T. Rowe Price Voted for the Dell Buyout by Accident
Sometimes companies are acquired by other companies. The way this works is that the target company's managers and board negotiate a deal with the acquirer, and then they call a meeting where the shareholders can vote on the deal. They book a big room, and the shareholders show up, and the managers say why the deal is good, and the shareholders discuss it for a bit, and then they vote. This is a real thing that actually happens. I went to one once.
But usually it is more complicated than that. Mostly, it is inconvenient for shareholders to show up at the meeting, so mostly they don't. Companies mail out proxy cards, and if you are a shareholder with better things to do that day, you fill out the proxy card and mail it back. The proxy card appoints someone who'll be at the meeting anyway -- typically someone who works for the company -- to vote on your behalf. You tell her how to vote, and then she does it, and it's as though you'd been there and voted yourself. You don't even have to pay postage; the company provides a self-addressed stamped envelope.
But usually it is more complicated than that. Most shareholders in most companies are not actually shareholders in those companies. That is, they are not "shareholders of record," who own shares that are reflected on the stock registry of the company, or physical paper stock certificates. Instead, they hold shares in "street name," meaning that their banks or brokers own the shares on their behalf. This makes a lot of things more convenient, especially if you want to sell your shares: You don't have to contact the company and tell it to update its shareholder records, or mail stock certificates anywhere. Your broker, who actually owns the shares, can handle it for you. But this means that your broker also has to handle voting for you: The broker is the legal owner, so it gets to vote, but it will ask you for instructions and then vote however you tell it to. (Incidentally, this means that, if you are a shareholder in street name and you want to show up and vote at the actual meeting, you have to get a proxy from your broker so that you can get in. )
But usually it is more complicated than that. Even the brokers mostly don't own the shares they hold for customers in "street name." Instead, they've all agreed to put their shares in one place, called the Depository Trust Co., which owns all their shares for them. This makes transfers even more convenient; instead of moving around physical share certificates between brokerages, everyone can just move around entries in DTC's central computer. It does make voting one level more complicated: Delightfully, DTC (or, strictly, its nominee, Cede & Co.) is the record owner for state law purposes, while the brokers are the record owners for federal law purposes, with the odd result that Cede (the Delaware owner) has to give proxies to the brokers (the federal owners) so that they can vote. By proxy. On behalf of their customers.
But usually it is more complicated than that. Most shares are held, not by individuals, but by institutions, pensions, asset managers, mutual funds, etc. Many of those institutions own shares in hundreds or thousands of companies, and can't keep track of all those companies' meetings. So often they hire someone to keep track of the meetings, submit their proxies, and generally be in charge of voting. The situation is similarly challenging for brokers, who have to keep track of even more meetings, and who similarly outsource their voting obligations. So T. Rowe Price is a gigantic institutional asset manager, and State Street is the broker for some of its funds, and both of them outsource the voting stuff, to two different companies:
State Street outsourced to Broadridge Financial Solutions, Inc. the task of collecting and implementing voting instructions from its many account holders, including the T. Rowe Petitioners. To carry out that task, State Street gave Broadridge a power of attorney which authorized Broadridge to execute proxies on State Street’s behalf. At that point, voting authority for the T. Rowe Petitioners’ shares rested with Broadridge.
To fulfill its contractual obligations to State Street, Broadridge communicated with State Street’s account holders and obtained voting instructions by mail, by telephone, or over the internet. With T. Rowe, the process involved an additional party: Institutional Shareholder Services Inc. (“ISS”). To facilitate the submission of voting instructions in connection with numerous meetings of stockholders each year, T. Rowe has retained ISS to notify T. Rowe about upcoming votes, provide voting recommendations, collect T. Rowe’s voting instructions, and convey them to Broadridge. To make the voting process more efficient, T. Rowe has a computerized system that automatically generates default voting instructions and provides them to ISS.
That's from Wednesday's Delaware Chancery Court decision about the 2013 leveraged buyout of Dell Inc., which features this lovely chart of how those shares were voted:
It isn't as simple as just showing up to a shareholder meeting when you're invited! There are wheels within wheels within powers of attorney.
One reason that I have told you this boring complicated story is to persuade you that it is boring and complicated. In human life, things that are boring and complicated tend to get messed up, because honestly, who wants to pay attention to all this stuff?
As it happens, T. Rowe Price's votes on the Dell buyout got messed up in two completely separate ways, requiring two separate Delaware court opinions. T. Rowe Price opposed the buyout, which was led by Michael Dell, the company's founder and chief executive officer. Several T. Rowe Price funds not only vocally opposed the deal, they also demanded appraisal of their shares, which is the process by which a Delaware court determines how much the shares were worth and orders the surviving company to pay the difference. The buyout price for Dell ended up being $13.88 a share. T. Rowe thought it should have been more, so it sued.
- You have to demand appraisal in writing before the merger closes, and hold your shares continuously until the closing, and
- You have to vote against the merger.
