Macroeconomic Effects of Public Pension
Reforms
Philippe Karam, Dirk Muir, Joana Pereira,
and Anita Tuladhar
WP/10/97
© 2010 International Monetary Fund WP/10/2
IMF Working Paper
Fiscal Affairs Department
Research Department
Macroeconomic Effects of Public Pension Reforms
Prepared by Philippe Karam, Dirk Muir, Joana Pereira, and Anita Tuladhar
1
Authorized for distribution by Benedict Clements, Manmohan Kumar, and Douglas Laxton
December 2010
Abstract
This Working Paper should not be reported as representing the views of the IMF.
The views expressed in this Working Paper are those of the author(s) and do not necessarily represent
those of the IMF or IMF policy. Working Papers describe research in progress by the author(s) and are
published to elicit comments and to further debate.
The paper explores the macroeconomic effects of three public pension reforms, namely an
increase in retirement age, a reduction in benefits and an increase in contribution rates. Using
a five-region version of the IMF‘s Global Integrated Monetary and Fiscal model (GIMF), we
find that public pension reforms can have a positive effect on growth in both the short run,
propelled by rising consumption, and in the long run, due to lower government debt
crowding in higher investment. We also find that a reform action undertaken cooperatively
by all regions results in larger output effects, reflecting stronger capital accumulation due to
higher world savings. An increase in the retirement age reform yields the strongest impact in
the short run, due to the demand effects of higher labor income and in the long run because
of supply effects.
JEL Classification Numbers:
H55, E6, H3, F41
Keywords:
OLG models, Aging, Pension Reforms, Fiscal Adjustment
Authors‘ E-Mail Addresses:
1
The authors are grateful to Carlo Cottarelli, Sanjeev Gupta, Benedict Clements, and Manmohan Kumar for
their guidance and to participants in the FAD seminar for useful comments. We thank Mauricio Soto for
providing pension spending projections in G-20 countries, Michael Kumhof for helpful discussions, Julia
Guerreiro for her assistance in compiling tax statistics, and Nadia Malikyar, Jeffrey Pichocki, and Mileva
Radisavljevic for help in formatting the paper. Special thanks to the Economic Modeling Unit of the Research
Department for sharing the GIMF‘s programs. Remaining errors are the authors‘ responsibility.
2
Contents Page
Introduction ................................................................................................................................4
I. Pension Spending Trends, Theory and Existing Studies ........................................................5
A. Current and Projected Public Pension Spending.......................................................5
B. Theory and Existing Studies .....................................................................................7
II. The Methodology For Modeling Public Pension Reforms ...................................................8
A. Overview of the Model‘s Key Features ....................................................................9
B. Quantifying Public Pension Reforms ......................................................................10
C. Caveats and Qualifications ......................................................................................12
D. Calibration ...............................................................................................................12
III. Results: Public Pension Reforms .......................................................................................13
A. Baseline ...................................................................................................................13
B. Region-by-Region Benchmark Public Pension Reform Scenarios .........................13
C. Benchmark Global Scenario—Simultaneous Reforms in all Regions ....................33
IV. Sensitivity Analysis ...........................................................................................................41
V. Conclusion ..........................................................................................................................42
Tables
1. Required Pension Age Extensions across the Regions ........................................................11
2. Cooperative Versus Regional Pension Reform—Increase in Retirement Age ....................38
3. Cooperative Versus Regional Pension Reform—Reducing Spending on
Pension Benefits...................................................................................................................39
4. Cooperative Versus Regional Pension Reform—Raising Contribution Rates ....................40
Figures
1. Change in Public Pension Expenditures, 2010–30 ................................................................6
2. Increase in the Retirement Age in the United States (Excluding Fiscal Consolidation) .....17
3. Increase in the Retirement Age in the United States ...........................................................18
4. Increase in the Retirement Age in the Euro Area ................................................................19
5. Increase in the Retirement Age in Emerging Asia ...............................................................20
6. Increase in the Retirement Age in the Remaining Countries Block ....................................21
7. Reducing Pension Benefits in the United States ..................................................................24
8. Reducing Pension Benefits in the Euro Area .......................................................................25
9. Reducing Pension Benefits in Emerging Asia .....................................................................26
10. Reducing Pension Benefits in the Remaining Countries Block.........................................27
11. Raising Contribution Rates in the United States................................................................29
12. Raising Contribution Rates in the Euro Area ....................................................................30
13. Raising Contribution Rates in Emerging Asia ...................................................................31
14. Raising Contribution Rates in the Remaining Countries Block ........................................32
3
15. Cooperative Versus Regional Pension Reform—Increase in Retirement Age ..................35
16. Cooperative Versus Regional Pension Reform—Reducing Pension Benefits ..................36
17. Cooperative Versus Regional Pension Reform—Raising Contribution Rates ..................37
18. Sensitivity Analysis Around the Benchmark Coordinated Global Reform Scenario—
Labor Supply Response .....................................................................................................55
19. Sensitivity Analysis Around the Benchmark Coordinated Global Reform
Scenario—Role of the Planning Horizon (Degree of Myopia) .........................................56
20. Sensitivity Analysis Around the Benchmark Coordinated Global Reform Scenario—
Share of LIQ Households at 50 Percent Worldwide ...........................................................57
21. Sensitivity Analysis Around the Benchmark Coordinated Global Reduction in Public
Pension Spending—Rapid Implementation .......................................................................58
22. Sensitivity Analysis Around the Benchmark Coordinated Global Reduction in Public
Pension Spending—U.S. Reduction Levels Applied Globally ...........................................59
23. Sensitivity Analysis Around the Benchmark Coordinated Global Reform Scenario—
Monetary Policy Accommodation .....................................................................................60
24. Sensitivity Analysis Around the Benchmark Reduction in Public Pension Spending by
the United States Only—Different Inflation Behaviors .....................................................61
Appendices
1. GIMF‘s Main Features and Calibration ...............................................................................44
2. Sensitivity Analysis .............................................................................................................51
Appendix Table
1. Selected Steady-State Values ...............................................................................................50
References ................................................................................................................................62
4
INTRODUCTION
The fiscal impact of the global crisis has reinforced the urgency of pension and health
entitlement reform.
2
Staff projections suggest that age-related outlays (pensions and health
spending) will rise by 4 to 5 percent of GDP in the advanced economies over the next
20 years, underscoring the need to take steps to stabilize these outlays in relation to GDP.
With the economic recovery not yet fully established, this paper emphasizes their short-run
macro impact in order to address concerns that these reforms can undermine short-run
growth.
3
We examine the preferred set of public pension reforms using the IMF's Global Integrated
Monetary and Fiscal (GIMF) model parameterized on data for five regions as representing
the entire world. We consider three policy reform options relating to pay-as-you-go public
pension systems that are commonly discussed in the literature. This analytical framework
allows us to approximately gauge the effects of these reforms on labor and capital markets
and growth in the short and long run.
4
(i) Raising the retirement age: this reduces lifetime
benefits paid to pensioners. Encouraging longer working lives with higher earned income
may lead to a reduction in saving and increase in consumption during working years. In
addition, increased fiscal saving will have long-run positive effects on output through
lowering the cost of capital and crowding in investment. (ii) Reducing pension benefits: this
increases agents‘ incentives to raise savings in order to avoid a sharper reduction in income
and consumption in retirement. It would reduce consumption in the short to medium run, but
would increase investment over the long run. (iii) Increasing contribution rates: this leads to
distortionary supply-side effects for labor, which combined with a negative aggregate
demand on real disposable income, depresses real activity in both the short and long run.
We assess how the policies compare in attaining the twin goals of strong, sustainable, and
balanced growth and fiscal stability (i.e., stabilizing the debt-to-GDP ratio against rising
pension entitlements). The key results show that increasing the retirement age has the largest
impact on growth compared to reducing benefits, while increasing contribution rates as
approximated by an increase in taxes on labor income has the least favorable effect on
output. Besides boosting domestic demand in the short run, lengthening working lives of
employees reduces the pressure on governments to cut pension benefits significantly or to
2
Pension and health entitlements already represent over a third of total spending in G-7 countries. Over the next
20 years, the net present value of pension spending increases alone is estimated at about 8 percent of GDP in
advanced and emerging countries. See IMF (2010a).
3
See Blanchard and Cottarelli (2010), ―The great false choice, stimulus or austerity,‖ Financial Times,
August 11 on attaining the twin goals of strong, sustainable, and balanced growth and fiscal stability.
4
Our analysis is undertaken relative to a ‗baseline‘ scenario, constructed such that an elevated long-run public-
debt-to-GDP ratio is reached in steady state based on staff‘s pension spending projections into the distant future.
5
raise payroll and labor income taxes. Reducing such benefits can lead to an increase in
private savings and an unwarranted weakening of a fragile domestic demand in the short run,
while raising taxes can distort incentives to supply labor. We also found that if regions
cooperate in pursuing fiscal reform, the impact will be greater than if only one or some of the
regions in the world undertake reform separately. In all, early and resolute action to reduce
future age-related spending or finance the spending could improve fiscal sustainability over
the medium run, significantly more if such reforms are enacted in a cooperative fashion.
The paper is organized as follows. Section II provides a background on past and projected
age-related pension outlays and discusses the reform options considered to offset them.
Section III provides a brief overview of the GIMF model while focusing on the details
pertinent to this exercise. Section IV presents the effects of the three different reforms taken
by each country at a time, then in all regions simultaneously. The global scenario highlights
the compounding effect of the reforms and their impact on external variables through trade
and (predominantly) financial spillover channels. Section V assesses in further detail the
possible reasons for the size of the macroeconomic impacts based on a sensitivity analysis
around the main results. Section VI concludes.
I. PENSION SPENDING TRENDS, THEORY AND EXISTING STUDIES
Age-related spending has been the main driver of current public spending increases over the
past two decades. These trends are expected to continue in the coming years for both
advanced and emerging economies pointing to needed entitlement reforms. Old-age
dependency ratios, which are already large in the advanced economies, particularly European
countries and Japan, are projected to double between 2009 and 2050, putting enormous
pressures on pension systems.
5
Furthermore, relatively high gross replacement rate of
pensions relative to average wages has also contributed to large pension spending and could
undermine the viability of pension system over the long run. In terms of contribution rates,
taxes on earnings are already high in a number of countries but in others, there is room for
raising payroll contribution rates (see IMF, 2010a). In this section, we examine the current
and projected pension spending and provide a short review and assessment of public pension
reform measures based on theory and existing studies.
A. Current and Projected Public Pension Spending
Within the advanced G-20 countries, pension outlays have risen by 1¼ percentage points of
GDP since 1990. Increases have been especially large for pensions in Japan and Korea in the
5
Based on the European Commission‘s Ageing Report (2009), it is evident that the anticipated demographic
transition will affect future pensions significantly. Despite recent pension reforms which may have strengthened
the counterbalancing impact of other factors, considerable spending pressures remain, however, in light of the
anticipated increase in dependency ratios.
