Papers exploring Angels vs. Venture Capital: how are they different, and are they complements?

Our reviews of SSRN papers this month have explored the overlap and differences between angel investing and venture capital. Here are two papers analyzing whether “VC” and “angel” are two separate funding tracks, and how deal terms differ between the two.

How do angel investors overlap, collaborate with, compete with, benefit from, or get punished by venture capitalists? There is a lot of opinion and research about this topic — here about pay to plays and VC behavior, for example. In their paper Angels and Venture Capitalists: Substitutes or Complements? Thomas Hellmann, Paul Schure, and Dan H. Vo provide an interesting review of how angels and VCs overlap. The article is available here on SSRN.

The authors look at how much, and how, angels and VCs work together on investments, or invest separately. Are “angels and VCs synergistic members of a tightly knit ecosystem,” with angels supporting companies first and then sending them over to VCs to scale (p.3), or are they “separate financing paths that rarely cross each other,” as angels back different types of companies or try to avoid their portfolio companies working with VCs (p.4)?

Using data on tax credits related to the British Columbia Investment Capital Program, so not the usual Crunchbase or Pitchbook sets, covering 469 companies (mostly technology-based) from 1995–2009 (a total of 18,925 investments by 9,424 unique investors across 2,184 financing rounds in 469 companies), the authors find a few interesting results:

1. “Dynamic persistence” within investor types — i.e. a company that raised capital from one type of investor is likely to raise more from the same type (p.24).
2. “Negative dynamic effect” between angel and VC financing — i.e. companies that obtained more angel financing in the past are less likely to subsequently obtain VC funding; and vice versa (p.24).
3. These choices are “company-led” (i.e. driven by the company characteristics), not “investor-led” (section 5).
4. There is no “quality contingent complements” — meaning that VCs were not just backing the “better companies” from the angels’ portfolio, though there is evidence that VCs had better results (section 6).
5. Overall the second description above, of separate financing paths, is much more the norm than the tightly knit ecosystem.

These results are summarized on pages 4–5 and derived in detail, with supporting statistics that for brevity we’ll skip here, in the rest of the article.

This is pretty fascinating stuff. If you read the Silicon Valley-dominated tech press or consume angel content, I’m sure you would selected the “ecosystem” and “quality contingent” hypotheses. The overwhelming public opinion is (i) angels fund early to set up a company for VC money and (ii) angels fund weaker companies that often fail to “graduate” to VCs. Turns out, these ideas might not be correct.

Does this translate to the VentureSouth experience?

Speaking personally and from anecdata, I (Paul) would agree with the article and side against the agreed wisdom, but I think it has changed over time and still varies by geography. Ten years ago, VentureSouth members absolutely prioritized businesses capable of achieving an “early exit” without necessarily seeking large amounts of venture capital funding, because that was simply not available in the southeast. This is consistent with the conclusion and data timeframe of the study.

But as more VCs have entered the SE market, and perhaps our members’ appetites have evolved to prefer slightly-later-stage, larger rounds, and more VC-coinvesting, this has changed a bit over time. We still prioritize time to exit, and hold rounds that intentionally bridge to large raises to stricter scrutiny. But we have coinvested with many smaller southeastern venture capital funds (and the boundary between “angel” and “micro VC” always gets more blurred), and appreciate the positive momentum from seeing a later-stage VC round well capitalize the company and increase our book value. Perhaps the dynamics change as ecosystems “mature” or get more VC capital locally or more interest from tech-hub VCs?

As we have noted in prior papers, getting good data is hard. This paper struggles, like all do, with quantifying exits. For example, resorting here in some instances to rating an acquisition successful “if there is some kind of press release with substantial and positive praise of the acquired company” (p.16) — which is a leap of faith! It also has to fill in a bunch of gaps (e.g. investors don’t necessarily claim or take their tax credits).

The jury remains out on whether recent trends will lead to better returns or not. In the paper, the VC investors had more frequent positive exits than the angel investors (with VC-only companies having a 27% higher probability of exit and 18% lower probability of failure). But this is not necessarily saying the VCs had higher returns — outcome depends on the entry valuation, ownership and dilution, challenges along the way, and much more. Either way, angel investing and VC investing is hard.

Our experience aside, the paper is a valuable reminder that good angel deals do not necessarily involve venture capitalists, and that positive exits can be achieved, and potentially more quickly and, though not covered in the paper, potentially with greater founder liquidity, without VCs. There are substitutes and complements to the “right way” to fund early stage companies! Overall, the literature remains inconclusive about the relationship between angel and venture capital funding (p.11).

