Introduction
END Capital facilitates access to “real world assets” (“RWA”) for DeFi investors/lenders. As opposed to financing more volatile crypto priced assets, we are focused on financing real world productive GDP loans using DeFi. We have in the past, and are currently, operating numerous mechanisms to facilitate access to RWA, including setting up and operating RWA access on the Centrifuge platform, as well as the real world asset market (“RWAM”) on the Aave platform. We also currently manage, operate and fund a specific pool of assets on the Centrifuge platform, called the “GigPool”, which finances companies who provide short-term lending products to ‘gig’ economy workers. We would utilize a TrueFi capital pool to expand the amount of GigPool funding, and intend to scale up by onboarding additional real world borrowers who finance the gig economy. We have a robust pipeline of asset originators which would require additional capacity in the GigPool, as well as provide diversification of a range of different credits within the pool. We have numerous other asset classes beyond the gig economy thesis that we are planning to expand into as well, but for our inaugural pool we will initially be focused on expanding the GigPool.
Who is managing the fund
Our team is comprised of a combination of software engineers who are deeply involved in the crypto space, and traditional finance (“TradFi”) team members who have >40 years of experience in originating, structuring, underwriting and servicing structured finance loans and securitized bonds for private credit funds (with AUM ranging from $400M - $5B). The range of assets which our structured finance team has transacted in is very broad, including everything from mainstream consumer credit (eg subprime auto, mortgages, unsecured lending, etc.), merchant cash advances, content/IP financing (film/music receivables), Amazon merchant financing, litigation finance, short term worker financing (‘gig’ economy) and many other types of assets. On deals which our team has been involved with originating, executing and closing, we have never had a default on a single loan, while generating low-teens returns for our investors. We have achieved this by utilizing extensive structuring and financial modeling expertise to stress test underlying default rates to understand at what point our loans become impaired and create structures which then mitigate that impairment if default rates increase (including early warning asset performance triggers, financial covenants, cash controls, turbo structures and other protective mechanisms).
What we do
The gig economy is a very large TAM of $1.5T and growing rapidly due to macroeconomic forces. There are many different types of gig workers, including delivery providers, ride sharing, home sharing, online merchants, content creators, consultants, and freelancers in many other sectors of the economy. Oftentimes these workers need short term financing to help them manage their cashflow needs. Thus the number and types of companies which are providing financing to gig economy is large and growing as well. The financing products are generally relatively high-yielding, short duration, self-liquidating assets, which lend themselves well to a TrueFi capital pool. We believe that with our current pipeline, we could onboard $50M of financing onto the TrueFi platform within the next 12 months. As mentioned above, due to our TradFi team’s expertise in underwriting and structuring loans in many different asset classes we envision expanding beyond the GigPool, but initially would focus on scaling up the GigPool.
What have you done?
As mentioned earlier, our TradFi team has in the past managed private credit funds on behalf of institutional LPs, generating low-teens returns on dozens of loans, with zero defaults. In DeFi, our current GigPool size is ~$2.5M and has generated an average levered cash on cash return of 12%-17%. We expect the pool size to grow up to ~$20M with an average cash on cash return of ~15%, based on current onboarding discussions with additional originators.
Portfolio description
We initially envision raising $5M in capital, to be immediately deployed into our GigPool. Generally speaking, we would envision using USDC for a term of 6 months and an average interest rate of 10%.
Transaction structure
We plan to utilize TrueFi capital to deploy into the mezzanine tranche of our GigPool. As currently structured, all Asset Originators into our GigPool are required to contribute first loss equity into our loans with them (currently approximately 20% of the underlying collateral asset value). This equity contribution by the Asset Originators aligns their interests with all lenders and will be subordinate to investors in this TrueFi pool. All TrueFi capital providers would receive principal and interest payments before any capital gets returned back to the Asset Originator. Also as discussed previously, our TradFi team is actively managing the risk through time-tested structural features, such as the incorporation of early warning triggers and covenants into our loans such that payment of principal and interest gets accelerated to TrueFi capital providers upon asset performance deterioration, such that our TrueFi pool is protected from deploying any new capital into an asset deterioration scenario (eg a higher risk of repayment scenario).
