Friday, September 4, 2020
I Just Read the Department of Labor's New ERISA Voting Proposals and Boy Are My Fingers Tired (from typing)
I’ve talked before in this space about how regulation of ERISA plans and ERISA plan voting is really part of a larger debate about the proper role of shareholders in corporate governance, and even whether the purpose of the corporation is to maximize shareholder value. And a few weeks ago, I blogged about how the Department of Labor had proposed new rules/interpretations to limit ESG investing for ERISA-governed retirement plans, and promote investments in private equity.
Apparently to honor Labor Day, the DoL took it a step further this week to limit ERISA plans’ ability to vote their shares, in large part because – it says – of excess costs due to “the recent increase in the number of environmental and social shareholder proposals introduced. It is likely that many of these proposals have little bearing on share value or other relation to plan interests…”
A lot of words after the jump. So many words.
It all begins when the Department of Labor began promulgating guidance in the late 80s and early 90s to the effect that votes are part of plan assets and fiduciaries need to take appropriate care in handling them. In a series of bulletins, the DoL made clear that votes should be cast when it is cost justified to do so – i.e., expected benefits exceeded costs – and that votes must only be cast in the financial interests of the plan rather than to advance other interests. (Some discussion and links here)
In the waning days of 2016 – essentially, just before Trump took office – the Department of Labor issued some clarifications to its policy. It continued to adhere to the position that proxy votes must only be cast to benefit the plan’s financial interests, and that voting would not be necessary at all unless it was cost-justified given the potential benefits. But, said the Obama administration, given the increasing popularity of ESG voting and engagement, a reasonable plan fiduciary could make a general judgment that engaging on ESG strategies would add to the value of the investment; additionally, because the plan would usually employ responsible professionals to manage voting and other matters, any particular proxy vote would usually not incur additional expenses and, absent special circumstances, there would probably be no need to conduct a vote-by-vote cost benefit analysis.
So, now the Trump Administration has proposed new rules. And in brief, they say that blanket voting policies are not cost-justified unless they favor management, and if by some chance a non-management vote is cost-justified, the DoL will attach as many recordkeeping requirements as necessary in order to ensure that is no longer the case.
I’ll get into the specifics in a moment, but the first thing to note is that the DoL begins by summarizing the previous interpretive bulletins and other guidance on this issue. And when it does so, it emphasizes its consistent policy of requiring that fiduciaries vote solely in the financial interests of the plan, and that votes are not required if they are not cost-justified based on expected benefits. But then it says that to avoid confusion, it will repeal the Obama-era bulletin – and no other one – and strike it from the CFR. Which would of course lead one to ask, why does it have to repeal the Obama bulletin if the DoL’s position has been so consistent over the years? The answer to that apparently – as I said above – is because Obama would permit blanket policies, and assumed individual votes are probably costless, and that’s what the DoL wants to target. I did not see anywhere in the release, however, where the DoL admitted that the new proposed rules represent a change from that 2016 interpretation.
Why are these new rules necessary, according to the DoL?
A couple of reasons, including:
Because so many ERISA fiduciaries incorrectly believed that they were required to always cast their ballots. What is the evidence that this is a common misapprehension? The DoL’s entire body of evidence is in footnote 12 of the release, and as far as I can tell, the only people who are arguing that ERISA fiduciaries believe they must always vote their shares are commenters representing organizations like the Business Roundtable and the Washington Legal Foundation.
Also, the new rules are needed because of “mixed” evidence on the value of casting a vote that has been developed since the earlier interpretations. Leaving aside my hurt feelings – shareholder votes don’t matter? What am I even doing with my life? – let’s just say that there’s been a lot of new thinking about this issue, especially in more recent years as institutional investors have begun to exert more influence. I’ll also highlight a new paper by Yaron Nili and Kobi Kastiel, Competing for Votes, which talks about how the mere fact that directors have to worry about shareholder votes makes them more responsive to shareholders and more willing to generate useful information. And, as discussed below, the DoL has a very narrow view of the value of a vote.
With that as background, what does the rule do?
