UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

ModernStats World Workshop

21-22 October 2024, Geneva

Incorporating AI into statistical standards: Enhancing GSBPM with

(generative) AI

Speaker: Riitta Piela, Statistics Finland

Author(s):

Abstract

Incorporating AI into statistical standards: Enhancing GSBPM with (generative) AI

In the evolving landscape of statistical production, AI stands out as a key force for change. Integrating AI with

GSBPM can significantly enhance statistical production. This abstract proposes a kind of ongoing

standardization and working model that utilizes generative AI across various stages of GSBPM, from data

collection to dissemination.

AI is capable of automating complex data processing tasks and acts as a versatile support system in different

production phases, speeding up and partly automating the GSBPM process. However, the adoption of

generative AI also brings challenges, especially in terms of maintaining transparency and adhering to ethical

standards. It is essential to develop standards that ensure all AI systems are efficient, fair and interpretable,

promoting responsible AI. The working model that is suggested in this abstract also emphasizes the importance

of ethical standards and transparency.

Additionally, it is important to harmonize AI applications with existing statistical standards and practices. Open

dialogue among statisticians, data scientists, machine learning engineers and ethical experts enables the

development of best practices. These practices leverage the possibilities of AI while protecting against its risks.

In this proposal will be presented how AI can be utilized in various sub-processes of GSBPM and how

principles of responsible AI can be applied to each implementation. It will also go through the different stages

of GSBPM and how AI is integrated into these various stages. AI is not only modifying statistical production

systems but is also likely to reform the processes themselves and even the standards of statistical production,

such as the generic statistical business process model (GSBPM).

In summary, the strategic integration of AI and especially generative AI into the GSBPM framework promises

to renew statistical production, making it more efficient and responsive. However, this integration must occur

carefully, committed to ethical practices and transparency, so that AI can support statistical integrity and not

compromise it.

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

ModernStats World Workshop

21-22 October 2024, Geneva

Streamlining statistical and data production

Speaker: Mikko Lindholm & Mari Ylä-Jarkko, Statistics Finland

Author(s):

Abstract

Statistical production has traditionally been built as a pipeline, so that each statistic has its own data collection,

serving the purposes of that particular statistic. The same way the IT-systems from validation tools to outputs

have been unique. The only common surface has been publication – with some variations for each statistic.

Statistics Finland has been transforming its production to more streamlined production system with common IT

tools and common data repositories. The work has started some 15 years ago from business statistics, but the

new strategic target is to have basic data repositories common for all products and thematic data repositories to

serve both statistical production as well as other data services, such as data for researchers.

Because most of our data comes from government registers, the data coverage is close to 100%. Also, all data

units have the same identifiers, such as person id, enterprise id, building id, etc. Thus, combining data from

different registers becomes possible, and the data repositories can be used to produce many other data products

besides official statistics.

In streamlining the data governance, also the data process is streamlined from data gathering to validation and

metadata. At the Statistical Office level, there is only one data process. On a more detailed level, there will be

some variations in the process, based on the data and collection method, but these variations will be kept to a

minimum.

This architecture aims at raising the productivity of statisticians, increasing the flexibility of resource use,

enabling quick production of temporary and experimental statistical products and improving data management

and regulatory compliance.

How to capture precariousness in the labour market by statistics? Dr. Anna Pärnänen & Dr. Hanna Sutela

Meeting of Experts on Measurement of Quality of

Employment

Geneve 14.-16.5.2024

In this presentation

• Measuring forms of employment by statistics

• Development in the 2000’s

• New forms of employment

• Debate on precariousness and insecurity

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Working life is changing, statistics keep on developing

63

37

Employed in Finland 1959

Statistics Finland (National LFS)

Employees Self-employed

• 1960’s dichotomy: self-employed and employees

• 1970’s self-employed in agriculture and forestry vs. other self-

employed; part-time vs. full-time

• 1980’s fixed-term employees annually;

contributing family workers

• 1990’s fixed-term employees monthly;

SE with employees vs. SE without employees, sole entrepreneurs,

freelancers, grant recipients

• 2000’s temporary agency workers

• 2010’s 0-hour-contracts, DPE, on-call workers, ”combi

employment”

• 2020’s (coming) independent and dependent workers,

dependent contractors (ICSE-18)

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Employees

Full-time Part-time

Unlimited

duration

Limited

duration

Unlimited

duration Limited

duration

0-hour-contracts

Temporary agency workers

Different types of self-employment

• Number of farmers keep on

declining

• The trend of employers is

decreasing since 2015, but

in the long run the changes

are quite modest

• Strongly increasing trend in

the number of self-

employed without

employees, esp. those aged

65-74 yrs. However, a clear

decrease from 2021 on.

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

0

50

100

150

200

250

SE with employees 15-74-v. SE without employees 15-74-v. Farmers 15-74-v.

SE with employees 15-64-v. SE without employees 15-64-v. Farmers 15-64-v.

Number of self-employed in Finland 1997–2023, 1 000

persons, LFS Light

entrepreneurs DPE

Development in the 2000’s

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Different forms of employment in 2000, 2010, 2019 and 2023 Share of all employees aged 15–64 years, % LFS

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

65.411.0

7.1

3.3 3.8

3.7 5.2

0.5

2000

Full-time employees with

a contract of unlimited

duration Full-time employees with

a contract of limited

duration Part-time employees

with a contract of

unlimited duration Part-time employees

with a contract of limited

durationn Self-employed in

agriculture and forestry

Other self-employed with

employees

Other self-employed

without employees

Contributing family

workers

65.410.1

8.4

3.5

2.5 3.7

6.0

0.4 2010

64.5 9.6

9.5

4.2

2.1

3.3

6.7

0.32019

64.2

8.8

10.0

5.7

1.8 2.9

6.3

0.4 2023

Different forms of employment in 2023 by gender Share of all employees aged 20–34 years, % LFS

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

64.813

7.3

7

0.8 1.6

5.3

0.4

MenFull-time employees with a

contract of unlimited

duration Full-time employees with a

contract of limited duration

Part-time employees with a

contract of unlimited

duration Part-time employees with a

contract of limited

durationn Self-employed in agriculture

and forestry

Other self-employed with

employees

Other self-employed

without employees*

Contributing family

workers**

48.7

16.9

18.4

10.4

4.7

Women

* Too small to be pudlished for women

**Too small to be publisehd

Different forms of employment in 2023 by origin Share of all employees aged 15–64 years, % LFS

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

64.2 9.1

9.7

4.6

2.1

3.2 6.7

0.4

Finnish originFull-time employees with a

contract of unlimited

duration Full-time employees with a

contract of limited duration

Part-time employees with a

contract of unlimited

duration Part-time employees with a

contract of limited duration

Self-employed in

agriculture and forestry

Other self-employed with

employees

Other self-employed

without employees*

Contributing family

workers**

53.8

14.2

10.7

8.4

2.9 8.9

Foreign origin

Full-time employees with a contract of unlimited duration, % of all employed in Germany, Italy and Sweden 2000/2004, 2022/2023

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

61.8 57.7 56.4 55

61.8

68.2

47.3 42.1

51.2 48.1

52.9

64.9

73.2 71.4

59.8 60.1

70 71.1

0

10

20

30

40

50

60

70

80

90

100

2000 2022* 2004 2023* 2000 2023*

Germany (15-64 yrs) Italy Sweden (16-64 yrs)

Total Women Men * Break in time series

New forms of employment

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

New forms of employment 2022

Zero-hour contracts • C. 5 % of all employees

• Especially those aged

under 25 and over 65 yrs.

• 40 % are students

• C. 80 % part-time workers

• More common among

people of foreign origin

Temporary agency work • 2,7 % of all employees

• The number doubled in ten years

• Mostly contracts of unlimited duration

• Men mainly full-time, women part-time

• C. 20% of foreign origin

• Accommodation and food services,

administrative a. support services,

transport and storage ,construction,

trade

Light entrepeneurs • C. 2 % of all employed

• Number grown over 60%

from year 2019

• Usually small scale work

• An average income from

billing service c. 5 300

e/yearly

• C. 20 % of foreign origin

Digital platform work

• 3,9 % of all population had

income from DPE during the

last 12 months

• 1,2 % during the last month

• More commmon among

those of foreign origin

• Trade, social media content,

renting out accommodation,

courier and taxi services

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

• ”New” forms of employment

are increasing, but still

marginal

0

20000

40000

60000

80000

100000

120000

2 0

0 9

2 0

1 0

2 0

1 1

2 0

1 2

2 0

1 3

2 0

1 4

2 0

1 5

2 0

1 6

2 0

1 7

2 0

1 8

2 0

1 9

2 0

2 0

2 0

2 1

2 0

2 2

2 0

2 3

Number of temporary agency workers

2009–2023, 15 to 74 years, LFS

0

20,000

40,000

60,000

80,000

100,000

120,000

2014 2018 2021 2023

Number of employees with 0-hour

contract, 15 to74 years, LFS

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Number of light entrepreneurs 2019-2022,

18 to 74 yrs, employment statistics

Why digital platform work?

0 10 20 30 40 50 60

Working this way gave an opportunity for

additional income and networking

Wanted to work flexibly

The work one wanted is available mainly or only

through platforms

No other work was available

Main reason for platform work

Those who had done platform work during the previous month, %;

DPE –pilot survey, LFS 2022

• The opportunity for additional income

and networking more commonly the

main reasons for women, those of

Finnish origin and those aged under

35

• The only job available or an

opportunity to work in a flexible way

more commonly the main reason for

men, those of foreign origin and those

aged 35 and more

• For 20 % of those having done DPE in

the previous month, DPE was the main

source of income

• For 54 %, less than ¼ of income from

DPE

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Debate on precariousness and insecurity

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Forms of employment in Finland, all employed aged 15 to 64 yrs 2014

%

2023

%

Change

%-points

Full-time employees with a contract of unlimited duration, excl. 0-hours and TAW 63,3 63,1 -0,2

Full-time employees with a contract of unlimited duration, 0-hrs-contract 0,8 0,5 -0,3

Full-time employees with a contract of unlimited duration, TAW 0,3 0,6 0,3

Full-time employees with a contract of unlimited duration, 0-hrs and TAW 0,1 0,1 0

Part-time employees with a contract of unlimited duration, excl. 0-hours and TAW 7 7,7 0,7

Part-time employees with a contract of unlimited duration, 0-hrs-contract 1,5 1,7 0,2

Part-time employees with a contract of unlimited duration, TAW 0,1 0,2 0,1

Part-time employees with a contract of unlimited duration, 0-hrs and TAW 0,2 0,3 0,1

Full-time employees with a contract of limited duration, excl. 0-hours and TAW 9,4 7,8 -1,6

Full-time employees with a contract of limited duration, 0-hrs-contract 0,2 0,1 -0,1

Full-time employees with a contract of limited duration, TAW 0,4 0,5 0,1

Full-time employees with a contract of limited duration, 0-hrs and TAW 0,1 0,1 0

Part-time employees with a contract of limited duration, excl. 0-hours and TAW 2,7 3,3 0,6

Part-time employees with a contract of limited duration, 0-hrs-contract 0,6 1 0,4

Part-time employees with a contract of limited duration, TAW 0,1 0,5 0,4

Part-time employees with a contract of limited duration, 0-hrs and TAW 0,1 0,2 0,1

Other employees 0,2 0,9 0,7

Farmers 2,4 1,8 -0,6

Self-employed with employees 3,7 2,9 -0,8

Self-employed without employees 6,5 6,3 -0,2

Contributing family workers 0,3 0,4 0,1 14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Nowadays nobody is sheltered from job insecurity

11 13

18 16

14

22

15

7 7

14 15

11

17

13

3

9 12

17 14

21

17

0

10

20

30

40

50

60

1984 1990 1997 2003 2008 2013 2018

Experienced threat of dismissal by socio economic group, %, Finnish

Quality of Work Life Surveys 1984, 1990, 1997, 2003, 2008, 2013 &

2018

Workers Lower level salaried Upper level salaried

• In the past decades, the level

of perceived job insecurity

has varied according to

economic fluctuations

• However, a clear change has

taken place as comes to the

distribution of perceived

inscurity

• High position in the labour

market, good education or a

job in the public sector do not

protect any more from threats

of unemployment and

dismissal

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

How well do the labour statistics capture precariousness in the labour market?

