Datos Económicos Italia 2003

8/19/2019 Datos Económicos Italia 2003

1/48

Project no: 028412

AIM-AP

Accurate Income Measurement for the Assessment of Public Policies

Specific Targeted Research or Innovation Project

Citizens and Governance in a Knowledge-based Society

Deliverable 1.3b Public health in Italy

Due date of deliverable: July 2007

Actual submission date: September 2007

Start date of project: 1 February 2006 Duration: 3 years

Lead partner: ECV

Revision [draft]


2/48

1

The distributional effects of public health

transfers in kind: the Italian case*

Conchita D'Ambrosio

Università di Milano-Bicocca and DIW, Berlin

and

Chiara Gigliarano

Università Politecnica delle Marche, Ancona

First version: August 2007

*We are indebted to Alberto Zanardi for many useful suggestions.


3/48

2

CONTENTS

Introduction………………………………………………………………………………….. 3

1. Results from other studies…...…………………………………………………………3

2. The Italian health care system……………………....…………………………………5

3. How to measure the health related transfers in kind.……………………………….13

4. Results from the empirical analysis …………………………………..……….…. …21

5. Concluding remarks…………………………………………………………….………31

Appendix A: Additional evidence from the empirical analysis.…………….………… 32

Appendix B: Results from alternative health expenditure data….……… ………….…42

References……………………………………………………….………………….….……46


4/48

3

1. Introduction

The analysis of the distributional effects of public policies is generally performed on

interventions aimed directly at the individual’s income, such as taxes and cash transfers. Equally

important is the evaluation of the impact of public expenses for services, i.e. transfers in kind, such

as education and health. The importance of this evaluation is particularly relevant for the policy

implications of monetary transfers vs. services supply.

Empirically, the evaluation of transfers in kind is more complex than the evaluation of monetary

transfers: the latter may be evaluated according to the costs incurred in producing them, an

amount that could differ from the profit enjoyed by the user because of possible inefficiencies in its

production.

Two alternatives to handle the problem are conceivable, as well discussed in Baldini et al.

(2006). On the one hand, we may look at the actual benefit occurred to the individual, i.e. whether

she has actually used health care services during the period of reference or not; in this case, the

variability of the benefits’ distribution would depend only on random effects (actual use approach).

On the other hand, we may consider the health related transfer in terms of its insurance function,

i.e. the benefit is a sort of insurance premium that individuals with particular socio-demographic

characteristics should pay in order to receive the services (insurance-based approach).

In an international perspective, several studies report on the distributional impact of health

public services in kind, applying both approaches. At the same time, much fewer are the works

specific on Italy; in Section 2 we briefly review its main results.

Aim of this report is to analyze the incidence and relevance of health related transfers in kind in

Italy and its effect on income distribution, applying the insurance-based approach.

We do not derive direct conclusions on the implication of such transfers in kind to inequality

and poverty, since we do not believe in an exercise of this type: transfers in kind to sick individuals

do not make them economically better off hence increasing their well-being.

2. Results from other studies

Quite rich is the literature on the distributional impact of transfers in kind, and health in

particular, in an international context; see, e.g., Smeeding et al. (1993), Steckmest, (1996) and

Marical et al. (2006), Garfinkel et al. (2006), Smeeding and Rainwater (2002).

Smeeding et al. (1993) study the combined effect on income distribution of non-cash incomegiven by health, education and housing, comparing seven western countries (Australia, Canada,

Netherlands, Sweden, U.K., U.S., West Germany), based on the LIS (Luxembourg Income Study)


5/48

4

data set, in the years between 1979 and 1983. They show that the most advantaged types of

households are the middle-aged families with children and the very elderly, while the most

disadvantaged socio-economic groups, in many of the analyzed countries, are the young singles

and younger families without children; the combination of the three non-cash transfers considered

have a levelling effect on income distribution.

Steckmest (1996) compares the distributional effects of health and education related non-cash

benefits in four countries (Norway, U.K., U.S. and Sweden). Adding education and health transfers

into the disposable income reduces inequality; higher decline in the degree of inequality are

monitored for Sweden and US, while the distributional effects in Norway and UK are limited.

Marical et al. (2006) enrich the analysis by providing estimates of the distributive implications of

public services (health, education and housing) in all the OECD countries. It constitutes one of the

few international studies where Italy is also included. Regarding health, in particular, they apply

both the insurance-based approach and the actual use approach, using the ECHP 2001 data set.

They show that, in general, public health care services are distributed uniformly across quintiles,

inducing a reduction in the inter-quintile ratio for all the countries. The distributive effect of health

care expenditures is, on average, significantly lower with the actual use approach than with the

insurance-based approach. According to the former, in particular, inequality slightly rises in Italy

and in Denmark.

Only a few studies focus on the distributional impact of health related public transfers in Italy.

The most recent studies, to the best of our knowledge, are Citoni (2001), Turati and Sonedda

(2005), Pacifico (2006) and Baldini et al. (2006).

Citoni (2001) applies the actual use method to the Bank of Italy Survey of Household Income

and Wealth (henceforth, SHIW) 1993 data set, which provides richer information about the usage

of healthcare services than the other Bank of Italy’s data sets.

Citoni (2001) concludes that the degree of inequality in the income distribution reduces after

including healthcare transfers; the same holds also if education transfers are added to the

individual income. He underlines that different results are obtained whether the benefits areequivalized or not (in the last case the effect is much smoother or sometimes even reversed) and

the distributional patterns differ according to the type of health care expenditure considered; e.g.

public hospital care is heavily concentrated on the lowest part of the income distribution while the

opposite happens for the specialists outpatient care. The overall distributional effect is a reduction

in the Gini index, with a redistribution of income in favour of the poor and the middle class.

Turati and Sonedda (2005), on the other hand, measure the health transfers in kind through a

mixture of the two approaches cited above. They apply the insurance-based approach to the non-

employed and the actual use approach for the employed individuals (taking into account the

information on the number of days of non-working due to illness); the data set employed is the


6/48

5

SHIW 2000. They conclude that inequality slightly increases after adding the health related

transfers. Moreover, they compare cash public transfers (including pensions) and transfers in kind,

concluding that the latter have a weaker redistributive effect than the former.

At the light of such controversial results, Baldini et al. (2006) and Pacifico (2006) try to better

clarify the distributional impact of health related transfers in kind in Italy, applying both the

insurance-based and the actual use method.

For the actual use approach, they apply a matching of two different datasets, the Bank of Italy

SHIW 2002 data set, that includes information on individual income, and the ISTAT survey on

living conditions and usage of health services (“Indagine multiscopo Istat sullo stato di salute e

ricorso a servizi sanitari”), related to the years 1999-2000. By matching the two datasets, based on

the propensity score approach, they assign to each individual in the Bank of Italy data a set of

variables on health conditions and consumption of health care services.

The total health expenditure is partitioned according to the kind of healthcare and assigned to

the individuals according to the consumption they have declared.

Their result affirms that inequality reduces more with the insurance-based approach than with

the actual use approach. They decompose the Reynolds-Smolensky measure of redistribution

(which is equal to the absolute difference of the Gini index before and after the introduction of the

transfers in kind) of the transfer in kind into the three components of incidence, progressivity and

reranking. Comparing transfers in kind with cash transfers (“trasferimenti assistenziali”), through

the micro-simulation model Mapp02, they conclude that the transfers in kind have a higher

redistributive effect than the cash transfers do. The Gini index reduces of about 4.1 percentage

points with the usage based approach and about 5.5 percentage points with the insurance-based

approach; the incidence of the cash transfers is higher than the one of the transfers in kind.

Beyond the difference in the levels, both the methods underline that the health related transfers in

kind have a redistributive effect in reducing the inequality in the income distribution. They conclude

that the tax system has a very low redistributive impact and a most important role is played by the

system of transfers (cash without pensions and in kind), which are much more able to act on the

income distribution. For the insurance-based method, they employ information on the total healthexpenditure for each region, related to the year 2002; this global amount is divided by the entire

population, so to obtain the per capita regional expenses. In a second stage, a class of age and

gender-related weights is employed, in order to take into account the differences in health needs

due to age and gender.

2. The Italian Health care system

In order to better understand the results from the empirical analysis, we summarize first the

main features of the Italian health care system.


7/48


8/48

7

• the provision of public contribution for public hospitals.

