Fiscal Institutions in New Zealand and the Question of a Spending Cap

New Zealand’s current fiscal policy framework has been in place for nearly 20 years. At its core is a set of principles around maintaining prudent levels of public debt and running fiscal surpluses on average over time. This framework, combined with an extended period of economic growth, contributed to New Zealand entering the global financial crisis with historically and internationally low levels of public debt. While the current fiscal policy framework has helped achieve and maintain defined, prudent levels of public debt, it is does not require the government to define a target level for government spending. Over recent years, government spending has increased as a share of GDP. Most of this reflects increased spending during the extended economic upturn through the middle of last decade. The recent recession has also played a small role in increasing spending, largely through the automatic stabilisers as New Zealand did not implement a substantive expenditure-based stimulus package. The Government therefore committed to investigating whether a spending cap would be an appropriate addition to the existing fiscal policy framework. This paper outlines the motivation for such a spending cap, presents a proposed design, including some of the potential challenges, drawing heavily on international experience. Reflecting on this analysis, the Government decided not to introduce a formal cap on total spending in Budget 2010. The benefits of the proposed spending cap are that it would have reinforced the commitment to the existing limit on new initiatives (via the $1.1 billion Operating Allowance) and placed an indicative limit on other forecasted expenses increases that go through the Baseline Update process. However, the complexity of the proposal may have led to significant communication challenges and some confusion about how it would operate alongside the existing system.


Introduction
New Zealand's current fiscal policy framework has been in place for nearly 20 years. At its core is a set of principles around maintaining prudent levels of public debt and running on average over time fiscal surpluses. This framework, combined with an extended period of economic growth, contributed to New Zealand entering the global financial crisis with historically and internationally low levels of public debt.
-----While the current fiscal policy framework has helped achieve and maintain defined, prudent levels of public debt, it is does not require the government to define a target level for government spending. Over recent years, government spending has increased as a share of GDP. Most of this reflects increased spending during the extended economic upturn through the middle of last decade. The recent recession has also played a small role in increasing spending, largely through the automatic stabilisers as New Zealand did not implement a substantive expenditure-based stimulus package. The Government therefore committed to investigating whether a spending cap would be an appropriate addition to the existing fiscal policy framework.
Section 2 considers the literature on fiscal rules, how they have been used internationally and how they have performed over the past few years. One thing that is apparent is that the appropriate design for a spending rule is dependent on the existing fiscal arrangements. Therefore, Section 3 outlines New Zealand's current fiscal institutions and Section 4 describes the evolution of Budget management processes. Section 5 provides some more context by outlining New Zealand's economic and fiscal performance over the past decade. Section 6 outlines some of the key design choices that would be relevant if a spending cap was to be introduced in New Zealand. Section 7 then discusses some the Government's reasoning for not going ahead with a cap on total spending at this point in time.

2
Fiscal rules -theory and international experience

Definitions and objectives of fiscal rules
Fiscal rules are a subset of fiscal institutions -the arrangements that form a nation's public finance framework. Institutions include the legislative framework for budgeting and fiscal planning, any policy guidelines or well-established norms, the public institutions involved in the planning and implementation of the budget process, and any independent entities that give advice or monitor performance. Kopits and Symansky (1998) define a fiscal rule as "a permanent constraint on fiscal policy through simple numerical limits on budgetary aggregates". Although the legal form can varyinternational treaty, constitutional amendment, legal provision, or policy guideline -a common theme, as the International Monetary Fund (IMF, 2009) has noted, is that fiscal rules are all mechanisms aimed at supporting fiscal credibility and discipline. Ongoing debate over the relative merits of rules versus the merits of other institutions, such as a fiscal policy committee or a fiscal advisory council, is outside the scope of this paper. 1 Fiscal rules can have various objectives, such as promoting debt sustainability, promoting macroeconomic stabilisation, containing the size of government, or supporting intergenerational equity. The key objective is usually the promotion of fiscal sustainability. The IMF (2009) has compiled a dataset of fiscal rules applied to central government in member countries, and characterised the rules into the following groupings: • budget balance rules -including rules that relate to the overall balance, the structural or cyclically-adjusted balance, or the balance over the cycle, with the aim of restraining the buildup of debt-to-GDP ratios; • debt rules -such as a limit or target for public debt as a share of GDP; • expenditure rules -also known as spending rules, may involve limits on total, primary or current spending, either in absolute terms, growth rates or as a share of GDP; and -----1 Wyplosz (2005), for example, argues that rules are often too flexible or too stringent, and that adequate incentives backed by institutions are the better option.
• revenue rules -may be ceilings to prevent an excessive tax burden, or floors aimed to boost revenue.

Prevalence of fiscal rules
Fiscal rules have become more prevalent among countries over the past two decades. The IMF (2009) has documented a rise in the use of fiscal rules; in 1990, only seven countries had national or supranational fiscal rules applying to central government, whereas by 2009 this had increased to 80 countries. This increased attention to fiscal rules was, at least in part, a reaction to a build-up of public debt in many countries through the 1970s and 1980s.
In recent years, spending rules (a subset of fiscal rules) have become more widespread, reflecting a trend for countries to move from a single rule (such as a debt or a balanced budget rule) to multiple rules. The choices and tradeoffs involved in a wider set of rules are discussed by Anderson and Minarik (2006) and Kumar and Ter-Minassian (2007). In 2009, 25 countries were making use of spending rules in some form -whereas only ten countries had been using a spending rule in 1999 (IMF, 2009). The increased prevalence of spending rules, in particular, reflects the fact that a debt target or balanced budget rule, on its own, places little discipline on the growth in government spending in the times of strong revenue growth during an economic expansion (Barker and Philip, 2007).