Glad to see more borrowers expanding into real world assets. You mentioned your DeFi AUM 2.5 million. What would you estimate your current AUM is outside of DeFi?
Thanks for your question. We are currently only aggregating assets using DeFi, so no other assets outside of DeFi currently. In the past, members of our team have managed portfolios at private credit funds (TradFi) in the range of between $400M - $5B.
In terms of current on-chain assets, yes. The amount of assets in the GigPool continues to grow monthly, though, and we expect it to quickly ramp up, in particular as we add more asset originators into the pool. Additionally, we have in the past originated and facilitated on-chain assets in aggregate of ~$15M aum in a range of different asset pools on Centrifuge.
My name is Jeremy, and I am a part of the END Capital team.
To clarify a few points - the way we will use the capital from this pool is a bit of a departure from how crypto trading shops are using the funds. Rather than funding an overall trading strategy, the capital that the investors in this pool will be mapped dollar for USDC to an off-chain fully secured financial instrument (either a note or a loan).
Currently, the notes that we’re originating and funding through DeFi have been overcollateralized by at least 20%, along with an additional cash reserve of 50%.
So what this means for the investors in our pool, is that each USDC that we borrow and deploy from this pool, will be overcollateralized by at least $1.25 worth of off-chain loan collateral assets, and currently an additional cash reserve from our borrower of $0.50 for a total of $1.75 worth of loan collateral assets and cash ensuring the principal of that USDC.
In other words, once we fully deploy the $5mm from this pool, there will be at least $6.25mm of yield bearing off-chain financial assets, and $2.5mm of cash reserves from our borrower to ensure that the principal amounts from the TrueFi investors are protected.
I will caveat that as we onboard larger fintech borrowers and/or our existing borrower’s origination quality improves, the overcollateralization and cash reserve structures may decrease, but never to the point of being undercollateralized (and most likely still remain overcollateralized with a safety buffer).
On the flipside - if our existing borrower(s)’ origination quality deteriorates, then there are structural features in the notes/loans we are structuring that not only provide early warning signs, but also reroutes cash flows to pay down the principal (i.e. derisk quickly). Then, subsequent notes/loans facing that borrower, if we decide to continue with this relationship, would be structured with higher overcollateralization and cash reserve levels.
Happy to clarify further if you have any additional questions!
@jeremykim83 thanks for this note. Can you please elaborate on the following questions:
Are you collateralizing with the real world assets themselves, or their future revenue streams? Is the borrower defaults, is there claim against the actual real world assets?
If yes, what is the valuation model used to price these real world assets? How are these assets being posted as collateral?
Also, can you please help me understand the “real world oracle” process here to determine changes in valuations of collateral, whether it’s the asset itself or the probability of those future revenue streams coming to fruition? Feels important in a macro climate like this.
Finally, just making sure I understand correctly. If I want to take out a loan of $1M, I have to post $1.2M of real world assets in addition to $500k of straight cash?
Each loan drawn from the TrueFi pool will be a levered tokenized version of a loan (“Loan A”) to a company that originates loans (“Loans B”) to its customers. Loan A is collateralized by a pool of the Loans B that are originated by the company. So in essence, the answer to this question is both: the Loan A (and thus each loan drawn from the TrueFi pool) will be collateralized by real world financial assets (Loans B), but the cash flows would be dependent on the future cash flow streams generated from Loans B. We’re finalizing a visualization of what this structure looks like, and will update our post shortly.
Yes, but this would be a very Draconian scenario (where “borrower” is the company borrowing from Loan A), as it would require bringing in a third party servicer to collect on the pool of Loans B collateralizing the Loan A. There are other structural features in Loan A that would entail a significant de-risking (i.e. paydowns on Loan A, which flow through as paybacks to the TrueFi lenders) before arriving at such a scenario. These structural features come in the form of protective covenants and performance triggers, but most importantly a borrowing base concept that essentially forces the company borrowing from Loan A to initiate paydowns before defaults on the pool of Loans B gets too high.
The END Capital team receives from the borrower company, on a monthly basis, a data feed representing all the loans (Loans B) that they’ve originated, and any collections made on all loans originated. With this data (and possibly other external data, if the company’s history is limited), we determine what the collection rates and default rates for each categorization of loans that this company originates. Based on this data, we adjust the structure to maintain a certain level of credit protection (my response to rattlecage represents the current structure we have in place).
