Improving compliance on tax deductions for work-related expenses
We partnered with the Australian Taxation Office to examine the extent to which behaviourally-informed communications to tax agents could reduce erroneous work related expense claims.
Recent unpublished ATO research has found that many taxpayers think it is wrong to over-claim work-related expenses and intend to do their best to report income and claim deductions correctly. However, in practice it appears some taxpayers think it is OK to over-claim ‘a little’. Ensuring deductions are claimed accurately is a priority for the Australian government and the ATO.
Approximately 65 per cent of Australian taxpayers claim work-related expense deductions. In 2015-16, 8.6 million taxpayers claimed nearly $22 billion in work-related expense deductions – representing a 15 per cent increase since 2010-11.
Over 70 per cent of Australians lodge their income tax return through a tax agent, so prompting a behaviour change among agents could lead to a larger increase in overall tax compliance.
The ATO uses data analytics to identify agents whose clients’ claims are higher than expected when compared with agents with similar clientele. Agents are then engaged directly to address compliance risks in relation to work-related expense claims in their client base.
In engaging with agents, the ATO uses a range of approaches to encourage compliance, from help and education through to enforcement activities. Unpublished work by the ATO has found that the most positive return on investment and sustainable compliance impact results from activities where interaction with agents is tailored (using data analytics) and involves direct dialogue between an ATO officer and agent.
In this trial we tested whether ATO data analytics could be coupled with behaviourally informed written communications to improve the accuracy of work-related expense claims in a low-cost manner.
We used a randomised controlled trial to test the impact of sending this letter to tax agents whose clients had higher than expected work-related expense claims compared with their peers. The letter specifically identified each agent’s clients who had lodged higher than expected claims (referred to as ‘identified clients’). There were over 2,000 tax agents in the trial, who were randomly assigned to a letter group or a ‘no letter’ control group.
In the letter group, 17 per cent of agents lodged amendments, compared with 1 per cent of agents in the ‘no letter’ group. On average, identified clients represented by agents in the letter group reduced their work-related expense claims by $191, compared with a $2 increase in the ‘no letter’ group. Overall, identified clients reduced their work-related expense claims by a total of $2.2 million. If we had sent the letter to all agents in the trial, including those who did not receive a letter, we expect this simple messaging intervention would have decreased work related expense deductions by about $4.4 million.
The letter led to an increase in the average total tax paid per client of $76, for a total of $0.9 million. If we had sent the letter to all agents in the trial, total tax revenue would have increased by an estimated $1.7 million. We cannot be sure about the contribution of the reduction in work-related expense claims to the increase in tax paid, because amendments to tax return labels other than work-related expenses could have influenced the net tax position.
Overall, these findings indicate low-cost, well-targeted and well-designed communications incorporating behavioural insights are effective in reducing erroneous work-related expense claims, although the length of time for which this behavioural change is sustained remains unknown.
So what does all this mean?
The results of the trial demonstrate coupling data analytics with communications incorporating behavioural insight techniques can influence compliance for work-related expense claims.
Identified clients represented by agents in the letter group reduced their work-related expense claims by a total of $2.2 million. On average, identified clients represented by agents in the letter group reduced their work-related expense claims by $191, compared with a $2 increase in the ‘no letter’ group.
While acknowledging factors other than work-related expense deductions may also be driving the amount of net tax paid, agents in the letter group also increased tax paid by $0.9 million. On average, tax paid increased per client by $76. Given there were as many agents in the ‘no letter’ group as in the letter group, we can estimate if the letter was sent to all agents in the trial, work-related expense claims would have reduced by $4.4 million and tax revenue increased by $1.7 million.
On the face of it, this result indicates an effective return on investment when the cost of sending the letters is compared with the additional tax revenue collected as a result of reducing deductions. However, noted alongside this result, unpublished ATO reports suggest other approaches employed within the ATO to influence compliance for work-related expense claims are resulting in a greater return on investment. The ATO has previously established that unless follow-up compliance work is conducted after letter campaigns, the effectiveness of future campaigns is decreased. The need for this follow-up action reduces the cost effectiveness of a letter-based approach.
Our trial shows it is important to test interventions because human behaviour is complex. Our trial showed the letter was effective at reducing work-related expense claims, but work-related expenses are just one part of the overall tax system and tax return process. As changes in deductions claimed in one category may be offset by increases in deductions claimed in other categories, future trials should connect and analyse data across all deduction categories to better identify shifts in behaviour around claiming deductions.
Michael Hiscox, founding director, Behavioural Economics Team within the Department of the Prime Minister and Cabinet