Introduction

2022, in the world of crypto, can be characterized by the unraveling of debt and fraud from centralized institutions. Between Terra/Luna, Three Arrows Capital, the collapse of CeFi lenders and the bankruptcy of FTX/Alameda, NAVs of close to $100 Billion were found to be gambled away and commingled.

With FTX/Alameda, the executives were allegedly manipulating the book value of their balance sheets, misappropriating funds and insider trading. Ultimately, a leaked balance sheet by CoinDesk and a sell order from a competitor exposed a $10B hole in FTX and collapsed the house of cards.

Context

Alameda was founded in 2017 as a quantitative crypto trading firm focused on arbitrage opportunities. FTX was started shortly after, in 2019, as a centralized crypto exchange. Both entities were founded by Sam Bankman-Fried (SBF). They became instant success stories as Alameda grew to over $10B in AUM and FTX grew to $30B in user deposits.

SBF quickly became the “golden boy” of crypto, amassing a wealth of over $20B. As a self-proclaimed “effective altruist”, he donated to numerous media publications and politicians. He became the front-man for the SEC’s efforts to regulate crypto and was a main contributor to the DCCPA, a controversial crypto-regulation bill currently proposed in congress.

The Collapse

On November 2nd, 2022, CoinDesk leaked a balance sheet of Alameda’s holdings. The results were alarming. The hedge fund held $14.6B in assets, mostly in illiquid tokens, to $8B in liabilities. Of those assets, $5.82B was in FTT, the made-up, no utility token of FTX. At the time of the leak, the circulating market cap of FTT was $3.4B, significantly less than Alameda’s position. Had Alameda been forced to sell the FTT, it would have caused the price to plummet well below its listed price of $26, meaning its book value did not reflect fair value. Shortly thereafter, it was uncovered that Alameda held or collateralized $802M of Serum (market cap of $280M), $508M of Maps (market cap of $10M) and $76M of Oxygen (market cap of $9M), among others. Put together, these assets had a book value of $7.2B but if they were sold on the open market, they would’ve likely sold for $1-2B. At fair value, it seemed likely that Alameda was insolvent and vulnerable to a margin call. Furthermore, these assets were all invested in by FTX Ventures.

What became abundantly clear was Alameda’s rise to prominence was predicated on an engineering hack that created an illusion of assets out of thin air. The formula was simple: FTX Ventures buys a majority stake (90%+) of a cryptocurrency or creates one in-house, releases a small percentage of it to the public (<10%) and promotes it to cause a buying frenzy. With all the demand concentrated on the small amount of public tokens available, the price of the tokens skyrocket, causing FTXs oversized position to list at a massive book value. FTX would then list this position to Alameda’s balance sheet, who would collateralize it to borrow funds, with FTX as its primary lender. To put it simply, Alameda was heavily leveraged, using illiquid cryptocurrencies FTX made out of thin air as collateral, with user funds on FTX as the primary source of leverage.

Shortly after the leak, Changpeng Zhao (CZ), the CEO of prominent FTX competitor Binance, announced on Twitter that his exchange was looking to sell its stake of FTT (est. $680M) “due to recent revelations that have come to light.” The announcement led to panic selling from FTT holders as they feared the sell pressure would lead to an Alameda liquidation, death spiraling FTT to $0. It also led to a bank run from FTX users as an Alameda liquidation meant the loss of billions of dollars in user funds. On November 8th, user withdrawals on FTX were frozen and FTT fell 80% to $5.20. By November 11th, FTX and Alameda filed for bankruptcy, posting a $10B hole in user deposits.

The $10B Hole

Although it is unclear exactly where the $10B of user deposits went, a general idea is beginning to emerge as more information leaks to the public.

The majority of funds seem to be lost through Alameda’s leveraged exposure throughout the 2022 bear market. As the market continued to go down, Alameda continued to “borrow” more user deposits to post margin on their loans, amounting to a hole of at least $4B.

