An analysis of Canadian defence capital spending

Bill Robinson Conventional Weapons, Defence & Human Security

Bill Robinson

The Ploughshares Monitor Summer 2012 Volume 33 Issue 2

Bill Robinson writes on Canadian defence policy and arms control issues. He was on the staff of Project Ploughshares from 1986 to 2001.

The Canada First Defence Strategy (CFDS) (DND 2008a) projected a total of $490-billion in Department of National Defence (DND) spending over the 20 years from FY 2008–09 to 2027–28 inclusive. This total is in nominal dollars (i.e., not adjusted for inflation) and does not include the incremental costs of major operations such as the Afghanistan mission. The CFDS pledged that additional funding would be supplied to cover these incremental costs.

The 2010 federal budget modified the CFDS plan, reducing DND’s baseline spending by about $15.5-billion over the 20-year CFDS planning period (i.e., to about $475-billion).1

The CFDS projected that “Equipment” and “Infrastructure” spending (i.e., capital spending) will total $100-billion during the 20-year period. This $100-billion represents capital spending as recorded under “accrual accounting” methods (discussed further below).

Personnel” ($250-billion) and “Readiness” ($140-billion) account for the rest of the $490-billion in originally projected spending.
It is not known how the reductions announced in Budget 2010 will be reflected in these categories. DND’s plan to reduce its civilian workforce may account for some of the cuts (Pugliese 2011). Given the difficulty of reducing personnel costs, however, it is the Readiness category, which includes spare parts, maintenance, and training, that is likely to absorb the lion’s share of the reductions. The full $15.5-billion would represent more than 10 per cent of that budget, however, so it may not be possible to accommodate the entire planned cut from this category.

Reductions in the capital spending budget are also possible, but any significant cut in that category would have a severe impact on DND’s capital acquisitions plans, which (as noted below) are already of questionable affordability. The analysis that follows is based on the assumption that the CFDS capital spending budget will remain unaffected by the Budget 2010 reductions. This assumption may be unrealistic.

CFDS capital spending projections
The CFDS divides DND’s projected capital spending into four subcategories: “Previous Announcements” (ongoing major equipment programs that predate the CFDS, such as the C-17 and C-130J projects), totalling $15-billion in spending; “New Major Fleet Replacements” (the flagship CFDS capital programs, including the New Generation Fighter and Canadian Surface Combatant projects), totalling $20-billion; “Other Capital” projects (all other equipment purchases during the 20-year CFDS horizon), totalling $25-billion; and “Infrastructure” projects (land, buildings, and works), totalling $40-billion.

Cash-based vs. accrual accounting 
To assess whether these budget numbers are sufficient to cover DND’s acquisition plans, it is necessary first to examine the difference between cash-based accounting and accrual accounting.

The Government of Canada is currently using two different systems of accounting to keep track of aspects of government spending. The government uses the accrual method of accounting to prepare its budget and present its current financial statements, but it uses the expenditure, or cash-based, method to track the spending authorities voted by Parliament (as recorded in the Estimates and Report on Plans and Priorities documents).

For the purposes of this analysis, the key difference between the two systems lies in the method and timing by which expenditures are recorded. Accrual-based accounting systems record as expenses the cost of the goods (or services) consumed during a given time period, while cash-based systems record the cost of the goods (or services) acquired during that period. Thus, rather than recording the purchase cost of an aircraft, for example, as an expenditure in a single year or over a couple of years, the accrual accounting system uses a depreciation schedule based on the planned life of the asset to record the value of the aircraft that is “consumed” during each year. Over the long run, the two systems show the same total expenditure on the asset, but the accrual system records that expenditure over a much longer period. (Note that this does not change the timing of actual payments made by the government, just the way those expenditures appear in the government’s books.)

The Auditor General of Canada strongly favours the adoption of full accrual accounting for the Government of Canada, arguing that the accrual system provides a more accurate picture of the state of government assets and thus provides an improved basis for financial planning. The government has begun to take steps in this direction. DND, for example, has reported some of its capital projects in both cash and accrual terms in some recent Reports on Plans and Priorities.2 However, full accrual budgeting is currently in place at DND only for infrastructure projects (DND 2010, p. 51), and even in that case it is not evident that accrual numbers are reported in the Reports on Plans and Priorities.

