Skip to main content
main content, press tab to continue
Article | Executive Pay Memo North America

How do TSR award valuations align with reality?

By Stephen Zwicker , Mark Daniels and Heather Marshall | June 20, 2024

Considering adopting or re-designing a total shareholder return plan? Don’t let fair value premiums catch you by surprise because they’re real.
Executive Compensation|Retirement
N/A

Relative total shareholder return (TSR) is the predominant performance metric in long-term incentive (LTI) awards according to results of the most recent WTW Long-Term Incentive Policies and Practices Survey Report – United States. LTI design comes in many shapes and sizes, and in those approximately 50% of companies using a TSR metric in their performance plan, it may be the only metric, one of multiple stand-alone metrics, or act as a vesting modifier to other performance conditions.

For compensation accounting and proxy disclosure purposes, the value of TSR-based awards is determined using the Monte Carlo simulation method. These valuation results often come as a surprise when a company first adopts a TSR metric or significantly changes an existing TSR metric. There almost always is a fair value premium over the underlying stock price for each target unit granted, and it often can be significant.

This article explores why those premiums exist in theory and tests whether reality supports them. To do this, we simulated that each company in the S&P 500 index granted a three-year relative TSR-based award on January 1 of each year between 2012 and 2021. For those 5,000 hypothetical grants, we compared the ultimate realized value to a typical grant date fair value.

Our strawman design

In the early days of TSR adoption, we often saw companies adopt the vesting schedule outlined in Table 1.

Table 1. Typical vesting schedule of relative TSR performance conditions in the early days of TSR adoption
TSR percentile ranking Percentage of target units vested
Below 25th percentile 0%
25th percentile (threshold) 50%
50th percentile (target) 100%
75th percentile and above (maximum) 200%
Vesting interpolated between 25th and 50th percentiles and 50th and 75th percentiles.

While this vesting schedule still exists in many plans, we see a lot more variation in design, including:

  • Caps on vesting for negative TSR performance
  • Higher percentiles required for threshold, target and maximum vesting
  • Lower threshold or maximum vesting percentages
  • TSR used as a modifier

These changes usually are adopted for governance reasons (e.g., executives shouldn’t be rewarded above target for negative TSR outcomes) and/or to control the accounting expense. For purposes of this analysis, we stayed with the vesting schedule in Table 1 as our strawman design because it is easy to understand and design consistency is needed to compare theory with reality. In addition, we used the following design features:

  • Three-year performance period starting January 1 with TSR measured from point-to-point (i.e. any price averaging was ignored)
  • Assumed that any dividends paid by the sponsoring company were accrued as additional units earned based on the same vesting schedule
  • Fixed the peer groups at the beginning of each three-year award cycle
  • Excluded any acquired companies from the final rankings
  • Treated any companies filing for bankruptcy as having a negative 100% return

Fair value premiums, in theory

For the S&P 500 companies used for this analysis, the current range of grant date fair values derived using the Monte Carlo simulation method would roughly be between 135% and 150% of the stock price. Where a company falls within that range depends on their individual volatility profile relative to other companies. This range can move slightly over time as the overall state of the economy and markets change, but it is a good representative range of fair values for purposes of this analysis.

So, why does the Monte Carlo simulation produce this valuation premium? In a nutshell, Monte Carlo is a technique that generates many random potential outcomes from an assumed distribution of possible outcomes to derive a fair value. Monte Carlo is a powerful tool used for a wide range of valuation purposes well beyond LTI awards.

For a TSR-based plan, we might simulate 100,000 TSR outcomes for the sponsoring company as well as each company in the peer group. The payoff value of the award in each individual simulation is equal to the percentage of target units vested based on the company’s simulated TSR percentile ranking multiplied by the company’s simulated stock price at the end of the three-year period. The fair value of the award is the average of the 100,000 simulated payoff outcomes discounted back to the measurement date. With this large number of simulations, the method is canvasing the range and likelihood of possible outcomes with a sufficient level of statistical significance.

But we still haven’t answered the question: Why is there a valuation premium for a TSR-based award? In a word, the answer is leverage. The design and intent of these plans is to reward strong stock price performance. When the sponsoring company’s stock price does well, higher vesting outcomes are likely to be achieved. Certainly, the company can do well on an absolute TSR basis and still perform poorly on a relative basis, but that is a less likely outcome. There is a high correlation between the company’s absolute and relative performance.

That correlation means that more shares are being earned when the company’s stock price is higher than today. The upside of more shares at higher stock prices will more than offset the downside of fewer (or no) shares at lower stock prices. Consider this simple example:

  • Company stock price is $10. With 50% probability, it will be either $5 or $15 in one year.
  • The company introduces a plan that pays two shares if the stock price increases or zero shares if it decreases.
  • The average stock price in one year is $10, consistent with today’s price, which should be the case (assuming a 0% interest rate environment for this simple example).
  • The average number of shares that will be earned is one.
  • However, the fair value of the plan is not the $10 average price multiplied by the one share earned on average (i.e., $10). Rather, it is $15, which equals a 50% probability of earning $0 ($5 stock price and zero shares), and a 50% probability of earning $30 ($15 stock price and two shares).