Well, the entire proposal is framed around this text from the proposed rule:
A plan fiduciary must vote any proxy where the fiduciary prudently determines that the matter being voted upon would have an economic impact on the plan … taking into account the costs involved…
A plan fiduciary must not vote any proxy unless the fiduciary prudently determines that the matter being voted upon would have an economic impact on the plan … taking into account the costs involved…
Plus, the fiduciary must document the calculations behind each vote. Per the rule, the fiduciary must:
Investigate material facts that form the basis for any particular proxy vote or other exercise of shareholder rights…. Maintain records on proxy voting activities and other exercises of shareholder rights, including records that demonstrate the basis for particular proxy votes and exercises of shareholder rights
If the fiduciary relies on a service provider to vote the shares or make recommendations, the fiduciary:
shall require such investment manager or proxy advisory firm to document the rationale for proxy voting decisions or recommendations sufficient to demonstrate that the decision or recommendation was based on the expected economic benefit to the plan…
I will leave it to others to discuss how vote-by-vote cost benefit determinations are to be made when a plan uses an advisor who serves multiple clients but I will point out that the DoL notes “Particular attention must be given to proxy advisory firms that provide both proxy advisory services to investors and consulting services to issuers on matters subject to proxy resolution” which is a not-exactly-subtle dig at ISS, and critically, ERISA fiduciaries cannot simply rely on SEC advice to investigate conflicts because “the SEC standards do not necessarily capture all the actions that ERISA may require…”
What leaps out here?
First, these are incredibly onerous requirements, if you consider a plan may hold interests in hundreds or thousands of companies, each of which has a shareholder meeting every year with multiple items on the ballot. DoL recognizes as such; it interprets the rule to require “an analytical process which could in some cases be resource-intensive (requiring, among other things, organizing proxy materials, diligently analyzing portfolio companies and the matters to be voted on, determining how the votes should be cast, and submitting proxy votes to be counted).” It recommends that fiduciaries consider such information as “the cost of voting, including opportunity costs; the type of proposal (e.g., those relating to social or public policy agendas versus those dealing with issues that have a direct economic impact on the investment); voting recommendations of management; and an analysis of the particular shareholder proponents.”
(That reference to shareholder proponents is, presumably, intended as a poke at “corporate gadflies,” even though – as Yaron Nili and Kobi Kastiel point out – these may propose widely popular corporate governance reforms.)
Put a pin in the “social or public policy agenda” thing for a moment.
Second, what is “economic impact” on which the requirement to vote, or refrain from voting, turns? Most of the time, DoL refers to the “economic interests of the plan,” but occasionally it refers to the impact on the particular investment standing alone. For example (emphasis mine in all cases):
“those relating to social or public policy agendas versus those dealing with issues that have a direct economic impact on the investment”
“the Department proposes several potential options for fiduciaries to consider that are intended to reduce the need for them to consider proxy votes thereby freeing resources for fiduciaries to focus on activities most likely to have an economic impact on the plan’s investment”
“even small voting costs may somewhat impair participants’ financial interest in their benefits if the votes pertain to issues that have little or no bearing on share value or are otherwise immaterial to the plan’s financial interest.”
“The expenditure of plan resources is generally warranted only when proposals have a meaningful bearing on share value or when plan fiduciaries have determined that the interests of the plan are unlikely to be aligned with the positions of a company’s management. In general, such proposals include those that are substantially related to the company’s business activities or that relate to corporate events…”
“plan resources may not be expended in circumstances where the fiduciary prudently determines that a proxy vote would not affect the economic value of the plan’s investment.”
“mixed evidence regarding whether proxy voting affects firm value”
In fact, the final rule almost seems to define the concept of economic impact on the plan by reference to the impact on the investment alone. Per the rule, a fiduciary must (emphasis mine):
(A) Act solely in accordance with the economic interest of the plan and its participants and beneficiaries considering only factors that they prudently determine will affect the economic value of the plan’s investment based on a determination of risk and return over an appropriate investment horizon consistent with the plan’s investment objectives and the funding policy of the plan;
(B) Consider the likely impact on the investment performance of the plan based on such factors as the size of the plan’s holdings in the issuer relative to the total investment assets of the plan, the plan’s percentage ownership of the issuer, and the costs involved
And when the DoL offers citations to support its argument that votes don’t matter, it cites as evidence David Yermack’s observation that “Activist institutions frequently state that their goal is not to improve the value of individual investment positions, but rather to create positive externalities by signaling optimal governance practices market wide,” as though this effect – little impact on the subject company but wide market effect – is proof that the vote itself may be immaterial.
This is not a minor distinction; as I’ve described in this space, and written an article about, shareholders may decide not to maximize wealth at a particular company in order to benefit other investments in the portfolio. We’re talking about everything from how investors who have stakes in multiple companies within an industry may prefer that they not compete too vigorously to the possibility that shareholders may prefer to see companies reduce their carbon emissions – even if doing so harms a specific investment – because climate change harms everything else. Indeed, this is one of the justifications for ESG: it addresses systematic risk across a portfolio (John Coffee just posted on that very issue).
(I’m not saying all of ESG does; I’ve argued before that demands for ESG disclosures are often not really investment related, but that’s not the same as indicting the entire field.)