• Changes in the labour market and digitalisation have brought along new forms of employment such as DPE

• However, various forms of employment equalling to various levels of (in)security have always existed …

• …all the ”new” is not new, but just more visible thanks to the development of labour legislation and labour statistics

• Labour statistics classify, draw boundaries and simplify the multifaceted reality with blurred boundaries

• According to statistics, ”normal employment relationships” have maintained their position surprisingly well in the 2000’s

• How well do quantitative statistical methods and classifications capture the (increased?) insecurity and (increased?)

fragmentation in the labour market?

• How well does the LFS (with decreasing response rates) include the most vulnerable groups and capture all forms of work ?

• What about qualitative changes? Does a contract of unlimited duration equal the same level of stability as it did earlier?

• Statistics provide tools to monitor the development, but do not tell the whole story

• Contextualisation and qualitative sources needed to understand where does the debate on precariousness stem from

14-16th May, 2024 Statistics Finland | [email protected], [email protected]

Thank you for your attention!

Legacy of COVID-19: post-crisis remote work in Finland Dr. Hanna Sutela & Dr. Anna Pärnänen

Meeting of Experts on Measurement of Quality of Employment

Geneve 14.-16.5.2024

Finnish Quality of Work Life Surveys • 1977, 1984, 1990, 1997, 2003, 2008, 2013, 2018, (2021), 2023 by Statistics Finland

• 3 000–5 800 employee respondents per data

• Response rate 68-91 %

• Face-to-face until 2018, web survey 2021 & 2023 (break in time series)

• Core questionnaire remains the same

• New themes added, old stuff removed/rotated; special themes

2018 digitalization of the working life

2021 Impact of COVID-19 on working conditions

2023 Ecological sustainability in work life

• Provides information for

policymaking

public debate

data for research (possibility to link survey data to register data)

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

One in three employee worked remotely in 2023

4 6 9

16

28

41

35

3 4 7

13

27

40 37

6 7 10

19

29

43

33

0

10

20

30

40

50

60

70

80

90

100

1997 2003 2008 2013 2018 2021 2023

Share of employees working remotely in Finland, % Finnish Quality of Work Life Surveys 1997, 2003, 2008, 2013, 2018 & 2023,

The Impact of COVID-19 crisis on work life 2021 (Break in time series 2018)

Total Women Men

• Remote work grew in strides in the

2010s, peaked during the COVID-19

crisis, but has fallen somewhat after

that

• More common for upper-level

salaried than low-level salaried;

among manual workers only a few

do remote work

• In 2023 more common for women

than men, for the first time

• Most common in the central

government sector, the least

common in the county government

sector (social services and health

care)

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

In addition, 8 % had

teleworked at some

point during the COVID-

19 crisis

More than one in five employee works remotely at least half of their working hours

1

1

1

7

7

7

7

8

8

6

6

6

6

7

6

6

7

7

67

63

65

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Men

Women

Total

Share of remote work in working hours, %, employees,

FQWLS 2023

All of the time Almost all of the time Around 3/4 of the time

Around 1/2 of the time Around 1/4 of the time Less

Does not work remotely

• 22 % of employees worked remotely

at least half of the working hours in

the autumn 2023…

• …equalling to 63 % of all remote

workers

• Down from 35 % or c. 80 % in spring

2021, respectively

• In 2021 intensive teleworking was

more common for men; in 2023

there were barely any gender

differences

• Highest intensity in the central

government sector, the lowest in the

county government sector

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Home has become the main place of work for almost half of the teleworkers…

80

22

51

6

1

2

1

0.5

1

0.6

0.5

12

76

45

1 1 1

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2018 2021 2023

The main place of work, employees in remote work, %, FQWLS 2018 & 2023, The Impact of COVID-19

crisis on work life 2021

Some other place

Home

Construction site, street, outside

Vehicle

Client's premises

Employer's premises

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

…and for one in six of all employees

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

88

62

73 71

47

61

7

4

6 9

4

6

1

1

1 6

6

7

1

1

8

5

8

3

30

17

4

35

16

2 1 2 2 2

0

10

20

30

40

50

60

70

80

90

100

2018 2021 2023 2018 2021 2023

The main place of work, all employees, %, FQWLS 2018 & 2023, The Impact of COVID-19 crisis on

work life 2021

Some other place

Home

Construction site, street, outside

Vehicle

Client's premises

Employer's premises

Women Men

Less ergonomics, but more peace in remote work

11

10

14

26

14

8

32

6

5

5

0 10 20 30 40 50 60

Poor or glaring lighting

Restless work environment

Repeated one-sided movements

Difficult or uncomfortable positions

Lack of space

Experienced adverse factors by work environment,

employees working remotely, %, FQWLS 2023

Workplace working environment Remote working environment

• Lack of ergonomics more

common for women than men in

remote work

• Eg. difficult or uncomfortable

positions: women 34 %, men 16

%

• In 2021 men in remote work

more often had their home office

or dedicated working place than

women; women more often at

the kitchen table etc.

• 20 % of those in remote work

has assessed the working

conditions and environment of

remote work together with their

employer in 2023

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Highest share of remote work among those having no workstation at the employer’s premises

9

11

17

33

31

37

43

50

60

51

38

18

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%100%

Own office room

Own workstation in a shared room (2-4 persons)

Own workstation in an open / activity-based office

Open/Activity-based office, no workstation

Share of remote work in working hours by the type of office

at employer’s premises, %, employees working both

remotely and at employer’s premises, FQWLS 2023

(Almost) all the time in remote work 1/2-3/4 of time in remote work 1/4 or less in remote work

• Increased share of remote work

and increasing share of open

offices create a circle:

• The more employees work remotely,

the less office space is needed and

the more open offices grow in

popularity…

• But the more restlessness in the

office, the more employees might

be tempted to stay at home in order

to be able to concentrate…

• …and the less office space will be

needed etc.

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Men have more commonly possibilities to influence when and how much in remote work than women

50

40

45

33

31

32

0 20 40 60 80 100

Men

Women

Total

Possibilities to inluence when and how much in

remote work, employees doing remote work by

gender, %, FQWLS 2023

A lot Quite a lot

• Upper-level salaried have more

possibilities than lower-level salaried

• Least possibilities in the county

government secrtor, the most in the

universities

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Every third can choose the amount of remote work almost

freely

22

26

24

24

30

27

12

12

12

42

31

36

0% 20% 40% 60% 80% 100%

Men

Women

Total

Amount of working at the workplace required by

the employer, employees in remote work, %,

FQWLS 2023

Around three to four days a week Around one to two days a week

At least on some days a month Can choose almost freely

• Weekly presence required more

commonly in women’s than men’s work

• Gender difference stemming (again)

from the gender segregation: female-

dominated county and local

government sectors require the most

presence, offer the least free choice

• Upper-level salaried having more choice

than lower-levels

• Among those, who can choose the

amount of remote work almost freely,

43 % work remotely (almost) all the

working hours, altogether 76 % at least

half of the hours

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Wish to work more remotely does not always exclude a wish to have more F2F contacts

22

30

34

36

71

64

60

55

7

6

6

9

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

1/4 or less

At least 1/2 of time

3/4 of the time

(Almost) all the time

Opinions of the sufficiency of face-to-face interaction

with co-workers by the share of remote work, %, FQWLS

2023

Would need more interaction This is enough Would need less

• Men in remote work more often than

women say they would need more

F2F interaction

• On the other hand, a good 10 % of

men doing only remote work would

cope well with even less F2F

interaction

• Those who had the best possibilities

to influence how much to work

remotely, most often work remotely a

large share of their working hours –

and they also long for F2F interaction

more often than the others ☺

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Every fifth does not record all the working hours done in remote work

79

81

80

0 20 40 60 80 100

Men

Women

Total

Share of those recording all working hours in

remote work, employees doing remote work

with obligation to record working hours by

gender, %

• Upper-level salaried employees, those in

supervisory positions and those working at the

universities more commonly not recording all the

hours in remote work, even if requested to do so

• The older the remote workers, the more often a

part of the working hours is not recorded

• About 20 % of men and 10 % of women do not

have an obligation to record hours (= managers

outside working time legislation or professional

working total hours)

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

COVID-19 crisis was a kick off to remote work for many

17

42

41

Employees in remote work in 2023 and

their status in the spring 2021, %, FQWLS

panel data (those employed as employees

both 2021 & 2023)

Not in remote work in the spring 2021

Remote work in 2021, started working remotely during the crisis

Remote work in 2021, had worked remotely already before

0 20 40 60 80 100

Not in remote work in the spring

2021

Remote work in 2021, started

working remotely during the crisis

Remote work in 2021, had worked

remotely already before

Share of those working remotely in 2023

by their status in the spring 2021, %,

Quality of Work Life Survey panel data

Works remotely in 2023

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Two out of five wage and salary earners are not interested in remote work

35 37 33

3 3

3

4 3 5

22 24

20

36 33 38

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Total Women Men

Remote work, interest in it and possibility to work remotely in one’s job, employees, %,

FQWLS 2023

Not interested and would not be

possible/would be difficult to arrange

Would be interested, but would not be

possible/would be difficult to arrange

Not interested, but would be possible

Would be interested, would be possible

Works remotely

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

To sum up: a profound change in the work life • The exceptional circumstances during COVID19 crisis boosted remote work: even those not interested in, and

some of those for whom it was in principle not possible, worked remotely (c. 49 % at the most)

• Afterwards, the situation has ”normalised” at the level of 35 %: almost all interested and almost all whose job

allows work remotely

• Not only the prevalence but also the intensity of remote work has increased compared to pre-COVID19

situation

• Home has become the main place of work for almost half of the remote workers, fewer and fewer evn have a

workstation at the employer’s premises any more

• Gender differences in remote work stem, for the most, from the gender segregation and structural changes in

the labour market

• Shortcomings in ergonomics, too few F2F contacts, unrecorded working hours… but flexibility, peaceful work

environment and better possibilities to reconcile work and private life

• Challenging situations especially as comes to the young newcomers of the labour market; what is happening

to the sense of community and belonging?

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

For more information

• The share of remote working has fallen in 2023

from the coronavirus period but was higher than in

pre-pandemic years - Statistics Finland

• Good one-fifth of wage and salary earners in remote

work in 2023 – home is the main place of work for

more and more people - Statistics Finland

Thank you for your attention!

