Insurance tax wasn’t complicated enough? Enter the REIT
Insurance tax wasn’t complicated enough? Enter the REIT
By Louis van Manen and Geraldene Goldner
For most tax professionals, corporate income tax follows a relatively familiar script.
There is one taxpayer. One set of tax rates. One set of financial statements. One tax computation.
Insurance tax professionals do not have that luxury.
While much of the corporate tax world debates transfer pricing, interest limitation rules and deferred tax movements, tax specialists within long-term and short-term insurers additionally spend their days navigating legislation that is often unique to the insurance industry and largely unfamiliar to the broader tax profession.
The irony is that most policyholders, investors and even many tax professionals will never appreciate just how complex the underlying tax calculations are.
It operates at the intersection of tax, accounting, actuarial valuations, investments and insurance regulation. Different types of insurance business have their own tax rules. Investment returns can have different consequences depending on where they arise. Accounting and actuarial measurements can feed directly into tax calculations. And, in the long-term insurance environment, even identifying the relevant taxpayer becomes more interesting.
Insurance tax teams have also spent the past few years navigating the introduction of IFRS 17 and the tax consequences arising from a fundamental change in insurance accounting.
One might therefore forgive the insurance tax manager for thinking that, having got on top of IFRS 17, the landscape might become a little quieter.
Not quite.
As of August 2026, qualifying unlisted companies wholly owned by certain institutional investors, including long-term insurers and short-term insurers, can potentially qualify as Real Estate Investment Trusts, or REITs.
That creates an interesting opportunity for insurers with significant property portfolios. It also introduces another specialist tax regime to an already unusually specialist world.
One insurer. Potentially five taxpayers and a REIT.
Perhaps the best illustration of how differently insurance tax operates is the taxation of long-term insurers under section 29A of the Income Tax Act.
A long-term insurer is required to establish and maintain up to five separate funds: the individual policyholder fund (IPF), company policyholder fund (CPF), untaxed policyholder fund (UPF), risk policy fund (RPF) and corporate fund (CF). These funds form the foundation of section 29A, with the tax position of the respective funds determined separately under the specialised legislative framework applying to long-term insurers.
So, there is one legal entity and one insurer, but potentially five different tax worlds within it.
The underlying architecture broadly distinguishes between different policyholder interests and the shareholder interest of the insurer. Even the risk policy fund, which at first glance looks more familiar because premiums and claims enter the calculation, has its own particular treatment. The RPF was introduced specifically to distinguish risk business from investment business within the five-fund system.
Yet, somewhat ironically, the RPF itself is not intended to bear normal tax. Section 29A(11)(bA) allows a deduction equal to the RPF’s taxable income before that deduction, effectively reducing the fund’s taxable income to nil. That does not mean that risk business is tax-free across the insurer as a whole: the mechanics of the five-fund system may still move value into the corporate fund, where the resulting amounts are brought into the insurer’s tax position. It does mean, however, that the fund created specifically to house risk business does not itself pay normal tax.
This is the first mind shift required in insurance tax: one company does not necessarily mean one conventional tax computation.
Assets, liabilities and transferring value to yourself
The five-fund system becomes more interesting when assets and policyholder liabilities enter the picture.
The section 29A framework requires the value of liabilities in the relevant policyholder and risk policy funds to be re-determined annually. Where the market value of assets exceeds liabilities withing the policyholder funds, assets equal to that excess are transferred to the corporate fund to be taxed in that fund. Where assets are insufficient to meet the liabilities, assets equal to the shortfall are transferred in the opposite direction.
These transfers operate as notional adjustments for purposes of calculating the insurer's annual tax liability and do not change the insurer's legal ownership of the assets concerned.
In less technical language, the tax system requires an insurer to transfer value from itself to itself, but between different deemed taxpayers.
Shortfalls can also give rise to return transfer credits, which influence the tax consequences when the relevant fund subsequently moves back into an excess position.
It is a good example of why "normal" corporate tax instincts can only take you so far in an insurance company.
The investment portfolio adds another layer
Insurers are significant investors, so investment taxation sits at the centre of the insurance tax function rather than at its edges.
Investment portfolios may contain cash, bonds, listed and unlisted equities, property and other asset classes. Those investments can generate interest, dividends, rental income and capital gains, each bringing its own tax considerations.
The five-fund framework then adds another question: in which fund does the return arise?
Capital gains provide a particularly good illustration. Different CGT inclusion rates can apply within different funds of the same legal entity, reflecting the particular tax treatment applicable to those funds.
That is a fairly unusual proposition in corporate tax: a single company can effectively find itself applying different CGT consequences across different parts of its investment portfolio.
Interest-bearing investments add their own complexity. Section 24J requires interest to be recognised for tax purposes on a yield-to-maturity basis rather than simply by reference to the receipt of cash.
Dividend-producing investments introduce further considerations around exempt income and the deductibility of associated expenditure. Section 29A contains specific apportionment formulas while SARS’s Binding General Ruling 30 (Issue 2) deals with the allocation of expenditure within and between an insurer's funds, necessitated by the unusual architecture of the five-fund system.
But the broader point is more important than any individual rule.
An insurance tax manager is not simply applying ordinary investment tax principles. Those principles have to be applied through the additional lens of the insurance tax regime.
