Peer Group Analysis – Are There Alternatives?
While in use for many years as a tool for setting compensation, peer groups developed to establish marketplace compensation levels have recently come under increased scrutiny by shareholders, advisory firms, academics and management. Some have even gone so far as to advocate for their elimination altogether.
For all of their flaws, elimination of peer groups is not practical. Doing so would leave compensation professionals with no market-based starting point in setting compensation and Compensation Committees and executives wondering whether pay is fair and appropriate. Additionally, shareholders and those who assess how well the company does in setting pay expect to see a list of companies used to determine pay as part of the annual Compensation Discussion and Analysis and in some instances use this group to evaluate the validity of the compensation process more generally.
While there are no viable alternatives to peer groups, by thinking creatively about how to derive and use the groups, shortcomings can be minimized.
The Basic Process
The concept of peer group analysis is deceptively simple: select a group of companies that are similar to yours, generally based on industry focus and revenue size, and determine what they pay executives in similar positions. Once the group is selected and the data aggregated, the next step is to calculate summary statistics to determine pay levels at the average, median, 25th and 75th percentiles for the components of pay. If your company has adopted a philosophy to target compensation at the median of the marketplace, determining market competitive pay for that position is clear.
However, there are a number of challenges with peer group analysis that must be recognized and addressed to ensure that the process is appropriate. First, many companies have difficulty identifying sufficient companies of similar size and industry focus. One of the major complaints about the work done by shareholder advisory firms is that their requirements for peer groups with a pre-determined number of companies forces the inclusion of companies which share neither potential customers nor employees. While groups of 12 to 18 companies are ideal, and particularly helpful for “smoothing out” or mitigating outliers, in some instances companies may be better served by a much smaller, and more targeted, group of truly comparable companies.
Second, when selecting a peer group, consideration should be given not only to those companies that are competition for revenues, but also those which are competitors for talent. While industry-focused groups work well for certain positions, such as the CEO and R&D positions, executives in finance, human resources and law generally have skill-sets that transfer across industries, giving them a much larger potential “peer” group.
Finally, in most cases, the statistics developed from peer groups do not recognize the tenure, experience, actual job responsibilities, or successes and/or failures of particular executives, all of which have material impacts on pay levels.
Rather than scrapping the concept of peer groups, consideration should be given to alternative ways of using the data revealed by peer groups. Instead of focusing only on the summary statistics, a focus on the big picture – all of the data – can yield more nuanced, and useful, answers.
This would include an understanding of the tenure and position scope of each executive in the peer group. While the company may be appropriate as a peer, is the executive? Executives who are new to the position may not be appropriate comparators for a long-tenured, seasoned and proven executive. Similarly, the peer group for the CFO position, for example, may be bifurcated into CFOs with strategic and banking experience, and those who simply keep the books. Pay levels will often reflect these differences and not all incumbents may be relevant comparators.
Further consideration may also be given to the pay levels of members of the peer group. Whether or not it is appropriate to include compensation outliers, either high payers or low payers, is another issue to be considered. One alternative approach is to eliminate pay data for any peer company executive with target pay levels less than one-half or more than two times that of the executive under review.
In instances where there are a limited number of size and industry-appropriate peers, it may be helpful to separately analyze compensation data for executives of companies that are a good industry match, but either too large or too small to meet the traditional size criteria. While not directly comparable, this data provides additional context with which to understand pay across peer companies, and can help decision-makers triangulate to the “right” number. Regression analysis is also sometimes used in these situations to control for variations in company size, but in many instances the correlations between company size and total compensation are low, and the results are not statistically significant.
It may be easy to dismiss these suggestions as too much work to serve as viable alternatives. However, it is important to keep in mind that beyond the CEO and CFO positions, peer group analysis has limited value, and compensation professionals must rely instead on published survey data. Detailed analysis for a few positions is possible and is not generally time or cost prohibitive.