The remainder seems to be lost, not through investing, but through commingling deposits for personal and business expenses. The expenses include but are not limited to $121M in personal real estate, $70M in political donations, a $1.4B acquisition of Voyager and an unsecured $1B loan to SBF himself.

Further Details and Evidence

As journalists investigate the situation, the alleged evidence to fraudulent activity accumulates. For example, it was found that FTX’s terms of service forbids the investing of user deposits. Section 8.2.6 states, “Title to your Digital Assets shall at all times remain with you… FTX Trading does not represent or treat Digital Assets in User’s Accounts as belonging to FTX Trading”. On November 7th, during the bank-run, SBF supported these terms in a tweet, stating “We don’t invest client assets”. We now know this wasn’t true.

It was always expressed by SBF that FTX and Alameda were separate entities and Alameda did not have special rights to using the FTX platform. It’s now known that Alameda had access to all real-time trading data and received liquidation exceptions. This is similar to playing poker while knowing your opponent’s hand. Furthermore, Alameda processed a significant portion of user deposits on behalf of FTX and often never sent them to the exchange, but instead kept them to be traded. They also joint-filed for bankruptcy.

Executives of FTX/Alameda led the public and investors to believe that the two companies were highly profitable. Renowned investment firms such as Sequoia Capital and Tiger Global invested hundreds of millions into the company noting its profitability. Evidence now shows that in 2021, when the crypto market hit all time highs, FTX had actually lost $3B.

The intent of SBF’s altruistic efforts may have been equally deceptive. It is now theorized that his donations to the media and politicians were attempts to control the public narrative and receive regulatory exemptions. His claims of “effective altruism” may have been attempts to obscure his commingling of deposits for personal use. Lastly, a draft of the DCCPA leaked in early November and seemed to call for stringent regulation on decentralized exchanges while seeking regulatory exemptions for FTX, further highlighting potential nefarious behavior.

** **

Part 2: The Implications and Moving Forward

Contagion

Because of FTX’s influence on the industry, its collapse is causing a ripple effect across centralized institutions. FTX was a popular custodian for institutions, who are now having to post losses on their funds. One of the largest hedge funds in the space, Multicoin, posted a 55% loss during these events. Furthermore, many of the largest lenders in the space provided loans to Alameda and are now risking insolvency. BlockFi, a retail-oriented crypto lender, had lent $671M to Alameda and deposited $355M on FTX. After already reporting solvency issues in Q2, they filed for bankruptcy on November 28th. Most notably, Genesis, crypto’s predominant prime brokerage firm, had $175M locked on FTX and supplied loans to Alameda, although the amount in loans is unknown. Genesis has frozen withdrawals and is expected to be illiquid. Its parent company, Digital Currency Group (DCG), is also the parent company of Grayscale, the suppliers of GBTC. DCG is expected to undergo a restructure to fulfill liquidity needs.

With decentralized applications, the contagion is fortunately less drastic. The crypto market cap fell 23% from $1.03T on November 5th to $790B on November 9th. It’s since stabilized to $850B. The Solana ecosystem saw the largest drawdowns as FTX was a prominent contributor. Solana’s coin, Sol, fell almost 60% and its total value locked decreased from $1B to $280M. Many of Solana’s wrapped assets, such as Sollet Wrapped Bitcoin, were in the custody of FTX and are now trading at cents on the dollar. Many smaller applications, such as Ren Protocol, used FTX to custody their treasuries and are unlikely to retrieve them. However, outside of Solana and some smaller applications, the decentralized sector was fairly immune with limited capitulation or bad debt.

The importance of blockchain and decentralization

The numerous black swans, FTX being the most prominent, within the digital asset space have damaged the industry’s reputation and have hurt many investors, retail and institutional alike. However, digital assets and decentralized protocols as technologies once again proved their resiliency. The second largest crypto exchange and one of the largest crypto hedge funds collapsed, leaving a hole larger than 1% the industry’s market cap. The contagion spread to the other institutions in the space, leading to further insolvencies. Yet, decentralized ecosystems remained intact.

