Relative TSR performance testing and governance
09/12/2019
mail.png

Many companies use a degree of discretion when assessing relative TSR performance. A few of these disclose that they exercised this discretion and have been surprised when their remuneration reports have not received proxy adviser support. The ones that do not disclose the application of discretion do not have an issue with investor support, primarily because of the absence of transparency in relative TSR measurement and outcomes.

Some company remuneration committees delegate the testing to management. We have found, in some cases, the discretion has not been applied by the board, but by management.

This begs the question, should discretion be used for relative TSR testing?

The answer is never by management, and in almost all cases, no by the board. It should not be necessary.

If the rules for calculating a company’s TSR relative to their peers are not clear, whoever is carrying out the calculations at the end of the performance period will be (or should be) asking a lot of questions. For example, should a spot price be used, or an average? If the latter, what kind of average? What happens to companies that were suspended, listed, merged or delisted during the period? Are dividends reinvested or just accumulated? If reinvested, at what price? To how many decimal places should rounding be?

If there are no ready answers to these questions, the master of the calculations will likely set their own rules. If they are at all captive to management, there may be a temptation to choose rules that will lead to a higher amount of equity vesting. Then when testing is completed the year after, if there are still no set parameters under which testing should occur, the tester might choose a completely new set of rules that result in the optimal vesting result for that year. If the company’s TSR performance is close to the 50th percentile of the peer group, these rule tweaks may mean the difference between vesting or not vesting.

It does not sound like good governance. And it is not. It was not even good governance nine years ago when we raised this (see HERE) .

Now that relative TSR is no longer regarded as the be-all-and-end-all, go-to metric that will guarantee strike free remuneration reports for years to come (we are glad that bubble has burst), it is even more important that when it is used, it is validly applied, correctly calculated, and consistent from year to year.

The TSR formula in a generic form is: Share price at end of the performance period less share price at start of the performance period plus dividends reinvested divided by share price at start of the performance period

No relative TSR test is complete until the following questions have been asked:

1 . Should the start and end price be averaged?

Should it be a spot price? Probably not. Averaging the share price used in the TSR calculation reduces effects of short-term volatility at the start and end of the performance period.

Averaging periods range from 5 to 180 days. Averaging over 30 calendar days should suffice for companies with low share price volatility, while 60 days would be better for higher volatility companies.

For more reliable results, the volume weighted average price (VWAP) over the selected period is used.  The average of the daily weighted average prices or closing prices over the relevant period can also be used.

2 . When is the averaging period relative to the start and end dates?

Does the averaging occur before the start date? After the date? Before and after the date? Is it inclusive of the date?

3 . When and how are dividends reinvested?

Some companies simply sum dividends. For those that reinvest the dividends, the reinvestment method can vary.  Some calculate TSR with dividends reinvested in shares at the share price at the dividend payment date (there still needs to be a decision as to what price – start, end etc). Others calculate dividends reinvested on the ex div date, while others re-invest the dividends in shares at the average share price for a specified days before the dividend payment date or ex div date.

Other factors to determine are whether net or gross dividends will be used. Are special dividends included or excluded from the dividends? Are dividends paid in shares included or treated as capital adjustment? If dividends are in a currency that differs from the stock listing currency, how should it be treated?

4 . What is the start date? Is it the grant date or another date such as the start of the financial year?

A danger with using the share price at grant date as the starting price for measuring TSR is that the actual grant date can vary for different participants in the LTI plan.  Grants are often made to most plan participants soon after board approval is received, but with grants to executive directors made after shareholder approval has been received. The terms and conditions of grants, including the performance periods, should be the same for all participants. Relative TSR can be calculated from any start date. It does not have to be calculated over financial years.  It may be considered better governance to calculate TSR from start and end dates when the market has been informed of annual results. The performance period should be clearly specified in the offer letter.

5 . To what accuracy will results be rounded?

Rounding to be applied to the TSR results needs to be specified – the nearest whole number? or to two decimal places?  Specifying it up front may save some pain in assessing results and vesting later. Rounding 0.5 upwards introduces positive bias and can be significant if outcome is 49.5 and threshold is 50th percentile.

6 . Which methodology of calculating percentiles will be used?

There are multiple different methods to calculate a company’s percentile rank against peers (open-source statistics software R has nine quantile algorithms). The testing methodology should be clear as to which will be used.

7 . Should the target company be included or excluded for percentile purposes?

Practice varies on whether or not the target company is included when the comparator companies are ranked according to their TSR performance.  Which approach is correct depends on how the vesting scale for the relative TSR test is described.

If the description is along the lines of “all rights will vest if Company X’s TSR over the performance period is at least equal to the TSR of the company which is at the 75th percentile of the comparator group of companies, ranked by their TSR performance”, then the target company would not be included in the ranking list.  The target company’s TSR is then compared with the ranked comparator group companies to determine the vesting outcome.

But if the description is more along the lines of “all rights will vest if Company X’s TSR performance over the performance period is at least at the 75th percentile of the comparator group of companies plus Company X, ranked by their TSR performance”, then the target company would be included in the ranking list.

8 . Is the peer group adequately defined?

The composition of the comparator group must be precisely defined.  If the peer group refers to the companies that comprise an ASX index group, the fact that the composition of an index may change every quarter means it is essential to specify that the comparator group consists of the companies that comprise the index at the start of the performance period.  The composition of the comparator group will then remain fixed for the duration of the performance period.

9 . How were peer companies treated if they were delisted? How were peer companies treated if they were suspended – or did this vary with the length or timing of the suspension (particularly at the start and end of the period)? What if trading in some companies has been illiquid?

Change is a constant among ASX-listed companies – it is not unusual to have a 25% churn rate in peer group companies during a 3-year performance period.

It is therefore important to specify what should happen if companies become delisted in a range of different circumstances. E.g. if they are acquired, merge, delist because they are insolvent, de-merge, delist during the performance period but are present at the start and end of the period, change name and ASX code, but are otherwise the same. These and other events have all happened before, and your TSR test supplier should be able to advise on how each should be treated and standardised into the method.

10 . How do you finalise TSR for peer companies listed on different foreign exchanges?

Does the methodology convert all comparator prices, including prices on ex-dividend dates for re-investment, to one currency before calculating TSR? If so, what rate is used – average, point in time, end of performance period?  Should TSR for each company be calculated in their local currency to avoid currency exchange effects?

The requirements will vary with each company, but once determined they should be explicitly stated.

Final checks:

11 . Are capital adjustments accounted for? Is the methodology the same for the peer company and tested companies?

12 . Is the methodology defined in both words and arithmetic symbols? Is it disclosed? If changing testing companies, what assurance is provided that the method is the same?

13 . Is the peer company and tested company TSR methodology the same? For valid results, performance for the target and comparator companies must be measured over the same period and using the same methodology.

© Guerdon Associates 2021
read more Back to all articles