All of the relevant T. Rowe Price funds more or less did both of these things. But not quite! Last year some of the funds' appraisal lawsuits were dismissed because they flunked the first test, for reasons that were entirely ridiculous: At some point, the shares beneficially owned by some T. Rowe funds had their official record ownership transferred from Cede & Co. to Kane & Co., a nominee of JPMorgan, which was the custody broker for those funds. The T. Rowe funds never sold those shares, but their entirely arbitrary legal ownership changed, which was enough to disqualify them from appraisal.
But the T. Rowe funds custodied with State Street avoided this problem, and stayed at Cede & Co. the whole time. They had an even dumber problem: They voted in favor of the deal. They didn't mean to, but they did, just because, you know, this stuff is so boring and complicated, who can keep track? The Dell buyout was so controversial that the shareholder meeting had to be rescheduled several times to get enough votes (and negotiate a higher price). Before the first scheduled meeting, ISS asked T. Rowe Price how it wanted to vote, and T. Rowe told ISS that some of its funds wanted to vote against the merger. So ISS recorded the votes against the merger. The first couple of times the meeting was postponed, nothing happened: ISS just kept the old voting instructions from T. Rowe, and T. Rowe logged back into the ISS voting system anyway to make sure the instructions were correct. Everything worked smoothly and redundantly. But then:
On September 4, 2013, the ISS Voting System generated a new meeting record for the re-scheduled meeting (the “September Meeting Record”). The T. Rowe Voting System showed both the July Meeting Record and the September Meeting Record. In the ISS Voting System, however, the September Meeting Record replaced the July Meeting Record. This had the effect of deleting the voting instructions that had been entered in the ISS Voting System.
The T. Rowe Voting System automatically pre-populated the September Meeting Record with the default voting instructions called for by T. Rowe’s voting policies. As a result, the T. Rowe Voting System populated the September Meeting Record with instructions to vote “FOR” the Merger, “AGAINST” the advisory resolution on golden parachutes, and “FOR” authority to adjourn the meeting.
No one from T. Rowe’s proxy team logged into the ISS Proxy System to check the status of T. Rowe’s voting instructions. As part of the routine operation of the two systems, the default instructions in the September Meeting Record were conveyed automatically to ISS.
Oops! I don't know whose fault this was. ISS's computer probably shouldn't have overwritten the old instructions. The T. Rowe proxy team probably should have double-checked that its instructions were right, like, the day before the vote. But it isn't hard to be sympathetic to both sides. The surprise is that this doesn't happen all the time.
By the way, you might assume that, since the law requires you to vote against the deal to get appraisal, and since T. Rowe Price voted in favor of the Dell deal, its appraisal lawsuit would be an obvious no-hoper and get tossed pretty quickly, but: nope. It took years, and the decision dismissing it is 70 pages long. This is in part because Vice Chancellor J. Travis Laster takes even more delight than I do in going through all of the complexities of the voting system, but also in part because T. Rowe had a real -- and paradoxical -- argument that it could demand appraisal, because it didn't own the shares, Cede did, and Cede voted enough shares against the deal to preserve its appraisal rights. The court didn't buy it -- I don't either -- but it wasn't a terrible argument.
When we talked about last year's T. Rowe-Dell decision, I said:
The financial system is built up in layers of abstraction over some vast and unwieldy machinery. The machinery is complicated in part in order to make the abstraction simple: You can buy stock with a click of a mouse because armies of people devote their careers to the legal niceties and operational maintenance and integration of all this back-office apparatus.
And sometimes they forget to check their voting instructions for the fourth time, and mild disaster ensues.
The obvious question is: Does it have to be this way? The machinery is complicated in part to make the abstraction simple, but also because it has accreted over decades to respond to particular problems in more or less ad-hoc ways, and because decisions made to simplify one thing, like stock transfers, have complicated other things, like voting. But computers are better now than they were when DTC was invented, or for that matter when proxy cards were invented. You could just, you know, build a big database of who owns shares, and transfer shares on that database, and not rely on a system in which people own shares in brokers' databases and brokers own shares in DTC's database and DTC owns shares in companies' (transfer agents') databases. The big database could send out voting instructions directly to the shareholders, cutting back on the outsourcing. You could build a new system, from the ground up, corresponding to the actual practices of finance rather than to the archaisms that they're built on.
This is of course more or less the dream of the blockchain, though there it is overlaid with a lot of mysticism about decentralization and cryptography. The mysticism can occasionally be annoying, but it serves an obvious function. All of this stuff is so boring. The people in charge of building the systems that transmit the voting instructions from the brokers' voting agents to the mutual funds' voting agents -- those people are not celebrities. Those people don't give keynote speeches at conferences. The blockchain's key benefit, I tend to think, is sociological: It makes this sort of back-end database technology stuff cool. A little overenthusiastic hand-waving may be a small price to pay to get people to pay attention to systems.
At the same time, though, technology is mostly good for solving technological problems. Building a really good computer system to keep track of share ownership is a good idea, but it isn't the hard part. (Building one simple system should be easier than the complex accretion of systems we have today.) The hard parts are sociological -- getting actors in the current system to give up control, and perhaps their market niches, to cooperate in building a new system -- and also, perhaps above all, legal. Laws exist, and many of them were written decades ago and are not well adapted to the conveniences of modern technology.
