Quantified perceived and Expected Inflation in the Euro Area European Commission

advertisement
European Commission
Directorate General Economic and Financial Affairs
Quantified perceived and Expected
Inflation in the Euro Area
How Incentives Improve Consumers Inflation Forecasts
Presentation by Staffan LINDÉN
Joint EC-OECD Workshop on International Developments of
Business and Consumer Tendency Surveys
14-15 November 2005
European Commission 2005
Purpose
• Investigate whether differences in
incentives to collect information can explain
the gap between the perceived and
expected inflation rates, and the official rate.
– Inflation is in general perceived and expected to
be higher than the official rate
• Present the data
European Commission 2005
Outline
• Costly information and how strong incentives to
form expectations improve projections
• Data
– Stylised facts from the qualitative data
– Comparison of qualitative and quantitative data
– Time series properties
• Perceived and expected inflation when there are
incentives to collect information
• Conclusions
European Commission 2005
Main message
• Incentives to collect information on inflation induce
respondents to produce inflation rates that are
closer to the official rate.
• It is important to ask questions that are relevant,
and that respondents have the information
necessary to give informed answers.
• The answers to the quantitative questions behave
in a similar way as the qualitative.
• The dataset is becoming unique, almost 400,000
observations just for the euro area
European Commission 2005
The role of costly information
• Hypothesis: incentives to collect information on
inflation induce respondents to produce inflation
rates that are closer to the official rate.
• There is a trade-off between informational
efficiency and the incentives to acquire
information.
• It is costly to collect and process information
• If respondents have no incentives to do so, they
will not inform themselves.
• This leads to noise in the indicators.
European Commission 2005
Three survey questions inducing people to
collect information of inflation
• Q13:
• Q14:
How likely are you to buy a car?
Are you planning to buy or build a
home?
• Q15:
How likely are you to spend any
large sum of money on home
improvements?
• It is assumed that these activities gives
respondents different incentives to obtain
costly information.
European Commission 2005
The Balance Statistic for perceived and 12
months lagged expected inflation
6.0
70.0
60.0
5.0
50.0
4.0
40.0
3.0
30.0
20.0
2.0
10.0
1.0
0.0
0.0
-10.0
91
92
93
94
95
96
97
HICP
98
99
Per BS
00
01
02
03
04
05
Exp BS
European Commission 2005
Stylised facts regarding the qualitative data
• There are structural shifts in perceptions and
expectations after 1 January 2002.
• There is a de-linkage between perceived inflation
and actual inflation.
• Perceived inflation is over stated after January
2002.
• Perceptions have been converging towards the
actual inflation rate.
• Expectations are more in line with the actual
inflation than perceptions after January 2002.
European Commission 2005
Perceived and expected inflation in the euro area
Income of the household
Income of the household
46.3
12.8
11.7
43.3
11.4
41.1
37.5
10.0
6.8
6.3
6.1
5.5
11.1
1st quart
2nd quart
3rd quart
Education
12.2
4th quart
1st quart
Gender
2nd quart
45.3
Secondary
Further
Male
Female
Primary
8.0
12.1
6.3
Secondary
6.3
Q51
Male
10.7
6.0
50-64
Q61
43.6
42.5
11.3
30-49
Further
8.1
Female
Age
43.8
37.7
5.4
5.1
16-29
8.5
8.1
Age
11.8
39.5
6.4
8.8
Primary
4th quart
Gender
35.9
5.7
5.1
3rd quart
41.7
10.6
6.0
8.0
46.2
9.3
6.3
8.9
Education
12.6
11.6
9.2
65+
16-29
10.5
8.4
30-49
50-64
Q5
8.0
65+
Q6
European Commission 2005
Stylised facts continued
• Inflation perceptions and expectations fall
as income increases.
• Inflation perceptions and expectations fall
as education increases.
• Perceived inflation seems to increase with
age, and expected inflation is humpshaped.
• Women perceive and expect higher inflation
than men.
