Fiscal consolidation and the slimmer state.
Barrell, Ray
Over the past two years governments in OECD countries have had to
ask themselves how they are to deal with the fiscal consequences of the
financial crisis. Output fell dramatically, as is discussed in Barrell
and Kirby in this Review, with some of the fall being a change in the
level of trend output. There is however a significant output gap, and a
case can be made for the government taking action to speed the process
of market-based adjustment and help close the gap by stimulating demand.
If there are other policy priorities then it is necessary to evaluate
actions both in terms of these priorities and their impact on demand and
output in the short term. When we consider issues around fiscal
consolidation we have to consider how much of the rise in deficits we
have seen is cyclical, and how much is structural. In addition we have
to ask to what extent the structural changes may require a slimmer
state. We first look at the case for reducing government borrowing, and
then at trends in spending and tax receipts. We distinguish between
decisions taken to produce a slimmer sate and those that would rebalance
the budget. In any decision making politicians have to balance
short-term costs of their actions against long-term benefit. Such
cost-benefit analyses are a normal part of the economist's toolkit.
Borrowing and borrowing costs
When governments borrow they have to pay interest to cover the real
cost of borrowing, the expected inflation rate and the expected value of
any losses that the lender might incur. Borrowing costs might rise if
the market changes its perception of the risk of default and adds a
premium to borrowing costs. If borrowing becomes more expensive, perhaps
because of default risk, it might be useful to do less of it, but
clearly this has not been the case. Figure 1 plots the government debt
stock as a per cent of GDP and government interest payments on the same
basis. The projections for the gross debt stock (excluding bank
deposits) come from our January 2011 forecast, as does the interest rate
projection, which is based on market expectations of future interest
rates and the normal structure of funding deficit funding. Government
interest payments as a per cent of GDP are likely to marginally exceed
their previous peak, although between 1997 and the first years of this
administration the debt stock as a per cent of GDP will have almost
doubled. Governments can, at the minute, borrow cheaply, especially if
they issue index-linked bonds.
We can see from figure 2 that long-term real interest rates on UK
government debt are currently lower than for most of the past two
decades. We have taken an average long-term bond rate and subtracted
from it the actual inflation rate over the next six years, and after
2004 we have gradually added more of our own forecast to this forward
projection. There appears to be no reason to think that the burden of
interest payments is currently expected to be excessive, and real
borrowing costs facing the government are very low. In addition, there
is little evidence to suggest that market perceptions of the risk of a
government default have been worryingly high. It is therefore hard to
justify a significant fiscal consolidation on these grounds, especially
in a period when there are significant spare resources in the economy.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The case for running a surplus
That it is not reasonable to make a case for reducing the deficit
in terms of the costs of borrowing does not mean we should not worry
about the debt stock. As Barrell and Weale (2010) show, borrowing now
reduces the resources available to future generations, and this transfer
of resources may not be fair. It is possible to calculate the level of
the deficit or surplus required in the medium term by looking at
spending and tax plans in relation to the structure of the population.
The requirement that the government's existing net assets plus the
present discounted value of future receipts net of payments should equal
zero makes it possible to calculate generational accounts. Future
generations are treated less favourably than current generations if the
present discounted sum of payments by future generations is larger than
that faced by current new-born children. Generational accounts also make
it possible to estimate the tax changes needed to ensure fairness. These
calculations are not particularly sensitive to the existing national
debt. This reflects the fact that the main driver of generational
imbalance is pay-as-you-go finance of age-related expenditures such as
health and welfare benefits for old people. Estimates of the
generational gaps are typically very sensitive to the real interest rate
used in the calculations. Nevertheless, the tax adjustments needed are
not very sensitive to these and are therefore more satisfactory as
indicators of possible budget imbalance.