The 20 regions, on the other hand, have the responsibility for the organization, funding and

administration of publicly financed healthcare.

Finance

The two main sources of funding, for the health care system, are constituted by the households

and the enterprises; the former pay for healthcare services and benefit from most of them, while

the latter benefit only from few services, such as public hygiene and job safety, and pay health

contributions (since 1998 converted into IRAP-Imposta Regionale sulle Attività Produttive).

The Italian health care system is a mixture of private, insurance-based and public health

system, since it is financed by prices, premium and taxes. Prices are paid for those health care

services that are not provided by SSN, such as particular categories of drugs and dental cure;

tickets, on the other hand, are partial prices that are paid partly by the patient and partly by the

SSN. Premiums are paid to private insurance companies that provide complementary or additional

insurance beyond SSN, while taxes are the most common way to fund the public health system.

More in details, the Italian national Health System is financed by:

• General taxation: Regional add-on to the national personal income tax (Addizionale

regionale IRPEF1), central grant (“Fondo per il fabbisogno sanitario ex D.Lvo 56/2000 “) financed

by the value-added tax (VAT2) and by other indirect tax (e.g. fuel) revenues;

• Compulsory social insurance, replaced in 1998 by a business tax, IRAP; 3

• Transfers from the State (ex FSN);

• Co-payments (or tickets) of patients for pharmaceutical and specialist ambulatory services.

After the devolution reforms of the years 1990 and 2000, the responsibility of financing the

healthcare system passed from the central government to the regions, implying that each region is

responsible for eventual deficit in its budget. The introduction of tickets and forms of co-

participation constitutes another way to reduce the regional costs for the health care system.

In order to increase their financial resources, the regions are allowed also to increase the rate

of the add-on to the national personal income tax or the rate of IRAP and to control the costs for

the staff, the pharmaceutical, specialist and hospital assistance.

1 Each region decides whether the rate for the regional add-on of the national personal income tax is a unique rate or is

different rate according to income brackets; the taxable income is the same as for the national personal income tax

IRPEF. Therefore the individuals exempt from the national tax are exempt also from the ADDIZIONAL REGIONALEIRPEF.2 In 2003, about 38.55% of the total national revenue from VAT was devoted to funding the public health system.

3 In 2003, the rate of IRAP was about 4.25% of the net value of the production; however, each region may decide the

entity of such tax rate.


9/48

8

Moreover, there exists a system of “perequation” among regions, so that the economically and

fiscally weaker regions receive a contribution from the economically stronger regions.

The new fiscal federalism, introduced by the law D.L.vo 56/2000, has been established,

together with the new responsibilities for the regions in the management of the health care system

and of eventual deficits, so to guarantee the implementation of the Title V of the Italian

Constitution, reformed by the law 3/2001.

An important step in the procedure of the establishment of the health care system finance is

constituted by the agreement between State and Regions to decide the aggregate funding and

how to allocate it among the regions; this often causes huge political friction. The formula for the

allocation of financial resources has changed several times; at the moment it is based on

population size, weighed by age-specific utilization rates for hospital care, drugs and residential

care, on standardized mortality rate and on an adjustment for interregional patient flows.

In particular, with the Agreement (“Accordo Stato-Regioni”) undersigned on August 8 th, 2001,

State and regions reached an agreement on the total amount of health care financial needs for the

years 2002 to 2004.

In 2003, which is the year of interest for our analysis, the total revenues of the SSN were

constituted for 39% by IRAP, for 3% by regional IRPEF, for 43% by VAT and other indirect taxes,

for 3% by co-payments of patients (tickets), for 7% by transfers from public and private sectors and

5% from integration from the State. (source: Ministero dell’Economia e delle Finanze, 2004); see

Figure 1.


10/48

9

Fig.1: Health care system finance, by resources. Year 2003.

39%

3%

43%

7%

3%5%

IRAP

ADDIZIONALE IRPEF

VAT+INDIRECT TAXES

TRANSFERS FROM PUBLIC AND PRIVATE SECTORS

OWN REVENUES

INTEGRATIONS FROM THE CENTRAL GOVERNMENT (only for Sicily and Sardinia)

Source: Ministero dell’Economia e delle Finanze (2004)

Providers

The SSN combines public financing with a mixture of public and private provision:

• Public structures, which are mainly focused on prevention, general public health and family

doctors;

• Private accredited structures related to and financed by the SSN, which are mainly focused

on diagnosis, hospital services and rehabilitation;

• Private structures not related to the SSN (financed by out-of-pocket and private insurance);

such kind of structure is not very popular in the Italian health care system.

Responsibility for delivery rests with the “Aziende Sanitarie Locali” (ASL).

The hospital structures are about 1400, of which 750 are public, 550 are accredited to SSN and

100 are private not accredited. On average, the public structures are bigger than the private ones.

In 2003, according to ISTAT (2006), the “hospital usage rate” was about 144.6 for 1000

inhabitants.

Among the OECD countries, in 2003 Italy has the highest percentage of old people (19%

versus 14% OECD mean), has the 2° position in number of doctors (41/10000 versus 29/10000

OECD mean) but the 23° position in number of nurses (54/10000).


11/48

10

Expenditure

According to the Italian General Accounting Office (Ministero dell’Economia e delle Finanze,

2004), in 2003 the total public health-care expenditure4 was equal to 83.2 billion Euro, i.e. a per

capita expense equal to 1407 Euro, with a deficit of 1.9 billion Euro. The incidence of the public

health care expenditure on the gross domestic product (GDP) was about 6.3%.

According to OECD (2005), 5 moreover, the total (public and private) health expenses in Italy in

2003 were 8.8% of the GDP.

In 2003, 75% of the total health-care expenditure is publicly financed (versus 72% OECD

mean) (4.0% less than in the year 1990), 21% is financed out-of-pocket and 0.9% is financed by

private insurance (source: OECD 2005).

Figure 2 shows that, in 2003, 10% of the public total health expenditure is for the hospital

sector, 34% for pharmaceutical care, 25% for services, 13% for drugs, 6% for general medicine

and 3% for outpatient specialist care, according to the Italian General Accounting Office.

Fig. 2: Health care system expenditure, by resources. Year 2003.

10%

34%

25%

3%

13%

6%

9%

HOSPITAL STAFF SERVICES SPECIALIST PHARMACEUTICAL GENERAL MEDICINE OTHER


4 Health expenditure is defined in Ministero dell’Economia e delle Finanze (2003, 2004) as the sum of all the costs for

the specific welfare services (see Figure 3) and net costs for the extraordinary management and for the private health

care provided in public health structures (named “intramoenia”).

5 The definition of health expenses provided by OECD (2005) is the following: “Total expenditure on health measures the

final consumption of health care goods and services (i.e. current health expenditure) plus capital investment in health

care infrastructure. This includes spending by both public and private sources (including households) on medical

services and goods, public health and prevention programmes and administration. Excluded are health-related

expenditure such as training, research and environmental health. Total expenditure on health does not include

compensation for loss in income due to health problems (sick pay and disability allowances).


12/48

11

The private health care expenditure is constituted mainly by co-payment and tickets for SSN

structures and private expenditure for services from structures not related to SSN. 6

It should be noted that the Italian fiscal system includes a tax-allowance on the personal

income tax for out-of-pocket expenditure, equal to 19% of the amount exceeding 129.11 Euro.

As already pointed out, the recent reforms of the Italian health care system allowed regions to

introduce co-payments for ambulatory care and for drugs (the so-called “ticket”). Each region was

free to choose the amount and level of such tickets. There exists, however, forms of full or partial

tickets exemption, for particular categories of patients, such as chronic ills, job-invalids, blind, deaf

and dumb, war-pensioners, prisoners, elderly with low income, unemployed, pregnant women.

Figure 3 synthesizes the structure of the Italian National Health System, in terms of administration

and management, of finance and of service provision.

Fig. 3. The Italian National Health Service 2002

6 As already pointed out, in 2003, 21% of the total healthcare expenditure was out-of-pocket expenditure, i.e. paid

directly by the patients, while 0.9% of the total healthcare expenditure was paid by private insurance.