European Commission 2005
Quantitative perceived and 12 month lagged
expected inflation rates
20.0
18.0
16.0
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
Jan-03
Jan-04
HICP
Jan-05
Per EA-9 +FR
Jan-06
Exp EA-9 +FR
European Commission 2005
What are the findings from the comparison
• Quantitative data gives the same results as
the qualitative.
• The euro area data replicates the results
obtained using other similar datasets.
• There has been a convergence process
going on.
European Commission 2005
Perceived and expected inflation when
incentive to obtain information is different
• The incentives are assumed to come from
the purchase of a car, a house, or spending
a large sum of money on home
improvement
• Figures to compare with
– Overall perceived inflation is 11.6%
– Overall expected inflation is 6.1%
– Official HICP rate of inflation is 2.2%
European Commission 2005
Perceived inflation rates depending on
likelihood to buy or invest
Buying a car
Buying a house
Home improvement
13.8
11.9
11.7
12.0
12.0
10.3
Very
likely
Fairly
likely
not likely not at all
likely
11.5
9.6
9.5
Don't
know
Very
likely
Fairly
likely
12.8
12.2
11.7
10.8
10.9
9.8
not likely not at all
likely
Don't
know
Very
likely
Fairly
likely
not likely not at all
likely
European Commission 2005
Don't
know
Expected inflation rates depending on
likelihood to buy or invest
Buying a car
Buying a house
Home improvement
7.0
6.2
6.0
6.2
6.1
Fairly
likely
6.2
6.1
6.1
6.2
5.9
5.8
5.3
Very
likely
6.8
6.0
5.2
not likely not at all
likely
Don't
know
Very
likely
Fairly
likely
not likely not at all
likely
Don't
know
Very
likely
Fairly
likely
not likely not at all
likely
European Commission 2005
Don't
know
Results from incentive induced inflation
perceptions and expectations
• The more likely respondents are to buy a car, a
house, or make home improvements, the closer
the reported inflation rate is to the official rate.
– Perceptions: respondents that are ‘fairly likely’ are 0.9
p.p. to 2.4 p.p. below the ‘not at all likely’
– Expectations: respondents that are ‘fairly likely’ are 0.9
p.p. below the ‘not at all likely’
• The exceptions are for those respondents that are
‘very likely’ to do so
European Commission 2005
Adjusting for sample size
• The relatively high inflation rates for the
‘very likely’ are caused by outliers in the
German data.
• Small samples are over represented in the
euro area aggregate.
• Two combined solutions
– Re-group the data into two categories
– When calculating the euro area inflation rates,
adjust for both country and sample size.
European Commission 2005
Perceived inflation rates depending on likelihood to
buy or invest – regrouped and weighted by size
Buying a car
Buying a house
18.8
8.9
likely
10.6
not likely
8.7
don’t
know
Home improvement
19.8
likely
10.9
not likely
18.7
9.4
don’t
know
likely
10.9
not likely
European Commission 2005
don’t
know
Expected inflation rates depending on likelihood to
buy or invest – regrouped and weighted by size
Buying a car
Buying a house
10.4
9.9
4.8
likely
Home improvement
6.4
5.8
4.9
4.6
not likely
9.6
don’t
know
likely
not likely
don’t
know
likely
5.9
not likely
European Commission 2005
don’t
know
Results from incentive induced inflation perceptions
and expectations – regrouped and weighted by size
• The more likely respondents are to buy a car, a
house, or make home improvements, the closer
the reported inflation rate is to the official rate.
• Perceptions: respondents answering ‘likely’ are 1.5
p.p. to 2.2 p.p. below the ‘not likely’
• Expectations: respondents answering ‘likely’ are 1
p.p. to 1.8 p.p. below the ‘not likely’
• The levels are still above the official rates, but…
• …for expectations the distance is almost halved.
European Commission 2005
Conclusions
• The quantitative data is of a high quality.
