October 31, 2013

CAPintel

Avoiding Pitfalls in Executive Compensation Benchmarking

Contact

Bertha Masuda
Partner Emeritus [email protected] 310-541-6233

Share

Competitive benchmarking is a critical step in the development of executive compensation programs. Companies typically define their pay philosophies, benchmark competitive pay and practices, design incentive programs and then set pay levels accordingly. Until recently, competitive benchmarking was one of the least contentious steps in the whole process.

Those days are over. Recent corporate governance reports criticized sloppy competitive pay benchmarking practices. One issue with competitive benchmarking is the fact that nearly all companies position executive compensation at median or above. In this case, the competitive data artificially escalates year after year. Second, many governance experts believe that compensation committees have relied too heavily on benchmarking studies and have failed to apply good judgment to the numbers.

Despite the recent backlash, executive compensation benchmarking remains a useful tool when it is carefully executed and balanced with other reference points. Some of the most common—and avoidable—benchmarking pitfalls we see:

  • Selection of the screening criteria that will produce an appropriate benchmarking peer group

A variety of criteria can be considered, including industry, size, business economics, business focus, business strategy, common pool for management talent, historical performance and geography. Identification of the critical screening criteria will lead to relevant peer group selection.

  • Review and selection of compensation surveys

Just as selecting peers is critical, so is the choosing the appropriate published surveys. Several factors should be considered, including the participation of direct peer or comparison firms, the availability of necessary data (for example, salary, bonus and long-term incentives) and the presence of relevant data cuts. It’s also important to review survey methodology to ensure consistency of approach when multiple surveys are used (three sources per position is ideal).

  • Blending proxy and survey data

Several pitfalls can arise when combining proxy and survey data: overweighting proxy data for some positions, using survey or proxy data that are outliers relative to other data sources, and overweighting sources with a small sample size. Look for large enough sample sizes and consistent data and assess whether the data you review is reasonable as part of your assessment.

  • Actual versus target bonuses

When analyzing competitive bonus levels, it is important to consider whether the competitive data is reporting target or actual bonus levels. If your company is lagging the industry, target bonus levels may be preferable for the analysis.  Your company may provide a competitive target opportunity, but actual payouts are lagging, as they should, because of performance.

For companies challenged by these and other pitfalls or those who do not want to rely entirely on benchmarking to set their executive pay levels, other analyses can be conducted:

  • Reviewing internal equity to ensure relative pay levels are reasonable
  • Taking performance into account when determining executive pay levels
  • Using wealth-accumulation analysis to assess the richness of executive compensation practices

While these analyses are quantitative and number-driven, they too depend heavily on subjectivity and judgment. Selecting the benchmarking approach that’s right for your organization requires an understanding of various resources and methodologies available and the application of them in the context of your business. Despite recent criticisms, competitive pay benchmarking continues to be a useful tool, although it is admittedly an imperfect one. 

The full article entitled “The Devil is in the Details: Analytical Pitfalls in Executive Compensation Benchmarking”, written by Bonnie Schindler, appears in the recently published book Survey Best Practices: A Collection of Articles from WorldatWork.