The Business Case for Workplace Mental Health: What the Numbers Do and Don't Show
In this article
Every budget season, someone in HR gets asked to justify spending on mental health, and the first instinct is to reach for the biggest number available. I spent a good part of my consulting career watching those presentations land, and watching some of them fall apart the moment a finance director asked a second question. The problem was rarely that the numbers were wrong. It was that they were being asked to prove something they were never designed to prove.
This piece walks through the figures that show up most often in the business case, what each one actually measures, and how to build an argument for your own organization that holds up under scrutiny. The short version: the evidence that poor mental health is expensive is strong. The evidence that any particular employer program will pay for itself is much thinner, and you are better off saying so out loud before someone else does.
The two numbers that appear in every deck
The cost of doing nothing
The figure quoted most often comes from the World Health Organization and the International Labour Organization, which estimate that about 12 billion working days are lost every year to depression and anxiety, at a cost of nearly US$1 trillion in lost productivity. The estimate was highlighted when the WHO released its guidelines on mental health at work in 2022, and the guidelines themselves are worth reading if you haven't.
What it does show: depression and anxiety are common, they hit people during their working years, and their effect on work is large enough to register at a global scale. It is a fair answer to anyone who still thinks mental health is a niche concern or a soft one.
What it doesn't show: what any of this costs your company. It is a global, modeled estimate built from prevalence data and assumptions about how much work is affected. Dividing it by the world's workforce and multiplying by your headcount produces a number that looks precise and means very little. Your industry, the age of your workforce, the benefits you already offer and local access to care all change the picture, sometimes dramatically.
The return on treatment
The second favorite is the return-on-investment figure. A WHO-led study published in The Lancet Psychiatry in 2016 estimated that every US$1 invested in scaling up treatment for depression and anxiety returns roughly US$4 in better health and ability to work.
What it does show: effective treatment for common mental health conditions is, at a population level, a sound investment. People who get evidence-based care tend to recover their ability to function, and the value of that recovery exceeds the cost of the care.
What it doesn't show: that an employer wellness program, a meditation app subscription or a lunch-and-learn returns four to one. The study modeled the scale-up of treatment, meaning psychological therapy and medication delivered through health systems across many countries, over a span of years. Its return also counts health gains, not only productivity. An employer contributes to that kind of outcome mainly by making sure employees can actually reach good treatment. That is a benefits-design and access question far more than a programming one, which is the part of the argument that tends to get lost.
Where vendor ROI claims usually go wrong
Vendors will often hand you their own return figures: three to one, five to one, sometimes higher. Some are carefully done. Many share a handful of weaknesses worth recognizing before you repeat the numbers to your CFO.
- Self-selection. People who sign up for a program differ from people who don't. Comparing users with non-users and calling the difference "impact" mixes the effect of the program with the effect of being the kind of person who enrolls.
- Regression to the mean. People tend to seek help when they feel worst, and many would improve somewhat over the following months regardless. A before-and-after comparison with no control group credits the program for that natural recovery.
- Soft inputs, hard outputs. A claim that converts a few points on a self-reported wellbeing scale into dollars of avoided turnover rests on a chain of assumptions. Ask to see each link.
- Short windows. Twelve months of data can show a dip in claims that turns out to be timing, not savings. Mental health care sometimes raises costs in the short run as people finally get treated.
- Who paid for the analysis. Not disqualifying on its own, but it matters, especially if the method isn't published.
There is also a broader caution from research on general workplace wellness programs. Two large randomized trials in the US, published in 2019, one at a warehouse retail chain and one at a large public university, found little measurable effect on medical spending or on most work outcomes over the periods studied. Neither was a test of mental health care specifically, and both had limited follow-up. They are still a useful reminder that when programs are tested rigorously, the effects tend to be smaller than the brochures suggest.
How strong is each claim?
Here is how I would roughly grade the arguments that come up most often in budget discussions.
| Claim | Strength of evidence | What to say instead of overselling |
|---|---|---|
| Depression and anxiety are common in working-age adults and affect job performance | Strong | State it plainly and cite the WHO/ILO estimate as a global figure |
| Effective treatment restores ability to work | Strong at the population level | Focus on whether your benefits let people reach effective treatment |
| A specific wellness program will return a fixed multiple | Weak to mixed | Present vendor figures as claims, with their method attached |
| Better working conditions (manageable workload, control, fair treatment) support mental health | Reasonably strong, and central to WHO guidance | Frame job design as risk management rather than perks |
| Mental health spending will lower health plan costs next year | Weak; costs may rise first | Set expectations for a multi-year view |
| Supportive managers improve retention | Plausible, with some support; hard to isolate | Track it internally instead of promising a number |
The pattern is consistent. The closer a claim gets to "this condition is costly and treatment works," the firmer the ground. The closer it gets to "this product will save you money," the more you should hedge.
