Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Aging and Deflation from a Fiscal Perspective
Hideki Konishi and Kozo Ueda
Waseda Univ
May 2014 @ Bundesbank
KU (Waseda) FTPL May 2014 @ Bundesbank 1 / 29
Negative Correlation bw Aging and Deflation in Japan
0
5
10
15
20
25
-‐4
-‐2
0
2
4
6
8
1980
19
81
1982
19
83
1984
19
85
1986
19
87
1988
19
89
1990
19
91
1992
19
93
1994
19
95
1996
19
97
1998
19
99
2000
20
01
2002
20
03
2004
20
05
2006
20
07
2008
20
09
2010
20
11
2012
CPI
Aging
KU (Waseda) FTPL May 2014 @ Bundesbank 2 / 29
Negative Correlation bw Aging and Deflation in OECDCountries
5
10
15
20
CPI inflation rate, average of 2000-13, %
Correlation ―0.53 (p-value 0.001)(excluding Turkey: ―0.51)
Turkey
Mexico
-5
0
0 10 20 30 40
Old (+65) dependency ratio (20-64), average of 2000-12, %
Japan
Mexico
SwitzerlandUS
Germany
KU (Waseda) FTPL May 2014 @ Bundesbank 3 / 29
Motivation
We examine how population aging influences fiscal balances andgeneral prices within a political-economic framework.
KU (Waseda) FTPL May 2014 @ Bundesbank 4 / 29
What We Do
We extend the standard fiscal theory of the price level (FTPL).
1 We embed the FTPL into a standard overlapping generation (OLG)model
I to make it possible to examine political and economic impacts ofdemographic changes.
2 We consider endogenous policy making
I by succession of short-lived governments,I who choose tax rates and government bonds outstandingI under the political influences of existing generations and strategic
responses by future governments.
KU (Waseda) FTPL May 2014 @ Bundesbank 5 / 29
What We Do NOT Do
1 Challenge FTPL
2 Endogenize interactions between monetary policy and fiscal policy
3 Do realistic quantitative analysis
4 Investigate other reasons of persistent deflation
1 Insufficient monetary policy, growth strategy, malfunctioning offinancial system, ...
KU (Waseda) FTPL May 2014 @ Bundesbank 6 / 29
Four Features in Constructing a Model
1 Around 90% of the Japanese government bonds (JGBs) are held bydomestic investors.
1 Closed-economy model.
2 Nominal interest rate has been fixed at almost zero.
1 Passive monetary policy is the key to FTPL.
3 A part of Japanese population aging is an unexpected phenomenon.
KU (Waseda) FTPL May 2014 @ Bundesbank 7 / 29
Revisions in the Japanese Total Fertility Rate Forcast
1
1.2
1.4
1.6
1.8
2
2.2
2.4
1965 1975 1985 1995 2005 2015 2025 2035 2045 2055Source: Ministry of Health, Labour and Welfare; Na!onal Ins!tute of Popula!on and Social Security Research.
(Total Fer!lity Rate)
Year
Replacement ra!oForecast in 1976
Forecast in 1986
Forecast in 1992
Forecast in 2012
Forecast in 1997
Forecast in 2002
Forecast in 2006
KU (Waseda) FTPL May 2014 @ Bundesbank 8 / 29
Revisions in the Japanese Life Expectancy Forcast
67
70
73
76
79
82
85
88
91
1965 1975 1985 1995 2005 2015 2025 2035 2045 2055
Actual Figure in 2012Forecast in 1992Forecast in 1997Forecast in 2002Forecast in 2006Forecast in 2012
Source: Ministry of Health, Labour and Welfare; Na!onal Ins!tute of Popula!on and Social Security Research.
(Life Expectancy)
Year
Female
Male
KU (Waseda) FTPL May 2014 @ Bundesbank 9 / 29
Four Features in Constructing a Model
1 Around 90% of the Japanese government bonds (JGBs) are held bydomestic investors.
2 Nominal interest rate has been fixed at almost zero.
3 A part of Japanese population aging is an unexpected phenomenon.
4 The voter turnout rates for the young generation especially 20’s, 30’s,and 40’s are declining and the gap between generations is widening.
KU (Waseda) FTPL May 2014 @ Bundesbank 10 / 29
Voter Turnout Rates by Age in Japan
30
40
50
60
70
80
90
100
31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
20's
30's
40's
50's
60's
70's
Note: The turnout rates by age in the Japanese lower house elections No.31–45 (from
1967 to 2009) are depicted.