6
past decade. Strong demographic factors were an important catalyst behind the increase in
pension outlays in France, Germany, Italy, and Japan where pension spending has already
surpassed 20 percent of total public spending. Looking forward, these trends are expected to
continue in all economies. Given the strong demographic pressures on these outlays,
reducing this spending would be difficult. A more realistic, if conservative, goal followed in
this paper would aim at stabilizing spending-to-GDP ratios—which would still require
significant structural reforms (IMF, 2010b).
Figure 1 shows pension spending projected to increase by an average of 1 percentage point
of GDP over the next 20 years. Large increases are projected in advanced countries that have
not substantially reformed their traditional pay-as-you-go systems, but in other advanced
economies, the increase would be less marked due to the projected impact of already
legislated reforms (IMF 2010a, Appendices IV and V). Adjustment needs may well be larger,
though, as the projections assume that these reforms will not be reversed, even when they
involve large cuts in replacement rates such as in Italy and Japan. Among emerging
economies, those with relatively high spending in 2010 are projected to experience the
steepest increase in outlays over the next 20 years. In other countries with currently low
coverage such as China and India, the projected increase is much less severe, but could rise
more rapidly if the system expands to cover a larger share of the population. Moreover,
beyond 2030, emerging economies are expected to experience a faster pace of aging
compared to the advanced economies.
Figure 1. Change in Public Pension Expenditures, 2010–30
(In percent of GDP)
-1
0
1
2
3
4
5
6
Luxembourg
Greece
Cyprus
Belgium
Finland
Slovenia
Norway
Netherlands
Iceland
New Zealand
Spain
Korea
Canada
Germany
Ireland
Australia
Denmark
Austria
United States
Malta
United Kingdom
Italy
Slovakia
France
Portugal
Czech Republic
Sweden
Japan
In percent of GDP
Advanced Economies
PPP Average: 1.1
-2
-1
0
1
2
3
4
5
6
7
Ukraine
Turkey
Russia
Mexico
Romania
Lithuania
Malaysia
Saudi Arabia
Brazil
Latvia
South Africa
Philippines
India
Pakistan
Argentina
Indonesia
China
Hungary
Bulgaria
Estonia
Poland
In percent of GDP
Emerging Market Economies
PPP Average: 1.1
Sources: Country authorities; EC (2009); OECD (2009); ILO (2010); and IMF staff estimates.
7
B. Theory and Existing Studies
A large body of research exists on fiscal consolidation in the face of demographic shifts and
the impact of public pension system reforms on growth and public debt dynamics. We focus
on three key reforms: (i) raising the retirement age, (ii) reducing benefits, and (iii) increasing
contribution rates.
Raising retirement age raises participation in the labor force beyond a certain age and slows
down the increase in the pension system dependency ratio.
6
This leads to a reduction in
transfer payments to pensioners, an increase in contributions and an increase in tax revenues
through increased income and consumption, therefore leading to higher public savings. In the
long run, output rises as firms demand more capital inputs to work with higher labor. Before
retirement, forward-looking consumers who will be providing more labor services and face a
shorter retirement period reduce their saving and increase consumption in anticipation of
increased future income. Earning income over a longer working period makes up for this
initial drop in savings and has a positive impact on their stock of wealth in the long run.
Reducing benefits has been the policy choice of several countries, with cuts of nearly
20 percent or more set to occur within the next 20 years. Benefits could be reduced by
modifying the base used to calculate benefits, modifying indexation rules, or taxing pensions.
Rules that link benefits to demographic and economic variables to maintain actuarial balance
could also lead to benefit cuts.
7
Based on theory, there are strong incentives for working
households to increase their savings in the face of announced decline in replacement rates of
the pension regime, in order to avoid a sharp reduction in their income and consumption after
retirement. A higher saving rate leads to stronger capital accumulation and an improvement
in the net asset position of the country, but the effect on short- and medium-run consumption
levels can be negative.
Increasing contribution rates needs to be assessed along with potential changes in the tax
rate on labor income, since it is their combination that determines the effective marginal and
average tax rates that are likely to affect decisions about labor participation and hours
worked.
8
These incentive effects of social contributions, however, might be less marked if
6
In this paper, raising the retirement age by a year induces individuals to effectively work one year longer on
average. This would cover those that retire at an age with full retirement benefits and those that choose early
retirement. It assumes that replacement rates stay constant, implying a decrease in total lifetime pension benefits
(see Section IV).
7
For example, in Japan, ‗macro indexing‘ is achieved by reducing pensionable earnings and benefits by the rate
of decrease in the number of contributors and increase in life expectancy at age 65. In Canada, benefits are
required to be reduced, or contributions increased, to address long-term actuarial imbalances.
8
A richer menu of taxation would allow a further distinction between personal income tax levied on labor and
social security contributions paid by workers and employees. Changing payroll tax on workers vis-à-vis
changing personal income tax can have different distortionary effects. We abstract from these details here.
8
their payment is seen as implying increased benefit entitlement. Overall, offsetting the
spending pressures from the pay-as-you-go regime based on an increase in contribution rates
usually reduces the potential output of the economy by distorting labor supply. A demand
effect through households‘ lower disposable income also adds to the negative impact of this
option.
Empirical findings on the other hand appear to be inconclusive, reflecting perhaps country-
specific and empirical methodology differences. Botman and Kumar (2007) look at age-
related reforms, but focused exclusively on the European Union (with Germany as an
example). They also analyze the impact of broader structural reforms, such as increasing
labor participation, product market liberalization, and higher R&D to help increase
productivity and find positive output effects in the short run. Nickel and others (2008) show
that timely tax-cut measures can moderate the adverse effect on consumption (and encourage
labor supply) of future announcement of cuts in pension benefits, and lower public debt;
while increasing retirement age, without cutting pension benefits, fails to lower public debt.
9
In contrast, Cournède and Gonand (2006) and Andersen (2008a, b) find that raising
retirement age is optimal based on a likely boost in growth and improved public debt
dynamics. Real GDP growth is stronger when rebalancing the pension regime by increasing
the retirement age (and containing spending) rather than lowering replacement rates and
raising taxes. Barrell and others (2009) have also demonstrated an improvement in public
debt dynamics following an increase in effective working life (for the United Kingdom and
the euro area countries taken together). As workers know that they will work longer, they
save less now and increase their consumption ahead of the prospective income increase. Over
the long run, labor and capital rise leading to an increase in GDP. Importantly, under constant
tax rates and spending, increasing the pension age would result in reduced budget deficits
and public indebtedness in European economies, on average.
II. THE METHODOLOGY FOR MODELING PUBLIC PENSION REFORMS
This section provides a summary of the methodology followed in addressing the aging-
related reforms of public finances, while strictly focusing on the main features of the model‘s
sectors (households, firms, and government) and parameters which have a direct and relevant
impact on our analysis.
10
Caveats and areas for future work remain given that the model, like
most others cannot reflect all complexities that can influence the effect of the considered
reform policies.
9
While the finding that a tax decrease aimed at offsetting the age-related fiscal consolidation effect on
consumption and labor is sensible, it should be assessed in light of the current situation with mounting fears of
unstable debt dynamics.
10
Appendix 1 provides further explanation of the sectors and the optimization involved. For a fuller exposition
of GIMF‘s properties and calibration, see Kumhof and others, KLMM (2010).
9
A. Overview of the Model’s Key Features
We use GIMF, a dynamic stochastic general equilibrium model widely used inside the Fund,
as a framework for analyzing the short- and long-run effects of the planned pension policy
actions. Key in analyzing the positive aspects of achieving fiscal sustainability in the face of
aging as well as the normative aspects of adjusting public policies to changes in
demographics, is GIMF‘s underlying overlapping generations‘ and finite horizons‘ structure.
It produces meaningful medium- and long-run crowding-out effects of government debt and
captures important life cycle income patterns, including age-dependent labor productivity.
Moreover, labor and capital markets are endogenous—the first allowing labor income taxes
to have distortionary effects and the latter providing an important channel through which
government debt crowds out economic activity. As such, a realistic supply side enables us to
consider the impact of public pension reforms on investment decisions.
The multi-country structure of GIMF allows an analysis of global interdependence and
spillover effects. The world in this model consists of five regions, the United States (US), the
euro area (EU), Japan (JA), emerging Asia (AS),
11
and remaining countries (RC). The
regions trade with each other at the levels of intermediate and final goods, with a matrix of
bilateral trade flows based on recent historical averages. International asset trade is limited to
nominally non-contingent bonds denominated in U.S. dollars. Importantly, the link between
regions through international financial markets provides the key channel for spillover effect
of aging-related spending at a global level while adding realism to the macro outlook and the
impact of policy response. The financial spillover effect is likely to dominate the trade
channel because of the compounding effects of cooperative public pension reform on real
interest rates, which in turn affect the cost of borrowing and overall debt dynamics
(Section IV.C.).
To emphasize the potential interaction role of fiscal and monetary policies, GIMF combines
sufficient non-Ricardian features with a number of nominal and real adjustment costs, such
that short-run dynamics of the model would be determined by the interaction of both of these
policies while longer-run dynamics are influenced mainly by fiscal policy. This combination
is missing from other new-open-economy macroeconomic models and fiscal models,
including the IMF‘s Global Fiscal Model (GFM).
There are three groups of agents and sectors in the model: households, firms, and the
government.
11
AS comprises China, Hong Kong S.A.R. of China, India, Indonesia, Korea, Malaysia, Philippines, Singapore,
and Thailand. For pension projections below a subset of those countries is considered, comprising China, India,
Indonesia, and Korea.
10
In the households sector, three parameters of interest determine the degree of non-Ricardian
behavior of agents:
,
and
.
is the share of liquidity-constrained households (LIQ) in
the economy, without access to financial markets, that are limited to consuming their
after-tax income in every period. The size of this group, assumed to differ significantly
across economic regions, can be crucial to the analysis of the effects of the labor income tax
reform measure for instance, as will be seen later. The remainder of the households, are
overlapping generations (OLG) households, who are fully optimizing agents. Each of these
agents faces a constant probability of death
1
in each period, which implies an average
planning horizon of
1/ 1
. In addition to the probability of death, households also
experience labor productivity (and hence labor income) that declines at a constant rate
over their lifetimes. Life cycle income adds another powerful channel through which fiscal
policies have non-Ricardian effects, as this along with
(probability of survival) produce a
high degree of myopia. Households of both types are subject to labor income, consumption
and lump-sum taxes and the presence of these taxes along with transfers and government
spending (see fiscal policy block in Appendix 1) allows us to relate the pension-related tax
and expenditure reforms to specific model‘s parameters and variables.
In order to represent an increase in retirement age in GIMF, we rely on two parameters in
particular:
(which corresponds to a decline in labor productivity over an average working
life—it defines agents‘ ―income profile‖) and
N
(which is an index of the population size
assumed to correspond to population of the work force age, ages 15 to 64).
Firms are managed in accordance with the preferences of their owners, myopic
OLG
households, and they therefore also have finite-planning horizons.
