Private Health Services Plan Rules for Canadian Corporations

A Private Health Services Plan lets a Canadian corporation pay for an employee's medical expenses and deduct the full cost as a business expense, with the reimbursement landing tax-free in the employee's hands. The plan has to actually qualify as a PHSP under CRA's rules. This article walks through what that qualification requires, step by step, from the legal definition through to the one scenario where CRA's own guidance contradicts itself.

What a Private Health Services Plan is for Canadian corporations

A PHSP is a specific legal arrangement defined in the Income Tax Act, and whether a corporate plan survives a CRA review starts with whether it meets that definition. CRA's payroll guidance is built around a single distinction: if a corporation pays or reimburses an employee for medical expenses and the plan qualifies as a PHSP, the payment is non-taxable. If the plan doesn't qualify, that same payment becomes a taxable benefit, no different from a bonus run through payroll.

Section 248(1) of the Income Tax Act defines a PHSP as a contract of insurance covering hospital expenses, medical expenses, or both, or as a medical care insurance plan or hospital care insurance plan. Provincial government health care insurance plans, as defined under the Canada Health Act, are explicitly excluded. That exclusion matters because it draws a clear line: a PHSP exists to cover what public health insurance doesn't, not to duplicate it.

In practice, most Canadian corporations set this up through a Health Spending Account, often called an HSA or HCSA. Different providers use different names, but they're describing the same CRA-recognized structure. The mechanics usually follow what's known as the cost-plus model. There's no premium and no deductible, and no insurance company sits between the employee and the claim approving or denying it.

The money moves in a fixed sequence. An employee incurs a medical expense and pays it out of pocket. The employee submits the receipt to the plan. The corporation's account reimburses the employee through the plan administrator. The corporation deducts the full reimbursed amount, plus the administration fee, as a business expense. The employee receives the money tax-free, and it never appears on a T4.

The tax treatment on both sides is what makes this structure attractive. The corporation deducts what it pays into or through the PHSP. Outside Quebec, the employee doesn't include the value of the benefit in income. No CPP or EI premiums apply to these amounts. The reimbursement avoids the payroll tax drag that a comparable cash bonus would carry.

None of this happens automatically just because a corporation calls its arrangement a PHSP. Qualification depends on meeting a specific set of CRA criteria, and those criteria are where most plans succeed or fail.

The five elements a plan must have to be considered insurance under CRA's definition

Before anything else is tested, CRA asks one threshold question: does this plan function as "a plan in the nature of insurance"? If it doesn't, nothing downstream matters. CRA's interpretation of that standard, summarized in Interpretation Bulletin IT-339R2, breaks the test into five elements, and all five have to be present for the plan to count as insurance.

The first element is an undertaking by one person. In a corporate PHSP, that person is the employer. The corporation doesn't need to buy a policy from an outside insurer to satisfy this piece. It simply has to be the party making the commitment.

The second element is an undertaking to indemnify another person. In the employment context, that person is the employee, along with any covered dependants, and the indemnification runs against the covered loss: the medical or hospital expense itself.

The third element is agreed consideration. The employee's side of the bargain is the promise to provide services to the corporation. In exchange, the employer provides PHSP coverage. That exchange, services for benefits, is what makes the arrangement an employment benefit rather than a personal gift.

The fourth element is a loss or liability tied to a specific event. The medical or hospital expense is that loss. Without a defined loss, there's nothing for the plan to indemnify against.

The fifth element is uncertainty: the event triggering the loss has to be genuinely unknown in advance. This is where most of CRA's scrutiny concentrates. If a plan is guaranteed to pay out regardless of whether any medical event occurs, there's no real insurance risk, and the arrangement stops qualifying under this test.

That uncertainty requirement is also where sole shareholder-sole employee situations run into trouble, a problem significant enough to warrant its own full treatment later in this article. When the person setting up the plan and the person receiving its benefits are the same individual, CRA questions whether any genuine risk-pooling exists, or whether the arrangement is simply the owner moving money to themselves through a different label. A plan covering only the owner and immediate family, with no arm's-length employee anywhere in the structure, has a hard time showing that the benefit flows from an employment relationship.