In October 2010 the Auditor General (p. 2.36) reported that the government intends to “complete an evaluation of the costs and benefits of accrual appropriations in the 2012–13 fiscal year.” She added, however, with some evident impatience, that it “has yet to commit to an implementation date for adopting accrual appropriations or to explain why it would not be prudent to do so.”

Thus, despite the fact that the spending numbers provided in the CFDS are accrual numbers, for the time being DND’s Reports on Plans and Priorities documents remain almost entirely cash-based.

Are the dollar figures projected in the CFDS sufficient to cover the named projects?
The CFDS projects $15-billion in spending on “Previous Announcements” (PA), $20-billion on “New Major Fleet Replacements” (NMFR), $25-billion in “Other Capital” (OC), and $40-billion in “Infrastructure” (I) over the 20-year CFDS planning period. The total cost of the NMFR projects over their entire service lives is projected to be $45–50-billion, but only $20-billion of that total is projected to be expended during the 20-year CFDS period (DND 2008a, p. 12).

The cost estimates that are publicly available are insufficient to assess the full probable cost of DND’s spending plans, but those cost estimates that are available, which cover the major PA and NMFR projects, do suggest that the CFDS totals could be sufficient, assuming that those estimates are accurate.

Two tables list publicly available estimates of the capital costs (excluding in-service-support costs) of the major PA and NMFR programs.

Most but not all of the cost of the projects in Table 1 would be expended in accrual terms during the 20-year CFDS period. Thus, the $15-billion in accrual-based expenditures estimated in the CFDS should be sufficient, assuming no major changes in costs. The $45–50-billion estimated in the CFDS for Table 2 projects covers their entire lifetimes and thus their full capital costs (it is estimated that $20-billion will be spent in accrual terms during the CFDS period).

These figures suggest that the spending numbers provided by the CFDS may be sufficient for the capital projects announced in it.3 But a couple of caveats are in order.

First, there is the question of inflation. DND’s spending numbers are provided in nominal dollars, but the CFDS acknowledges that inflation will occur during the CFDS planning period: an estimated average rate of 2.1 per cent per year. Such inflation would have a significant effect on the cost of projects purchased in the later years of the period, unless (as is unlikely) estimated inflation is already built into the cost estimates currently available in the public domain.

Assuming the $26-billion estimated for the Canadian Surface Combatant project, for example, is what it would cost to build the ships today, their actual cost would be closer to $34-billion (nominal dollars) at the prices that the CFDS projects for the 2020–2030 period, when most of the ships are likely to be procured. This factor would be less important in accrual accounting terms than in cash terms, however, since only a small proportion of the total cost of projects begun late in the CFDS planning period would be deemed to be expended, and thus be counted, during that period. It would mean, however, that the total cost of NMFR projects ($45–50-billion) is almost certainly significantly underestimated.

Second, there is the possibility of cost overruns unrelated to inflation. If the government’s current cost estimates turn out to be significantly lower than the eventual actual costs of its procurement projects (i.e., the eventual cost is higher even after allowing for inflation), then the capital spending projected in the CFDS will certainly be insufficient and either the budget will have to be increased or procurement projects will have to be delayed or cancelled.

Just two large projects, the Next Generation Fighter and the Canadian Surface Combatant, account for 70–80 per cent of the total spending projected for New Major Fleet Replacement projects. Cost overruns in either or both of these two projects would thus pose an especially large risk to the CFDS capital spending plan.

Given the regularity with which large military procurement projects tend to cost significantly more than originally projected, the affordability of the CFDS capital acquisitions plans is questionable at best.

Third, there are the costs to sustain the equipment procured under these projects. These costs fall mainly into the personnel and readiness spending categories and so mostly do not directly affect the capital spending numbers discussed here.4 But sustainment costs can involve huge sums of money (often two or more times the capital cost of the procurement), and it is essential that the full life-cycle costs of equipment projects be taken into account when procurement decisions are made. As the controversy surrounding the cost of the proposed F-35 purchase has amply demonstrated, there is a significant risk that these costs may be substantially underestimated in DND’s spending plans. As a result, it is likely that the CFDS’s projected personnel and readiness spending numbers will also be insufficient. In the past, such shortfalls have often been met in part by reducing capital spending; similar responses in future years would place the affordability of the CFDS capital acquisition plans in even greater doubt.

