Aging and Deflation from a Fiscal Perspective Hideki Konishi and - - PowerPoint PPT Presentation

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 ¡ 1981 ¡ 1982 ¡ 1983 ¡ 1984 ¡ 1985 ¡ 1986 ¡ 1987 ¡ 1988 ¡ 1989 ¡ 1990 ¡ 1991 ¡ 1992 ¡ 1993 ¡ 1994 ¡ 1995 ¡ 1996 ¡ 1997 ¡ 1998 ¡ 1999 ¡ 2000 ¡ 2001 ¡ 2002 ¡ 2003 ¡ 2004 ¡ 2005 ¡ 2006 ¡ 2007 ¡ 2008 ¡ 2009 ¡ 2010 ¡ 2011 ¡ 2012 ¡ CPI ¡ Aging ¡

KU (Waseda) FTPL May 2014 @ Bundesbank 2 / 29

Negative Correlation bw Aging and Deflation in OECD Countries

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

10 20 30 40

Old (+65) dependency ratio (20-64), average of 2000-12, %

Japan Mexico Switzerland US Germany

KU (Waseda) FTPL May 2014 @ Bundesbank 3 / 29

Motivation

We examine how population aging influences fiscal balances and general 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

◮ to make it possible to examine political and economic impacts of

demographic changes.

2 We consider endogenous policy making ◮ by succession of short-lived governments, ◮ who choose tax rates and government bonds outstanding ◮ 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 of financial 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 by

domestic 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 2055

Source: Ministry of Health, Labour and Welfare; Naonal Instute of Populaon and Social Security Research.

(Total Ferlity Rate) Year Replacement rao Forecast 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 2012 Forecast in 1992 Forecast in 1997 Forecast in 2002 Forecast in 2006 Forecast in 2012

Source: Ministry of Health, Labour and Welfare; Naonal Instute of Populaon 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 by

domestic 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

Model

KU (Waseda) FTPL May 2014 @ Bundesbank 12 / 29

Fiscal Theory of Price Level (FTPL)

Today’s price level Pt is determined to balance govt’s intertemporal budget in real terms [Leeper (’91), Woodford (’01) etc.] . RBt−1 Pt = Tt − Gt +

∞

∑

s=t+1

• s

∏

k=t+1

rk −1 (Ts − Gs) Assumptions of FTPL

◮ Nominal debts: govt has liabilities predetermined in nominal terms. ◮ Passive monetary policy: CB keeps a nominal interest rate constant

• ver time.

◮ Active govt policy: govt is not constrained by its budget balance eqn. ◮ 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.

◮ Fiscal surplus is expected to deteriorate due to aging.

The standard FTPL takes account of no political factors.

◮ Policy choice will also respond to demographic aging through

intergenerational politics.

◮ Need to incorporate intergenerational politics and endogenize policy

choice 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 survival probability θj

t.

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 tax rates to maximize the weighted average of the young’s and the old’s utility in each period, taking account of the effects on current and future prices as well as next govt’s policy responses.

◮ Solve the Markov-perfect equilibrium of the dynamic policy choice

game.

KU (Waseda) FTPL May 2014 @ Bundesbank 15 / 29

Household

Each young household in period t chooses cy

t and ℓt to maximize the

expected utility: uy(cy

t , ℓt) + βEt

• θj

t+1uo(co t+1)

• ,

where co

t+1 = rt+1

θt+1

• (1 − τt)ℓt − 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 vy

t (τt ;

rt+1) ≡ uy cy

t (τt ;

rt+1), ℓt(τt ; rt+1)

• +βEt
• θj

t+1uo

rt+1at(τt ; rt+1) θt+1 + gT

t+1

• ,

where at represents saving per young and rt is the real interest rate given by RPt/Pt+1. The indirect utility of the old in period t is vo

t (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 through insurance 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 weighted average of indirect utilities: Wt = γtvo

t (rt, bt−1) + vy t (τt ;

rt+1), taking account of

◮ bonds market clearing condition: at(τt ;

rt+1) = bt.

◮ the budget balance:

rtbt−1 = nt (bt + τtℓt(τt ; rt+1)) − (nt + θt)gC

t − θtgT t .