With our current company borrower, the loan collateral (Loans B), along with the cash reserve, are pledged. As we shift towards larger transactions, we would be exploring an SPV structure and a true-sale of collateral (Loans B), though this would entail significant legal costs (SPV creation docs, DACAs, servicing agreements, trustees, etc) - hence, this is a structure we plan on exploring when we start pursuing larger transactions (which should be very soon).
As I mentioned in my response to the valuation question above, we review the company’s originations on a monthly basis. For TrueFi lenders, we intend on producing monthly reports compiling the highlights of our analysis, distributed to lenders via a data room. On a side note - we do have a technology team taking learnings/insights from the work that END Capital does, to bring this process on-chain.
This is correct, assuming in this example that you are an early stage fintech lender. This structure is quite common when it comes to private credit lending facing early stage fintech lenders (i.e. post seed, series a, series b) with limited and/or volatile origination performance history. These companies will typically have large slugs of capital from their latest raise, a significant portion of which is earmarked for originating loans. Naturally, as the company matures and their loan performance history becomes more established and more importantly, becomes more stable, the cash reserve portion would decrease. The terms of someone borrowing from one of these companies would be based on the actual loan product that these companies offer (not based on the deal structure we have in place with the company). Currently the GigPool is working with one company at the moment, but is in discussions with two others; for the one company that has already been onboarded onto the GigPool, a borrower would present this company an invoice for services rendered, and the company would provide a short term advance on the face value of that invoice (at a discount), and would collect on the full invoice.
Also, @Codeknight, @Saltines123, and to the broader TrueFi community: we will be hosting an AMA next Monday at 3pm est to go into more detail on our short term and long term strategy with the TrueFi pools, and to field any questions from the community. We will share the Zoom link for the AMA shortly.
Topic: END Capital TrueFi AMA
Time: Jun 13, 2022 03:00 PM Eastern Time (US and Canada)
Join Zoom Meeting
Meeting ID: 893 3727 3451 Passcode: 204023
One tap mobile
+16468769923,89337273451#,*204023# US (New York)
+13126266799,89337273451#,*204023# US (Chicago)
Dial by your location
+1 646 876 9923 US (New York)
+1 312 626 6799 US (Chicago)
+1 301 715 8592 US (Washington DC)
+1 346 248 7799 US (Houston)
+1 408 638 0968 US (San Jose)
+1 669 900 6833 US (San Jose)
+1 253 215 8782 US (Tacoma) Meeting ID: 893 3727 3451 Passcode: 204023
Find your local number: Zoom International Dial-in Numbers - Zoom
As an Originator, it sound like the cost of debt would be in the low teens. What is the typical APR for a Borrower?
In addition to the high APR, the Borrower would need to post equity collateral. Where in the risk curve would you say a typical Borrower lies? Another way to put it, why wouldn’t a gig worker seek TradFi borrowing at a much lower APR if they had the means/credit worthiness to do so?
The collateral being posted is the ‘Loan B’ to the Borrower from Originator. Does this represent security over the cashflows from the creditor to the Borrower? e.g. are we taking credit risk on Spotify for royalty payments or Uber
Can you provide an example of a covenant/trigger mechanism that provides structural protection prior to security enforcement? You mentioned a borrowing case concept.
Can you step through the mechanism and priority of payments in a default scenario? My understanding is as follows:
END provides Loan A to Originator
Originator triggers covenant due to non-performance by Borrower
Originator required to repay Loan A based on agreed profile/structure until back in compliance
If breach cannot be remedied, Borrower 1st loss position is enforced by END to repay Super Senior first, then TrueFi mezz
Security over Loan B enforced and used to repay remaining TrueFi mezz position
Not quite - for the Originator, the cost of debt is lower than low teens (closer to mid-to-high single digits). The reason why this we’re targeting a higher yield on this TrueFi pool is that the lenders in this TrueFi pool will be in a levered senior position. If you reference the diagram in one of my prior responses (and the AMA recording as well), the weighted average interest rate on the Super Senior Financing and the Levered Senior Financing (i.e. the TrueFi Pool) will be the cost of debt to the Originator.
