Transcript Alternative inflation measures 2013 May

Alternative Inflation Measures
Comparison: core vs. overall
• Overall measures: very volatile, especially
because of food & energy shocks
• Core = excluding food & energy here; but
research works on alternatives
Comparison across measures
• Examine GDP price index, PCE price
index, CPI
Differences between measures
• Coverage (scope)
– CPI: out of pocket spending of urban
consumers
– PCEPI: prices of all personal consumption
expenditures in NIPA
– GDPPI: price index of prices of all goods
produced in economy (70% of which is in
PCEPI)
Differences between measures
• Weights:
– PCEPI and GDPPI: weights updated every
quarter; based on business surveys
(production measures)
– CPI: weights updated every two years; based
on household surveys (out-of-pocket
expenditure measures)
Differences between measures
• Revisions:
– PCEPI and GDPPI: revised as time passes to
better reflect true weights across sample of
businesses
– CPI: never revised, except for seasonal
adjustment
Differences between measures
• Bias:
– PCEPI and GDPPI: small bias because of
quality adjustment bias; around 0.5% for
PCEPI and 0.4% for GDPPI
– CPI: larger bias because of quality adjustment
bias and substitution bias; at least 1.0%
Differences between measures
• Examples
– Housing prices are 15% of PCEPI, but 32% of
CPI (urban)
– Medical prices are 20.5% of PCEPI, but only
6.2% of CPI (out of pocket)
Data revisions
• GDPPI and PCEPI are revised; CPI is not
• That means the CPI is worse, not better
Data revisions
• Policy implications: the Fed could be
confused by initial data, as it was in 2003
Data revisions
• In 2000, Fed switched main variable for
inflation to PCE price index (PCE
inflation); in 2004 switched to PCE price
index excluding food and energy prices
(core PCE inflation); in 2007 it began
using both
• Problem: these variables get revised
• Issue: are the revisions large enough to
worry about?
Data Revisions
• May 2002: data show decline in core PCE
inflation from 2.0% in 2000Q3 to 1.2% in
2002Q1
• Academic research on deflation and the
zero bound are fresh in policymakers’
minds
Core PCE inflation rate from 1997Q1 to 2002Q1, as observed May 2002
2.4
2.2
Inflation rate (percent)
2.0
1.8
1.6
1.4
1.2
1.0
1997
1998
1999
2000
Year
2001
2002
Data revisions
• Perhaps as a consequence of worry about low
inflation, Fed drives real fed funds rate to
negative levels for first time since early 1970s
• Fed adds phrase to FOMC statement that it
worries about an “unwelcome fall in inflation”
• But: revised data by December 2003 show that
inflation wasn’t declining after all
v=May2002
v=Dec2003
2000Q3
2.0%
1.7%
2002Q1
1.2%
1.5%
Core PCE inflation rate, May 2002 and December 2003
2.4
2.2
Inflation rate (percent)
2.0
OBSERVED
DECEMBER 2003
1.8
1.6
1.4
OBSERVED MAY 2002
1.2
1.0
1997
1998
1999
2000
Year
2001
2002
Data revisions
• The Fed gets rid of the “unwelcome fall”
language by May 2004. Revised data by
August 2005 show Fed should have
worried about an unwelcome rise in
inflation
2000Q3
2002Q1
v=May 2002 2.0%
1.2%
v=Dec2003
1.7%
1.5%
v=Aug2005
1.6%
1.8%
Core PCE inflation rate, May 02, Dec 03, & Aug 05
2.4
AUGUST 2005
2.2
DECEMBER 2003
Inflation rate (percent)
2.0
1.8
1.6
1.4
MAY 2002
1.2
1.0
1997
1998
1999
2000
Year
2001
2002
Data revisions
• Policymakers need to understand
revisions to inflation
Figure 5
Four Quarter Inflation Rate in PPCEX
Initial to August Revision and Actuals
3.0
Actuals and Revisions (percent)
Initial
2.5
August
2.0
1.5
1.0
0.5
0.0
Revision
-0.5
-1.0
1995
1996
1997
1998
1999
2000
Date
2001
2002
2003
2004
2005
Figure 6
Revisions to Four Quarter Inflation Rate in PPCE and PPCEX
Initial to August Revision
0.8
Revisions (percent)
0.6
0.4
PPCE
0.2
0.0
-0.2
-0.4
1995
PPCEX
1996
1997
1998
1999
2000
Date
2001
2002
2003
2004
2005
Chained CPI
• The chained CPI uses chain weighting to
improve on the overall CPI
• It works, but to do so, it must be revised
• Data are final after two years
• Revisions cause problems for changing
payments, as in Social Security
Conclusions
• There is no perfect measure of inflation
• Fed uses multiple measures
• Using core inflation prevents overreaction
to short-term shocks
• Fed should be aware of revisions
• The chained CPI is more accurate than
the overall CPI