PARLIAMENT

GOVERNMENT

MINISTRY OF HEALTH

REGIONAL GOVERNMENTS

REGIONAL HEALTH DEPARTMENTS

PUBLICHOSPITALENTERPRISES

ASLs

PHARMACIES GENERALPRACTICTIONERS

PRIVATE ACCREDITEDHOSPITALS ANDSPECIALISTS

RESEARCHHOSPITALS

CITIZENS/PATIENTS

NATIONALBUDGET

REGIONALBUDGET

Co-paymentsfor ambulatorycare

List prices Capitation

Regional taxes

National taxes

Tariffs

Co-payment for ambulatory care

Tariffs

Administ rat ion and p lanning flowsFinancial flowsService flows

Source: France, Taroni, Donatin i (2005)


13/48

12

Interregional mobility

The Italian health system is characterized by the free mobility of the patients throughout the

country, so that patients can choose the structure where to be cured. This implies the existence of

interregional mobility of patients. Figure 4 shows the net interregional mobility related to the year

2003; the net interregional mobility of a region A is measured as the difference between the total

revenue that region A receives from all the regions whose patients are cured in region A, minus the

money that the region A gives to those regions where its patients are cured. Figure 4 shows that

the regions with highest positive interregional mobility are Lombardia and Emilia Romagna, while

the regions with highest negative interregional mobility are Campania, Sicilia and Calabria.

A recent study on the interregional mobility for admissions at hospitals, by Assr (2003) shows

that Lombardia is the region with the highest net mobility, followed by Emilia Romagna, Lazio,

Veneto and Toscana; the regions with the lowest net mobility (negative) are Campania, Calabria,

Sicilia, Puglia and Basilicata. Only 5.5% of the health assistance in mobility refers to highly

specialized healthcare, while 58.8% of the hospital admissions are for ordinary healthcare. In this

report it is highlighted that most of the hospital admission in mobility, in the year 2003, was not due

to the inadequate healthcare service quality of the structure in the own region.

Fig. 4: Net interregional mobility (in Euro). Year 2003.

Net interregional mobility (in Euro)

-300

-200

-100

0

100

200

300

400

500

C a m p a n i a

S i c i l i a

C a l a b

r i a

P u g l i a

B a s i l i c a

t a

S a r d e g n a

M a r c h e

P i e m

o n t e

V a l l e

d ' A o s t a

L i g u r i a

T r e n t i n

o

M o l i s e

A b r u z z

o

F r i u l i - V

e n e z i a

U m b r i a

L a z i o

T o s c a n a

V e n e t o

E m i l i a

- R o m a g n a

L o m b a r d i a



14/48

13

4. How to measure the health related transfers in kind

The aim of the paper is to translate into a monetary form the benefits received by the provision

of health related public goods and study the implication on the income distribution. The mostcommon way to measure health benefits in kind is to use the costs due to produce the health

transfers as a proxy of such benefits. We hypothesize that the marginal costs to produce such

goods by the state are equal to the marginal benefits received by the agent that asks for such

goods. In case of inefficient operators in the market of such goods, the marginal cost is not

anymore an adequate proxy for the marginal benefit.

In the literature, we find different alternative methods to measure the benefits from transfers in

kind. As already discussed in the Introduction, the main methods are two: the actual use-based (or

usage-based) approach and the insurance-based approach.

The usage based method assigns to each individual a transfer (based on the costs) according

to her effective use of the healthcare service; as a result, individuals are more differentiated from

each other. There exist, however, difficulties in finding data set that include both information on

income and on the usage of the healthcare services and the general health condition of the person.

The second problem is due to the difficulty in finding unitary costs for the different kind of health

services; see Pacifico (2006) for a wider discussion.

The insurance-based approach, on the other, is easier to implement, since it does not require

information on the individual health condition nor on the usage of healthcare services; the method

assigns health related transfers to the individuals proportionally to some socio-economic

characteristics, such as gender, age and geographical area. The individual transfer may be

considered as the premium an individual should have paid in order to be insured against the risk of

illness. The amount of such premium should be determined according to such individual

characteristics. The transfer is defined as proportional to the probability of loosing good health

condition, where the probability is a function of observable variables such as age and gender.

The main disadvantage of such approach is that it does not take into account the individual’s

differences in health condition and needs nor in income, since we hypothesize that the health

conditions depend not only on demographic characteristics, such as age and gender, but also on

socio-economic characteristics. The main advantage of such approach, on the other hand, is its

easy implementation, since it needs information on the aggregate expenditure on health for the

entire country and on the population distribution by age and gender.

In this report we explore two different analyses, following in both cases the insurance-based

approach.

The first analysis is conducted for the comparative report for the AIM-AP project, which is

based on the definition of health-care benefits as the health-care amounts per capita per age group

provided by OECD; the results from such analysis are discussed in Appendix B.


15/48

14

In this section we focus on the second analysis, which defines the health-care benefits as

health-care amounts per capita per age group calculated for different regions and different kinds of

health-care services.

EU-SILC 2004

The dataset employed for the analysis is the “IT-SILC XUDB 2004-versione Febbraio 2006”,

which contains the Italian data of the European Survey of Income and Living Conditions (EU-

SILC), based on the European Union Regulation (no. 1177/2003) which defines the EU-SILC

project. In particular, it contains extra variables beyond the ones common to all the European

countries that are part of the project.

This survey replaces the former European Community Household Panel (ECHP) with the main

scope to provide, through harmonized definitions and methods, comparable data, cross-sectional

and panel, in order to analyse jointly income and welfare distribution among the households and to

monitor the effect of European and national socio-economical policies.

The sample design of the Italian SILC data set consists of a two-stage sampling, according to

which, for each region, municipalities are first divided into the so-called auto-representative

municipalities (with larger population size) and the non auto-representative municipalities (smaller

in size). All the auto-representative municipalities belong to the sample and within them

households are systematically drawn from the register office records. For the non auto-

representative municipalities, a sample design with two stages is conceived, according to which a

sample of municipalities is chosen (first stage) and households are selected randomly within each

municipality, from the register office records (second stage). The probability of selecting a

municipalities is proportional to its population size, while households are drawn with the same

probability.

The Italian EU-SILC sample contains 24,204 households and 61,429 individuals (52,509 aged

15 and more years old at the end of the referring income period of time) living in 731 municipalities.

For the analysis of this paper we take into account only the individuals with strictly positive income;therefore, we reduce to 24,048 households and 61,107 individuals.

In EU-SILC 2004 information on income refer to the year 2003, while information on the living

conditions refer to the moment of the interview, i.e. the year 2004. The income variable considered

in EU-SILC 2004 for the baseline income is the total disposable income “HY020”, given by the

sum, for all household members, of gross personal income components, gross cash benefits (self-

employment, sickness, survivor, unemployment, disability), income from rental of property, family

allowances, housing allowance, interests and profits from capital investments, minus taxes on

income, wealth, social insurance contributions. Such baseline income, therefore, includes already

all the direct public transfers related to health condition, if any.


16/48

15

Units of analysis are individuals living in private households with positive disposable income.

In order to take into account the differences in needs among households with different sizes,

we apply the modified OECD scale for both disposable income and health transfer in kind. 7

Since our analysis is based on the individual’s age, we should note that in EU-SILC 2004 data

set the individuals aged more than 80 years old were assigned the age of 80; therefore, we are not

able to distinguish the differences in age for the over80s.

Health care expenditure per capita per age group

The health benefits in kind are approximated by the costs that the state has to pay to produce

them. In order to obtain the per capita health expenses weighted by age and gender, we collect

first the following information:

• Data on the expenditure of National Health System (SSN), which are taken from “Relazione

Generale sulla Situazione Economica del Paese” by Ministero dell’Economia e delle

Finanze (2003, 2004).8 Such costs are partitioned into the following categories of

expenses: for staff, for services and commodities, for extraordinary costs, for general

medicine, for drugs, for specialists, for rehabilitation, for prosthesis, for hospital and for

other kind of health assistance.

• The demographic population, i.e. the number of residents in each region at the beginning of

the year 2003 (data available on www.demo.istat.it ). Figures 5, 6 and 7 show the

distribution of the population by age, respectively, for the overall population and for each

region. We observe that the distribution of the overall population is skewed to the left, with

much higher percentage of elderly than that of young people. Figure 6 shows, on the other

hand, that the regions in South Italy and Trentino Alto Adige are characterized by higher

percentages of young inhabitants (less than 44 years old) than the rest of Italy. In Figure 7,

instead, we see that higher proportion of middle aged individuals live in the bigger cities,while a higher percentage of elderly live in small villages.