Regarding regulation, the FTX debacle was not a product of bad regulation, but a product of poor transparency. It didn’t take the SEC or an auditor to strongarm FTX into releasing their financials. All it took was their balance sheet to be leaked on Twitter. From there, crypto influencers conducted investigative journalism to uncover the truth, leading to its demise. Had the balance sheets of both companies been publicly available and updated real-time, the entire scheme would have been uncovered well before it became an empire.

One of the primary value propositions of blockchain technology is its transparency. It has the ability to, in real-time, prove a company’s solvency and use of funds. Had FTX been required to custody user deposits on-chain, on publicly known wallets, any commingling of funds would’ve been posted on a blockchain explorer for users to immediately catch. Due to the events in November, exchanges, such as Binance, are beginning to post more of their finances on chain to verify their proper use of funds. Blockchain technology offers regulatory oversight that far exceeds current regulations, and should be considered as a regulator’s friend, not enemy.

As blockchain technology continues to develop, the expectation is that more of the financial world moves on-chain for operational and compliance purposes. Regulating institutions will be more feasible. Transferring and verifying assets will be instantaneous and immutable. And custodying assets will be unmediated and trustless. The recent events only emphasize the importance of blockchains, and with proper regulation, can accelerate its adoption.

Passing proper regulation that uses the regulatory benefits of blockchains and encourages onshore innovation will require an in-depth understanding of the technology. If regulators work with developers and allow them to experiment, uncovering its long-term potential, regulation that best protects consumers and investors while encouraging innovation will emerge. The DCCPA, in its current state, written by SBF, fails in that regard.

Arana Ventures

Arana Ventures was founded with the belief that decentralized systems have the potential to provide immutability and security at an unprecedented level, and after these events, our premise remains stronger than ever. We are continuing to roll out our fund of funds and look forward to releasing a direct investment fund in late 2023.

For our fund of funds, we will be doing additional due diligence to understand how portfolio funds custody their assets and protect themselves from vulnerable points of centralization. Most of the funds we’re looking to work with either use decentralized services or have diversified exposure to centralized entities.

For our direct fund, we currently have internal capital we’re testing with. Because we conduct all of our investments directly on-chain, our exposure to the collapse was limited and we outperformed beta. Prior to the crash, we had 50% of our assets in cash or cash equivalents. We are now 100% long as the market’s resiliency has impressed us.

Note to Potential LPs

Overall, of all the funds we’re in contact with, the ones with the most exposure seem to be the net-neutral arbitrageurs and market makers. These funds depend on centralized exchanges and lenders to maximize their opportunities and drive “safe”, consistent returns. The highly volatile, net long funds tend to utilize decentralization the most and, therefore, had the least exposure.

Throughout the bear market, we have seen increased interest from investors to work with the “safer” net-neutral funds and capitalize on the inefficiencies of the digital asset space without the risk. The current situation is a reminder that this industry, in its adolescence, does not favor risk aversion. Investors of this space, including prospective LPs of Arana, may not want to invest with the intent of earning consistent returns. Instead, they may want to invest a moderate amount of capital with the goal of generational upside but the risk of losing their investment altogether. **Not Financial Advice**

Conclusion

The FTX/Alameda collapse, as alleged, is one of the greatest examples of fraud in the modern day. It has financially affected millions of people and damaged the reputation of crypto technology. However, the digital asset space has once again proven its resiliency and should be viewed as the future of financial regulation. With the right leadership, blockchain technology and crypto can shape a more transparent, self-sovereign future. The recent events have not undermined that vision, but strengthened it.

FYI, just because institutional whales are shorting Tether doesn’t mean it’s worth shorting. DOYR, read their attestations, etc. Also, this is not financial advice, just a fun little shorting procedure I found that could pay dividends down the road.

How shorting works

Let’s begin by giving a quick overview of how shorting a position works.

You borrow 100 units of stock/commodity X at $1 per unit.

You then sell the 100 units of X for $100 in the open market.

You now have $100 and you owe your lender 100 units of X.