14th-16th May 2024 Statistics Finland | [email protected], [email protected]

Digital economy Supply and Use Tables in Finland Eljas Tuomaala, 23-25 April 2024, UNECE: Group of Experts on National Accounts

Background

• Grant funding by Eurostat

• Project from 2021 to 2023

• Reference year 2018

• Following the OECD Guidelines

• Excluding products beyond SNA2008/ESA2010

https://digital-strategy.ec.europa.eu/en/policies/desi

11 April, 2024 Statistics Finland | [email protected]

Methodology

11 April, 2024 Statistics Finland | [email protected]

Main principles

• Focus on compilation of the complete tables and high priority indicators

• Use readily available data sources only • Identify data gaps • As easily reproducible as possible • Experimental by nature

11 April, 2024 Statistics Finland | [email protected]

Digital industries

Digitally enabling industries

• Standard SUT aggregate industry level data

• Unit level data when needed to split industries

Digital intermediary platforms charging a fee • A list of manually

identified units combined with unit level data

Data and advertising driven digital platforms • A list of manually

identified units combined with unit level data

Firms dependent on intermediary platforms • ICT Usage survey

data for incorporated • Indirect (short term

accommodation, food delivery) data sources for unincorporated

11 April, 2024 Statistics Finland | [email protected]

Digital industries

E-Tailers

• ICT Usage survey data combined with unit level data

• ISIC rev. 4 Category 4791 “Retail sale via mail order houses or via Internet”

Digital only firms providing financial and insurance services • A list of units

identified as digital by Bank of Finland

• Not fully comprehensive

Other producers only operating digitally

• Research on most well-known firms

• Variety of issues: identifying, part of the main business, foreign units

• Industry excluded 11 April, 2024 Statistics Finland | [email protected]

• Included: • ICT goods and Priced

digital services • Priced cloud computing

services • Priced digital

intermediary services

• Excluded: • Products outside of

SNA2008/ESA2010 production boundary, such as data

• In the Finnish SUTs the total number of products used at the working level is 836

ICT goods and Priced digital services

• Reclassifying the suitable products within the standard SUTs

• Overlapping with cloud computing and digital intermediary services is checked and adjusted accordingly

Priced cloud computing services

• Identifying potential product categories (CPA 58.2, 62.01, 63.11)

• Company study: breakdown between cloud and non-cloud

• Combining values and ratios on relevant NACE activities, final CCS output estimate

• No data sources for imports and use categories, therefore calculated proportionally based on the above estimate

Priced digital intermediary services

• Domestic output is based on the Digital intermediary platform industry estimate

• Imports estimated as case studies, not reliable enough

• Industry breakdown for intermediate consumption is challenging

Digital products

11 April, 2024 Statistics Finland | [email protected]

Transaction types Digitally ordered

• ICT Usage survey with strong assumptions - Questions on sales via digital means (split into direct

from counterparty and via digital intermediary platforms)

- Strong assumptions: for example, the percentage of digital ordering in telecommunication industry (NACE 61) was used on the similar product (CPA 61)

- Did not provide plausible results on the use side

• Household final consumption expenditure based on a report on e-commerce in Finland in 2018 by Paytrail

- Paytrail is a payment processing company that produces a report on e-commerce annually

• Some other minor data sources

Digitally delivered

• No decent data available

• Potential value of digitally deliverable services instead of actually digitally delivered

• Trade in Services by Mode of Supply (MoS) - Strong assumptions: mode of supply category

one classified as digitally delivered - Some other adjustments

• Very difficult in practice

11 April, 2024 Statistics Finland | [email protected]

Results

11 April, 2024 Statistics Finland | [email protected]

Results: Digital industries Digital industries’ gross value added 7.4 %* of the total economy

Digital industries in Finland, year 2018, EUR million, share of the digital economy

Digital industry Output Gross value

added Output

Gross value added

Data and advertising driven digital platforms

90 50 0 % 0 %

Digital intermediary platforms charging a fee

110 20 0 % 0 %

Digital only firms providing finance and insurance services

150 90 0 % 1 %

Digitally enabling industries 27,400 12,500 85 % 84 %

E-tailers 4,320 2,070 13 % 14 %

Firms dependent on intermediary platforms, corporated

30 10 0 % 0 %

Firms dependent on intermediary platforms, unincorporated

200 140 1 % 1 %

Other producers only operating digitally

NA NA - -

Digital economy 32,300 14,880 100 % 100 %

11 April, 2024 Statistics Finland | [email protected]

*Digital Economy is not deducted from other industries in this figure

Results: Digital products Share of digital products of all products in 2018 Digital products in Finland, year 2018, EUR million,

share of all products

Products Total supply and use in purchasers prices, EUR

million Share of all products

ICT goods 10,270 1,8 %

Priced digital services 25,810 4,6 %

Priced cloud computing services 3,740 0,7 %

Digital intermediary services 130 0,0 %

Total digital products 39,950 7,1 %

All products 561,103 100 %

11 April, 2024 Statistics Finland | [email protected]

• Digitally ordered: • Estimates on product

level • Household final

consumption expenditure based on a separate estimate: 9 % of total

• Digitally delivered: • 14 % of total output

potentially digitally deliverable

19%

14%

6%

22%

16%

0%

5%

10%

15%

20%

25%

Imports of goods and services

Output Final consumption Exports of goods and services

Intermediate consumption

Digitally ordered, % of total

Results: Transaction types

11 April, 2024 Statistics Finland | [email protected]

Release

National publication in October 2023: an article on Statistics Finland’s website

Also in English: https://www.stat.fi/tietotrendit/artikkelit/202 4/digital-economy-rivals-construction-in- finland/

Questions from users: Data on employment (productivity) International comparison

11 April, 2024 Statistics Finland | [email protected]

Lessons learned

11 April, 2024 Statistics Finland | [email protected]

Opportunities

Even though the existing data are limited a lot can be done Focus on main indicators Time series more useful than the full tables? Some parts easier to compile for multiple years than others

Challenges

Data gaps exist: • Classifications in existing inquiries

and statistics (CCS, DIS) • International trade, foreign units • Transaction types very difficult • Lack of coverage: micro/small

units, product level data Quality ranging from reasonably reliable to highly experimental

What could be done?

Expanding current inquiries: new questions and/or improve coverage Find new data sources Elaboration of guidance and practical examples

Thank you!

11 April, 2024Statistics Finland | [email protected]

1

UNITED NATIONS ECONOMIC COMMISSION FOR EUROPE

CONFERENCE OF EUROPEAN STATISTICIANS

Workshop on Ethics in Modern Statistical Organisations

26-28 March 2024, Geneva, Switzerland

15 March 2024

Structure of ethical issues in new data ecosystems (pre-workshop version) Marianne Johnson, Timo Koskimäki, Markus Sovala (Statistics Finland)

[email protected]

Abstract

The current ethical frameworks for statistics, and for social research, rely conceptually on the idea of an organisation – or researcher – collecting data directly from the units they aim to study. In the era of data ecosystems and frequent sharing of data, the assumption of direct data collection has become obsolete.

The statistical community has recognised that the new data-era generates new ethical issues. Trust-enhancing measures like ethical commissions, ethics-related web segments and ethics communication strategies has been set up to overcome the everyday distrust situations.

The paper analyses the new ethical issues using the concept of professional ethics as a tool for analysis. A standard professional ethics code consists of a value statement, a set of claims to subjects and stakeholders, and a set of promises (code of conduct). The most prominent example is the medical profession, but this kind of deconstruction can be applied to any other professions, enabling a systematic study of ethical issues related to professional practices.

We analyse examples of recent ethical debates, related to data and statistics, from the point of view of professional ethics, to better understand the reasons for the emergence of negative public debates and other indications of distrust.

The key result is that current ethical codes for statistics do not recognize the new issues related to use of data. A proposal is made to amend the codes by adding descriptions and guidance related to (at least) the following situations:

- Secondary and tertiary use of data Example on secondary use is NSO using privately held or administrative data for statistics, example on tertiary use of data is researchers using micro-data provided by NSO

- Combining data from different sources In 2024 UN Statistical Commission meeting very many participating countries indicated that they are going to base their 2030 census on administrative data. We urgently need ethical guidelines on combining administrative data.

- Use of micro-data (or very granular statistics) in knowledge-based decision making. Data-based decisions typically impact specific groups of individuals or even individuals. There are already now many borderline cases on ISI principle 12, protection of the subjects. One example is to determine taxable value of a dwelling using micro-data on dwelling prices and statistical modelling; many decisions relating to social benefits impact (smallish) groups of people, and statistics, even micro data, are commonly used to allocate benefits.

2

Structure of ethical issues in new data ecosystems (pre-workshop version) Marianne Johnson, Timo Koskimäki, Marianne Johnson (Statistics Finland)

[email protected]

Paper

Introduction

1. The statistical community has recognised that the new data-era generates new ethical issues. Trust-enhancing measures like ethical commissions, ethics-related web segments and ethics communication strategies has been set up to overcome the everyday distrust situations1. However, it has been difficult to determine what was the cause that triggered the social distrust on statistics.

2. In this paper, we analyse the causes using the concept of professional ethical standard as a tool to understand the distrust situation. The study of professions, including professional ethics, is a well-established field of study2. As the ethical codes are intended to demonstrate core values and preferred code of conduct to stakeholders, including citizens and the media, they are a natural reflection surface to emerging distrust- situations. As a result of emerging distrust, there has also been a vivid discussion on ethical standards among professional (official) statisticians3.

3. First, we reflect the basic ideas of the study of professions and relate that to ethical codes and principles related to official statistics. Then we provide two examples of recent public debates on ethics and statistics. We conclude by reflecting our examples with the professional codes of statisticians.

Features of a professional ethics

4. A standard professional ethics structure consists of the following elements:

 a value statement, specifying the professional promise; what is the specific common good that the profession claims to provide to the society and individual people, e.g. health professionals are promoting public health and curing diseases.

 a set of claims to subjects and stakeholders, e.g. permission to violate intimacy, cause pain when diagnosing, professional integrity and education system, resources

 a set of promises (code of conduct), e.g. respect for the patient's right to self-determination, the duty to 'do good', – the duty to 'not do bad', to treat all people equally and equitably.

5. In the formal ethical codes this structure of values, promises and claims is supported by practical guidance highlighting how professionals should behave (e.g. Statistics Finland 1993). The current professional ethics for statistics – the ISI Declaration - also recognises this structure: “[… ] the Principles inherently reflect the obligations and responsibilities of – as well as the resulting conflicts faced by – statisticians to forces and pressures outside of their own performance, namely to and from: • Society

• Employers, Clients, and Funders

• Colleagues

• Subjects

1 ECE/CES/2023/24 2 For an overview see Suddaby and Muzio 2015, Bateman 2012 3 e.g. IAOS 2022; ECE/CES/2022/2

3

The ISI Declaration on professional ethics then elaborates these dimensions to four shared values - respect, professionalism, truthfulness and integrity - and then continues with a list of 12 ethical principles4. One can find similar ethical characterisations from many statistical organisations (ECE/CES/2023/24)

6. For this paper, we will use the structure of the ISI declaration of professional ethics as tool for analysis. The examples we use here come from Finland and Norway. The Finnish case is about the use of statistical micro data for knowledge-based decision making: Should the state know, behind your back, how much you will potentially cost for the health-care service provider. The Norwegian case is about the National Statistical Office access to new, privately held data sources5. How much does the National Statistical Office need to know about the individuals in the country. The Incidents Micro-data for knowledge-based decision-making

7. In Finland it has taken decades of planning to find a solution to restructure public social- and health care services. One aim has been to take the responsibility to provide these services from the municipalities and instead have new welfare areas, consisting of ten ore more local governments (municipalities). This change was realised in 2023. At the same time, there was also a plan on having private health care providers take a bigger role in health care system. The welfare areas would buy the basic health –care services from private enterprises. The price to be paid to private enterprises would be based on the health-related characteristics of the client. The clients would be free to choose between different service providers, and the service-provider would be compensated by the welfare area depending on the health-related characteristics of the clientele. This latter part of the plan has not realised and is currently not on the political agenda. However, it provides an interesting case of an ambitious plan to apply knowledge-based decision making in the health-care system.

8. Establishing a system of private production of publicly provided services would be implemented by channelling central government funding for services in advance. This would be achieved by the state compensating for the provided services in advance. A model was to be developed for calculating how much funding should be distributed to the different health care actors. The task was given to the Finnish Institution for Health and Welfare (THL), from where an application was sent out to Statistics Finland to get use of individual level data gathered by Statistics Finland from different registers for statistical production.

9. The plan was to use background information from Statistics Finland (such as age, education, language, occupation, place of residence, socioeconomic status etc) and link this data to information on the persons health conditions from healthcare and medication registers from THL and the Social Insurance Institution KELA. With this data on individual level THL would be able to assess for each citizen a risk value, that would give an estimate of the persons upcoming costs as a user of the health care system. The plan was that the calculated risk value would only be known by THL and KELA and that not even individuals themselves would be told the results of the calculations.

10. There were many legal problems with the plan. The EU general data protection regulation states that the use of data should be transparent, and everybody should have the right to check their data. The Statistics Act states that data gathered for statistical purposes can only be given to another statistical agency (which the THL is) for statistical purposes. Statistics Finland can give access to pseudonymized data to be used in scientific research and statistical analyses. The law explicitly rules out that data obtained from Statistics Finland could be used for decisions relating to the individual. The case as such did not involve such decisions, but the sole existence of the coefficient was perceived as a risk that it could be used on individual decision making.

11. There was wide media coverage of the plan that came to be known as the Capitation reimbursement model. First the articles were neutral and brought forth the ideas behind the model with informative interviews with the

4 ISI 2023 5 The description of the Norwegian case is made by the authors of the paper and not Statistics Norway

4

experts at THL. The benefits of the model were put forth and it was implied that the government had much better possibilities to calculate the individual risk values than insurance companies, as the government agencies already have so extensive data on individuals available.

