I’ve been thinking a lot about a twitter thread from Patrick McKenzie back in late January about the potential for SaaS companies turning to securitization as a source of financing. On the one hand, I think he’s definitely right, especially as we see a larger system of VC-alternative investment models pop up; others have also chimed in and written intelligently in a bit more detail on what the idea of debt coming to the tech industry means as well.
At a high level, the point is pretty simple – most SaaS businesses have, for a given cohort of customers, a pretty predictable and regularly recurring cashflow stream. That, fundamentally, looks a lot like a fixed income instrument like a mortgage or a student loan or any other asset which spits off similar-ish amounts of cash every month and has some nonzero chance of getting cut off at any point in time (like a mortgage borrower defaulting, or a SaaS customer cancelling a subscription).
In a prior life, I used to trade consumer and esoteric ABS, and this is certainly far from any of the most unorthodox things I can think of securitizing. That said, I also think it’s important to think through how one gets from point A to point B, and I’m not sure I’ve really seen much detail from anyone on what such a securitization would actually look like, or who would be buying, or how the market should evolve, and I thought it might be worthwhile to try and flesh that out a little.
How
Let’s talk about what a SaaS securitization would actually look like. I think there’s probably three types of securitization that make sense to consider:
- Standard asset-backed securitization – This is your vanilla ‘static pool’ type securitization, which is probably easiest to understand and what most people are thinking about when they refer to securitization. Take a fixed/static pool of assets, package them together, put it in some SPV and sell debt against the cashflows which flow into the SPV. Here your credit risk/analysis is specifically considering the fixed set of cashflows owned by the SPV and what you expect their value to be. The debt is simply paid down over time beginning on day 1 based on the cashflows that come in, and once the debt is paid off the assets are returned to the borrower. 1
- Term facility securitization – This is somewhat more nuanced, and is currently how a lot of younger lending companies finance their loans. You similarly have an SPV/facility which issues debt against a pledged pool of assets, but you don’t begin paying things down on day 1; instead, you make interest payments from the cashflow coming in, but you have a ‘revolving period’ during which the borrower can pledge new assets over time to replace old assets which are going away (e.g. cancelled subscriptions or the like), such that the overall amount of collateral in the facility stays relatively fixed. After this revolving period ends, you begin to amortize the same way you would in the standard ABS.
- Whole business securitization – This is definitely a more ‘esoteric’ version of securitization, but has been a pretty popular type of financing for companies that have very stable recurring bond-like cashflows from their assets but don’t explicitly have bond/loan type assets generating those cashflows (for example, fast food franchise fees, or gym membership subscriptions). Rather than picking out some assets to go into an SPV, here the SPV ends up owning all of the IP and revenue-generating assets for a company, and the company then owns the equity of the SPV. The SPV issues debt, and the debt owners get their payments prior to the company receiving any of its revenue for the month.2 If you squint a little bit, this actually sort of looks like what a Shopify Capital merchant cash advance or a Stripe Capital loan looks like (since both those companies basically have control of the revenue channel for a business taking out a loan), though the format here would involve a more regular debtlike payment schedule rather than the ISA-like structure of a merchant cash advance (fixed payment terms vs. fixed percentage of revenue).
Institutional buyers
Implicit in the argument around SaaS securitization, I think, is that at least to some extent the same buyers of current MBS or consumer/esoteric ABS should be buyers of SaaS ABS as well. In the long run, I think that makes sense – it’s a (somewhat) uncorrelated stable cashflow stream, you can set it up to be whatever maturity you want, and the underlying business concept isn’t that hard to understand, all of which are positive things. That said, I still think there are some shorter-term hurdles to getting that institutional interest.
Chief among these to me is the question of ratings. The vast, vast majority of institutional buyers in the ABS space are funds looking for fixed income investments with IG ratings3 from a major ratings agency4; if it’s an unrated investment their target returns are significantly higher. Offhand, I’m not sure how difficult it will be to convince ratings agencies of the appropriate risk profile for subscription cashflows? Unlike an actual loan or debt instrument, there’s no legal or contractual obligation for a customer to continue to pay, which makes underwriting the cashflows a bit more challenging. On the other hand, churn and retention figures are pretty straightforward metrics to track, so estimating the total lifetime cashflows of a given cohort shouldn’t be particularly difficult; it’s just a matter of making the case for those metrics. Again, this is where I think structure is quite important. In the ‘static pool’ (Option 1) type of securitization, you’ve probably got a period of relatively stable estimated cashflows for a cohort, with some tail at the end which is harder to quantify – depending on the term and loan-to-collateral ratio of the debt you’re trying to issue, that strikes me as harder than a facility or WBS type approach where it’s easier to make the case that stable cashflows at the business level merit low risk, even if the probability of any given subscriber’s cashflows has too imprecise a tail to merit a safe rating. It’s possible I’m overstating this challenge – Kroll has put out some info on what they look for in WBS deals, which explicitly talks about the durability and strength of the underlying business cashflows, so it’s not something they’re unfamiliar with – but it strikes me as a challenge.