Government’s intertemporal budget constraint is discussed in Appendix 1. We suffice by
highlighting the role of fiscal policy in stabilizing deficits and the business cycle, through a
typical fiscal rule. The latter stabilizes the government deficit-to-GDP ratio at a long-run
target (structural) level, which rules out default and fiscal dominance (dynamic stability).
It also stabilizes the business cycle by letting the deficit fall with the output gap. Finally,
monetary policy in the model is based on an inflation-forecast-based interest rate reaction
function in which the central bank sets interest rates in order to stabilize inflation at an
announced target level.
B. Quantifying Public Pension Reforms
For all public pension reform measures, we use 2014 as the starting point for our benchmark
scenario. This is near the end of the current version of the IMF‘s World Economic Outlook,
when most economies are forecasted to have returned to stable output gaps around zero, and
inflation close to their target levels. Starting from such a position, pension reforms are likely
to generate short-run increases in output leading to a monetary policy reaction (as will be
11
seen in Figure 15 later). However, we also assess how the results might change if monetary
policy remained accommodative for a year or two, as a pre-announced and conscious
decision to accommodate a stimulative fiscal policy measure. This is dealt with in the
sensitivity analysis Section V. The results below show that delaying a monetary policy action
would boost short-run consumption and real GDP (considerably more relative to the
benchmark), and public finances improve as government deficits decline faster in light of
lower debt service payments. When normal conduct of monetary policy returns, there is the
usual dampening effect on demand in the medium run (Section V and Appendix 2).
We consider differentiated retirement age increases which are sufficient to stabilize pension
spending as a share of GDP at its 2014 level, over the next three to four decades. Differences
in necessary retirement age increases stem from different baseline projections of the pension
gap (size determined exogenously based on country-specific demographics and pension
parameters) across the five regions, implying different consolidation needs such that the
resulting debt trajectory as a share of GDP is stabilized. Based on staff estimates of the
projected pension spending in the five regions, the required ‗pension age extension‘ is shown
in Table 1. The estimates suggest, for instance, that Japan would not require additional
reforms.
Table 1. Required Pension Age Extensions across the Regions
US
JA
AS
EU
RC
(Number of years)
2015–2030
2.5
...
1.0
1.5
3.0
2030–2050
0.0
...
0.5
0.5
1.5
Source: IMF (2010a) and staff estimates.
In line with the change in the lifetime income horizon, we implement a two-year extension of
working lives on average, globally, by lengthening agents‘ income profile (
), and assuming
an increase in the working-age population (N).
12
The income profile is assumed to increase
immediately following the start of the reform, with the labor force gradually increasing over
the next 15 years. The increase in the income profile is consistent with a decrease of private
saving as a percent of GDP, as found in studies focused on European countries.
13
Both
measures are consistent with a gradual phase in of increases in retirement age as well as
12
In GIMF, by assuming that the working-age population spans the ages 15–64, a two-year extension is roughly
equivalent to a 4 percent increase in this population size. This can be made more moderate, since cohorts
usually congregate near the middle of the age distribution (or in the case of aging Western societies, between
40 to 60 years old).
13
See, in particular, Khoman and Weale (2008) and Barrell and others (2009).
12
evidence which suggests that agents react by delaying retirement (several years ahead of the
change itself) thereby leading to an increase in labor supply over their lifetime horizon.
The two other considered reform options, a reduction in pension benefit payments and an
increase in contribution rates, are simply modeled as non-distortionary lump-sum transfers
to all households and an increase in the labor income tax rate, respectively.
C. Caveats and Qualifications
The fiscal block of the GIMF model does not allow for an explicit breakdown of working-
age and retired population,
14
nor does it feature an elaborate pay-as-you-go pension regime.
In light of this, it is not possible to interpret a rising dependency ratio as due to reduced
fertility or increased longevity. Other issues like labor force perception of the pension system
as a tax-and-transfer system versus an insurance mechanism, and movement to more
actuarially-based public programmes
15
may affect individuals‘ saving and labor supply
behavior differently, influencing in turn the normative assessment of reform—we leave these
interesting extensions for future work. But unlike other large simulation models dealing with
full-blown demographics and pension systems which may complicate the interpretation of
results, GIMF focuses on the dynamics and long-run equilibrium of the main variables in a
transparent way, and these dynamics and equilibria can be changed by modifying a few
essential parameters. The structure is flexible enough to compensate for missing households
who ‗really‘ retire by treating an extension of working lives with regard to agents‘ income
profile coupled with a potential increase in ‗working age‘ population.
Another caveat lies in examining the contribution rate hike scenario which we proxy by an
increase in the labor income tax rate. This can be seen as a lower bound on this issue, since
we have only captured the effect on the labor supply decision and not the direct effects on
labor demand from higher contributions by employers. The decline in pension payments to
retirees is then captured through lower pension transfers and pension deficits.
D. Calibration
Relevant steady state ratios and parameters of particular importance for this exercise are
discussed in Appendix 1, with a brief summary of the important ratios provided in Appendix
Table 1. The model is calibrated to reflect key macro features in the five regional blocs
(including key expenditure ratios of consumption, government, investment, net exports, and
14
Changes in the population structure are not captured in a detailed way in GIMF compared to models with a
richer cohort distribution. A change in participation rate of older workers is calibrated in GIMF through some
relative measures of working-age population, as discussed.
15
This is a purely design question of social security programmes, to emulate a private retirement saving plan.
See Disney (2005).
13
factor incomes) as well as key fiscal variables reflecting the fiscal structure of the regional
blocs (revenues and spending, net debt- and deficit-to-GDP ratios). More detailed calibration
tables are presented in KLMM (2010). Unless otherwise stated, similar behavioral parameter
values apply to all regions and are based on microeconomic evidence. We use an annual
version of the model because the critical pension-related fiscal issues stressed are of a
medium- to long-run nature.
Calibrated government debt-to-GDP ratios are based on 2014 net debt projections for the
five regions from the IMF's World Economic Outlook Update, February 2010 (IMF, 2010c)
but have taken into account the mounting pressure of pension gaps. The real global growth
rate is 2.5 percent, the global population growth is 0.5 percent, and the long-run global real
interest rate is 4.0 percent.
III. RESULTS: PUBLIC PENSION REFORMS
A. Baseline
The baseline scenario is based on the IMF‘s February 2010 World Economic Outlook for
public debt in G-20 countries, up to 2014–15, close to what the May 2010 update shows. It is
also based on Fiscal Affairs Department staff‘s projections of public pension spending and
primary fiscal balances over the next four decades, which translate in a very distant future
into higher steady state debt-to-GDP ratios and a higher world real interest rate than usually
used in GIMF.
B. Region-by-Region Benchmark Public Pension Reform Scenarios
This section discusses the effect of reforms undertaken in each region on its own, beginning
in 2014. While similar behavioral parameter values apply to all regions, country-specific
variation in the demographics is reflected in the size and pace of the adjustment as described
in Table 1 above. Such an individual action is then compared to a cooperative action (albeit
still allowing for country differences) in Section IV.C.
Three reform options relating to pay-as-you-go public pension systems are assessed. They
are broadly equivalent in terms of their fiscal impact, all of them being broadly sufficient to
offset the projected increase in pension spending over the long run, excluding their possible
and distinct effect on growth. With the economic recovery still under way, it is important to
assess the short- and medium-run impact of such reforms on the pace of activity as well as
their budgetary impact. Results show that the type of reform matters: increasing the
retirement age has the largest positive impact on real GDP, while increasing contribution
rates has the least favorable effect on output.
16
16
As will be seen, while pension gaps are reduced steadily with the adoption of any of the planned pension
reforms discussed below, we observe that along the transition path to higher long-run real GDP, the short-run
(continued…)
14
Increasing retirement age
To anticipate the results, increases in the retirement age are the most effective tool: on
average across regions, raising the retirement age by two years on average
17
would raise
GDP by almost 1 percent in the short to medium run and 4¼ percent by 2050 above the
baseline scenario. It reduces the debt-to-GDP ratio by 30 percentage points over the same
period.
Figure 2 illustrates the effect on the United States of an increase in retirement age in the
United States alone, while first keeping public pension spending (transfers) constant. As
expected, a delay in the retirement age boosts labor supply and labor income. Agents reduce
their saving and their demand for assets during working years, while increasing consumption.
Future earning incomes over a longer working period are higher and are brought forward
through higher consumption by optimizing agents. The private saving rate as a ratio of GDP
declines immediately by 0.2 percentage points, while consumption rises above baseline by
close to 2 percentage points in the short run, preceding the increase in real GDP.
18
In the
short run, the increase in real GDP is 0.75 percent above baseline in period 2, and public
finances improve slightly (the debt-to-GDP ratio is only 4 percentage points below baseline
after 20 periods), a direct result of increased tax revenue collected on income and
consumption. Considering Figure 2 alone, despite providing a partial analysis of the overall
reform scenario (no implied transfer reductions are modeled thus far), it gives a clear
interpretation of the boost in consumption as stemming from an increase in lifetime income
horizon (the effect of which is brought forward) and working age population. A cut in
benefits as embedded in Figure 3, will dampen this effect.
Figure 3 depicts the complete analysis of this reform as the concurrent fiscal consolidation
implied by the cuts in public pension spending, as the number of years over which pensions
are paid are reduced, is added to Figure 2. The budget deficit improves as a result by close to
3 percentage points of GDP after 30 years and settles around 2.2 percentage points once
reached in the long run, given the target we impose in the distant future. Equivalently, the
debt-to-GDP ratio declines by roughly 43 percentage points in the long run—more than
tenfold the improvement shown in Figure 2. At the same time, a lower government debt
which is perceived as a decline in
OLG
agents‘ net wealth along with the decline in transfer
dynamics of the macroeconomic adjustment differ substantially. Size, speed, and timing of the fiscal
consolidation plan, alongside the incentive effect of reforms on consumption and savings all play a role.
17
The two-year average reflects variation across regions in the increase in the retirement age needed to stabilize
the debt-to-GDP ratio against rising pension entitlements.
18
Such a strong response in labor supply is assumed to be absorbed quickly with minimal effect on the
unemployment rate and the associated unemployment benefits to be paid by the government.
15
payments (with a more pronounced effect on
LIQ
agents) both work to depress consumption
markedly, which now only rises modestly above baseline, compared to Figure 2.
In the short run, there is a tightening of interest rates (by 80 basis points in year 4) in
response to inflationary pressure emanating from a short-run increase in domestic demand.
As the demand pressures continue from the stimulative increase in labor supply, the
monetary authority maintains a tight stance for a long period. This interest rate effect
dampens domestic demand in the short run. But in the medium to long run, investment is
boosted as real interest rates fall in response to the fiscal consolidation, leading to visible
improvement in output. In addition, output rises with the increase in labor supply (and a fall
in marginal cost from the falling real wage) which in turn attracts more capital. This rises
marginally less than labor supply
19
and output continues on an upward trend, reaching over
3.5 percent above baseline in the long run.
