Forecasting Deficits Ontario Style

One of my most piercing electoral memories is the 1993 federal election debate that featured then Prime Minister Kim Campbell, Jean Chretien, Audrey McLaughlin, Preston Manning and Lucien Bouchard.  The part that always sticks in my mind is Lucien Bouchard’s persistent questioning of Campbell with respect to the size of the deficit in the 1993 budget.  Campbell seemingly dodges the question as Bouchard intones:  “What has been the real deficit?”  “Its a simple question, yes or no”

Of course, Campbell politically either did not want to pin herself down with an answer or was cerebral enough to realize that there is no simple answer to what is a complex question and indeed a complex number.  By the early 1990s, it was a common occurrence to have federal deficit projections that quickly expired as the numbers for the actual deficit came in and missed by a large margin.  Accurate forecasting of the government finances appears to have improved but it is still an issue as an assessment in the Globe and Mail last fall in the wake of the 2009 recession compared the accuracy of the Federal Finance Department with the Parliamentary Budget Office. In a comparison of 15 reports from 2008 between the PBO and Finance, the difference between the average projected surplus or deficit and the final number was 12.6 billion for Finance and 13.2 billion for PBO.  Personally I thought given the greater resources available to Finance to make its projections compared to the PBO, the difference was minor.

While “Forecast that Deficit” is unlikely to make it as the next television reality series, it is nonetheless an interesting issue not only at the federal level but also at the provincial level – and particularly Ontario.  Ontario recently announced that its final deficit figure for 2011-12 as reported in the public accounts was 13 billion dollars – 3.3 billion dollars less than the estimate in last year’s budget.  While the spin on this could be that Ontario was doing better than expected in the war against the deficit, it could also lead to the question of how good is the Ontario government at forecasting its revenues and expenditures and ultimately its deficit?

The accompanying table presents a range of deficit estimates and forecasts starting from the 2009 Ontario budget and up to the 2012 Budget along with the 2009 update in the immediate wake of the worsening recession and the most recent update in this summer.  At the dawn of the recession, in the 2009 budget, Ontario was forecasting a budget deficit of 14.1 billion dollars, which was quickly upgraded to an estimate of 24.7 billion and then settled in finally at 19.3 billion by the 2011 budget.  The value of the 2009-10 deficit figures between 2009 and 2012 ranged from a high of 24.7 billion to a low of 14.1 billion with both off from the final figure by 5.4 and 5.2 billion dollars respectively.  It does remind me a bit about the old joke about the three econometricians going out hunting in which one shoots left of the target, the other shoots right and the third says “We got it! We got it!”

If we look specifically at the deficit numbers for 2011-12, in the 2009 budget, the deficit was estimated two years down the road at 9.7 billion dollars.  By the 2009 update it had jumped to an estimate of 19.4 billion.  In the 2010 budget, it was down to 17.3 billion dollars while the 2011 budget had it down to16.3 billion dollars.  The number continues to decline until the release of the public accounts.  It is interesting that the 2012-13 first quarter update released in June puts the deficit at 15 billion while the public accounts released three months later then has it at 13 billion.  For 2011-12, the deficit numbers between 2009 and the present ranged from a low of 9.7 billion dollars to a high of 19.4 billion settling in at 13 billion with the max off by 6.4 billion dollars and the minimum off by 3.3 billion dollars.

The 2010 to 2012 budgets also all had forecasts of where the Ontario deficit would be until 2017-18 – the target date set for its elimination.  Note that all three budgets have practically identical forecasts for the deficit for the 2013-14 fiscal year and forward. It does seem somewhat amusing that the deficit estimates become more stable as you project into the future.  Given how much the numbers change from year to year, why have these forecasts not varied? 

Forecasting a deficit is a tough job especially when the economy is undergoing a recession - and Ontario was hit very hard - but this range of deficit figures and their iteration to a much lower final numbers than originally forecast suggests that Ontario’s deficit could actually be eliminated much sooner than you think.  I would not be surprised to see it almost gone in the 2015-16 fiscal year – two years ahead of schedule.

