Retention bonus vs pay raise: what each actually costs a small business employer

A retention bonus and a pay raise of the same face value cost almost exactly the same in year one. The difference shows up in years two and three, and in one place most employers never price at all: for a non exempt employee, a retention bonus is a nondiscretionary bonus under federal wage law, so it has to be folded back into the regular rate and it raises the overtime you already paid. A $5,000 bonus to someone working 50 hour weeks costs about $5,500 before payroll taxes. The same $5,000 as a raise creates no retroactive obligation at all.

The usual framing of this decision is cultural: a bonus is transactional, a raise is a commitment. That is true enough, and it is also unfalsifiable, which makes it useless when you are the one signing off the money. Below is the version with the arithmetic filled in, the federal rules that change the answer for hourly staff, and an honest account of the four situations where the bonus really is the right instrument.

Year one is a tie, so ignore it

Take a salaried, exempt employee at $75,000 who has told you they are interviewing. You can offer a $5,000 retention bonus for staying twelve months, or a $5,000 raise. In the first twelve months these cost the same: $5,000 of wages plus employer Social Security and Medicare at 7.65 percent, which is $382.50. Both land at $5,382.50. Anyone who tells you the bonus is cheaper is comparing a one time payment to a permanent one and calling that a saving.

The comparison only becomes real when you extend it. Assume a 3.5 percent merit increase each year, applied to whatever the base happens to be. The raise lifts that base permanently, so every future percentage is calculated on a bigger number. The bonus resets to zero and you face the same conversation again.

Over three years the repeated bonus and the compounding raise land within about $570 of each other on a $75,000 salary. That is the finding worth keeping, because it kills the argument that a bonus is the fiscally responsible option. If you expect to need the retention twice, you are spending raise money and getting bonus outcomes. What you buy for the extra $570 is a base that moves, which is the only version of this that changes the employee position against the market. You can size the same comparison on your own numbers with the salary increase calculator.

The part almost nobody prices: overtime

Now run the same $5,000 past a non exempt employee, and the arithmetic stops being symmetric.

Section 7(e)(3)(a) of the Fair Labor Standards Act lets an employer keep a genuinely discretionary bonus out of the regular rate. A retention bonus never qualifies, and the regulation is unusually blunt about why. 29 CFR 778.211(b) states that the employer must retain discretion over both the fact and the amount of payment, and then: "If the employer promises in advance to pay a bonus, he has abandoned his discretion with regard to it." A retention bonus is a promise made in advance. That is the entire mechanism. Stay until this date and you receive this amount. There is no version of it that preserves discretion.

So it goes into the regular rate, and 29 CFR 778.209(a) sets out what that means when the bonus covers more than one workweek. The bonus "must be apportioned back over the workweeks of the period during which it may be said to have been earned," and the employee "must then receive an additional amount of compensation for each workweek that he worked overtime during the period equal to one-half of the hourly rate of pay allocable to the bonus for that week multiplied by the number of statutory overtime hours worked during the week." Where the bonus cannot be tied to particular weeks, 778.209(b) permits assuming an equal amount was earned each week.

Work that through for someone at $25.00 an hour on a steady 45 hour week, receiving $5,000 for a twelve month stay. Allocate $5,000 across 52 weeks and each week carries $96.15. Spread over the 45 hours worked, that is $2.1368 an hour. Half of that is $1.0684, and with five overtime hours in the week you owe an extra $5.34 for that week. Across 52 weeks, $277.78.

The general form is simpler than the worked example suggests, and it is worth writing on the wall next to whoever approves these:

A $5,000 retention bonus to that 50 hour a week employee is therefore $5,000 of bonus, $500 of recalculated overtime, and 7.65 percent employer payroll tax on both, for $5,920.75. You are paying $1.18 for every dollar the employee perceives. A raise does none of this. The higher rate flows into the overtime premium prospectively, week by week, as payroll runs. There is no back calculation, no amended pay periods, no reopening a year of payroll records. If your overtime arithmetic is not already solid, the mechanics live on FLSA overtime rules and you can check any single week against the overtime calculator.