KU (Waseda) FTPL May 2014 @ Bundesbank 11 / 29
Fiscal Theory of Price Level (FTPL)
Today’s price level Pt is determined to balance govt’s intertemporalbudget in real terms [Leeper (’91), Woodford (’01) etc.] .
RBt−1Pt
= Tt − Gt +∞
∑s=t+1
(s
∏k=t+1
rk
)−1(Ts − Gs)
Assumptions of FTPL
I Nominal debts: govt has liabilities predetermined in nominal terms.I Passive monetary policy: CB keeps a nominal interest rate constant
over time.I Active govt policy: govt is not constrained by its budget balance eqn.I Price adjustment: Intertemporal budget is balanced only in equilibrium
through current price adjustment.
KU (Waseda) FTPL May 2014 @ Bundesbank 13 / 29
What does the Standard FTPL Predict?
Price level in Japan should rise sooner or later.
I Fiscal surplus is expected to deteriorate due to aging.
The standard FTPL takes account of no political factors.
I Policy choice will also respond to demographic aging throughintergenerational politics.
I Need to incorporate intergenerational politics and endogenize policychoice into FTPL.
KU (Waseda) FTPL May 2014 @ Bundesbank 14 / 29
Model Features
OLG model consists of the young and the old.
Labor income tax only. No physical capital.
Each individual will live for two period with an uncertain survivalprobability θjt .
Monetary policy is passive, keeping a fixed nominal interest rate.
Govt remains in power only for one period.
A succession of short-lived govts choose debt issues and income taxrates to maximize the weighted average of the young’s and the old’sutility in each period, taking account of the effects on current andfuture prices as well as next govt’s policy responses.
I Solve the Markov-perfect equilibrium of the dynamic policy choicegame.
KU (Waseda) FTPL May 2014 @ Bundesbank 15 / 29
Household
Each young household in period t chooses cyt and `t to maximize the
expected utility:
uy (cyt , `t) + βEt
[θjt+1uo(co
t+1)
],
where
cot+1 =
rt+1
θt+1
[(1− τt)`t − cy
t
]+ gT
t+1.
KU (Waseda) FTPL May 2014 @ Bundesbank 16 / 29
Household 2
Indirect utility function of the young in period t is
v yt (τt ;~rt+1) ≡ uy
(cyt (τt ;~rt+1), `t(τt ;~rt+1)
)+βEt
[θjt+1uo
(rt+1at(τt ;~rt+1)
θt+1+ gT
t+1
)],
where at represents saving per young and rt is the real interest rategiven by RPt/Pt+1.
The indirect utility of the old in period t is
vot (rt , at−1) ≡ uo
(rtat−1(τt−1 ;~rt)
θt+ gT
t
).
KU (Waseda) FTPL May 2014 @ Bundesbank 17 / 29
Market Clearing for Government Bonds
All the government bonds are held by domestic investors throughinsurance companies, ultimately by the young.
at(τt ;~rt+1) = bt ,
where bt is the real government bond supply in t.
KU (Waseda) FTPL May 2014 @ Bundesbank 18 / 29
Optimization Problem Facing Short-lived Governments
Govt in period t chooses τt , bt , and rt to maximize the weightedaverage of indirect utilities:
Wt = γtvot (rt , bt−1) + v y
t (τt ;~rt+1),
taking account of
I bonds market clearing condition: at(τt ;~rt+1) = bt .I the budget balance:
rtbt−1 = nt (bt + τt`t(τt ;~rt+1))− (nt + θt)gCt − θtgT
t .I and the next-period government’s policy decision embodied in~rt+1.
where nt ≡ Nt/Nt−1 and θjt represent young population’s growth rateand the survival probability. If γt = θt/nt , the government is amyopic utilitarian who maximizes the sum of utilities.
KU (Waseda) FTPL May 2014 @ Bundesbank 19 / 29
Markov Perfect Equilibrium
In a multi-period OLG model, extremely complex to calculateequilibrium, even its steady state.
I The entire path of past and future policies influence the behavior ofcurrent households and the current policy.
I ∂ajt/∂τt?
However, simple in the 2-period OLG model.
Eliminate rt and the optimization problem is reduced into
maxτt ,bt
Wt = γtuo
(nt(bt + τt`t)− (nt + θt)gC
t
θt+ gT
t
)+ v y
t (τt |~rt+1)
subject to the bonds market clearing condition at(τt ;~rt+1) = bt .
bt−1 does not appear here!