Design features
The IMF has suggested that there are three components of effective fiscal policy rules: 1) an unambiguous and stable link between the numerical target and the fiscal objective; 2) sufficient flexibility to respond to shocks, so that a rule should at least not exacerbate the macroeconomic impact of a shock; and 3) a clear institutional mechanism to map deviations from the rule into incentives to take corrective actions (e.g., by raising the cost of deviations, or mandating the correction of a deviation).
The legal form of fiscal rules may vary. With regard to spending rules, although in some (predominantly developing) countries these are embedded in national legislation, the IMF (2009) has found this is not necessarily a requirement for a rule to endure. Ljungman (2009) examines spending rules in three countries -Finland, the Netherlands and Sweden -and found that each has the status of a political commitment with no predefined sanctions in the event of a breach, other than reputational costs for the Government. Ljungman concludes that any spending rule that is not perceived as serving the interest of the Government and Parliament will inevitably be circumvented, and that "in the absence of this widespread political support, it is doubtful that the legislative status of a spending rule will have any impact on actual policy formulation".

Effectiveness of fiscal rules
Research into the effectiveness of fiscal rules is ongoing, but in reviewing available empirical studies, the IMF recently concluded that fiscal rules have generally been associated with improved fiscal performance (IMF, 2009). In addition, Badinger (2009) has found tentative evidence across a sample of OECD countries that the fiscal rules introduced since 1990 reduced the extent to which governments have made use of discretionary fiscal policy, although no New Zealand-specific results are reported. Intuitively, the effectiveness of a rule depends on the institutional context into which the fiscal rule is being applied and the existing macroeconomic environment, as well as the design of the rule itself.
In terms of spending rules, countries such as Finland, the Netherlands and Sweden appear to have had positive experiences. Ljungman concluded that the general impression in each of those countries has been that a spending rule has contributed to maintaining stable public finances. However, as Ljungman notes, an unambiguous correlation between the spending rules and the robustness of public finances is difficult to establish, particularly since economic growth had been relatively strong in the period between their introduction in the mid-1990s and the time of his review in 2008. In addition, Finland and the Netherlands are part of the euro area, so it is plausible that improvements in the conduct of their fiscal policy have been influenced by requirements of the Stability and Growth Pact associated with that monetary union.
The global financial crisis in 2008-09 and the associated macroeconomic shocks have posed challenges for fiscal institutions in many countries. There are signs that even countries with established spending rules have substantially increased spending in an environment with lowerthan-expected economic growth and decisions to implement fiscal stimulus packages. For example, the OECD's Economic Outlook from May 2010 forecasted general government spending as a share of GDP to have increased between 2007 and 2011 in Finland (+8.2 percentage points), the Netherlands (+6.4 percentage points) and Sweden (+2.8 percentage points). 2 It will be interesting to see how countries with spending rules fare in managing spending growth over the next few years.

3
New Zealand's legislative framework 3 Reflecting a combination of external factors and policy choices, New Zealand's fiscal position deteriorated considerably from the mid 1970s until the early 1990s, with net public debt rising from around 5 per cent of GDP in 1974 to above 50 per cent of GDP in 1992. In response, the Government adopted a number of practices that aimed to improve fiscal management, with a large emphasis on transparency. The Fiscal Responsibility Act (FRA) of 1994 codified the initial practices, including the shift to accrual accounting, the publication of short-term fiscal forecasts and the publication of a pre-election economic and fiscal update.
The FRA aimed to address the earlier poor fiscal performance by: • strengthening the incentives on Ministers to set Budget priorities and to follow an agreed fiscal strategy; and • providing more regular information to the public on the medium-term fiscal outlook and the decisions that underpinned that outlook.
In 2005, the FRA was incorporated into the Public Finance Act (PFA) of 1989. The fundamental principles of responsible fiscal management contained in the 1994 Act were retained (see below). The intention of the merger was to consolidate legislation regarding public finance and it also provided the opportunity to make some amendments to the FRA.
The amendments were introduced to align New Zealand's fiscal reporting with best international practice after assessing legislation in the United Kingdom and Australia, reviewing the best practice guidelines issued by the IMF and OECD and drawing on experience with the legislation since its introduction. The key addition was a legislated requirement for the Treasury to -----2 The comparable figure for New Zealand is +4.4 percentage points. 3 The section draws on New Zealand Treasury (2005). Scott (1996) and Janssen (2001) discuss the development of the Fiscal Responsibility Act and its relationship to wider public sector reform such as the State-owned Enterprises Act 1986, the State Sector Act 1988 and the Public Finance Act 1989. produce a regular statement on the long-term fiscal position covering at least 40 years (New Zealand Treasury, 2009).
The PFA sets out five principles of responsible fiscal management. The two that are most relevant for this paper are those associated with debt and fiscal balance: 4 • Reducing total debt to prudent levels, so as to provide a buffer against factors that may impact adversely on the level of total debt in the future. Until prudent levels of debt have been achieved, the Government must ensure that total operating expenses in each financial year are less than total operating revenues in the same financial year. • Once prudent levels of total debt have been achieved, maintaining those levels by ensuring that, on average, over a reasonable period of time, total operating expenses do not exceed total operating revenues.
Definitions such as "prudent" level of debt or "reasonable period of time" are not specified in the PFA. It is left to the Government of the day to interpret the relevant fiscal terms. Importantly, although a Government can depart from the principles, the PFA requires any such departure to be temporary and that the Minister of Finance specify the reasons for departure, the approach to be taken to return to the principles and the period of time that this is expected to take.
In addition, the PFA requires the Government to annually state long-term (ten or more years) fiscal objectives and short-term (three year) fiscal intentions for the following variables: 5 • total operating expenses; • total operating revenues; • the balance between total operating expenses and total operating revenues; • the level of total debt; and • the level of total net worth.
With the exception of the principles of responsible fiscal management that relate to debt and the operating balance, the PFA is not prescriptive about what the fiscal objectives and fiscal intentions should be. Rather, it requires the Government to state its objectives and intentions, whether they have changed and how they accord with responsible fiscal management. This means that a trend increase in government expenses as a share of GDP is permissible under the PFA provided that the principles relating to debt, the operating balance, and revenue are adhered to.