7 Different equivalence scales have been applied, in literature, for the health related transfers; see, e.g. Citoni et al.

(2006).8 In particolar, we employ the Appendix SA-1 “Costi del SSN per funzioni di spesa. Anni 2001-2004”.

http://www.demo.istat.it/http://www.demo.istat.it/


17/48

16

Fig. 5: Distribution (in %) of the overall Italian population in 2003, by age.

Distribution (in %) of

population by age

1.582.65

9.28

39.93

26.63

10.789.16

0

5

10

15

20

25

30

35

40

45

75 years

Source: ISTAT, in www. demo.istat.it

Fig. 6: Distribution (in %) of the population in each region, by age. Year 2003.

0%

20%

40%

60%

80%

100%

P i e m

o n t e

V a l l e

d ' A o s t a

L o m b a r d i a T r

e n t i n o

V e n e t o

F r i u l i - V

e n e z i a

L i g u r i a

E m i l i a

- R o m a g n a

T o s c a n a

U m b r i a

M a r c h e L a z i o

A b r u z z o M

o l i s e

C a m p a n i a P u

g l i a

B a s i l i c a t a

C a l a b r i a S i

c i l i a

S a r d e g n a

I T A L I A

OVER75 YEARS

65-74 YEARS

45-64 YEARS

15-44 YEARS

5-14 YEARS

1-4 YEARS


18/48

17

Fig. 7: Distribution (in %) of population, by age and size of municipality. Year 2003-2004.

14.26 14.19 14.27

67.07 66.78 64.47

18.67 19.03 21.26

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

>50000 inhabitants 2000-50000 inhabitants 75 4.290 3.111 0.064

Source: Ministero della salute (2006) and AA.VV. (1999).


19/48

18

After collecting all such information, we proceed in the following way.

First, we group the different kinds of healthcare expenditure into seven categories (see Table 2):

I. Hospital expenses: the total hospital-related expenses plus 85% of the 75% of expenses for

commodities and services plus 85% of the 75% of expenses for staff and plus administrative costs

(proportional to the share);9

II. Expenses for specialists: the entire specialists-related expenditure plus 15% of the 75% of

commodities and services expenses plus 15% of the 75% of expenses for staff plus administrative

costs (proportional to the share);

III. Pharmaceutical expenses: the total pharmaceutical costs plus administrative costs

(proportional to the share);

IV. General medicine and other assistance: general medicine-related costs plus expenses for

rehabilitation plus expenses for prosthesis plus other kinds of assistance plus administrative costs

(proportional to the share);

V. Expenses for prevention: 40% of the 25% of the expenses for commodities and services

plus 30% of the 25% of expenses for staff plus administrative costs (proportional to the share);

VI. Expenses for national health assistance: 40% of the 25% of expenses for commodities

and services plus 30% of the 25% of expenses for staff plus administrative costs (pro-rata);

VII. Extraordinary expenses: extraordinary expenses plus administrative costs (proportional to

the share).

Table 2: Re-partition of the health care expenditure and corresponding weights.

EXPENSESFOR: HOSPITAL SPECIALISTS DRUGS

GEN ERALMEDICINE PREVENTION

HEALTH ASSISTANCE

EXTRAORDINARY

- total hospital-related

expenses;

- 85% of the75% of

expenses forcommoditiesand services;

- 85% of the75% ofexpenses for

staff;

- administrativecosts

(proportional tothe share).

- specialists-related

expenditure;

- 15% of the75% of

commoditiesand servicesexpenses;

- 15% of the75% of

expenses forstaff;



- pharmaceuticalcosts;



- generalmedicine-

related costs; - expenses for

rehabilitation;

- expenses forprosthesis;

- other kinds ofassistance;


(proportionalto the share).

- 40% of the25% of the


- 30% of the25% of

expenses for

staff;



- 40% of the25% of


- 30% of the25%of

expenses forstaff;

- administrativecosts (pro-

rata).

- extraordinaryexpenses;



WEIGHTSUSED:

HOSPITALWEIGHTS

SPECIALISTSWEIGHTS

DRUGSWEIGHTS

SPECIALISTSWEIGHTS

PER CAPITASPECIALISTS

WEIGHTSPER CAPITA

9 The total expenses for commodities and services and for staff are partitioned into the main types of health services,

according to estimations provided in Ministero dell’Economia e delle Finanze (2003). It has been estimated that 75% of

the expenses for commodities and services and for staff are for the public expenses, 85% of which are for hospital and15% of which are for healthcare of specialists. The remaining 25% of expenses for staff is partitioned into: 40% foradministrative costs, 30% for costs for prevention and 30% for health assistance. The remaining 25% of expenses forcommodities and services is partitioned into: 40% for prevention, 40% for health assistance and 20% for administrativecosts.


20/48

19

Figure 8 shows the total health care expenditure in the year 2003 for each region, according to

the above partition. We see that the main source of expenses is due to the hospital care, followed

by the pharmaceutical and by the general medicine care.

Fig. 8: Health care expenditure of each region, by kind of expenditure. Year 2003.

KIND OF HEALTHCARE EXPENSES

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

V a l l e

d ' A o s t a

M o l i s e

B a s i l i c a

t a

T r e n t i n

o

U m b r i a

F r i u l i V

e n e z i a G i u l i a

A b r u z z

o

M a r c h e

S a r d e g n a

L i g u r i a

C a l a b

r i a

T o s c a n a

P u g l i a

E m i l i a

R o m a g n a

P i e m

o n t e

V e n e t o

S i c i l i a

L a z i o

C a m p a n i a

L o m b a r d i a

I T A L I A

HOSPITAL SPECIALIST PHARMACEUTICAL GENERAL MEDICINE PREVENTION GENERAL HEALTH CARE EXTRAORDINARY


The second step is to take into account the expenses for the interregional mobility of the patients,

by subtracting from the health care expenditure of each region the corresponding total mobility;10 in

particular, we partition the total amount of net mobility into the seven categories of health

expenditure, according to the corresponding share of total expenditure, in the following way:

ij

ij i

i

E M M

E = ,

where Mij is the net mobility of region i corresponding to the kind j of health expenditure, Mi is the

total net mobility of region i, Eij is the health expenditure of region i of the type j and Ei is the total

expenditure for health care of region i.

Figure 9 shows how the total expenditure of each region changes whether we take into account

the interregional mobility or not. In general, very small changes occur and, in particular, Lombardia,

Emilia Romagna and Lazio, which are characterized by high mobility, show the highest decrease in

expenditure when the mobility is taken into account; the highest increases occur, instead, for

Campania, Calabria and Sicilia.

10 Data on the net interregional mobility are taken from Table SA3 in “Relazione Generale sulla situazione economica

del Paese” by Ministero dell’Economia e delle Finanze (2004).


21/48

20

Fig. 9: Total health care expenditure of each region, with and without mobility effect.

Year 2003.

0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

P i e m

o n t e

V a l l e

d ' A

o s t a

L o m b a

r d i a

T r e n

t i n o

V e n e

t o

F r i u l

i - V e n

e z i a

L i g u r

i a

E m i l i a

- R o m

a g n a

T o s c

a n a

U m b r

i a

M a r

c h e

L a z i o

A b r u

z z o

M o l i

s e

C a m

p a n i a

P u g l i

a

B a s i l

i c a t a

C a l a b

r i a S i

c i l i a

S a r d

e g n a

total expenses without mobility total expenses with mobility

Source: Ministero dell’Economia e delle Finanze (2004).

The third step, in order to obtain the per capita health expenses per age, is to assign to each

individual the regional health expenses net from mobility.

For each region, the health expenses of kind V. and VII. (see the above list) are allocated

among individuals per capita. For health costs of type I., we employ the weights for hospital

expenses, indicated in Table 1; for health costs II., IV. and VI., we use the weights for expenses for

specialists; for costs III., we use the weights for pharmaceutical expenses. Such procedure is

shown also in the last row of Table 2.

For each region and for each of the seven types of health care expenditure discussed above,

we reproduce the following Table 3, which allow us to obtain, in its last column, what we want: the

expenses per capita differentiated for age classes.