As always, we like to pick out a couple of interesting nuggets of data or ideas. Here we learned:

• In this dataset, VCs invested only slightly later than angels (p.18 and Panel A) — at 2.3 years for angels and 2.7 for VCs. The received wisdom is VCs invest much later than angels; even in early/non-SV markets like British Columbia 1995–2009, this was not true. (This supports the main “substitute” hypothesis, but even by itself is an interesting stat.)
• 24 companies IPO had an IPO — 5%, more than I would have guessed. More angel-only deals (6) IPOd than VC-only deals (5) — the opposite of what you guessed?

Of the 7215 angels in the study, 6801 are casual angels — meaning they only invested in one company in the dataset. (214 were serial angels and 200 angel funds.) Yikes. Please be diversified!

Given these differences, let’s explore how angel investments and venture capital investments have in the past been structured different — using Are Angels Different? An Analysis of Early Venture Financing by Brent Goldfarb, Gerard Hoberg, David Kirsch, and Alexander Triantis from 2013. Paper available here.

Goldfarb et al look at 182 Series A preferred stock deals from 1993–2002 into high-growth / tech companies from the records of a defunct law firm saved in a “Closed Archive” in the Library of Congress — an interesting set of source data for sure! — documenting the varied deal terms and participation from investors. The records cover 458 founders (across 165 companies), 482 different VCs (across 150 deals), and 2,528 angels (mostly individuals, but also including non-VC institutions like universities or banks, across 144 deals). They also crunched the usual sources for internet presence, later rounds, and exit data.

“We find support for the conventional wisdom that angels invest with fewer protections than do VCs” (p.2). These are deals involving a prominent law firm, so the angels are most likely to be “sophisticated” and “most like VCs” (p.3). Even being sophisticated, they accepted fewer downside cash-flow and control protections than VCs.

Why do they do that? In section 2, the authors outline prior scholarship about why, despite the earlier-stage, “higher-risk” investments angels make, they do not require more protections than later-stage VC investors. This has been a puzzle for some time.

The authors consider three potential explanations: 1. transaction costs — legal fees are expensive, and on smaller deals no-one can afford them, so the deals have fewer terms including protections; 2. “preferences” — angels are often entrepreneurs, and are more closely aligned with entrepreneurs, and can influence them socially rather than contractually (or are just more naïve); 3. “selection” based on some kind of unobserved variable (which, honestly, I don’t get!). And conclude that #2 (angels align with entrepreneurs more than VCs do) is best supported by their data.

Let’s quickly review what these protections are. The first is dividends: VCs have them, angels do not (p.18, and none of the angel-only deals had them per p.20). The authors argue this suggests VCs are less patient and use this tool (“a strong incentive for the firm to accelerate to a successful exit event” (p.18)) to get things moving. Not sure we agree this is really that powerful a tool — though maybe that proves the authors’ point!

Second is liquidation preference. Honestly we don’t quite follow what the descriptions here mean — they don’t seem to describe “liquidation preferences” as we understand them. But whatever exactly is being measured, here the protections for angels are “much less favorable” (p.20) — with the difference looking fairly stark (0.579 for VC, 0.125 for angel-only, significant at the 1% level) on Table IV.

Lastly, redemption rights. Again VCs have them (72 of 150 deals), angels do not (4 of 32 angel-only deals) (p.20).

Today, we would expect 1x liquidation preference on all deals, no dividend, and only sometimes a redemption right. Why are the structures in this dataset so different from that expectation? Have things changed over time, so that angels have learned the lessons from forgoing some protections (liquidation preference), or have angel investors’ preferences become less aligned with founders as angels have become more “professional”? Have VCs learned to be more investor friendly (forgoing redemption rights)? Is the sample misleading?

At VentureSouth we propose “VC-style” structures in our deals — priced preferred equity rounds using NVCA model documents — even if (in the wisdom of some) that is “overkill” for what’s needed in an angel round. We try to avoid SAFEs and convertible notes, wholly avoid common stock, and judge equity rounds lacking protections harshly. Not all funds do. Whether this leads to better or worse returns, we’ll leave for you to judge — but this paper definitely suggests limited protections can lead to worse returns.

This is a fascinating paper overall, and leaves us with more questions than when we started! In addition, there are some random interesting nuggets that we thought you would enjoy:

• The superior performance of VC-only, larger deals is only present when the board of directors is not firmly under the control of either common or preferred shareholders. (p.4). A balanced board of directors is the way to go. We mentioned this in Episode 99 of the Venture In The South podcast here.
• Warrants were also sold in 15% of the rounds, and more so in the smaller deals (20%) (p.17). How does that square to how often your deals include warrants? Seems high to me. Wonder why?
• We find that the average time between first and second closings is 153 days, though it is much longer for large deals (198 days on average) (p.17). It’s absolutely standard practice for rounds to take some time to fill entirely (despite FOMO and overnight deals you might read out), but this length is surprising even so.

And any attorneys reading might be interested to learn the average billing was for 169 hours of work (p.17). Ooof.

If you like these reviews, have feedback, or have anything we should add to our reading list, please get in touch!