Sunday, June 11, 2017

Why Plato and Ed Schreyer Were Right: Those at the Top Should Be Paid No More than Three or Four Times What We Pay Those at the Bottom

Through the 1950s and early 1960s—considered by many to be a golden age of corporate well being and of growth for the economy generally—CEOs of large businesses typically made about 20 times what the average worker in those organizations made. But the seeds for the explosion in executive pay that followed had already been sown. In 1951 Arch Patton, an executive at the management consultant firm McKinsey & Company, had developed a survey of executive compensation that compared executive pay not with the pay of others lower down in an organization, but with executives at other organizations. Understandably, the survey came to be closely followed by executives, and the results were published annually in The Harvard Business Review. Patton expanded on the ideas underlying the survey in an influential 1961 book, Men, Money and Motivation, arguing that corporations would be well served by knowing how the pay of their top people compared with the pay of other corporations; “when compensation of executives is above the industry average, above-average demands can be made on the executive group.” Boards of Directors listened, and compensation levels for those at the top increased significantly through the 1970s and 1980s. Much of that compensation, though, was opaque; it was difficult to assign a monetary value to deferred payment schemes and complicated stock option plans. The 1992 decision by the US Securities and Exchange Commission to require full and transparent disclosure by corporations of the compensation paid to their top five officers changed all that. Often described as an attempt to “help investors keep compensation in line” (as The Globe and Mail put it in a 5 December 2015 article), the move was in fact designed merely to increase transparency, with no judgement made as to the desirability of reining in executive wages. The previous system of providing information as to what companies were actually paying their top people was described by SEC Chair Richard Breedon described as “impenetrable”; he complained that it “insulated management from accountability to shareholders.”

Whatever its intent, the introduction of the new system became a landmark in the history of inequality. Suddenly boards and shareholders could see clearly the value of total compensation packages, and compare them to the packages other companies were offering. Far from choosing to rein in compensation, boards became ever more determined to ensure that compensation for their organization’s top people was “above the industry average.” With everyone continually striving to be above the average, the result was not hard to foresee. Average executive levels soared ever further beyond those of the average worker, and further still beyond those of workers at the bottom of the wage scale. We are now faced with unprecedented levels of inequality—inequality that is not only unfair in itself, but also undermines the cohesiveness of society, and undermines the economy itself, as more and more capital is devoted to savings socked away by the rich, and less and less to expenditures that drive further economic activity. (That such expenditures have continued at relatively high rates is largely attributable to many people carrying debt loads that are, in the long run, unsustainable.)

In such circumstances it seems obvious that we should return to measuring compensation for people at the top by comparing them to people in the middle or at the bottom of their own organizations—and that we should aim to set maximum multiples. But just where should those maximums be set?

In the fourth century BCE Plato suggested than in human societies governments should “permit a man to acquire double or triple, or as much as four times the amount [that is deemed to be at poverty levels]” (Plato, Laws V, sect 744). In the early twentieth century the financier J.P. Morgan—the epitome of American business establishment values of the day—is said to have recommended that those at the top should be paid no more than twenty times what the average worker made. In 1940 George Orwell called for “limitation of incomes, on a scale that the highest tax-free income in Britain does not exceed the lowest by more than ten to one” (“The Lion and the Unicorn,” Part 3, Section 2). In the 1970s management guru Peter Drucker argued in The Wall Street Journal for “a published corporate policy that fixes the maximum compensation of all corporate executives…as a multiple…of the lowest paid regular full-time employee,” and suggested maximum multiples ranging from 15 to 1 for smaller businesses to 25 to 1 for large multinationals. Also in the 1970s, Manitoba premier Ed Schreyer (later to become Canada’s Governor General) aimed “to bring about greater equality…and reward the dignity of work,” proposing that the ratio of incomes for those at the top of the income scale to the incomes of rank-and-file workers be no more than 2.5 to 1. (Errol Black and Jim Silver, “Manitoba's NDP: time to return to its social democratic roots, Canadian Centre for Policy Alternatives, 15 August 2012).

A proposed multiple of 2.5 between the highest and the average might translate into a multiple of 3.5 or 4 to 1 between the highest and lowest paid—or, after tax, a multiple perhaps closer to 3:1. That range has always felt intuitively appropriate to me, but I’ve never seen any attempt to present an extended argument for the appropriateness of that or of any other multiple. That’s what I want to do here.

I should preface the argument by making clear that such multiples should, to my mind, apply to all organizations—not just to corporations. Given the extraordinary multiples at many large corporations, it’s understandable that they have been the focus of most discussions. University professors or hospital administrators may think of their organizations as being relatively free of inequality. Yet the levels of inequality in those organizations are now typically greater than were corporate levels of inequality in the 1950s and 1960s. A hospital cleaner in America makes $11.19 an hour or $22,380 a year, while an orthopedic surgeon working in the same hospital may make well over $400,000. A cleaner at an American university working 40 hours a week makes, on average, $10.76 an hour or $21,520 a year, and at some universities an adjunct instructor working full time (teaching 8 one-semester courses per year) will make barely more than that—$22,000-$28,000 at institutions where adjunct professors are paid $3,500 or less per course taught. Meanwhile, the average for American university presidents has reached $500,000 a year.

