I didn’t have any intention of writing again about the Labour Party’s fiscal plans. I’d already done so, prompted by the infamous Steven Joyce “fiscal hole” here and here.
But on Monday evening I was sent some analysis of the fiscal outlook prepared by a group of former senior Treasury officials, who were keen that I should give it some coverage. They are keen to retain anonymity, but I know all those involved, and have a considerable regard for most of them. Most, in my observation, would also seem considerably more likely to vote for ACT than for, say, Labour or parties to its left. But what they sent me wasn’t particularly value-laden; it was an attempt at a technocratic assessment of some of the basic pressures on government finances over the next few years.
I’m not going to swamp you with numbers And trying to unpick and explain here the differences between core Crown and total Crown expense items, and which is relevant where, is sure to have your eyes glaze over, and (frankly) on a couple of points I couldn’t get clear answers myself.
But their main point is a simple one. Budget projections of government tax revenue, including the PREFU ones that both Labour and National are relying on, include the effects of forecast wage and price inflation, and forecasts of a rising population. On the other hand, most line items for government expenditure, as presented in the PREFU, do not do so.
But things governments purchase, and people governments employ, will become more expensive over time (that’s inflation). And a rising population will also, over time and all else equal, require more public employees – perhaps no more Treasury officials, but certainly more nurses and teachers, police and perhaps even Corrections officers. Given the outlook for inflation and population growth, those cost pressures are largely inescapable – at least without specific decisions to cut real per capita spending in some or other areas. (When I say “inescapable”, of course governments can change inflation and non-citizen immigration targets, but given the targets they choose, there are future – largely inevitable or inescapable – spending consequences.
In fact, the PREFU documents would be much more useful – as I noted a couple of weeks ago, picking up a comment on an earlier post – if they included some analytical tables showing expenditure numbers that adjusted for these all-but-inescapable inflation and population pressures. If political parties presented their alternative plans relative to that sort of baseline we would all be much better off.
Instead, at present, we are given a table like this (or its total Crown equivalent)
In this table, some specific line items show considerable increases over the four year period. In particular, where there is a statutory requirement to make expenditures in a particular way, explicit allowance is made in the PREFU line item numbers. New Zealand Superannuation expenditure for example – eligibility and formula in statute – is explicitly shown as increasing from $13043 million in 2016/17 to $16085 million in 2020/21 (included in the first line above).
But most government spending isn’t like that. It is simply subject to appropriations after each year’s budget, and each government agency must make an explicit bid for any new spending, even that which (in effect) simply results from inflation or population increases. So those – largely inescapable – prospective spending increases aren’t identified in specific line items in the table above. Instead, it is all lumped together in the line labelled “Forecast new operating spending”, which – in PREFU – captures the cumulative total of the annual “operating allowances” the current government has identified for the next few years (cumulative because the decision to spend $1000 million extra next year, and another $1000 million extra the following year means that in the second year, total spending would be $2000 million higher than what has been appropriated this year).
In other words, that new operating spending line can deceive. From a Treasury budget management perspective it is all new money available to spend – you don’t want government departments counting on what ministers and Parliament haven’t approved. From a bigger picture perspective though it is a mix of money available for genuinely new initiatives, and money that will end up having to be spent to keep up with wage and price inflation and population. Quite how much governments want to have available for genuinely new stuff is up to them – and political and market tolerance for debt and taxes. But in recent years, that proportion hasn’t been much. Government spending has been falling as a share of GDP. In fact, as I’ve noted previously, both parties tell us they expect to reduce government spending as a share of GDP further over the next few years.
Labour less so than National, but it is still a reduction.
And the other half of the point the former senior Treasury officials are making is that Labour seems to have promised to do quite a lot of new stuff over the next few years. Specifically, in 2020/21 – the final PREFU year, although not the last year of Labour’s plan – they have announced specific policy promises that, on their numbers, would have total spending $6058 million higher than under PREFU.
Of course, they are partly doing this through revenue measures – primarily not proceeding with National’s promised tax cuts. On their numbers – and the same economic assumptions – revenue will be $2341 million higher in 2020/21 than is provided for in the PREFU. Since the projected surplus in 2020/21 is almost identical to that in PREFU, they have precommitted a net additional $3700 million to fund the (net) new policy promises.