New Zealand's budget management process
As with the legislative framework, the Budget management process has evolved over the past 20 years. This evolution can be split into three distinct phases: fixed nominal baselines; fiscal provisions; and the Fiscal Management Approach. 6

Fixed nominal baselines
Prior to the introduction of the PFA in 1989, the Budget process involved making regular -----4 The others relate to net worth, fiscal risks, and the predictability and stability of tax rates. 5 The reporting requirements in the PFA relate to a definition of "total" government that includes the Core Crown, Crown entities, and State-owned Enterprises (SOEs). Given the central role of the budget, fiscal policy has focused on the Core Crown and Crown entities. 6 More detail and evaluation is provided in Barnes and Leith (2001); OECD (2002); the New Zealand Treasury (2003); and Wilkinson (2004). adjustments to personnel costs based on public sector wage negotiations. Operating and capital spending were generally adjusted annually to reflect expected cost movements. Government Budgets were made only for the year ahead with no forecasts of spending in subsequent years.
The early 1990s saw a shift to fixed nominal baselines, where the "baseline" is the agreed Budget allocation over the forecast period. Government spending was split into "formula-driven" and "fixed" (i.e., no change to nominal baseline amounts). Formula-driven indexation applied to non-departmental spending on benefits (e.g., inflation indexation of unemployment payments, wage indexation of public pensions) and volume adjustments. A specific policy decision was required to change non-indexed spending. A key issue to emerge was the effect of fixed nominal baselines on the short-term fiscal forecasts. For example, three-year fiscal forecasts between 1994 and 1996 included increases in government spending only for those areas affected by indexation. With all other spending assumed to remain constant over time, this yielded a profile of rising forecast surpluses. Together with concerns about agencies' abilities to meet rising costs this created pressure to increase nominal baselines.

Fiscal provisions
In its 1997 Budget, the Government adopted a $5 billion (cumulative) cap on new spending over the three fiscal years 1998 to 2000. This cap was on top of expenses already included in the fiscal forecasts (i.e., on top of the fixed nominal baselines and formula-driven indexed items). The cap evolved into a mechanism known as the fiscal provisions, which also included a set of rules for identifying which items would be treated as specific policy decisions and therefore "counted" towards the cap on spending. Formula-driven increases in expenses that did not "count" would still be permitted but did not impact on the amount available for new initiatives. For example, an increase in unemployment benefit payments due to higher unemployment would not be financed by (or "count against") the fiscal provisions.
A capital provision, linked to the debt objective, sat alongside the operating provisions. The capital provision generally provides for new investments or where maintaining current operations cannot be funded from accumulated depreciation on balance sheets.

