COMPILATION OF ITALIAN HICP BY DIFFERENT GROUPS OF HOUSEHOLDS
Geneva, 2023, June 7-9
UNECE CPI Expert Group meeting
Ilaria Arigoni, Istat (Italy) ([email protected])
Alessandro Brunetti, Istat (Italy) ([email protected])
Valeria de Martino, Istat (Italy)([email protected])
Federico Polidoro, World Bank ([email protected])
Outline
2
• Inflation in Italy and in the Euro area in 2022-2023: an overview
• Current Istat methodology to compile HICP by five groups of households
• Changing from expenditure to income the variable to identify the groups of
households: main outcomes
• The impact of inflation on the income-based groups of households
• Characteristics of the households in the extreme groups and comparison
between their distributions in the five groups by expenditure and by income
• Is it enough focusing on the weights to measure the actual impact of inflation
on the poorest people?
• Some concluding remarks and perspectives
UNECE CPI Expert Group meeting 2023, June 7-9
Inflation in Italy and in the Euro area in 2022-23: an overview
3
• 2022, as well the final part of 2021, have been characterized, in Italy, in the
European Union (EU) and in the world, by a sharp increase of the rates of change
of consumer price indices that are slowly decreasing in the first part of 2023
• Italian inflation measured by HICP has raised from +1.0% in July 2021 (+2.2% in
the Euro Area) to +12.5% in November 2022 (+10.1% in the Euro area), slowing
down respectively to +7.0% and +8.7% in April 2023.
• Given the impact of energy prices on the sharp raise and on the recent slowdown
of inflation, the overall HICP excluding energy has gone on speeding up (arriving in
March 2023 at +7.9% in the EA and at +6.9% in Italy) and starting declining only in
April 2023 (+7.4% the EA; +6.7% in Italy)
• Yearly rates of change of food prices are still very high (in April 2023 +13.5% in the
EA, +11.0% in Italy)
UNECE CPI Expert Group meeting 2023, June 7-9
Inflation in Italy and in the Euro area in 2022-23: an overview
4
Figure 1. HICP Indices and annual rates of change. Italy and Euro area. 2016 – 2023. Percentage values
7.0
8.7
123.13
121.4
80.0
85.0
90.0
95.0
100.0
105.0
110.0
115.0
120.0
125.0
130.0
-2.0
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
Ja
n-1
6
Ap
r-1
6
Ju
l-1
6
Oc
t-1
6
Ja
n-1
7
Ap
r-1
7
Ju
l-1
7
Oc
t-1
7
Ja
n-1
8
Ap
r-1
8
Ju
l-1
8
Oc
t-1
8
Ja
n-1
9
Ap
r-1
9
Ju
l-1
9
Oc
t-1
9
Ja
n-2
0
Ap
r-2
0
Ju
l-2
0
Oc
t-2
0
Ja
n-2
1
Ap
r-2
1
Ju
l-2
1
Oc
t-2
1
Ja
n-2
2
Ap
r-2
2
Ju
l-2
2
Oc
t-2
2
Ja
n-2
3
Ap
r-2
3
Euro area m/m-12 (left axis) Italy m/m-12 (left axis) Euro area index Italy Index
UNECE CPI Expert Group meeting 2023, June 7-9
Current Istat methodology to compile HICP by five groups of households
5
• Since 2005, Istat has been compiling and disseminating a measure of the impact of
the inflation on five different groups of households of equal dimension ordered by
their spending power (from the lowest of the first group to the highest of the fifth) used
as a proxy of their income conditions.
• Indices of consumer prices are compiled considering the different structure of
consumption expenditure of each group of households (summarized in the system of
weights).
• HICPs by population subgroups are “satellite” indices of HICP: they share the set of
basic information (basket of products and price elementary data) and the
methodology of Italian HICP, but they are different each other for the system of
weights used for their calculation.
UNECE CPI Expert Group meeting 2023, June 7-9
6
• Weights for the five subgroups of households based on HBS data
• To estimate the weights, consumption expenditures are equivalized by using an
appropriate equivalence scale (Carbonaro scale), that considers the effects of
economies of scale and makes them comparable to that of a two-member household,
and, as such, among different-size households
• Households ordered by equivalent consumption expenditure, are organized by specific
cut-point values and divided into five groups of equal size (equivalent-expenditure
fifths)
• In a situation of perfect equality, a share of 20% of the total expenditure sustained by
all the households would be placed in each fifth: actually, in 2021, in terms of
equivalent expenditure, that of the last fifth was about 5 times that of the first fifth
(inequality measure on expenditure side).