The Originator then turns around, and should logically lend at a higher rate than their cost of debt/capital. For the Originator that we are currently working with (i.e. invoice factoring for social media influencers), the Originator effectively purchases invoices for marketing services rendered, at a discount to the face value of the invoice (i.e. they pay $96 for a $100 invoice, then collect the full $100).
So there may be some confusion, since we lend to a company that in turn lends to a borrower (so the company is both a borrower and a lender). I’ll address the second question in this block first - with regards to the Originator we are currently working with (invoice factoring), the gig worker would provide marketing services to a company, and would send them an invoice. That company may take anywhere between 30-120 days to process that invoice internally before sending out that payment. The Originator basically provides the gig worker on day 1, the opportunity to receive nearly the full amount of that invoice. As you mentioned, if they have savings and/or access to other loans/credit cards etc to tide them over, perhaps they wouldn’t need this service; however, traditional banks and lending institutions typically struggle with properly pricing these services for 1099/gig workers, not because they don’t have the capacity, but rather, the market for non-1099 workers is so large, that it is simply easier to categorize these borrowers in a higher risk bucket, and call it a day.
In terms of where these borrowers would sit in the risk curve - given that the services are already rendered, and that the counterparty to the invoices are typically companies (some large, some small), the best answer is that they sit at various points in the risk curve, with the expectation that the pricing will account for where each borrower sits on that risk curve. However, when talking about where the Originator sits on the risk curve - the Originator, being an early stage fintech lender, is by nature fairly risky. Early stage lenders in general are still working out the best approach to their underwriting models - not just optimizing output given input, but actually integrating that model into their business practices without drastically changing their customer base (i.e. risk of negative selection). For early stage fintechs, we do target those fintech lenders that have raised at least a Series A, as they would have a large slug of capital waiting to be deployed (not just on hiring, but also to originate loans). So while that capital sits, we can apply that capital as collateral support to our debt instrument. Moreover, as a lender to the fintech lender, we can gently (or, not so gently) nudge them towards originating the types of loans that we see a higher probability of success (i.e. profitability given realized losses). These sort of structural features in the financial instrument (i.e. the cash reserve, the borrowing base concept, etc) would protect both the super senior and levered senior lenders, so even if the early stage fintech lender is fairly risky, the debt instrument itself is very much protected.
Typically, there is a true sale of the collateral assets into an SPV, so that there is indeed a security over the cash flows of those assets. There is typically also a DACA (Depository Account Control Agreement) that allows the lender to effectively enforce that security. However, specific to the Originator we are working with and the size of the financial instruments, the collateral is pledged (so remains on the Originator’s balance sheet), along with a corporate guarantee and the cash reserve. As we (as END Labs) scale and are able to provide larger amounts of capital, we intend to institutionalize our current Originator. Additionally, we would on-board larger, later-stage fintech lenders, where we can implement these structural procedures early on.
As for what credit risk TrueFi lenders are exposed to - basically, the risk that the realized cash flows of the underlying collateral pools, along with the additional structural protections (including the cash reserve and corporate guarantee), are not enough to fully pay out on the contractual obligations of that financial instrument. When looking specifically only at the collateral pool - the risk is actually on both the creditor to the borrower, as well as the borrower themselves (since there is a risk that the borrower provides erroneous or falsified invoices).
The borrowing base concept is a very strong source of structural protection, but I will go over this in my response to your last question. In the meantime - common triggers are:
collections based triggers - has the collateral pool been collecting slower than we expected? then we de-risk.
static pool collections triggers - has the Originator been playing fast and loose with their recent underwriting? then we de-risk
default/dq based triggers - have defaults or dqs been higher than expected, then we de-risk
any others - typically there are other triggers in place taking into account the collateral types, as well as the growth strategy of the Originator.
Your understanding is slightly off, so I will try to clarify through several different example scenarios:
Scenario 1: Losses are slightly higher than expected
END provides Loan A to Originator at an 80% advance rate with expected losses at 10%
END reviews the Originator’s loan data, and notices delinquencies are coming in higher than historical trends. No covenants/triggers tripped yet, but delinquent loans are kicked out of the borrowing base, so the Originator must post either additional cash or additional newly originated loans to account for the removal of the delinquent loans from the borrowing base. Any cash flows/recoveries from these delinquent loans are still passed through the financial instrument in accordance with the distribution waterfall. However, the Originator in this instance decides to post additional collateral.