12. Quite fast the articles started to question if giving ‘health points’ to persons was the right way to go ahead. The Data Protection Ombud took part in the discussions, as well as other legal scholars, and the plan was condemned as among other things citizens’ rights had not been taken into account. One of Statistics Finland directors also pointed out the ethical issues on a semi-official blog-platform of Statistics Finland6

13. Statistics Finland sent out a letter of inquiry to THL asking for a better description on how the data from Statistics Finland was planned to be used and on what grounds they would have the right to handle the data.

THL came around and submitted an answer to Statistics Finland as well as a new data application that differed in many ways from the previous application. THL still requested much of the same data from Statistics Finland to be linked to health data from THL and medication data from KELA, but now the data was to be used for a research project to come up with a model for distributing funding to each Welfare area. The data would not include identifiers and would be used over Statistics Finland’s remote access system by researchers stated in the data permit.

14. As the public discussion was critical towards purely individual coefficients, the legislators ended up to a solution where no individual data would be permitted for the calculation of coefficients. Instead, a rather limited set of variables to be used was defined in the draft legal act. The legal act was never passed, due to resignation of the Government.

Access to new, privately held data sources

15. Norway introduced a new Statistics Act in 2019. The Act granted Statistics Norway rights to access to privately held data for statistical purposes. One of the first attempts to get access to privately held data was to gather data on private consumption using cash-register records combined with payment card data. However, one key actor in the process, the supermarket chains contested the Norwegian statistical office's authority to request regular submission of purchases data collected by these stores. The data was intended to be collected directly from point-of-sale systems and would encompass nearly all grocery purchases made by the entire Norwegian population. Although the purchase data does not include personal identification numbers, Statistics Norway would be able to link more than 70 % of all grocery purchases (receipts) to persons and households through debit card transactions from the banking systems. After this it is possible to link the purchase data to other data on individuals and households already held by Statistics Norway, and thus to e.g. generate information on categorized product purchases by household size, income, education level and geographical region, as well as possibly produce new statistics on dietary habits.

16. One of the supermarket chains filed a complaint with the Norwegian Data Protection Agency (DPA). After investigation, the DPA concluded that the public authority (the State) was intruding upon citizens' privacy by collecting such data. It emphasized that there are limits on what data that official agencies should handle when it comes to personal data, even though the intentions are good. Laws state that everybody has the right for respect of their and their families’ private life. The DPA found Statistics Norway lacked sufficient legal grounds to process transactional personal data as proposed and, under Article 58(2)(f) of the GDPR, banned the data processing.

17. Primarily, the issue revolves around concerns about government overreach in data collection rather than mistrust in Statistics Norway's utilization of individual data. Even political parties took part in the debate saying that they do not want Norway to become a surveillance society. Statistics Norway contends that it is wrong to identify the agency as "the state" and asserts its’ authority to gather necessary data under the Norwegian

6 Koskimäki 2018

5

Statistical Act. They are not interested in individuals but in statistics. Developing new methods is an integrated part of statistics production. There are also high standards set for producing high quality statistics which Statistics Norway aims to achieve by using the best data available. However, there is a recognition within Statistics Norway of the critical importance of maintaining high levels of public trust.

18. The challenge of accessing purchase data is a setback to Statistics Norway's efforts to explore using supermarket receipts for the Household Budget Survey (HBS). Future requests would imply continuing to engage in open dialogue about data usage and conducting necessity assessments for each new planned data collection, as well as continued consultations with the Data Protection Agency.

Reflection

19. Our data – reflection of the case studies with respect to ISI ethical principles and professional values – is presented in Annex 1. According to our judgement, 6 out of the 12 ethical ISI principles were relevant from the point of view of our cases. Considering the ISI professional values, three instances related to value “truthfulness”, two to “professionalism” and one to “respect”

The following principles were classified as truthfulness issues:

2. Clarifying Obligations and Roles

The respective obligations of employer, client, or funder and statistician regarding their roles and responsibility that might raise ethical issues should be spelled out and fully understood.

8. Maintaining Confidence in Statistics

In order to promote and preserve the confidence of the public, statisticians should ensure that they accurately and correctly describe their results, including the explanatory power of their data.

9. Exposing and Reviewing Methods and Findings

Adequate information, including open-source software, should be provided to the public to permit the methods, procedures, techniques, and findings to be assessed independently.

The following principles were classified as professionalism issues:

3. Assessing Alternatives Impartially

Available methods and procedures should be considered, and an impartial assessment provided to the employer, client, or funder of the respective merits and limitations of alternatives, along with the proposed method.

10. Communicating Ethical Principles

In collaborating with colleagues and others in the same or other disciplines, it is necessary and important to ensure that the statisticians’ ethical principles are clearly understood by all participants, and properly reflected in the inquiry.

One principle was classified under the value Respect:

12. Protecting the Interests of Subjects

Statisticians are obligated to protect subjects, individually and collectively, insofar as possible, against potentially harmful effects of participating.

20. Our first observation is, that all the issues studied relate to the conflict between expectations of the audience (people and media) and the actions of data institutions. In both cases the audience was taken by surprise. In Finland this was probably because throughout the entire process, it was, even for professional audience, unclear whether the research methodologies would be publicly available and whether it would be possible to evaluate the quality of the results. Also, it was unclear who exactly would have access to these sensitive coefficients. Both THL and especially KELA, the two institutions that would have access to coefficients, have also administrative and surveillance functions. They were, in a way, not perceived as research institutes but more as parts of state apparatus. The Norwegian case was more transparent, the approach was quite straightforward to produce official statistics; perhaps the description of the new statistics to be produced was a bit loose. Despite

6

this, the Statistical Office was perceived as part of – potentially repressive - state apparatus, not as research institute or independent statistics producer.

21. All the statisticians and researchers acted in good faith and carefully followed their ethical codes of practice. Despite this, these data- actions were considered by the audience suspect, intrusive and threatening. Even the cases that we have here classified as ethical issues, are in essence cases where the statistician or researcher has carefully followed the code, but in the eyes of the public, they were not acting ethically. We think the root cause for the distrust we have analyzed here are not the behavior of the institutions or their staff, the issue is that the ethical codes are not fit for the new data situations we face.

22. Statistical institutes and the statistical community have done a lot of work to grasp the new ethical challenges. Trust centra has been established, strategic communication on data confidentiality and data security has been enhanced and explanatory memoranda on the new data ecosystems are being produced. These actions may fail, as they only try to remedy the symptoms, not the root cause which is the fact that statistical institutions have entered to the new areas – privately held data and data services – that are not familiar to our audiences. These new types of tasks are not reflected in our ethical code, not even in the most recent one, ISI code that was revised 2023. Thus, there is no authoritative norm to support data decisions in these new settings. To update the norms would not only benefit the statistical community but all institutions and experts dealing with data and analysis.

23. The ethical codes should (at least) cover the following new situations and provide guidance on how act: - Secondary and tertiary use of data – example on secondary use is NSO using privately held or administrative data for statistics, example on tertiary use of data is researchers using micro-data provided by NSO

- Combining data from different sources – in 2024 UN Statistical Commission meeting very many participating countries indicated that they are going to base their 2030 census on administrative data. We urgently need ethical guidelines on combining administrative data - Use of micro-data (or very granular statistics) in knowledge-based decision making. Data-based decisions typically impact specific groups of individuals or even individuals. There are already now many borderline cases on ISI principle 12, protection of the subjects. One example is to determine taxable value of a dwelling using micro-data on dwelling prices and statistical modelling; many decisions relating to social benefits impact (smallish) groups of people, and statistics, even micro data, are commonly used to allocate benefits.

24. It should be noted, however, that the Finnish case is an excellent example on evidence-based decision making. As such, it would not have violated the integrity of the subjects. It would probably also fall in to the “do good” category as the application of the coefficients would result in better allocation of resources and thus better services. The statistics Norway case would also fall in the “do good” -category. To do as planned would have saved a lot of taxpayers’ money and, at the same time, improved technical quality and relevance of official statistics.

25. We must be prepared for the ethical discussion about the right way to use data gathered for statistical purposes. As the statistical offices are becoming data repositories for a vast array of different governmental registers that can be linked to each other, it is inevitable that society has other use for the data than only the production of statistics. The statistical offices could be a one stop shop to provide much needed micro data for evidence-based decision making. The data needed within government, and the laws concerning the use of data collected for statistics as well as the data protection legislation, do not always meet. There are e.g questions , concerning the need to protect statistical units in cases when the data has not been obtained by the statistical office directly from the statistical unit, but from other sources. Statistics Finland has e.g. turned down an application by the Ministry of Education for school-wise information on the student’s parents’ socioeconomic status (income, national origin, education). The ministry would have wanted datato be able to give more funding to schools by applying positive discrimination. According to current national legislation, Statistics

7

Finland cannot disclose the names of our statical units (in this case schools) so the schools needing more funding could not be identified.

Annex: Key elements of ISI code of ethics, Case studies, and Violated values

Case Finland

Case Norway Violated values

1. Pursuing Objectivity  Statisticians should pursue objectivity without fear or favour, only selecting and using methods designed to produce the best possible results.

Not an issue Not an issue None

2. Clarifying Obligations and Roles  The respective obligations of employer, client, or funder and statistician regarding their roles and responsibility that might raise ethical issues should be spelled out and fully understood.

Role of NSI not Clear, role of the research organization not clear

Use of the requested information not specified in detail.

Truthfulness - processes were not transparent.

3. Assessing Alternatives Impartially  Available methods and procedures should be considered, and an impartial assessment provided to the employer, client, or funder of the respective merits and limitations of alternatives, along with the proposed method.

No alternatives were considered to the basic idea

No alternatives to micro-linking were considered.

Professionalism – social acceptability was not considered carefully enough.

4. Conflicting Interests  Statisticians avoid assignments where they have a financial or personal conflict of interest in the outcome of the work.

Not an issue Not an issue None

5. Avoiding Preempted Outcomes  Any attempt to establish a predetermined outcome from a proposed statistical inquiry should be rejected, as should contractual conditions contingent upon such a requirement.

Not an issue Not an issue None

6. Guarding Privileged Information  Privileged information is to be kept confidential. This prohibition is not to be extended to statistical methods and procedures utilized to conduct the inquiry or produce published data

Perceived as risk by subjects?

Perceived as risk by subjects?

None

7. Exhibiting Professional  Statisticians shall seek to upgrade their professional knowledge and skills, and shall maintain awareness of technological developments, procedures, and standards which are relevant to their field, and shall encourage others to do the same.

Not an issue Not an issue None

8

8. Maintaining Confidence in Statistics  In order to promote and preserve the confidence of the public, statisticians should ensure that they accurately and correctly describe their results, including the explanatory power of their data

The calculated coefficients would not be made available to subjects, details of the models used unclear

The usage of the requested data unclear

Truthfulness – lack of transparency

9. Exposing and Reviewing Methods and Findings  Adequate information, including open source software, should be provided to the public to permit the methods, procedures, techniques, and findings to be assessed independently.

Was perceived inadequate by subjects

Was perceived inadequate by subjects

Truthfulness – lack of transparency

10. Communicating Ethical Principles  In collaborating with colleagues and others in the same or other disciplines, it is necessary and important to ensure that the statisticians’ ethical principles are clearly understood by all participants, and properly reflected in the inquiry.

Failure to communicate to subjects and stakeholders; differing professional codes between statistician and economists

Failure to communicate to subjects and stakeholders

Professionalism – social acceptability not considered carefully enough

11. Bearing Responsibility for the Integrity of the Discipline 

Not an issue Not an issue None

12. Protecting the Interests of Subjects Statisticians are obligated to protect subjects, individually and collectively, insofar as possible, against potentially harmful effects of participating

Was perceived as threat, unclear whether there is impact to the subjects

Was perceived as intrusive

Respect – the promise to “not do bad “not convincing, the promise to “do good” not convincing from the subjects’ point of view.

9

References

Bateman, Connie (2012): Professional Ethical Standards: The Journey Toward Effective Codes of Ethics Work and Quality of Life, 2012, pp 21 – 34

ECE/CES/2022/2 Core values of official sta s cs (downloaded 11.3.2024)

ECE/CES/2023/24 An ethical approach to the development of social acceptance strategies for national statistical offices by Canada, Ireland, UK and Eurostat (2023). (downloaded 11.3.2024)

IAOS (2022): IPS02.Core values of official sta s cs – What are they and how do we demonstrate them?

