Editor’s note: This is general educational information about how Australia’s dividend imputation provisions are written. It is not tax advice or investment advice, and no reader should treat it as a statement of their own tax position. It is based on the legislation and company disclosure listed at the end.
Analysis: three separate gates, not one benefit
The usual summary, that franking removes double taxation, describes the intent but not the structure. The Act creates three gates in series. The company must be eligible to frank and must allocate a credit. The credit is capped by reference to tax the company could actually have paid on the underlying profits. The recipient must clear residency and anti-avoidance tests before the gross-up and offset apply at all. A dividend can pass the first two and fail the third.
That is why the value of a franked dividend cannot be quoted as a single number. The cash is fixed and identical for every holder of the same parcel. The credit is fixed by the company within a statutory ceiling. The use of the credit is determined by provisions that look at the recipient, not the payment. The cap in section 202-55, tied to the tax the paying entity could have paid at its own corporate tax rate for imputation purposes, is the reason a company that has not paid Australian tax on the profits behind a dividend cannot manufacture a credit for it.
The benchmark rule is the piece most often left out, and it explains something visible in disclosure. Because a listed public company paying an ordinary dividend to all holders under one resolution falls outside section 203-5, the franked and unfranked split it discloses reflects the state of its franking account and its underlying profits, not a policy choice about which shareholders to favour. A payment that is partly unfranked, like the 0.01 unfranked component in Telstra’s 1H26 interim dividend, is a split the company disclosed for that dividend, and on the mechanism above the franked portion is limited by the maximum franking credit in section 202-55.
For a reader working through a dividend statement, the useful questions follow the statute in order. How much of the payment is franked and how much is not. What credit the distribution statement states, remembering that section 202-65 caps it at the maximum regardless. And whether the recipient is the kind of entity Division 207 lets use it at all.
What the documents say
A franked dividend is often described as a dividend with tax already paid on it. The Income Tax Assessment Act 1997 is more precise, and the precision matters. Part 3-6 of that Act, headed The imputation system, sets out three separate questions: whether a company may attach a credit at all, how large the credit is allowed to be, and which recipients are entitled to use it. The answers are not the same for every shareholder, and none of them is decided by the size of the cash payment.
Franking is an allocation, and it has a ceiling
Section 202-5 states the conditions. An entity franks a distribution if the entity is a franking entity that satisfies the residency requirement when the distribution is made, the distribution is a frankable distribution, and the entity allocates a franking credit to the distribution. All three limbs have to hold. A note to the provision records that the mechanism by which an entity allocates the credit, whether by resolution or some other means, is determined by the entity itself.
The size of the credit is capped by law rather than by choice. Section 202-55 defines the maximum franking credit for a distribution as equivalent to the maximum amount of income tax the entity making the distribution could have paid, at its corporate tax rate for imputation purposes for the income year in which the distribution is made, on the profits underlying the distribution. Section 202-60 then says the amount of the franking credit on a distribution is that stated in the distribution statement, unless that amount exceeds the maximum, and gives a formula worked from the amount of the frankable distribution and the applicable gross-up rate, defined as the corporate tax gross-up rate of the entity making the distribution.
Overstating the credit does not achieve anything. Section 202-65 provides that where the credit stated in a distribution statement exceeds the maximum, the amount of the franking credit on the distribution is taken to be the maximum, and not the amount stated.
Partial franking is common, and it shows up plainly in company disclosure. Telstra’s published dividend history splits each payment into a franked amount and an unfranked amount per share. Its final dividend for 2H25, with an ex dividend date of 27-Aug-2025 and paid on 25-Sep-2025, carried a franked amount of 0.095 and no unfranked component. Its interim dividend for 1H26, ex dividend on 25-Feb-2026 and paid on 27-Mar-2026, carried a franked amount of 0.095 and an unfranked amount of 0.01. The same headline cents per share can therefore carry different amounts of attached credit.
What the shareholder does with the credit
The receiving side sits in Division 207. Section 207-20 states the general rule in two subsections. If an entity makes a franked distribution to another entity, the assessable income of the receiving entity for the income year in which the distribution is made includes the amount of the franking credit on the distribution, in addition to any other amount included in relation to the distribution. The receiving entity is then entitled to a tax offset for that income year equal to the franking credit.
That is the gross-up and offset in its statutory form. The credit is added to income and then subtracted from tax. Nothing in the section adjusts for the recipient’s own rate, which is why the outcome differs between recipients rather than between dividends.
The general rule is heavily qualified. Section 207-15 excludes partnerships and most trustees, routing them into Subdivision 207-B instead, and applies subject to Subdivisions 207-C, 207-D, 207-E and 207-F. Subdivision 207-C sets residency requirements that an individual or corporate tax entity receiving a franked distribution must satisfy. Subdivision 207-D removes the gross-up and offset in cases where the distribution, or a share of it, would not have been taxed in any case, with Subdivision 207-E carrying the exceptions to that removal. Subdivision 207-F disapplies the rules where the imputation system has been manipulated in a way the income tax law does not permit.
Whether an unused credit becomes cash is dealt with elsewhere again. Division 67 of the Act contains the refundable tax offset rules, and section 67-25 within it is headed Refundable tax offsets, franked distributions.
The benchmark rule stops selective generosity
A company cannot frank one distribution heavily and the next one lightly at will. Section 203-5 provides that a corporate tax entity must frank all frankable distributions made within a particular period at a franking percentage set as the benchmark for that period, and calls this the benchmark rule. Section 203-10 sets the benchmark franking percentage by reference to the franking percentage for the first frankable distribution made during the relevant period, and adds that an entity has a benchmark franking percentage even if it is not subject to the rule.
The object is stated at section 203-15: to ensure that one member of a corporate tax entity is not preferred over another when the entity franks distributions. Section 203-20 then exempts a company in a franking period where it is at all times during that period a listed public company and cannot make a distribution on one membership interest during the period without making a distribution under the same resolution on all others. Listed companies whose ordinary dividends necessarily go to every holder alike sit outside the rule, because the mischief it addresses cannot arise.
