The numbers The Treasury will release in its PREFU next week will make it fairly easy to follow some bits of New Zealand fiscal policy over time, less so others, but do almost nothing to facilitate international comparisons, and discourage New Zealand users and analysts from looking at fiscal policy in the way most other economic analysts and the international agencies do. It is fairly transparent, but in an insular way.

New Zealand’s government accounting framework often wins plaudits. As a matter of accounting, no doubt the plaudits are largely fair (although I’m not an accountant, and have previously suggested that some accountants rather overrate the contribution of the New Zealand model). It just isn’t a framework used widely for economic analysis or as a good basis for holding governments to account for the conduct of fiscal policy.

The focus here tends to be on the operating balance (excluding gains and losses). Last year, for example, the current government articulated its primary fiscal goal as being to (eventually) deliver and maintain an average (OBEGAL) operating surplus of between 0 and 2 per cent of GDP. It is a fairly theory-free measure

attractive mainly for being relatively more controllable year to year than a simple operating balance would be. But it is the first of the handful of “Fiscal strategy indicators” in the tables Treasury publishes. It was the measure that was forecast to turn positive (surplus) in 2025/26 in this year’s Budget (before they realised they’d forgotten to allow for the near-certain loss of lots of tobacco excise revenue), but in 2024/25 in last year’s Budget (and perhaps in 2026/27 in next week’s PREFU). It is the measure that was still in material deficit in 2022/23 and 2023/24 even at Budget time, before the deterioration in the tax revenue this (calendar) year became apparent.

But an operating balance measure (OBEGAL or total) isn’t widely used for fiscal analysis purposes elsewhere. In fact, if you turn to the two main compilations of comparable cross-country fiscal data – the IMF Fiscal Monitor and OECD Outlook – you won’t find any operating balance data at all (although you can put together such numbers yourself from components in the OECD database).

There are a couple of good reasons why that isn’t the focus.

Operating balances sound good and commercial, somewhat akin to a profit and loss account for a business. Capital expenditure typically doesn’t appear in a profit and loss account, and nor should it (as with New Zealand government accounts depreciation appears in a profit and loss account, and in the government’s operating balance). But when a business does capital expenditure it does so with the goal of generating future income (or maybe reducing future costs) but that is rarely the case with government capital investment spending. New Zealand’s total Crown OBEGAL does take account of the Crown’s ownership interests in various profitable businesses (which is not the case with the standard general government concepts used in the national accounts and in most cross-country fiscal analysis, where commercial business interests are excluded by construction) but most total Crown capex is just another form of spending. It might – like operational spending – be done for good reasons, it might have expected benefits down the track, but any such benefits are rarely fiscal in nature. And today’s capex is tomorrow’s depreciation (with no new matching revenue). In principle, the worse the project the larger and sooner the economic depreciation.

The second reason is the role of interest (financing costs). It is very common in overseas and international agency fiscal analysis to focus on primary balances, which exclude financing costs. These series are prominent in IMF and OECD tables. In New Zealand the Treasury publishes no primary balance measures at all and never seems to discuss the concept (one can construct some yourself, although probably only as reasonable approximations)

There are two good reasons for a primary balance focus. The first is that interest payments mostly reflect past fiscal choices rather than today’s (if your country ran up a high debt a decade ago, you are most probably still servicing that debt now). It is a real fiscal burden now (and thus primary balances aren’t the only relevant metric) but what matters more for the future debt trajectory is what choices are being made now re non-interest spending, and revenue. Broadly speaking, if you are running a primary surplus debt (to GDP) is likely to be heading in broadly the right direction, and if you are running a material primary deficit, then debt probably isn’t going to be heading in the right direction (it depends on your respective interest rates and GDP growth rates).

The other reason for a primary balance focus hasn’t been very relevant in advanced countries for several decades but is now: inflation. When inflation is low nominal interest rates tend to be low. When inflation is high nominal interest rates tend to be high. If interest rates are 2 per cent and inflation 2 per cent, the substantive economic implications for government finances are no different than if inflation is 5 per cent and interest rates are 5 per cent (real rates are the same). But nominal interest is what is recorded in the government accounts (and in those New Zealand government operating deficit numbers I mentioned earlier). It makes a difference: at present government bond rates and the OCR (much of the Crown’s debt is currently those RB settlement account balances) have increased a lot, but are actually both still below the rate of inflation. In economic substance terms, the operating deficit numbers in the Budget (and those to be published next week) are materially overstated.

(Inflation adjustment issues are pretty pervasive – company accounts and individual debt burdens, not just the government’s – are just another reason why we shouldn’t be tolerating a central bank doing such a poor job on core inflation that we have start thinking again about how inflation distorts both behaviour and accounts.)

