Input Method – Humans vs Machines
The most striking difference between available KMP remuneration databases arises from the input method. A number of providers are using “OCR” and machine learning, which is equivalent to a computer taking a photo of the PDF data tables and inputting the numbers directly into a database. This means that while the resulting database will generally reflect the actual disclosures of companies (depending on the accuracy of the system), many of the data points will not be valid for benchmarking purposes, such as in the case of part-year incumbents, or where exceptional payments are being made. These systems generally struggle with unwinding accounting treatment and exceptional payments such as one-off takeover retention arrangements.
Manual Input Data Quality
The quality of data can vary significantly between human-powered databases too. Some like GRG have the data extracted and analysed by Australian degree qualified analysts who are supported by experienced consultants who exercise judgement on the many interpretation issues that inevitably arise when data is being extracted from Remuneration Reports. Many other data providers outsource data entry to teams in large foreign data houses that generally do not specialise in remuneration, and the inputting is not overseen by Australian consultants who use the data to provide information and advice to clients.
There are around 900 ASX listed companies with market capitalisation of more than $25 million. While a given client would never need all of these, the analytics that are able to be undertaken when a large set of data is available helps ensure that data interpretation, extrapolation and analysis delivers the best possible results for clients. Databases which are comprised of say ASX300 companies only, are often challenged to produce benchmarking samples that are relevant and closely aligned to the client. Such smaller samples also cannot support intelligence to derive the correct answer for circumstances where there are gaps in the market, for example in the case of rare roles.
Consistency in the valuation approaches used for different elements of remuneration is critically important. The first point to note is that the values being input to the database need to be reflective of the amount paid to KMP for a full year of work.
There are generally three main elements of remuneration being: Fixed Pay, short term variable remuneration (STVR) or short term incentive (STI) and long term variable remuneration (LTVR) or long term incentive (LTI). Fixed Pay is comprised of salary, superannuation contributions, other benefits and fringe benefits tax (FBT). Variation in conclusions regarding Fixed Pay usually relate to when part-year incumbents are included, and accruals for leave to be taken in future years are not part of annual remuneration should not be added to the other elements of remuneration. Consulting payments unrelated to KMP role can also often arise and need to be excluded. It is not uncommon for Australian companies to employ some KMP in other countries and for them to pay such KMP allowances or additional Fixed Pay to compensate them for differences in the costs of living, living standards or remoteness. Such roles should be excluded from benchmark data in relation to Australian based KMP if the exceptional remuneration cannot be excluded.
The calculation of STVR as disclosed in Remuneration Reports relates to the actual amount paid during or for the relevant year. Arguably the target levels of STVR award opportunity could be more informative, however STVR targets are poorly disclosed and there are considerable differences is the definition of target used by different companies. Analysis within companies has revealed that actual STVR awards tend to be mainly in the range of 80% to 120% of perceived target levels. Accordingly, statistics based on statutory disclosure tend to approximate target award opportunities. However, the deferral of STVR can substantially complicate disclosures and may require adjustment to provide an appropriate benchmark.
LTVR valuations require the reporting of values calculated under AASB2, the accounting standard dealing with share-based payments. Because this standard requires different valuation and reporting approaches to be applied to different instruments with different terms, it is GRG’s experience that the reported values, which often include negative values tend to understate target LTVR levels except when KMP have received 3 or 4 annual LTVR grants and they are based on market conditions only.
A concerning practice with LTVR value reporting is when total grants are valued at face value. Typical LTVR practice involves grants at a stretch level that is normally double the target level. This gives rise to overstatement due to the value being based on double the target (note that STVR values are almost always based on target values) and face value relates to the market value of a share whereas rights and options are the dominant form of instrument used for LTVR purposes and such instruments have values of less than the share price. This approach produces LTVR values that are inconsistent with STVR values and can result in grossly inflated LTVR values and target total remuneration packages. In some cases these may be presented as “Target” remuneration, when in fact they are stretch.
A tool that is often used to look at internal relativities between jobs is a job sizing methodology. There are several such tools, but each has the same objective of assessing the relative size of jobs both relative to external market references, and within a given company’s organisation structure. Factors often considered in job sizing include: know-how, problem solving, and accountability. The outcome of job evaluation is typically a points score where the higher the score the bigger the job.
In surveys of market remuneration practices the outcomes are often presented as pay per point in the job evaluation score. This approach has some validity when dealing with non-KMP roles. However, it can be misleading when looking at remuneration for SLT roles. These roles in ASX listed companies are usually paid differently to other roles as there is a stronger emphasis on STVR and particularly LTVR than is applied to similar sized role which are not KMP roles, and a higher level of scrutiny for Fixed Pay relativity between companies that are publicly disclosing.
When considering externally obtained market remuneration data for SLT members it is important to ensure that all of the roles in the sample are KMP roles. Inclusion of other roles can taint the data and give misleading results. Often a KMP role can end up being compared to a role that is one or two layers down in the organisation structure of a much larger company, simply because the evaluation score is equivalent. The remuneration structure for these roles should be materially different.