The second big risk is simply size. Most ABS deals are quite big. There are a few reasons for this. There are some fixed costs, for one, so it’s easier to do a bigger deal to reduce those costs per dollar borrowed, but also the vast majority of institutional buyers are very large entities who manage hundreds of billions of dollars. If you’re managing that much money, you need to focus on opportunities to invest a lot of money at once. If you have a deal which is both (1) not particularly big and (2) something new and unfamiliar, it’s going to be hard to get those guys to care. Lastly, liquidity – larger deals mean more bonds in circulation, and likely more owners of said bonds. That’s generally viewed as a positive for liquidity, since any seller has more potential buyers who are already familiar with the deal.
How large are we talking about? If I look at every ABS deal Kroll5 rated in the past year, I see a total of six issuances which are under $100mm in bonds. Four of these are pass-through securities of pools of Upgrade and Upstart online consumer loans, which have very well-defined collateral paydown schedules. One of these is an SBA loan securitization, which is again an easily understood collateral pool. The last is a tax lien securitization, which is actually a pretty funky concept but was issued by New York City (a well known entity) and is the 25th such deal they’ve done in the past few decades (a long track record). Those are all pretty easy concepts for people to get on board with, so if you’re a big institution you can make a decision on these deals pretty quickly.
Okay, so how large a business does it take to issue $100mm in ABS bonds? Picking a random recent WBS deal, debt-to-securitized net cash flow tended to be in a 5.5 to 6.6x range. These are for very very large and well known companies with larger issuances, so to be conservative I’ll use a 2.5x to 3.5x multiple range for a SaaS company – that puts you in the $30 to $40 million range for annual cashflow to issue $100mm in WBS. Thinking on a static pool instead and taking a more granular approach, assuming something like a 5% interest rate (pretty conservative) and debt that fully amortizes over 5 years, you’d need revenues to cover about $1.9 million per month, or $22.6 million per year, in debt payments. Unless you want all of your revenue going towards servicing your debt, that again probably puts you in a $25mm+ range. Are there SaaS companies that can do that? Without question; I am familiar with Patio11’s Law. But that is still a hurdle.
The other buyers
That’s not to say you need the large institutional investors to make this happen – there are still quite a few interested parties I’d imagine should have interest in something like this. The first coming to mind is hedge funds, private credit funds, or other non-ratings constrained institutional investors. These are the same types of funds which are often sponsoring private term facilities for younger lending companies today, or buying consumer debt from online, so they’re already often involved in some kind of startup or earlier stage financing. They’re comfortable with no ratings, they’re comfortable with smaller deal sizes and less liquidity, but they’re looking for higher rates of return. This is sort of the ‘vanilla’ option of who makes sense as a SaaS ABS funder.
Next coming to mind is the accredited investor community. On the generic side, you don’t have to look very far to see that there’s been a lot of success in selling illiquid debt instruments to the AI community. Yieldstreet has had $1 billion invested on their platform online. Peerstreet has over $3.5 billion transacted. Fundrise lists over $4.9 billion in transactional value.6 Beyond the generic AI interest, it’s also interesting to remember that there are a lot of new AIs coming around every year in the tech industry – and if an angel want to make a lot of startup equity investments in businesses they know, it should make sense to get a bit of portfolio diversification with some startup fixed income investments in businesses they know, too.
Lastly – and maybe the most natural funder – would be the Stripe Capitals and Shopify Capitals of the world. It’s hard to imagine someone with more intimate familiarity with a business’s revenue streams and cashflows than the platforms which are the primary conduits of those cashflows. It’s hard to imagine someone with better servicing capabilities around those revenues than the primary conduits of those cashflows. It’s hard to imagine someone with better data on what SaaS revenues look like than the engines that drive SaaS revenues. And it’s hard to imagine someone in a better position to do securitized lending to a SaaS company than the guys who are already lending to them! This just makes too much sense to not happen.
Who’s selling
Offhand, anyone with a SaaS product? The main question is around what the next best alternative source of financing is. We’ve already talked about the necessary size to probably get institutional interest, but at the big end, for example, I’m still surprised and almost disappointed that something like this hasn’t happened already for someone like a Netflix or Spotify (very large, very well known, very stable subscription model). The odds strike me as pretty close to zero that it hasn’t been considered by somebody somewhere; these are big companies with sophisticated banking partners and I’m am close to 100% certain that somewhere in some investment bank there is an associate or a vice president who has run the numbers on how the spreads in the esoteric ABS space compare to those in the high yield market and which makes more sense for the big issuer – the reason it wouldn’t have happened is because your other debt options are really good, not because this option is bad.
If you’re a small guy, on the other hand, who’s nowhere near that institutional level but is mostly bootstrapping or relying on credit card debt, you’re almost certainly better off with some kind of ABS solution marketed at private credit or accredited investors. It’s just a matter of someone building that marketplace.
Lastly
The last thing I want to be clear on is that I think the long-term size of this opportunity is huge. Any skepticism above is around how to get from Point A to Point B, not whether/if it is possible to reach Point B. I’m not sure how big the SaaS industry as a whole is, and am afraid to even put a number down since I know I’ll be far off, but suffice it to say it is big and growing. The esoteric ABS space is something like $50 billion in issuance per year right now, and there’s no reason that can’t or won’t grow significantly. There is a huge level of demand for steady cashflow streams from uncorrelated assets, and there is a huge base of companies which are being built and could use another alternative source of funding which aligns well with the revenues they generate. Getting there is the interesting challenge.