Comments

[Neil]

"Especially in Ontario" is a pretty Ontario centric view of the universe. I don't know about other provinces, but Alberta is routinely off its deficit projections by billions of dollars, on a substantially smaller budget. Last year's projected $3.4B deficit turned out to be (wait for it) $23m. That's like a nickel in government spending terms.

Of course, anyone who's been involved in budgeting in any organization - public, private, non-profit...doesn't matter - knows that any and all forecasts rely heavily on current trends persisting. But that can only take you so far...there's just too many variables that affect the future to be able to accurately predict anything more than a month out. Even the weathermen - who can get away with being wrong most of the time - won't predict anything past 14 days...and the 14 day forecast is usually weirdly consistent with seasonal norms.

When it comes to budget forecasts, long term numbers are more stable because the long term trends are more stable. Shocks tend to even themselves out over time, so it's really only on the one-year time span that we learn anything that materially changes our expectations for the future.

[Jacques René Giguère]

In 1981,as a young staffer at the Québec National Assembly, in a long gone youth in a past century, I was asked to give a workshop to newly-elected MNAs. It was the early 80's high inflation and recession, a confusing time.
I tried to explain various types of deficits ( budgetary or national acount basis, with or without transfer payments, with or without interest payments primary, secondary, conjonctural or fulll-employment) with or without capital budget, . The audience was not of idiots: MBAs ( some from Harvard), a PhD in international corporate law...) One even becaame Prime Ministerof Québec..
At the end, one of the guy(Harvard MBA, former very-high level staffer in Trdeau's gov't) rose and while leaving, said :"Well Jacques, a deficit is a deficit isn't it".
In that environment,I grew up fast.

[JR Hulls]

How does Canada value it's infrastructure and depreciation and maintenance costs? I was stunned to find out that the US does not look at this as a line item in Federal budgets, which given all our problems with roads, water and sewer systems etc. ( I realize some of these are local) looks like a huge book keeping error. http://www.gpo.gov/fdsys/pkg/GAOREPORTS-AIMD-95-34/pdf/GAOREPORTS-AIMD-95-34.pdf

From the preface: "We found widespread agreement among accounting and budgeting experts that the federal government generally does not depreciate transportation infrastructure, research and development (R&D), and human capital for either accounting or budgeting purposes. In our research we found no evidence that states or private sector businesses use depreciation in budgeting for any of these types of investments. However, economists depreciate infrastructure and R&D investments in their more global analyses to make rough estimates of national wealth."

We have to turn to other sources such as the American Society of Civil Engineers to see the enormity of this problem. Please see http://somewhatlogically.com/?p=489 for discussion and links to the ASCE report card, mostly 'd's resulting in multi-billion dollar deficit costs that don't seem to be taken into budget accounting.

[JR Hulls]

And a deficit isn't a deficit when a bridge collapses, it's a disaster with considerable human and economic costs.

[Livio Di Matteo]

Good question JR. I suspect the Canadian accounting practice may be the same. I do not recall seeing a depreciation item on government infrastructure in either the federal or provincial budgets.

[Bob Smith]

I can't say I've ever seen a depreciation figure in a budgt.

I can't be bothered to look for them, but the federation of Canadian Municipalities (and others) are forever publishing estimates of the "infrastructure deficit", understood as the cost of replacing existing infrastructure that is now at, or beyond, its service life. Depending on who you ask that number ranges from anywhere from low 100 billion to 200+ billion (which reflects the reality that capital spending hasn't reflected depreciation over the past 30 years). Mind you, depreciation is only one of the MANY expenses/liabilities that goverment's don't reflect on their books (unfunded liabilities for health care, pensions?). In any other sector that sort of record-keeping would earn you a hefty jail term.