If X goes down in price, you can then repurchase 100 units of X for less than $100. For example, if X goes down 20% to $0.80, you can repurchase 100 units for $80.

You then pay back the 100 units of X to your lender and pocket the $20.

Just like that, you just made $20 by betting against an asset.

How does this work off-chain?

Shorting off-chain with leverage is incredibly simple as many CEXs offer margin trading or leveraged tokens that do the heavy lifting for you.

FTX, for example, allows you to buy a 3x short position by purchasing an ERC-20 token called USDTBEAR. With this, you can open a $1000 short position by buying $333.34 of the token. The token itself does all the borrowing and swapping mentioned earlier for you.

If USDT falls to $0.50, you earn $500, but if USDT rises to $1.33, you lose your position altogether. In the case of a large market cap stablecoin like USDT, hitting $1.33 is highly unlikely so, in my opinion, it looks like a solid position.

But this method carries a prominent risk: you’re using a centralized exchange. If the past couple of months have taught us anything, it’s that if “it’s not your keys, it’s not your coins.” FTX, for example, has a history of questionable behavior with inconsistent liquidations and market manipulation all on the retail investor’s dime.

What about on-chain?

Leveraged shorting on-chain gets a bit more complicated as everything needs to be done manually. In order to open and close positions, you need a lending protocol like Aave and a DEX like Uniswap. When shorting pegged assets like stablecoins, I’d recommend using Aave V3 as you can take advantage of their high efficiency mode (e-mode), allowing for a 97.5% liquidation threshold and 97% LTV.

*Note: High efficiency mode, or e-mode, is a feature on Aave V3 where pegged/backed assets like stablecoins can be loaned and borrowed at significantly higher thresholds.

To keep things simple, let’s look at how you can short $1000 of USDT without leverage:

1. Collateralize 1100 USDC ($1100) on Aave V3

2. Borrow 1000 USDT ($1000) against that using e-mode

3. Go on Uniswap and swap the 1000 USDT for 1000 USDC

4. Store the 1000 USDC in your wallet

If tether depegs to $0.50, you simply buy back the 1000 USDT using 500 of the 1000 USDC you received earlier, pay back the loan and withdraw the 1100 USDC. You now have 1600 USDC, which is a $500 profit!

Now, let’s short with leverage

Creating an on-chain short position with leverage takes the example we just went through and adds a layer of complexity.

You may have noticed that after swapping the 1000 USDT for 1000 USDC, the USDC was just sitting in your wallet. What a lot of people do to create a leveraged position is put that USDC back into Aave as collateral to borrow even more USDT.

Let’s take the previous USDT example, but this time you only collateralize 100 USDC. You can then borrow up to 97 USDT (97% LTV) and swap that to 97 USDC. You collateralize the 97 USDC, allowing you to borrow another 94 USDT. You swap the 94 USDT for USDC, collateralize that and so on. If you do this a total of 13 times, you can short $1057 of USDT with $1157 in collateral, all from your original collateral of $100.

Let’s take the previous USDT example, but this time you only collateralize 100 USDC:

1. Collateralize 100 USDC ($100) on Aave V3

2. Borrow 97 USDT ($97) against that using e-mode

3. Go on Uniswap and swap the 97 USDT for 97 USDC

4. Collateralize the 97 USDC on Aave V3

5. Borrow 94 USDT against that using e-mode

6. Go on Uniswap and swap the 94 USDT for 94 USDC

7. Continue the loop…

If you follow the above loop another 11 times, for a total of 13 loops, you can short $1057 of USDT all from your original $100. You will have 1157 USDC as collateral and 1057 USDT as your debt obligation.

However, there are some serious problems with this method. For starters, opening this position requires 39 transactions. The gas fees, effort and time it would take to go through this is simply ridiculous. In addition, it would be vulnerable to MEV bots and price changes.

But the real issue is closing the position. In the scenario where USDT doesn’t depeg, closing the position would require undoing the circular collateralization. And at a 97.5% liquidation threshold, you can only withdraw a small portion of collateral at a time.