A worked example: building the case at a mid-sized firm
Picture a hypothetical 220-person engineering and surveying firm in Colorado. The HR director, call her Joanne, wants funding for three things: a better EAP, manager training, and a review of the health plan's behavioral health network. Her CFO is skeptical of soft benefits and has sat through wellness pitches before.
Joanne's first draft opens with the US$1 trillion figure and the four-to-one return. She cuts both from the opening and moves them to a single context slide, clearly labeled as global estimates. Then she builds the rest from what she can see inside the firm.
What she knows locally. Voluntary turnover last year was 14 people. Exit interviews mention workload and deadline pressure again and again. Short-term disability claims included several for mental health conditions, though the numbers are too small to report by category without risking identifying anyone. Three employees told HR they had given up trying to find an in-network therapist who was accepting new patients.
What she estimates, and labels as estimates. Using the firm's own recruiting and onboarding costs, she puts a rough replacement cost on each departure. She doesn't claim the package will prevent a set number of exits. Instead she shows a break-even point: if the changes help the firm keep two people who would otherwise have left, they roughly cover the first-year cost. The CFO can judge for himself whether that's plausible.
What she proposes to measure. EAP clinical cases per quarter, the share of employees who can say how to reach the EAP, time to first appointment for behavioral health care, and a short annual survey item on whether people feel able to raise workload concerns. She also has her leadership team complete the site's 16-statement self-check and records the four area scores as a baseline to revisit.
What she doesn't promise. Lower health plan costs next year. She says the opposite: better access may increase behavioral health claims at first, and that should be read as people getting care they needed rather than as failure.
The CFO approves the EAP upgrade and the network review, and asks to revisit manager training after six months of data. That is a reasonable outcome. A case built on honest local figures tends to get partial yeses that grow over time. A case built on borrowed global returns tends to get either a polite no or a funded program that is quietly cut the following year.
Arguments that don't depend on ROI at all
Some of the strongest reasons to invest in mental health at work don't need a multiplier.
Risk and compliance. Employers already carry obligations around reasonable accommodation, disability discrimination and safe working conditions, and mental health conditions fall within them. Psychosocial risks such as excessive workload, unclear roles and harassment are increasingly treated in official guidance as workplace hazards like any other. The US Surgeon General's framework for workplace mental health and well-being sets out the components, and it reads more like an operating model than a wellness brochure.
Money you already spend. Most organizations already pay for an EAP and for behavioral health coverage inside the medical plan. If employees can't find or use those services, part of that spend is wasted. Improving access to what you already fund is an efficiency argument, not a request for new money, and finance teams tend to hear it that way.
Talent. Candidates and employees increasingly ask what mental health support looks like, and managers who handle it badly cost you people. That is hard to quantify cleanly, but most CFOs will accept it as a qualitative factor if you don't dress it up with invented precision.
Leadership credibility. If senior leaders talk about wellbeing while workloads stay unmanageable, employees notice quickly. The leadership commitment section of the four key areas describes what consistent backing looks like in practice, and it is often where a business case quietly succeeds or fails.
Presenting it to finance
When you take the case upstairs, a few habits make a real difference.
- Lead with your own data, even if it's thin. Turnover, disability claims, survey results, exit-interview themes, time to care. Local evidence beats global evidence in a budget meeting.
- Label every estimate as an estimate. Show your assumptions. Finance teams respect a transparent rough number more than a polished one with no workings behind it.
- Use the global figures as context, not proof. One slide, clearly attributed. The WHO/ILO estimate establishes scale; the Lancet Psychiatry return establishes that treatment works. Neither is a forecast for your firm.
- Offer a break-even framing instead of an ROI promise. "This pays for itself if it helps us keep two people" invites a judgment call. "This returns four to one" invites an argument you will probably lose.
- Commit to measurement and a review date. Agree in advance what you will track and when you will report back. Our step-by-step guide to running a workplace mental health assessment covers how to set a baseline, and the resources library includes free survey and measurement tools.
- Be clear about what won't change quickly. Culture and access improvements usually take more than a year to show up in retention or claims data. Saying so up front protects the program at its first review.
The honest version is the persuasive one
The numbers on mental health and work are genuinely striking, and it is tempting to stretch them. Resist that. A finance leader who catches one inflated claim will discount the rest of the deck, including the parts that are solid.
The defensible position is simpler: poor mental health is common and costly, effective treatment works, and the organization's job is to make sure people can reach that treatment and aren't made worse by how work is organized. Everything beyond that should be measured locally, reported plainly, and adjusted as you learn what your own people need.
