A couple of months ago, David Larcker, Professor of Accounting at Stanford University, published his Seven Myths of Board Directors, one of which questioned whether classified boards were always detrimental to shareholders. Now comes a new academic study that claims that classified boards actually improve long-term value creation.
Classified boards are those where non-executive directors are split into three or four classes, only one of which comes up for annual election. It is typically thought of as an anti-takeover device because challengers must wait and win two or three election cycles. Classified boards are also used to protect directors from the threat of early removal by shareholders. Such boards have been the target of a campaign over the last few years by Harvard law professor Lucian Bebchuk that has very successfully declassified many of the boards in the S&P 500.
The new study, by Martijn Cremers of the University of Notre Dame and Simone M. Sepe of the University of Arizona, is based on a “new and expanded” dataset compared to the one previously used by a multitude of studies (including a seminal one by Bebchuk) which claim to show the exact opposite – that classified boards destroy value.
Before we go any further, let me first make the distinction, that is nowhere made in this study, between, say, three types of shareholder – activist hedge funds, large institutional shareholders and day (millisecond!) traders. Each has different aims and modes of operation – a distinction apparently lost on the study’s authors. It is important to bear this in mind as we consider some of the claims made by the authors.
The study claims that classified or staggered boards where directors have three-year terms outperform those with annual elections by 3.7% over 34 years. These results are used to defend board authority in the form of a classified board against short-term shareholder and market pressures, promoting long-term value creation, because directors do not have to worry about being unseated every year. A classified board, the study says, is a “rational response to market imperfections that are more complex and more significant than shareholder advocates generally realize”. Shareholder advocates, on the other hand, claim that classified boards diminish accountability and allow value destroying behaviour to go unchecked.Prior studies claiming that classified boards destroy value were based on a 1995-2002 timeframe that, says the study, does not have enough variation in classifying and declassifying boards. The dataset used by the new study runs from 1978 to 2011 and is that much broader.
The dataset is based on data gathered by the Investor Responsibility Research Center and, subsequently, RiskMetrics from 1990 to 2011 and on the authors’ own dataset from 1978 to 1989, constructed for a prior study. The study also notes that the classified board is “far from dead” because even at the end of the study period 47% of its sample of around 3,000 companies still has a classified board. Figures from MSCI for the largest 2,800 companies in the US, show a smaller proportion with a classified board, but this is largely because almost a 100 large cap companies have declassified in the last two to three years.
Furthermore, the study claims that the greater proliferation of declassified boards in the later period is based not on shareholders simply not consenting to them but rather on boards’ actions to declassify themselves. While this claim is not borne out by the number of successful shareholder proposals to declassify boards in recent years, it is also not possible to distinguish between unilateral board decisions to declassify and a decision to declassify because of confidential shareholder engagement.
Support for shareholders consenting to classified boards is put forward in the report by data that shows that some companies have classified their boards even during recent years when it would be seen as antithetical to shareholder power. However, this calls into question what kinds of companies these were, i.e. controlled or newly-floated companies perhaps?
I’m not qualified to question the statistical analysis of this paper, but I am qualified to question the assumptions that lie behind the theoretical pronouncements.
There are numerous problems with the hypotheses in the paper. For one, the fear of being unelected. How many directors get unseated every year? The study goes back to 1978 when no regular institutional shareholder group could unseat even the most unpopular director because of plurality voting – when directors just had to get one vote. It’s only in the last five years that majority voting has come in, so that shareholders voting against a director, rather than simply withholding votes, could have any effect.
But for around 30 years of this 34-year period, there was no real fear of being replaced at all – except in a hedge fund proxy fight. This undermines one of the central assumptions of this paper.
The study also claims that its new dataset allows it to control for firm variables and isolate board structure changes and their influence on firm value; something it says was not available to earlier contradictory studies. It posits that earlier studies actually revealed reverse causality. So that what was really happening was that companies were undervalued, they adopted a classified board, and this new board protection caused company value to rise. If this were indeed the case, this would seem to be a pretty major – and hitherto unnoticed – flaw in the earlier studies.
The third contentious claim is that classified boards commit shareholders to a longer-term engagement thus increasing wealth for all. While shareholders cannot quickly replace a classified board, they can still sell their shares, so it is difficult to understand how merely having a classified board commits them to a longer-term engagement.
The problems with a short-term commitment outlined by the paper, however, do not describe the actions of institutional shareholders. The shareholders that the paper routinely describes – those who focus on short-term market pricing – are not institutional shareholders at all, who own the vast majority of US stocks.
There is also a long discussion in the paper about how classified boards rely on long-term projects and optimal stakeholder (customers/suppliers) relationships as the main channels to growing value. How can it be claimed that these would also not be the main channels for value growth for a declassified board? It is also claimed that classified boards have a greater commitment to R&D, human capital, large customers and committed employees. It is the opposite, apparently, for declassified boards because short-term commitment from shareholders – who can, yes, again, remove the board or sell shares – focuses managers on the short term. But, again, how often does this actually happen?Finally, the paper also studies the single effect of the adoption of a classified board compared to the other – now infamous – Gompers ‘governance index’. Harvard Business School professor Paul Gompers’ 2003 study of shareholder empowerment and its effect on firm value claimed that if shareholders sold shares in firms with low shareholder empowerment and bought shares in firms with high shareholder empowerment they would see a marked investment gain. The new paper attempts to isolate the effects of the adoption of a classified board.
Its findings show that while the other negative shareholder characteristics do adversely affect a firm’s value, a classified board does not. In fact, it actually mitigates the negative effects of the other characteristics. It is again difficult to understand how this can be tested based on the data available; since so many of the firms studied had all the characteristics of management entrenchment it would seem difficult, if not impossible, to isolate the effects of a classified board in a large enough sample for results to be reliable.
So what are the conclusions of the study? The introduction of a “quasi-mandatory” rule to make classified boards the default governance structure, to outlaw shareholder proposals calling for board declassification, and a return to plurality voting (one vote and you are re-elected). The rule called for is “quasi-mandatory” because the authors believe that if there is agreement among shareholders, directors and management to adopt a declassified board, then they should be allowed to do so, though only with a two-thirds majority vote. The rule is aimed at preventing shareholders from coercing a board to declassify.
I can only imagine the reaction of most long-term shareholders to this attempt to roll back shareholder democracy and board accountability.
Paul Hodgson is an independent governance analyst.