Image by Gilles Lambert on Unsplash

Taxation is an issue that galvanises Australian public and political opinion. Public debate has recently included persistent claims across a spectrum of commentators that some are ‘not paying their fair share’ of tax—in particular, corporate tax, and more specifically, corporate tax avoidance.

A number of recent Government initiatives attempt to constrain corporate tax avoidance. Some might be characterised as trying to increase the costs of avoiding taxes, such as increased public transparency for large corporations (for instance, Tax Laws Amendment (2013 Measures No.2) Act 2013). Others try to constrain strategies or corporate structures historically used to avoid corporate tax (like the Tax Laws Amendment (Combating Multinational Tax Avoidance) Act, 2015).

However, one policy that has been in place in Australia since 1987 has received much less critical attention – dividend imputation. The recent changes to address corporate tax avoidance tend to be superficial and motivated by populism, while insufficient regard is paid to the indirect impacts of dividend imputation on the operation of the corporate tax system generally. Our paper looks at the impact of dividend imputation on corporate tax avoidance.

Double taxation and dividend imputation

In the so-called ‘classical tax system’ of the United States, the European Union and most other countries, corporate profits are simply taxed at the corporate tax rate applicable to the specific jurisdiction.

For example, if the corporate tax rate is 30 percent for large corporations, $30 would be paid on a profit of $100. After-tax earnings would be $70. If the corporation distributed these profits as a dividend, then the shareholder would pay tax on this income at their personal marginal tax rate. A marginal tax rate of 47 percent means an additional tax of $32.90. The combined impact is that, on profit of $100, $62.90 would be payable as tax. This is the problem of the ‘double taxation’ of corporate profits.

Many countries address the double taxation issue by taxing dividend income for shareholders at a lower rate. But this approach can be problematic. Taxing income at different rates for individual taxpayers creates potential anomalies relating to the structuring of income. This is already well-demonstrated in Australia with negative gearing, which allows the recognition of capital gains that are concessionally taxed, while permitting expenses to be deductible against income at higher rates. A more extreme problem is that it incentivises firms to engage in corporate tax avoidance, to maximise after-tax profits and dividends to shareholders.

Dividend imputation is an alternative solution. Shareholders get a tax credit for corporate tax already paid on the dividends received. Continuing the above example, the shareholder would receive a dividend of $70, with an imputation credit of $30 reflecting the corporate tax paid. The shareholder would report the combined value of $100 as income and tax of $47 would be due. But with the $30 tax credit (known as a ‘franking’ or ‘imputation’ credit), only the balance of $17 would be payable. This is similar to the taxes withheld on wages, where the dividend is like net wages and the imputation credit is akin to Pay-As-You-Go (PAYG) tax withheld.

This solution maintains a consistent tax rate for all income for an individual, so there are no incentives to structure income in different ways. A more critical feature, which is generally not well understood, is that it reduces incentives for corporate tax avoidance since shareholders benefit from corporate tax paid.

After 30 years, there is evidence that dividend imputation in Australia has increased the payment of dividends as corporations pass the imputation credits on to shareholders. It has also led to reduced leverage as the value of the tax shelter arising from debt is reduced (increasing the cost of debt) and the after-tax cost of equity reduced.

We investigated whether dividend imputation has a further potential impact of limiting the benefits of corporate tax avoidance. Put simply, our main hypothesis was that if shareholders receive a benefit commensurate to the costs of corporate tax payments, there is little incentive to engage in corporate tax avoidance.

Evaluating differences in the level of tax avoidance

Using a sample of 4,729 firm-year observations between 2004 and 2015, we analyse variations in effective tax rate as a proxy to corporate tax avoidance across firms (i) not paying dividends, (ii) paying dividends without tax credits, and (iii) paying dividends with tax credits. If dividend imputation indeed reduces the incentives for tax avoidance, different levels of tax avoidance are expected between these three groups.

We find that the results are economically significant. Firms paying dividends with tax credits attached have an average cash effective tax rate almost 17 percent higher than those paying dividends without tax credits. Additionally, the latter have an average cash effective tax rate approximately 2 percent lower than firms that do not pay dividends. These results are presented in Table 1 and support the main hypothesis: dividend imputation reduces the incentive for tax avoidance.