– It behaves well according to the benchmark.
• Incentives to obtain information on inflation
improves the survey responses.
– The expectations decline by as much as 1.8
p.p. (the overall rate falls from 6.1% to 4.3%,
which still is above the official rate of 2.2%).
• The gap to the official rate is thus partly
explained
European Commission 2005
Conclusions continued
• It is important to ask questions that are
relevant, and that respondents have the
information necessary to give informed
answers.
• It can be useful to “cut” the data in different
ways to improve the survey results.
European Commission 2005
Relevance of the questions
• For most questions in the harmonised
questionnaires respondents have the
information readily available.
• For the questions on price developments,
less so.
– Different consumption baskets
– Macro-variable that requires a lot of information
to forecast
– The more detailed the question, the more
information is necessary
European Commission 2005
Quantifying qualitative data
Three different methods are used
• The Balance Statistic
• The Carlson-Parkin Method
– The probability approach
• The Anderson Method
– The regression approach
European Commission 2005
Perceived inflation quantified by using the
Carlson-Parkin and the Anderson methods
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
91
92
93
94
95
96
97
HICP
98
99
C-P
00
01
02
03
04
05
Anderson
European Commission 2005
Expected inflation lagged 12 months, quantified by
using the Carlson-Parkin and the Anderson methods
6.0
5.0
4.0
3.0
2.0
1.0
0.0
91
92
93
94
95
96
97
HICP
98
99
C-P
00
01
02
03
04
05
06
Anderson
European Commission 2005
Perceived and 12 month lagged expected inflation the Anderson method adjusted for the structural break
6.0
5.0
4.0
3.0
2.0
1.0
0.0
91
92
93
94
95
96
HICP
97
98
99
Per And+Dum
00
01
02
03
04
05
06
Exp And+Dum
European Commission 2005
Correlations between perceived and
expected inflation, and actual inflation
Correaltion between
HICP rate of inflation
Jan 1991 to Sep 2005
Jan 1991 to Dec 2001
Jan 2002 to Sep 2005
Apr 2004 to Sep 2005
Perceived inflation for the past 12 months
BS
C-P
And
And & Dum
0.48
0.22
0.55
0.55
0.89
0.80
0.82
0.80
-0.19
-0.05
0.01
0.03
0.40
-0.17
0.24
0.25
Correaltion between
HICP rate of inflation
Jan 1991 to Sep 2005
Jan 1991 to Dec 2001
Jan 2002 to Sep 2005
Jul 2004 to Sep 2005
Expected inflation for the next 12 months 1
BS
C-P
And
And & Dum
0.64
0.65
0.71
0.71
0.83
0.80
0.80
0.79
0.08
-0.03
0.15
0.20
0.11
-0.09
0.04
0.06
1 Expected inflation is lagged by 12 months
European Commission 2005
Comparison with other national datasets
• The above mentioned stylised facts are
present in similar datasets from the US and
Sweden.
• For income, education, and gender the
results are the same.
• It differs for age.
European Commission 2005
Correlations between quantitative,
qualitative, and actual inflation
Correlation between
Percieved inflation Q51
HICP rate of inflation
Perceived inflation for the past 12 months (May03-Sep05)
Q51
BS
CP
And
And & Dum
1.00
0.91
0.58
0.78
0.74
-0.42
-0.43
-0.57
-0.33
-0.29
Correlation between
Expected inflation Q61
HICP rate of inflation
HICP rate of inflation (lag 6)
HICP rate of inflation (lag 9)
HICP rate of inflation (lag 12)
Expected inflation for the next 12 months (May03-Sep05)
Q61
BS
CP
And
And & Dum
1.00
-0.15
-0.14
0.04
0.06
-0.12
0.53
0.46
0.23
0.17
-0.03
0.01
0.01
0.11
0.11
0.39
0.51
0.49
0.09
0.01
-0.35
-0.14
-0.30
-0.17
-0.13
European Commission 2005
Download