[whitfit]

I think that governments book expenditures all in the current period. IE, build a bridge in 2012 and count it all as an expense in 2012. There is no accounting for depreciation because it doesn't count its capital assets, and treats its capital expenditures as expenses in the current year. The only longer term balance sheet item I see is the debt.

[Bob Smith]

I think that's right.

[JR Hulls]

Then it is indeed a very large hole in the necessary budgets, and as the costs of deferred maintenance are huge, as Bob Smith pointed out earlier, you go to jail if you do this sort of thing in the private sector as you are very badly overstating your assets if they are not maintained. Maybe we need an accounting reform first, then a sensible budget policy.

[Jacques René Giguère]

accounting reform: without basic understanding, it would be useless.
A few years ago, I was on my college Board. A local business man, also a member, once asked the Dean why the college was not making a profit.

[Determinant]

JR:

Government budgets at the provincial and federal level are one a cash basis, not an accrual basis because governments have the power to tax. Unlike private businesses, there is little risk of the customer walking away. If you need more money, you raise it through taxes or borrowing. Getting the money isn't such a problem.

That brings me to my next point. Governments, at least in Canada, budget backwards. First they produce the Main Estimates, which is the list of authorized expenditures for the whole government, itemized by Department. When these are passed, a Department may draw on the Consolidated Fund for the money. If Parliament isn't sitting and hasn't passed the budget yet (for instance the government lost a budget vote, it has happened) the Governor General may issue a warrant against the Estimates (again, it happens).

Unlike the private sector, governments first decide on what to spend and then decide how to pay for it. This is helped because most of the time the interest on government debt is below the combined rate of growth and inflation, and thus not a burden.

I am in the process of seeking employment in the Public Service of Canada and the Estimates are a good way to learn what happened to Departments and Programs I have applied for.

Headline "Public Service Cuts!!!" My reaction: Meh. Not in the programs I applied for.

[JR Hulls]

I'm not sure that I buy the argument that having one's head in the sand because one has the power to tax is an excuse for a misleading evaluation of value and costs. The fallacy comes in when, as Defiant says, the government effectively says, I've purchased a new bridge, and I'm not even going to think mandating the costs of maintaining it and replacing it when it reaches the end of it's life because I can just raise taxes. These are real costs, and while a certain amount of the budget is allocated to maintenance, there is no transparent link to the aforementioned factors, and certainly no direct link to capital expenditures. One can see this in the budget process for public schools in many of the U.S. states, where they are restrained in their ability to set aside and hold monies for replacement of school buildings and end up holding substandard/unsafe buildings together with chewing gum and bailing wire, long after their economic service life has expired. With no accounting linkage an open invitation to deferring maintenance costs to 'balance' budgets, which in the case of buildings, bridges and highways etc. dramatically reduces the service life, thus requiring more tax monies for the earlier replacement. There are many other real economic losses from failure to maintain infrastructure, in the case of highways, lost travel time in traffic jams, increased vehicle maintenance, as documented by the American Society of Civil Engineers as referenced in my earlier response.

And I wish Defiant good luck with his job hunt, and hopefully he will be proved wrong about the public being totally cynical. Maybe they are annoyed at being mislead as to the true costs of running the country, but transparency in the costs of maintaining infrastructure to the public benefit might make the politicians much more careful about capital expenditures.

[Bob Smith]

Yeah, I'm with JR Hulls, having the power the tax shouldn't affect how governments record their costs. THe purpose of accurate accounting is to ensure that stakeholders (be they shareholders in a company, or voters in a polity) have an accurate understanding of what is going on. Moreover, accurate accounting is different from deciding how much governments spend (or businesses) spend. It doesn't matter how governments (or businesses) decide what to spend, since the accouting is only intended to record what their actual expenses.

