare even with these criticisms, and only time will indicate whether the service will be useful without being exploited. Still, there are more open questions in light of this three month time span: What is the structure of the business model for fiat-backed stablecoins and the impact on market participants?; Are there regulatory implications associated with building a more open financial system; What are the pitfalls to providing a service that does not fully consider market implications?; What is the role of liquidity for the markets between seemingly valued instruments?

Over here at Neutral, we’ve noticed issues with stablecoin models/services, and we’re seeking to create something better. The Neutral Dollar is a tradeable stablecoin demonstrating transparency and lower volatility. The token is powered by underlying stablecoin swaps that are liquidity- and exchange-neutral for more efficiency, which means that our solution is more marketable and beneficial to all participants. The Neutral Dollar also benefits from more dynamic liquidity due to its nature of being an aggregated, non-centralized basket. The mechanics of our platform align trading behavior and incentivize collateral to balance, resulting in stability of the product through all market cycles. We view the Neutral Dollar as a more fair and transparent solution for all stablecoin use cases, hoping to popularize the use of these tokens and cryptocurrencies altogether.

— — — —

If you’re interested in learning more about our team and products, check out our website or start a conversation with us on Telegram or Twitter.