So why a multiple of 3? Why that number specifically?

Let’s start by thinking of a normal work day of 8 hours, and two salaried workers. Perhaps one of the two may worker harder than their colleague—perhaps even twice as hard. So to provide fair compensation, arguably, she should receive an additional eight hours worth of pay. Suppose on top of that she doesn’t just work harder; she also works longer. Let’s say she always works twelve hour days, compared to the eight hour days worked by her colleague. In my experience (now more than 40 years in the workplace), there are in fact very, very few people who actually put in twelve hour days as a regular practice—and most of the ones who actually do so are not working flat-out for the full twelve hours. But for the sake of argument let’s say this person is one of those very, very few; let’s allocate to her on that basis a further 4 hours worth of pay.

We’re now up to 2 ½ the level of the fellow worker. But perhaps that’s still not enough. Maybe the person in question manages somehow to work more than twice as hard as her colleague, or somehow does manage to put in more than 12 hours of flat-out work per day. As compensation for this extraordinary individual, let’s arbitrarily allocate to her a further amount equivalent to 50% of her fellow worker’s pay, bringing her up to three times what the fellow worker makes.

But surely that’s still not enough, you may say. It’s not just that this person—let’s say she is the CEO or the President of the organization—it’s not just that she works much harder than others and much longer hours; it’s also that she bears much more responsibility. Another argument for paying more to heads of organizations is based on weight of responsibility: the larger the responsibility, according to this argument, the higher the rewards should be. Most of us feel intuitively that there is something to this argument, and to some extent I think so too; if I didn’t think so, I would not have felt it reasonable for the company that I have headed to pay me almost three times as much as our lowest paid employee when I am sure that I have not worked more than twice as hard, or more than a twelve hour day (an average of more like 9.5 is more like it). But there’s far less to the argument that greater responsibility deserves greater pay than is commonly assumed. The CEO or President of any organization bears substantial responsibility, of course—but that responsibility is arguably less than the responsibility borne by many ordinary workers in everyday occupations. Someone who heads up a company that manufactures Kleenex, or that processes accounting information shoulders important responsibilities—but are they an greater than those shouldered by a day care worker or a camp counsellor—let alone those shouldered by a pediatric nurse? People working in those sorts of capacities constantly shoulder life and death responsibilities involving small children. If weight of responsibility is to the basis on which pay differentials are determined, surely a pediatric nurse should make far, far more than the CEOs of a great many companies. And let’s not just look to jobs with obviously important responsibilities. Consider for a moment the job of hospital cleaner. If these workers don’t do their jobs very thoroughly and very well, the consequences to human health are almost incalculable; the spread of c difficile and other deadly “superbugs” becomes far, far more likely. Yet these are among our lowest paid workers. (Over the past generation they’ve become even lower paid than they were, as hospitals have outsourced jobs such as cleaner to companies that will pay the workers less than the hospital would be obliged to; instead of being valued members of the hospital team, such workers are now as marginalized as they are poorly compensated.)

What about the unpleasantness of the job, and the stress? I think there’s a lot to be said for the idea of paying more to people whose jobs are very unpleasant to compensate for that. But again, if we truly believed in that principle, the extra rewards would go to those who clean the toilets in airports and hospitals, and who scrounge the garbage dumps of the developing world for a living—not to the CEOs of large corporations. And stress? Yes, it’s stressful to lead an organization—but numerous studies have shown that the sort of stress that damages health and shortens lives is far more common among those who struggle to make ends meet than it is for those who head large organizations.

But wait, many of you may say. Those who are paid large amounts often have far more education than the average worker. Surely they should be paid more in recognition of the level of education they have attained.

But should they? Let’s think the matter through from first principles. Higher levels of education typically enable the beneficiaries to have much more interesting jobs than do those with little education. Should they benefit from their education both by being able to obtain more interesting work and by being paid much more for doing it?

Doctors often argue that they deserve very high levels of pay because they suffered through (and paid for) so many years of school. But by far the greater part of the expense of their education is borne by the taxpayers—most of whom make no more than a fraction of what doctors make for the rest of their lives once they do finish their education. Does a doctor really deserve to make more than three times what the average taxpayer makes?

What, then, of one’s ability to do a job effectively? It’s fair to point out that those who are paid huge amounts often have huge talents—and as a result do indeed work more effectively than many others. But why, precisely, should humans be compensated for ability? No doubt some people are able to achieve considerably more than others, but let’s set to one side that part of their achievements that has resulted from their education and that part of their achievement that has resulted from working harder than others. What we’re left with are the abilities one is born with, and the opportunities one’s family was able to provide during one’s upbringing. To be sure, genetic difference and differences in family background create large differences in the levels of ability of different individuals. But such differences are surely a matter of good or bad luck, not of dessert. Does someone who has had the good luck to be born with an extraordinarily high level of intelligence (or with extraordinarily good hand-eye coordination, or with extraordinary talent of any other sort), or someone who has had the good luck to be born into a wealthy family that is able to provide for them a privileged upbringing, deserve to be paid more because of the higher level of ability that is the result of those genes and that upbringing? To pay such people more on those grounds amounts, in essence, to paying them more for having been lucky. What possible justification could there be for building luck money into our system of compensating workers? Arguably, it would be at least as fair in the other direction to compensate people for having been disadvantaged in the talent they were born with, or the upbringing they had—or the opportunities they never had.