Now, where there is arguably a distinction between governments and businesses is, in theory, governments are so large and have so many "capital" projects on the go, that their current expenditures might actually accurately reflect depeciation. Think about a government that has a portfolio of 1000 roads, each with an expected life expectancy of 50 years. If the government is rebuiding 20 roads a year, and booking that expense as a current expense (whereas, typically, in the private sector something like that would be a capital expenditure - which is why you need depreciation in your income statement to get your income calculation right), the budget will reflect economic reality. Indeed, in that context, arguably booking road repairs/replacements as current expenses is a fair characterization. But, of course, if the government decides to only rebuild 19 (or 18, or 17) roads for a few years, then its current expenses don't reflect its underlying costs.

[Simon van Norden]

As it happens, there are good economists who systematically study budget forecasts, their precision and accuracy, and the systematic influence various variables have on such things. (You can find one such summary at http://www.philadelphiafed.org/research-and-data/publications/working-papers/2011/wp11-25.pdf)

Just think what we might learn about how Ontario compares if someone were to read some of this literature....

[JR Hulls]

Thanks for the interesting link, Simon. I thought the conclusions were quite telling in light of the US situation, and seem to indicate that accurate real time data would greatly underscore budgeting for depreciation and 'savings' for eventual replacement at the end of infrastructure economic service life to accurately reflect revenue requirements.

From the conclusions: "As regards the first set of papers, it emerges that revisions in fiscal data tend to be large and can often be predicted based on ex-ante information; i.e., revisions seem to ‘reduce noise’ rather than ‘add news.’ In addition, initial releases by the national statistical authorities are biased estimates of the final values. Most papers in the second group show that strong fiscal rules (for example, establishing tight expenditure ceilings) and institutions (such as medium- term budgetary frameworks) tend to lead to relatively accurate releases of fiscal data and small deviations of fiscal outcomes from governments’ plans. As regards the third set of papers, it emerges that fiscal plans reported at the time of budgeting are now more frequently used to capture the ex-ante stance of fiscal policy. In this context, the ex-ante reaction of fiscal policies to the economic cycle is estimated to be more ‘counter-cyclical’ when real-time data are used instead of ex-post data."

If I am correctly interpreting the paper, long term budgeting for infrastructure costs would tend to be a stabilizing effect. However, getting governments to show that they have failed to see a huge hole in the road to realistic budgets, infrastructure replacement probably being one of many such potholes, would take some doing.

[whitfit]

The other thing is that this problem with government budgets provide one of the biggest incentives for P3 projects. The government can usually borrow much cheaper than the service provider, but because they have to book the cost of the infrastructure as an expense in the current period, they are more likely to use a third party to finance and provide the infrastructure. The government pays for the infrastructure over time, but pay more than they would pay if the government just borrowed the money and paid it off over the same period. This provides an incentive for the current government to build things (the public likes new shiny buildings and roads!) and not account for it in the current budgets.

Now, there may be several very good reasons for using the private sector to provide infrastructure and services - but exploiting a difference in accounting between the public and private sectors (accounting arbitrage?) should not be one of them.

[JR Hulls]

Determinant, sorry to get your name wrong in my response to your post, but you do sound a little Defiant.

[Livio Di Matteo]

Thanks for the link Simon. Certainly an excellent survey of international literature. Studies of the reliability of forecasts have been done for Canada as a whole though it has been some time since I've looked at one - there was one in the CJE in the mid 1990s - by Campbell and Ghysels. I think the biggest obstacle to doing the kinds of comparisons you might like to see is the absence of a comprehensive comparative data set for Canada that contains budget data for the provinces and the revisions to their budget projections over a period of time. The authors of the working paper you mentioned state a similar issue with respect to international data in their conclusion: "Looking ahead, the publication and maintenance by central banks and international institutions of datasets including key real-time scale indicators are fundamental preconditions for the future development of this strand of empirical research." You did work on the reliability of output gap estimates for Canada - have you ever had a chance to look at the data by province?

[Determinant]

JR:

Well, I have never been accused of not having a personality! ;)

Canada is different with respect to the US in that provinces do not have constitutionally-entrenched borrowing limits, balanced-budget requirements or referendum requirements for bond issues. They also have access to a large slice of the income tax pool, plus transfers from the Government of Canada. The use of special entities to hide borrowing, which features prominently in the US, such as in New York, does not happen in Canada on anywhere near the level is does in the US.