These calculations inevitably depend on assumptions about future
population structure and about spending associated with people of
different ages. Some components of spending, such as pension payments,
are clearly policy driven. Others, such as health spending, are
sensitive to need. There are a number of alternatives, including
different assumptions about benefit policies, and patterns of health
spending on the elderly and the infirm. Recent calculations by the
Institute of the gap in tax receipts using a real interest rate of 3 per
cent per annum suggest that taxes need to rise by about a sixth, or
about 6 per cent of GDP, to deliver intertemporal budget balance. This
suggests that, in the medium term, the government should aim for a
surplus, and either raise taxes or change spending plans. However,
raising taxes or cutting spending in the short run reduces the level of
activity in the economy, and short-run costs have to be compared to
long-run benefits.
Even if the government were to reduce spending on health and
education, much as is planned, and shift the burden to the private
sector, the case for running a surplus would be little changed, but the
need for significant increases in taxes would be reduced. The need for a
surplus depends in part on the chronic lack of net saving by the private
sector in the UK. Barrell and Kirby in this Review show that national
wealth has been falling in the UK for two decades, and given the
national net saving rate it will continue to do so. Saving at a national
level is largely needed to fund retirement, and the need for saving
depends in part on our generosity to ourselves both in terms of length
of retirement and the level of income in retirement.
The fall in national wealth has happened despite current
projections of rising old age dependency rates. The population above
retirement age (or state pension age) is expected to rise by 1 1/2 per
cent of the population of working age over the next five years despite
the planned rise in the state pension age of women. Given that this is
not unexpected, the fall in national wealth from three times income in
1991 to just twice income in 2010 is hard to justify. There is a strong
case to be made for raising saving, but this would reduce the level of
activity in the economy in the short run and short-run costs have to be
compared to long-run benefits. However, there are alternatives to
raising the saving rate. If wealth is needed to pay for future
retirement incomes, we can either raise wealth or reduce needs by
raising the age at which people are expected to enter retirement. Each
additional year on working lives reduces the need for saving by around 1
per cent of GDP and, as Barrell, Kirby and Orazgani (2011) demonstrate,
it raises trend growth in the medium term by up to 0.2 per cent. (1) In
addition, each year on working lives could raise net government revenues
by 0.7 per cent of GDP in the medium term, reducing the need for fiscal
consolidation noticeably.
There are other reasons for running a surplus in addition to these.
Financial crises are inevitable, and at some point another will occur
and the deficit will rise sharply and the debt stock increase. It is
important that we leave space for this to happen, although it is better
to take action to avoid crises. (2) Deficits flow on to the debt stock,
but debt also changes because of privatisations and financial
interventions either to rescue companies or bail out banks. One might
expect the process of privatisation and nationalisation to approximately
pay for itself and hence there should be no addition to debt in the
longer term. However, the overall debt accumulation process from
financial transactions is likely to add to debt, with financial crises
in particular shifting the mean of this process. Over the period 1975 to
2009 excess debt issuance averaged 0.1 per cent of GDP a year, but these
figures do not include the costs of the recent financial crisis, which
could amount to a net 10 per cent of GDP. (3) Even if governments run
balanced budgets, they will accumulate debts as long as bad accidents
are worse than accidental good fortune. In the public sector this is
inevitably the case, and hence a targeting regime of budget balance and
a steady state debt stock of between 10 and 20 per cent would be
consistent, and if we wanted no debt we would have to run a surplus.
In these circumstances a budget deficit of 10 per cent of GDP is
undesirable, and plans have to be in place to reduce the deficit. Some
of the reduction would take place as the economy returns to equilibrium
and the output gap is closed, but in current circumstances this cannot
be expected to reduce the deficit significantly. We would judge that the
output gap is currently around 4 per cent of GDP, and given marginal tax
and benefit rates we can expect no more than a 2 per cent of GDP
increase in revenues and reduction in expenditures as the gap closes.
Hence, if working lives are not extended, then either tax rates have to
rise or spending has to be cut. However, in the current situation we
would continue to suggest that no action to tighten policy should be
taken this year.
The size of the government
There are many reasons why governments exist, and not just because
life would be nasty, brutish and short without them. Governments in a
democracy exist to mediate the various wishes of the citizens. These
seem to involve us in redistributing incomes between people at any point
in time and also redistributing them over time. In addition, governments
can raise welfare and output by undertaking activities that are less
efficiently done by the private sector because of market failure that
cannot be rectified by good regulation. As Barrell and Hubert (1999)
stress, changes in technology and in preferences change the desirable
boundaries of the state, and the optimal share of government spending in
GDP and the scale of transfers between individuals change over time.