Where is the amortization of existing assets?
Although the spending numbers provided by the CFDS are possibly sufficient for the capital projects announced in it, they may not cover all of the capital accrual expenditures that will have to be amortized over the 20-year CFDS period. As of 2008, DND’s existing “tangible capital assets” had a net book value of $28-billion (Auditor General of Canada 2008, pp. 10.4-10.5). Proper accrual accounting would show the depreciation of these existing assets (probably at close to zero value by 2028) in addition to the depreciation of those assets acquired during the 20-year CFDS planning horizon. Thus, there is about $25-billion in depreciation—more than $1-billion per year—that is not obviously accounted for in the CFDS numbers.

If this expenditure were added to the $100-billion in accrual-based capital expenditure cited in the CFDS, we would get a total accrual expenditure during the 20-year period of approximately $125-billion.

One possibility is that the missing $25-billion is in fact included in the “Other Capital” and “Infrastructure” categories. If “Other Capital” covered a large proportion of the depreciation of DND’s prior tangible capital assets, there would be little additional money available under that category for new equipment purchases during the 20-year CFDS planning period. The plan would thus call for DND to receive very little new equipment of any kind other than the projects listed in Tables I and 2. This is clearly not the case.

On the other hand, it is not obvious that a large proportion of the missing $25-billion could be included in the $40-billion allotted to the “Infrastructure” category. Recording the depreciation of capital equipment such as CF-18s, Halifax-class frigates, and light armoured vehicles under this category would be both highly misleading and inappropriate, but I have not found sufficient information to rule out this possibility entirely.

Cash-based estimates
Another way to examine DND’s capital plans is to consider cash-based spending.

The CFDS does not provide cash-based estimates of its spending plans, so it is not possible to assess the adequacy of planned spending in cash terms. But it is possible to make at least crude estimates of the amount of cash needed to pay for the acquisition plans laid out in the CFDS.

Some of the cash expenditure on projects in the “Previous Announcements” category was made prior to 2008. Most, however, took place in 2008 or later. It seems fair, therefore, to estimate cash spending in the PA category of at least $10-billion during the CFDS period.

Cash spending on projects in the “New Major Fleet Replacements” category is estimated by the CFDS to be $45–50-billion.

Cash spending on the “Other Capital” and “Infrastructure” projects is more difficult to estimate. Cash expenditures are likely to be considerably higher than accrual expenditures during the CFDS period, since virtually all of the cash expenditures will occur during the CFDS period, while a considerable proportion of the amortization (and thus the accrual expenditures) of the purchased assets will occur outside the CFDS period. This difference can be clearly seen in the NMFR category, in which the cash cost of the planned purchases is estimated to be $45–50-billion, while the accrual expenditure during the CFDS period is estimated to be only $20-billion, a mere 40–44 per cent of the cash cost.

The difference between cash and accrual expenditures may be especially stark in the NMFR category, given the relatively long amortization period assigned to large ships and aircraft and the comparatively late procurement of the very large Canadian Surface Combatant project. But it is probably safe to estimate that cash expenditures in the OC and I categories will exceed accrual expenditures by 1.5 to 2 times during the CFDS period. This suggests that cash expenditures on OC/I will be in the $60–80-billion range (or as high as the $100–130-billion range if depreciation of existing assets is not included in the CFDS numbers).

DND’s total cash requirement for capital spending during the CFDS period is likely, therefore, to be in the $115–140-billion range, which is $15–40-billion more than a casual reader of the CFDS document might expect. This larger sum won’t be shown in the Estimates documents if the switch to accrual-based accounting eventually takes place. But whether or not that change is made, the money will still be spent, Canadian taxpayers will still be footing the bill, and the expenditure of the cash will still appear (albeit less prominently) in the government’s books.

Assuming that the ratio of capital spending to overall spending will remain roughly constant during the CFDS period, these figures suggest that DND should currently be allocating about $5.0–6.0-billion per year to capital spending. The 2010–11 Report on Plans and Priorities projects $5.4-billion in capital spending in 2010–11. These figures suggest that DND’s planned capital spending may currently fall within the required range, assuming its cost estimates are correct, but only if its actual level of capital spending does not fall substantially below the planned level (in recent years, the actual level of capital spending has often fallen far short  of the planned level).