Earlier this week, I wrote about LendingClub's recent legal troubles, in which its push to simplify and disintermediate lending may have led it to cut some legal corners with the loans it was selling to customers. But I didn't want to put too much blame on the "financial technology" mindset. I said: "The fintech industry, after all, didn't invent fintech." Securitization, the MERS mortgage registry: These were attempts to transcend the old-timey legal structures of lending and to create abstract, technology-enabled, fast and convenient systems for trading and owning loans. They mostly worked! But there were glitches, because the clever new systems were built on top of inconvenient old rules. And just because your new system is cleverer and faster, that doesn't mean you get to ignore the old rules.
The DTC/Cede/street name/etc. system is another case of pre-fintech fintech, a privately coordinated system for simplifying and electronifying the cumbersome traditional processes of owning and trading and voting shares. It's just that the way to simplify those processes often made them more complicated in other ways. Perhaps the next round of financial technology will avoid that difficulty. But complexity is a stubborn survivor; often, when you think you are getting rid of complexity, you are only moving it somewhere else.
I mean, in the normal merger case, it's someone (or someones) who work for the company. In proxy fights, like where a hedge fund is running its own nominees for the board, the hedge fund can solicit shareholders to sign its own proxy card, giving voting authority to someone who works for the hedge fund.
This post will eventually be about Dell, so I refer you to the proxy statement for Dell's 2013 buyout. The proxy card is at the very end; it says:
You hereby (a) acknowledge receipt of the proxy statement related to the above-referenced meeting, (b) appoint Lawrence P. Tu and Janet B. Wright, and each of them, as proxies, with full power of substitution, and authorize them to vote all shares of Dell Inc. common stock that you would be entitled to vote if personally present at such meeting and at any postponement or adjournment thereof and (c) revoke any proxies previously given.
This proxy will be voted as specified by you. If no choice is specified, the proxy will be voted according to the Board of Director recommendations indicated on the reverse side, and according to the discretion of the proxy holders for any other matters that may properly come before the meeting or any postponement or adjournment thereof.
On the back, there are some check boxes to vote For, Against, or Abstain on the merger and a couple of other matters.
Also phone and Internet voting instructions, because who sends snail mail?
By the way, it's usually more complicated than that, too: Most companies have outside transfer agents to keep track of their shareholder records. From Wednesday's Dell decision:
Dell outsourced the obligation to maintain its stock ledger and generate a stock list to its transfer agent, American Stock Transfer & Trust Company, LLC (“American”).
See, e.g., page 136 of the Dell proxy statement:
Stockholders of record will be able to vote in person at the special meeting. If you are not a stockholder of record, but instead hold your shares in “street name” through a bank, broker or other nominee, you must provide a proxy executed in your favor from your bank, broker or other nominee in order to be able to vote in person at the special meeting.
Most of them. In the Dell case, the proxy says that there were 1,755,951,717 shares outstanding as of the record date. Wednesday's Chancery decision says that Cede & Co. was the record holder of 1,535,558,891 of them, or about 87 percent.
There were previous cases in which appraisal-arbitrage investors bought shares after the record date for voting on the merger (which is usually weeks before the meeting date). So they didn't get to vote, and companies argued that they shouldn't get appraisal because they couldn't prove that the shares they bought had been voted against the deal. The courts decided that this was unfair, and since Cede actually owned all the shares anyway, the appraisal arbs could count Cede's votes against the deal. That is, as long as Cede voted enough shares against the deal to "cover" all the people seeking appraisal, the arbitrageurs could assume that all the shares they bought were shares that voted no, in the absence of evidence the other way.
That was rough justice, but again: in the absence of evidence the other way. Here, we know that the T. Rowe funds actually voted for the deal, so there is not much of a fairness argument for pretending that they'd voted against. From the decision:
The evidence showing how Cede voted particular blocks of shares provides a basis for distinguishing the Appraisal Arbitrage Decisions. Under those opinions, an appraisal petitioner that held in street name can establish a prima facie case that the Dissenter Requirement was met by showing that there were sufficient shares at Cede that were not voted in favor of the merger to cover the appraisal class. This showing satisfies the petitioner’s initial burden and enables the case to proceed. If there is no other evidence, then as in the Appraisal Arbitrage Decisions, the prima facie showing is dispositive.
The analysis, however, need not stop there. Once the appraisal petitioner has made out a prima facie case, the burden shifts to the corporation to show that Cede actually voted the shares for which the petitioner seeks appraisal in favor of the merger.
And since we can trace how T. Rowe/State Street/Broadridge/Cede/etc. voted these shares, and since they voted for the merger, they lose.
Incidentally, record dates -- where you have to own the shares weeks before the meeting to vote the shares at that meeting -- are kind of another weird archaism.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
To contact the author of this story:
Matt Levine at email@example.com
To contact the editor responsible for this story:
James Greiff at firstname.lastname@example.org