Table 3: Procedure for obtaining the expenditure per capita by age (“e”=per capita expenditure)

Agerealpopulation weight

EquivalizedPopulation Ratio %

re-partitionedpopulation

expenditure percapita,by age

75 years p8 c8 p8*c8 r8=p8*c8*100/S2 rp8=r8*S1/100 e*rp8/p8Total S1=p1+…+p8 S2=p1*c1+…+p8*c8 S3=rp1+…+rp8


22/48

21

The last two columns of Table 3 show that the population in each age group is recalculated, by

using the weights of the third column (which are the coefficients of Table 1); then the per capita

expenditures weighted by age are obtained multiplying the per capita expenditure (indicated with

“e” ) by the ratio of such re-partitioned population over the real population, for each age group.

The last step consists of summing, in correspondence of each age group and each region, the

seven different kinds of per capita health expenses, weighted by age and gender.

The per capita expenses that we obtain for each age group, for each region and for each type

of expenses represent the individual health transfers in kind and are shown in Figure 10.

Figure 10 shows that the regions assigning higher per capita weighted amount of health

transfers are Trentino and Valle d’Aosta, while Campania has the lowest per capita weighted

transfer.

Fig. 10: Health care amounts per capita per age groups (Euro). Year 2003.

HEALTHCARE AMOUNTS PER-CAPITA PER AGE GROUP (EURO)

500

1000

1500

2000

2500

3000

3500

4000

4500

75

ABRUZZO

VAL D'AOSTA

BASILICATA

CALABRIA

CAMPANIA

EMILIA

FRIULILAZIO

LIGURIA

LOMBARDIA

MARCHE

MOLISE

PIEMONTE

PUGLIA

SARDEGNA

SICILIA

TOSCANA

UMBRIA

VENETO

TRENTINO

Source: Own elaboration of data in Ministero dell’Economia e delle Finanze (2004).

5. Results from the empirical analysis

In the following analysis, we focus only on the change in the distribution of income after health

transfers in kind. We underline in particular which are the groups that are the most affected by

such public intervention, comparing their final income distributions.


23/48

22

We do not discuss the implications of health transfers in kind on inequality and poverty, since,

according to our opinion, the increase in mean income attributed to some groups of individuals due

of their health condition cannot be considered as an improvement in their socio-economic state.

IR relevance on the overall population

When we add the healthcare transfers to the cash-income of each individual, the overall

disposable income obviously increases; the degree of such increment, i.e. the health related

transfers relevance, is described in Table 4.

Table 4 compares the income share, by quintile, of the baseline distribution (based on cash

income) with the distribution of income including the healthcare transfers in kind. The share of

income increases for the three lowest quintiles and reduces for the two highest quintiles.

Therefore, health transfers in kind are able to reduce the overall dispersion in the income

distribution.

Table 4 shows, moreover, the percentage increase in income due to health transfers, by

quintile; we observe that the relevance of such transfers decreases drastically with income. In

particular, looking at the subgroups defined by age, we note that when health transfers are added

only to the group of individuals aged between 15 and 64, which are about the 67% of the

population, the entire society benefits much more than in case of health transfers added to the

other two groups. Figure 11 provides a clear insight into such trend.

In absolute terms, same results hold when looking at the absolute equivalized health transfer

mean, shown in the last columns of Table 4. When health benefits are transferred either to the

youngest group or to the sole elderly, the absolute mean transfer decreases with income, because

young and old people are mainly located in the lowest quintiles; for the middle-aged individuals, the

average amount of the transfers increases with income, revealing that such group is mainly located

in the highest quintiles. When the health transfers in kind are added to the entire population, such

amount slightly increases when moving from the lowest to the middle-income quintiles and then

reduces from the middle to the highest income quintile.

Tab. 4: Relevance of health related transfers in kind. Year 2003.

INCOMESHARE

% INCREASE IN DISPOSABLEINCOME DUE TO TRASFERS TO

PERSONS AGEDMEAN TRASFER PER CAPITA DUE TOTRASFERS TO PERSONS AGED

QUINTILE A B 0-14 15-64 65+ ALL 0-14 15-64 65+ ALL

1 7.4 8.8 5.7 20.3 10.3 36.3 317 1127 574 2019

2 12.7 13.7 3.1 11.5 8.3 22.9 297 1088 790 2175

3 17.1 17.6 1.9 9.1 5.8 16.9 245 1167 747 2159

4 22.6 22.3 1.1 7.7 3.7 12.6 191 1301 627 2118

5 40.1 37.6 0.6 4.5 1.7 6.9 190 1352 518 2060 ALL 100.0 100.0 1.7 8.1 4.4 14.1 248 1205 650 2101

Source:EU-SILC 2004


24/48

23

Fig. 11: Increase (%) in disposable income due to transfers to persons aged

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

1 2 3 4 5 ALL

0-14 15-64 65+ ALL

Source: EU-SILC 2004.

The following Figures 12 to 15 show the kernel density estimation for the income distribution

before and after health transfers in kind, both for the overall population and for subgroups defined

by age; we note that, in each graph, the income distribution is shifted to the right after the transfers,

with the most evident impact attributed to the above65 years old group.

Fig. 12: Kernel estimation of the income distribution before (red line) and after (blue line) health

transfers in kind, overall population.


25/48

24

Fig. 13: Kernel estimation of the income distribution before (red line) and after (blue lin

e) health transfers in kind, for group aged 25 years old and below.


transfers in kind, for group aged 26 to 64 years old.


26/48

25


transfers in kind, for group aged 65 years old and above.

Finally, Figure 16 shows the change in the relative income position (defined as the ratio of the

income mean of the subgroup over the overall income mean) before and after the introduction of

the health related transfer to the entire population; we observe clearly that the higher increase in

income position is in correspondence of the older individuals, while a slight loose in income

position happens for young and middle-aged individuals.


27/48

26

Fig. 16: Relative income position before and after health related transfers. Year 2003.

Change in relative income position due to health transfers by age group

0

20

40

60

80

100

120

75 years

baseline income income+health


IR relevance by groups

It is interesting now to decompose the percentage increase of income by subgroups,

partitioning the population according to different socio-economic characteristics; Table 5 reveals, in

particular, in which subgroups of the population the health transfers in kind have greater impact.

Splitting the population by type of household, we observe that the highest increase in income

mean is attributed to the old singles or couples, although they do not have the lowest income

mean. The lowest increase in income mean is attributed to the couples or singles that live alone

without children or old family member, who have the highest baseline in income mean.

Among the groups based on the socio-economic conditions of the householder, the

unemployed and the pensioner benefit the most from health care transfers, the former mainly

because their baseline income mean is very low and the latter because they are older than the

other kind of householders, in particular than the white collar or self employed.

According to the education level, the least educated householders show the greatest increase

in income; the main reason is due to the fact that the householder with high level of education are

younger on average that the householders with very low level of education and income mean.

When the population is split into age levels, we observe, in both data sets, that the elderly are

the most affected by the health transfers in kind, followed by the group of the below25 years old;

this is due to the shape of the health care amount per capita per age group shown in Figure 10.

The geographical partitions, finally, show that the South of Italy and the smallest villages

register the highest health transfers relevance; the reason of such result is mainly due to the lower

income mean of such beneficiaries groups.


28/48

27

Tab.5: Change in income mean by groups. Year 2003.

Income position

Characteristic of householdor household head

Pop.

share in%

Income

Mean(Baseline)

% increase

in meanequiv.Income

BaselineIncluding

health

Household type

Older single persons or couples 17.2 13916 24.2 93 102

Younger single persons or couples 14.8 17524 8.3 118 112

Couple with children up to 18 34.9 13845 12.4 93 91

Mono-parental household 2.7 11789 13.0 79 78

Other household types 30.4 15695 14.0 105 105

Socioeconomic group of HH head

Blue collar worker 20.9 12085 13.8 81 81

White collar worker 18.6 17597 9.7 118 113

Self-employed 17.3 17939 9.9 120 116

Unemployed 3.2 8003 21.1 54 57

Pensioner 29.1 15116 18.5 101 105

Other 11.0 12415 19.1 83 87

Educational level of HH head

Tertiary education 9.1 23796 8.1 160 151

Upper secondary education 28.8 16547 11.2 111 108

Lower secondary education 30.4 13315 14.1 89 89

Primary education or less 31.7 12398 21.0 83 88

Age of HH member

Below 25 24.8 13312 13.3 89 89

25-64 55.9 15845 11.5 106 104

Over 64 19.3 14251 23.6 96 104

Area

North 45.3 17090 12.3 115 113

Center 19.3 16000 13.6 107 107

South+islands 35.4 11527 17.9 77 80

City size

>50000 inhabitants 41.9 15912 13.0 107 106

2000-50000 inhabitants 39.5 14709 14.3 99 99


29/48

28

Disaggregating by regions

Figure 17 orders the regions according to their income mean before the health transfers; the

poorest regions are located in the South of Italy (Calabria, Sicilia and Basilicata), while the richest

are in the North (Lombardia, Valle d’Aosta and Emilia Romagna). After health transfers in kind,

almost every region maintains the same rank, with very slight re-ranking such between Valle

d’Aosta and Emilia Romagna.