Now it is true that The Treasury doesn’t just publish (total Crown) operating balance (OBEGAL) numbers. Second on that table of Fiscal Strategy Indicators is a measure called “Core Crown residual cash”. Unlike the OBEGAL it is (a) measured on a cash basis (not accruals), b) is a core Crown measure (not total Crown), and c) it treats all cash (capex and opex) alike. It is, in other words, much more akin to the series that show up in those international databases (but note that neither core Crown nor total Crown is the same as “central government” for national accounts purposes, let alone “general government” – which is vital for cross-country comparisons with countries where state/provincial and local governments make up a much bigger share of overall government activity)

When I went to work for a while at The Treasury I vividly recall one of their more astute analysts encouraging me to focus on this measure. Few in New Zealand seem do (and I only inconsistently across time). Neither political party seems to. And in case you were wondering, in this year’s Budget – recall, before the fall off in tax revenue – the core Crown residual cash deficits were expected to be 5.7 per cent of GDP in 22/23 and 6.5 per cent in 23/24.

The ups and downs of the economic cycle affect fiscal balances. Expenditure is affected to a small extent (eg unemployment benefits) but the main impact is on the revenue side (tax receipts). Consistent with that it is common to look at, and the international agencies, publish estimates of cyclically-adjusted deficits (total and primary).

The New Zealand Treasury does typically publish estimates of the cyclically-adjusted operating balance, but not of even any other measures, and notably not of any primary balance measure. One could do one’s own rough and ready adjustments to the alternative series, but how much better if Treasury were doing the adjustments themselves, using disclosed the contestable methodologies. There is also no long-term consistent time series (eg published with the annual long-term fiscal data on The Treasury website). There is no one right answer to how much allowance should be made for cyclical factors in doing cyclically-adjusted numbers – it is the same contests around the output gap and the NAIRU as monetary policy makers repeatedly face (and estimates are likely to be revised, perhaps quite a bit, over the years) – but in big booms and deep recessions it is important to try (normally we would not want deep fiscal cuts if we could be reasonably confident a deficit was mostly a reflection of a deep cyclical downturn).

Until relatively recently, The Treasury used to publish (but buried very deep in an Additional Information document) the operating balance on a terms of trade adjusted basis (a higher/lower terms of trade – variable and largely outside New Zealand control – can flatter/punish fiscal outcomes independent of government fiscal policy choices. Here is an example, from the 2017 PREFU, of how much difference Treasury thought the rising terms of trade had made in improving fiscal outcomes over the previous decade.

As the terms of trade trended upwards fairly consistent over 15 years, the adjustment perhaps became somewhat moot. In the last few years, on the other hand, the terms of trade have been falling back.

Much of this post so far has been about how we might better analyse New Zealand numbers in isolation, but international comparisons also matter. Why? Because there are no single right and wrong answers about the appropriate deficits or surpluses or debt levels, but we can often make better sense of our own data if we can compare and contrast with developments in other advanced economies. But comparing and contrasting needs to involve data compiled as consistently as possible across countries (as, say, with cross-country comparisons of GDP growth and productivity, and where comparisons of inflation are often complicated by the absence of such consistently compiled measures). For those purposes, the fiscal databases and forecasts done by the IMF and the OECD are the standard. In those databases we can find raw and cyclically-adjusted primary and overall fiscal balances, gross and net debt measures, and measures of net lending (savings less investment), all on a general government basis.

Both the IMF and the OECD produce numbers for New Zealand, including forecasts for the coming years. These are typically done starting Treasury’s fiscal numbers and adjusting for (a) different views on the state of the economy, and b) the methodological differences between the New Zealand measures and these internationally conventional ones. At times – eg the IMF Article IV report last week – they will provide some updated estimates, but generally the IMF and OECD numbers are published only twice a year.

But in the Treasury numbers accompanying the EFUs there is no attempt made to present their updated numbers in an internationally comparable format (not even just as an appendix), whether flow or stock measures. On the face of it, doing so should be a relatively mechanical exercise that would require a one-off investment to set up but almost no marginal resource cost beyond that (especially when the non central government components of general government are small and slow moving in New Zealand). And yet they choose not to do it, and as a result serious international comparisons are harder than they need to be. There will be some new IMF WEO estimates for New Zealand in October, but they are likely to emerge just a couple of days before the end of our election and won’t be able to inform discussion/debate of the fiscal situation. But Treasury could have produced their numbers in IMF and OECD formats.

This has been something of a background post with little data and just one chart. Tomorrow I will do a post looking at some of these better measures (NZ specific and internationally comparable) across time and across countries (OECD and IMF advanced economies) to put some light on how things stand right now, and have evolved in recent years. PREFU will then give us some more – apparently gloomier (although how much of that will be judged cyclical?) – data next week.