Understanding ever increasing executive pay

Ed Balls announces that, if Labour wins the next election, he will reintroduce a 50p top tax rate (reduced to 45p by Osborne). Assorted captains of industry say that this would be disastrous for the UK economy. And I remembered a recent post by Justin Fox about why “We Can’t Afford to Leave Inequality to the Economists”. While I would not object to his point that inequality involves political and moral issues as well as economic ones, I think he misses a key point. We need economists to provide a narrative of why the pay of the top 1% has surged ahead in the US and UK since the 1980s, a narrative to counter the claims from the 1% themselves that it just represents the market rewarding skill and productivity. This post talks about one narrative that I think has great power, and also has important implications for that top tax rate.

This narrative matters because change has to be mediated through politics. Janan Ganesh in the FT says that although restoring the 50p rate in itself is popular in the UK, in the end voters may be more swayed by business leaders saying it (or a Labour victory more generally) will damage the economy. As I have noted before, Labour under Blair, Brown and to an extent Balls went out of their way to be business friendly and woo the business sector, because they thought this was essential to electoral success. (Robert Peston, who is a better position to know, agrees.) Most of the reasons why they thought this have not gone away.

There is a widespread belief that there is too much inequality in the UK and US, while at the same time the public underestimate the degree of inequality that actually exists. Yet arguably elections get won or lost on who the electorate believes is competent to ‘manage’ the economy. If political parties that aim to do something about growing inequality also appear not to enjoy ‘the confidence of business’, then they may not get elected. We need people who have some knowledge and objectivity about the economy and markets to argue that those that speak for business are actually just speaking for their own personal interests.

 

So what is the alternative story to the argument that executive pay reflects the market rewarding the rising productivity of CEOs? The first obvious point is that executive pay is not determined in anything that approximates an idealised market where prices are set to balance supply and demand. Instead it is set within a bargaining framework between employer (the firm) and employee (the CEO). Even if we imagine the employer in this case to be someone that genuinely reflects the interests of shareholders, the costs associated with losing your CEO, together with informational problems in assessing their true worth (which can lead to the age old problem of judging quality by price, and the ‘arms race’ that Chris describes), mean that the CEO potentially has substantial bargaining power.

Yet this situation did not suddenly arise in the 1980s, and it will be true in most countries, and not just in the US and UK. So why did executive pay start taking off in the 1980s in these two countries? Well something else happened at the same time: tax rates on top incomes were also substantially reduced. Why does reducing the tax rate on top incomes lead to a rise in those incomes pre tax? With lower tax rates, the CEO has a much greater incentive to put lots of effort into the bargaining process with the company. They, rather than the tax man, will receive the rewards from being successful.

This is the idea set out in this paper by Piketty, Saez and Stantcheva [3]. They call this a “compensation bargaining” model. The paper backs up this theoretical model with evidence that there is a “clear correlation between the drop in top marginal tax rates and the surge in top income shares”. In addition, they present microeconomic evidence that CEO pay for firm’s performance that is outside the CEO’s control (i.e. that is industry wide, and so does not reflect personal performance) is more important when tax rates are low. (Things like stock options.)

Now one reaction to this model is that it ignores many other social/economic factors that may also have been important. Things like changing social norms and political changes (loosely, the rise of neoliberalism), reduced union power, changes in financial regulations, growing financialisation etc. I think this reaction is correct, but as the authors themselves say, such explanations are “multi-dimensional and it is difficult to estimate compellingly the contribution of each specific factor”. Economists like simple models that can be tested against the data. That is what the compensation bargaining model set out by Piketty et al does. I don’t think it is too much of a stretch to think about bargaining effort as a proxy for all these other factors. [1]

There is a nice parallel between the compensation bargaining model and the union bargaining model popular outside the US in the 1970s/80s, which made many economists somewhat antagonistic to growing union power. There is a difference. There union power distorted the economy by raising the real wage, and generating involuntary unemployment. In the compensation bargaining model, increasing executive pay is just a rent-seeking redistribution, and is socially costly only because effort is wasted on bargaining. However as it involves redistribution to the 1% from the 99%, I don’t think many besides economists will worry about that too much. (Economists and others have, of course, begun to discuss some of the perhaps more important indirect costs of this inequality. The social costs are documented in a comprehensive way here, but there is also the distortion of representative democracy (see here, here, here, here and here) or encouraging the portrayal of poverty as self-induced.)

The compensation bargaining model has a clear policy implication. The problem with lowering the top rate of income tax is that it encourages the executive class to engage in efforts to raise their pay at the expense of everyone else. We need a high top rate of tax to discourage this, even if this rate might not actually bring in more income. [2] Perhaps most importantly, it provides a plausible alternative to the market rewarding effort narrative that is so frequently used, and it also has the advantage of being closer to the evidence. 

[1] For an analogy, think about central bank independence. There are many reasons why you might be concerned about politicians being able to set interest rates, and why control by independent central banks is preferable. Macroeconomics settled on one particular idea, that of time inconsistency and inflation bias. Was it because all economists thought this was really the most important problem. I suspect for some at least it became a proxy for rather more general, but therefore vaguer, stories.

[2] Technically, the tax rate is raised above the conventional optimum by an amount that represents a Pigouvian correction to the rent-seeking externality.

[3] Now published in American Economic Journal: Economic Policy, Vol. 6, Issue 1, February 2014