Current Istat methodology to compile HICP by five groups of households
UNECE CPI Expert Group meeting 2023, June 7-9
7
Figure 2. Expenditure weights by 5 households groups and main special aggregates in 2023 (HBS year 2022)
UNECE CPI Expert Group meeting 2023, June 7-9
Current Istat methodology to compile HICP by five groups of households
8
• The methodology adopted using households’ expenditure data has been transferred to
households’ income data derived from HBS
• In this case to detect the households’ groups, households’ income data are equivalized
by using OECD-modified equivalence scale
• Households ordered by equivalent income, are organized by specific cut-point values
and divided into five groups of equal size (equivalent-income fifths) from the poorest
one (the first) to the wealthiest one (the fifth)
• Thus, the weights are estimated using the expenditure data of each income group of
households
• The inter quintile ratio between the fifth and the first income group is equal to 3.9,
whereas the fifth group spends 2.07 times what the first group spends (in 2021)
From expenditure to income the variable to identify the groups of households: main outcomes
UNECE CPI Expert Group meeting 2023, June 7-9
9
Figure 3. Expenditure and income weights. First group of households. Main special aggregates (2022, HBS 2021)
UNECE CPI Expert Group meeting 2023, June 7-9
From expenditure to income the variable to identify the groups of households: main outcomes
219,419
112,662
145,527
207,524
100,205
115,613
0
50,000
100,000
150,000
200,000
250,000
P
ro
c
e
ss
e
d
fo
o
d
(
in
c
l.
a
lc
o
h
o
l
a
n
d
to
b
a
c
c
o
)
U
n
p
ro
c
e
ss
e
d
fo
o
d
E
n
e
rg
y
N
o
n
-e
n
e
rg
y
in
d
u
st
ri
a
l
g
o
o
d
s
S
e
rv
ic
e
s
re
la
te
d
t
o
h
o
u
si
n
g
S
e
rv
ic
e
s
re
la
te
d
t
o
c
o
m
m
u
n
ic
a
ti
o
n
S
e
rv
ic
e
s
re
la
te
d
t
o
re
c
re
a
ti
o
n
S
e
rv
ic
e
s
re
la
te
d
t
o
tr
a
n
sp
o
rt
s
S
e
rv
ic
e
s
m
is
c
e
ll
a
n
e
o
u
s
Exp weights first group
Income weights first group
10 UNECE CPI Expert Group meeting 2023, June 7-9
From expenditure to income the variable to identify the groups of households: main outcomes
Figure 4. Expenditure and income weights. Fifth group of households. Main special aggregates (2022, HBS 2021)
115,474
49,340
67,376
124,251
53,387
79,555
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
P
ro
c
e
ss
e
d
f
o
o
d
(
in
c
l.
a
lc
o
h
o
l
a
n
d
t
o
b
a
c
c
o
)
U
n
p
ro
c
e
ss
e
d
f
o
o
d
E
n
e
rg
y
N
o
n
-e
n
e
rg
y
i
n
d
u
st
ri
a
l
g
o
o
d
s
S
e
rv
ic
e
s
re
la
te
d
t
o
h
o
u
si
n
g
S
e
rv
ic
e
s
re
la
te
d
t
o
c
o
m
m
u
n
ic
a
ti
o
n
S
e
rv
ic
e
s
re
la
te
d
t
o
re
c
re
a
ti
o
n
S
e
rv
ic
e
s
re
la
te
d
t
o
tr
a
n
sp
o
rt
s
S
e
rv
ic
e
s
m
is
c
e
ll
a
n
e
o
u
s
Exp weights fifth group
Income weights fifth group
11
• The structure of weights is similar but not the same between those referred to the
groups detected by expenditure data and those referred to the groups detected by
income data
• Specifically:
✓ For the first group (low income/low expenditure) the weights of the aggregates affected by
higher increase of consumer prices in 2022, decrease in relative terms (by 1.19
percentage points for unprocessed food, by almost 3 p.p. for Energy)
✓ For the fifth group (high income/high expenditure), vice versa the weights of the aggregates
affected by higher increase of consumer prices in 2022, increase in relative terms (by
0.88 percentage points for unprocessed food, by almost 1.22 p.p. for Energy)
From expenditure to income the variable to identify the groups of households: main outcomes
UNECE CPI Expert Group meeting 2023, June 7-9
The impact of inflation on the income-based groups of households
12 UNECE CPI Expert Group meeting 2023, June 7-9
14.5
10.7
17.9
9.9
-5.0
0.0
5.0
10.0
15.0
20.0
F
eb
-1
8
A
p
r-
18
Ju
n
-1
8
A
u
g
-1
8
O
ct
-1
8
D
ec
-1
8
F
eb
-1
9
A
p
r-
19
Ju
n
-1
9
A
u
g
-1
9
O
ct
-1
9
D
ec
-1
9
F
eb
-2
0
A
p
r-
20
Ju
n
-2
0
A
u
g
-2
0
O
ct
-2
0
D
ec
-2
0
F
eb
-2
1
A
p
r-
21
Ju
n
-2
1
A
u
g
-2
1
O
ct
-2
1
D
ec
-2
1
F
eb
-2
2
A
p
r-
22
Ju
n
-2
2
A
u
g
-2
2
O
ct
-2
2
D
ec
-2
2
income weights first group Income weights fifth group Exp weights first group Exp weights fifht group