Comparator Group Selection
GRG has undertaken research in respect of company size and levels of remuneration. That research revealed the strongest correlations were with market capitalisation and net profit. In each case remuneration grew as market capitalisation and profit grew. Of these two, market capitalisation had the stronger correlation. Other size metrics had lower degrees of correlation with levels of remuneration e.g. revenue and assets. Number of employees had a poor correlation with levels of remuneration.
Industry sector alone had no correlation with levels of remuneration for KMP.
Size Range for Comparator Group
Larger companies tend to pay more than smaller companies. They also provide greater STVR and LTVR as percentages of Fixed Pay than smaller companies. As a consequence, when selecting companies to comprise a comparator group for benchmarking purposes it is important that they fall within a reasonable range of the size of the company being benchmarked. As a guide GRG’s view is that companies should be selected from a size range of approximately 50% to 200% of the size of the company being benchmarked. Of course, the closer the size of the comparator group companies to the size of the company being benchmarked the better.
Balanced Comparator Group
More important than the number of companies in a comparator group is that there be an equal number smaller and larger than the company being benchmarked. If the comparator group is composed of more companies that are smaller than the company being benchmarked, then the mid-point data is likely to be lower than an appropriate reference point for the company being benchmarked. Conversely, if the comparator group is composed of more companies that are larger than the company being benchmarked, then the mid-point data is likely to be higher than an appropriate reference point for the company being benchmarked. These outcomes arise because of the earlier observation that larger companies pay more than smaller companies.
Number of Companies in Comparator Group
In selecting the number of companies to comprise a comparator group it is important that there are sufficient data points to produce meaningful statistics. GRG has found that around 20 companies will usually comprise a useful comparator group.
Industry and Business Challenges
When selecting the companies from within the size range it is also important to focus on companies that operate in the same industry sector as the company being benchmarked and facing similar business challenges. GRG’s experience is that limiting comparator group companies to only those that operate in the same industry sector as the company being benchmarked will produce generally very small samples and therefore need to be supplemented with additional companies or additional comparator groups. Exceptions include when benchmarking small resources companies.
At the broad industry categorisation level there are differences between the pay practices of companies in the Resources, Financial and Industrial & Services sectors. However, within those sectors there tends to be little difference in pay practices, only the quantum and mix.
Overseas Companies and Companies with Overseas Operations
Some companies that are listed on the ASX have domiciles in overseas countries. Often these companies adopt remuneration practices that relate to the country of domicile rather than the practices of Australia. Obviously, such companies are not appropriate for use in comparator groups for companies that are benchmarking the remuneration of Australia based KMP.
Complete Sample for Anchor Role
When benchmarking remuneration for members of the SLT it is important to ensure that each company in the comparator group has data for the Chief Executive Officer (CEO) role. The remuneration data for this role needs to robust and can be used as an anchor when market competitive remuneration for other roles are set as percentages of the Fixed Pay for the CEO role.
Data Alone is not Always Sufficient
Inadequate Role Data
Even when a robust comparator group is selected it is not uncommon for the data in respect of some of the roles being benchmarked not to be adequate. This can be because:
- Some roles are unique to the company and equivalent roles are not available in other companies,
- For some roles the data sets are too small to be seen as a reliable reference point, and
- For some roles the data sets are clearly skewed with the majority of the data from smaller or larger companies than the company being benchmarked.
When this occurs, the consultants providing remuneration recommendations will need to utilise other tools in their kit bags. A well informed consultant should be able to produce a valid and reliable recommendation for the remuneration of such roles, by drawing on proximal information available and applying robust methodologies. This is where a large database and experience can assist greatly, since relying on data only will produce useless invalid outcomes.
Fixed Pay Models
Consulting firms will have at their disposal remuneration models which may be used to underpin recommendations on Fixed Pay levels and target STVR and LTVR award opportunities. These can be banding or pay level models that are mainly used to ensure that appropriate internal remuneration relativities are achieved. They can enable recommendations to be made for unique roles and for roles where market remuneration data is too thin to be relied upon alone.
Remuneration Data Analytics
Provided that remuneration databases have adequate quantities of remuneration metrics it is possible to undertake in depth analysis to identify remuneration relationships that would not otherwise be evident. These relationships can be powerful tools in helping remuneration consultant to develop remuneration recommendations for clients.
An example of the use of analytics is where a role occurs sporadically across a wide range of company sizes. In these circumstances analytics can reveal the market competitive rate for a company when its comparator group reveals no data.
Understanding the quality and methodology behind market data you may have access to will be a key driver of whether the benchmarking information you receive will produce a valid and reliable result. In some cases, this simply will not be possible given the limited size of the ASX market, and in those cases, advice may be required to obtain a logical solution. In many cases, what appears to be a simple task is subject to many nuances that need to be properly understood before a conclusion is drawn and action taken. After all, there is a lot riding on KMP remuneration for a wide range of stakeholders.