KU (Waseda) FTPL May 2014 @ Bundesbank 20 / 29
Result 1
bt−1 does not appear here!
The government’s optimal choices of τt and bt should be independentof bt−1.
I Only the optimal choice of rt , or the price level Pt , depends on bt−1.
Burdens of public debt are not passed to future unborn generationseven in the absence of altruistic bequests. [cf. Bowen, Davis, andKopf (1960), Barro (1979)]
I Larger bond issues in period t end up with higher prices in period t andt + 1.
I Burdens of public debt are fully paid by current old though reductionsin the real value of their assets and by young generation throughreductions in the real interest rate.
KU (Waseda) FTPL May 2014 @ Bundesbank 21 / 29
Result 2
Prices respond to aging in opposite directions, depending on whetherit is caused by longer lifetime or by lower birth rate.
Suppose that γt = θt/nt , population ratio. Then,
I Lower birth rate nt inflates prices as the standard FTPL predicts.I Longer life-expectancy θt is likely to inflate prices as lower birth rate
does, if it is expected.I Longer life-expectancy θt is likely to deflate prices as lower birth rate
does, if it is unexpected.
KU (Waseda) FTPL May 2014 @ Bundesbank 22 / 29
Intuition
Economic impact of aging
I Fiscal surplus declines due to a decline in tax revenues from the young,raising tax and price.
Political impact of aging
I Govt opts to decrease deficits, increase taxes, and lower price.
Moreover, unexpectedly long lifetime makes the old worse-offbecause their savings turn out insufficient.
I This strengthens the government’s distributional concerns, leading todeflation.
KU (Waseda) FTPL May 2014 @ Bundesbank 23 / 29
Concluding Remarks
Negative correlation
I A mild deflation and population aging in Japan’s lost decades.
Our result here suggests that this puzzling observation might becaused by
I the combined political and economic effects of unexpected populationaging having occurred from extension in longevity.
Future work
I address the accumulation in government bonds outstanding that isobserved in Japan
I introduce an endogenous monetary policy responseI introduce foreign investors to buy the government bondsI make a quantitative analysis
KU (Waseda) FTPL May 2014 @ Bundesbank 24 / 29
Responses of Prices to Extension of Longevity
The perfect foresight case experiences inflation whereas theunexpected case experiences deflation, implying that the publicexpectation for aging is a key to understanding the response of theprice levels.
0.98
0.99
1
1.01
1.02
1.03
0.94
0.95
0.96
0.97
0 1 2 3 4 5 6 7 8 9 10 11 12
Expected shock
Unxpected shock
KU (Waseda) FTPL May 2014 @ Bundesbank 26 / 29
Calibration
2 period OLG
I 1 period is 40 years.I Utility is given by
u(c lt , `lt) =
(c lt
)1−σ
1− σ− χ
(`lt
)1+1/υ
1 + 1/υ,
where σ = 1 and υ = 0.5.I β = 0.9940 (1% annually), gT
t = 0.01, and gCt = 0.01.
I Variables associated with demography, θt and nt , are derived fromJapan’s official statistics and forecasts by National Institute ofPopulation and Social Security Research (IPSS).
F As an innitial state, in 1997, θt = 0.620, nt = 0.304, andγyt /γo
t = 0.828.F As a final state, according to long-run forecasts in 2060 made in 2012,
θt = 0.781, nt = 0.555, and γyt /γo
t = 0.856.
KU (Waseda) FTPL May 2014 @ Bundesbank 27 / 29
Transition Path
Deflation due to recovered birth rate
0 1 2 3
0.4
0.45
0.5
0.55
0.6
0.65
Tax rate
0 1 2 3
1.01
1.015
1.02
1.025
1.03
1.035
Real interest rate (annualized)
0 1 2 3
0.94
0.96
0.98
1
1.02
Real bond
KU (Waseda) FTPL May 2014 @ Bundesbank 28 / 29
Transition Path under Exogenous Tax Rate
Multi-peiod OLG (16 generations, 5-year interval)
0 10 20 30 40 50
5
10
15
20
Real government bond
Transition period
0 10 20 30 40 50
-0.02
-0.01
0
0.01
0.02
Real interest rate (annual)
Transition period
0 10 20 30 40 50
-0.02
-0.01
0
0.01
0.02
Inflation rate (annual)
Transition period
0 10 20 30 40 50
0
0.1
0.2
0.3
0.4
Income tax rate
Transition period
KU (Waseda) FTPL May 2014 @ Bundesbank 29 / 29