And how much as-yet-unallocated spending provision is there in PREFU for 2020/21? Well, we can see that number in the table above: $5495 million. But if around $3700 million is already committed to meet policy promises, that leaves only around $1800 million.
Again, the question the former senior Treasury officials are raising in whether that is enough to cover the inescapable cost pressures.
Over the three years (from the current Budget numbers for 2017/18) to 2020/21, those “inescapable” pressures include projected:
- Population growth of 4.6 per cent
- Cumulative increases in the CPI of 6.2 per cent, and
- Cumulative wage increases (QES measure) of 8.5 per cent.
Across core Crown agencies and Crown entities (the latter includes schools and DHBs) the government expects to spend $20142 million this year on “personnel expenses”. Using those cost pressures:
- forecast wage inflation alone would lift annual spending by around $1700 million by 2020/21, and
- if the number of public employees kept pace with population growth that alone would add another $900 million to the wage and salary bill.
- for an overall increase (the new workers will also get the higher salaries) of around $2700 million
But there was only about $1800 million left after the specific policy promises Labour has made.
Another item we could look at is Other Operating Expenses. These are around $39000 million this year (2017/18 Budget). Assume that the cost of whatever the government is purchasing increases at the rate of CPI inflation, and that the volume of purchases will increase with the forecast increase in population, and that would raise government spending on this item (simply to deliver the same volume of real services per capita) by about $4200 million by 2020/21.
Across just those two (large) items, the “inescapable” cost pressures would add $6900 million by 2020/21 to annual spending. And yet, the government’s operating allowance – passed to Treasury as current policy to use in PREFU – is for only $5495 million more spending by 2020/21.
We know the policy promises the Labour Party has made: they’ve committed about $3700 million to meet specific policy promises. We don’t have any decent estimate (that I’m aware of) of the cost of the National Party’s promises, although they seem almost certain to be less – for now anyway – than those the Labour Party has put out.
But the bottom line is that, on the macroeconomic assumptions both parties are using, the existing operating allowances (and revenue projections) are not sufficient to cover even what former senior Treasury officials would regard as “inescapable” cost pressures on the Budget over the next few years. (In the spreadsheets they sent me, they didn’t explicitly allow for the population growth, but in my subsequent exchanges with their representative, they accepted that the population pressures on spending are just as real as the inflation ones.) The gap – hole if you like – appears likely to be materially larger for Labour than for National, but it is there for both sides of the political divide.
Of course, there are ways through that. A government led by either main party could decide to make material cuts to services or other spending. Labour seems to have talked of the Defence budget and perhaps Corrections. National – probably faced with a smaller gap – has given us no clues. And none of those possible cuts – from either party – has been outlined in any detail and debated in this election campaign. Perhaps public service salaries could be held below general wage inflation. Perhaps no more teachers or nurses could be hired even as the population increased? Perhaps some extraneous (but mostly rather small) government agencies could be closed?
Then again, perhaps the spending as a share of GDP tracks just aren’t very credible, probably for either party (but probably more so for Labour). As I’ve said before, it is a little hard to understand how, for example, a party can campaign on the idea of years of underfunding of core services, and yet suggest that government spending will fall as a share of GDP over the next few years if they are elected.
And when (properly measured) net core Crown debt is about 9 per cent of GDP, and the macro outlook suggests rising surpluses from here, it isn’t entirely clear why it would be sensible to do so, given the stuff that a left-wing party says that it wants to deliver.
The future is uncertain. We could have a recession in the next few years, or a period of really strong growth. But if we take the Treasury PREFU macro outlook as given, it is hard not to conclude that under whichever party we elect there will be more total spending (than in either PREFU or the Labour plan) as a share of GDP and somewhat higher public debt. And before anyone starts hyperventilating about higher interest rates, a reminder of the Orr/Conway results I linked to on Monday: on those estimates, even a 10 percentage point increase in net debt to GDP might be worth about 6.5 basis points on long-term bond yields. Realistic differences over the next three years would not be worth anything like an additional 10 percentage points on the debt to GDP ratio.
