with Adam Butt and Gaurav Khemka
The Australian Labor Party has proposed a change in policy under which imputation tax credits can only be offset against existing tax liabilities, with some exceptions such as pensioners and charities. Retirees are the major group that would be impacted by this policy, given that most are untaxed and hence currently able to claim the full value of imputation credits as a tax refund.
Such a policy change would effectively reduce the returns that such retirees receive from investing in Australian equities by the amount of imputation credits, which average 1.3%-1.4% per annum for the Australian market overall. This is a significant number, noting that the expected long-run equity market return might be in the order of 7%-8% per annum. It is no wonder this policy is a subject of heated discussion and much consternation from those nearing or in retirement.
The portfolio effects of franking credits
Our research addresses what full access to imputation tax credits means for Australian retirees in two ways.
First, we ask how imputation could affect how they might invest. Specifically, we find that retirees are justified in having a considerable bias toward Australian equities in their portfolio to capture the imputation credits.
Second, we estimate how valuable imputation credits are to retirees. We confirm they are potentially the equivalent of a 5%-6% increase in spending during retirement.
Our approach involves modelling rational behaviour for retirees who are funding their retirement out of an account-based pension, and may access the age pension under existing eligibility rules. We model retirees with starting balances at age 65 ranging from $25,000 up to $1.6 million (i.e. the cap on tax-free retirement accounts). We assume retirees form their portfolios and drawdown on their pension accounts to maximise their spending outcomes until they die.
We also model two types of retirees with differing preferences. One type prefers a higher level of spending spread over the course of their retirement. The other type has a target spending level, based around either the ‘comfortable’ or ‘modest’ retirement spending standards of the Association of Superannuation Funds of Australia. (In technical terms, the first type is modelled using power utility, and the second using a reference-dependent utility function). The model is run both excluding and including imputation credits, and the difference compared.
Our first finding is that access to imputation credits can support holding a portfolio with a considerable ‘home bias’ to Australian equities, largely at the expense of lower exposure to world equities. The exact portfolio breakdown depends on how the analysis is set up, including the assumed type of retiree, their starting balance, and their age.
To illustrate the tenor of the results, consider a retiree starting with a balance of $500,000 at age 65, who targets spending at the 'AFSA comfortable' of $42,764 per annum. Excluding imputation credits, our modelling suggests that this retiree should allocate their portfolio on average over the course of retirement to 26% in Australian equities, 33% in world equities and 41% in fixed income. When imputation credits are included in the analysis, the portfolio breakdown comes out as 46% in Australian equities, 15% in world equities and 39% in fixed income – a notable home bias.
The reason for the sizable switch away from world equities under imputation is that Australian equities offer substantially higher returns for a retiree who can claim the full credits, but without a meaningful increase in overall portfolio risk. The limited impact on portfolio risk arises because Australian and world equities are substitutes to a large extent. The retiree is swapping one form of equity market risk for another in order to improve their outcomes on a risk/return basis.
The value of imputation credits
We then estimate the value to retirees of having access to the full tax refunds from imputation credits. We do this through converting the uplift in benefit (utility) arising from imputation into three measures that can be readily interpreted. Again, the exact estimates vary with modelling set-up, so we will convey broad averages across retiree types and starting balances.
We find that imputation delivers a value equivalent to an average 5%-6% increase in spending over the course of retirement, an 8%-9% larger superannuation fund balance at the point of retirement, or a 0.6%-0.8% per annum increase in returns on the portfolio during retirement. Such not-insignificant numbers underwrite the consternation among those in or nearing retirement about a potential change in policy.
Portfolio and policy implications
Our study suggests the policy change will have important implications.
First, academics have tended to view home bias as a ‘puzzle’ to be explained. Our findings suggest that equity home bias might be at least partly explained as a rational response to tax effects that lead to differential returns on investment choices which contribute similar amounts to overall portfolio risk.
Second, the benefit removal would likely have some substantive effects. To the extent that imputation credits supplement income in retirement, the loss of tax credits could exacerbate the problem of the adequacy of superannuation balances for supporting a reasonable level of retirement spending. To some extent, access to imputation credits in retirement might be seen as an alternative to making higher superannuation contributions while at work in order to generate retirement income.
Third, a change in policy might also result in retirees providing less support to Australian companies via the investments they make.
We also highlight the net cost to the government of providing access to imputation tax credits to retirees, accounting for the fact that there will be some offset through reduced age pension payments if tax refunds are retained. For example, we estimate a total expected net cost per individual over the course of their retirement of about $30,000 for retirees that retire with a $100,000 balance, and around $80,000 for those retiring with a $500,000 balance (in 2017-8 dollars). The largest benefit in dollar terms accrues to retirees with the largest initial balances, raising some questions around equality of the policy.
Dr. Gaurav Khemka and Associate Professors Adam Butt and Geoff Warren are from College of Business and Economics at The Australian National University. For the full research paper, click here.
The Australian Investors Association recently asked its members to send in their views on Labor's franking credits proposal. The replies have been compiled in a single document, linked here.