Fig. 17: Mean income before and after health transfers (weighted by age).

0

5000

10000

15000

20000

25000

C a l a b

r i a

S i c i l i a

B a s i l i c a

t a

P u g l i a

C a m p a n i a

M o l i s e

S a r d e g n a

A b r u z z

o

I T A L I A

M a r c h e

U m b r i a

L i g u r i a

L a z i o

V e n e t o

T o s c a n a

P i e m

o n t e

F r i u l i - V

e n e z i a

T r e n t i n

o

L o m b a r d i a

V a l l e

d ' A o s t a

E m i l i a

- R o m a g n a

income mean before health transfers income mean after health transfers


Figure 18, on the other hand, shows the percentage increase in disposable income by quintile,

for each region. In correspondence of the lowest quintiles, there are differences in the proportions

of income’s increase among regions; the poorest inhabitants of Trentino, Valle d’Aosta and Molise,in particular, receive higher relative increase in their income mean than the poorest inhabitants of

the other regions do. More similar level, among regions, of the percentage increase in income are

registered in correspondence of the highest income quintiles. For all the regions, the percentage

increase in income mean decreases with income. Table 6 reports the same values illustrated in

Figure 18.


30/48

29

Fig. 18: Increase (in %) in disposable income, by quint ile. Year 2003.

0

5

10

15

20

25

30

35

40

45

50

1 2 3 4 5

QUINTILE

% I

N C R E A S E I N D I S P O S A B L E I N C O M E

Piemonte

Valle d'Aosta

Lombardia

Trentino

Veneto

Friuli-Venezia

Liguria

Emilia-Romagna

Toscana

Umbria

Marche

Lazio

Abruzzo

Molise

Campania

Puglia

Basilicata

Calabria

Sicilia

Sardegna

ITALIA



31/48


32/48

31

6. Concluding remarks

In this report, we convert into monetary amounts the benefits in kind related to the public

supply of healthcare, taking into account both age and gender of the individuals.

In particular, very useful are the weights that are usually employed by the regions and the State

in order to allocate among the regions the national funds for healthcare; such coefficients weigh

the population differently according to age and gender. We are therefore able to obtain per capita

healthcare costs, weighted by age and gender, and add them to the individual cash income, in

order to obtain an accurate value of personal income that takes into account also the benefits from

the public service of health.

Health benefits are easy to predict from a life cycle analysis; health care benefits are, in fact,disproportionately high for the elderly. Since single elderly on average have low income, these

benefits represent a large proportion of the latter.

The most affected groups, indeed, that emerge from our analysis are the over65, the old

singles and couples and the households with unemployed, pensioner or less educated

householder.

Disaggregating by regions, after the health related transfers in kind, very slight re-ranking of the

regions occurs, in term of income position.

Implications on the levels of inequality and poverty are discussed in the Appendix A; in Appendix B a robustness analysis is conducted, which is alternative to the one conducted in the

main part of the report and by employs the per capita health care expenditure differentiated by age

and gender that are provided by Marical et al. (2006).


33/48

32

Appendix A: Addi tional evidence from empirical analysis

In this appendix, we discuss the implications of the health related benefits on the degree of

inequality and poverty, both for the overall population and for several subgroups.

Effects on overall inequality and poverty

Different inequality measures are considered in order to capture differences along the entire

distribution of income; as absolute inequality measure,11 we consider the Gini mean difference and

as relative indices,12 we compare the Gini index and the Atkinson index for different value of the

sensitivity parameter.

Table A1 shows that inequality has decreased when including health transfers in the income ofthe entire population; this is true for all the relative measures considered. In particular, the

percentage in the reduction is high for the inequality measures that are more sensitive to changes

in the tails of the distribution, i.e. the highest reduction is registered in correspondence to the

Atkinson index with the highest value of the inequality sensitive parameter. According to the

absolute Gini measure, instead, inequality slightly increases when adding health transfers to the

whole population. Such opposite trend of the relative and absolute inequality measures is because

the relative health transfers are decreasing with income, while the absolute health transfers are

slightly increasing with income (see Table 4). The same holds also in case of health transfersadded only to the 15-64 aged, while for the remaining two groups both absolute and relative

inequality decrease.

The result is in accordance with most of the studies on health transfers in kind (see, among

others, Baldini et al., 2006).

The reduction in income inequality reflects both the greater importance of health care services

for those at the bottom of the income distribution (i.e. the ratio of health care benefits to disposable

income declines across the distribution) and a distribution of public health care services that tends

to benefit most those in the lower quintiles.

The last three rows of the Table A1 show, instead, the Foster-Greer-Thorbecke (FGT) poverty

index for three different parameters; the poverty line is not fixed, in this analysis, i.e. it has been

updated with health transfers. The result on the overall distribution is a consistent decrease in all

the poverty indices; moreover, the percentage of reduction increases as the parameter of

sensitivity increases.

11 “An index of absolute inequality demands invariance to equal additions to all incomes.” (Lambert (2001),page 111)12

An index that satisfies the scale invariance, i.e. an index that is not “affected by equiproportionate (scale)changes in all incomes” is a relative index. (Lambert (2001), page 110)


34/48

33

Focusing on the subgroups, when health transfers in kind is added only to the youngest and

middle aged groups poverty decreases, while if the transfers are added only to the elderly, keeping

constant all the other groups, the headcount ratio reduces but the remaining measures of poverty

increase.

Tab. A1: Change in the overall inequality and poverty indices. Year 2003.

VALUE OF THE INDEXPROPORTIONAL CHANGE DUE TO

TRANSFERS TO PERSONS AGED

A B 0-14 15-64 65+ ALL

Gini meandifference

4846 4878 -10.9 5.3 -3.0 0.7

Gini 0.325 0.287 -1.9 -5.8 -2.9 -11.8

Atkinson 0.5 0.091 0.070 -3.9 -12.1 -4.7 -22.4

Atkinson 1.5 0.272 0.192 -5.0 -17.6 -2.4 -29.2

FGT0 0.186 0.144 -4.3 -8.9 -2.9 -22.8

FGT1 0.055 0.037 -6.5 -19.8 2.3 -33.1

FGT2 0.027 0.015 -9.8 -28.8 4.3 -43.1

Source:EU-SILC 2004

Effect on inequality and poverty by groups

Table A2 decomposes the changes in inequality by subgroups. For each different partition, the

inequality within groups is always much higher than the inequality between groups, i.e. itspercentage contribution to aggregate inequality is always greater than 90%; moreover, the

introduction of health transfers decreases both the within groups and the between groups

inequality, causing a decrease in total inequality.

When health transfers are added to the whole population, inequality decreases more for older

singles and couples and for the mono-parental households, for the families with unemployed or

pensioner householder, for families whose head has primary or no education, for the individuals

aged more than 64 years old. Moreover, inequality reduces more in the South Italy and islands and

in the towns with less than 2000 inhabitants. Such results coincide with what has been discussed

about the decomposition of the percentage increase in mean income by subgroups.

The impact of health transfers in kind on inequality can be analyzed graphically through the

Lorenz and Generalized Lorenz curves.

Figure A1 shows that before and after health transfers in kind, the group of over65 is the most

equal, while the Lorenz curves of the other two groups intersect. From Figure A2, moreover, we

see that, while there is no dominance among the groups in terms of Generalized Lorenz curves,

after the health transfers in kind such dominance occurs, with the group of over65 enjoying the

highest welfare level, followed by the 15-64 and by the 0-14 aged groups.