Passing the insurance-nature test is necessary, but it isn't sufficient on its own. A plan can clear all five elements and still fail to qualify if it covers the wrong things, which is the next test.

The 90% rule and eligible expenses

Once a plan clears the insurance-nature threshold, CRA applies a second test based on what the plan actually pays for. The rule requires that all, or substantially all, generally interpreted as 90% or more, of what's paid under the plan relate to expenses eligible for the Medical Expense Tax Credit. Fall below that threshold, and the plan loses its PHSP status regardless of how well-documented or well-administered it is.

How the test is measured depends on the type of plan. For an insured plan, CRA looks at the composition of premiums paid during the calendar year, not at which specific benefits were paid out. For a self-insured plan, which is what a standalone HSA typically is, there are no premiums to measure, so CRA instead looks at the benefits actually paid to employees over the calendar year. In both cases, the question is the same: does the money flowing through this plan overwhelmingly match the kinds of expenses the METC was built to cover.

That eligible expense list tracks section 118.2(2) of the Income Tax Act, the same section used to calculate the METC on a personal return. It covers prescriptions and drugs, dental treatment, vision care including glasses, contact lenses, and laser eye surgery, paramedical services such as physiotherapy, massage therapy, and chiropractic care, mental health services from psychologists and licensed therapists, medical devices prescribed by a physician, fertility treatments, and hospital expenses.

What falls outside that list is just as important to understand. Gym memberships don't qualify. Cosmetic procedures without a genuine medical purpose don't qualify. Over-the-counter products bought without a prescription don't qualify either. These are the categories that trip up plans administered loosely, because they look like health spending in a general sense without meeting the specific statutory definition.

CRA also allows for what it calls connected expenses: costs incurred within a reasonable window following a medical expense, directly tied to it, even if the cost itself isn't medical in nature. Transportation home after a procedure that leaves someone unable to drive is the kind of example CRA recognizes under this category. These connected costs can count toward the plan's eligible spending even though they wouldn't independently qualify as a medical expense.

The reason corporations use a PHSP instead of just having employees claim the METC personally comes down to the math. The METC only returns a partial credit on a personal return. A PHSP, by contrast, lets the corporation cover the eligible expense in full and deduct the entire amount as a business expense. For the same underlying medical cost, the PHSP structure produces a better tax outcome for both the corporation and the employee.

The stakes of the 90% rule are structural. A plan that lets even a modest stream of ineligible claims through, whether through a lax administrator or an owner approving their own spending, risks losing PHSP status across the board, not just on the ineligible claims themselves. That risk is why CRA puts so much weight on how the plan is administered, which is the next piece of the puzzle.

Structural requirements a corporation must meet before the first claim is filed

A plan can cover the right expenses and still fail a CRA review if it wasn't built correctly from the start. The plan has to be established, documented, and administered before any claim is submitted.

The first requirement is predetermined limits. A corporation is allowed to set different annual limits for different classes of employees, so management and staff can have different caps. What isn't allowed is leaving the limit undefined. A plan with no cap at all doesn't function as a plan in the nature of insurance, because there's no bounded risk for the employer to indemnify against.

The second requirement is third-party administration. CRA strongly favors, and in practice all but requires, that a qualified third-party administrator run the plan. Self-administered plans draw significantly more audit scrutiny, because self-administration undermines the arm's-length character the insurance-nature test depends on. A third-party administrator processes claims, checks each one against the METC-eligible expense list, issues reimbursements by electronic funds transfer, and produces the annual reporting the plan needs. Those are tasks a business owner can't credibly perform on their own claims without collapsing the very structure that makes the plan qualify. The administration fee charged for this work, typically a percentage of each approved claim, is itself a deductible business expense.

The third requirement is record retention. Every medical receipt submitted under the plan needs to be kept for a minimum of six years, so the corporation can support its claims if CRA opens a review.

