Do Sign-On Bonuses Improve Employee Retention?

Evidence grade: Mixed. Credible causal studies exist and disagree, or the effect holds only under conditions the studies can name. This grade describes the published evidence. It is not a prediction about your workforce and not a result RetainScore has produced.

A recruiting tool asked to do a retention job

A sign-on bonus is money for saying yes. The research on whether it changes who says yes is reasonably good, and it says the money works: it moves people toward jobs and places they would otherwise have passed on. The question this page answers is narrower and asked far more often in hiring meetings than in the literature: once the person has said yes, does the bonus keep them?

The honest answer is that it keeps them for as long as some of it has not been paid, and not much beyond that.

Where the money moved people, and where it did not

The cleanest test is Californian. Between 2000 and 2002 the state offered a competitively allocated $20,000 Governor's Teaching Fellowship to new teachers who committed to four years in a low-performing school. Steele, Murnane and Willett used an instrumental-variables design on the career histories of 27,106 licence applicants to separate the fellowship's effect from the kind of person who applied for it. On placement it worked: "acquiring a GTF increased their probability of doing so by 28 percentage points". On retention it did nothing the comparison group was not already doing. "75 percent of both GTF recipients and non-recipients who began working in low-performing schools remained in such schools for at least four years". The state paid $20,000 for a commitment three-quarters of comparable teachers made for free.

Massachusetts ran the more famous program and got the worse result. From 1999 the state paid $20,000 signing bonuses, $8,000 up front and the rest across three further years, to recruit new teachers, many from out of state. Fowler tracked each cohort through state records: "The bonus program has lost teachers at rates well above comparable national averages". Of the first 59 recipients, 46% had stopped teaching after three years, "raising this group's attrition rate to 46% (27 of 59), more than twice the national 3-year attrition rate of 20%". The second cohort fared better and still lost more than a quarter, "raising their 2-year attrition rate to 28% (29 of 105)". Three-quarters of the money was still to be paid when most of those teachers left.

Liu, Johnson and Peske interviewed thirteen of that first cohort. The bonus, it turned out, had barely figured in the decision: "the bonus money had very little influence in recipients' decisions to enter teaching". The fast-track certification route created to deliver it mattered far more. Their verdict on the program was that it "focused too narrowly on recruitment and not enough on retention", which is a fair summary of how most employers use the instrument.

The one condition under which a signing bonus retains

The largest study of signing bonuses anywhere is the Army's. RAND analysed enlistment bonuses for every Army accession from October 1999 whose term ended before October 2008, and started by naming the two forces that pull in opposite directions. Bonuses above a threshold are paid in installments and forfeited on early exit, so "recruits receiving more than the threshold amount have a financial incentive to fulfill their service obligation". But bonuses also draw in people who wanted the job less, and "this selection effect of bonuses would tend to generate a positive relationship between enlistment bonus receipt and first-term attrition". In the raw data, bonus recipients did leave more. Once the design handled the selection, the sign flipped: "Enlistment bonuses decrease attrition by 5 percent".

Five percent is a real effect, and it comes from the installment structure. Read together with the teacher studies, the pattern is not complicated. A signing bonus retains to the extent that it has not yet been paid. At that point it is a retention bonus with a different name, and the evidence on retention bonuses applies: modest, conditional, and inclined to stop working when the payments stop.

Hospital data shows what stopping looks like. Laudio's analysis of nurse hiring across more than 50 US hospitals found that "each RN who received a signing bonus was 5.3% more likely to be retained by the end of the second year", worth about $2,800 against bonuses of $10,000 and up. The shape of the departures is the useful part. Very few nurses leave in month eleven. Instead, "they leave at the end of the 12th month, after the 12-month bonus milestone is paid out". The bonus did not change whether they left. It scheduled when.

Why the grade is mixed

Mixed is the grade for credible studies that disagree, or for an effect that holds only under conditions the studies can name. Both apply. The Army analysis finds a modest reduction in early exits and the California analysis finds none, and the difference between them is the condition: whether money is still owed. Where the whole bonus is paid at signing, no study on this page finds a retention effect, and the Massachusetts record runs the other way.

The honest boundaries

• Teachers, soldiers and nurses. The evidence comes from three occupations with public data and unusual labour markets. The mechanism, unpaid money as a reason to stay, should travel; the sizes may not.
• The Massachusetts figures are descriptive. Fowler compares bonus recipients with national averages, not with a matched group of similar recruits. The program brought people into teaching who might not otherwise have entered, and some of the attrition is that. The comparison still tells you what the bonus did not do.
• The hospital analysis is a vendor's. It is not peer reviewed, and the authors say so themselves: "Note that this is not a perfect approach", because hospitals that pay bonuses may also run better residency programs. Treat the 5.3% as a ceiling.
• One source is read from its abstract. The Liu, Johnson and Peske interviews sit behind a publisher's bot check and the Harvard copy is blocked too; the claims drawn from it are the ones the abstract states.
• The Army effect is small. Five percent fewer early exits, in a setting where leaving forfeits money and breaking a contract has consequences a civilian employer cannot impose.

If you are paying them anyway

• Book it as a recruiting cost. That is what the evidence supports. If the vacancy would not fill without it, the bonus is doing its job; if it would, you are paying for a commitment you would have had for free, which is the California result.
• If you want retention, do not pay it at signing. Installments with forfeiture are the only structure on this page with a measured retention effect. Expect a cliff at each payout date and count survival past month thirteen, not month twelve.
• Do not read the raw numbers. Bonus recipients often leave more, because the bonus recruited people who wanted the job less. The Army data shows the effect only after that selection is handled, and yours will not be.
• Ask why the vacancy needed a bonus. A signing bonus is the price of a job people are not choosing at the going rate. The retention diagnostic is built to find the reasons people leave it, which are usually the reasons others will not take it.

Read the companion grade on whether retention bonuses work, which is where the evidence for any unpaid instalment of a signing bonus actually lives.