ISI (2023): Interna onal Sta s cal Ins tute Declara on on Professional Ethics (downloaded 11.3.2024)

Koskimäki, Timo (2018): Sinunkin elämällesi oikea hinta. Blogi Tieto & Trendit 11.4.2018.

United Na ons (2014): UN Fundamental Principles of Official Sta s cs (downloaded 11.3.2024)

Suddaby, Roy and Daniel Muzio (2015): Theore cal Perspec ve on the Professions. In: The Oxford Handbook of Professional Service Firms. Oxford University Press 2015

Sta s cs Finland (1993): Toimi oikein lastoalalla. Tilastokeskuksen amma ee nen opas. Käsikirjoja 30, Tilastokeskus, Helsinki 1993. (Guide on professional ethics in Official Sta s cs, in Finnish only)

Structure of Ethical Issues in New Data Ecosystems

Markus Sovala

Director General, Statistics Finland

UNECE Workshop on Ethics in Modern Statistical

Organizations

March 26-28, 2024

Geneva, Switzerland

Outline: Structure of Ethical Issues

•Methodology

•Data and results

•Conclusions

•Discussion

22 March, 2024 Statistics Finland | [email protected]

Methodology – study of professions

• Study of professions is an established field of research in social sciences.

• Professional ethics is an established subject in studies of professions.

• Study design:

- Two recent incidents of distrust analyzed.

- Public distrust emerges, when perceived behavior (of the professionals) does not

comply with the expectations of the audience (expected ethics).

- Research question: What were the main differences between perceived and

expected behavior.

• ISI Declaration on professional ethics used as framework to analyze the incidents

- which aspects of the ISI code emerged in incidents, and which not.

- which core values were perceived being violated.

22 March, 2024 Statistics Finland | [email protected]

Case 1: Attempted healthcare reform in Finland (… one of many attempts, this was in 2015- 2019 …)

• New regional administrative layer (welfare areas)

- To manage public health services and social care (read:

old people care).

• Allocation of central government funding

- To private and public service providers in new welfare

areas.

• Capitation reimbursement model

- An individual health risk coefficient for all citizens

- Micro-data on health records and socio-economic status.

• Public debate on the approach

- Plan not accepted by the parliament and Government

resigned, but not only for data related reasons.

22 March, 2024 Statistics Finland | [email protected]: Helsingin Sanomat

Case 2: New data sources

• New Statistics Act of Norway:

- Powers to NSI to access privately held data

• The data request:

- Detailed information on consumption, based on cash

registers, to be statistically linked with payments data

(credit cards).

- To replace Household Budget Survey and to provide

new statistics on health aspects of the diets.

• Public debate of the approach:

- (Many) Data holders rejected the request, being

worried of their own reputation.

- Data Protection Supervisor gave negative opinion.

- Plan halted, negotiations with DPA and data holders

continue.

22 March, 2024 Statistics Finland | [email protected]: E24

ISI declaration of Professional Ethics – perceived violations

• Core values:

- Respect, professionalism, truthfulness and integrity

• Ethical principles

- 1. Pursuing Objectivity

- 2. Clarifying Obligations and Roles

- 3. Assessing Alternatives Impartially

- 4. Conflicting Interests

- 5. Avoiding Preempted Outcomes

- 6. Guarding Privileged Information

- 7. Exhibiting Professional

- 8. Maintaining Confidence in Statistics

- 9. Exposing and Reviewing Methods and Findings

- 10. Communicating Ethical Principles

- 11. Bearing Responsibility for the Integrity of the

Discipline

- 12. Protecting the Interests of Subjects

• Respect:

- 12. Protecting the interest of the subject

- Perceived as a risk by the audience

• Professionalism

- 3. Assessing alternatives impartially

- 10. Communicating ethical principles

- No alternative to intrusive methodologies were

considered

- Tertiary users (in healthcare-case) failed to

understand the sensitivity of their approach

• Truthfulness

- 2. Clarifying Obligations and Roles

- 8. Maintaining Confidence in Statistics

- 9. Exposing and Reviewing Methods and

Findings

22 March, 2024 Statistics Finland | [email protected]

Conclusions

• The researchers and institutions acted by the book, no (traditional)

professional misconduct.

• The audience perceived the actions quite differently: attempts to

create a new, potentially better funding system became difficult or

impossible (Case #1). Data providers didn’t trust NSI’s good

intentions (Case #2).

• The stakeholders and supervisors perceived the actions quite

differently

- The legislators (Case #1) and Data protection supervisor (Case

#2) disagreed.

• Conclusion: The current professional code does not support

statisticians in emerging new data ecosystem roles.

22 March, 2024 Statistics Finland | [email protected]

Discussion

• Possible reason for missing guidance: The basic assumption behind current codes is that all

data is collected by NSIs, mainly by surveys. This not true any longer.

- In Finland about 99 per cent of data points used for statistics are coming from admin

registers.

- Most countries will face the same situation as the censuses and (other statistics as

well) are moving rapidly towards use of administrative data.

- It’s big question how this rich data can be used efficiently for different purposes and

what’s role of NSI in this.

• Proposal: Our profession should aim at broader code, including advices for emerging data

ecosystem roles:

- General principles.

- Guidance for practical cases:

- Secondary and tertiary use of data (use of admin data for statistics and for research).

- Combining data from different sources (both admin data and privately held data).

- Use of micro-data (or very granular statistics) in knowledge-based decision making.

22 March, 2024 Statistics Finland | [email protected]

Comparison of EGSS and structural business statistics data on measuring circular economy 9th Joint OECD-UNECE Seminar on SEEA Implementation Geneva, 18 – 20 March 2024 Nina Hiltunen, Niko Olsson, Johanna Pakarinen

Contents

• Defining circular economy • Circular economy indicators in Finland • Structural business statistics in measuring

circular economy • EGSS and circular economy • Industries and products • Review of EGSS circular economy figures • Differences between environmental goods

and services and structural business statistics data

18-20 March, 2024 Statistics Finland2

Defining circular economy business

• Circular economy is a comprehensive system change

• Defining is essential for measuring

• In Finland circular economy business is defined through activities that are pivotal from the perspective of a product or service life cycle

• Eight activities, which encompass a total of 18 indicators

18-20 March, 2024 Statistics Finland3

Circular economy indicators in Finland

• First indicator set produced in 2020 • Updated in 2022 and 2023 with added

indicators • The goal was to produce indicators that

demonstrate Finland’s progress towards the circular economy with an emphasis on a business perspective

• Indicators are primarily based on statistical data already collected for other purposes

• 18 indicators with time series from around 2010 onwards

18-20 March, 2024 Statistics Finland4

Circular economy business indicators in Finland

18-20 March, 2024 Statistics Finland5

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

-

10 000

20 000

30 000

40 000

50 000

60 000

2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

EU R

m ill

io n

Number, turnover and personnel of circular economy establishments

Number of personnel Number of establishments Turnover (EUR million)

Circular economy business indicators in Finland • Structural Business Statistics (SBS) serve as the

data source

• Circular economy is defined through industries

• The Eurostat list of industries serves as the foundation, and the category of ‘Other circular economy industries’ has been added to include industries not originally in the Eurostat list

• Industry classification helps identify which industries fall under the circular economy

• Measuring all circular economy activities across the entire economy is challenging

• The data is compiled from establishment-specific information, covering enterprises where only part of their activity aligns with the circular economy

• Recycling • 381 Waste Collection • 383 Materials recovery • 4677 Wholesale of waste and scrap • 4779 Retail sale of second-hand goods in stores

• Repair and reuse • 331 Repair of fabricated metal products, machinery and equipment • 4520 Maintenance and repair of motor vehicles • 4540 Sale, maintenance and repair of motorcycles and related parts and

accessories • 95 Repair of computers and personal and household goods

• Other circular economy industries • 771 Renting and leasing of motor vehicles • 772 Renting and leasing of personal and household goods • 773 Renting and leasing of other machinery, equipment and tangible goods

18-20 March, 2024 Statistics Finland6

SBS as a data source

Positives: •Finland has excellent data coverage in

SBS •Detailed data on an establishment level •Covers product and service value chains •Possibility for regional data

Negatives: •Industry classification a good

starting point but misses a lot of circular economy activity

18-20 March, 2024 Statistics Finland7

EGSS and circular economy • EGSS in Finland is calculated from national accounts’ economic figures

• To calculate CE from EGGS data, the following categories of Classification of Environmental Protection Activities and Expenditure (CEPA) and Classification of Resource Management Activities (CReMA) categories could be considered:

18-20 March, 2024 Statistics Finland8

•02 Wastewater management •Mostly NACE 22: products related to

wastewater management •03 Waste management

•381 Waste Collection •382 Waste treatment and disposal

•07 Protection against radiation •382 Waste treatment and disposal

CEPA

•11B Minimisation of the intake of forest resources •16 Renovation of wooden packaging, pallets,

etc. for reuse •13C Minimisation of the use of fossil energy as

raw materials •Mostly NACE 22: products such as

replacement bags for plastic bags and regenerative rubber

•14 Management of minerals •Mostly NACE 24: iron made from recycled

materials •383 Materials recovery

CReMA

Comparison of CE industries & products in EGSS and SBS

EGSS

• Recycling • 381 Waste Collection • 382 Waste treatment and disposal • 383 Materials recovery • 4677 Wholesale of waste and scrap • 4779 Retail sale of second-hand goods in stores

• Products • 08 Reuse of slag in construction activities • 13 Textile bags and sacks as replacements for plastic bags • 16 Renovation of wooden packaging, pallets, etc. for reuse • 17 Paper bags as replacements for plastic bags and paper made from

recycled materials • 22 Products for water and wastewater management and products such

as replacement bags for plastic bags and regenerative rubber • 24 Iron from recycled materials

Circular economy business indicators (SBS)

• Recycling • 381 Waste Collection • 383 Materials recovery • 4677 Wholesale of waste and scrap • 4779 Retail sale of second-hand goods in stores

• Repair and reuse • 331 Repair of fabricated metal products, machinery and equipment • 4520 Maintenance and repair of motor vehicles • 4540 Sale, maintenance and repair of motorcycles and related parts and

accessories • 95 Repair of computers and personal and household goods

• Other circular economy industries • 771 Renting and leasing of motor vehicles • 772 Renting and leasing of personal and household goods • 773 Renting and leasing of other machinery, equipment and tangible goods

18-20 March, 2024 Statistics Finland9

CE output, value added and export based on EGSS data • Figures are based on

national accounts’ economic data

• Chosen NACE industries and CEPA/CReMA categories are previously mentioned

• Circular economy figures then follow trends in the general economy

18-20 March, 2024 Statistics Finland10

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

EU R

m ill

io n

EGSS Output, value added and export of circular economy

Output Value Added Exports

NACE and CEPA/CReMA review

18-20 March, 2024 Statistics Finland11

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

EU R

m ill

io n

EGSS output of circular economy by CEPA/CReMa

02 03 07 11B 13C 14

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

EU R

m ill

io n

EGSS output of circular economy by NACE

08 13 16 17 22 23 24 38 46 47

• Largest industry by far is NACE 24 Manufacture of metals • Recycled metals • CReMA 14 consist mostly of recycled metals

• Second largest industry is NACE 38 Waste collection, treatment and materials recovery • CEPA 3 consist of only waste collection and management

NACE and CEPA/CReMA review

18-20 March, 2024 Statistics Finland12

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

16 000

18 000

20 000

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

St af

f-y ea

r

EGSS staff-year for circular economy by NACE

08 13 16 17 22 23 24 38 46 47

0

2 000

4 000

6 000

8 000

10 000

12 000

14 000

16 000

18 000

20 000

2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

St af

f-y ea

r

EGSS staff-year for circular economy by CEPA/CReMa

02 03 07 11B 13C 14

Differences between environmental goods and services and structural business statistics data EGSS data (based on national accounts data)

Share of EGSS is estimated by using expert

assessments, internal statistics, information

requests, websites

No regional data

Consistent with national accounts

SBS data (based on tax data and enterprise surveys)

Data coverage in Finland is excellent

Data compiled on LKAU level •Possibility to cover

enterprises where only part of their activity aligns with the circular economy

Regional data

18-20 March, 2024 Statistics Finland13

Thank you!

