View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

D ISCUSSION OF “D EBT AND I NCOMPLETE
F INANCIAL M ARKETS ,” BY K EVIN S HEEDY
James Bullard
President and CEO
Federal Reserve Bank of St. Louis

Brookings Institution
Washington, D.C.
21 March 2014
Any opinions expressed here are mine and do not necessarily reflect those of others on the Federal Open Market Committee.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

M ONETARY POLICY RATIONALES

Leading rationale: “Sticky price friction prevents the market
solution from being fully optimal.”
Associated policy advice: “Keep prices stable along a price level
path.”

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

A N ALTERNATIVE MONETARY POLICY RATIONALE

An alternative rationale: “Non-state contingent nominal
contracting friction keeps the market solution from being fully
optimal.”
Associated policy advice: “Move the price level
counter-cyclically in response to aggegate income shocks."
This advice is quite different in nature!

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

N ICE PAPER

This is a very nice paper that lays out considerable intuition for
the alternative rationale.
Prof. Sheedy has set the standard for future analyses in this area.
The paper includes commentary on an extensive related
literature.
The paper also includes a tug-of-war between sticky prices and
non-state contingent nominal contracting.
In a calibrated model with both frictions present, the non-state
contingent nominal contracting friction is the more salient for
policymakers.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

T HREE QUESTIONS

I will organize my discussion around three questions.
The model seems “special.” Would these results hold in a general
equilibrium life-cycle model with many heterogeneous
participants in a large private credit market?
Tentatively, yes.

What are some of the key questions on which future research in
this area should focus?
Whither nominal GDP targeting?
The paper has much to say in framing the debate on the wisdom of nominal
income targeting.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

A More Realistic Version

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

I S THE MODEL SPECIAL ?

The Sheedy model has two types of households, relatively
impatient and relatively patient.
Since there are just two types of agents, there is only one set of
marginal conditions that requires “repair.”
The policymaker has just one tool, the price level, which neatly
fixes the marginal conditions.
Question: Would the results carry over to a more realistic
environment with more heterogeneity in the private credit
market?

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

L IFE - CYCLE VERSION

To investigate: Consider a stripped-down, endowment general
equilibrium life-cycle economy.
I will describe a “quarterly” specification, with households living
241 periods.
The odd number of periods provides a “middle” period at which
life-cycle income peaks.

Interpretation: Cohorts begin participation in the economy at
age 20, die at age 80, and are most productive in the middle
period, age 50.
See Sheedy (2013) for an analysis of a three-period model.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

K EY FRICTION

Sheedy: Loans are dispersed and repaid in the unit of account—that
is, in nominal terms—and are not contingent on income realizations.
Other assumptions I am making: Within-cohort agents are
identical, no population growth, inelastic labor supply,
time-separable log preferences, no discounting, no capital, no
default, flexible prices, no borrowing constraints, no government
other than the central bank.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

L IFE - CYCLE PRODUCTIVITY

All agents are endowed with an identical productivity profile
over their lifetime.
The profile begins at zero, rises to a peak at the middle period of
life, and then declines to zero.
Agents can sell productivity units in the labor market at the
competitive wage.
The productivity profile is symmetric.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

L IFE - CYCLE INCOME PROFILE
1.0

0.8

0.6

0.4

0.2

5

10

15

20

F IGURE : A schematic life-cycle income profile of a typical household. About
50 percent of the households earn 75 percent of the income.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

A GGREGATE SHOCKS
Sheedy: The only source of uncertainty is an aggregate shock.
Accordingly, assume the real wage is exogenous and grows at
gross rate λ (t) , where we think of λ (t) as near unity and
8 H
w.p. 1/3
< λ
λ w.p. 1/3
λ (t) =
: L
λ
w.p. 1/3
and λH

λ

λL and where λ = λH + λL /2.

Thus, the mean growth rate of national income is λ.
If λH = λ = λL , there is no uncertainty.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

T HE POLICYMAKER

The policymaker completely controls the price level, which is a
unit of account in this model.
To see a two-period example along this line, see Evan Koenig
(2013, IJCB).
The within-period timing protocol is as follows: (1) nature
chooses the growth rate, (2) policymaker chooses a price level,
and (3) households make decisions to consume and save.
For now, let’s think of the policymaker choosing P (t) = 1 8t,
“price stability.”

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

N ATURE OF THIS ECONOMY

The economy I have described has 241 different households,
each with its own level of asset holding.
To calculate the full stochastic equilibrium, in principle, one has
to keep track of the evolution of the distribution of asset
holdings over time.
If there was no uncertainty, a non-stochastic steady state has
interesting properties that fit well with the intuition offered by
Prof. Sheedy.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

Steady State

W HITHER NOMINAL GDP TARGETING ?

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

N ON - STOCHASTIC STEADY STATE NET ASSET HOLDING
2

1

5

10

15

20

1

2

F IGURE : Schematic asset holding by cohort when the system is in a
non-stochastic steady state. About 25 percent of the population holds about
75 percent of the assets.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

H OW LARGE ARE THESE MARKETS ?