One payment, three different rulebooks

The reason this catches people is that the same $5,000 is classified differently depending on which agency is looking at it, and no single guide covers all three.

To the IRS it is a supplemental wage. Publication 15 confirms the withholding rate on supplemental wages "remains 22%," rising to 37 percent once supplemental wages to that employee exceed $1 million in the calendar year. That flat 22 percent is optional and applies when the bonus is paid separately from regular wages; aggregate it with a normal paycheck and you withhold at the employee's W-4 rates instead. To the Department of Labor the same payment is a nondiscretionary bonus that reopens the regular rate. And to your compensation structure it is not pay at all: it sits outside base salary, so it does not move the employee within a band and it is invisible the next time the role is benchmarked.

That last one has a consequence worth stating plainly. The flat 22 percent frequently over withholds relative to what the employee actually owes, so a $5,000 bonus can arrive as roughly $3,500 after federal withholding and FICA. The employee compares that number, not the gross, against the raise they did not get. You paid $5,000 and change, and they felt $3,500, once. A raise of the same size is felt in every paycheck for as long as they stay.

When the bonus is genuinely the right call

None of this makes retention bonuses a mistake. There are four situations where a bonus is clearly the better instrument, and they share a shape: the need has an end date, or the pay is already correct.

The need is genuinely time boxed. An acquisition closing, a system migration, a plant wind down, a customer transition. You need a specific person through a specific date and then the requirement disappears. A raise granted for a nine month need is a permanent cost for a temporary problem.

The person is already at or above market. If their compa ratio is above about 1.05, a raise pushes them toward the top of the band for reasons that have nothing to do with the market or their level, and it creates a gap against peers doing the same work. That is how pay compression starts. Check where they actually sit with the compa ratio calculator before you assume a raise is deserved.

You are holding a date, not changing a career. Bonuses answer "will you stay until March." They do not answer "is this a job worth staying in," and using one to answer the second question buys a delay at full price.

The team has a band and this person is inside it. When you can show the role is paid correctly, a bonus is a clean recognition of a temporary burden rather than a quiet admission that the salary was wrong.

The inverse is the expensive case. Give a retention bonus to someone who is genuinely underpaid and you have bought a date, not a solution. On that date the market has moved again, they have another offer, and you are negotiating from a worse position with a precedent already set that leaving conversations produce cash.

Decide with a number, not an instinct

Both options are expensive next to doing nothing, and both are cheap next to replacement. That is the comparison that should frame the decision, and it is one you can actually compute. Start with what you are losing: run your employee turnover rate and split it into voluntary and involuntary, because a total number mixing quits with terminations cannot tell you whether you have a pay problem or a hiring problem. Voluntary turnover running well above the BLS quits rate for your industry is usually pay or management, and no bonus fixes either.

Then price the alternative honestly. Replacement is not just a recruiter fee. It is the vacancy weeks, the manager time spent on running first round screening interviews, the onboarding, and the months before a new hire is producing at the level of the person who left. Against a total like that, both a $5,000 bonus and a $5,000 raise usually look cheap, which is precisely why the choice between them should turn on evidence rather than on which one feels more affordable this quarter.

The evidence you need is small. Where does this person sit against the market rate for their role and level, and is the need temporary or permanent. Below market and permanent means a raise, and it means the band was wrong before this conversation started. At market and temporary means a bonus. Everything else is a negotiation you are having without data, and the employee, who has an offer in hand, has more of it than you do. Building the salary bands that answer the first half is the part you can do before anyone resigns.

Sources: 29 CFR 778.208, 778.209(a) and (b), 778.211(b); FLSA section 7(e)(3)(a); IRS Publication 15 (Circular E), supplemental wages; Social Security Administration contribution and benefit base, $184,500 for 2026; BLS Job Openings and Labor Turnover Survey, June 2026. Employer payroll tax figures assume the employee is below the Social Security wage base. This is general information about published law, not legal or tax advice. Retention agreements, clawback terms and permitted wage deductions turn on your own facts and your own state's law, so confirm the analysis with counsel before acting on it.