14 September 2005
Figures show interest free student loans policy easily affordable
“Costings released today show Labour’s interest free student loans policy is easily affordable and, even over a 15 year
timeframe, would have a negligible impact on government debt,” Finance Minister Michael Cullen said today.
“This is in sharp contrast to the $11 billion plus in debt servicing costs which would accumulate over the same 15 year
period to support the additional $1.3 billion a year each year from 2009 National would have to borrow to fund its tax
cuts,” he said. [Refer separate press statement.]
“The documents make it clear that many of the assumptions on which the first round of Treasury costings were done were
unrealistic. They assumed, for example, that course fees would rise 3.6 per cent a year on average which is well above
the rate of inflation and totally ignores the impact of the government’s fee maxima policy.
“They also took no account of Labour policy to expand access to student allowances to 50 per cent of students. This led
to a very large overestimation of costs, particularly over the longer term,” Dr Cullen said.
“Labour has already made public its own costings and Treasury’s initial costings over four years; this being the normal
forecast horizon used by the Treasury in preparing its budget forecasts and by political parties in costing their
manifesto promises.
“No formal Treasury reports were prepared on the issue because it is Labour Party for the election campaign rather than
government policy. Instead the data were communicated to a member of my staff via e-mail.
“As I told the House on 2 August, I asked Treasury to cost a range of scenarios, including around interest free student
loans and universal student allowances. It provided these in the context of the Long Term Review of Tertiary Education
Expenditure.
“This was listed as a specific fiscal risk in both this year’s budget and the Pre-Election Economic and Fiscal Update.
Because the review is ongoing, I thought it should be withheld under the normal constitutional conventions applying to
ongoing work.
“I did, however, release both sets of four year costings to assist the public debate and to debunk some of the more
extravagant comments being made by National and their allies,” Dr Cullen said.
“The Ombudsman has indicated to me that he takes a slightly different view of the application of the Official
Information Act in this circumstance and, as I promised at the outset, I am cooperating with his interpretation although
I do have serious concerns about the precedent it creates. We will be taking these up after the election”
Dr Cullen said it was important to understand correctly the status of the figures provided by the Treasury. As the
Secretary of the Treasury explained to the Ombudsman [letter attached]; they were of a highly preliminary nature, had
not been subjected to any of Treasury’s usual quality assurance controls and were scenarios intended to provide a sense
of the order of magnitude of different policy options rather than formal costings.
“I have already said publicly that I did not agree with the assumptions underpinning the first set of costings. These
were that the take up rate for full time students would rise to 95 per cent for fees, 85 per cent for course costs and
75 per cent for living costs with the rates for part-time students rising to 40 per cent for fees and 30 per cent for
course costs. [Part-timers are not eligible for living costs.]
“I considered these estimates all too high, not least because officials hopelessly over-estimated the increase in the
take-up rate when Labour made loans interest free during study, predicting that it would rise from 50 per cent to 90 per
cent when in fact it rose by only five points to 55 per cent.
“I asked Treasury to do a second set of costings based on a lower set of take-up assumptions and it is these that the
Labour Party drew on for its costings. I am fully satisfied that these assumptions are realistic and that Labour’s
costings are robust,” Dr Cullen said.
“Were I not, I would not have allowed the policy to go forward.”
******************
Attached: Letter from the Secretary to the Treasury to the Ombudsman
13 September 2005
John Belgrave
Chief Ombudsman
Level 14
70 The Terrace
PO Box 10152
Dear John
You have recently initiated an urgent inquiry into the decision of the Minister of Finance to withhold information
relating to the Official Information Act request from Nicola Willis of the National Party Research Unit regarding “All
advice prepared in relation to the Government’s policy proposal to abolish all interest changes on student loans for
those who reside in New Zealand”.
In the event that you form a view that the Treasury papers captured by this request should be released, I thought it
would be helpful to clarify the status of the costings set out in the papers, so that any comments associated with their
release are factually accurate. Namely, they represent broad cost estimates of different student support scenarios and
were intended to provide the Minister with a sense of the order of magnitude of different policy options rather than
formal Treasury costings of particular student support policies.
This is important because should any of these policy options, including the policy to remove interest on all student
loans, be presented for active consideration as government policy in the future, the Treasury would need to undertake a
more formal exercise to cost the policy. I cannot say at this stage whether this would result in a higher or lower cost
of that policy, or how significant the magnitude of these changes might be.
As you know, the Treasury was asked by the Minister of Finance to provide estimates of the fiscal costs of various
scenarios involving changes to student support over a period of five weeks or so.
Those scenarios costed later in this time period related to a policy to remove all interest on student loans. Such
requests are not uncommon in the preliminary stages of policy development, where Ministers are interested in getting a
sense of the order of magnitude of different options.
In this case, we were also asked to provide these cost estimates with relatively short turnaround times. Accordingly, we
were constrained in the information we could bring to bear and had to limit the extent to which we involved other
agencies in developing and testing the assumptions and costings.
The status of the costings is captured in our first email to the Minister’s office of 9 June 2005, in which we stated:
“This estimate is rudimentary and can only be relied upon to show the order of magnitude of costs.” We reinforced this
qualification in our email to the Minister’s office of 22 June where we describe the costing as “rudimentary” and
“basic”.
To provide some context for these comments, I thought it would be useful to set out the approach that was adopted in
producing the cost estimates, and identify the nature and scope of the refinements that would be expected in order to
produce an acceptable policy costing.
First, the cost estimates were produced assuming that all existing student loan policy parameters remained unchanged. In
practice, when designing a policy such as removing interest on student loans, Ministers have a number of parameters they
can adjust depending on their policy objectives. These include (but are not limited to) loan eligibility criteria, term
limits that may be placed on the loan, restrictions on access to tertiary education, and changes to income repayment
thresholds.
Once the broad direction of policy is agreed, we would expect Ministers to make decisions on the parameters as part of
the normal government policy development process. These policy details can potentially have a significant impact on the
costs of a proposal.
Secondly, we were unable to use the full resources we would normally draw on to cost such policies. In this instance
time constraints meant that the Ministry of Education was only able to use the Tertiary Education Student Loan Analysis
(TESLA) model for some of the variations, and some estimates (including those for the policy to remove interest on
student loans) had to be done using TESLA data, but outside of the TESLA model.
Thirdly, the cost estimates do not include any estimate of the implementation costs associated with the design,
development, or delivery of the policies.
In addition, the earlier cost estimates were produced within the Treasury without consulting any other agencies, while
for the later estimates (including the cost of removing interest on student loans) consultation was limited to the
Ministry of Education. This meant that we were not able to access other key information, such as the Inland Revenue
Department’s data relating to the Student Loans Scheme, or the Student Loans Fair Value model currently under
construction. These information tools would be used to more accurately cost policy changes of this nature.
Finally, none of the cost estimates were subjected to the full range of internal quality assurance mechanisms we would
normally use to test assumptions, such as sensitivity analysis.
I understand the Minister of Finance has also expressed his concerns to you that the release of costings such as these,
prepared in the formative stages of policy development as a way of helping Ministers shape the direction of policy, may
deter Ministers from asking Treasury for this type of exploratory information in the future, to the detriment of the
quality of policy decision-making. You will no doubt consider this in your deliberations.
I trust you find this information useful in the context of your current investigation. Please feel free to contact me
directly should you have any questions on the issues raised in this letter.
Yours sincerely,
John Whitehead
Secretary to the Treasury