35/48

34

Fig. A1: Lorenz curves before (a) and after (b) health transfers, by age groups (65: green)

(a) (b)

Fig. A2: Generalized Lorenz curves before (a) and after (b) health transfers, by age groups

(65: green)

(a) (b)

Figure A3 and A4 compare the Lorenz and generalized Lorenz curves of the subgroups

created by the regional partition of Italy into North, Center and South. Both before and after the

health related transfers, South of Italy is the most unequal subgroup, while no dominance occurs

for the two remaining groups. In terms of generalized Lorenz curves, both before and after the

transfers a welfare dominance exists among the groups, revealing that the highest welfare level is

enjoyed in the North, followed by the Center and, much further, by the South; after the transfers,


36/48

35

the welfare of all the groups increases, but much more in the south, reducing, therefore, the

distance from the rest of the country.

Fig. A3: Lorenz curves before (a) and after (b) health transfers, by geographical area (North

Italy: blue, Center Italy: red, South Italy: green)

(a) (b)

Fig. A4: Generalized Lorenz curves before (a) and after (b) health transfers, by geographical

area (North Italy: blue, Center Italy: red, South Italy: green)

(a) (b)

A last observation regards the impact of health benefits on poverty levels, by subgroups, is

synthesized by Table A3. According to each of the measures FGT0, FGT1 and FGT2, poverty

reduces in each of the groups defined by different partitions. Each of the poverty measures

reduces mostly for old single or couples, for pensioner householders, for householders with


37/48

36

primary or less education, for the over64, for North and Center of Italy and for the very small

villages.

Tab. A2: Change in inequality, by subgroups. Year 2003.E F G H

Mean LogDeviation

(MLD) Baseline

% contribution toaggregate inequality

Characteristic of household or household head

% changein

inequality

BaselineIncome

andHealth

Household type

Older single persons or couples (at least one 65+) 0.138 -35.6 12.3 10.7

Younger single persons or couples (none 65+) 0.208 -20.0 16.0 17.3

Couple with children up to 18 0.205 -25.8 37.0 37.2

Mono-parental household 0.273 -30.7 3.8 3.6

Other household types 0.181 -25.3 28.5 28.8

% Within groups inequality 0.188 -26.2 97.7 97.6% Between groups inequality 0.004 -23.7 2.3 2.4


Blue collar worker 0.133 -25.9 14.3 14.4

White collar worker 0.121 -17.8 11.7 13.0

Self-employed 0.292 -21.3 26.2 27.9

Unemployed 0.379 -41.6 6.2 4.9

Pensioner 0.129 -30.6 19.4 18.2

Other 0.236 -33.2 13.5 12.2

% Within groups inequality 0.176 -26.7 91.2 90.5

% Between groups inequality 0.017 -20.2 8.8 9.5


Tertiary education 0.209 -16.3 9.8 11.1

Upper secondary education 0.164 -20.6 24.5 26.3

Lower secondary education 0.183 -26.8 28.8 28.5

Primary education or less 0.161 -32.5 26.4 24.1



Age of HH member

Below 25 0.216 -27.0 27.7 27.4

25-64 0.196 -23.6 56.8 58.8

Over 64 0.140 -35.3 14.0 12.3



Area

North 0.155 -23.3 36.5 37.9

Center 0.161 -24.7 16.0 16.3

South+islands 0.214 -30.1 39.3 37.2



City size

>50000 inhabitants 0.204 -24.6 44.3 45.2

2000-50000 inhabitants 0.184 -26.3 37.7 37.6


38/48


39/48

38

The remaining graphs illustrate the differences in percentage reduction of inequality and

poverty among regions, after the introduction of the health benefits. In Figure A5 we observe that

the regions may be grouped into three clusters, according to the levels of their pre and post Gini

index: regions with high pre and high post Gini, regions with low pre and low post Gini and regions

with middle pre and middle post Gini index.

Fig. A5: Gini index before and after health transfers, for each region. Year 2003.

ITALY

Piemonte Aosta

Lombardia

TrentinoVeneto

Friuli

Liguria

Emilia-Romagna

Toscana

Umbria

Marche

Lazio

Abruzzo

Molise

Campania

Puglia

Basilicata

Calabria

Sicilia

Sardegna

0.23

0.24

0.25

0.26

0.27

0.28

0.29

0.3

0.31

0.32

0.23 0.25 0.27 0.29 0.31 0.33 0.35 0.37

GINI BEFORE HEALTH TRANSFERS

G I N I A F T E R

H E A L T H

T R A N S F E

R S

Source:EU-SIL 2004

Fig. A6: Gini index before and after health transfers, by region. Year 2003.

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

F r i u

l i - V e n

e z i a

V e n e

t o

B a s i l i c a

t a

T r e n

t i n o

M a r

c h e

T o s c

a n a

M o l i

s e

L i g u r i a

U m b r

i a

E m i l i a

- R o m

a g n a

L o m b a

r d i a

P i e m

o n t e

S a r d

e g n a

V a l l e

d ' A

o s t a

P u g l i

a L a

z i o

I T A L

Y

A b r u

z z o

S i c i l i a

C a l a b

r i a

C a m

p a n i a

Gini baseline Gini baseline+health



40/48

39

Figures A6 and A7 show a slight re-ranking of the regions in terms both of Gini index and of

FGT2 poverty index. From Figure A8 we note that, among all the regions, Molise, Basilicata,

Trentino register the highest decrease in inequality, while Trentino, Abruzzo and Molise in poverty.

Fig. A7: Foster-Greer-Thorbecke (FGT2) index of poverty before and after health transfers, byregion. Year 2003.

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

T r e n

t i n o

V e n e

t o

F r i u l

i - V e n

e z i a

E m i l i a

- R o m

a g n a

L o m b a

r d i a

M a r

c h e

T o s c

a n a

U m b r

i a

V a l l e

d ' A

o s t a

P i e m

o n t e

A b r u

z z o

L i g u r i a

L a z i o

I T A L

Y

S a r d

e g n a

M o l i

s e

B a s i l i c a

t a

P u g l i

a

C a m

p a n i a

S i c i l i a

C a l a b

r i a

FGT2 baseline FGT2 baseline+health


Fig. A8: Decrease (in %) of Gini and FGT2 after health transfers, by region. Year 2003.

-70

-60

-50

-40

-30

-20

-10

0

I T A L Y

P i e m

o n t e

V a l l e

d ' A o s t a

L o m b a r d i a

T r e n t i n

o

V e n e t o

F r i u l i - V

e n e z i a

L i g u r i a

E m i l i a

- R o m a g n a

T o s c a n a

U m b r i a

M a r c h e

L a z i o

A b r u z z

o

M o l i s e

C a m p a n i a

P u g l i a

B a s i l i c a

t a

C a l a b

r i a

S i c i l i a

S a r d e g n a

Gini FGT2 Source:EU-SIL 2004


41/48

Tab. A4: Inequality and poverty changes, by region. Year 2003.

GINI ATKINSON 0.5 ATKINSON 1.5 FGT0 FGT1 FGT

REGION Baseline%change Baseline

%change Basel ine

%change Basel ine

%change Basel ine

%change Basel ine

%c

Piemonte 0.309 -11.0 0.081 -20.3 0.223 -23.4 0.166 -27.6 0.040 -42.3 0.016

Valle d'Aosta 0.317 -12.5 0.086 -23.3 0.252 -29.6 0.102 -40.7 0.028 -45.7 0.014Lombardia 0.303 -10.5 0.082 -19.1 0.220 -23.5 0.088 -29.7 0.022 -38.0 0.011