Discussing next the external variables, if we only focus on the effect of the increase in the
retirement age without the fiscal adjustment (Figure 2), the United States experiences an
appreciation of the real exchange rate, and, therefore, a deterioration in the trade balance and
the current account. Here, the saving-investment perspective predominates, as real wages
decline (as more labor is supplied), the higher return on capital attracts capital inflows and
leads to a current account deficit which needs to be closed through a depreciation in the
exchange rate. However, the effect on external variables is dominated by the fiscal
consolidation that is occurring simultaneously (Figure 3). Now, lower real interest rates from
increased world saving crowds in investment in external assets, leading to an accumulation of
net foreign assets. In the very long run, with declining interest payments to foreigners,
current balances are above baseline which means that the real effective exchange rate begins
appreciating again. However, relative to the baseline, the real effective exchange rate has
depreciated, albeit much less in 50 years into the future, relative to only 30.
Figures 4 to 6 show the same ‗package‘ of reforms (akin to Figure 3) being undertaken by
each of the three other regions facing notable challenges to their pension systems—the euro
area, emerging Asia, and remaining countries (recall, in the case of Japan, no pension age
extension was needed). Although the quantitative results are different, the story behind each
scenario is intrinsically the same as that of the United States, with some qualification.
For the euro area (Figure 4), results are qualitatively similar to the United States, but there is
a smaller required pension-age increase to attain given budgetary saving (this is primarily
due to the fact that in the euro area a pensioner receives on average larger benefits), more
rigid prices and a more aggressive monetary rule, leading to a weaker consumption profile
19
Private capital rises less than employment due to an expected marginal upward pressure on interest rates from
increased demand which may lead to a partial crowding-out.
16
relative to the United States, in the short run.
20
Over the long run, consumption improves by
more as pension transfers are cut more aggressively in the later periods, bringing with them a
larger drop in interest rates, and therefore lower debt level (close to 47 percentage points
below baseline). Driven by higher domestic demand, real GDP rises 5.8 percent above
baseline.
An exception is emerging Asia, because it pursues a fixed nominal exchange rate vis-à-vis
the United States, instead of inflation targeting. In this case (Figure 5), there is a depreciation
of the real exchange rate, as occurs in the cases of other regions such as the United States
(Figure 3) or the euro area (Figure 4). But emerging Asia‘s nominal exchange rate peg
constrains them to importing their short-run interest rate profile from the United States (as
uncovered interest rate parity holds in GIMF). In order that the real effective exchange rate in
emerging Asia depreciates in the long run, there will be downward pressure on domestic
prices, resulting in sustained disinflation, relative to the baseline scenario. Here, GIMF may
overstate the actual inflation and interest rate dynamics in emerging Asia, as there is no role
for capital or credit controls, which may play some role in the actual conduct of exchange
rate policy.
This story related to the conduct of monetary policy in emerging Asia will also hold in the
other two public pension reform options discussed below. A policy aiming at greater
exchange rate flexibility in emerging Asia over the medium run, whether through a nominal
appreciation or through higher inflation is not assessed here.
20
A stylized Taylor-type interest rate reaction function is adopted, where the central bank adjusts the policy rate
on the basis of the deviation of inflation from its target to stabilize inflation at a pre-specified target level. The
rule matters in the response to offset inflationary pressures arising from a boost in domestic activity. A
persistent underlying inflation process with monetary policy being tightened as a result would put downward
pressure on growth. Reduced price rigidities can mitigate this effect by effectively speeding the response of
inflation and shortening the period of tighter policy. Delaying the response of monetary policy will also boost
short-run consumption and real GDP (for further clarification, please see Section V and Appendix 2).
Moreover, while the size and time profile of the fiscal adjustment play a role, initial conditions and market
responses to the adjustment plans are crucial.
17
Figure 2. Increase in the Retirement Age in the United States
(Excluding Fiscal Consolidation)
18
Figure 3. Increase in the Retirement Age in the United States
19
Figure 4. Increase in the Retirement Age in the Euro Area
20
Figure 5. Increase in the Retirement Age in Emerging Asia
21
Figure 6. Increase in the Retirement Age in the Remaining Countries Block
22
Reducing benefits
21
This reform generates rewards over time following the transitory short-run initial costs of
fiscal tightening on aggregate demand. Consider if the decline in benefit payments (and the
consequent decline in government debt) occurs only in the United States. Figure 7 shows the
simulated effects of reduced government debt-to-GDP ratio brought about by decreases in
pension benefit payments (that behave in GIMF as non-distortionary lump-sum transfers) to
reverse past promises of enlarged public pension spending. Although consumption drops by
about 1 percent below baseline in the short run, this is largely outweighed by the persistent
benefits of lower real interest rates and higher real GDP—over time, real GDP rises and
settles at a higher level in the long run, almost 0.5 percent above the baseline scenario.
World real interest rates decline, moderately, beginning in period 10, before they hit a trough
close to -0.4 percentage points below the baseline after 40 years.
22
Such an effect is
transmitted to the global economy with all countries experiencing a boom in investment,
varying between 0.5 and 2 percent in the euro area and emerging Asia, respectively, and a
permanent expansion in real GDP, varying between 0.3 to 0.5 percent. In sum, the U.S.
policy scenario generates a positive and large effect in other regions as long-run real interest
rates are equalized internationally to a lower level and capital investment is boosted.
So, following a reduction in pension benefit payments, U.S. domestic demand (consumption
and investment) decreases for a rather prolonged period of time, while real GDP experiences
an uptick for a brief period,
23
buoyed by improved external balances, but decreases
moderately thereafter before increasing and settling at higher levels in the long run.
Consumption declines in light of the non-Ricardian nature of the model whereby a fiscal
consolidation reduces the net wealth of
OLG
agents (as the value of taxes for which they are
now expected to be responsible has increased, if taxes were to be used as an instrument). For
a given marginal propensity to consume, these reductions in (human) wealth lead to a
reduction in consumption, accompanied by a decline in real interest rates. During the initial
phase, real interest rates are predominantly driven by the monetary policy response to excess
21
The average reduction in benefit payments is high exceeding 20 percent as driven by large projected pension
spending in the remaining countries block. This differs across regions depending on the savings needed to
stabilize the debt-to-GDP ratio against rising pension entitlements.
22
Potentially reduced sovereign risk premia associated with favorable market responses to improved public
finances, as a result of pension reform, are not taken into account.
23
The short-lived increase in GDP occurs since demand for foreign goods (imports) falls rapidly in response to
the quick movement in the real exchange rate (a property of the standard risk-adjusted uncovered interest rate
parity condition determining exchange rate), but domestic demand falls more slowly, as it is driven by the
slower decline from the fiscal consolidation. In other words, imports fall more quickly than consumption and
investment, and this, by simple accounting, leads to higher real GDP.
23
supply in the economy and deviation of inflation from target. Externally, reduced import
demand (part of the consumption demand decline is absorbed by trading partners) leads to
improvement in trade balances and a real depreciation.
Over the longer run, real GDP increases relative to the baseline. Higher fiscal saving leads to
an increase in both in U.S. and world savings, given the size of the U.S. economy. Real
interest rates decline by close to 40 basis points in order to re-equilibrate world saving and
investment. The non-Ricardian
OLG
structure of the model and the endogenous capital
formation provide the channels through which government debt crowds in investment in U.S.
physical capital, so that real output increases. Moreover, agents‘ decreased investment in
government debt instruments frees up resources to other forms of investment, including
foreign assets. This implies that current balances improve subsequently necessitating a real
appreciation in the exchange rate, which only comes gradually.
In other regions (euro area, emerging Asia, and the remaining countries block) which
undertake similar reforms, the effects are similar (Figures 8 to 10). However, the spillover
effects are different as they are driven by their responsiveness to movements in the world real
interest rate. For instance, the spillover effects of reforms initiated by a large economic
region (i.e., the United States or the euro area) on other regions‘ real GDP is four times the
spillover effect if a smaller region (i.e., emerging Asia constitutes is calibrated in GIMF to be
13 percent of world nominal GDP, versus 27 percent for the United States) undertakes
reforms, since a smaller region will have less of a long-run impact on world interest rates,
and by extension on investment and output on those regions which do not undergo reform.
24
Figure 7. Reducing Pension Benefits in the United States
25
Figure 8. Reducing Pension Benefits in the Euro Area
26
Figure 9. Reducing Pension Benefits in Emerging Asia
27
Figure 10. Reducing Pension Benefits in the Remaining Countries Block
28
Raising contribution rates
24
We proxy the increase in contribution rates in GIMF, by considering an increase in the labor
income tax rate, in this case, the United States only (Figure 11). Consequently, there is a
decline in the supply of labor with negative effects on actual and potential output. Besides
this supply-side effect, this policy measure affects the demand side of the economy indirectly
through a decline in households‘ real disposable income—this income and wealth effect is an
important channel given the myopic nature of both
LIQ
and
OLG
agents (in light of the
planning horizon parameter and
finite remaining working life of 20 years on average).
Given that the
LIQ
households consume at most their after-tax current income, the size of
this group, which is assumed to be significantly different across economic regions, is critical
for the analysis of this labor income tax measure. Moreover, a decline in potential output is
likely to exert upward pressure on inflation. In all, the effect on U.S. real GDP is notably
worse than in the benefit-reduction scenario in the short run, and even in the long run. This
should not be surprising, since GIMF, like most models in the literature, find that the
multiplier effect of a change in the labor income tax rate is higher than an equivalent shift in
a lump-sum transfer, such as a pension benefit cut (Coenen and others, 2010).
Therefore, the results of this form of public pension reform are similar to those found under a
cut in pension benefits, but the distortionary nature of this reform means the short-run losses
are more significant—real GDP declines by about 0.7 percent below baseline by period 10.
The negative effect of distortionary taxes on potential output also means significant losses in
the long run. Also, the consequent decrease in the real world interest rate does not play as
effective a role in raising real GDP in the long run as in scenario 2 above—real GDP remains
close to 0.4 percent below baseline versus an increase of 0.4 percent when cuts in pension
benefits are the fiscal measure of choice.
Once again, these results also hold in the other three regions, the euro area, emerging Asia,
and the remaining countries block (Figures 12–14).
24
On average, the necessary increase in contribution rates is 2¾ percentage points or over 10 percent. Again,
this differs across regions depending on the requirement to stabilize the debt ratio against rising pension
entitlements.
29
Figure 11. Raising Contribution Rates in the United States
30
Figure 12. Raising Contribution Rates in the Euro Area
31
Figure 13. Raising Contribution Rates in Emerging Asia
32
Figure 14. Raising Contribution Rates in the Remaining Countries Block
33
C. Benchmark Global Scenario—Simultaneous Reforms in all Regions
A cooperative strategy to pursuing fiscal reform has a larger impact on output and fiscal
sustainability than if regions undertake reform alone. The cooperative benefit is greater than
the sum of individual country/region benefits. The magnification effect of global reforms on
key variables is driven by the significantly stronger decline in the world real interest rate
corresponding to larger compounding effect on word savings under the cooperative strategy.