In the above scenario, the health factor is 1.067. Let’s say as you undo your position, you only withdraw as much at a time without reducing the health factor below 1.02 (already a dangerously low number). This would take a minimum of 15 loops, or 45 transactions. The risk of being liquidated by price fluctuations during this unwind is very high.

Fortunately for you, this short position is between two stablecoins. Imagine doing this with non-stable tokens where e-mode is unavailable and a measly 1.02 HF can liquidate you in a moment’s notice. Sure, you can set your HFs higher, maybe having them start at a 1.3 with a floor of 1.1. But you’re still undertaking massive risk just to close a position, not to mention your leverage capabilities would be limited to 2.1x.

Long story short, don’t do this. It’s one of the most common ways people get liquidated. It’s stupidly risky, a pain in the ass to deal with, expensive and time consuming.

Now that we’ve gotten all of this out of the way, let’s get to the good stuff.

TUDS - the way shorting should be done

TUDS, or Trustless Undercollateralized Decentralized Shorting, is an on-chain shorting method our investment firm, Arana Ventures, came up with. It’s a simple method that uses a flashloan to provide upfront collateral. It takes 5 transactions to open and close any position, and with the use of a batching protocol, it feels like 1.

Let’s once again use the example of shorting $1000 of Tether using $100 of USDC. How it works is the following:

1. Borrow a flashloan of 1000 USDC ($1000) from Aave

2. Deposit the 1000 USDC + 100 USDC of your own as collateral on Aave (1100 USDC total)

3. Use e-mode to borrow 1000 USDT ($1000) against the collateral

4. Go on Uniswap and swap the 1000 USDT for 1000 USDC

5. Return the 1000 USDC to the flashloan

In the end, you effectively have a debt obligation of 1000 USDT collateralized by 1100 USDC, of which only 100 USDC came from your wallet. That’s another way of saying you’ve shorted $1000 of USDT from only $100 of USDC.

Because you’re using a flashloan, you need to use a batching protocol to complete the entire process within the same block. To do this, we use Furucombo, a drag and drop batching protocol. The picture below illustrates how this works.

As you can see on the top left, I’ve provided $100.73 of collateral and will receive 1100 aUSDC, which indicates that I have 1100 USDC locked up on Aave. On the right side, the boxes outline the process I described earlier and show a 1002 USDT short position. The reason for the additional 2 USDT is due to a slight price discrepancy between the two stablecoins.

A few weeks ago, we tried our first test case, opening a $50 USDT short position with $25 USDC on Aave V2 using ethereum. After I submitted the batched transaction, it showed up on our Aave account within a few seconds.

Had I used V3, I would’ve been able to spend as little as $5 to open this same position.

In total, opening a short position using TUDS takes 5 transactions, 6 if you include the return funds, and the process all happens in one block. This allows for CEX level speed and simplicity while preventing the risk of price fluctuations or MEV. The position was also opened completely on-chain with self-custodied funds.

Now that we’ve covered how to open a short position, let’s go over how to close one.

Essentially, all you do is the same 5 step process but backwards. Using the 1000 USDT example, in the event it does not depeg, closing a position would look like this:

1. Borrow a flashloan of 1000 USDT ($1000) from Aave

2. Repay the 1000 USDT loan from the short position

3. Withdraw your 1100 USDC ($1100) in collateral

4. Go on Uniswap and swap 1000 of the 1100 USDC for 1000 USDT

5. Return the 1000 USDT to the flashloan and walk away with 100 USDC (p/l of $0)

Because USDT didn’t change in price, you walk away with the original 100 USDC you invested when opening the position, giving you a p/l of 0. The below picture illustrates how this works.