Fiscal management approach
In Budget 2002, the Government signalled a change to the fiscal provisions framework that: • shifted the focus to the paths of the operating balance and debt rather than just the nominal new spending amount; and • sought to ensure that spending intentions remained relevant as the economic and fiscal outlook evolved. Spending plans would be reviewed twice yearly with reference to updated forecasts and progress against fiscal objectives.
These new procedures were termed the Fiscal Management Approach (FMA), with the amounts for new initiatives being relabelled as the Operating Allowance (for expense and revenue initiatives) and the Capital Allowance (for capital initiatives). This is the system that remains in place today.
Under the FMA there are three ways that the levels of expenses, revenue and capital items can change.
The first is changes in the profile of the expected values of expenses, revenue and capital resulting from current policy settings (referred to as the "profile"). For expenses, these changes will generally result from existing demand-driven programmes. For example, the current forecasts will build in an expectation of the rising cost of New Zealand Superannuation (NZS) as more people reach retirement age. In Budget 2010 the forecasted cost of NZS in 2010 is $8.287 billion and in 2011 is $8.822 billion and in 2014 is $10.781 billion. This expected rising profile is built into the expense forecasts.
The second way in which expenses, revenue and capital can change is via the addition of new discretionary initiatives which are included as part of the Operating Allowance (for revenue and expenses) and the Capital Allowance. These are referred to as "new discretionary initiatives". The focus of Budgets has tended to be on allocating those allowances to the Government's priority initiatives. The allowances are set with a view to achieving the Government's medium-term operating balance and debt objectives. So, if the Government decided to increase the rate at which NZS is paid or change the eligibility criteria which increased uptake, those discretionary policy decisions would be counted against the Operating Allowance in the year the decision was made. New discretionary initiatives are then incorporated into the base or the profile of forecasted spending for future years.
The third way in which expenses can change is when there are revisions to the forecasted expenses of existing programmes which are seen to be outside the direct control of government because they are demand, volume or index driven (these are referred to as "changes in forecasted costs"). These changes in forecasted costs are incorporated automatically through the Baseline Update process. This occurs twice a year as part of the updating of the fiscal estimates during the forecast round. Many of the non-welfare related Baseline Updates were originally envisaged as "counting" against the Operating Allowance. Overtime this practice has changed, and some spending increases have not been counted against the Allowance, e.g., the increased costs of KiwiSaver, a subsidised saving scheme, due to higher than forecast uptake. The Baseline Update process also incorporates other changes to baselines, such as those due to policy decisions (e.g., a decision to bring forward forecast expenditure) or valuation changes relating to impairments (mainly of student loans and tax receivables, and reflecting changes in future collectability of these assets).
This separation between demand-driven items that are automatically incorporated into the forecasts via the Baseline Update process and discretionary initiatives that count against the fixed Operating and Capital Allowances puts some pressure on the boundary between the two categories. The FMA specifies a set of rules as to what types of new initiatives must be agreed to within and outside the Operating and Capital Allowances. In addition, the government is ultimately responsible for setting the allowances in each Budget so as to achieve its fiscal objectives.
In setting the Operating and Capital allowances under the FMA, information on the macroeconomy is also considered. The weight put on macro-stability issues ("macro headroom") relative to sustainability issues ("fiscal headroom") has varied through time depending on the stage of the cycle.

5
New Zealand's economic and fiscal performance over the past decade

The 1998 to 2007 economic expansion
Between the September quarter 1998 and the December quarter 2007, New Zealand experienced its longest period of economic expansion since 1945. Although the expansion was not as long as those experienced in countries such as Australia and the United Kingdom, the length of the expansion still made it difficult to establish at the time how much of the increase in economic activity was sustainable and how much was cyclical. Figure 1 presents the estimated output gap for that period, from the perspective of 2010.
A l o t o f t h a t growth was based on fundamentals, such as population growth, a strong global economy and rising terms of trade. H o w e v e r , a s t h e expansion continued, there was increasing concern about the buildu p o f i m b a l a n c e s , reflected in excess credit growth, increased net foreign liabilities and h i g h n o n -t r a d a b l e inflation.
Throughout this period, the Government's fiscal strategy was to strengthen the fiscal position, both through d e b t r e p a y m e n t t o a c h i e v e t h e d e b t objective and through accumulating financial assets in the New Zealand Superannuation F u n d ( N Z S F ) . T h e Government established the NZSF in 2001 as a means to prefund out of current tax revenue some of the projected increase in fiscal costs associated w i t h t h e a g e i n g population (e.g., public pensions). This meant running successive operating surpluses - Output Gap (percent of GDP) * GAAP data for Total Crown Operating Balance for these years has not been backdated on IFRS basis. Source: New Zealand Treasury.  1995* 1996* 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Note: History based on a multivariate filter. Forecasts based on a production function. Source: New Zealand Treasury, Budget 2010.

Figure 2
Operating Balance Before Gains and Losses (percent of GDP) something that occurred up until 2008-09, as Figure 2 shows. This approach was in lieu of relying solely on increased future debt levels and future tax revenue or decisions to alter the public pension liability by changing eligibility or entitlements.
In the early 2000s, the fiscal strategy was achieved by relatively tight fiscal discipline. By the mid-2000s, the extended period of strong economic activity meant that the Government was presented with a series of upward revisions to its revenue forecasts (see Figure 3). For example, actual revenue for the 2008 financial year was about $2.5 billion higher than the forecast figure produced at Budget 2007. These revenue surprises saw the fiscal position strengthen faster than planned.
The Government's response to the strongerthan-expected revenues included faster debt repayment (see Figure 4) and an associated downward revision of its longterm debt objective, and increasing government spending. In addition, the corporate tax rate was reduced in 2007 and personal tax rates were reduced in 2008 with a reduction to the top threshold rate in 2009.   1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 The process for increasing spending and reducing taxes was primarily by increasing the Operating and Capital Allowances. When the Budget management process was changed to the FMA, the allowances were expected to be medium-term concepts that were set with a view to achieving the Government's medium-term operating balance and debt objectives. They were not expected to be revised frequently. However, in practice, the Government tended to use the positive revenue surprises and lower-thanexpected levels of other expenses (see Barker, Buckle and St Clair, 2008) to increase the size of the Operating Allowa n c e . T h u s , t h e Operating Allowance tended to be revised (usually upwards) twice y e a r l y w h e n t h e economic and fiscal forecasts were done.   Much of this increase reflected Budget decisions to direct new discretionary resources to expand existing services (e.g., health care, education and justice) and to increase transfers in the form of income subsidies for low and middle income working families, interest-free student loans and a subsidised saving scheme (KiwiSaver).
But a considerable share of the growth in Core Crown expenses over this period -around 40 per cent -occurred as a result of both the changing profile of costs o v e r t i m e a n d t h e changes in forecasted costs. For example, the actual cost of NZS grew by $190-$540 million per a n n u m . F o r e x i s t i n g programmes like NZS it is not straightforward to distinguish between the changes due to the rising profile and the forecast changes in the historic d a t a . F o r n e w e r initiatives like KiwiSaver, it is possible to identify the changes to forecasted costs because the initial forecasts were counted against Operating Allowance in the year in which -----7 The chart focuses on the final year impact as the profile across the forecast horizon varies. A s w i l l b e discussed below, it is these sorts of changes to forecasted costs that could have been subject to an indicative limit and the associated trade-offs of a spending cap.