The fourth requirement is consistency across employee classes. If a corporation has more than one employee, the plan has to be offered consistently within each class of employee. Offering the PHSP to the owner while leaving out another employee who sits in the same class isn't permitted. CRA reads inconsistent treatment within a class as evidence the benefit is really flowing to someone as an owner, not as an employee, which is the exact distinction the next section turns on.

The owner-employee: salary requirement and shareholder benefit risk

Most corporate PHSPs that get successfully challenged by CRA don't fail because of what they covered. They fail because of who received the benefit. When a shareholder receives a benefit as a shareholder rather than as an employee, that payment becomes a taxable shareholder benefit, and the corporation loses the deduction it claimed.

The foundational rule is straightforward: a PHSP is an employee benefit. To receive it tax-free, the owner has to actually be an employee of the corporation, actively engaged in the business and compensated for that work in an employment capacity, not simply holding shares in the company.

This is where the dividend-only compensation strategy creates a real problem. Many incorporated owners pay themselves entirely in dividends, often to reduce CPP contributions or manage their personal tax bracket. That's a legitimate compensation choice on its own terms, but a person paid only in dividends is a shareholder, not an employee. CRA can and does reclassify PHSP benefits paid to dividend-only owners as shareholder benefits under section 15(1) of the Income Tax Act. When that happens, the corporation loses the deduction, the owner owes personal tax on the amount reimbursed, and interest or penalties can apply on top of that. The result functions as double taxation on money that was meant to move tax-free.

This is a fixable structural issue, not a trap waiting to spring on unsuspecting owners. The fix is to put the owner on payroll: T4 employment income, a real employment contract with the corporation, and compensation that reflects actual work performed for the business. What salary level satisfies CRA depends on the owner's province, their broader compensation structure, and how much the PHSP itself gets used. That's a calculation for a qualified accountant to run against the specific numbers involved, because the Income Tax Act doesn't set a fixed dollar threshold here.

What CRA does require is reasonableness. The benefits a shareholder-employee receives through a PHSP have to be consistent with what an arm's-length employee doing similar work would receive. There's no statutory cap on PHSP deductions, but the benefit can't be wildly disproportionate to the employment relationship it's supposedly tied to. That reasonableness standard is genuinely fact-specific, and no single salary figure makes a plan automatically safe.

This problem is hardest when there's no arm's-length employee to compare against at all, because the owner is the only person on payroll.

The sole shareholder–sole employee problem and CRA's current position

When the sole shareholder of a corporation is also its only employee, two separate tests collide at once: the insurance-nature requirement and the shareholder benefit rules. Both point at the same weakness. With only one person in the entire arrangement, the plan has a hard time proving that its benefits flow from an employment relationship.

CRA's position on this exact scenario is on the record. Technical Interpretation 2014-0521301E5, dated June 25, 2014, states that a plan covering a sole employee-shareholder does not qualify as a PHSP, because it lacks the necessary elements of insurance, and that a cost-plus plan in that situation would not likely constitute a plan in the nature of insurance. CRA's stated conclusion goes further: where the sole shareholder is also the sole employee, CRA will treat that person as receiving the benefit in their capacity as a shareholder, unless they can show that employees with similar duties and responsibilities at a similarly sized corporation receive similar benefits under a similar plan.

A later interpretation pushes the position further still. Technical Interpretation 2022-0928901C6, from the CALU Q10 roundtable, concludes that a self-insured HSA set up for a sole employee-shareholder and their family would likely not qualify as a plan in the nature of insurance, and therefore would not qualify as a PHSP.

That creates a real tension for professional corporations and other one-person operations. CRA's public-facing guidance confirms that a corporation with as few as one employee can be eligible to set up an HSA. Its own technical interpretations, meanwhile, cast serious doubt on whether a sole shareholder-employee structure can ever satisfy the insurance-nature test. These two positions sit uneasily next to each other, and resolving which one applies to a given corporation depends entirely on the specific facts of that corporation's structure and operations.