And then there’s the more mundane stuff, like how an investor who owns both stock and debt of a single company may vote against risky maneuvers that boost share prices in order to protect the value of the debt. So, point being, if economic impact is defined on an investment-by-investment basis, it would mean that, for example, anyone who held shares in both Tesla and SolarCity would have been advised to vote the Tesla shares against the merger and the SolarCity shares for. That kind of irrational voting is something Leo Strine, for example, has specifically warned against. See Leo E. Strine, Jr., Securing Our Nation’s Economic Future: A Sensible, Nonpartisan Agenda to Increase Long-Term Investment and Job Creation in the United States, 71 Bus. L. 1081, 1093 (2016).
More generally, this is the problem with thinking of voting as a single action within a single context. Votes have power because they are a collective act, both with other shareholders in the context of the particular matter under consideration, and for a signal they send to the market with respect to investor preferences more generally. Shareholders at a few companies, for example, cast ballots for majority voting, and declassifying boards, and proxy access, and it led to widespread and voluntary adoption of these measures across a large swath of the market. So the value of a vote goes well beyond that particular contest and that particular company, in ways that can broadly impact a portfolio. Indeed, this fact is acknowledged in the very Yermack paper that the DoL cites in its release, but the DoL omits the relevant portion. The full Yermack quote reads, “Activist institutions frequently state that their goal is not to improve the value of individual investment positions, but rather to create positive externalities by signaling optimal governance practices market-wide, potentially improving the value of the institutions’ other diversified investments.” (emphasis mine)
The DoL does not acknowledge this point in its discussion of the “mixed” evidence of the value of votes, and elides it by intermittently referencing only the “economic value of the plan’s investment.”
The point being, the Obama-era interpretation would allow plans to make a general judgment that certain kinds of arrangements - whether reduced carbon emissions or single-classed boards - add value to the portfolio generally, and adopt a voting policy that favors those arrangements. The DoL rule does not allow that kind of policy-based voting, so that each vote must be individually calculated anew, which necessarily adds to plan burdens and expenses, to the point where each individual vote may no longer be cost-justified. And, at least in some parts of the release, the DoL does not seem to recognize how the value of a vote with respect to one company may in fact have plan-wide effects.
One note, though: The DoL does say that if the plan engages an outside advisor to vote the shares, that advisor must consider the needs of each individual plan it advises and justify votes on that basis – which is clearly a requirement that is more onerous when votes are considered as part of the whole plan, and not on a company-by-company basis, but I don’t disagree with the general idea; I’ve argued before that mutual fund families, for example, should not too readily assume that all funds have similar voting preferences.
Third, the DoL says: “fiduciaries must be prepared to articulate the anticipated economic benefit of proxy-vote decisions in the event they decide to vote,” which I loosely translate to mean, “We are going to make an example out of a couple of plans, so are you sure you want to cast that ballot?”
So these rules would make it very very hard to vote, is the point, and thus may functionally disenfranchise ERISA fiduciaries.
But that’s not all.
As above, the DoL recognizes the documentation costs may themselves exceed the benefits of a vote, and so helpfully, provides for certain “permitted practices” which plans can adopt as general rules for voting. As DoL puts it, these are “intended to reduce the need for fiduciaries to consider proxy votes that are unlikely to have an economic impact on the plan, thereby allowing plans to focus resources on matters most likely to have an economic impact.”
So as I understand it, these are general voting policies for matters unlikely to have an economic impact. And they were developed, the DoL says, in part based on its belief in the mixed evidence that voting has value.
The first such “permitted practice” is “A policy of voting proxies in accordance with the voting recommendations of management of the issuer on proposals or particular types of proposals that the fiduciary has prudently determined are unlikely to have a significant impact on the value of the plan’s investment…”
And which proposals are these? They include, DoL elaborates, precatory proposals. As DoL puts it, “a fiduciary could determine not to spend plan assets on proxy votes for nonbinding proposals, unless it is aware that such a proposal will somehow still have an economic impact on the value of the plan’s investment.” And, of course, there’s ESG in general: per the DoL, “It is likely that many of these proposals have little bearing on share value or other relation to plan interests.”
Let me stop you right there.
First, obviously, there’s a lot to question in the idea that precatory proposals have no economic impact. I mean, Congress requires a precatory say-on-pay vote, and I think it would be rather offended by the suggestion that the vote has no relevance. And we all can talk about how precatory proposals influence management even if they aren’t binding, etc, etc, and DoL knows this as well as anyone. I mean, you could argue that voting for directors is functionally precatory in a plurality voting system when there’s no contest on the table.
But also note, the rule – from the beginning of all of this – is that fiduciaries must not vote unless the vote will affect economic value. (Other quotes from the release: “the determination of whether or not the vote will affect the economic value of a plan’s investment portfolio is critical in triggering a fiduciary’s obligations under ERISA to vote or abstain from voting”; “If the proposal has no or negligible implications for the value of the plan’s investment, it would be better for the plan to simply refrain from voting than to incur even small costs making this determination.”; “fiduciaries must not vote in circumstances where plan assets would be expended on shareholder engagement activities that do not have an economic impact on the plan”).