In the event USDT depegs to $0.50, closing a position would entail the following:

1. Borrow a flashloan of 1000 USDT ($500) from Aave

2. Repay the 1000 USDT loan from the short position

3. Withdraw your 1100 USDC ($1100) in collateral

4. Go on Uniswap and swap 500 of the 1100 USDC for 1000 USDT

5. Return the 1000 USDT to the flashloan and walk away with 600 USDC (p/l +600%)

Because USDT went down by 50%, you now exit your position with 600 USDC instead of 100 USDC. That’s a $500 profit from a $100 investment!

Do you see how much simpler this is than the circular method of leverage? It’s 5 transactions that happen within 1 block while leaving the health factor completely untouched. And you can customize this to whatever level of leverage you’d like. 2x, 5x, 8x, 15x, and so on.

Plus, the TUDS method operates on a level of simplicity that rivals centralized exchanges but with increased flexibility and self-sovereignty. There’s no all-knowing, all-seeing SBF  👁👁 randomly liquidating you because he wants to buy the dip during his mid-morning shit.

TUDS Limitations

Despite being the best way to short cryptos on leverage, TUDS has its limitations.

For starters, it’s simpler than the circulating method, but it’s still a bit tricky to understand and requires a level of comfort with DeFi protocols. If I had to guess, a good portion of you are still trying to wrap your heads around it. However, once you get the hang of it, this method becomes fairly straightforward.

Secondly, there are limitations to the amount of leverage you can provide. If you use e-mode, which requires lending and borrowing between assets that are pegged to one another, you can achieve up to 32x leverage. However, this gives a health factor of 1.005, which is laughably low. Without e-mode, the max leverage depends on your collateral. But with relatively “safe” and liquid assets like ETH and USDC, you can expect a 5-6x max leverage with a ~1.03 HF, which is also dangerously low.

At Arana, we have a 1.1 HF minimum when shorting stablecoins w e-mode and a 1.3 HF minimum when shorting non-stablecoins without e-mode. This gives us a max leverage of 7.8x and ~2.1x respectively, which is still competitive to FTX’s 3x leveraged tokens.

Probably the biggest limitation has to do with the chains you can use and the assets you can short. Because Aave V3 mostly only uses the L2 chains, and Furucombo uses only Polygon and Ethereum, the only way to create short positions with e-mode is through the Polygon chain. Without e-mode, you can use Aave V2 with the Ethereum chain. But you’re still limited to two chains and can only collateralize and short the assets Aave recognizes.

Lastly, the variable interest rates on Aave, or for any DeFi lending protocol, can get pretty extreme. For the most part, rates are actually pretty low, usually between 0.5-5%. This is miles better than FTX’s 10.95%. However, if supply goes down and/or borrows go up drastically due to some anomaly (see solend whale), that 0.85% interest can quickly turn into 85%. In the short term, this really isn’t a problem. But if you’re not paying attention, you could end up accumulating insane interest payments over the matter of a few months.

The potential impact of TUDS

Even with its limitations, the long-term viability of TUDS is close to limitless.

Finance is incomplete without a way to create leveraged short positions, and DeFi is incomplete if those positions can’t be safely created on chain. Without shorting, we wouldn’t be able to curb overvalued assets, provide hedging opportunities or shield commodity providers from market volatility.

This is especially true in the crypto world, as during bull markets, valuations are unjustifiably high, whales are overleveraged and commodity providers, such as miners, hold treasuries with only the commodity they’re providing. This leads to a massive bull run followed by a nightmarish crash into a bear market. Sound familiar?

With a strong crypto shorting market, the bull runs will stay within the bounds of reason, leading to smoother transitions into the bear markets. Overvalued assets are shorted, gradually bringing them back to justifiable prices. Retail investors can be hedged from whale shenanigans. And miners, instead of dumping their treasuries at the least profitable price, further hurting the market, can exit their profitable short positions.

TUDS is the first way to create short positions efficiently, trustlessly and safely. The infrastructure to simplify the process and expand its usability is not there yet, but my hope is that a better developer than myself sees the potential in this and finds a way to make TUDS accessible to everyone. If you think that person could be you, feel free to email me at parth@arana.ventures, and I’d be happy to hop on a call.

Have a great day,

Parth