Thus far, we have only considered reforms in each region of the world in isolation. While it
is in each country‘s interest to pursue reforms regardless of what other countries or regions
do, there can be a clear advantage from promoting global policy cooperation. In the cases of
individual action, the effects of the policy measure will often leak abroad, which, while
benefitting other regions, reduces the potential impact domestically. Countries can simply
delay reforms and free ride on adjustments undertaken elsewhere.
25
Faced with a common
and an unavoidable demographic shift and a future of possibly muted growth and high
unemployment, cooperative action for public pension reform among all regions can be key.
Cooperation can buttress the twin goals of growth and fiscal stability by stabilizing the
debt-to-GDP ratio against rising pension entitlements. It is expected that the world real
interest rate will over time change by more than when an individual region engages in reform
alone, which leads to a larger effect on capital accumulation and potential and actual output
levels in the long run.
Figure 15 clearly shows those benefits by looking at the effects on the United States when it
undertakes an increase in retirement age pension reform alone (the right column), versus a
cooperative global effort (the left column), both beginning in 2014. Under the cooperative
case, real GDP is 50 percent higher in the long run. A cooperative action results in an interest
rate decline that is about five times that under an individual action. As a result, a permanent
expansion in real GDP worldwide (average over the five regions) in the order of 7.2 percent
above baseline follows (Table 2)—this table also shows that this is about 40 percent larger
than the sum of benefits from individual country reforms.
The magnified effects of simultaneous public pension reforms by all regions, relative to
reform by each region alone, are further illustrated in Table 2 (the increase in retirement age
case). It highlights the effect of reform on real GDP, consumption, the real interest rate, and
the government debt-to-GDP ratio in the five regions, which are clearly larger in every case
under a cooperative policy action (the first set of rows) versus the individual action (the
second through fifth set of rows). While all regions benefit relatively more from a
25
Botman and Kumar (2007) analyze the macroeconomic effects of a coordinated policy response to global
demographic pressures in a four-country version of GFM (Germany, the rest of the euro area, the United States,
and the rest of the world) and show clear benefits from a cooperative fiscal adjustment. Freedman and others
(2009a,b) also show the benefits of a coordinated approach to fiscal activism in a GIMF setting.
34
cooperative action, the euro area, a large and relatively less open region benefits relatively
less than a smaller and more open emerging Asia (40 percent and 110 percent improvement,
respectively).
Figures 16 and 17 make a similar point for the other two policy options. In the case of an
increase in labor income taxes to finance the pension gap, real GDP no longer falls relative
to the baseline scenario in the long run. However, real GDP in the United States in this case
increases by only 0.5 percent, while rising above baseline by more than three times that
amount in the long run (1.6 percent) if the fiscal measure was a decrease in pension benefits.
Tables 3 and 4 summarize the results of these corresponding reforms if carried out
individually (along with their spillover effects into other regions) or globally.
To sum up, promoting a global cooperative increase in retirement age appears to yield the
largest impact on activity—the relative improvement in real GDP worldwide is four and over
10 times larger than under reform options 2 and 3.
In terms of external balances, the global cooperative scenario yields a weaker external
balance in each country, and corresponding less accumulation of net foreign assets. The
current account improves by less under a cooperative action than when a policy is taken by
each country or region on its own. Under a global scenario where only one country does not
reform, improvements in the current account balances of the reforming countries would be
reflected in a deteriorating balance of the non-reformer—private saving declines or
consumption rises due to an appreciating real exchange rate and an investment rises due to
lower interest rates.
As for improving public finances, stabilizing the GDP share of age-related (pension)
expenditures leads to a sizable decline in the debt-to-GDP ratio. Early and resolute action to
reduce future age-related spending could significantly improve fiscal sustainability in several
countries over the medium run and more so if such reforms are again enacted in a
cooperative fashion; for instance, debt-to-GDP ratios decline by 40 to 50 percentage points
(depending on the undertaken reform) below baseline on average across the regions, an
improvement of approximately 30 percent relative to a non-cooperative strategy
(Tables 2–4). This is due to the magnified effect of fiscal consolidation efforts on world
savings and world real interest rates with larger attendant crowd-in effect on investment.
35
Figure 15. Cooperative Versus Regional Public Pension Reform—Increase in
Retirement Age
36
Figure 16. Cooperative Versus Regional Public Pension Reform—Reducing
Pension Benefits
37
Figure 17. Cooperative Versus Regional Public Pension Reform—Raising
Contribution Rates
38
Table 2. Cooperative Versus Regional Public Pension Reform—Increase in
Retirement Age
US
EU
JA
AS
RC
Global Scenario
Real GDP
5.4
7.9
2.1
7.2
13.5
Consumption
6.4
10.5
2.1
8.2
14.3
Real Interest Rate
-0.72
-0.72
-0.72
-0.72
-0.72
Government Debt to GDP
-53.6
-64.8
-0.0
-21.2
-77.1
United States Scenario
Real GDP
3.6
0.3
0.4
0.5
0.5
Consumption
4.0
0.2
0.2
0.4
0.6
Real Interest Rate
-0.13
-0.13
-0.13
-0.13
-0.13
Government Debt to GDP
-43.0
0.0
-0.0
0.0
-0.0
Euro Area Scenario
Real GDP
0.4
5.7
0.4
0.6
0.8
Consumption
0.2
6.5
0.2
0.6
1.4
Real Interest Rate
-0.14
-0.14
-0.14
-0.14
-0.14
Government Debt to GDP
0.0
-47.4
-0.0
0.0
0.0
Emerging Asia Scenario
Real GDP
0.1
0.1
0.1
3.4
0.2
Consumption
0.3
0.3
0.4
2.6
0.4
Real Interest Rate
-0.01
-0.01
-0.01
-0.01
-0.01
Government Debt to GDP
0.0
0.0
0.0
-13.5
-0.0
Other Countries Scenario
Real GDP
1.0
1.3
0.9
2.2
11.7
Consumption
1.6
3.1
1.3
4.1
11.6
Real Interest Rate
-0.25
-0.25
-0.25
-0.25
-0.25
Government Debt to GDP
0.0
0.0
-0.0
0.0
-67.0
39
Table 3. Cooperative Versus Regional Public Pension Reform—Reducing
Spending on Pension Benefits
US
EU
JA
AS
RC
Global Scenario
Real GDP
1.6
1.7
1.7
2.0
1.9
Consumption
1.3
1.7
-0.3
0.8
2.4
Real Interest Rate
-0.82
-0.82
-0.82
-0.82
-0.82
Government Debt to GDP
-53.2
-58.1
-0.0
-21.3
-68.8
United States Scenario
Real GDP
0.5
0.3
0.4
0.4
0.4
Consumption
1.3
-0.2
-0.0
0.1
-0.0
Real Interest Rate
-0.18
-0.18
-0.18
-0.18
-0.18
Government Debt to GDP
-41.0
0.0
-0.0
0.0
0.0
Euro Area Scenario
Real GDP
0.2
0.4
0.3
0.4
0.3
Consumption
-0.1
1.5
-0.0
0.0
-0.0
Real Interest Rate
-0.15
-0.15
-0.15
-0.15
-0.15
Government Debt to GDP
-0.0
-42.5
0.0
0.0
0.0
Emerging Asia Scenario
Real GDP
0.0
0.1
0.1
0.1
0.1
Consumption
-0.0
-0.0
-0.0
0.5
-0.0
Real Interest Rate
-0.03
-0.03
-0.03
-0.03
-0.03
Government Debt to GDP
-0.0
0.0
-0.0
-14.3
0.0
Other Countries Scenario
Real GDP
0.5
0.6
0.6
0.6
0.7
Consumption
-0.2
-0.0
-0.2
-0.1
2.1
Real Interest Rate
-0.29
-0.29
-0.29
-0.29
-0.29
Government Debt to GDP
0.0
0.0
0.0
0.0
-59.8
40
Table 4. Cooperative Versus Regional Public Pension Reform—Raising
Contribution Rates
US
EU
JA
AS
RC
Global Scenario
Real GDP
0.5
0.0
1.2
1.0
0.0
Consumption
0.1
-0.5
-0.6
-0.3
-0.1
Real Interest Rate
-0.62
-0.62
-0.62
-0.62
-0.62
Government Debt to GDP
-45.9
-50.2
-0.0
-19.2
-62.3
United States Scenario
Real GDP
-0.4
0.2
0.3
0.3
0.2
Consumption
0.4
-0.2
-0.1
-0.0
-0.2
Real Interest Rate
-0.14
-0.14
-0.14
-0.14
-0.14
Government Debt to GDP
-37.8
-0.0
-0.0
0.0
-0.0
Euro Area Scenario
Real GDP
0.1
-1.2
0.2
0.2
0.1
Consumption
-0.2
-0.3
-0.1
-0.1
-0.4
Real Interest Rate
-0.11
-0.11
-0.11
-0.11
-0.11
Government Debt to GDP
0.0
-40.4
-0.0
0.0
0.0
Emerging Asia Scenario
Real GDP
0.0
0.0
0.1
-0.1
0.1
Consumption
-0.1
-0.1
-0.0
0.3
0.0
Real Interest Rate
-0.03
-0.03
-0.03
-0.03
-0.03
Government Debt to GDP
0.0
-0.0
-0.0
-14.5
0.0
Other Countries Scenario
Real GDP
0.3
0.3
0.4
0.2
-1.2
Consumption
-0.4
-0.5
-0.4
-0.8
0.1
Real Interest Rate
-0.22
-0.22
-0.22
-0.22
-0.22
Government Debt to GDP
0.0
0.0
-0.0
0.0
-57.0
41
SENSITIVITY ANALYSIS
The benchmark reform scenarios depend on many assumptions, ranging from agents‘ degree
of impatience and myopia, to the timing of fiscal consolidation. As such, we test the
sensitivity of those benchmark results to changes in selected key parameter values and
investigate the potential shifts in the impact of the public pension reform on the economy.
We focus on the results of the benchmark cooperative global reform scenarios
26
and apply the
tests to the benefit cut scenario except for two tests, the sensitivity of hours worked to a raise
in retirement age and the role of monetary policy accommodation in the short run. To keep
the paper concise, we report the qualitative results of the main tests and refer the reader to
Appendix 2 for a quantitative assessment accompanied by Figures 18 to 24. The sensitivity
tests are the following:
A smaller increase in labor supplied in response to an increase in retirement age
reduces real activity. The benchmark case assumes no change in hours worked and
effort.
A shorter planning horizon with more myopic agents leads to a larger drop in
consumption initially but higher real GDP in the medium run driven by a boost in
investment resulting from an increased labor supply. Under a much longer planning
horizon, agents become far more Ricardian in their behavior. The demand effect is far
weaker with consumption behavior barely changing in response to reform in both the
short and long run. The supply and the capital intensity of production respond a lot
less which in turn reduces the effect of fiscal reform on real interest rates—the latter
does not respond in the long run to the global reform which lowers debt permanently.