A further, less visible consequence of the five-fund framework lies in the treatment of assessed losses. The manner in which the IPF and CPF are taxed has, in many cases, resulted in significant assessed losses accumulating within those funds. Because those losses remain confined to the fund in which they arose, they cannot simply be deployed elsewhere in the insurer. Depending on the nature and performance of the underlying insurance products, there may also be little prospect of generating sufficient future taxable income in the relevant fund against which to utilise them. The result can be substantial tax losses that are not merely ring-fenced, but effectively trapped—sometimes indefinitely.
Then IFRS 17 arrived
If the five-fund system was not sufficiently specialised already, insurance tax principles also has an unusually close relationship with financial reporting.
Section 29A expressly incorporates concepts derived from IFRS measurements, including insurance and investment contract liabilities, reinsurance contract liabilities, liability for incurred claims and deferred acquisition costs.
Against that background, the introduction of IFRS 17 was never going to be simply an accounting exercise.
IFRS 17 became effective for annual reporting periods commencing on or after 1 January 2023 and fundamentally changed aspects of the recognition and measurement of insurance contracts. Its introduction accordingly had implications for the tax calculations of insurers and required consequential changes to the South African tax framework. These changes include a 6-year transitional phasing-in mechanism for tax.
For the tax professional, it also changed the language and architecture of the financial information feeding into the tax calculation. Familiar concepts and balances do not necessarily appear in familiar places or in the same form.
The consequence is that an insurance tax manager increasingly needs to understand not only the tax legislation, but enough of the accounting and actuarial mechanics to understand where the numbers feeding into the tax computation originate and what they represent.
Short-term insurance has its own rules too
Long-term insurance may provide some of the more colourful examples, but the specialised environment does not end there.
Short-term insurance business is governed by its own tax framework under section 28 of the Income Tax Act, including insurance-specific rules relevant to determining taxable income. Long-term insurers licensed as micro-insurers also fall within the tax framework of section 28.
IFRS 17 affected this environment too, requiring short-term insurers and their tax teams to consider how changes in accounting measurement interact with the applicable tax regime. Here a 3-year transitional phasing-in mechanism applies for tax.
Whether dealing with long-term or short-term insurance, therefore, the recurring theme is much the same: ordinary tax concepts frequently acquire another layer when they enter an insurance company.
Which makes the latest addition particularly interesting.
And now we add a REIT
Until recently, South Africa's REIT tax regime was exclusively the domain of listed property groups.
That changed on 19 August 2026, when the first stage of the expansion of the regime to qualifying unlisted companies came into operation. Importantly for the insurance sector, this includes qualifying unlisted companies wholly owned by specified institutional investors, including long-term insurers and short-term insurers.
An insurer with a substantial property portfolio held through a wholly owned subsidiary may therefore now have another question to consider:
Should the property company become a REIT?
There are, inevitably, conditions.
The new framework includes an 80% property-asset threshold, a 75% rental-income requirement, a 60% debt-to-asset ceiling, minimum gross assets of R300 million and a requirement to distribute at least 75% of distributable profits within six months after year-end. These requirements also need to be monitored to ensure continued qualification.
But satisfying the REIT requirements is only half the story.
Where the REIT calculation ends, the insurance calculation begins
At its simplest, the REIT regime has a flow-through character.
Section 25BB provides a deduction for qualifying distributions made by a REIT or resident controlled company, subject to the requirements and limitations of the section. It also provides capital gains tax relief for qualifying property disposals while restricting specified allowances on immovable property.
That summary can make section 25BB sound deceptively straightforward. In practice, many of its complexities only emerge when its definitions, qualifying requirements, limitations and timing rules are worked through carefully against the facts of a particular structure or transaction. Seemingly modest differences between an existing arrangement and a contemplated one can produce materially different outcomes, making scenario testing an essential part of the analysis rather than a final sense-check. The point is perhaps best illustrated by SARS’s Interpretation Note 97 (Issue 3) on the taxation of REITs and controlled companies, which runs to roughly 100 pages—an unusually substantial interpretive treatment for a regime that can appear relatively simple when reduced to its headline principles.
For an insurer, however, the analysis cannot stop at the property-company level.
If the REIT obtains a deduction for a qualifying distribution, the next question is what happens when that distribution reaches the insurer. In which fund does it arise? What is its tax treatment in that fund? And does the combined outcome across the structure remain attractive once the insurance tax consequences are taken into account?
Existing deferred tax positions and historic allowances may therefore also need to be reconsidered.
The important question is consequently not simply whether REIT status reduces tax in the property subsidiary.
It is whether the combined outcome makes sense across the insurance structure as a whole.
And just like that, what started as a specialist property-tax question becomes an insurance-tax question.
One more specialist regime for the desk
The expansion of the REIT regime is a welcome and timely development for institutional property investors.
For insurers with significant property portfolios, it provides another option when considering how those assets should be structured and how their returns should flow through the wider insurance business.
But it also illustrates something broader about modern insurance tax.
The insurance tax function already operates across the five-fund system, investments, capital gains, deferred tax, accounting developments, actuarial information and the specialist regimes applicable to long-term and short-term insurance. Increasingly, it must also understand how the tax regimes applying to the assets in which insurers invest interact with that framework.
That breadth is what makes insurance tax challenging.
It is also what makes it interesting.
Having navigated IFRS 17, an insurance tax manager might therefore be forgiven for looking at the fund calculations, actuarial information, investment schedules and deferred tax workings covering the desk and thinking:
"Right. I think we have this under control."
And then somebody walks in and asks:
"Have we considered making the property subsidiary a REIT?"
Perhaps there really is always room for one more file on the desk.