And finally, my suggestion – indeed plea – to The Treasury, and to possible new Fiscal Council (which Labour is proposing if elected) – is that for the next PREFU we should have analytical tables that take explicit identifiable account of the likely – largely “inescapable” (in the words of the senior Treasury officials) cost pressures resulting from inflation and population growth. There are real debates to be had on what level of real per capita services/spending governments should provide – that is the stuff of politics – but for these sorts of purposes a much more sensible baseline – and a more enlightening one for voters – is one that explicitly takes account of those pressures, and thus more clearly identifies how much is “left over” for genuinely new initiatives. Anything beyond that amount has to be funded by either cuts in other spending, or lower surpluses. But inflation means things cost more, and more people means more government spending (all else equal), and it doesn’t help to bury those pressures, as current practice – led by Treasury, which decides what to publish – tends to do.
croaking cassandra 
Mathew Hooten said he thought adding costs from CPI increases to budgets would be inflationary on Twitter this morning. I said I thought the impact would be small, and easily offset by the reserve bank. What do you think?
The other thing I wonder about with the inevitable increase in costs both parties are not keen to discuss is do they plan to borrow to meet these costs?
Where Labour has been depicted as the party of “tax and spend” I sometimes wonder if National should be depicted as the party of “borrow and pretend”. Public debt has not been an issue this election. Should it be?
LikeLike
I presume Matthew had in mind some increase in total spending, beyond what Labour is already proposing, to cover inflation pressures? Of course, at the moment at a bit more inflation would be welcome – we’ve been below target for the last five years. But, in fairness, the numbers aren’t small – could easily add another 1% of GDP to spending, which probably would take the RB a bit closer to tightening at some point.
You say that public debt hasn’t been an issue this election. that is largely so, altho Steven JOyce tried to make it one. To me, we should probably have some sort of kumbaya moment where we celebrate that thru the ups and downs of the last 25 years, under govts led by both parties, we’ve had pretty responsible fiscal management and low debt. As I said in the post, net debt is about 9% of GDP at present – higher than it was when Labour left office, but there are only half a dozen OECD countries with lower debt. With a rapid rate of population growth – which i think is in itself crazy – we probably should be having a thoughtful discussion as to whether a bit more public debt would not be appropriate. Of course, the counter to that is the poor quality of much govt capex, but a rapidly growing population needs more schools, hospitals, roads etc, and it is perfectly sensible to consider borrowing to finance a fair chunk of that – as a fast-growing business might do/
As I said in the post, if the macro forecasts are roughly right i suspect we’ll see both parties spend more than they are now saying, and as a result surpluses will be lower than forecast (but still surpluses). On that sort of revised track, debt/GDP might still be trending down slowly, but if the ratio did rise – from 9% – it wouldn’t be much. It is more a risk with Labour, but I wouldn’t overstate the differences – after all, if the Nats survey this time they’ll be very concerned about the next election and will want to be seen to “do stuff”, which often costs quite a lot. Labour would be a first term govt and have put a lot of their “stuff” upfront.
LikeLike
The problem is that older people are just living longer and therefore there is increasing dependency on the health system and Universal Superannuation and therefore also not vacating existing houses. Jacinda Ardern has already drawn a line that Superannuation age will not increase beyond 65 which basically means that we actually do need population increasing basically to increase the tax pool to fund the health system, to fund the Universal Superannuation system as well as to provide care to the old. Again it is a simple mathematical equation which seem to escape our NZ economists conceptually.
LikeLike
MH said: “It seems to me very poor public policy practice to factor in #CPI changes in budget baselines. That’s how you cause inflation.”
In this thread: https://twitter.com/MatthewHootonNZ/status/910244305305866241
LikeLike
But they are two different issues. I certainly think there should be an automatic translation into formal baselines that depts have authority to spend. Doing so wouldn’t cause inflation, but it would take some pressure off depts to justify spending. My argument is simply that the “as-if” numbers should be presented at PREFU time, alongside the existing tables, to show what things would look like if all relevant spending lines were rated forward for forecast inflation and population growth. Whether they should actually get that money is of course a political decision after the relevant election.