What of what some will call the “practical argument”—that, regardless of the rights or wrongs of it, one will not be able to attract and keep top talent unless one pays top dollar?

Those who make this argument overrate three things. They underrate, first of all, the importance of financial incentives to human beings. There will always be many talented individuals willing to try for interesting and demanding jobs, even if at they are not paid astronomical sums for doing so; if that were not true, countries would have no one willing to serve in Cabinet positions, or in the senior levels of the bureaucracy (where most people make only a fraction of what they could make with large corporations). For most people—including most talented people—money is not everything. It’s not even the most important thing.

Those who make the so-called “practical argument” underrate too the degree to which talent is scarce; in any large organization there are almost always many individuals capable of filling the top jobs. Finally, they underrate the degree to which organizations owe their success or failure to one individual (or to a handful of individuals) at the top. During the start-up phase, to be sure, many organizations are disproportionately reliant on the abilities and the vision of the founder(s). But the evidence suggests that, in the case of large, established organizations, who the CEO is usually makes relatively little difference to how the organization performs.

Does a company, then, really need to pay its CEO tens of millions of dollars a year? The facts of the matter simply don’t bear that out. Many North American companies have done exceedingly well without ever paying their CEO at anything like the multiples that most large institutions offer their top people. And in the other direction there are many, many companies that have seen their fortunes decline drastically despite paying the CEO tens or even hundreds of millions of dollars. And again, North American corporations thrived in the 1950s and 1960s despite far, far lower multiples of CEO to average worker compensation.

One should also look to outside North America. American corporations may have gotten into the habit of paying their CEO a fortune every year, but that’s been much less the case in other countries. In Japan the head of a large organization can expect to make ten or twenty times more than what those lower down the ladder make—not four hundred times more, as in the United States. But no one I’m aware of has suggested that Honda or Toyota would be far more successful if those organizations paid their top people in the way that American companies pay their CEOs.

* * *
Many of the above numbers are multiples of the average worker—not of the lowest paid. What is the rationale for referencing the lowest paid rather than the average? About that there can be no mystery. Western societies have become far too much in the habit of focusing on the incomes of the middle class—and far too little on the incomes (and the wealth, or lack thereof) of those at the bottom of the ladder. If we focus only on a comparison of the highest paid with those in the middle, it becomes ever easier for society as a whole to keep forgetting about its least fortunate members—and for inequality between the middle and the bottom to keep growing, even as inequality between the middle and the top begins to shrink.

* * *
I have been trying to put forward a case for 3:1 being an appropriate multiple. Is it possible to prove that this is precisely the appropriate number? Of course not. By their very nature these issues will always involve a large subjective element. But I cannot believe that something in the range of three or four is not a more appropriate multiple than J.P. Morgan’s 20—let alone than the 350-400 that has become the average in large American corporations.

Apparently I’m not alone in my intuitions. Psychology professor Kevin Payne’s new book The Broken Ladder: How Inequality Affects the Way We Think, Live, and Die (the subject of a 4 June 2017 Nicholas Kristof column in The New York Times) reports on surveys as to what members of the general public feel to be appropriate pay multiples. Interestingly, there is no great divide between left and right: “liberals said CEOs should be paid four times a much as the average worker, while conservatives said five times.”

A multiple of three (or four, or five) is still a very long way from Marx’s “from each according to his abilities, to each according to his needs.” A multiple of three or four does not eliminate inequality; it still accepts the value of monetary incentives for people to work hard, to strive to succeed. But a multiple of three or four would mean the end of “luck money” as a guiding principle of compensation. It would be the basis for a far fairer, far more livable society. Let’s go there.

CEO Compensation

(The following is the full version of a letter to The Globe and Mail that was published in slightly edited form on June 1. I have been developing a longer argument on this topic; I will post that shortly.)
I may be unusual among CEOs in entirely agreeing with Mark Roberts (letters, May 30) that CEO compensation should be limited to a certain multiple of the compensation of a corporation’s lowest-paid employee. I remember that many years ago a Manitoba premier suggested this idea—and suggested that three would be an appropriate multiple. That sounds about right to me—and it’s a multiple that I don’t think our corporation has ever exceeded in its 32 years. It’s just about impossible to work more than three times as hard as someone else; there are only 24 hours in a day. Should one be paid more for having more ability or education? Perhaps, but arguably it’s at least as fair in the other direction to compensate someone for having been disadvantaged in the talent they were born with, or the upbringing they had, or the educational opportunities they never had.