Incentives: Can the right incentives lead to the right behaviour?

It’s never easy to change a habit. And when that habit is enjoyable, it becomes extremely complex and difficult to change behaviour patterns¹. So, how do we get people to choose the celery stick instead of the chocolate bar, the treadmill instead of the couch and saving instead of spending?

For many South Africans, participation in a retirement fund or medical scheme is a condition of employment. However, employees often have a plan choice on the medical scheme and some may have choices regarding the level of contribution made to their retirement fund, either explicitly or by adjusting the percentage of their salary defined as pensionable. In addition, members can choose to make additional voluntary contributions (AVCs) to retirement funds.

Employees also have significant flexibility in their private provision for retirement and medical expenses. The problem is that left to their own devices, they may not choose appropriately. South African savings rates are notoriously low. Household debt to disposable income is currently 76.3%² and local households have had a negative net savings rate since 2005. In other words, South Africans borrow more than they save.

As of 31 December 2011, members contributed 13.5%³ of pensionable salaries towards retirement funding, well short of the 20%⁴ required to provide for a 75% replacement ratio at that date. With subsequent declines in bond yields, the required contribution rate shot up to 27% as at 30 September 2012. Considering the potential cost of medical scheme contributions during retirement, even the targeted replacement ratio of 75% might be too low, given that only a few individuals can still rely on their employer for postretirement subsidies towards medical scheme contributions. So could tax and other incentives change this picture?

Tax incentive

During the 2009 tax year, tax incentives for retirement savings and medical scheme contributions totalled R17bn⁵ and R15.7bn⁶ respectively. But despite the cost to the fiscus, the jury is out as to whether a tax incentive can boost benefits participation. If a household responds to a tax incentive by reducing spending or increasing their income by say working overtime so that they can save more, then the tax subsidy might be considered to be successful at boosting net private saving⁷. However, net savings do not increase in response to incentives if:

• The savings are funded by redirecting contributions from less tax-efficient to more tax-efficient vehicles.
• Savings would have been made anyway due to a strong savings culture.
• The household goes into debt in order to finance the saving.

In the absence of a strong savings culture or household budgeting, both of which are lacking in South Africa, incentives might be ineffective in raising savings rates to the required levels. 

Similarly, with medical scheme coverage, there is some US evidence to suggest that where medical insurance participation is voluntary, tax incentives may increase coverage but tax incentives alone are unlikely to result in full coverage⁸.

However, there are some arguments for offering tax incentives even if there’s mixed evidence of their effectiveness.

Retirement savings

​​​​​​​Firstly, authors like Douglas Bernheim suggest that people’s savings rates are influenced by what they think the ‘authorities’ suggest that they should save, and that people take the tax incentives as a suggestion from the ‘authorities’. 

Secondly, if a tax subsidy is presented clearly so that savers can understand the benefit, then it may improve savings behaviour. 

Under the new proposals⁹ for taxation of retirement fund contributions, employer contributions will be taxed as a fringe benefit and individuals will be permitted deductions of up to 22.5% of their income if aged under 45 and 27.5% of their income if aged 45 and over. These deductions will be limited to R250 000 per year if aged under 45 and R300 000 per year if aged over 45. 

However, if you start from the premise that tax incentives change behaviour, and bigger incentives yield larger results, the equity of the tax subsidy becomes important. In other words, policymakers would want to consider whether the subsidies are targeted at the right groups of people.

The proposed tax subsidy system benefits wealthier individuals more than lower-income earners. The graphic below shows the tax benefit for two savers both saving 22.5% of their income. In one US study, lower and middle-income households were much more likely to make voluntary retirement savings and to contribute more towards these plans if offered an additional contribution towards their retirement savings instead of a traditional tax subsidy¹⁰.

An alternative way to express the tax subsidy may be a 25% contribution match, based on the tax profile in the 2011 Tax Statistics¹¹. The contribution match would involve the current tax benefit on contributions being removed entirely, which would mean that retirement savings are made from post-tax money. However, the South African Revenue Service (SARS) would then contribute 25c for every R1 each saver contributes from their post-tax income to their retirement fund, with a maximum annual contribution from SARS of say R100 000. The figure of 25% and the maximum annual contribution would need to be revised from year to year as the taxpayer demographics evolve.

In 'Low-income earners', we suggested that at least some South Africans with low levels of income would benefit from saving. The incentive scheme could be extended to low-income earners who do not pay tax, provided that they register with SARS.

Although there may be some administrative challenges in moving to a contribution-match system, this way of presenting the subsidy may maximise the impact of the tax revenue foregone because it is a far easier system to understand than the tax deduction.

Healthcare

From a medical scheme point of view, the tax incentives for medical expenses were changed from tax deductions to tax credits with effect from 1 March 2012. Previously, medical scheme contributions could be used to reduce taxable income subject to certain limits. Now, the tax payable is reduced by the medical contribution tax credits, subject to a limit. The result is that individuals of different income levels, who are subject to different marginal tax rates, now receive tax incentives of equal rand value. This is illustrated in the table below.

The new system could be seen as incentivising lower-income earners to join medical schemes more than the old system did. However, due to the recent implementation of the new system, there is currently no way to evaluate how successful it has been in achieving this.

Tax deduction for medical scheme contribution 

Incentives programme

Incentives and rewards programmes have become increasingly popular but one can be sceptical about whether they truly change behaviour or are just gimmicks. For an incentive programme to be successful,
three things are required:

• Registration
• Initial use of the programme
• Sustained use.

The easier the benefits of the incentive programmes are to understand, the higher the registration, the initial use and the sustained use will be. Registration statistics vary but even if a fee is involved, about 60% of eligible members may enrol¹². The level of engagement can be surprisingly low. Only 20%-57% of enrolled members use loyalty programmes.

In terms of sustained use, most studies show some sort of attrition over time, either due to drop-out or waning enthusiasm. A total of 40% of consumers drop out of consumer loyalty programmes, citing reasons such as:

• The rewards are unappealing, unavailable or of little value.
• Difficulty accessing the rewards. 
• Rules changing or points expiring. 
• Privacy issues¹².

However, for the minority who stay the course, we may observe sustained positive behaviour change¹³.

Implications

The change to the retirement tax subsidy proposed by the National Treasury is unlikely to have an impact on retirement savings rates. A cleverly designed savings incentives programme may have the desired result,
provided that employees remain sufficiently engaged. Designing incentive schemes where the man in the street understands the impact on his wallet remains a challenge for policymakers and the financial services industry.

References

1 Patel et al (2011)
2 South African Reserve Bank (2012)
3 Member WatchTM 2012 data set
4 Alexander Forbes Research and Product Development
5 National Treasury (2012a)
6 National Treasury (2011)
7 Engen (1996)
8 Gruber (2000)
9 National Treasury (2012d)
10 Duflo et al (2005)
11 South African Revenue Services & National Treasury (2011)
12 Lambert et al (2009)
13 Patel et al (2011)