According to Mian and Sufi (AER, 2011), the ratio of household
debt to GDP was about 1.15 before the increase during the 2000s
when it ballooned to 1.65.
In today’s dollars, that would be about $19.5 trillion to about $28
trillion, comprised mostly of mortgage debt.
Messing up these markets might be quite costly for the economy,
so this friction could be quite important.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

N ON - STOCHASTIC STEADY STATE CONSUMPTION
1.0

0.8

0.6

0.4

0.2

5

10

15

20

F IGURE : Non-stochastic steady state consumption by cohort. The private
credit market completely solves the point-in-time income inequality problem.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

K EY FEATURE OF THE NON - STOCHASTIC STEADY STATE

By careful choice of assumptions, the general equilibrium gross
one-period real interest rate is equal to the gross real output
growth rate in the steady state; that is,
R = λ.
This value for the real interest rate effectively means that all
income earned within a period is divided equally among all
participants alive in the economy at that time.
Prof. Sheedy’s excellent intuition: This means all households have an
“equity share” in the economy.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

Main Finding

W HITHER NOMINAL GDP TARGETING ?

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

T HE STOCHASTIC CASE

Now allow aggregate shocks, maintaining the price stability
policy P (t) = 1 8t.

State-contingent loan contracts are ruled out by assumption.
When there is a shock, consumption will now be allocated
unevenly across households alive at date t.
To bear risk appropriately across the economy’s participants, the
price level policy should produce something like an “equity
share” of any surprise movements in income.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

C OUNTER - CYCLICAL PRICE LEVEL POLICY
In the three-period version, the gross real interest rate R that
prevails between dates t and t + 1 is:
R=

P (t + 1) w (t + 1)
.
P (t)
w (t)

(1)

Take P (t) = 1 and w (t) = 1 as given, allow nature to choose a
rate of growth of real wages (and hence national income), and
then let the policymaker choose P (t + 1) to make R = λ, the
average growth rate in the economy.
If growth is high, P (t + 1) < 1, which is a decrease in the price
level—hence policy is “counter-cyclical.”
This policy is “real interest rate smoothing,” as R never moves in
response to shocks.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

E QUITY SHARES

This (perfectly credible) policy will again cause all consumption
available in the economy at date t to be split evenly between
living households.
If there is more income than expected, all get more to consume,
or if there is less income than expected, all get less.
Everyone bears the risk of real income shocks equally—the “equity share”
contract.

Because shocks have a simple structure in this example, this
policy also works for the 241-period model.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

B OTTOM LINE

The bottom line: Versions of the key result presented in this paper
may also hold in a wide class of general equilibrium life-cycle economies
with many different participants in the private credit market.
This is encouraging for this line of research.
Still, there are many questions ...

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

Some Directions for Future Research

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

C HANGES IN THE LONG - RUN GROWTH RATE

To implement the policy suggested here, the policymaker has to
know the long-run growth rate of the economy.
An incorrect estimate of that growth rate would cause the
policymaker to inadvertently, but persistently, distort the private
credit market.
This might cause considerable damage relative to simply
ignoring the non-state contingent nominal contracting friction.
A better handle on this issue would be a helpful addition to this
literature.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

G OOD - BYE TO POSITIVE INFLATION TARGETS ?

The simple formula above suggests that positive inflation targets
are ill-advised in this class of models, as they permanently
distort the real interest rate away from the real output growth
rate.
In models like the one presented here, average inflation should
be zero.
Messing with the price level is serious business in this economy and must
be done only in response to a shock.

I think it would be interesting to get a better handle on this
aspect of the suggested monetary policy.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

D EFAULT

Many have argued that the presence of default in actual
economies indicates that a type of state contingency does exist
after all, albeit perhaps in more extreme cases.
Understanding the role of default arrangements, especially
endogenous debt constraints, is the subject of a large literature.
In life-cycle economies, for instance, see Azariadis and Lambertini
(2003, RES).

Future research should try to integrate that literature more fully
with the arguments for a counter-cyclical price level policy.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

H ARMING CASH USERS ?

This model has no money demand.
In the U.S. economy, perhaps 15 percent of households are
unbanked and perhaps another 15 percent are nearly unbanked.
These intensive cash users tend to be poor and may not
appreciate a policy aimed solely at credit markets.
A fully satisfactory model would have this segment of the
population included.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

Whither NGDP Targeting?

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

W HITHER NOMINAL GDP TARGETING ?

This paper describes the optimal policy in the non-state
contingent nominal contracting world as “nominal GDP
targeting.”
This policy is the polar opposite of what one would obtain from a
purely sticky price framework.
Yet many today argue from a sticky price perspective that nominal
GDP targeting may provide a good policy benchmark.
See, for instance, Woodford (2012, Jackson Hole Symposium).

We need to sharpen up this debate—NGDP targeting cannot be
all things to all people.

I NTRODUCTION

L IFE - CYCLE VERSION

S TEADY S TATE

M AIN F INDING

S HARPENING THE NGDP

M ANY Q UESTIONS

W HITHER NOMINAL GDP TARGETING ?

TARGETING DEBATE

My unsolicited advice:
Those advocating NGDP targeting have long needed a model.
Advocates of NGDP targeting should embrace the Sheedy class of
models.
Variations in inflation would be deliberate and systematic.

Those wishing to remain with the sticky price rationale should
argue for stable prices.
Variations in inflation would be limited, as they have been in the U.S.
since 1995.