Except now fiduciaries may vote on matters that do not affect economic value, but only if they adopt a blanket policy to vote with management. So there’s an exception, then, to the “don’t vote unless it has an economic impact” rule, and the exception is, voting with management is fine.
As far thumbs on scales go, this one’s pretty blatant, and it again signals that the object of the DoL’s ire is not wasted votes or valueless votes, but the mere concept of shareholder influence within the corporate form.
It’s genuinely unclear to me whether a blanket vote in favor of management even for impactful matters is okay; the phrasing suggests it’s focusing on nonimpactful matters only, but either way, the basic justification for the vote-for-management rule, the DoL explains, is that shareholders may rely “on the fiduciary duties that officers and directors owe to a corporation based on state corporate laws.” As I read that, shareholders don’t have to exercise oversight of management generally (unless, the DoL allows, they actually determine there’s a specific conflict) because state corporate law does that for them. The DoL even has a fairly hilarious footnote where it notes that shareholders can rely on managerial fiduciary duties because shareholders can sue to enforce them. Of course, as we all know, those rights are extremely limited because state corporate law relies on the shareholder franchise; it’s why the courts can take a hands-off approach to management matters, because they assume shareholders will do the monitoring, not the judiciary.
The second permitted practice is “A policy that voting resources will focus only on particular types of proposals that the fiduciary has prudently determined are substantially related to the corporation’s business activities or likely to have a significant impact on the value of the plan’s investment such as proposals relating to corporate events (mergers and acquisitions transactions, dissolutions, conversions, or consolidations), corporate repurchases of shares (buy-backs), issuances of additional securities with dilutive effects on shareholders, or contested elections for directors.”
Again, the focus is only on this particular company rather than portfolio-wide concerns, and the not-subtle implication is that shareholders have no role in monitoring management more generally.
The final permitted practice is that “a fiduciary may adopt a policy of refraining from voting on proposals or particular types of proposals when the plan’s holding of the issuer relative to the plan’s total investment assets is below quantitative thresholds that the fiduciary prudently determines, considering its percentage ownership of the issuer and other relevant factors, is sufficiently small that the matter being voted upon is unlikely to have a material impact on the investment performance of the plan’s portfolio.”
In other words, feel free not to vote if you don’t have a large stake. And as with other aspects of the rule, the general direction here is to treat each vote in isolation. But, again, votes are not isolated acts. They communicate preferences across the portfolio, thus forcing managers to attend to shareholders’ interests more generally. And the DoL knows this, of course; the goal of the proposal is to eliminate that possibility.
The DoL does say that its “permitted practices” are not exclusive; ERISA fiduciaries may develop their own general voting policies to avoid the expense associated with case-by-case analyses. But given the bent of the proposed rules, I suspect the DoL would not look kindly on, say, a fiduciary deciding to adopt voting policies favoring independent boards or disclosure of political spending (unless they had the backing of management, subject to those famously-enforceable state law fiduciary duties).
Upshot: The fairly transparent goal of these rules is to take ERISA plans out of the voting and engagement business, except for things in the mergers/spinoffs/contests category. DoL is trying to turn back the clock to the corporate governance ecosystem that existed in oh, the 1980s or so.
Whew. As per the subject line, my fingers are tired from typing all of that, so quick concluding thoughts:
First, I’ve said before in this space that Delaware and the federal government operate symbiotically; one pushes in a direction and it prompts the other to act. Delaware has given managers a freer hand in recent years precisely because it believes institutional investors can protect themselves. The more that the federal government discourages routine shareholder voting, the more that Delaware may feel compelled to step up its oversight. The 1980s, after all, were a really aggressive time for Delaware courts.
Second, ERISA itself doesn’t actually cover all that much anymore; it only applies to private pension plans, and as the DoL itself notes, ERISA plan assets are far, far smaller than they used to be, and a lot of those are in things like 401(k) mutual funds to which these rules would not apply. ERISA’s influence has tended to extend more broadly because a lot of state/local plans look to ERISA for guidance as to best practices. The more restrictive ERISA becomes, however, the more states may choose to break away; maybe a Calpers or a NYSCRF or a nonprofit might consider developing an alternative model fiduciary code for public pension plans, especially since – as Paul Rose has ably argued – a fiduciary standard that requires only consideration of the financial interests of the plan (ignoring negative externalities caused by profitable corporate behavior) is a poor fit for public plans.
Third, as I’ve said before, to call our modern corporate governance regime “private ordering” borders on the Kafka-esque.