Comparing NGDP targets to optimal
discretion and to UK policy
Simon Wren-Lewis
Economics Department and Merton College
Oxford
Structure of talk
No new research – just my take on current evidence
Compare a NGDP level target to an ideal
discretionary monetary policy regime
Assume the discretionary regime cannot undertake history dependent policy
LNGDP targets provide the commitment mechanism required to undertake time inconsistent policy
How far does current UK monetary policy depart
from ideal discretion, and does this change the argument?
Lack of dual mandate in the UK is a real problem
Zero Lower Bound is a real problem (given perverse fiscal actions)
Context
UK macroeconomic performance since 2010
has been disastrous, both in comparison to previous recoveries and the US.
I have not seen any remotely persuasive
‘structural’ reasons why this has to be so, but there are some fairly obvious policy related
explanations. (Austerity + ZLB + a string of cost-push shocks.)
For a policy mistake of this magnitude, it seems
unlikely that either fiscal or monetary policy
bears sole responsibility. It also seems sensible to ask to what extent the macropolicy regime caused or enabled this mistake.
Optimal response to a cost-push shock
With a New Keynesian Phillips curve, and if c/a=d=1, the optimal policy at
all times is to target LNGDP.
This policy is time inconsistent: a positive one period shock will imply
inflation and output below target for t>0. A LNGDP target could provide the commitment to enact this policy. (Woodford)
Is the parameter ‘d’ near 1? Estimates of ‘a’ appear small, particularly at
low inflation (e.g. IMF), but ‘c’ derived from utility can also be small (and are related). But is the latter robust, or indeed remotely plausible?
http://mainlymacro.blogspot.co.uk/2013/05/microfounded-social-welfare-functions.html
Objective (Ignore discounting) Constraint
When bygones should be bygones
If we have inflation inertia, then LNGDP or price
level targeting is costly.
For example, see various Bank of Canada studies
If the target measure of inflation is not a good proxy
for the welfare relevant (‘core’) measure of inflation, optimal discretion can ignore unexpected noise
shocks
http://mainlymacro.blogspot.co.uk/2012/06/but-which-inflation.html VAT changes, or administered prices
Commodity price shocks?
But GDP deflator may be closer to core inflation than CPI
More difficult to handle shocks to natural rate, or
Stabilising debt contracts
Kevin Sheedy: Debt and Incomplete Financial
Markets: A Case for Nominal GDP Targeting
As most life-cycle borrowing is based on fixed
nominal repayments, markets are incomplete, and agents face uncertainty due to aggregate NGDP changes.
Sheedy integrates these welfare costs
alongside standard Calvo costs of inflation, and finds the former are much more important.
NGDP vs optimal discretion: summing up
Pro
Allows history dependent policy
Stabilises debt contracts
Con
Implicit non-optimal weighting, which implies that
we should accommodate inflation shocks by much more than NGDP targets do.
If there is inflation inertia, bygones should be bygones
Discretion can ignore welfare irrelevant changes
How does current UK policy depart from
discretionary ideal?
(1) Lack of effective dual mandate
Monetary policy has two roles: setting inflation and equating
aggregate demand and supply
http://mainlymacro.blogspot.co.uk/2013/04/why-dual-mandate-is-essential.html Latter will not happen independently of monetary policy
An inflation target undercuts expectations based adjustment, but
levels targets do not.
Critical at ZLB. Inflation targets institutionalise the mistake Friedman highlighted
for the Great Depression
http://mainlymacro.blogspot.co.uk/2013/05/the-liquidity-trap-and-macro-textbooks.html
Targeting a two year inflation forecast has not proved flexible enough
to allow output gap stabilisation , but super flexible inflation targeting (Treasury 2013) risks making monetary policy opaque.
Inflation bias justification for single inflation target oversold.
Adding dual mandate will not unleash inflation bias. Public pressure to reduce
involuntary unemployment weak (contrast to stabilisation bias)
Single inflation target also puts too much weight on a single measure
of inflation. No clear justification from a social welfare perspective for focus on CPI.
How does current UK policy depart from
discretionary ideal?
(2) Fighting the last war
Even where a dual mandate exists, policy is far too cautious
about inflation
Spectre of 1970s
The strange belief in ‘over the cliff’ expectations
In fact repeating the same mistake of fighting the last war
If the 1970s was due to a desire to avoid the unemployment of the
1930s, is recession today a result of a desire to avoid the 1970s?
NGDP targets are a corrective to either the absence of a
dual mandate, or over concern about inflation (e.g. Frenkel)
Although this appears to be an argument for growth
rather than levels targets, levels targets make a rerun of the 1970s less likely
See 1970s Germany
http://mainlymacro.blogspot.co.uk/2013/04/myths-and-realities-of-Is the ZLB special?
Carney: while benefits of LNGDP targets unclear in normal times,
clearer at ZLB
http://www.bankofcanada.ca/2012/02/speeches/monetary-policy-framework-all-seasons/
But using LNGDP targets as a way of dealing with the ZLB is second
best to
fiscal policy
Central banks that encouraged austerity at the ZLB were/are guilty of the worst
type of macropolicy error.
Central banks that suggest that monetary policy at the ZLB can be as effective
as conventional policy, and that do not ask for fiscal help, are either not much better, or delusional
http://mainlymacro.blogspot.co.uk/2012/09/zero-lower-bound-denial.html
While there is a clear argument that monetary policy is the preferred stabilisation tool
when unconstrained (Eser, Leith and WL), case has not been made that QE is superior to fiscal policy
Central banks have a duty to provide what is textbook macroeconomic
wisdom to the public, and to be honest about the unpredictability of unconventional monetary policy
In the UK, it is arguable (Osborne) that a more optimal fiscal policy would have been offset by a more restrictive monetary policy, given the over reation of monetary policy to cost-push shocks.
But currently fiscal policy is not optimal
Now monetary policy is the only game in town.
The Treasury seems also to have ruled out a
dual mandate
QE and FLS useful but not enough.
So need to raise expectations about future
inflation in the short term
How can this be achieved?
A Fed style policy of forward guidance one
possible route, but for the UK a commitment to allow inflation to go to, say, 3% might not
A proposal
Christine Romer
Quoted here:
http://macromarketmusings.blogspot.co.uk/2012/02/christina-romer-we-need-regime-change.html
“.. one way out of a recession at the zero lower bound is by changing
expectations. To do that, often what is needed is a very strong change in policy – something economists call a “regime shift”. The most effective way to shake an economy out of a terrible downturn when we’re at the zero lower bound is an aggressive change in policy that makes people wake up, say “this is a new day” and change their expectations. What the Fed has done since early 2009 is much more of an incremental change.”
Will Japan prove to be a case study of this point?
The proposal: within current framework, adopt LNGDP as an
intermediate target as “a guide to achieving inflation target in most appropriate and efficient manner.”
Of course to shift expectations, the LNGDP path adopted would have
to start from pre-recession levels, with a conservative guess at the amount of UK output permanently lost as a result of the recession.
Conclusion
Arguments for LNGDP rather than optimal discretion are
finely balanced
However monetary policy is currently not optimal
because it is haunted by the 1970s, and outside the US does not operate a dual mandate
The optimal response to the ZLB problem is fiscal
expansion, but in its absence current monetary policy has not produced an optimal inflation/output gap mix.
To change this mix given existing tools will probably
require a shift in expectations beyond what forward guidance might provide.
Adopting LNGDP as an intermediate target to efficiently