Figure 5. Inflation impact on the first and the fifth (by expenditure and income) groups of households. All-item
index. M/M-12 rate of change, January 2018 – December 2022
The impact of inflation on the poorest group of households in Italy
13 UNECE CPI Expert Group meeting 2023, June 7-9
• The differences highlighted in the weights between the extreme groups considered
either in terms of equivalent expenditure or in terms of equivalent income, do not
produce gap in terms on impact of inflation between the first and the fifth group in the
years when inflation is relatively low (2018 – 2020)
• As soon as the price increase becomes heterogeneous across the different product
aggregates, with energy prices sharply growing at rate strongly higher than that of
other aggregates, followed by food products, the inflation gap between the first and the
fifth group (considered both in terms of expenditure and income) starts enlarging (end
2020)
• Given the differences in the structure of weights of the 2 extreme groups, the gap
between the two groups detected by expenditure data becomes gradually wider than
that between the two groups detected by income data (in December 2022 it is equal to
8 p.p. in the first case and to 3.8 p.p. in the second case)
Characteristics of the households in the extreme groups and comparison between their
distributions in the five groups by expenditure and by income
14 UNECE CPI Expert Group meeting 2023, June 7-9
• In 2021 in the first fifth of households’ expenditure group (but not of income):
✓ about 72% of the households range from 2 to 4 members
✓ in 28.8% the breadth of households is equal to 2, are mainly couples without children in
which the reference person is elderly (14.6%) and retired from work (95.0%)
✓ Households of 3 or 4 members are mainly couples with two children
• In 2021 in the first fifth of households’ income group (but not of expenditure):
✓ there are mainly one-component households (37.1%), followed by those households of a
size equal to 2 (22.9%)
✓ A fifth of the households (54.7% of all single-component) are elderly alone, retired from
work (49.8%) or inactive but in other condition (different from retired) (46.3%)
✓ With respect to what is observed in households belonging to the first fifth of expenditure but
not income, in the first fifth of households’ income group, the over-represented are mainly
single people (1.8 times) and single-parent families (1.3 times)
Characteristics of the households in the extreme groups and comparison between their
distributions in the five groups by expenditure and by income
15 UNECE CPI Expert Group meeting 2023, June 7-9
Table 1. Cross distribution of households by expenditure/income fifths. Absolute value and percentage points. 2021
Income fifths
Exp fifths 1 2 3 4 5 Total
1
2,692,049 1,375,267 699,709 297,137 137,507 5,201,670
10.35 5.29 2.69 1.14 0.53 20
51.75 26.44 13.45 5.71 2.64
51.75 26.44 13.45 5.71 2.64
2
1,219,493 1,518,011 1,261,158 785,691 415,794 5,200,147
4.69 5.84 4.85 3.02 1.6 19.99
23.45 29.19 24.25 15.11 8
23.44 29.19 24.24 15.1 7.99
3
710,222 1,149,206 1,363,005 1,182,498 798,416 5,203,348
2.73 4.42 5.24 4.55 3.07 20.01
13.65 22.09 26.19 22.73 15.34
13.65 22.1 26.2 22.73 15.35
4
417,855 796,851 1,127,863 1,523,926 1,334,976 5,201,472
1.61 3.06 4.34 5.86 5.13 20
8.03 15.32 21.68 29.3 25.67
8.03 15.32 21.68 29.3 25.67
5
162,527 361,463 750,225 1,412,514 2,514,584 5,201,313
0.62 1.39 2.88 5.43 9.67 20
3.12 6.95 14.42 27.16 48.35
3.12 6.95 14.42 27.15 48.35
Total
5,202,147 5,200,798 5,201,960 5,201,766 5,201,277 26,010,000
20 20 20 20 20 100
F
ifth
s o
f in
co
m
e b
y fifth
s o
f exp
en
d
itu
re
Fifths of expenditure by fifths of income
Characteristics of the households in the extreme groups and comparison between their
distributions in the five groups by expenditure and by income
16 UNECE CPI Expert Group meeting 2023, June 7-9
• In 2021:
✓ 21.8% of the first fifth of households by expenditure is allocated in the last three fifths of
households by income
✓ 24.8% of the first fifth of households by income is allocated in the last three fifths of
households by expenditure
✓ 24.5% of the fifth fifth of households by expenditure is allocated in the first three fifths of