Transfers can take place either through taxes or through benefit
payments to individuals. The government's consolidation programme
has concentrated on spending on goods and services, and figure 3 gives
details of these in real and nominal terms as a share of national
income, and we also include spending on state pensions and benefits to
those over state pension age as a share of nominal GDP. Figures for 2009
in particular are affected by the collapse in income and the rise in
unemployment, and do not necessarily represent changes in programmes.
[FIGURE 3 OMITTED]
The largest share of government spending is on goods and services,
and this involves employing people and purchasing in--for instance,
aircraft carriers are produced by the private sector but used by the
public sector. In real terms government consumption has been falling for
decades, with the largest decline after the Korean War followed by
retrenchments at the start of the Thatcher and Major administrations. As
we can see from figure 3, real spending as a share of GDP changed little
between 1998 and 2008, but nominal spending rose by around 4 per cent of
GDP. This reflects, at least in part, the difficulties of measuring
productivity in the public sector, which has been flat or falling for a
decade. It also reflects the 8 per cent rise in public sector regular
pay relative to the private sector between 2001 and 2010. These two
trends may be related, as an increase in the number of nurses, say,
relative to the number of patients, may improve the quality of care, but
it will be measured as a decline in productivity and it may require an
increase in relative wages. Some of the increase in relative wages since
2000 may reflect catching-up and will have been associated with
improvements in productivity. (4) However, in other areas, such as parts
of the health service, the improvements in relative wages can at best be
described as fortuitous, and the case for major reductions in some areas
is clear. Transfers to pensioners have been rising as a share of
national income since 1999 although their share in the population has
not risen. There may be scope for cutting public sector real wages
rather than spending, and it would be much less damaging to the economy
in the short term. There also appears scope for reducing generosity to
pensioners, and this may induce people to work longer.
Tax receipts as a share of GDP are influenced by the cycle in
output, and the fall in revenues as a share of GDP of 2.8 per cent
between 2007 and 2009 is partly explained by this. Calculations in
Holland, Barrell and Fic (2010) suggest that revenues fall by around 25
per cent more than GDP in a downturn, and therefore we might expect that
about half of the decline is cyclical. Hence around half of this fall
may be structural, with half of that coming from declines in taxes on
transactions in assets, in particular on property transactions. As it is
to be hoped that in future the property market is less feverish than in
the past, this may be seen as permanent. If people pay less tax as a
share of their incomes then they are better off, and if it is necessary
to offset the consequent worsening of the deficit then it would be wise
to raise taxes on the same group of people, perhaps by a tax on property
designed to reduce transactions and property prices.
How should we consolidate?
Given that the deficit has to be reduced in the medium term, and
that the government feels it should be seen to act now, it is useful to
ask how it might act on the deficit. There are three issues to consider
when choosing between cutting spending and raising taxes: what are the
short-term effects on output if all are equally credible, are there
differences in the long-term credibility of different programmes, and
were initial levels of tax and spending desirable?
If we look at the impacts on output of a temporary 1 per cent of
GDP reduction in spending, or increase in taxes sustained for two years,
using our model NiGEM we can see from table 1 that the cheapest way, in
terms of lost output at least, to reduce the deficit is to raise direct
taxes. These include taxes on property and stamp duties on financial
transactions as well as income taxes and national insurance. A temporary
rise in direct taxes would induce no significant supply-side response,
and would be largely absorbed by a reduction in saving. Cuts in state
benefits and pensions (transfers) and increases in indirect taxes would
have marginally more impact on output. Cutting spending is noticeably
more costly in terms of output than raising taxes (or cutting public
sector wages which has similar effects to cutting transfers) and hence
the current consolidation programme which is based largely on real
spending cuts and on reducing benefits will have much larger short-term
impacts than a programme based on direct tax increases. Given the
current scale of the output gap this is hard to justify unless there are
major gains from increased credibility.