END has a brief chat with the Originator on this trend to see if the Originator shifted their growth strategy and/or collections procedures.
Realized Losses on the underlying collateral pool come in at 12%. No covenants/triggers tripped.
END takes a deeper look into the data to see what was driving the increase in losses.
END has a thorough discussion with the Originator on the findings from our data deep dive, and allows the Originator to clarify/explain any shortcomings in their underwriting/execution.
Originator receives any excess cash flows after meeting all debt obligations (both interest and principal) of the financial instrument.
END decides to continue lending, but with caution, and explores lowering the advance rate to account for recent performance, especially if we see the trends continuing.
OVERVIEW/RESULTS: The TF lenders in this scenario suffer no change in expected returns, since the overcollateralization was enough to make us (END and the super senior and levered senior lenders) whole.
Scenario 2: Losses are much higher than expected
END provides Loan A to Originator at an 80% advance rate with expected losses at 10%
END reviews the Originator’s loan data, and notices delinquencies are coming in higher than historical trends. No covenants/triggers tripped yet, but delinquent loans are kicked out of the borrowing base, so the Originator must post either additional cash or additional newly originated loans to account for the removal of the delinquent loans from the borrowing base. Any cash flows/recoveries from these delinquent loans are still passed through the financial instrument in accordance with the distribution waterfall. However, the Originator in this instance decides to post additional collateral.
END has a brief chat with the Originator on this trend to see if the Originator shifted their growth strategy and/or collections procedures.
Realized Losses on the underlying collateral pool come in at 17%. Loss covenant trigger tripped (i.e. 15% loss trigger).
END takes a deeper look into the data to see what was driving the increase in losses.
END has a thorough discussion with the Originator on the findings from our data deep dive, and allows the Originator to clarify/explain any shortcomings in their underwriting/execution.
Originator receives no excess cash flows (given the trigger breach), and all excess cash flows are applied towards paying down the principal balance of the financial instrument (first the super senior, then the levered senior).
END decides to pause lending, allowing the principal balance of the financial instrument to pay down. During the pause, END will work with the Originator on amending the structure of the financial instrument to increase structural protections (tightening concentration limits/eligibility criteria, lowering the advance rate, etc) before lending to this Originator again.
OVERVIEW/RESULTS: The TF lenders in this scenario suffer no change in expected returns, since the overcollateralization was enough to make us (END and the super senior and levered senior lenders) whole.
Scenario 3: Losses are extremely high
END provides Loan A to Originator at an 80% advance rate with expected losses at 10%
END reviews the Originator’s loan data, and notices delinquencies are coming in higher than historical trends. No covenants/triggers tripped yet, but delinquent loans are kicked out of the borrowing base, so the Originator must post either additional cash or additional newly originated loans to account for the removal of the delinquent loans from the borrowing base. Any cash flows/recoveries from these delinquent loans are still passed through the financial instrument in accordance with the distribution waterfall. However, the Originator in this instance decides to post additional collateral.
END has a brief chat with the Originator on this trend to see if the Originator shifted their growth strategy and/or collections procedures.
Realized Losses on the underlying collateral pool come in at 25%. Loss covenant trigger tripped (i.e. 15% loss trigger).
END takes a deeper look into the data to see what was driving the increase in losses.
END has a thorough discussion with the Originator on the findings from our data deep dive, and allows the Originator to clarify/explain any shortcomings in their underwriting/execution.
Originator receives no excess cash flows (given the trigger breach), and all excess cash flows are applied towards paying down the principal balance of the financial instrument (first the super senior, then the levered senior).
END decides to pause lending, allowing the principal balance of the financial instrument to pay down. During the pause, END will work with the Originator on amending the structure of the financial instrument to increase structural protections (tightening concentration limits/eligibility criteria, lowering the advance rate, etc) before lending to this Originator again.