The 2008-09 recession and the global financial crisis
Although the New Zealand economy has performed much better than many other developed economies during the global financial crisis, it still contracted 3.4 per cent in real terms from the beginning of 2008 to the middle of 2009. As well as bringing the earlier expansion to an abrupt end, it prompted most forecasters to significantly revise down their projections for trend economic activity going forwardincluding the Treasury, as Figure 8 shows.
Therefore, not only did the fiscal position deteriorate as revenues declined through the recession and as a result of the tax cuts, but s t r u c t u r a l d e f i c i t s emerged because some  Year ended 30 June of the previous fiscal expansion was premised on the earlier -but ultimately overly optimisticview of trend economic activity. As a result, net debt was projected to rise faster and further than previously projected.
Whatever the cause of the structural deficits, it was apparent at the time of Budget 2009 that a significant period of fiscal restraint was going to be required to return the forecast fiscal accounts to a sustainable position (see Figure 9). Budget 2009 included the postponement of scheduled personal tax cuts, a temporary suspension of contributions to the NZSF and a downward revision of future Operating Allowances.

Overall assessment of the past decade
Over the past decade New Zealand's fiscal position has strengthened considerably as a result of a combination of fiscal consolidation, improved institutional arrangements that had been established earlier, and improved economic performance.
In particular, the debt objective has been a key fiscal anchor that has helped communicate the Government's fiscal strategy and acted as a Budget management tool. By 2006, net debt had returned to below 10 per cent of GDP, where it remained until the advent of the global financial crisis. However, the fiscal framework did not constrain expenditure growth during a period of sustained economic expansion. Although a trend increase in government expenses as a share of GDP is permissible under the PFA, self-imposed expenditure objectives were either not achieved or revised upward, and there was insufficient attention paid to the base of spending -both its level and composition. 8 These broad conclusions are reflected in a number of papers assessing New Zealand's fiscal framework (see Janssen, 2001;OECD, 2002;Wilkinson, 2004;and Buiter, 2006).
The macroeconomic stabilisation role of the FMA, particularly in an environment of revenue surprises, and the potential role of alternatives is considered by Barker and Philip (2007). Barker and Philip conclude that the challenges of identifying and adjusting to permanent changes in the fiscal outlook are likely to have remained under any alternative Budget management approach.
In its 2008 Briefing to the Incoming Minister the Treasury wrote: "Given your priority around disciplining government spending we think there would be merit in adopting an additional fiscal anchor in the form of a medium term expenditure or revenue constraint (e.g., as a share of GDP)". The benefits to the Government of adopting such an anchor were seen as: • signalling an intent to restrain the growth in spending and commitment to particular revenue levels to better manage expectations over the next three years and beyond; • potentially increasing the contribution of fiscal policy to macroeconomic stability by providing more certainty and better supporting monetary policy over the longer term; and • assisting the government to achieve a slowing in expenditure growth from current rates over the longer term to manage future spending pressures.
Similarly, the OECD also recommended consideration of a spending cap for New Zealand (OECD, 2009). This focus was reinforced by the Minister of Finance, who stated in the 2009 Fiscal Strategy Report and the 2010 Budget Policy Statement that the Government was investigating a spending cap as a way of strengthening its fiscal strategy. The next section outlines some of the key design choices that the Treasury considered when preparing advice on whether or not a spending cap would be appropriate for New Zealand.  (1) Excludes the impact from the tax package.

6
Designing a spending cap for New Zealand

Objectives of the cap
The main objective of the proposed spending cap was to help the Government deliver on its fiscal strategy. The fiscal strategy is focused on achieving the debt objective by managing the operating balance and capital spending. For a given revenue track, the way to manage the operating balance is to control government spending. For example, the Budget 2010 fiscal strategy projects a reduction in core Crown expenses from a peak of 34.7 per cent of GDP in 2011 to 28.4 per cent by 2024 -the final year of the projection period.
There are several ways in which a spending cap could potentially achieve that fiscal control: • Increase transparency around the total level of spending (in 2010-11 around $70 billion), with more focus on baselines and less on the new discretionary initiatives (the $1.1 billion Operating Allowances). The cap would have been (in theory) a simple number against which the public could assess the actual level of government spending. • Provide some built-in inertia in response to revenue surprises. Any upside revenue surprise would not immediately translate into higher spending, although it could have been factored in when resetting the cap. • Improve fiscal management by putting a cap on total spending not just on discretionary new initiatives. The expenses that currently go through the Baseline Update process are subject to a lower degree of scrutiny than those expenses that count against the Operating Allowance as they are seen as outside the direct control of Government. However, many of the changes in costs are flow-on effects of policy choices made by the Government (e.g., benefit indexation is a policy choice). However, as discussed below, many of these other items would have remained outside the spending cap for various reasons.