◮ and the next-period government’s policy decision embodied in

rt+1.

where nt ≡ Nt/Nt−1 and θj

t represent young population’s growth rate

and the survival probability. If γt = θt/nt, the government is a myopic 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 calculate equilibrium, even its steady state.

◮ The entire path of past and future policies influence the behavior of

current households and the current policy.

◮ ∂aj

t/∂τt?

However, simple in the 2-period OLG model. Eliminate rt and the optimization problem is reduced into max

τt,bt Wt = γtuo

nt(bt + τtℓt) − (nt + θt)gC

t

θt + gT

t

• + vy

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 independent

• f bt−1.

◮ 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 generations even in the absence of altruistic bequests. [cf. Bowen, Davis, and Kopf (1960), Barro (1979)]

◮ Larger bond issues in period t end up with higher prices in period t and

t + 1.

◮ Burdens of public debt are fully paid by current old though reductions

in the real value of their assets and by young generation through reductions in the real interest rate.

KU (Waseda) FTPL May 2014 @ Bundesbank 21 / 29

Result 2

Prices respond to aging in opposite directions, depending on whether it is caused by longer lifetime or by lower birth rate. Suppose that γt = θt/nt, population ratio. Then,

◮ Lower birth rate nt inflates prices as the standard FTPL predicts. ◮ Longer life-expectancy θt is likely to inflate prices as lower birth rate

does, if it is expected.

◮ 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

◮ Fiscal surplus declines due to a decline in tax revenues from the young,

raising tax and price.

Political impact of aging

◮ Govt opts to decrease deficits, increase taxes, and lower price.

Moreover, unexpectedly long lifetime makes the old worse-off because their savings turn out insufficient.

◮ This strengthens the government’s distributional concerns, leading to

deflation.

KU (Waseda) FTPL May 2014 @ Bundesbank 23 / 29

Concluding Remarks

Negative correlation

◮ A mild deflation and population aging in Japan’s lost decades.

Our result here suggests that this puzzling observation might be caused by

◮ the combined political and economic effects of unexpected population

aging having occurred from extension in longevity.

Future work

◮ address the accumulation in government bonds outstanding that is

• bserved in Japan

◮ introduce an endogenous monetary policy response ◮ introduce foreign investors to buy the government bonds ◮ make a quantitative analysis KU (Waseda) FTPL May 2014 @ Bundesbank 24 / 29

Appendix: quantitative analysis

KU (Waseda) FTPL May 2014 @ Bundesbank 25 / 29

Responses of Prices to Extension of Longevity

The perfect foresight case experiences inflation whereas the unexpected case experiences deflation, implying that the public expectation for aging is a key to understanding the response of the price levels.

0.98 0.99 1 1.01 1.02 1.03 0.94 0.95 0.96 0.97 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

◮ 1 period is 40 years. ◮ Utility is given by

u(cl

t, ℓl t) =

• cl

t

1−σ 1 − σ − χ

• ℓl

t

1+1/υ 1 + 1/υ , where σ = 1 and υ = 0.5.

◮ β = 0.9940 (1% annually), gT

t = 0.01, and gC t = 0.01.

◮ Variables associated with demography, θt and nt, are derived from

Japan’s official statistics and forecasts by National Institute of Population and Social Security Research (IPSS).

⋆ As an innitial state, in 1997, θt = 0.620, nt = 0.304, and

γy

t /γo t = 0.828.

⋆ As a final state, according to long-run forecasts in 2060 made in 2012,

θt = 0.781, nt = 0.555, and γy

t /γo t = 0.856.

KU (Waseda) FTPL May 2014 @ Bundesbank 27 / 29

Transition Path

Deflation due to recovered birth rate

1 2 3 0.4 0.45 0.5 0.55 0.6 0.65 Tax rate 1 2 3 1.01 1.015 1.02 1.025 1.03 1.035 Real interest rate (annualized) 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)

10 20 30 40 50 5 10 15 20 Real government bond Transition period 10 20 30 40 50

• 0.02
• 0.01

0.01 0.02 Real interest rate (annual) Transition period 10 20 30 40 50

• 0.02
• 0.01

0.01 0.02 Inflation rate (annual) Transition period 10 20 30 40 50 0.1 0.2 0.3 0.4 Income tax rate Transition period

KU (Waseda) FTPL May 2014 @ Bundesbank 29 / 29