OVERVIEW/RESULTS: The TF lenders in this scenario suffer no change in expected returns - the overcollateralization itself was not enough to protect the financial instrument from risk of principal loss, so we enforce the corporate guarantee and cash reserve to make us (END and the super senior and levered senior lenders) whole (both principal and interest).
Scenario 4: Losses are ridiculously high
END provides Loan A to Originator at an 80% advance rate with expected losses at 10%
END reviews the Originator’s loan data, and notices delinquencies are coming in much higher than historical trends. DQ covenants/triggers tripped, but delinquent loans are kicked out of the borrowing base, so the Originator must post either additional cash or additional newly originated loans to account for the removal of the delinquent loans from the borrowing base. Any cash flows/recoveries from these delinquent loans are still passed through the financial instrument in accordance with the distribution waterfall. Given that the DQ trigger is tripped, the Originator is not allowed to add additional collateral loans, only cash to pay down the loan to meet the borrowing base requirement.
END takes a deeper look into the data to see what was driving the increase in delinquencies. END has a thorough discussion with the Originator on the findings from our data deep dive, and allows the Originator to clarify/explain any shortcomings in their underwriting/execution.
Originator receives no excess cash flows (given the trigger breach), and all excess cash flows are applied towards paying down the principal balance of the financial instrument (first the super senior, then the levered senior).
Realized Losses on the underlying collateral pool come in at 45%. Loss covenant trigger tripped (i.e. 15% loss trigger), but this is redundant since the DQ trigger is also tripped.
END decides to pause lending, allowing the principal balance of the financial instrument to pay down. During the pause, END will work with the Originator on amending the structure of the financial instrument to increase structural protections (tightening concentration limits/eligibility criteria, lowering the advance rate, etc) before lending to this Originator again. However, given the high realized losses, END will likely opt to continue de-risking and not to amend the instruments structure going forward.
OVERVIEW/RESULTS: The TF lenders in this scenario suffer no change in expected returns - the overcollateralization itself was not enough to protect the financial instrument from risk of principal loss, so we enforce the corporate guarantee and cash reserve to make us (END and the super senior and levered senior lenders) whole (both principal and interest).
So what scenario would TF Lenders be exposed to principal risk? Essentially a scenario where defaults come in absurdly fast (i.e. too fast for the triggers, monthly reviews, and regular discussions with the company to catch early signs) and absurdly high (i.e. greater than the OC + excess spread + cash reserve, for our current Originator, is a scenario where they collect ~$30 for every $100 deployed).
It has clarified a few things for me so thank you. Hopefully has been of benefit for others as well.
As a final comment, the thinking behind my line of questioning was to get a better handle on the risk that is being taken here. It sounds like from a structural perspective, there are adequate measures in place to mitigate these risks, which need to be appropriately priced.
However, unless I am mistake (happy to be corrected) - this is a very high risk segment of lending to end customers who have little track record and high risk of default in a volatile industry, which is probably why there is very little Tradfi appetite to aggressively pursue such lending (unless we have direct recourse to the creditors).
Defi presents opportunities to bank underbanked segments, whether it be through use of better technology to assess risk, or simply being more native and frictionless to its participants, but we should be aware of the risks and make sure we apply appropriate discipline to protect our capital, which END Capital appears to understands well.
There is actually a lot of TradFi appetite for these types of opportunities. The way TradFi structured credit managers assess risk isn’t so much on whether default rates are high - high default rates isn’t as big of an issue, if the originator is sophisticated enough to price these sort of loans properly, and have robust servicing operations in place to handle various stages of delinquency.
During my TradFi career, we’ve seen opportunities where the expected default rates of the underlying collateral pools were even north of 40%. In those instances, not only did the structure protect us, but the originator was able to price the risk appropriately (this is directly related to what Ed was referring to when he mentioned “excess spread” in the AMA).
However, the main risk is not so much high default rates, but high volatility levels in performance. If default rates are at 1% for one cohort of loans, but 30% for the next cohort, then there isn’t a clear indication as to what is driving the default rates. The structural protections that I mentioned (especially the triggers) provide us with early warning signs, but more broadly, there needs to be someone who actively monitors and manages the risk, which we are positioning ourselves to do on behalf of the TrueFi lender community, through our new pool.