Gains from increased credibility can exist in a forward-looking
world. If fiscal policy is tighter, then interest rates can be lower,
and in a world of flexible exchange rates the exchange rate would be
weaker. Both might help offset the contractionary effects of fiscal
policy on output and both are in operation in the results in table 1. If
fiscal policy is credibly tighter next year, interest rates can be lower
this year. The longer fiscal policy is expected to be tight, the lower
interest rates might be and hence the greater the offset. Indeed, if
fiscal policy is expected to be tighter in the future but is not
tightened now, then it is possible with forward-looking financial
markets to have an expansionary fiscal contraction. However, the policy
has to be credible to get a large offset, and the less credible a policy
is the smaller the offset in the short run. Policies that are hard to
reverse, such as a change in the retirement age, are clearly more
credible than those that will almost certainly be reversed.
It is sometimes claimed that fiscal consolidations based on
spending reductions have been more successful than those based on taxes,
and hence the offsets would be larger. In their recent paper,
revealingly entitled Received wisdom and beyond; lessons from fiscal
consolidation in the EU', Larch and Turrini (2008) suggest that
this received wisdom may not be as useful as it once was. Over the past
forty years, evidence from fiscal consolidations across the OECD does
suggest that spending-based consolidations have been more effective but,
if we split the European Union programmes at 1995, it is clear that
those before this date were better done with spending cuts, but after
that date the evidence is not so supportive of this position. Indeed,
consolidations now appear to be more effective with institutions such as
the Office for Budget Responsibility in place than without, whether the
consolidation is based on spending or on taxes. Given the ambiguous
nature of the evidence on the relative benefits in the long run of
cutting spending and the clear evidence that the costs are greater in
the short run, it may be the case that this path has been chosen mainly
to produce a slimmer state.
The decision to construct a slimmer state has been complicated by
the decision to ring-fence some areas of spending. This decision can
either be seen as the result of political expediency in the run-up to an
election or of a clear change in priorities on the part of society. (5)
Prior to the election we argued (6) that, if existing spending plans
were properly balanced, then the case for ring-fencing did not exist.
Spending on each part of the government should perhaps be seen as an
optimisation exercise where marginal social benefits, whatever they may
be, are the same across spending departments. This means that a cut
would have the same costs wherever it was applied. If the crisis has
left us worse off compared to where we thought we would be (a permanent
loss of output), then spending plans need to be re-evaluated. On top of
this we need to ensure the fiscal costs of the crisis are paid off
through a fiscal consolidation programme. Both of these considerations
led Barrell and Kirby (2010) to suggest that the optimal response to the
crisis would be to cut spending and raise taxes in equal measure. The
current split between taxes and spending in the consolidation programme
can be seen as evidence that there is a desire for a slimmer state. It
is perhaps unwise to start a discussion of the slimmer state with what
should be ring-fenced. Ring-fencing some areas has meant that cuts have
had to be deeper in others, and the incremental costs of those cuts in
terms of loss of benefit to society could be large. The state may as a
consequence be slimmer, but its shape may be rather odd.
Conclusions on the future of the state
There is no doubt that a fiscal consolidation plan should be in
place, and we have argued for one both before and after the election in
Barrell and Kirby (2010 and 2010a). Increasing the national debt
transfers resources from our children to us, and leaves us unprepared
for the next crisis. It reduces national income as higher government
borrowing will be reflected, at least in part, in increased foreign debt
and hence in interest payments to foreign residents. Its impact on the
capital stock will be much more limited as the private sector
'takes' the world rate of return, and our debt will have
little effect on this. (7)
However, the debate over the timing and scale of the consolidation
is not over. Borrowing is cheap, debt default risks in the UK are low,
and there is a significant output gap in the economy. In these
circumstances we would argue for a delay in consolidation. Governments
act slowly, and expenditure cuts on goods and services appear to be
delaying themselves, with a reduction in their scale between the Summer
Budget and the Comprehensive Spending Review. This process is likely to
continue and, if cuts cannot be made, either taxes should rise now or
tax rises should be announced for the near future. The economy would
benefit from such a shift. Better still, the Pensions Bill before
parliament should be altered and retirement ages raised by two further
years by 2020. (8)
Behind much of the current spending agenda there is an underlying
desire for a slimmer state, Some parts of this agenda are wise, and for
instance the move towards higher student fees will reduce subsidies to
those who will be wealthier than those who pay for the subsidy, and they
will also increase efficiency in university teaching. However, the
handling of the increase may be described as at best rushed. The
restructuring of benefits to remove complexity will also be beneficial
in the long run, and hopefully can be used to achieve the objectives of
redistribution for which benefits were designed. The reduction in the
scale of the state has been described as reducing bureaucracy, removing
waste and raising productivity. In each case it is not clear that this
will mean any more than a reduction in service provision by the public
sector and an increase in private provision in areas such as health and
education.