Johanna Pakarinen [email protected]

Niko Olsson [email protected]

Nina Hiltunen [email protected]

*Prepared by Anna Pärnänen & Marjut Pietiläinen.

NOTE: The designations employed in this document do not imply the expression of any opinion whatsoever

on the part of the Secretariat of the United Nations concerning the legal status of any country, territory, city

or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

United Nations Economic Commission for Europe

United Nations Children’s Fund

Expert meeting on statistics on children

Geneva, Switzerland, 4–6 March 2024

Item 8 of the provisional agenda

Statistics Finland’s progress in a holistic mapping of data and statistics on children in line with the National Child Rights Strategy

Note by Statistics Finland*

Abstract

This paper describes Statistics Finland’s task on the Measure 24 of the first National

Child Strategy, which is based on the UN Convention on the Rights of the Child and

promotes the implementation of the Convention. The Strategy was published in

2021. In accordance with the measure, Statistics Finland was tasked with producing a

comprehensive picture of the state of knowledge about child wellbeing, identifying

blind spots in the knowledge base, and making a proposal for a child data portal. The

paper gives an overview of the background and the aims of the task, the work done

so far, indicators covered and the conclusions.

The paper is a shortened version of the final report on Measure 24 of the National

Child Strategy Knowledge about children – Current status and development needs

prepared by Johanna Lahtela and Anna Pärnänen, Statistics Finland. The report is

available online: https://www.lapsenoikeudet.fi/wp-

content/uploads/2022/08/Knowledge-about-children_measure24.pdf.

I. Introduction

1. Finland’s first National Child Strategy was published in 2021. The vision of the strategy is a

child and family-friendly Finland where the rights of the child are respected. The aim is to

Working paper 24

1 March, 2024

Working paper 24

2

mainstream children’s rights and status so that children are consistently taken into

consideration in all policies and activities alongside other members of society, and that

children are informed of their rights. The strategy pays special attention to securing the

status of vulnerable children and better recognising their needs (Finnish Government, 2021).

2. In accordance with the Measure 24 of the National Child Strategy, Statistics Finland was

tasked with producing a comprehensive picture of the state of knowledge about child

wellbeing, identifying blind spots in the knowledge base, and making a proposal for a child

data portal. The aim was to combine information about children on a single website to

promote the use of the data in a way that supports children’s rights. To achieve a

comprehensive picture of the data needs and use of data, the measure was carried out in

close cooperation with stakeholders. The strong commitment of the stakeholders and the

common will was crucial for the project.

3. The work has been rewarding and very necessary. Despite previous efforts to compile child

data, no comparable overviews of the state of knowledge about children have been done

before. This overview has been necessary because information about children is scattered,

and there are many data producers. In addition, a wide range of users have different data

needs. Although there is a wealth of information available about children — even a

surprisingly large amount — there are still blind spots that need to be better addressed in the

future. This paper provides an overview of the tasks carried out under the measure 24 of the

National Child Strategy.

II. How is the wellbeing of children measured?

4. The wellbeing of children can be measured in many ways. For example, the OECD has

outlined multiple domains and possibilities for measuring child wellbeing. The OECD

divides the different dimensions of child wellbeing into three tiers: the outermost tier covers

public policies, while the inner tiers concern children’s living environment and their

activities, behaviours, and relationships. These dimensions are divided further into different

aspects, each with its own dashboard (OECD, 2021, see also the OECD Child Wellbeing

Dashboard1.) The indicators should be age-sensitive and stage-sensitive, reflect children’s

own views on wellbeing, capture inequalities, and be responsive to the needs of children

from different backgrounds, for example.

5. The different domains of child wellbeing form a multidimensional network (Figure 1).

Multiple aspects of life such as living conditions, individual experiences and social

protection affect wellbeing. In the case of children, social protection is emphasised because

the younger the child, the more they depend on the adults around them. The obligation to

protect children is also enshrined in the Convention on the Rights of the Child.

Figure 1. Measuring children’s wellbeing is complicated

1 https://www.oecd.org/els/family/child-well-being/data/dashboard/.

Working paper 24

3

6. Childhood settings and experiences have an impact throughout the individual’s life cycle.

From a societal perspective, measuring child wellbeing is also important to ensure the

functioning of society in the long term. For example, the number of adults actively engaged

in society in the future can be influenced by limiting the factors of disadvantage that

contribute to the risk of social exclusion.

7. However, children should not only be thought of as future adults. There are more than one

million children in Finland, which means that around a fifth of Finns are under the age of 18.

It is therefore important to know how children are doing right now. All children have the

right to a good life and a society that supports their growth. This also means that efforts must

be made to reduce illbeing.

8. Children’s wellbeing is embedded in their growth environment. It is influenced by the

family’s income level and family relationships, the school environment, hobbies and the

living environment. Indicators such as family wellbeing therefore play an important role in

measuring child wellbeing. Children themselves also identify the family as an important

source of their wellbeing (Poikolainen, 20142)

9. Measuring child wellbeing is challenging because it is impossible to choose indicators that

could measure the wellbeing of all children. For example, the needs of an infant are very

different from those of a teenager. This means that different indicators are needed for

children of different ages.

2 Poikolainen, J. (2014) Lasten positiivisen hyvinvoinnin tutkimus – metodologisia huomioita. Nuorisotutkimus 32:2.

Indidators should be parly age- sensitive

Data collection can leave blind spots, e.g. regarding young children and those in a particularly vulnerable position

Children-s well-being

embedded in their growth environment

Living conditions,

experiences, protection

Measuring children's well-being

The different aspects of well-being format multidimensional network

Childhood settings and

experiences have an effect

throughout the individual's life

cycle

Indicators that change

over time

Protective and risk factors Well-being and disadvantage

Children's own views on

matters important to

their well- being

Working paper 24

4

10. The perception of what is considered wellbeing is constantly changing, and it is influenced

by societal developments, political priorities, and common values and standards. As the

world changes, new aspects of wellbeing also emerge. For example, because of the changes

in the concept of family and the digital transformation, we now need indicators that did not

exist in the 1990s. The indicators should therefore change in line with changes in society. On

the other hand, permanent indicators are also necessary to monitor changes in wellbeing over

time.

11. Wellbeing indicators are divided into objective and subjective indicators. Objective

indicators measure resource-based wellbeing, while subjective indicators reflect individuals’

own perception of their wellbeing. (Haanpää, Toikka & af Ursin, 20203.) Subjective

wellbeing cannot be measured using register-based data alone, but it also requires access to

survey-based data.

12. Although we talk about measuring wellbeing, indicators often describe “illbeing”. For

example, in the National Indicators of Child Wellbeing, the indicators of “No close friends”

and “Difficulties in communicating with parents” have been selected as indicators of social

relationships. One of the reasons given for emphasising indicators of illbeing is that it is

easier to measure illbeing than wellbeing (Ministry of Education and Culture, 2011).

Indicators of illbeing can help identify the risk of social exclusion, for example.

13. However, it is also important to measure wellbeing. Indicators of wellbeing do not just

measure whether children are doing well. Good family relationships or a healthy lifestyle are

also protective factors that make it easier to cope with life’s challenges (Poikolainen, 2014).

Wellbeing indicators can thus reinforce good practice by highlighting strengths and positive

factors. Particular attention should be paid to measuring the wellbeing of vulnerable

children. The Finnish Terminology Centre defines vulnerable persons as follows: “A group

of people who, due to factors beyond their control, do not have the same opportunities as

other population groups and are therefore at risk of inequality.”

14. All children are inherently vulnerable, as they do not have the same opportunities as adults

to make decisions about their lives. However, some groups of children are more vulnerable

than others. For example, migrant children, children with disabilities, children who have

experienced violence, and children belonging to gender or sexual minorities are in a more

vulnerable position than other children.

15. Nevertheless, indicators have their limitations. A single indicator does not necessarily tell

the data seeker anything until the data are put into context and analysed. Contextualisation

may be based on time series or on comparative data. On the other hand, indicators may also

need to be supported with a broader interpretation and analysis of the phenomena. For

example, does the increase in the number of child welfare notifications reflect an increase in

general illbeing or a lower threshold for reporting issues?

16. The mapping of the knowledge about children’s wellbeing has been an important data policy

exercise. The report describes the current state of knowledge about child wellbeing and

makes suggestions for improvement. The suggestions put forward in the report will help

develop knowledge about children further.

3 Haanpää, L., Toikka, E. & af Ursin, P. (2020) Alakouluikäisten lasten moniulotteinen elämääntyytyväisyys

Suomessa. Yhteiskuntapolitiikka 85.

Working paper 24

5

III. Progress of the work

17. Measure 24 comprised three sets of tasks. The aim was first to produce a comprehensive

description of the knowledge base on the wellbeing of children and young people. The

second objective was to identify the data needs, data content and blind spots in the data.

Third, Statistics Finland was tasked with planning a data repository with stakeholders that

describes the situation of children and young people and facilitates finding and using data

and monitoring the status of children in Finland. Another task was to make a proposal for a

data portal, its implementation method, and its content, as well as for the implementation

schedule. The work was carried out between March 2022 and February 2023.

18. The main objective of the measure was to produce a comprehensive description of the

knowledge base. This was carried out by compiling all available indicators in a single

roadmap (Excel file). The preparation of the roadmap required an in-depth investigation of

different indicator websites, data sources, the indicators themselves and their production.

The end result was a roadmap that provides a comprehensive picture of what kind of data on

child wellbeing are produced in Finland, how the data are produced, and by whom. The

roadmap served as a basis for the planning of the indicator website.

19. The mapping of blind spots was carried out throughout 2022, and a workshop on data gaps

was organised for stakeholders in the spring of 2022. Statistics Finland also cooperated with

Measure 25 of the National Child Strategy, led by the Ministry of Finance. In the measure, a

survey was conducted among municipalities and hospital districts, and the questionnaire also

included questions on what kind of data gaps had been identified at regional level. Blind

spots were also mapped in other stakeholder meetings, and further observations were made

over the course of the indicator work.

20. One of the tasks of the measure was to make a proposal for a data portal, its content, and its

implementation method and schedule. The work started in the spring of 2022. The location

of the data portal was discussed with the steering group4. This involved investigating

whether Statistics Finland’s website would be a suitable location for the portal. The task was

carried out in cooperation with Statistics Finland’s ongoing website renewal project to find a

workable solution for the implementation of the child data portal.

21. In addition to the tasks described above, several awareness-raising measures were

undertaken. The launch of the measure was announced by publishing a joint news release

together with the Ministry of Social Affairs and Health. The measure was also presented in a

webinar and at various stakeholder meetings. The awareness-raising measures carried out in

the measure have been very successful. The articles, blogs and infographics have attracted a

large number of views on Statistics Finland’s website, as well as on Twitter.

22. A second workshop was organised on the outputs of the measure, findings on the current

state of the knowledge base and suggestions for improvement. The workshop was attended

4 Steering group members represented following organizations: Finnish National Agency for Education, Mannerheim

League for Child Welfare, Secretary General for the National Child Strategy / Prime Minister’s

Office, Finnish Institute for Health and Welfare, Itla Children’s Foundation, Diverse Families

Network, Central Union for Child Welfare, Statistics Finland, State Youth Council, the Ministry of

Social Affairs and Health, Social Insurance Institution of Finland and SAMS–Samarbetsförbundet

kring funktionshinder

Working paper 24

6

by a wide range of experts from different stakeholders, including data producers, data users

and data administration staff. The results of the workshop were also utilised when drawing

conclusions on the improvement of the knowledge base.

IV. Child wellbeing indicators

23. Indicators play a key role in the use of statistical data. Indicators are key figures that at best

enable broad and complex matters to be presented in a simple way. They are also needed in

target setting, monitoring, planning and decision-making, including in the monitoring of

children’s wellbeing.