Design of a spending cap
This section outlines the main design features of a possible spending cap designed to work within New Zealand's existing institutional framework. We have drawn on the experiences of the Netherlands, Sweden and Finland, adopting the aspects that best suit our objectives and New Zealand's economic and fiscal environment.
On the face of it, the idea of a cap on government spending sounds relatively simple. However, as noted below, many of those countries with existing expenditure caps have a range of exclusions. On reflection Treasury considered that some exclusions would likely be appropriate in the New Zealand context, for the reasons outlined below.
The proposed spending cap would have been for an absolute dollar figure for government spending based on core Crown expenses -this is a measure of operating expenses. The measure would have therefore excluded capital spending and the spending undertaken by State Owned Enterprises (SOEs). Crown funding of Crown entities would fall under the cap. The rationale for excluding capital spending was so that governments would be less likely to cut back on potentially productive capital projects instead of stopping or scaling back ongoing programmes out of operating expenditure. While this runs the risk of expenditure that should be considered as operating expenditure being classified as capital spending, prudent accounting practices and the maintenance of the debt objective would likely have helped limit such practices.
To reduce the risk of the spending cap making fiscal policy more pro-cyclical (e.g., to prevent the need to cut spending during times of recession in order to reduce the deficit), we thought it would have been appropriate to exclude unemployment benefit spending and debt finance expenses from the coverage of the cap.
We also thought it would have been appropriate to exclude remeasurements, losses and debt impairment because these are large and volatile items of spending which are seen to be outside the direct control of the Government.
Given data limitations and the compliance costs of overcoming those limitations, tax expenditures would not have been included. However, the Treasury is working to improve the accountability and transparency of tax expenditures (Fookes, 2009), which will likely make it more difficult and transparent for Governments to use tax expenditures to circumvent other budgetary processes. As part of Budget 2010, the Government released some information about tax expenditures as a step towards increasing transparency. 9 The proposed spending cap would have been set in nominal terms to avoid the need to deflate a target set in real terms. In addition, a nominal target would tend to result in less procyclicality of fiscal policy than would a real target or a short-term ratio to GDP target.
Under the proposed design, the expenditure cap would have been set for three years with the third year out being set on a rolling basis. The Operating Allowance for new operating initiatives would have been retained. The Operating Allowance seeks to limit new discretionary spending and revenue initiatives, while the spending cap would have sought to limit total spending. However, there is a link between the two. The expense forecasts assume that all of the Operating Allowance will be used for expenses rather than revenue. If a portion of the Operating Allowance was subsequently used for revenue initiatives, that amount would not be available for new operating spending. Thus, the new path of forecast expenses would be lower than the original forecast. As a result, with an unchanged spending cap, there would appear to be extra room under the cap (i.e., a larger margin) equal to the size of the revenue initiative. Therefore it would be important to ensure the Government did not revise the Operating Allowance to try to make use of the extra room under the cap.

Setting the cap
Consistent with the intent of the PFA, the level of the proposed cap would have been set by the current administration, rather than prescribed in a way that attempts to set the cap for future, yet-to-be-elected governments. Although an incoming Government would have the ability to reset the spending cap, the transparent nature of New Zealand's fiscal framework means that the new Government would have been expected to explain and justify any change.
To set the cap, the Government would have started with the forecasts of expenses being subject to the cap. These forecasts would have included the base as well as the expected profile over time plus the Operating Allowance for new operating initiatives -The forecasted amount is the amount the Government expected to spend. The Government would then add a margin (itself not in the forecasts) to that forecast level of spending. That margin would be designed to provide a buffer for unforeseen movements in forecast expenses (e.g., those that go through the Baseline Update process).The forecast amount plus the margin would determine the level of the cap -this is the amount the Government promises not to exceed.
The spending cap would have reinforced the limit on new discretionary spending imposed by the Operating Allowance as well as placing an indicative limit on the changes to forecasted costsdescribed in Section 4. However, because the calculation of the cap is based on the existing forecasts, the spending cap would not have placed any limit on the increase in expenses due to changes in the profile of existing spending. For example, it would have incorporated the existing forecast increase in NZS, expected over time as increasing numbers of people reach 65 years of age.
The level of the cap, and therefore the margin, would have essentially been an explicit commitment by the Government not to increase spending above that level. As such, the cap (and the margin) would not have represented an amount of money that is available for spending (unlike the Operating Allowance). Even if the Government only used a small amount of the margin (i.e., did not exceed the cap), it would still have been spending more than it originally forecast.
The size of the margin would have been an important element in the credibility of the spending cap. If it was set too tight, the Government may have been required to make significant cuts to spending in other areas to accommodate forecast changes, or risk revoking the cap. If it was set too loose, the spending cap would exert no effective fiscal discipline.
-----10 Some countries do allow for revisions for technical changes or changes with justification.
But the appropriate size of the margin is dependent on the other measures used to provide flexibility within the cap. If most of the cyclical or other volatile elements were excluded from the coverage of the cap, the size of the margin would be smaller than if those elements remained. The rules around what happens if the Government exceeds the cap are also pertinent. If exceeding the cap was not permitted or was reputationally costly, we would expect the margin to have been higher than if there were softer penalties for breach.
In assessing the size of the margin we looked at the size of the margin in other countries. The largest margin of 1 per cent of government expenditure in any one year is used by Sweden, which does not exempt any items from its expenditure ceiling, but governments there are able to use some of the margin for new discretionary spending. Their experience suggests that the lack of other exclusions significantly helps with the communication and monitoring of their cap. The Netherlands' ceiling covers about 85 per cent of government expenditure and has a margin of about 0.5 per cent. Additional leeway was provided by a deliberate policy of using conservative forecasts. Finland's ceiling covers 75 per cent of government expenditure and their margin is about 0.25 per cent.
To help determine an appropriate margin for New Zealand we undertook an analysis of past changes in expense data to assess how large a margin would have had to have been to cover the fluctuations that occurred. This could only be a hypothetical analysis given that a spending cap was not in place at the time and fiscal circumstances were different (i.e., the revenue surprises discussed in Section 4).
In assessing the size of the margin, we also considered other differences between New Zealand and the countries that currently operate spending caps. For example, New Zealand is a small open economy, meaning that the economy and the fiscal position are likely to be more volatile than in larger, less open, economies. Furthermore, New Zealand is one of just a handful of countries that reports its fiscal accounts on an accrual basis rather than a cash basis. This has the potential to add to the complexity of communicating outturns relative to a cap.
Weighing up all of these factors, our preference was for a margin of around 1 per cent of spending covered by the cap. For 2008-09 this would have been $550 million. A margin of 1 per cent would have been at the upper end of the margins used in other countries. This largely reflects the fact that the proposed New Zealand cap captures a larger share (95 per cent in 2008-09) of total spending than many of the caps of these other countries.