DOI: 10.1177/0027950111401132
REFERENCES
Barrell, R., Davis, E., Liadze, I. and Karim, D. (2010), 'Was
the subprime crisis unique? An analysis of the factors that help predict
banking crises in OECD countries', National Institute Discussion
Paper no. 363.
Barrell, R. and Hubert, F. (1999), Modern Budgeting in the Public
Sector, London, NIESR.
Barrell, R. and Kirby, S. (2010), 'Medium-term prospects for
the public finances', National Institute Economic Review, 212,
April, pp. F60-7.
--(2010a), 'UK fiscal prospects', National Institute
Economic Review, 213, July, pp. F66-70.
Barrell, R., Kirby, S. and Orazgani, A. (2011), 'The
macroeconomic impact from extending working lives', Department for
Work and Pensions Economics Paper no. 95.
Barrell, R. and Weale, M. (2010), 'Fiscal policy, fairness
between generations and national saving', Oxford Review of Economic
Policy, Spring.
Holland, D., Barrell, R. and Fic, T. (2010), 'Fiscal positions
in the EU and prospects for consolidation', 7th Euroframe
Conference on Exit strategies for EU economies in a globalised world,
Brussels, June.
Honohan, P. (2008), 'Risk management and the costs of the
banking crisis', National Institute Economic Review, 206, October,
pp. 15-24.
Larch, M. and Turrini, A. (2008), 'Received wisdom and beyond;
lessons from fiscal consolidation in the EU', European Economy,
Economic Paper 320.
Nickell, S. and Quintini, G. (2002), 'The consequences of the
decline in public sector pay in Britain: a little bit of evidence',
Economic Journal, 112, 477, Special Issue, February, pp. F107-18.
Reinhart, C.M. and Rogoff, K.S. (2009), This Time is Different:
Eight Centuries of Financial Folly, Princeton University Press.
NOTES
(1) If working lives were to be extended by five years sustainable
output would be up to 5 per cent higher after ten years and hence trend
growth would be 0.5 per cent higher for a decade.
(2) The increase in core capital requirement for banks set out in
Basel III should reduce the probability of crises, but, as Barrell et
al. (2010) point out, it would be hard to reduce the probability to
zero.
(3) Honohan (2008) suggests that this is the cost of an average
crisis. The cost of the crisis will not be known for some years.
(4) Nickel and Quintini (2002) discuss this issue in teaching.
(5) Both the Labour and Conservative parties announced they would
introduce some ring-fencing of departmental spending during the course
of the election campaign.
(6) See Barrell and Kirby (2010) published in April 2010.
(7) Reinhart and Rogoff (2009) produce suggestive evidence that
debt stocks in excess of 90 per cent of GDP reduce output. This may have
been the case in a world with limited capital mobility, but it is
unlikely to be the case now, except for very large economies such as the
US.
(8) Barrell, Kirby and Orazgani (2011) set out the costs and
benefits of such a move.
Table I . Effects on GDP of a temporary fiscal consolidation
(per cent difference from base--1% of GDP impulse,
interest rates fixed for 2 years)
Date Real spending Transfers Direct tax Indirect tax
2011 -0.62 -0.16 -0.13 -0.18
2012 -0.71 -0.25 -0.17 -0.26
2013 0.20 -0.07 -0.14 -0.20
2014 0.24 0.03 -0.05 -0.16
2015 0.16 0.04 0.00 -0.13