Increasing the share of liquidity-constrained agents in the population to 50 percent in
all regions leads to a reduction in consumption in the United States and the euro area
(regions where this share has doubled), contrary to the benchmark results. Over the
medium run, interest rates fall relative to benchmark, boosting investment and the
long run level of real GDP. Emerging Asia and the block of remaining countries
(regions for which this share does not change) benefit from lower world interest rates,
which actually leads to a small improvement in consumption starting in the first year.
As for the timing of fiscal consolidation, having the public pension spending adjust
immediately to its lower long-run level depresses consumption in the short run for
most regions. The difference is larger in the remaining countries block due to a larger
share of
LIQ
households. The increase in government savings crowds out the need
for private savings much quicker, but investment picks up faster, which means that
26
For the most part, individual country reform scenarios (as discussed in Section IV. B.) rely on the parameters
discussed below in a similar fashion.
42
the recovery in real GDP happens much earlier than in the benchmark reform
scenario.
Monetary policy accommodation for one or two years boosts consumption and real
GDP at the expense of added inflation volatility. Short-run demand pressure results in
movements in inflation away from its target and a further decline in real interest rates,
prolonging the initial positive impact of the public pension reforms. Over the short
run, real GDP and consumption rise considerably more relative to the benchmark
scenario, and public finances improve as government deficits decline faster in light of
lower debt service payments. When normal conduct of monetary policy returns, there
is the usual dampening effect in the medium run. In the long run, the economy
follows the same path as in the benchmark scenario.
The formulation of inflation behavior can have an effect on real variables. In the case
where the United States alone carries out public pension reform, there is a short-run
decline in consumption, because of inflationary pressures that lead to an increase in
interest rates by the monetary authority. However, by either cutting the degree of
inflation persistence, or altering how inflation persistence is determined, consumption
will be positive. Such a change to the calibration of the model would be at the
expense of other properties such as the volatility of exchange rates, and the possibility
for effective monetary accommodation.
27
IV. CONCLUSION
We considered reforms to the pension system that can help ensure the long-run viability of
public finances, while mindful of their short-run effect on economic activity in the midst of a
global financial crisis. This is carried out within a dynamic general equilibrium model
(GIMF) that captures the important economic interrelationships at a national and
international level. We emphasized measures to contain and fund the rising costs of age-
related spending in the medium to long run. We find that reforms which lead to short-run
adverse effects on real GDP (i.e., benefit reductions) are largely outweighed by the benefits
of declining real interest rates and the positive effect on future potential productive capacity.
The reform which has the most positive effects in the long run is lengthening the working
lives of employees, effectively raising the size of the active labor force relative to the retiree
population. It helps boost domestic demand in the short run but also eases off the pressure on
governments to cut pension benefits alone—which can lead to additional private savings and
cause fragile domestic demand to fall in the short run—or to raise payroll and labor income
taxes—which can distort incentives to supply labor. We also found that the impact on real
27
The latter property is a linchpin of many results in other work on fiscal stimulus and consolidation with
GIMF, such as Freedman and others (2009a).
43
GDP of a cooperative approach to age-related fiscal reforms is greater compared to a case
where one but not all regions undertake reform.
In terms of public finances, our results generally show that stabilizing the GDP share of age-
related expenditures leads to a sizable decline in the debt-to-GDP ratio. Early efforts and
resolute action to reduce future age-related spending or finance the spending through
additional tax increases and other measures (preferably through an increase in retirement age)
could significantly improve fiscal sustainability in several countries over the medium run,
and more so if such reforms are enacted in a cooperative fashion.
44
Appendix 1. GIMF’s Main Features and Calibration
The model economies feature three main sectors: households, firms (and entrepreneurs) and a
government. An external sector is also fully developed. The appendix describes households‘
optimizing behavior including aggregation across generations, highlights the firms‘
constrained profit maximization problem and discusses the government sector‘s
intertemporal budget constraint and fiscal policy aimed to stabilize deficits and the business
cycle. Real aggregate variables, say
t
x
, are rescaled by dividing by the level of labor
augmenting world technology
t
T
and by population growth; hence the notation
t
t t t
x x Tn
. The steady state of
t
x
is denoted by
x
. The world economy‘s technology
grows at the constant rate
1tt
g T T
and world population grows at the constant rate
n
.
Under the assumption that all consumers in the five regions face identical survival
probabilities, the relative size of the populations remain constant.
Households’ Sector and Optimality Conditions
The share of liquidity-constrained
LIQ
households agents in the population is
()z
where
z
refers to a country. In each period
( ) ( ) 1
t
N z n z
n
of such individuals are born, where
()Nz
indexes absolute population sizes in period
0
. Second, there are
OLG
households
with finite planning horizons with a share equals to
1 ( )z
; in each period
( ) (1 ( )) 1
t
N z n z
n
of such individuals are born. Each of these agents faces a constant
probability of death
1
in each period, which implies an average planning horizon
1/ 1
. In addition to the probability of death households also experience labor
productivity (and hence labor income) that declines at a constant rate
over their lifetimes,
reaching near zero in 20 years.
28
Life-cycle income adds another powerful channel through
which fiscal policies have non-Ricardian effects, as this along with
produce a high degree
of myopia. Households of both types are subject to uniform labor income, consumption and
lump-sum taxes.
28
This stylized treatment of life-cycle income is made possible by the absence of explicit demographics in
GIMF, which means that an assumption of declining labor productivity suffices to be correct for the average
worker.
45
Overlapping Generations (OLG) Households
Analytically, the key relation for
OLG
households is their optimal consumption rule which
states that consumption equals the marginal propensity to consume out of wealth times real
wealth. The current aggregate consumption is expressed as a function of real aggregate
financial wealth and human wealth. The latter is composed of the expected present
discounted value of households‘ time endowments evaluated at the after-tax real wage, and
of the expected present discounted value of capital or dividend income net of lump-sum
transfer payments to the government. The implication is that government debt adds to agents‘
net worth, that the time profile of taxes affects the time profile of consumption, and that in
the long run government debt crowds out private capital and net foreign assets. The reason is
that a household with finite planning horizon attaches less importance to higher tax payments
in the distant future, by discounting future tax liabilities at rates that are higher than the
market real interest rate due to their myopia
1
and the rate of decline of their labor
income
1
. Government debt is therefore net wealth to the extent that households, due to
short planning horizons, disregard the future taxes necessary to service that debt.
A fiscal consolidation through higher taxes represents a tilting of the tax payment profile
from the more distant future to the near future, resulting in a reduction in the debt stock. The
present discounted value of the government‘s future primary deficits has to remain equal to
the current debt under this scenario. But for households the same tilting of the tax profile
represents a decrease in human wealth because it increases the expected value of future taxes
for which the household expects to be responsible—this is true, for example, for the direct
effects of labor income tax on labor income receipts. For a given marginal propensity to
consume, this decrease in human wealth leads to a decrease in consumption. This is
particularly relevant in explaining the likely negative demand effects of raising contribution
rates on consumption.
A key parameter for the marginal propensity to consume, the intertemporal substitution in
consumption
1
is in line with the business cycle literature—this parameter affects the
sensitivity of consumers to changes in the real interest rate. For a conventional assumption of
1
, the income effect of an increase in the real interest rate is stronger than the substitution
effect and tends to increase the marginal propensity to consume, thereby partly offsetting the
contractionary effects of a higher real interest rate on human wealth. A larger
therefore
tends to give rise to larger interest rate changes in response to fiscal shocks, which is possible
to analyze given the more general utility functions added in GIMF.
Consumption, Wealth, and Labor
The household‘s budget constraint, derivation of the first-order conditions for each
generation and aggregation across generations are discussed in detail in KLMM (2010),
Appendices 1–3. The lifetime expected utility of a representative household is such that a
46
representative member of the
OLG
group of households and of age
a
derives utility at time
t
from consumption
,
OLG
at
c
and leisure
,
L OLG
t a t
Sl
where
L
t
S
is the stochastic time endowment
and
l
is the supply of labor by households,
1
1
,,
0
1
()
1
OLG OLG
s OLG L OLG
a s t s t a s t s
s
c S l
(5)
is the discount factor,
1
determines the planning horizon,
0
is the coefficient of
relative risk aversion, and
01
OLG
determines the consumption-leisure choice.
The key result for
OLG
households—their optimal consumption rule, states that consumption
equals the marginal propensity to consume out of wealth times real wealth. The current
aggregate consumption of
OLG
households is expressed as a function of their real aggregate
financial wealth
t
fw
and human wealth
t
hw
, with the marginal propensity to consume out of
wealth given by
1
t
. Human wealth is in turn composed of
L
t
hw
, the expected present
discounted value of households‘ time endowments evaluated at the after-tax real wage, and
K
t
hw
, the expected present discounted value of capital or dividend income net of lump-sum
transfer payments to the government. After rescaling by technology we have
OLG L K
t t t t t
c fw hw hw
(6)
where
*
1 1 1 1 1 1 1 1
1
1,
N T f
t t t t t t t t t t
t
fw i b b b i f e
gn
(7)
,1
1 1 ,
LL
t t L t t
t
g
hw N w hw
r
(8)
,
1
,
K ls OLG OLG K
t t t t t t
j
t
g
hw d rbr hw
r
(9)
,
1
R
t c t
t
tt
OLG
t
p
j
r
(10)
1
1
1
11
1
1
1 , 1 ,
,
1 , 1 1
, 1 , 1
1
1
OLG
OLG
v
R
R
OLG
t L t t c t
t c t
t
tt
R OLG
R
t c t t
t L t t c t
wp
p
cg
j r g
pc
wp
(11)
Liquidity Constrained (LIQ) Households and Aggregate Households
These agents can consume at most their current income, which consists of their after-tax
wage income plus net government transfers. This group of households therefore has a
47
marginal propensity to consume out of present income of unity, so that fiscal multipliers of
revenue-based fiscal stimulus measures such as tax cuts and increases in transfers are
particularly high whenever such agents constitute a large share of the population. The
objective function of
LIQ
households is assumed to be identical to that of
OLG
households.
LIQ households take their wage as given, accepting the OLG households‘ wage. Similarly,
LIQ households supply labor in an amount proportionate to the amount supplied by OLG
households in response to their consumption-leisure optimization problem. Aggregate
consumption and labor supply are given by
OLG LIQ
t t t
C c c
and
OLG LIQ
t t t
L l l
.
Firms
Firms are managed in accordance with the preferences of their owners, myopic
OLG
households, and they therefore also have finite-planning horizons. The productive sector
features a conventional setup with standard features for the tradables, nontradables,
intermediate goods, final consumption and investment goods. An added financial accelerator
mechanism amplifies the propagation mechanism for standard shocks and creates a role for
new shocks. So except for capital goods producers, entrepreneurs and retailers, firms are
monopolistically competitive and subject to nominal rigidities in price setting. Each firm
maximizes the present discounted value of net cash flow or dividends. The first-order
conditions are standard except for the presence of the probability of survival term,
, in the
discount factors. Again, KLMM (2010) contains the complete details for all firm and union
sectors for the interested reader.