LikeLike
It may well be that improvements in health, increases in longevity, gradual natural increases in the population make increasing but gradual demands on superannuation. Bringing in abnormal numbers of migrants together with their own families is an exponential increase. When those same migrants bring in their 4 parents the exponential rate of increase becomes steeper and sooner
LikeLike
New migrants cannot bring in their parents now under National’s many tweeks in Immigration policy. The previous ones that are here need 10 years permanent residence before being allowed access to all these health and social welfare benefits. However in most cases migrants do bring in a large amount of assets from the prior working lives, selling up existing houses to bring into NZ.
LikeLike
I’m definitely with you on the need for PREFU to incorporate CPI increases, but not so sure on the population ones (harder to predict accurately and more likely/able to be influenced by government policy – as our immigration setting become more politicised).
I was a Vote budget manager in central government during the two years that Winston Peters was Treasurer – and I can’t recall the exact details of the parameters he gave us but I think it was something like a requirement for a 1% decrease in our Vote each year (and any pay rises intended had to also be provided for within that parameter). I can’t recall whether/how CPI projections were treated. The intention being, as a Vote budget manager you had to find real cost savings in your operational budget – not big ones, but real ones. New money requested had to be accompanied with a business case and these requests were responded to as a separate matter, and then added back into your 1% reduction budget, if approved.
Having entered government from the private sector all this made perfect sense to me and I had no trouble meeting the requirement in each of those two years. (The requirement for the 1% reduction was dropped once the coalition fell apart if I recall correctly). I put in a proposal for new money in year 1 whilst Peters was Treasurer – it was refused, but the feedback given was that it was a considered a worthy initiative and I should consider a way to fund it out of the existing/agreed 1% reduction in baseline (and resubmit my budget if I chose to do that).
I guess my point is, where core (i.e., departmental head office) budgets are concerned (at that time) I found there were relatively easy savings to be made. No idea whether departmental head offices have the same amount of ‘fat’ in them these days.
It also occurs to me that it is likely time for another Quango hunt.
LikeLiked by 1 person
On the first para, the argument for including population growth isn’t getting the forecasts right, but simply treating revenue and spending symmetrically. At present, we are presented with revenue numbers that include the effect of immigration/popn change, while the expenditure consequences are buried in the operating allowance money bag.
I don’t have a strong view on what fat there is in the system – altho I suspect it must be much less than in say 2009. Equally, sometimes those sometimes arbitrary rules – zero operating allowances, requirements on agencies to save 1% can be useful – but it is always worth illustrating them against a baseline in which nothing is done – no hard choices, no generous ones – and that is the merit of starting with PREFU numbers that include estimate of the future spending implications of inflation and population growth.
Agree re Quangos – although of course each have their own vested interests, so always easy to let them slowly wither than to actually kill them.
LikeLike
Thanks, I hadn’t caught that rationale regarding symmetry, i.e., that TSY are making the assumption that revenue growth is dependent on maintaining our high volume, low skills base immigration settings – a sort of sad ‘fact of the matter’ where most of our GDP improvement is coming from.
Which leads me to a question. What if we had instead been achieving productivity (GDP per capita) gains under a very low population growth scenario – am I right to then assume that such productivity gains wouldn’t necessarily flow through to spending?
LikeLike
Re revenue, of course we would need so much revenue if the population were growing more slowly. In the short-term, immigration is probably a net fiscal positive, but we need to keep in mind constantly the effects on both sides.
Re productivity, it is hard to answer. I’ve made the point recently that real wage inflation has been running ahead of (nil) productivity growth, so arguably from the cost side expenditure pressures might not have been any greater. But higher GDP flowing from higher productivity growth would have lifted revenue, and thus probably improved the sort of baseline I’m talking about.
Of course, really fast productivity growth might well have been reflected in higher interest rates, raising (our modest) debt service costs.