Trentino 0.280 -14.6 0.064 -27.2 0.198 -34.1 0.094 -55.3 0.018 -59.3 0.007

Veneto 0.274 -12.2 0.064 -22.3 0.196 -32.2 0.088 -27.9 0.020 -45.2 0.008

Friuli-V.G. 0.273 -11.3 0.062 -21.1 0.185 -25.5 0.086 -35.2 0.019 -40.9 0.008

Liguria 0.295 -12.1 0.075 -23.2 0.253 -34.2 0.128 -29.3 0.038 -35.3 0.019

Emilia R. 0.296 -10.9 0.076 -19.9 0.213 -23.7 0.077 -29.1 0.019 -37.4 0.009

Toscana 0.282 -11.3 0.075 -13.1 0.206 -17.1 0.125 -29.0 0.031 -34.7 0.013

Umbria 0.296 -11.6 0.075 -13.1 0.206 -17.1 0.125 -29.0 0.031 -34.7 0.013

Marche 0.281 -12.2 0.068 -22.3 0.191 -24.8 0.122 -39.5 0.025 -39.7 0.011

Lazio 0.319 -11.4 0.086 -22.1 0.260 -28.6 0.152 -26.3 0.043 -35.0 0.021

Abruzzo 0.329 -12.9 0.096 -22.8 0.234 -24.0 0.169 -31.0 0.042 -42.9 0.018

Molise 0.288 -15.0 0.070 -29.4 0.227 -35.1 0.204 -28.4 0.067 -38.8 0.035

Campania 0.351 -12.5 0.107 -24.1 0.303 -28.9 0.355 -15.6 0.118 -27.4 0.060

Puglia 0.319 -13.2 0.088 -26.6 0.303 -39.3 0.329 -14.6 0.111 -30.6 0.055

Basilicata 0.279 -14.6 0.065 -28.2 0.208 -32.4 0.314 -31.3 0.082 -36.8 0.038

Calabria 0.350 -13.3 0.107 -25.9 0.322 -34.8 0.392 -13.0 0.133 -29.2 0.067

Sicilia 0.349 -14.4 0.102 -27.9 0.322 -37.1 0.396 -20.8 0.129 -32.2 0.065

Sardegna 0.312 -13.0 0.084 -24.5 0.254 -32.3 0.223 -29.2 0.057 -36.7 0.027

ITALY 0.325 -11.76 0.091 -22.41 0.272 -29.23 0.186 -22.76 0.055 -33.08 0.027



42/48

41

Appendix B: Results from alternative health expenditure data

For comparative purpose, within the AIM-AP project, we apply the same method, insurance-

based, on the same data set, EU-SILC 2004, but using different source for the health care

expenditure amounts per capita per age group, the one used in Marical et al. (2006). The health

care costs per capita per age group employed in this second analysis are illustrated in Table B1.

The definition of total health expenditures considered is the following: “total expenditure on

health measures the final consumption of health care goods and services plus capital investment in

health care infrastructure. This includes spending by both public and private sources on medical

services and goods, public health and prevention programmes and administration. Excluded are

health-related expenditure such as training, research and environmental health” (OECD (2005)).

Total expenditure in 2003 is about 82,382 millions Euros, according to the definition ofhealthcare expenditure provided by the OECD.

Tab. B1 Health care costs per capita per age group (with correction factor) 2003

AGE HEALTHCARE COST

0-4 1027.72

5-9 682.85

10-14 596.57

15-19 640.78

20-24 730.74

25-29 818.1430-34 884.24

35-39 932.98

40-44 990.84

45-49 1086.49

50-54 1251.70

55-59 1508.19

60-64 1895.15

65-69 2370.35

70-74 2938.46

75-79 3556.47

80-84 4197.7285+ 3949.44

Source: Marical et al.(2006)


43/48

42

Tab. B2: Relevance of health transfers in kind. Year 2003.

INCOME SHARE

% INCREASE IN DISPOSABLEINCOME DUE TO TRASFERS TO

PERSONS AGED

MEAN TRASFER PER CAPITADUE TO TRASFERS TO

PERSONS AGED

QUINTILE A B 0-14 15-64 65+ ALL 0-14 15-64 65+ ALL1 7.44 8.58 4.79 18.32 11.23 34.33 266 1018 624 1908

2 12.73 13.68 2.69 10.59 9.17 22.45 255 1004 869 2128

3 17.12 17.59 1.65 8.52 6.42 16.59 210 1089 821 2120

4 22.60 22.45 0.98 7.41 4.08 12.47 166 1250 688 2103

5 40.10 37.70 0.54 4.33 1.87 6.74 161 1295 559 2015

ALL 100.00 100.00 1.42 7.57 4.77 13.74 211 1129 711 2049

Source:EU-SILC 2004

Tab. B3: Inequality and poverty changes. Year 2003.

VALUE OFTHE INDEX

PROPORTIONAL CHANGEDUE

TO TRANSFERS TOPERSONS AGED

A B 0-14 15-64 65+ ALL

Gini 0.325 0.290 -1.7 -5.1 -3.0 -10.9

Atkinson 0,5 0.091 0.072 -3.4 -10.9 -4.8 -20.8

Atkinson 1,5 0.272 0.198 -4.6 -16.3 -2.3 -27.0

FGT0 0.186 0.149 -3.9 -7.3 -1.8 -19.9

FGT1 0.055 0.040 -6.0 -17.6 3.8 -28.4

FGT2 0.027 0.017 -8.8 -26.0 5.8 -37.6

Source:EU-SILC 2004

The results are very similar to the ones obtained from the analysis discussed throughout the

report and in Appendix A, even if the levels of the impacts are slightly different.

In particular, the health related transfers in this second analysis induce lower increase in

income share for the lower quintiles and lower decrease in income share for the higher quintiles

than the first analysis does; therefore, the reduction in inequality due to health transfers will be

lower with this second method, as shown in Table B3.

Table B2 shows, moreover, that the percentage increases in income mean are lower with thissecond method, if compared to the method described in the report; we should note that when

transfers are added to the sole elderly, inequality increase much more with the second than with

the first approach.

Finally, looking at the Tables B4 and B5, this analysis underlines that the groups that benefit

the most from the health related transfers are the same underlined by the previous analysis; higher

reduction in inequality and poverty is registered with this second approach, for the groups that

include old individuals, such for the age-related, household type-related and educational level-

related groups.


44/48

43

Tab. B4: Mean income and inequality changes, by subgroups. Year 2003.

A B D E F G H

% contribution toaggregate inequality

Characteristic of household orhousehold head

Pop.share

in %

IncomeMean

(Baseline)

%increasein meanequiv.

Income

Mean LogDeviationBaseline

% changeIn

inequality

BaselineIncome

andHealth

Household type

Older single persons or couples 17.2 13916 26.7 0.138 -37.5 12.3 10.1

Younger single persons or couples 14.8 17524 8.5 0.208 -20.0 16.0 16.9

Couple with children up to 18 34.9 13845 10.6 0.205 -22.9 37.0 37.6

Mono-parental household 2.7 11789 10.0 0.273 -24.6 3.8 3.8

Other household types 30.4 15695 13.6 0.181 -23.4 28.5 28.8

% Within groups inequality ./. ./. ./. 0.188 -24.5 97.7 97.2

% Between groups inequality ./. ./. ./. 0.004 -6.7 2.3 2.8


Blue collar worker 20.9 12085 12.1 0.133 -22.4 14.3 14.7

White collar worker 18.6 17597 8.5 0.121 -15.1 11.7 13.0

Self-employed 17.3 17939 9.0 0.292 -18.8 26.2 28.0

Unemployed 3.2 8003 19.1 0.379 -38.1 6.2 5.1

Pensioner 29.1 15116 19.8 0.129 -31.2 19.4 17.6

Other 11.0 12415 19.9 0.236 -31.6 13.5 12.1




Tertiary education 9.1 23796 7.5 0.209 -14.7 9.8 11.1

Upper secondary education 28.8 16547 10.4 0.164 -17.8 24.5 26.5

Lower secondary education 30.4 13315 13.1 0.183 -23.7 28.8 29.0Primary education or less 31.7 12398 22.0 0.161 -30.5 26.4 24.2



Age of HH member

Below 25 24.8 13312 11.3 0.216 -23.6 27.7 27.9

25-64 55.9 15845 10.9 0.196 -21.8 56.8 58.5

Over 64 19.3 14251 25.9 0.140 -36.8 14.0 11.7


% Between groups inequality ./. ./. ./. 0.003 2.2 1.4 1.9

Area

North 45.3 17090 12.1 0.155 -22.0 36.5 37.5Center 19.3 16000 13.3 0.161 -22.8 16.0 16.3

South+islands 35.4 11527 17.2 0.214 -27.3 39.3 37.7



City size

>50000 inhabitants 41.9 15912 12.7 0.204 -22.9 44.3 45.0

2000-50000 inhabitants 39.5 14709 14.0 0.184 -24.4 37.7 37.5


45/48


46/48

45

References

AA.VV. (1999), “Pronto il riparto del FSN’99”, Agenzia Sanitaria Italiana (ASI) 1/2, 7/14

January 1999

Assr on line (2005), “Mobilità ospedaliera- Anno 2003”, in Monitor 14, 76-94.

Baldini, M., Bosi, P. and Pacifico, D. (2006), “Gli effetti distributivi dei trasferimenti in kind: il

Datos Económicos Italia 2003

Documents

Transcript of Datos Económicos Italia 2003