Do Startup Valuations Matter for Investment Returns?

If you raise at a low valuation now and you exceed your expectations, it will make it much easier to raise the next round. If you raise at a high valuation now and exceed expectations, you may be at a disadvantage in fundraising because the prior round set a precedent that will be hard to match. In some sense, fundraising for startups is like a frog jumping between lily pads until they get big enough that capital sources become more continuous. We want to have a good feel for the size and distance of the next lily pad and your odds of hitting it. Current valuation is a big piece of determining that.

On the other hand, all company attributes are never equal. My past work has suggested early-stage financings are relatively efficient (i.e., asset prices reflect all available information). That means higher valuations are associated with companies that have more auspicious indicators of future success.

So the two competing factors are:

• Startups with higher prices are typically better companies that are more likely to get big markups, and
• Higher prices make it harder for a company to get marked up, while increasing entry prices produce lower return multiples.

The Data on Markups

Our methodology is to segment deals by time and look at the performance of seed deals grouped by the quintiles of pre-money valuations within that time span. Note that these deals will all have about the same amount of time to season, so it’s coherent to directly compare markup rates between these valuation segments. We’ll also look at typical markup rate multiples within the marked-up deals of those quintiles.

For simplicity, we define a “marked up” deal as a deal that has a positive gross multiple as of July 1, 2023. That means companies that, for instance, raised a Series A but then shut down prior to July 1, 2023, would not be counted as a “mark up.”

We’re looking at data as of July 1, 2023, so these are deals that have had 2-5 years to season. In general, as deals age, their markup rates and multiples tend to increase, as more companies raise later-stage rounds at valuation step-ups. Note this data only reflects seed deals that have a defined valuation or cap in USD (uncapped SAFEs are excluded).

Takeaways

There are two clear trends in the data:

• Seed deals have gotten more expensive over time, and
• Two years is not enough time to observe markups at seed, while years 3-5 of seasoning seem to be less important (put another way, if your seed deal hasn’t been marked up after three years it appears unlikely to ever be marked up).

The data doesn’t show a clear relationship between valuations and markup rates within time periods. The highest quintile of seed deals by valuation have the second-lowest markup rate in 1H18, but the highest markup rates in 2Q20 and 2Q21 deals.

If there is an observable trend in the data, it’s that the median multiple decreases as markup rate increases. For example, in 1H18 the third quintile has the highest median markup and lowest markup rate, while the fourth quintile has the highest markup rate and lowest median markup. This is a sign of market efficiency because it suggests the marginal deal that gets marked up (i.e., the seed-stage startup that just makes it to Series A) will get a relatively small markup. The negative relationship between markup rates and multipliers is evidence that VCs are context sensitive and not just blindly applying a “standard deal” for their Series A companies.

In her response to Ryan’s query on valuations, Linda Xie of Scalar Captial said she views passing on a startup because of a high valuation as a missed opportunity. She quoted Paul Graham’s advice on early-stage investing:

Valuation matters far, far less than the decision of whether to invest or not. The spread between bargain and outrageous startup valuations can't be more than 5x, in a world where the best investments can return 1,000x.

Our data backs up Xie’s and Graham’s perspective: we see about a 5x spread between typical top and bottom quintile seed deals and no apparent difference in those deals’ performance. This result might call into question the nature of “alpha” in venture capital, as it suggests there are no opportunities for an investor to generate excess returns by only allocating capital into “cheap” or “expensive” deals. This is why, in a previous study, we found startups led by founders from “top” founder schools (e.g., Stanford, Harvard, and MIT) don’t necessarily generate the highest markup rates, as the founders may use their credentials to obtain higher valuations than their startups otherwise would have.

Instead, AngelList data suggests broadly indexing across all viable seed-stage deals, regardless of valuation, is the most reliable way to potentially hit on an outlier investment as an LP.