Wage Drift

What wage drift is

Wage drift is the difference between how much an organisation’s wage base actually grew and how much the approved pay increase alone would have made it grow. It is the amount by which pay rose outside the increase decision, not the size of the increase itself.

In sequence, wage drift comes after the salary review. The review decides the amount and how it is shared out; wage drift measures how much was paid out on top of what the review decided. An organisation that shares out a fixed increase budget well but never measures wage drift is controlling one decision out of the many that move pay.

One point of scope comes before any figure. The approved increase is a rate the organisation decides for itself, and nothing below presents it as a prescribed rule or a required amount.

Where the gap comes from: wage drift worked through

Take an organisation with an annual wage base of SAR 10,000,000 across 200 employees, which approved an increase of 4 percent at its salary review. The expected base is therefore SAR 10,400,000.

By the end of the year the base stood at SAR 10,830,000. Actual growth was 8.3 percent, wage drift was 4.3 percentage points, and in riyals it came to SAR 430,000. That sum is made up of separate decisions, each of which was paid on its own and none of which was ever looked at in total:

• Promotions: 14 employees on an average wage of SAR 55,000, with an average rise of 9 percent on promotion. The effect is 14 times 55,000 times 0.09, or SAR 69,300.
• Replacement effect: 22 leavers on an average wage of SAR 44,000, replaced by 22 new joiners on an average of SAR 52,000. The difference is SAR 8,000 per head, or SAR 176,000 in total.
• Adjustments outside the review cycle: 9 cases averaging SAR 6,000 each, or SAR 54,000.
• Variable pay elements, those that follow the volume of work, above what was paid in the previous year: SAR 130,700, the remainder that completes the SAR 430,000.

No item on that list is wrong in itself. Each promotion was decided for its own reason, each hire was accepted at its price, and each adjustment was put forward with its own case. Nothing in the definition makes wage drift an error to be corrected. It is a measure of where pay decisions are actually being made, and a total that nobody sees, because every item in it is approved somewhere other than the salary review. Once that total has been seen, it may turn out to be acceptable.

The figures in this example are assumed, chosen to show how the gap is composed. They are not a benchmark, they do not set any rate or amount, and they are not taken from any source. In the sources we reviewed we found no general figure for an acceptable level of wage drift, and some drift is intended and planned for.

The largest item in wage drift is not promotions

The largest item in the example is the replacement effect: SAR 176,000 of the SAR 430,000, or 40.9 percent. It is also the only item in which no decision was taken under the name of pay at all. What decided it was 22 separate interviews, each concluding that a particular candidate was worth a particular amount.

It follows that, where the replacement effect is the largest item, wage drift is brought under control by managing where new joiners enter their pay range, which is what the compa ratio measures, rather than by tightening promotions. An organisation where people join at the middle of the range carries this item in any year in which those leaving were paid less; one where they join at the start of the lower third of the range avoids it, provided those leaving were paid more than those joining.

Wage drift enters the base and compounds

The SAR 430,000 is not one year’s spending, because it has become part of the base on which next year’s increase is calculated:

• Without any drift, a 4 percent increase in the second year would be 4 percent of SAR 10,400,000, or SAR 416,000.
• On the base as it actually stands, the same increase is 4 percent of SAR 10,830,000, or SAR 433,200.

The difference is SAR 17,200 in the second year alone, although the decision taken that year did not change. The effect repeats and grows, which is why wage drift is measured every year and presented alongside the request to approve the next increase, not after that increase has been approved.

Negative wage drift looks like discipline, and is not

The gap can come out negative, meaning the base grew by less than was approved. Take a year with no promotions and no adjustments, in which 22 employees on an average of SAR 52,000 left and were replaced by joiners on an average of SAR 44,000, and in which variable pay elements fell by SAR 90,000:

• Effect of the approved increase: plus SAR 400,000.
• Replacement effect: 22 times 8,000, or minus SAR 176,000.
• Variable pay elements: minus SAR 90,000.

The base reaches SAR 10,134,000, growth is 1.34 percent, and wage drift is minus 2.66 percentage points. In a report this reads as pay discipline, and it is not. The organisation is not paying its employees less; it is paying different employees.

Wage drift is therefore read by its direction and its cause together. A negative figure that comes from controlling where joiners enter the range is an intended result. A negative figure that comes from experienced people leaving is a result that can show up in pay before it shows up in the work.