households by income
✓ 26.0% of the fifth fifth of households by income is allocated in the first three fifths of
households by expenditure
Characteristics of the households in the extreme groups and comparison between their
distributions in the five groups by expenditure and by income
17 UNECE CPI Expert Group meeting 2023, June 7-9
• The heterogeneity of the allocation between the two groups of households at the basis of the
differences in the structure of weights
• In the first fifth of households by income there are households belonging to groups of
households by expenditure from the second (23.44%) to the fifth (3.12%)
• In the fifth group of households by income there are households belonging to groups of
households by expenditure from the second (26.44%) to the fifth (2.64%) group
• It means that the breakdown of expenditures in the first group of households by income is
different from that of the first group of households by expenditure, bringing behavior of
consumption typical of households that spend wider amount and reducing the relative weight of
food and energy products and vice versa for the fifth group of households by income
• This brings closer in 2022 the lines of inflation that affect the two extreme groups by income
(specifically lowering that related to the poorest) with respect those that affect the two extreme
groups by expenditure
Focusing only on the weights to measure the actual impact of inflation on the poorest?
18 UNECE CPI Expert Group meeting 2023, June 7-9
• Till now, in Italy, the analysis to estimate the differentiated impacts of inflation on groups of
households broken down by their economic condition has focused on the structure of weights
• In 2022 the government supports to poor households (detected in the basis of their income or
other indicators of their economic conditions) related to energy products (in particular, electricity
and gas) have been wide and in the form of reduction of prices
• This was considered in the compilation of the energy product consumer price indices that are
the results of weighted mean (with weights given by the number of households that have
benefited of the government support) of different inflation profile
• The aggregate consumer price indices of energy products are considered as such to estimate
the impact of overall inflation on the poorest and on the richest groups of households
• Should we start considering different profiles of inflation in addition to different structure of
weights?
• Moreover, how should we consider the impact of government support to households on the
weights of different groups, given the traditional temporal lag in the weights’ estimation?
Some concluding remarks and perspectives
19 UNECE CPI Expert Group meeting 2023, June 7-9
• The capacity of the extreme groups of households by expenditure to be a proxy of the poorest
and of the wealthiest households is mitigated by these results
• The outcomes of the estimation of the impact of inflation on different groups of households are
interesting and encourage further analysis. Specifically:
✓ Of the socio-demographic and socio-economic characteristics of the households’ groups by income
to be further analyzed
✓ Of the relationship between income and expenditure in the different groups
✓ Considering different profile of inflation for the 2 extreme groups to complement the approach based
exclusively on the weights
✓ Refining further the work on the weights given the effects on the structure of expenditure of the
government support to poorest households on energy products
• Starting the dissemination of an experimental statistics to open the debate (in 2024?)
• The new frame regulation on the social statistics that will harmonize HBS in the EU under a
common legal umbrella since 2026 will enhance the possibility to use HBS data to compare
across the European countries the impact of inflation on the different groups of households
Thank you
Ilaria Arigoni, Istat (Italy) ([email protected])
Alessandro Brunetti, Istat (Italy) ([email protected])
Valeria de Martino, Istat (Italy)([email protected])
Federico Polidoro, Istat (Italy) ([email protected])