Breaching the cap
Under the proposed design, if spending exceeded the cap, the Government would have stated either in the Budget Policy Statement or in the Fiscal Strategy Report the reasons for the breach and what steps it would take to reduce spending to ensure it did not breach subsequent caps. There would not have been any explicit sanction for breaching the cap, but unless action was taken to reduce spending by the amount that the cap was breached, there would be an increased likelihood of further breaches. A breach of the cap in any one year would have used a portion of the margin available for subsequent year(s). Any spending above the forecast level of expenses (even if it did not breach the cap) would have, subject to a given revenue track, reduced the operating balance (i.e., reduce a surplus or increase a deficit) and increased debt. If spending increased to a level close to but not above the cap, this would have been revealed in the Budget Policy Statement or Fiscal Strategy Report documents. There would have been an expectation that the Government would comment on the likelihood of a breach and what the Government would do to avoid the breach occurring.
The cap would have been monitored at the aggregate level so it would be a collective Cabinet decision about where spending is reduced to address any excess. There would be a number of options for Cabinet; for example, it could: • require the department with higher-than-expected expenditure to reduce baseline spending to accommodate the additional costs; • find baseline savings in another vote; or • reduce new operating initiatives (i.e., the Operating Allowance).
Thus, if spending was higher than expected because of higher-than-forecast school enrolments, the Cabinet might choose either to reduce baseline spending in Education or find savings elsewhere to increase the Education baseline by the amount of the overspend or charge the overspend against the Operating Allowance.

Main changes from the current system
The biggest change from the current system would have been the inclusion under the cap of changes in forecasted costs that currently go through the Baseline Update process such as higher than expected costs of benefit indexation. This would mean that large increases in those items could potentially have resulted in tradeoffs with other spending, which does not occur in the current system.
The spending cap process would have put a lot more focus on the generation of the spending forecasts. There might have been an incentive for departments to pad their forecasts of spending to provide additional room for unexpected expenditure. However, this would have to have been balanced by the risk that if Ministers consider a department's spending to be inefficient they could be a target for savings to be made.
The spending cap would also have been a fixed commitment to an annual level of spending over a three year period. Given that the cap would have been introduced under the existing PFA, revisions to the cap could not have been ruled out, but any increase in the cap would have to be transparent and would have needed to be justified.
The commitment to the spending cap would also have committed the Government to a maximum level of the Operating Allowance in those years. Revisions to the Operating Allowance would generally have required revisions to the spending cap as well. The main implication of this is that temporary increases of revenue above the forecast level would not have been able to be used to increase spending during the period of the cap. The main reason for this was to ensure that increases in revenue that occurred for cyclical/temporary reasons were not spent. While the increases in revenue may be structural or permanent, it can take a number of years to identify the change in trend. If those revenue increases are in fact structural, they could then have been built into expectations about increased spending and tax cuts when the cap was reset for the third year out.

Risks around adopting a cap
The adoption of a spending cap would have carried some risks, as outlined below.
• It could have reduced the flexibility to deal with shocks as the spending cap could have reduced a Government's ability to engage in counter-cyclical spending during times such as the recent global financial crisis. The placement of unemployment benefit spending outside the cap helps to mitigate against this risk because this is the main cyclical item of expenditure. Countries such as Sweden and Finland have come through the global financial crisis without technically breaching their expenditure ceilings. In Sweden, this was assisted by the fact that some of the margin can be used for new discretionary spending which has been counter-cyclical in recent years. Others, such as the Netherlands, have made temporary amendments to their spending cap during the recent recession. • It could have hampered the Government in dealing with other shocks such as a population shock where a migration boom lead to a spike in economic growth and revenue but also health, education and other spending. While a sharp increase in population could happen quickly, the spending implications are likely to follow over time. The occurrence of such a shock may be an instance where the Government could have been prepared to explain a revision to the cap. • It could have been complex to communicate, in simple terms, the entire design specification of the cap. This could have undermined its effectiveness. • Implementing the cap within the existing framework of the PFA might have meant the cap was not durable as any incoming Government would not have been bound to follow the same protocol. • The spending cap would not have solved the problem of the inability to accurately differentiate temporary and permanent revenue surprises. Governments might still have decided to increase spending in the third year in response to a surprise increase in revenues, only to find by the time the third year came around that those revenues were temporary. The Government would still have had the option of revising down the cap if they chose. • The cap could have become a target rather than an upper limit -the Government might have faced pressure to increase spending up to the maximum permissible even in situations where it would have been prudent to reduce spending.

