July 27, 2008
Gordon Smith on Broughman's Independent Directors
Posted by Christine Hurt

[The following comments were edited by Gordon after Gordon and Brian discussed the paper offline. Gordon misread the section containing Brian's hypothetical facts, and that misreading provoked some comments that are not applicable to the paper.]

Brian Broughman’s paper is an important and provocative contribution to the literature on venture capital contracts and independent directors. His attention to detail in the empirical section of the paper is suggestive, but inconclusive. Nevertheless, this paper is a must-read for anyone interested in the governance of venture-backed firms. In the following paragraphs, I focus on Part IV (”Data”) in an attempt to provoke a more thorough examination or explication of the data in a future draft. All in all, I think this paper takes a valuable step forward.

Brian begins with hypothetical facts that illustrate the possibility that a venture-backed firm with only two control options – VC control or Entrepreneur control – may fail to pursue the “efficient” strategy when that strategy does not produce the highest expected return for either party. In the hypothetical case, the “efficient” option, i.e., the option with the highest combined expected value, is a low-risk strategy, but the VC receives the highest expected payoff when the company is sold immediately, while the Entrepreneur receives the highest expected payoff when the company pursues a high-risk strategy. Brian wants to know how venture-backed firms get to the efficient option, and he suggests a crucial role for the so-called “independent directors.”

Before we get to the discussion of that role, I pause to comment on Brian’s notion of an “efficient” outcome. As noted above, Brian identifies the “efficient” option by reference to the combined expected value of each option to the VC and Entrepreneur. Two problems with his approach:

(1) From a societal standpoint, there is no difference between an immediate exit and the “efficient” (low-risk) strategy. If the low-risk strategy were the optimal course for the company, as posited in the hypothetical, presumably the Buyer in the immediate sale would pursue that strategy. Thus, the only strategy that would result in “inefficiency” is the high-risk strategy, and the only party who has an incentive to pursue that strategy is the Entrepreneur. Fortunately, as shown by Brian and others, entrepreneurs rarely exercise outright control of venture-backed firms. If the company pursued the low-risk strategy, the only issue would be whether the return from that strategy would be earned by the Buyer or the Entrepreneur. Under the facts of the hypothetical, the VC cannot gain from the low-risk strategy, but stands to lose $20 if the strategy fails. The obvious question: why would the VC place itself in such a position? (Several possible answers suggest themselves, but I think Brian needs to develop this point.)

(2) From the standpoint of the two parties in the venture-backed firm, comparing an immediate exit with two future exits is misleading unless you account for the possibility of an alternative investment after the sale. Presumably, both parties would take the proceeds from the immediate sale and reinvest those proceeds. Whether the low-risk or high-risk strategies were attractive, therefore, depends not only on the terms of the immediate sale, but on the expected returns from these alternative investments. Given that the VC has a negative expected return from continuing the current company, we know that the VC has a preference for the immediate sale. But the Entrepreneur might also have a preference for the immediate sale if the return from that sale plus the return from an alternative investment exceeds the return from both of the future exits.

I haven’t work out all of the implications of the foregoing comments, but I assume that Brian can smooth out the kinks in the hypothetical facts and retain the basic conflict between VC and Entrepreneur. That leads us to the so-called “independent directors” and the heart of Brian’s paper. Brian argues forcefully – using a combination of venture capital contracts, interviews, and survey data – that these directors act as “arbiters” of the conflict between VC and Entrepreneur. And he has almost convinced me that he is right. But I have a few comments and questions about the argument …

The argument turns on the assertion that “independent directors” – i.e., directors not affiliated with VC or Entrepreneur – are chosen by mutual agreement of the VC and Entrepreneur. This is important because an arbiter must be unbiased. If “independent directors” are actually chosen by one party or the other, we would expect those directors to favor the party who selected them.