In fact there was a study done by the federal Department of Finance a few years ago that certain borrowers like the Canada Mortgage & Housing Corporation had interest rates a few basis points higher than the Government of Canada, yet their debts were explicitly obligations of the Government of Canada anyway. The Government terminated the issuance of debt directly by some of these borrowers and had them get funds directly from the Government of Canada under general Government of Canada borrowing.

Governments in Canada have no problem booking expenses in full at the time incurred and that is exactly what they do.

As for unfunded pension liabilities for the Government of Canada, that is a canard put forth by the CD Howe Institute. Some of their rhetoric is almost libellous. There is no liquidation value for the Public Service Pension Plan, the Government of Canada can't go bankrupt. It only has a going-concern cost which is in fact well under control.

[Bob Smith]

"Governments in Canada have no problem BOOKING EXPENSES in full at the time incurred and that is exactly what they do."

That's the problem, a lot of the government's costs are not current cash expenses. Depreciation is a great example. It's a real cost. Not a cash expense, true, but a cost nevertheless. That's why businesses have to reflect it in their financial statements. A government that purports to be running a balanced budget while not replacing/maintaining its infrastructure is misleading its voters in the same way that a company reporting its income without depreciation is misleading its shareholders (the latter, of course, is illegal, the former is par for the course).

[Sean]

Ontario does accrual accounting, having switched from cash accounting some time ago. the province introduced depriciation for moveable / minor tangible capital assets as well (cars, planes, boats, computer software and hardware, etc) within the past five years.

[Kathleen]

The Federal government adopted accrual accounting several years ago under the Financial Information Strategy. I think it was roughly 2002.

[JR Hulls]

Kathleen, not being contentious, but can you point this debate as to how they determine the value of existing infrastructure and how they account for service life costs? Thanks.

[Sean]

Ontario falls under the rules of the Public Sector Accounting Board (PSAB) of the Canadian Institute of Chartered Accountants (CICA). i would imagine the Federal Government does as well but i do not know that for a fact. Presumeably PSAB/CICA have issued policies on the correct accounting treatment for major and minor tangible capital assets but how that is operationalized across the government is challenging given the scope of assets purchased and consumed each year. I am not sure that these policies have been applied retroactively to the province's major capital infrastructure (roads, bridges, etc) given the difficulty of trying to re-state the value of legacy assets and the fact that they were fully expensed on a cash basis when they were built.

[JR Hulls]

I'm neither economist or accountant, and got dragged kicking and screaming into economic matters with a project at Dominican University in California to develop an 'economic flight simulator' to simplify looking at effects of resource utilization and ecological costs. That said, it would seem to be a relatively simple, albeit time consuming project to come up with a reasonable way of realistically budgeting for infrastructure costs. For example, there is a great amount of data within the US Fed Highway System on bridges and inspections, albeit a little shaky at the edges, as seen in this FHWA presentation.

http://msnbcmedia.msn.com/i/msnbc/Sections/NEWS/PDFs/080801_bridge_concern.pdf

If one were to take the remaining service life of the bridges (and similar federal data exists for water supplies, sewers, etc) and budget costs on that basis, at least we would start to get some transparency as to these multi-billion dollar unavoidable costs. As Simon van Norden pointed out earlier with his link, the budget process would become more 'real time' and accurate. Looking at it from a logical point of view, and to perhaps be a little controversial, what is the difference between failure to properly fund pension liabilities and failure to budget for infrastructure costs?

I strongly recommend that those interested look at the ASCE economic studies here: http://www.asce.org/economicstudy/ which show multi billion dollar costs and resultant GDP impact from the failure to maintain infrastructure. If anyone has contact with PSAB, I'd love to hear what they say about the fact that these costs are nearly invisible to the average government representative and the voting public.
(except when they hit a pothole)