24. The indicator work started with a review of the theoretical framework for the wellbeing of

children and young people. For example, the methods for the measurement of child

wellbeing and the concept of wellbeing as a whole were outlined based on the OECD report

entitled “Measuring What Matters for Child Wellbeing and Policies”. The early stages of the

work involved getting acquainted with previous indicator work carried out at national and

international level.

25. In addition, different sources of indicators were mapped. The initial mapping already

showed that the data volume was likely to be large, as more than 40 initial sources of data

were identified.Due to the large amount of data, it was deemed necessary at an early stage to

establish a reference model for the classification of the wellbeing indicators. The reference

model served as the basis for a knowledge base that consists of child wellbeing indicators.

26. The reference model was based on earlier models of wellbeing, but the aim was to keep the

different domains of wellbeing manageable. It was therefore decided that the model should

have a total of eight wellbeing domains (Figure 2) including: health and wellbeing, hobbies

and leisure, social relationships, inclusion and participation, school and early childhood

education and care, housing and living conditions, safety, and services, benefits and social

support. The ninth domain, demographic indicators, describes the demographic structure of

the child population.

Figure 2 Reference model for the domains of wellbeing

Demographic indicators

•Children in the population

•Children and immigration

Health and wellbeing

•Physical health

•Mental health

•Lifestyle

•Functional capcity

Hobbies and leisure

•Hobbies

•Leisure

•Housework

Social relationships

•Family and relatives

•Friends

•School community

Inclusion and participation

•participation

•Participation at school

•Societal trust

Working paper 24

7

27. The next step was to start the actual compilation of indicators in an Excel file to serve as a

roadmap. Initially, the mapping covered all indicators describing children and young people

aged 0–29. However, it soon became clear that the volume of data would be considerably

larger than anticipated and could become unmanageable. It was therefore decided to limit the

indicators to children aged 0–17. Second, it was decided to focus on nationally produced

data because combining internationally coordinated data resources with other child data is

challenging. Third, the indicators selected for the roadmap had to be based on data that were

regularly produced for a time series to be available. This restriction excluded individual and

one-off studies from the mapping. Fourth, the indicators were limited to those directly

related to child wellbeing. This meant that the object of measurement of the indicator had to

be either the child or the family of the child. As a result, indicators such as the cost of

various services or measures taken by municipalities and schools were excluded.

28. Even after these restrictions, the volume of data describing child wellbeing was huge:

overall, around 2,400 indicators were compiled for the roadmap. When looking at the

indicators as a whole, it is important to note that the indicators do not form an immutable

database. New background variables may be added to the register data, or the data content of

the surveys may change. Sometimes even the data producer can change, as has been the case

with the Child Victim Survey and statistics on early childhood education and care, for

example. The list of indicators therefore continuously evolves over time, at least to some

extent.

29. Finally, the indicators were further categorized into smaller sets within the different domains

of the reference model. This enabled more detailed examination of the data content, blind

spots, and data overlaps in the domains to be carried out. For each domain, the distribution

of the indicators by age group was also examined. In addition, the main sources of data were

identified for each domain, as well as weaknesses in the knowledge base.

30. The classification stage revealed the limitations of the earlier “siloed” reference model: One

indicator can belong to more than one wellbeing domain at the same time. For example,

bullying at school can belong to the school and early childhood education and care domain,

but it can also belong to the safety domain. The roadmap also indicates all alternative

wellbeing domains to which the indicator could belong. The roadmap was also colour-coded

to indicate whether the indicator described wellbeing/protective factor or ill being/risk

factor. This provided an overall picture of the content of the indicators.

31. The age range of the children in question is also indicated in the indicator roadmap. The age

data are divided roughly into three groups: 0–6, 7–12 or 13–17. The comment field of each

indicator provides detailed information about whether the data are available by age group or

broken down by another age-based classification, for example. The age breakdown can

sometimes create obstacles for using the data. For example, in many registers, the data are

available for young people aged 15–19 or 15–24. In this case, it would be difficult to use the

School and ECEC

•Participation in education

•Participation in ECEC

•Enjoying school or ECEC

•Learning

Housing and living

•Housing

•Income level

•Employment

•Material standard of living

Safety

•Violence and crime

•Sexual violence and harassment

•Safety in intimate relationships and at home

•Bullying

•Accidents

Services, benefits and social support

•Social services

•Healthcare services

•Social benefits

•Other social support

Working paper 24

8

data to describe child wellbeing. Especially individuals who are closer to the end of the age

range of 15–24 are at a very different stage in their lives than the minors in the group.

32. The gender variable is available for around two thirds of the indicators. The gender

breakdown does not take into account non-binary persons. The indicators measuring only

one gender describe birth and abortion among adolescents.

33. The background variables of language, citizenship or socioeconomic status are only

available for a few indicators. The availability depends on the topic and background data.

For example, the language variable is more often available for indicators related to education

than for other topics.

34. The largest producers of indicators are THL (approx. 850 indicators), Statistics Finland

(approx. 800 indicators), various recurrent surveys of universities (approx. 450 indicators),

and Kela (approx. 200 indicators). Other data producers include various authorities and

organisations.

35. Surveys are used to collect information about the subjective experience of wellbeing.

Subjective wellbeing indicators measure life satisfaction, exercise habits, experiences of

violence and social relationships, for example. Slightly more than half the indicators in our

dataset are based on only 15 surveys. Eight of these surveys focus solely on children and

young people, while the remaining seven are either population-wide surveys or surveys that

ask adults questions about their children or family. Surveys mainly collect information about

a specific topic such as hobbies or experiences of violence. An exception to this is the

School Health Promotion Study, which produces extensive data on several wellbeing

domains.

36. The survey-based indicators are updated less frequently than the register-based indicators.

Surveys focusing solely on children are carried out every two years at most. The longest

update interval is ten years (see e.g. Statistics Finland’s Time Use Survey). Due to lack of

permanent funding, the continuity of surveys is also less certain than in the case of register-

based statistics. The challenge with survey-based data is that they inevitably have blind spots

because some children are unable to answer the surveys themselves due to their age, literacy

or other reasons.

37. Register-based indicators describe the use of services, becoming a victim of crime or living

conditions, for example. The indicators compiled in this measure are based on more than 60

different registers. Some aspects of wellbeing, such as hobbies and leisure or inclusion and

participation, are difficult to measure using register-based data alone.

38. In Finland, register-based resources are extensive and often cover almost all children,

regardless of their age or background. Most registers exclude only children who do not

reside permanently in Finland. These include asylum seekers and undocumented children.

39. In addition to their good coverage, the advantage of using register-based indicators is that

they are regularly updated, typically once a year. Some indicators are updated several times

a year, even monthly. These include Kela’s benefit data.

40. However, not all register-based data available have been compiled into indicators. For

example, a lot of wellbeing data are collected at maternity and child health clinics and in

school healthcare that are currently unavailable for research purposes. There is also a lack of

information about children of prisoners or children who have run away from substitute care,

for example. It is likely that the data exist somewhere in the customer files of the prison

administration or the police, but they have not been compiled into statistics. The reason for

this may be the lack of harmonisation of recording practices and information systems, which

Working paper 24

9

makes it difficult to compile statistics, or simply that the register data are not publicly

available.

41. The following section describes the existing child wellbeing knowledge base by domain.

The figures present the distribution of the indicators by age group and background data. The

description of each domain starts with an overview of the type of indicators included in the

domain. This is followed by a description of the main data sources and a few examples of

the indicators. The purpose of the example indicators is to give an idea of the different types

of indicators and background data available. Finally, the domain’s strengths, weaknesses and

data gaps are summarised.

42. The following restrictions were followed when compiling the indicators: the indicator must

describe children aged 0–17, the indicator must describe either a child or a family with

children, and the indicator must be produced on a regular basis.

V. Blind spots in knowledge about children

43. One of the tasks of the measure was to identify the blind spots in knowledge about children.

This was done in several ways during the work. In May 2022, a workshop on data gaps was

organised. The participants included people who used child data in their work and

representatives of various children’s organisations.

44. Measure 24 of the National Child Strategy collaborated with Measure 25, led by the

Ministry of Finance. The aim of Measure 25 was to create models to promote the

implementation of child-oriented budgeting and the monitoring of outcome data in

municipalities and wellbeing services counties. The survey of regional operators carried out

under Measure 25 also included questions related to the work in Measure 24.

45. The questions concerned the identified gaps in knowledge about children and young people,

and what kind of data is needed. Forty responses were received in the survey. Information

about data gaps was also obtained in various stakeholder meetings. For example, the

implementers of the measure met with representatives of the Sámi Parliament, participated in

the work of the working group on the knowledge base on violence against children and met

with various actors involved in child wellbeing projects.

46. Data gaps were also identified during the compilation of the indicators and the preparation of

the roadmap.

VI. Challenges in the current state of knowledge

47. The mapping work highlighted many challenges concerning the current state of the

knowledge base. The challenges be summarised in eight themes. The first theme relates to

the extent of the knowledge base. The large number of child wellbeing indicators, more than

2,400, shows that a wealth of information is available about children. However, this

information is very scattered in multiple places, which means it is difficult to use and makes

it difficult to create an overall picture of the state of children’s wellbeing. Information about

children is scattered because there are so many data producers. The main data producers are

Working paper 24

10

THL, Statistics Finland, universities and Kela. Other data producers include research

institutes, higher education institutions, and the central government and local

administrations.

48. Another finding was that although a lot of information is available, there are clear gaps in the

knowledge base. The lack of information is clearer among certain groups of children such as

immigrant or disabled children, children under school age, or sexual and gender minorities.

49. The third point related to the coordination of data production. There is no single body that

coordinates the production of data on children. This lack of coordination is reflected in many

ways in the state of the knowledge base. On the one hand, a lack of coordination leads to

data gaps, where no particular body is responsible for satisfying specific data needs. On the

other hand, it can lead to overlapping data. Different surveys ask the same questions in

slightly different words, thus creating an unnecessary burden on respondents. Due to the

long-term nature of the work, sufficient resources should be allocated to coordination.

50. Fourth, there is limited cooperation between data producers. This lack of cooperation leads

to a situation where some data producers compete for the same data resources. This

especially applies to schools and data collections organised in schools. According to reports,

schools find implementing various surveys somewhat burdensome in their daily school life

because the surveys always involve informing parents and dealing with data protection and

authorisation procedures.

51. The fifth point is related to the continuity of data, which is not always guaranteed, and

depends on the resources available. In Finland, there are currently only a few subjective

child wellbeing indicators whose continuation is secured with funding.

52. The sixth key aspect describing the knowledge base on children is the poor combinability of

the data. Because the information is scattered in multiple places, it is difficult to combine

different data sources. When one data producer has access to variables that measure

subjective wellbeing and another to variables that expand the possibilities of using the data,

but these data are not combined, the possibilities of using the data become limited.

53. Furthermore, the needs for regional-level data are currently insufficiently met. Such data

would enable the monitoring of the development of child wellbeing, the use of services and

resources, and costs at the regional level. In addition, comparative regional-level data is

needed to identify good practices better within the reference group and to exchange

experiences between regions.

54. In Finland, a lot of information about children’s wellbeing is collected at maternity and child

health clinics and in school healthcare. This information is used in healthcare for monitoring

child wellbeing. However, it is not used at the national level to create an overall picture of

the state of children’s wellbeing. The information collected at maternity and child health

clinics and in school healthcare forms an untapped data resource, the use of which would

significantly improve the state of the knowledge base.

55. Some of the above aspects of the state of the knowledge base have already been discussed

previously, for example, in the report on the National Indicators of Child Wellbeing (2011).

However, many of the issues that were raised at the time have not been followed up. There is

a lot of use for information about children. Monitoring children’s wellbeing is important

because childhood experiences are reflected well into adulthood. Many resources are also

invested in children through education and healthcare. The improvement of the knowledge

base is therefore of paramount importance, and efforts should be made to continue this work

in the future.

Working paper 24

11

VII. Proposals for improving the knowledge

56. The implementation of Measure 24 – the overall description of the knowledge base, the

identification of blind spots and the designing of the data portal – has highlighted clear needs

for development. The state of the knowledge base on children could be improved though

several measures described below.