Why comparing averages misreads wage drift

A common way of measuring wage drift is also a weak one: comparing average pay on two dates. The problem is that an average moves whenever the group itself changes. Take the same organisation, which gave nobody a single riyal of increase, but from which 20 employees on SAR 40,000 each left and were replaced by 20 on SAR 56,000 each:

• Change in the base: 20 times 16,000, an increase of SAR 320,000.
• New base: SAR 10,320,000, with average pay at SAR 51,600 instead of SAR 50,000.

Average pay rose by 3.2 percent without a single increase being granted. The sound measurement is made on a matched group: the same people on both dates. That separates the effect of a changing workforce from the effect of increases, and the composition effect is then shown as an item of its own, as in the first calculation.

The reverse also happens, and it hides drift that is real. When higher earners leave and are replaced by people paid less, the average falls, even though those who stayed may have received increases above the approved rate.

What wage drift is measured against: the base, not spending

Wage drift is measured against the wage base on a set date, not against what was spent during the year. The base is a future commitment that recurs; spending is something that has already happened.

For that reason, measuring against annual payroll spending alone does not work. An organisation that granted its increases in the eleventh month shows only two months of them in that year’s spending, and their full effect in its base. Anyone measuring against spending sees a small drift, and is then surprised the following year to find the base carrying twelve months of what had appeared as two.

In practice, the base is captured on a fixed date each year and the gap is calculated on it, with spending shown alongside as a second figure that answers a different question. The same reasoning is why wage drift is read together with the salary structure. The base moves within known pay ranges, and a movement that takes many employees outside their ranges is a question for the pay table, not for the increase rate.

Where wage drift sits among its neighbours

The ideas that sit closest to wage drift sort along one axis: what each one describes. An annual increment describes a decision, pay compression a shape and market movement a cause, while wage drift is a measured result.

An annual increment is a rise granted by a decision. What wage drift measures is the gap between the total that was paid and what the increment decision set.

Pay compression is a shape. It describes how narrow the distance between two levels has become, while wage drift describes how the whole base moves over time. Either can occur without the other: drift concentrated among new joiners produces compression, and drift spread evenly does not.

Market movement is a cause. Rising pay in the market pushes some of the decisions described above, but it is not wage drift. Wage drift is what happened inside the organisation: it is measured there and explained by the organisation’s own decisions.

What makes wage drift measurable

• A defined base. What goes into the wage base and what stays out is written down and held constant on both dates. Changing the definition between two measurements produces arithmetical drift that does not exist.
• A recorded reason for every change. The breakdown above cannot be drawn from a payroll statement showing that an employee’s pay changed. It comes from a record showing why it changed: a promotion, an adjustment or a replacement. That is a matter of record design, and it has to be in place a year before the measurement.
• Fixed and variable pay kept apart. Pay that follows the volume of work moves with that volume. Adding it to fixed pay in a single figure makes a busy season look like drift in pay policy.
• Measurement by unit. An overall drift of 4.3 points may be ten points in one unit and zero in another, and the remedy differs entirely between the two.

Defining the base is a measurement choice. The definition of wage drift settles no legal question about which pay elements count as wage, or about pay increases. Those questions have their own sources, and the base used to measure drift should not be borrowed as an answer to either.

The wage drift figure should also not be shown as a bare percentage. A drift of 4.3 points on a base of SAR 10 million is one thing, and on a base of SAR 100 million quite another, with the percentage unchanged. The riyal amount should always be written next to it, together with the number of people each item covered, because 14 promotions among 200 employees is a different situation from 14 among 2,000.

Before the next increase is approved

The most useful test before the next increase is approved is to present last year’s wage drift alongside the request for it, broken down by item rather than as a single figure, in the way a variance analysis breaks the gap between budget and actual into its components. The question then being answered is not “how much do we grant?” but “how much will actually be paid once this rate is added to the promotions, replacements and adjustments that are going to happen?”

The largest of those items is then examined, because that is where the decision sits. If it is the replacement effect, the remedy belongs in the rules for placing new hires within their pay range at appointment, not in the increase rate. Lowering the approved rate to control a figure whose source lies elsewhere controls the reported number and leaves its cause in place, and the cost falls on people who had no part in causing it.

This is an explanation of the concept, not legal advice.

Related terms

• Salary Structure
• Payroll Statement
• Payroll Accounting
• Variance Analysis
• Budgeting