Government
Budget Constraint
Fiscal policy consists of a specification of government consumption and investment spending
cons inv
t t t
G G G
, lump-sum taxes on both
LIQ
and
OLG
households, lump-sum transfers
to both
LIQ
and
OLG
households
OLG LIQ
t t t
, and tax rates on labor income (
,Lt
),
consumption (
,ct
) and corporate income (
,kt
). This rather detailed representation of taxes,
transfers and government spending enables us to analyze the impact of pension-related tax
and expenditure reforms. Pensions, in particular, are part of the untargeted lump-sum
transfers to households, whereas social security contribution rates are embedded in taxes on
labor income.
Government consumption spending is unproductive, while government investment spending
augments a stock of publicly provided infrastructure capital that depreciates at some rate. Tax
revenue is endogenous and given by the sum of labor, consumption, capital, and lump-sum
taxes. The real government budget constraint is
48
11
11
tt
t t t t t t t
tt
ii
b b G b s
gn gn
(1)
where the government issues nominally non-contingent one-period nominal debt
t
b
at the
gross nominal interest rate
t
i
and
t
s
is the primary surplus.
The government‘s policy rule for transfers partly compensates for the lack of asset ownership
by
LIQ
agents by redistributing a small fraction of
OLG
agents‘ dividend income receipts (in
various sectors, retail, unions, etc.) to
LIQ
agents. Government lump-sum transfers and
lump-sum taxes are received and paid by
LIQ
agents in proportion to their share in aggregate
consumption, but this rule can be easily changed, for example, to allow for transfers that are
100 percent targeted to
LIQ
agents.
Fiscal Policy
A fiscal policy rule stabilizes deficits and the business cycle. First, it stabilizes the interest-
inclusive government deficit to GDP ratio
rat
t
gd
at a long-run target (structural) level
rat
t
gdss
,
which rules out default and fiscal dominance (dynamic stability). Second, it stabilizes the
business cycle by letting the deficit fall with the output gap. We have
ln
rat rat gdp
t
tt
pot
gdp
gd gdss d
gdp
(2)
Here
0
gdp
d
, and
rat
t
gd
is given by
11
1
1
100 100
tt
t
tt
rat
tt
t
tt
ib
b
sb
gn gn
gd
gdp gdp
(3)
Shocks to
rat
t
gdss
represent changes in government saving preferences. We denote the
current value and the long-run target of the government debt to GDP ratio by
rat
t
b
and
rat
t
bss
.
The relationship between
rat
t
bss
and
rat
t
gdss
follows directly from the government‘s budget
constraint as
1
rat rat
tt
gn
bss gdss
gn
(4)
where
is the inflation target of the central bank. In other words, for a given trend nominal
growth rate, choosing a deficit target
rat
t
gdss
implies a debt target
rat
t
bss
and therefore keeps
debt from exploding. We note that the long run autoregressive coefficient on debt, at
1 gn
,
is quite close to one. An identical relationship holds between the long run net foreign
liabilities to GDP ratio and the long run current account deficit to GDP ratio.
49
The value of
gdp
d
can be thought of as quantifying the automatic stabilizers in the economy.
So for a given one percent increase in the output gap, the fiscal authority increases its target
surplus by
gdp
d
percent. GIMF allows for permanent shocks to technology and to private
saving, which have permanent effects on potential GDP due to the non-Ricardian features of
the model. Potential output is therefore modeled as a moving average of past actual values of
GDP to allow for the gap to close over time. Fiscal policy can typically be characterized by
the degree to which automatic stabilizers work. Finally, the fiscal rule equation above is not
an instrument rule but rather a targeting rule. Any of the available tax and spending
instruments can be used to make sure the rule holds. The default setting in this paper is that
this instrument is lump-sum transfers.
Monetary Policy
The model uses a stylized Taylor-type interest rate reaction function in which central bank
set interest rates on the basis of the deviation of current and one-year ahead inflation from the
inflation target, to stabilize inflation at a pre-specified target level.
Calibration
The steady state ratios which are of particular importance for this exercise are presented in
Appendix Table 1. Steady state GDP decompositions and trade flows are based on recent
historical averages. Tax revenues are decomposed into its four components, with the assumed
shares, based on somewhat detailed data, used to infer the model‘s steady state tax rates.
As for other parameter of interest, household utility functions are equal across countries. The
intertemporal elasticity of substitution is 0.5 (describing the sensitivity of consumption to
changes in the real interest rate), or = 2, and the wage elasticity of labor supply is
0.5
. The
parameters
,
and
are critical for the non-Ricardian behavior of the model. The shares
of
LIQ
consumers are 25 percent in the United States, the euro area, and Japan, and 50
percent in emerging Asia (reflecting among other things the nascent or underdeveloped
financial markets for domestic consumers) and other countries. The planning horizon is equal
to 20 years for
0.95
and the average remaining time at work is 20 years for
0.95
.
The main criterion used in choosing
and
is the empirical evidence for the effect of
government debt on real interest rates.
29
29
Our model is calibrated so that a 1 percentage point increase in the government debt to GDP ratio in the
United States leads to a long-run increase of approximately 1 to 2 basis points in the U.S. (and world) real
interest rate. This value is on the low end of the range of estimates provided by Laubach (2003), Engen and
Hubbard (2004), and Gale and Orszag (2004).
50
As for production technologies, the elasticities of substitution are equal to
1
between capital
and labor,
0.75
between domestic and foreign goods, and
0.5
between tradables and
nontradables. Steady state markups equal
1.1
in manufacturing and wage setting,
1.05
in
retailing, investment and consumption goods production, and
1.025
for import agents.
In calibrating the fiscal rule, we use OECD estimates (Girouard and André, 2005) for the
output gap coefficients,
gdp
d
which proxy for the automatic stabilizers in the economy. We
assume the overall target deficit-to-GDP ratios are equal to the projected 2014 government
debt to GDP ratios (IMF, 2010c). Finally, the calibration of monetary rule parameters is
conventional for an annual model. For emerging Asia, we assume a fixed nominal exchange
rate vis-à-vis the U.S. dollar.
Appendix Table 1. Selected Steady-State Values
A. Expenditure Shares (In percent of GDP)
US
AS
EU
JA
RC
Consumption
65.3
59.0
58.2
59.5
59.0
OLG Consumption
51.8
33.7
46.5
46.9
34.0
LIQ Consumption
13.5
25.3
11.7
12.6
25.0
Private Investment
17.2
25.0
18.3
21.0
19.0
Government Expenditures
17.5
16.0
23.5
19.5
22.0
Government Consumption
15.0
12.0
20.5
17.0
20.0
Government Investment
2.5
4.0
3.0
2.5
2.0
Exports
11.5
27.0
17.4
11.0
21.9
Imports
11.5
27.0
17.4
11.0
21.9
B. Deficit, Debt, Taxes, and Transfers
US
AS
EU
JA
RC
Net Debt / GDP
128.7
48.5
136.2
156.0
94.4
Deficit / GDP
6.8
2.6
6.6
6.1
4.6
Labor Income Tax Rate
29.3
8.3
40.8
26.3
22.3
share in tax revenues
61.9
20.1
60.5
55.2
33.8
Capital Income Tax Rate
15.8
12.2
13.4
19.3
13.4
share in tax revenues
11.3
13.9
6.9
17.5
7.3
Consumption Tax Rate
7.0
17.0
18.4
8.7
17.0
share in tax revenues
16.3
46.3
26.4
18.2
25.2
Lump-Sum Taxes / Total Tax Revenue
10.2
18.9
6.1
9.1
33.8
Government General Transfers / GDP
9.5
5.1
15.6
7.5
16.7
C. Factor Shares (In percent of GDP)
US
AS
EU
JA
RC
Labor Income
60.0
54.0
60.0
60.0
60.0
Source: KLMM (2010); IMF (2010c); and staff estimates.
51
Appendix 2. Sensitivity Analysis
Labor Supply Response to an Increase in the Retirement Age
Figure 15 above shows the impact of a global increase in retirement ages assuming that, in all
countries, workers will effectively stay longer in the labor force for as many years as the
official increase in the retirement age. Arguably, labor supply among older workers would
decrease in such an event. Figure 18 considers the same reduction in pension benefits as in
the benchmark reform scenario, but the increases in hours worked are
23
and
34
the
amount assumed to be associated with the extension of working life presented earlier.
As expected, a smaller increase in the labor supplied by older workers reduces real GDP by
roughly 20 to 25 percent worldwide in the
34
case, and 27 to 33 percent in the
23
case.
This is the result of both a direct effect on production and an indirect effect through a lower
marginal product of capital with a negative effect on investment growth. Human wealth
decreases cumulatively over time, leading to a decline in consumption under an unchanged
marginal propensity to consume.
Role of the Degree of Myopia or the Planning Horizon
The short-run adjustment of the global economy is influenced by the behavior of households
and firms, and depends to a large extent on
OLG
households‘ degree of myopia, exemplified
in the length of their planning horizon, assumed to be 20 years (
0.95
) under the
benchmark scenario. Figure 19 shows the results under two alternative values of
, (i)
0.90
reflecting a shorter planning horizon or more myopic
OLG
agents and, (ii)
0.98
reflecting a much longer planning horizon. The latter case is expected to bring the results
closer to a typical representative-agent Ricardian model in stark contrast to the standard
GIMF non-Ricardian
OLG
structure.
Under a 10-year planning horizon with more myopic consumers, consumption initially drops
notably for the United States and the euro area which are characterized by a lower share of
LIQ
households in their population.
OLG
agents do not take into account the crowding-in
benefits and the higher future income from later retirement that is foreseen in the long run,
and while simply consuming their (lower) after-tax current income they react more on impact
to a fall in pension benefits. By a wealth effect, labor supply also rises more (agents consume
less leisure). This, coupled with a lower marginal propensity to consume, brings investment
up by a substantial amount, and in the long run the real GDP boost leads to lower debt to
GDP ratios and real interest rates.
Under a much longer planning horizon of 50 years, agents become far more Ricardian in
their behavior and their debt holdings are no longer treated as a source of wealth.
Accordingly, the demand effect is far weaker and saving and consumption behavior barely
52
changes under the considered public pension reform in both the short and long run. The
supply and the capital intensity of production respond a lot less which in turn reduces the
effect of fiscal reform on real interest rates. Importantly, the real interest rate does not
respond in the long run to the global reform which lowers debt permanently. However, such
an extreme and unrealistic near-Ricardian behavior does not garner much empirical support.
Share of Liquidity—Constrained Households
The introduction of
LIQ
consumers adds further realism to the dynamic responses of macro
variables, with changes in their share value in the population
expected to have a notable
impact on the results.