Companies often ask if the valuation premium reflects an expected or predicted vesting outcome under the plan. As demonstrated in the example, that is not the case. The plan pays a single share on average (target outcome) but has a fair value that is 150% of the stock price. That is due to the leverage that exists between the vesting and stock price outcomes.

Fair value premiums, in reality

Companies (and especially plan participants) often are skeptical that the valuation premiums are realistic. It’s easy to view the awards much like the example we provided. The most likely (or average) outcome from the plan is to earn the target number of shares, and they should be worth the company’s stock price on the date of grant. Hopefully, the theory discussion above helped dispel why that would be the case, but let’s look at some real-world examples.

As described in the introduction, we reviewed 5,000 hypothetical three-year TSR awards granted by the S&P 500 constituents in the most recent 10-year period ending December 31, 2023. Table 2 summarizes the realized results from those awards.

Table 2. Realized results from a review of 5,000 hypothetical three-year relative TSR awards
Award cycle Number of companies completing three-year cycle1 Average realized payoff value relative to grant date stock price
2012-2014 477 238%
2013-2015 473 207%
2014-2016 465 165%
2015-2017 456 177%
2016-2018 457 165%
2017-2019 464 189%
2018-2020 472 181%
2019-2021 477 251%
2020-2022 478 165%
2021-2023 482 178%
Average 470 191%
1 Excludes acquired companies that did not complete the three-year award cycle.

 


Wow. There is not one single year when the realized values were lower than the fair value range (135% to 150%) cited earlier. There are some reasons why these values are higher, but this demonstrates that the Monte Carlo valuation technique is accomplishing its intended purpose to capture the value of TSR-based awards in the compensation accounting expense and proxy disclosure.

To further illustrate the leverage of these awards, Table 3 summarizes the realized values based on the ultimate percentile rankings.

Table 3. Realized values by quartile ranking
Percentile ranking Percentage of companies Average vesting percentage Average TSR Average payoff value1
Below 25th percentile 25% 0% -22% 0%
25th – 50th percentile 25% 75% +24% 94%
50th – 75th percentile 25% 150% +54% 233%
75th percentile and above 25% 200% +119% 438%
Total/average 100% 106% +44% 191%
1As a percentage of the grant date stock price and target number of units.

 


Table 3 shows the correlation between vesting outcomes and stock price outcomes we would expect. In the 50% of plans that end up with rankings below the 50th percentile, the payoffs, on average, are below the target award value. But for plans above the 50th percentile, payoffs are more than two-times (between 50th and 75th percentiles), or four-times (75th percentile and above) the target values. These are the potential payoffs that drive the value of these awards, with both shareholders and executives winning.

To further illustrate this point, Figure 1 shows outcomes for just the 2021-2023 awards derived from WTW’s TSR Performance Monitor software application. The graph highlights the range of returns from each quartile of performance and their effect on value.

Screenshot of WTW's TSR Monitor screen showing a sample peer summary
Figure 1. Outcomes for 2021-2023 awards highlighting the range of returns

The payoff values from this 10-year sample of awards are significantly higher than the Monte Carlo fair values. Does this mean the model is actually understating fair values? No, for several reasons:

  • LTI awards, including TSR-based plans, are a risky form of compensation. The Monte Carlo simulation method is run in a risk-neutral framework that discounts the fair value for this risk. The payoff values for our hypothetical S&P 500 plans are real-world outcomes that are not risk-adjusted.
  • Grant date fair values are the beginning of period values. The payoff values we calculated are end-of-period values. We ignored the effects of discounting during the three-year performance period.
  • We looked at 5,000 potential outcomes for our hypothetical plans. As discussed, the Monte Carlo method might use 100,000 simulations so, even though we looked at many award outcomes, they still fall short of everything that might happen with these awards.
  • Table 3 shows an ending stock price value 44% greater than the grant date price, on average. That corresponds to a 13% annual return for the companies in the S&P 500 index from January 1, 2012, to December 31, 2023. By and large, this was a good period for the markets that helped increased the payoffs from these plans. Other time periods may not yield as favorable of results, tying with the aforementioned number of simulations.

Analysis that supports theory

As relative TSR prevalence in LTI award designs continues, valuation questions and skepticism inevitably will continue among companies and executives. This analysis supports the theory that valuation premiums for TSR-based awards are appropriate. The Monte Carlo method is working as intended to capture the true value of these award designs. If your organization is concerned about the level of valuation premium, consider exploring plan design alternatives to evaluate the tradeoff between a plan’s potential value and the associated accounting value.

Authors

Senior Director, Retirement
email Email

Director, Retirement
email Email

Senior Director, Executive Compensation and Board Advisory
email Email

Contact us