LikeLike
If we get rid of the 7 Maori seats and all the other Maori specific or Maori biased programmes and all this Te Reo waste of time and money nonsense plus disband the Treaty of Waitangi Tribunal, the budgets should stack up quite nicely and probably deliver another billion in surpluses.
LikeLike
As I say, every bit of spending has its own vested interest and vocal constituency. I’d get rid of the Ministry for Women, the Retirement Commission (or whatever it is now called), and raise the NZS age, but I don’t fancy my chances……
LikeLike
1) So tax on increased wages of government employees is included as income, but the increased wages of those same government employees aren’t included in expenditure.
2) Name and shame the person/people responsible for asymmetric treatment of income & expenditure. Is it Treasury, or a government that put this into law?
3) For simplicity, although not as accurate, I could argue that inflation, population and demographic changes should be excluded entirely. But then perhaps some of the benefit of the PREFU process is lost.
LikeLike
On 1), yes I guess that is right, at least at the line item level. The provision for the increased wages is, in effect, in the operating allowance line, but we get no sense from the published number of how much is needed for, eg, those wage increases and how much is available for genuinely new programmes/services.
on 2), what Tsy do is a reflection of the way budget management works (ie there is an operating allowance, and it is divvied up at each budget between, in effect, inescapable cost pressures and new stuff), but equally they put out a lot of entirely discretionary material in the PREFU documents, and could easily include tables along the lines I’m suggesting – illustrative scenarios showing what-ifs if inflation and population pressures, as forecast, translated into spending on various functions over time. It isn’t a specific elected government’s fault that such info is not provided.
3) I wouldn’t favour that option. After all, we have an official inflation target, and inflation does create fiscal drag (an overall net gain to the Crown). And we do have population growth, and need to think about what the means for future spending neeeds (and revenue).
LikeLike
Hi Michael,
Thank you for an enlightening post. A couple of questions (apologies if already asked answered):
1. Am I right in understanding that the prefu operating allowances are essentially set by the government of the day, and are not the result of specific bottom up analysis? Presumably there must be some behind the scenes level headed analysis, but that is kept from public view? Or is it reasonable to assume it is quite politically driven given your ex treasury people think it’s not enough.
2. Your analysis above assumes (fairly, given the way Labour have framed it) that all of the $6b is ‘new’ spending, and any inflation or popultion adjustment to existing spending is separate. Alot of the $6b clearly is genuinely new spending, however there are some big chunks in Health and Ed (possibly some of the others) which aren’t really specific new policies and so might reasonably be able to be assumed to be this inflationary adjustment?
LikeLike
On 1, yes the operating allowance is govt policy, and Treausry then reports the fiscal outlook based on all aspects of govt policy communicated to them. Often there isn’t any detailed bottom-up analysis for the out-years, and the operating allowance has served as a macro top-down way of signalling/expressing overall intentions. In many ways, the out year numbers aren’t very meaningful anyway – inflation can be higher or lower than forecast, and population growth is almost completely unforecastable (esp over 1-2 year horizons) given the huge variability in net migration (esp the NZ citizen component). That is partly why I don’t get over-excited about these issues, whether they relate to Labour or (less seriously) National. Budgets are made each year, the operating allowance itself has no macroeconomic significance (absolute numbers mean very different things if inflation is 1% than if it is 2% and so on), and even prudent debt levels depends partly on population growth rates. If we were close to some fiscal cliff, small changes might matter more. But we aren’t – and haven’t been under either govt for decades.
On 2, yes fair comment. there may implicitly be some cost/popn factors implicit in the Labour numbers. I suspect they have taken account of what they see as past shortfalls on that score (their number more or less says that) ,and perhaps for their new initiatives’ costing they have allowed for population growth and inflation from here. But nothing really suggests they have allowed for inflation and population pressures from here across the broad base of (non-welfare) govt spending. That, when all boiled down, is the argument of the ex Tsy officials. I complemented that by pointing out that, on the macro foreasts, even the govt’s operating allowances wouldn’t cover inescapable cost pressures, and – quite clearly – they want to do new stuff too (eg the now more-or-less promised “families package” in 2020)
LikeLike
Thanks. I finally feel like I understand this issue!
LikeLike