We analyze the change in results when the share of
LIQ
agents equals 50 percent for all
countries. As Figure 20 portrays, consumption now declines for the United States and the
euro area, contrary to the benchmark results. The supply side is not affected in the short and
medium run, but after 10 periods, interest rates will have fallen more than in the benchmark,
propelling investment and long run growth. Emerging Asia and the block of remaining
countries (regions for which the share of
LIQ
households did not change) benefit from lower
world interest rates, which actually leads to a small improvement in consumption already in
the first periods.
Timing of the Fiscal Consolidation
Pension gaps are reduced steadily with the adoption of any of the planned public pension
reforms. In Section IV, the dynamics of the macroeconomic adjustment were shown to have
been strongly influenced by the way the profile of fiscal consolidation was phased in,
pursuant to an increase in retirement age. Under the benchmark scenario above, pension
spending cuts by the government, on average, were more tilted toward the longer run with
relatively less adjustment undertaken in the medium run. This, and to the extent that agents
are non-Ricardian, had attenuated the impact of the reform on reducing agents‘ wealth in the
short run, and consequently the expected negative impact on consumption. Moreover, there is
less of a crowding-out effect on private saving in comparison with a more front-loaded fiscal
consolidation scenario.
To illustrate this point, Figure 21 shows an immediate and permanent decline in public
pension spending while preserving the magnitude of the fiscal consolidation itself in the long
run. As a result, consumption drops in the short run for the United States, the euro area and
the block of remaining countries. The difference is larger in the latter, due to the larger share
of
LIQ
households. As expected,
OLG
agents‘ consumption is not affected to the same
extent. As just noted above, the increase in government savings crowds out the need for
53
private savings much quicker, but investment picks up faster, which means that the recovery
in real GDP happens much earlier than in the benchmark reform scenario.
Finally, Figure 22 considers the case where all countries mimic the fiscal consolidation
profile adopted in the United States (that is, benefit reduction is similar over time). The
magnitude of responses in real GDP, consumption, and private saving are much closer to
what is seen for the United States case, particularly in the long run. However, the external
adjustment patterns—imports, exports and real effective exchange rate—across the various
economic regions remain unchanged, as they are mostly dictated by trade linkages as
calibrated in GIMF.
Monetary Policy Accommodation
The monetary policy rule can play an important role in moderating the short-run effects of
public pension reforms. When reforms are announced, forward looking consumers
immediately anticipate a higher future income. The expected rise is particularly large and
faster when retirement ages increase, as discussed in Section IV. This leads to higher
consumption and inflationary pressures, which under the benchmark scenario triggers a usual
policy rate increase, dampening the increase in consumption and real GDP.
In Figure 23, we consider the effect of a monetary accommodation of the initial boost in
consumption and real GDP. Nominal interest rates are kept fixed at their initial levels for
either one year, or two years. This in turn reduces real interest rates because the nominal
interest rate is lower than it would be otherwise, and because inflation is higher than it would
be otherwise. There is a crowding in effect on consumption and investment, driving up real
GDP. Public finances are improved as public debt declines further and fiscal deficits improve
on the account of lower debt service payments and the effect of larger real GDP on the
relevant tax bases. Eventually, the role of the stabilizing impact of the interest rate reaction
function is restored and policy rates now rise by a larger magnitude than was the case under
the benchmark case, as monetary policy must bring a larger deviation from the inflation
target under control.
This exercise demonstrates how a short-run boost in activity can be extended through a pre-
announced monetary accommodation. This may be adopted as a policy choice to sustain a
confidence effect in structural (entitlement) reforms, without deviating for too long from the
price stability mandate of central banks. The long-run impact of the public pension reforms
does not change with monetary accommodation.
Behavior of Inflation
The way in which we determine the behavior of inflation can affect the short-run
macroeconomic effects of the public pension reform scenarios under consideration. A useful
54
example, explored in Figure 24, bases its results on the benchmark public pension reform
scenario where the United States alone undertakes public pension reform by increasing the
pensionable age accompanied by a reduction in public pension spending (see Figure 3).
Inflation dynamics are affected in GIMF by the level and the formulation of nominal
rigidities; we study their implications by conducting two experiments: cutting the level of
nominal rigidities (by half) and changing the way nominal rigidities enter in the model. On
the latter, we deviate from the usual and modern practice of allowing an increase in inflation
from year to year (Ireland, 2001) to a hybrid formulation where half of the rigidities come
from the Ireland fomulation and the other comes from the use of a more traditional
Rotemberg (1982) formulation, where adjustment costs depends on the deviation of inflation
each year from the stated inflation target.
Recall that under the benchmark scenario, consumption weakens during years 2 to 7. Under
lower inflation persistence (as a result of lower nominal rigidities or adopting a different
formulation of the inflation process), it is shown here that consumption can actually
strengthen. This can be explained as follows. First, by effectively speeding up the response of
inflation, the time horizon over which the monetary authority has to offset the inflationary
impact is shortened. Second, inflation is now more responsive to interest rates, since inflation
either moves faster (lower rigidities) or naturally tends towards the long-run inflation target
(a weight of 0.5 on the Rotemberg formulation of nominal price rigidities); as such, interest
rates would need to increase by less to get the same movement in inflation. Consequently,
lower interest rates exert less downward pressure on the real economy in the short run.
Note that the perceived benefits of these experiments on domestic activity may suggest a
change in the inflation process calibration. We should however carefully consider their
implications on the overall model properties and not just the impact on domestic activity
(a subject for future research). To draw an example, monetary accommodation under these
experiments, would be less effective in the model (and consequently, in light of other policy
work with GIMF, less realistic). Moreover, some variable responses of the model are less
sensible; for example, the REER moves much less now (less inflation persistence) at odds
with observed movements in the data.
55
Figure 18. Sensitivity Analysis Around the Benchmark Coordinated Global
Reform Scenario—Labor Supply Response
56
Figure 19. Sensitivity Analysis Around the Benchmark Coordinated Global
Reform Scenario—Role of the Planning Horizon (Degree of Myopia)
57
Figure 20. Sensitivity Analysis Around the Benchmark Coordinated Global
Reform Scenario—Share of LIQ Households at 50 Percent Worldwide
58
Figure 21. Sensitivity Analysis Around the Benchmark Coordinated Global
Reduction in Public Pension Spending—Rapid Implementation
59
Figure 22. Sensitivity Analysis Around the Benchmark Coordinated Global
Reduction in Public Pension Spending—U.S. Reduction Levels Applied
Globally
60
Figure 23. Sensitivity Analysis Around the Benchmark Coordinated Global
Reform Scenario—Monetary Policy Accommodation
61
Figure 24. Sensitivity Analysis Around the Benchmark Reduction in Public
Pension Spending by the United States Only—Different Inflation Behaviors
62
References
Andersen, T. M., 2008, ―The Scandinavian Model: Prospects and Challenges‖ International
Tax and Public Finance, Vol.15, No. 1, pp.45–66.
Andersen, T. M., and L. H. Pedersen 2008, ―Distribution and Labour Market Incentives in
the Welfare State: Danish Experiences,‖ Swedish Economic Policy Review, Vol.14,
pp.175–214.
Barrell, R., I. Hurst, and S. Kirby, 2009, ―How to Pay for the Crisis or: Macroeconomic
Implications of Pension Reform,‖ Discussion Paper No. 333 (London: National
Institute of Economic and Social Research).
Botman, D., and M. Kumar, 2007, ―Global Aging Pressures: Impact of Fiscal Adjustment,
Policy Cooperation, and Structural Reforms,‖ IMF Working Paper 07/196
(Washington: International Monetary Fund).
Cournède, B., and F. Gonand, 2010, ―Restoring Sustainability in the Euro Area: Raise Taxes
or Curb Spending,‖ OECD Working Paper No. 520 (Paris: Organization for
Economic Cooperation and Development).
Coenen, G., and others, 2010, ―Effects of Fiscal Stimulus in Structural Models,‖ IMF
Working Paper 10/73 (Washington: International Monetary Fund).
Disney, R., 2005, ―Household Saving Rates and the Design of Social Security Programmes:
Evidence from a Country Panel,‖ CESifo Working Paper No. 1541 (Munich: Ifo
Institute for Economic Research).
Engen, E.M., and R.G. Hubbard, 2004, ―Federal Government Debt and Interest Rates,‖
NBER Macroeconomics Annual, Vol. 19, pp. 83–138.
European Commission, 2009, The 2009 Ageing Report: Economic And Budgetary
Projections for the EU-27 Member States (Brussels).
Freedman, C., and others, 2009a, ―The Case for Global Fiscal Stimulus,‖ IMF Staff Position
Note 09/03 (Washington: International Monetary Fund).
Freedman, C., and others, 2009b, ―Fiscal Stimulus to the Rescue? Short-Run Benefits and
Potential Long-Run Costs of Fiscal Deficits,‖ IMF Working Paper 09/255
(Washington: International Monetary Fund).
63
Gale, W., and P. Orszag, 2004, ―Budget Deficits, National Saving, and Interest Rates,‖
Brookings Papers on Economic Activity, Vol. 2, pp. 101–87.
Girouard, N., and C. André, 2005, ―Measuring Cyclically-adjusted Budget Balances for
OECD Countries,‖ OECD Working Paper No. 434 (Paris: Organization for Economic
Cooperation and Development).
International Monetary Fund, 2010a, ―From Stimulus to Consolidation: Revenue and
Expenditure Policies in Advanced and Emerging Economies,‖ Departmental Paper
No. 10/03 (Washington).
———, 2010b, ―Strategies for Fiscal Consolidation in the Post-Crisis World,‖ Departmental
Paper No. 10/04 (Washington).
———, 2010c, World Economic Outlook Update, February, World Economic and Financial
Surveys (Washington).
Ireland, P., 2001, ―Sticky Price Models of the Business Cycle: Specification and Stability,‖
Journal of Monetary Economics, Vol.47, pp. 3–18.
Khoman, E., and M. Weale, 2008, ―Are We Living Beyond Our Means? A Comparison of
France, Italy, Spain and the United Kingdom,‖ Discussion Paper No. 311 (London:
National Institute of Economic and Social Research).
Kumhof, M., and D. Laxton, 2007, ―A Party without a Hangover? On the Effects of U.S.
Government Deficits,‖ IMF Working Paper 07/202 (Washington: International
Monetary Fund).
Kumhof, M., D. Laxton, D. Muir, and S. Mursula, 2010, ―The Global Integrated Monetary
and Fiscal Model (GIMF)—Theoretical Structure,‖ IMF Working Paper 10/34
(Washington: International Monetary Fund).
Laubach, T., 2003, ―New Evidence on the Interest Rate Effects of Budget Deficits and Debt,‖
Finance and Economics Discussion Series 2003-12, Board of Governors of the
Federal Reserve System.
Nickel, C., P. Rother, and A. Theophilopoulou, 2008, ―Population Ageing and Public
Pension Reforms in a Small Open Economy,‖ ECB Working Paper 863 (Brussels:
European Central Bank).
Rotemberg, J., 1982, ―Sticky Prices in the United States,‖ Journal of Political Economy,
Vol. 90, pp. 1187–211.