57. The production of child data needs to be coordinated. Finland should have a body that

regularly monitors the state of the knowledge base on children. The coordinator’s role

should also include ensuring that the knowledge base is improved so that blind spots are

covered, and data needs are met. This would be done in close cooperation with the different

data producers in a designated coordination group. The data coordination should also aim to

investigate opportunities for closer cooperation in the collection of survey-based data to

avoid overlapping data collection, to reduce the burden on respondents, to minimize

competition for data resources, and to free resources for data analysis.

58. Opportunities and barriers to using previously untapped data for secondary purposes should

also be investigated. Untapped data refers to the main potential sources of data for secondary

purposes, such as data collected at maternity and child health clinics and in school

healthcare. These data are collected primarily for monitoring children’s health and for use by

healthcare professionals. This would be a reliable source of data because the data are

collected by professionals, and sufficient guidance could also ensure consistency in data

registration. The data collected at maternity and child health clinics and in school healthcare

would be valuable because data are collected on the entire age group in principle. This

would also allow better monitoring of the state of wellbeing of vulnerable children, while

respecting data protection requirements.

59. In terms of the improvement of the knowledge base as a whole, the main thing would be to

compile knowledge about children as comprehensively as possible in a single data resource

to create a child data repository. If all data were combined in a single data repository, the

accessibility of the data would significantly improve. In principle, the data repository would

contain comprehensive background data. The possibility to combine register-based and

survey-based data would also enable more detailed analyses of children’s wellbeing.

Building a data repository would also improve the production of regional-level data on child

wellbeing, as comprehensive regional data would be available as background variables.

60. Regarding register-based data, efforts should be made to harmonise the definition and

measurement methods of the key wellbeing indicators. This would support the regions in

monitoring child wellbeing, allocating resources and adopting good practice, both at the

municipal level and at the level of wellbeing services counties.

61. Since easy access to information is essential, it was proposed that a data portal for child

wellbeing indicators be built that is linked to the child data repository. One of the

bottlenecks of data portals is the updating of data, which often must be done manually. A

data portal connected to a data repository could be automatically updated.

62. Ideally, the data resources of the different data producers would be combined in a shared

data repository. A child data portal could then be built on the data repository. The data portal

would be linked to an indicator service, which would enable the publication of the child data

portal (Child wellbeing indicators). The data repository would enable more effective use of

the data to support policymaking and to provide researchers with customised datasets.

Working paper 24

12

63. Qualitative data and links to individual reports could also be added to the portal. This would

further strengthen the knowledge base on children and make it easier to find information

about vulnerable children, for example.

64. Overall, the improvement of the knowledge base is a process which should start with the

designation of the coordinating body and the establishment of the coordination group. To be

successful, the improvement of the knowledge base on children requires extensive

cooperation between experts, research institutions, ministries, and agencies. The work

depends on good and innovative cooperation between the different data producers to achieve

a workable outcome.

VIII. Indicator website

65. One of the tasks of Measure 24 was to outline a child data portal, its implementation method,

and its content, as well as the implementation schedule. In practice, this meant considering

where the indicator website for data on child wellbeing should be located, and what technical

solutions should be used.

66. The implementation of Measure and the design of Statistics Finland’s indicator website

coincided. Statistics Finland is in the process of renewing its website. One of the tasks on the

agenda in the autumn of 2022 was to design a website for the production and publication of

indicators. Combining the two projects had both beneficial and limiting effects on the

website redesign and the National Child Strategy measure. The projects were combined so

that the sets of indicators identified in the National Child Strategy measure could be used to

pilot the redesign team’s user interface design. The ready-made indicator sets made the work

of the user experience designer easier, as the sets could be used in the design work to help

identify which indicator sets would be easy to find and access and would be interesting and

necessary from the end users’ perspective.

67. The design of the user interfaces of the indicator service and the underlying technical

solutions were based on feasibility. The designs are therefore based on existing open

database interfaces. Without a clear link to a specific site and its constraints, the interface

design of the indicator site could remain detached, and its feasibility could not be ensured.

The aim was that the interface design would be realistic and feasible to implement.

68. Statistics Finland’s own site constraints determine the functionalities and visual look of the

website. The design in the pilot phase was therefore primarily based on the child data in the

databases of Statistics Finland. The starting point for the implementation is to use data

available in databases and the possibility to use interfaces, which allows the automation of

the data content update process. Automatic updating of the data content via open interfaces

would ensure the continuity of the site, as sites based on manual updating often fail due to a

lack of resources. In principle, the site will therefore not support the importing of data in

Excel format, for example. The inclusion of child wellbeing indicators other than those

produced by Statistics Finland would require the use of shared databases and interfaces.

69. More detailed specifications determining the use of interfaces and the requirements for the

data provided through them should be developed in a follow-up project. Combining data

from other data producer organisations on the website will be resolved later if the

construction of the indicator website for child data is to be pursued. However, this work will

require additional resources. A precise schedule for creating a service that includes all the

Working paper 24

13

key indicators of child wellbeing cannot yet be determined, as the work would first require

the harmonisation of data and implementation of shared interfaces.

70. Regarding the objectives of Measure 24, it has already been possible to think about the

technical solutions for the presentation of the data. The preliminary design of the indicator

website is based on the idea of displaying 20–30 key indicators per wellbeing domain. The

view may contain key figures, graphs and/or tables. The indicators for the pilot may be

selected from those that are available in Statistics Finland’s existing database tables.

Database tables can include many different indicators with different background variables.

The aim is to make data available on the portal at multiple levels. For example, the user

could check only the key indicators but could easily find more detailed data by background

variable if necessary. The visual design is still at the conceptual stage.

IX. Conclusions

71. The report presented the main outputs of Measure 24 of the National Child Strategy. More

than 2,400 child wellbeing indicators were identified. Despite the abundance of information,

there are blind spots in the knowledge base on children. They concern vulnerable children

and children under school age, in particular. In the measure, the following proposals for

action to improve the knowledge base on children were made.

72. The production of child data should be coordinated.

73. Cooperation on the collection of survey-based data should be enhanced.

74. Opportunities and barriers to using untapped data for secondary purposes should be

identified.

75. Regarding register-based data, efforts should be made to harmonise the definition and

measurement methods of the key wellbeing indicators at the regional level.

76. A child data repository should be created.

77. A data portal for child wellbeing indicators that is linked to the child data repository should

be constructed.

78. Data gaps and development needs have already been highlighted previously in various

contexts, and individual projects have been carried out to develop knowledge about children.

Many actors have their own aspirations and goals in this area. However, child wellbeing is a

very broad subject area, both in terms of data content and how the data are produced. A

comprehensive change cannot be achieved through individual efforts, but close cooperation

between different actors is needed to achieve the objectives. The first step should be to bring

together experts on child data to set common objectives and measures to achieve them.

79. Inclusion and cooperation should also be extended to children. Currently, the knowledge

base on child wellbeing does not sufficiently take into account children’s own perspective on

their wellbeing. Additionally, not all children are able to report on their wellbeing in surveys.

The right of the child to be heard should also apply to information about children.

The National Child Strategy and child data in Finland Marjut Pietiläinen, 6 March, 2024

UNECE Expert meeting on statistics on children

TransMonEE network session

This presentation is based on the report “Knowledge about Children”

prepared by Anna Pärnänen and Johanna Lahtela, Statistics Finland.

https://www.lapsenoikeudet.fi/wp-content/uploads/2022/08/Knowledge-about-

children_measure24.pdf

The 1st Finnish National Child Strategy

• The goal is a society that respects the rights of children.

• Is based on the UN Convention on the Rights of the Child, which has been in force

at the level of an act since 1991 in Finland.

• Promotes the implementation of the Convention on the Rights of the Child.

• Records the current state of the wellbeing and rights of children and young people

as well as the key objectives and measures to promote them.

• The implementation plan for the Child Strategy set out 30 measures. The measure

24 concerned producing a comprehensive knowledge base for monitoring the well-

being of children and young people.

• https://stm.fi/en/child-strategy

6 March, 2024 Statistics Finland2

Statistics Finland’s task on child strategy

• The work was carried out between March 2022 and February 2023.

• Created a framework to help categorize the indicators

• Collected all indicators on children’s well-being into one road map (excel) to help plan the web page

• Made a proposal for a web page on indicators on children, its content, the place and execution

• Mapped information gaps

• Took a closer look at the needs of information on vulnerable groups

• Made conclusions on the state of the child data in Finland & suggested how to develop it in the future

6 March, 2024 Statistics Finland3

Co-operation methods

6 March, 2024 Statistics Finland4

Project team

Steering group

Strong commitment

Workshops

Synergies with other measures of the National Child Strategy

Presentations and sharing the information

6 March, 2024 Statistics Finland5

Demographic indicators

Children in the

population

Children and

immigration

Health & Wellbeing

Physical & mental health

Lifestyle

Functional capacity

Hobbies & Leisure

Hobbies

Leisure

Housework

Social relationships

Family and relatives

Friends

Social community

Inclusion and participation

Political & organizational

activity

Participation at school

Societal trust

School & Early childhood

education and care (ECEC)

Participation in education & ECEC

Enjoying school or ECEC

Learning

Housing & Living

conditions

Housing

Income level

Employment

Material standard of

living

Safety

Violence and crime

Sexual violence and harassment

Safety in intimate relationships and

at home

Bullying

Accidents

Services, benefits and

other support of society

Social services

Healthcare services

Social benefits

Other social support

The indicator framework

The indicators

• Nationally produced indicators with time series

• Total ~ 2 400 indicators

• ~ 850 are produced by The Finnish Institute for Health and Welfare

• 800 by Statistics Finland

• ~ 200 by the Social Insurance Institution of Finland

• Some indicators fall on more than one sector in the framework

• Indicators that were left out

• Indicators which don’t directly measure the wellbeing of the child (e.g. costs of

services)

• Are one-time-only indicators

6 March, 2024 Statistics Finland6

7Statistics Finland6 March, 2024

Education & Early

childhood education and

care

347 indicators

Hobbies & Leisure

244 indicators

Source of data

Survey Register

Source of data

Survey Register

0

100

200

0-6y. 7-12y. 13-17y.

Housing & Living

conditions

154 indicators

Source of data

Survey Register

0

200

400

0-6y 7-12y 13-17v.

0

50

100

0-6v. 7-12v. 13-17v.

8Statistics Finland6 March, 2024

Health and Wellbeing

336 indicators

Services, benefits and

other support of society

467 indicators

Source of data

Survey Register

Safety

608 indicators

Source of data

Survey Register

Source of data

Survey Register

0

100

200

300

0-6y. 7-12y. 13-17y.

0

200

400

0-6y. 7-12y. 13-17y.

0

500

1000

0-6y 7-12y 13-17v.

9Statistics Finland6 March, 2024

Social connections

70 indicators

Population statistics

168 indicators

Source of data

Survey Register

Source of data

Register

Participation

72 indicators

0

50

100

0-6y. 7-12y. 13-17y.

Source of data

Survey Register

32

34

36

38

0-6y. 7-12y. 13-17y.

Conclusions about the state of the child data in Finland

• There is an enormous amount of data on children, but it is dispersed in different places.

• There are a great number of data producers, but data production is not coordinated by anyone.

• Lack of coordination causes both overlaps and gaps of information (e.g. small children and

vulnerable groups).

• Dispersed information leads to weak combinability of the data.

• There is limited collaboration among data producers and in some cases competition for the same

informant resources.

• Data continuity is only seldomly ensured.

• There is a lot of untapped information collected in child health care centers and in school health

care.

• Needs for data on a region level are not always met & there is a lack of region level follow-up

indicators.

6 March, 2024 Statistics Finland10

Proposals to develop the child data base

• Production of child data needs a coordinator & co-operation group.

• The possibilities for co-operation in survey data production should be investigated.

• The possibilities and restrictions to use untapped data should be investigated.

• For register data, the most important indicators for regional use should be selected in co- operation with the data users & common practices agreed on.

• All data should be gathered in the same data base to create a common data source.

• The child data portal could be built on the child data base.

6 March, 2024 Statistics Finland11

Thank you!

More information:

marjut.pietilainen(at)stat.fi

anna.parnanen(at)stat.fi

https://childstrategy.fi/publications/researches-and-reports/

List of tables

Table1

Table2

Table 2a

Table 2b

Table 3

Table 4

Table 5

Table 5a