Of course, all of this presumes that directors chosen by mutual agreement of the VC and Entrepreneur would actually be unbiased. Brian argues this point by reference to reputational constraints: “If an independent director develops a bad reputation among the broader community of entrepreneurs or among VCs, he is unlikely to be appointed to serve on future boards.” (p. 21) When I see arguments that depend on reputational constraints, I wonder about the mechanisms for developing and transmitting reputation. How would anyone know whether a particular director systematically favored VCs or entrepreneurs? The data are likely to be incredibly noisy – much noisier than the arbitration decisions in the Bloom and Cavenaugh study cited by Brian – and even if one could make the case against a particular director rather strongly, how would that director’s reputation be conveyed to future parties? Yes, Silicon Valley is tight, but I have my doubts about the effectiveness of this constraint.

Now to the central question of the paper: are independent directors chosen by mutual agreement of the VC and Entrepreneur? Or is control contingent on respective levels of stock ownership by VC and Entrepreneur, as I argued in my 2005 UCLA Law Review article? Before going any further, it is worth reminding ourselves that the VC and Entrepreneur need an arbiter only when they disagree. All of us who have written on this issue, as far as I know, recognize that VCs and entrepreneurs in fact choose directors by mutual agreement as long as all parties are on the same strategic page. But what happens when the parties diverge? Who has the power to elect the independent directors? Do they exercise that power?

Let’s turn to the evidence. Brian reports data on 54 venture-backed companies located in Silicon Valley that were sold to an acquirer in 2003 or 2004. Sources of data include the corporate charter, interview responses, and survey responses for 32 of the companies. On the issue of whether independent directors are chosen by mutual agreement of VC and Entrepreneur, Brian reasons that “the charter … is generally not the decisive document for the selection of independent directors.” Rather, he asserts, the key document is the voting agreement, and the “typical” voting agreement requires mutual agreement of VC and Entrepreneur. Brian acknowledges that he did not request voting agreements from the firms and that “only a small number of firms” supplied him with voting agreements – three to be exact – but all of these voting agreements required mutual agreement.

Brian bolsters his argument by appealing to the model venture capital financing documents of the National Venture Capital Association, which include a voting agreement specifying mutual agreement of the parties. Interestingly, the NVCA model charter includes the sort of voting provision that I found in my study, namely, a provision allocating a specified number of directors to each of VC and Entrepreneur, with the balance of the directors to be elected by VC and Entrepreneur voting together as a single class. This, of course, is not a provision requiring “mutual agreement,” but rather a provision allowing the party with the most votes to prevail in a contested election. (Of course, most elections in venture-backed firms are not contested, but remember that we are only interested in case of disagreement.) Thus we seem to have an unexplained disconnect between charters and voting agreements. Brian observes that the voting agreements comply with the charters, and that is true enough, but it doesn’t explain why the charters do not require mutual agreement.

One possibility is that voting agreements are not as pervasive as Brian suggests. Thus, the charter may be drafted with majoritarian default rules, and the parties may contract around those default rules only occasionally. This seems unlikely, when you consider that Kaplan and Stromberg also found “mutual agreement” provisions in their sample. Another possibility is that the corporate charter supplies the rules for all shareholders, while voting agreements apply only to certain shareholders and not to others. Perhaps early investors tend to bond with entrepreneurs through voting agreements to protect themselves against later investors. Still another possibility is that the corporate charter supplies the rules when the voting agreements fail or are terminated. Note that the NVCA model voting agreement may be terminated upon agreement of specified percentage of the “Key Holders” of the company’s stock, thus holding out the possibility that large investors could dispense with mutual agreement and take control of the company for themselves under the voting rules in the charter.

The bottom line is that we still don’t know the bottom line about the control of venture-backed firms. The rights that really matter are those that adhere when VC and Entrepreneur disagree, and the firms in Brian’s sample do not give us very good data on this situation. Nevertheless, Brian has provided more grist for the mill. And he may very well be right, in the end.

Junior Scholars | Bookmark

TrackBacks (0)

TrackBack URL for this entry:

Links to weblogs that reference Gordon Smith on Broughman's Independent Directors:

Recent Comments
Popular Threads
Search The Glom
The Glom on Twitter
Archives by Topic
Archives by Date
January 2019
Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    
Miscellaneous Links