Adding the capital spending numbers to the Personnel ($250-billion) and Readiness ($140-billion) spending numbers pledged in the CFDS produces a total spending figure in cash terms in the $505–530-billion range. This total is for the baseline budget only and does not include the incremental costs of major operations. To reach this total, DND would currently have to be spending about $21.8–22.8-billion a year, excluding incremental operations costs, and its budget would have to grow to about $32–33-billion by 2027–28. DND’s 2010–11 budget, excluding incremental operations costs, is about $20.6-billion.

Apparently DND is not currently on track to spend the amount of money that the CFDS seems to require. Since the government has not published a cash-based CFDS spending plan, it is not evident how the Budget 2010 decision to remove $15.5-billion from planned spending during the CFDS period will affect the amount DND actually spends in cash terms on equipment projects. But it would seem on its face to make full acquisition of the CFDS wish list even less likely. The review of the CFDS that DND reportedly will shortly undertake may indicate that DND itself does not believe that its procurement plans are feasible within projected spending totals (Pugliese 2011b).

Next steps
The capital spending plans outlined in the CFDS might be affordable within the spending levels specified by the CFDS, but only if a number of unlikely assumptions were to prove correct.

First, the publicly available cost estimates for major procurement programs would have to take sufficient account of the effects of future inflation. Second, there would have to be no significant cost overruns in those programs. Third, the sustainment costs of the new equipment would also have to be sufficiently accounted for in DND’s spending projections, and those estimated costs would have to account adequately for inflation and future cost overruns. Finally, the question of the depreciation of existing assets would have to be resolved in a way that did not reduce DND’s ability to pay for its procurement plans. If any of these assumptions turns out to be invalid, the CFDS procurement plan will not be affordable without additional money.

The Budget 2010 decision and subsequent government statements highlighting the necessity of government spending cuts suggest that less money will be available for DND than was originally anticipated in the CFDS. The analysis in this article was based on the assumption that the Budget 2010 cuts would fall exclusively in areas that would not affect procurement spending. This assumption is probably not realistic.

Analysis of the CFDS procurement plan from a cash-based perspective also suggests that either the plan is drastically underfunded or the planned level of spending in cash terms is much higher than has generally been assumed. The latter interpretation is probably correct, but the current level of spending is not likely sufficient to support the entire CFDS plan, even if the inflation, cost overrun, sustainment, and Budget 2010 assumptions made above hold true.

Since the likelihood that all of these assumptions are correct is slight, current funding plans will not be sufficient to support the CFDS’s procurement ambitions. We can thus anticipate that the government will eventually be forced to seek increases in military spending substantially greater than those already planned or that it will be forced to trim the CFDS procurement plan to fit the money available.

The size of the probable shortfall is difficult to estimate with available information. It would be helpful if the government would make public more details about its procurement plans, including cash estimates of total life-cycle cost, inflation assumptions, and cost overrun contingencies. The question of the depreciation of pre-2008 assets also needs to be resolved to properly assess the accrual accounting-based numbers provided in the CFDS.



1. Based on the CFDS-projected inflation rate, this modified CFDS baseline budget will correspond to about $370-billion in 2008 dollars over the 20 years. To maintain comparability with the figures used in the CFDS, however, this analysis will restrict itself to nominal-dollar figures.

2. See DND 2008b, p. 60 and Table 13, Section III: Supplementary Information; DND 2009. The 2010–2011 Report on Plans and Priorities contains no such information, however.

3. I have not found sufficient information to assess the adequacy of the $25-billion in projected “Other Capital” spending or the $40-billion in projected “Infrastructure” spending.

4. Sustainment costs frequently do include some capital expenditures, however, including the cost of upgrades and life-extension projects and the replacement of equipment lost to attrition.


Auditor General of Canada. 2008. Observations of the Auditor General on the Financial Statements of the Government of Canada for the Year Ended March 31, 2008. Public Accounts of Canada, 2007–2008.

———. 2010. Observations of the Auditor General on the Financial Statements of the Government of Canada, Public Accounts of Canada, 2009–2010,

Department of National Defence (Canada). 2008a. Canada First Defence Strategy.

———. 2008b. Report on Plans and Priorities 2008–2009.

———. 2009. Report on Plans and Priorities 2009–2010.

———. 2010. Report on Plans and Priorities 2010–2011.

Pugliese, David. 2011a. DND to cut 2,100 jobs, plan shows. Ottawa Citizen, May 26.

———. 2011b. Robert Fonberg To Head Review of Canada First Defence Strategy? What Might Be Dumped from the CFDS? Defence Watch blog, May 29.

Spread the Word