Table 1: Association between dividend imputation and corporate tax avoidance
  Coef. t-stat.  
Constant       -0.104 -2.590 **
DivTC          0.147 5.432 ***
DivNTC         -0.022 -1.793
Outside%       -0.076 -2.038 *
Outside%*DivTC -0.044 -0.911
Size           0.012 6.004 ***
ROA            -0.001 -1.350
Leverage -0.032 -1.385
PP&E 0.079 3.030 **
R&D 0.082 1.088
Intangibles 0.108 6.372 ***
Total Accruals  -0.026 -1.373
Observations 4729
R-squared 0.154
Adjusted R-squared 0.149
F-Stat. 36.33
Year & Industry Fixed Effects
* p<0.05, ** p<0.01, *** p<0.001


The remaining three hypotheses relate to management behaviour for maximising shareholder wealth. These were proposed because of unsubstantiated assumptions made in the literature about management pursuits in companies that pay dividends with tax credits.

Our results show that the costs and benefits of tax avoidance are heterogeneous across firms. Most importantly, there is significantly less tax avoidance for firms paying dividends with tax credits, and this provides empirical evidence of how dividend imputation reduces the benefits of corporate tax avoidance. However, even when controlling for this there is still variation in tax avoidance across firms. Some of this may be explained by governance characteristics, but further study is required.

Policy implications

The main result in confirming the downward effect of imputation on tax avoidance is intuitive. However, it is important because it has a number of policy implications.

First, there have been calls to dismantle dividend imputation in recent years, as canvassed in an Australian Treasury Department discussion paper in 2015. This correctly identifies that Australia is one of the few countries with a system of dividend imputation. However, there appears to be a significantly lower incidence[1] of corporate tax avoidance by large domestic corporations in Australia than in other countries. Hence, any benefit to Government revenues gained by dismantling or limiting dividend imputation would likely be countered by increased corporate tax avoidance.

Second, if large Australian corporations benefit little from corporate tax avoidance, and there is limited scope for small and medium corporations to engage in corporate tax avoidance (because their operations are primarily in Australia and there are significant transaction costs associated with corporate tax avoidance), then regulatory attention focussed on foreign multinationals with operations in Australia would likely yield improved overall results for net tax collections.

Third, limiting the dividend imputation system (e.g. limiting refunds of imputation credits as proposed by the Australian Labor Party) will impact corporations’ incentives to pay dividends. This could lead to different or additional shareholder ‘clienteles’, whereby those shareholders unable to benefit from imputation credits invest in firms not paying dividends. For these firms not paying dividends, or paying dividends without imputation credits, there will be increased incentives for corporate tax avoidance. Hence, the budget impacts of limiting the operation of dividend imputation will be overstated.

Much recent debate over corporate tax cuts (as proposed by the Liberal/National Party) has been ill-informed. The public discourse has generally assumed homogeneity of the effect of tax cuts across firms. This ignores the incidence of corporate tax and the off-setting effects of dividend imputation.

Essentially, for large Australian corporations, there is little evidence of corporate tax avoidance and they are paying dividends with imputation credits. The budget impacts of a reduced corporate tax would be offset by reduced imputation credits and higher tax payments (smaller refunds) by individuals. The benefits of a tax cut for any Australian company will be constrained by the extent to which they presently pay corporate tax. That is, the lower the effective tax rate, the lower the benefit and the less the impact on the budget. Thus, as suggested by anecdotal evidence that the Australian subsidiaries of foreign multinationals are aggressively avoiding corporate tax, the impact on them and the budget from a tax cut will be negligible. That leaves small business and the evidence is that for these firms any extra profit (and cash) is likely to be reinvested.

In Australia, there is a need for tax reform, and this should be informed by rigorous analysis in terms of its detailed effect on the Federal Budget. However, changes to dividend imputation would be best characterised as an ‘ad hoc’ change to the tax system and focused on a particular group of tax payers. Such changes are highly problematic as they inevitably have indirect and unexpected outcomes.

[1] The 2014 Tax Justice Network report, ‘Who Pays for Our Commonwealth?’ reported a mean of 26% (87% of the statutory tax rate (STR)) for Australian companies, compared to Dyreng, Hanlon & Maydew (2008) who reported the equivalent from the U.S. of 27% (77% of the STR).

Leave a comment

Your email address will not be published. Required fields are marked *