Workers should not be discriminated against on the basis of their gender. This is axiomatic and is enshrined in the Fair Work Act. Unfortunately, workers have, and do, face discrimination from some employers.
The WGEA recently released data suggesting that female CEOs are paid $159k less than male CEOs. The WGEA CEO is quoted as saying:
‘It is concerning that even when you’re the most senior role in an organisation, there’s such a massive difference between, on average, what men and women are being paid.’
There’s just one problem: comparing compensation between CEOs is like comparing apples and oranges. And, there’s a corollary: this same problem applies across the workforce. Not all academics produce the same quality of work. Not all equity analysts are equally skilled. Not all investment bankers bring the same deal flow. The CEO compensation figures are a microcosm of a broader issue in the gender pay data.
CEO compensation is not based on an arbitrary dollar amount and is not standardised across CEOs. The overwhelming weight of academic literature shows that CEO pay should vary between CEOs. To that end, the academic literature sometimes finds that female top managers are underpaid, but sometimes it does not.
Let’s take a very basic example: if you beat a counterfactual you should be paid more. That counterfactual will vary between companies. For example, if the company is struggling and you are appointed to turn it around, then avoiding bankruptcy might be a good outcome. If the company is a market leader, then lukewarm performance might be a bad outcome. It varies based on the company’s circumstances and its goals.
What about the compensation structure? Not all compensation is equal.
CEOs can be paid in cash, stock, options, or in other ways. Some portion of the cash is fixed and is risk-free. It is how many employees are paid. But, another portion is variable and might be subject to personal or company performance. This is common in companies, including quasi-government ones, such as Australia Post. It is also common for many professionals: indeed, I have a variable component of my salary.
Many CEOs and startup employees also receive stock and options as part of their compensation. The stock and options are risky. Stock prices go up and down. The options might expire worthless. The purpose of these compensation components is to align CEO’s objectives with those of shareholders and to force the CEOs to bear the costs (or reap the rewards) of their actions.
Being paid $1 of stock is much more risky than being paid a $1 of cash. Thus, the compensation figures reported by the WGEA are, in my opinion, meaningless. Without knowing the compensation structure, it is impossible to know whether female CEOs are underpaid, even if they work at companies of the same type in the same circumstances.
CEO compensation is also tailored to achieve specific outcomes or to address specific behavioural attributes. A basic example is outside opportunities: If a CEO is in demand, possibly because of his/her skillset or the nature of the industry, they can pay for more. But, it can also vary based on the CEO’s personal circumstances. For example, compensation might vary based on whether the CEO is ‘overconfident’, the CEO’s life-stage, or the CEO’s latent risk aversion. Tailoring to cater to, or offset, CEO attributes are important and well documented in the academic literature.
The net result is that comparing sample averages, or medians, can lead to mistaken assumptions about pay. It is disquieting when a government department does not appear to grasp the idea of performance targets, bonuses, compensation calibration, and compensation structuring.
How then might we determine whether a CEO is underpaid? Well, a basic starting point is to collect data on CEO compensation and its structure. We also collect data on the firm, its performance, peer compensation, and all other factors that might influence pay. We identify the CEO’s gender and other personal attributes that might influence compensation. We then run a multiple regression to determine whether female CEOs do genuinely get paid less after controlling for these factors. This is a starting point, but not the end. We would also need to ensure that other econometric quirks, such as small samples and outliers do not bias the results. But, it would at least be better than comparing summary statistics between groups.
As far as my opinion goes, I find the figures in this report to be unhelpful when it comes to determining if there is a gender gap in CEO pay. They do nothing to address genuine problems with gender equality.