Last week the prime minister was asked about changes to the HELP loan system.  In response he referred to Accord recommendations that ‘the system can be made simpler and be made fairer’ (emphasis added), and that ‘we’ll be making announcements pretty soon on that’.

‘Fairer’ probably means a lower-of HELP debt indexation formula, moving the indexation date so that more recent compulsory repayments can be taken into account, and a marginal rate repayment system to remove high effective marginal tax rates.

But what did the PM mean by ‘simpler’? I’m guessing that this refers to moving away from the current 18-threshold and rate repayment system to a system with fewer thresholds.

Marginal repayment systems

Simplicity is not an inherent feature of marginal rate systems. Both the current repayment model, which charges a % of a HELP debtor’s total income at each threshold, and a marginal rate system, which charges a % of income above the threshold, can work on single or multiple rates.

England and New Zealand have simple marginal repayment systems, 9% and 12% respectively above a single threshold income.

But the Accord report impliedly suggested a multi-rate marginal system, something made more explicit in media comments by Bruce Chapman, who had been advising the Accord panel. A March blog post of mine cites the evidence from the time. But it wasn’t clear exactly what marginal rates were intended.

The Chapman proposal

An article last week by Peter Martin in The Conversation linked to Chapman’s previously unpublished Accord paper on a multi-rate marginal system. It presented 3-rate and 4-rate marginal options, shown below. Three-rate or 4-rate systems are simpler than an 18-rate system, but more complex than the English or NZ systems.

The first threshold is from 2022-23. Option 1 may have been what Jason Clare was referring to when he said that someone on $75,000 a year could end up repaying about $1000 less than now, although on my calculations on 2022-23 thresholds it would be $727 less. Option 2 would be $152 less.

But indexation of the current thresholds between 2022-23 and 2023-24 has already effectively delivered an equivalent saving for someone on $75,000. Admittedly this was due to a quirk of the current system, which caused someone on $75K to fall back a full percentage point of total income repayment, instead of the more typical half a percentage point.

Quirks aside, just as a marginal rate system is not inherently simpler than the current system nor does it inherently reduce repayments, except for people with incomes just above the threshold for each repayment rate.

Chapman’s options compared to mine

In my post on possible marginal rates I gave two three-rate examples. The most important difference between my examples and Chapman’s options is that he offers only a small income range between the first and second rates of about $7500, while my range was more than $20,000.

Chapman’s options recognise a point that I made, which is that in marginal systems the lower rates matter much more for repayment revenue than they do in total income systems. This is because all repaying debtors in marginal rate systems get the lower rate.

With a marginal rate system governments have a greater incentive to lower the first threshold. When the previous government reduced the first threshold from $51,997 in 2018-19 to $45,881 in 2019-20 it only affected people within incomes between those two amounts. With a 1% of total income rate the revenue gain was not large. On a marginal rate system every repaying debtor would pay an additional 6% of their income between those two amounts.

The politics

For the politics of reform, a downside of low marginal rates on lower income levels is that, to maintain repayment revenue, they require high top rates and/or applying the top rate to relatively low incomes. Chapman’s option two would put most repaying debtors into the top 15% repayment category.

The marginal rate system would remove high HELP EMTRs for people just above each repayment threshold. But for debtors with incomes in between threshold levels the situation is less favourable. The HELP EMTRs in the 15% Chapman examples would be higher than under the current system, as seen in the chart below.

As I said in my post on the politics of moving to a marginal rate system if we could go back to 1988, when the loan system was first legislated, I would support a marginal rate system. But the practical and political complexities of transitioning to a marginal rate system now are significant. I would count the repayment system as a lower-priority reform than indexation levels and dates.