Other proposals for managing government spending
The above-mentioned questions about the attention paid to the base of spending, as well as questions around how a cap on total spending could bolster existing arrangements, have prompted discussion around alternative approaches to managing government spending. There are a range of alternative proposals. Two that have been discussed within New Zealand are detailed below.
A recent Government-initiated taskforce proposed that the PFA be amended to require the Minister of Finance to specify a five-to-ten year target for future operating spending -either the real per capita level of spending, or spending as a share of GDP (2025( Taskforce, 2009). The Minister would also be required to report publicly on progress relative to that goal. The proposal seeks to put the spotlight on the implications of the fiscal strategy for the size of government. The Taskforce holds the view that growth in government spending should be restrained, so that core Crown expenses decrease as a share of GDP -initially to 2005 levels (30 per cent of GDP), with the medium-term goal being 20 per cent of GDP. The PFA allows for spending intentions and objectives to be couched as a target share of GDP. The Minister of Finance set such a target in the 1995 Budget Policy Statement, although this practice has not been consistently applied.
A more prescriptive spending rule, in the form of a Taxpayer Rights Bill, has been proposed by the ACT Party, one of the governing National Party's support parties in Parliament. A similar Bill was proposed in Wilkinson (2004), drawing on the experience of Colorado in the United States. Such a Bill would limit spending growth to the rate of inflation plus the rate of population growth, with any proposal for higher spending being subject to a referendum. Furthermore, it would require any revenue above that limit to be refunded to taxpayers, unless retention of this excess revenue is approved by referendum. A legislated limit on expenses and revenue would require the PFA's principles of responsible fiscal management to be revisited. This is because the principles are based on requiring governments to be transparent when setting their fiscal strategy, whereas a highly-prescriptive fiscal rule would, in effect, largely be determining the fiscal strategy.
While this report has focussed on one possible design for a cap on total spending, there are other possible designs which may be relevant, depending on the objectives of the cap. For example, a cap could be used to place a limit on a particular type of expenditure rather than total spending.

7
The Government's response Reflecting on the above analysis, the Government decided not to introduce a formal cap on total spending in Budget 2010. Although a cap on total spending could have brought some benefits, there are also some risks, particularly associated with the complexity of the proposal. The Government considers that the current system, which includes a cap on new initiatives (including both revenue and expenses), namely the Operating Allowance, achieves some of the key objectives of a total spending cap. In particular, the Government's commitment to maintaining the Operating Allowance of $1.1 billion (increased by 2 per cent per annum from 2011-12) suggested that any future revenue surprises will not be used to increase spending.
Meanwhile, the Government continues to look for ways to address the other issues identified such as increasing the range of expenses subject to an effective limit and increasing the focus on the base rather than just the marginal spend. For example, for Budget 2010, the Minister of Finance initiated a reprioritisation process that resulted in $1.8 billion of savings within existing baselines being redirected to higher priority areas over the four-year forecast period. Budget 2010 also indicated that various aspects of the current FMA will be reconsidered with a view to improving the Government's ability to scrutinise expenditure increases that at present are not counted against the Operating Allowance.

Conclusion
New Zealand's existing fiscal framework -centred around the principles embedded in the Public Finance Act -contributed to New Zealand entering the global financial crisis with historically and internationally low levels of public debt. However, the focus on debt did not prevent Government spending increasing as a percent of GDP. This paper considered whether a spending cap would be a beneficial addition to the fiscal tool kit.
To be effective, a spending cap needed to fit into the existing FMA and be supported by a strong political will to be bound by the cap. The proposal considered in this paper entailed a rolling three-year nominal target for core Crown expenses, as set by the government. It was designed to have a range of exclusions, such as unemployment benefit expenses (which are cyclical and part of the automatic stabilisers). In addition, there was to be a margin to accommodate unexpected changes in forecast expenses.
The benefits of the proposed spending cap are that it would have reinforced the commitment to the existing limit on new initiatives (the Operating Allowance) and placed an indicative limit on changes to forecasted expenses that go through the Baseline Update process. However, the complexity of the proposal would have led to significant communication challenges. There may have been some confusion about how it would operate alongside the existing system.
The review of the FMA, signalled in Budget 2010, will assess whether more of the changes to forecasted expenses should be "counted" within the Operating Allowance. Ideally, future arrangements will also allow the fiscal pressures associated with the rising profile of some categories of demand-driven expenses (e.g., New Zealand Superannuation, some categories of welfare benefits) to be more clearly identified and compared at the same time as decisions are being made around new spending initiatives. A simple and transparent approach will ensure that the underlying trade-offs around current policy settings and their long-term fiscal effects are visible. This will contribute to New Zealand having a sustainable future fiscal path and being well-placed to respond to long-term fiscal challenges.