Presentation is loading. Please wait.

Presentation is loading. Please wait.

SILC implementing regulation - flags

Similar presentations


Presentation on theme: "SILC implementing regulation - flags"— Presentation transcript:

1 SILC implementing regulation - flags
24-25 September Task-force On the revision of the EU-SILC legal basis

2 On the revision of the EU-SILC legal basis
Current situation Relatively complex rules Inconsistencies in the use of flags New need – information on the data source NSIs opinion divided between keeping the system and simplifying it Changing rules requires some initial investment only in the first year 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

3 On the revision of the EU-SILC legal basis
Progress October 2017 – first proposal to the TF June 2018 – presentation to WG and written consultation Summer 2018: Eurostat + volunteer countries (CZ, SE, SI, FI) September 2018: presentation of 3 scenarios to the TF – internal consistency November 2018: final choice of scenario Changing rules requires some initial investment only in the first year 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

4 On the revision of the EU-SILC legal basis
Scenarios Scenario A Based on current use Small simplifications/consistency imporvements Scenario B New information: source of data Simplification concerning gross/net values Scenario C New information: main source of data Simplifications concerning gross/net data 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

5 Questions for all scenarios
Keep or drop the imputation factor? If keep, how to calculate it? Concatenate flags or have several variables? Treat other monetary/numeric variables similarly to income variables? Is there a need for the 0 = no income flag? 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

6 Scenario A: non-income variables _F, 1 digit
-1 missing value -3 non-selected respondent -7 not applicable because the question was not used that year Other negative values to be defined according to the question 1 value is filled and no other information is necessary/possible Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

7 Scenario A: income variables _F, 1st digit – reason for missing
-1 missing value -3 non-selected respondent -7 not applicable because the question was not used that year Other negative values to be defined according to the question no income 1 value is filled and no other information is necessary/possible Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

8 Scenario A: income variables _F, 2nd digit – collected net/gross
1 Net of tax on income at source and social contributions 2 Net of tax on income at source 3 Net of social contributions 4 Gross 5 Unknown 6 Mix Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

9 Scenario A: net income variables _F, 3rd digit – recorded net/gross
1 Net of tax on income at source and social contributions 2 Net of tax on income at source 3 Net of social contributions 4 Unknown 5 Mix 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

10 Scenario A: income variables _I, 1st digit – imputation method
no imputation 2 deductive imputation 3 statistical imputation 4 gross/net conversion 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

11 On the revision of the EU-SILC legal basis
Scenario A: income variables _I, 2nd digit and next – imputation factor Imputation factor = collected value / recorded value (in percent with no decimal) No imputation = 100 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

12 On the revision of the EU-SILC legal basis
Comments on scenario A 0 = no income is a value itself and should have flags concerning the source Imputation method is relevant for other monetary variables too 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

13 Scenario B: non-income variables _F, 1 digit
Filled from survey/interview with respondent 2 Filled from survey/interview with proxy 3 Filled from administrative sources 4 Filled by construction 5 Filled by imputation -1 missing value -3 non-selected respondent -7 not applicable because the question was not used that year Other negative values to be defined according to the question 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

14 Scenario B: income variables _F, 1st digit
Filled from survey/interview with respondent 2 Filled from survey/interview with proxy 3 Filled from administrative sources 4 Filled by construction 5 Filled by imputation -1 missing value -3 non-selected respondent -7 not applicable because the question was not used that year Other negative values to be defined according to the question Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

15 Scenario B: income variables _F, 2nd digit – collected net/gross
1 Net of tax on income at source and social contributions 2 Net of tax on income at source 3 Net of social contributions 4 Gross 5 Unknown 6 Mix Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

16 Scenario B: income variables _I, 1st digit – imputation method
no imputation 2 gross/net conversion 3 deductive imputation 4 statistical imputation 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

17 On the revision of the EU-SILC legal basis
Scenario B: income variables _I, 2nd digit and next – imputation factor Imputation factor = collected value / recorded value (in percent with no decimal) No imputation = 100 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

18 On the revision of the EU-SILC legal basis
Comments on scenario B Filled by construction” would not result in information about the underlying source or sources of information 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

19 Scenario C: all variables _F, 1st digit
main source is survey/interview with respondent 2 main source is survey/interview with proxy 3 main source is administrative data 4 it is not possible to establish a main source -1 missing value -3 non-selected respondent -7 not applicable because the question was not used that year Other negative values to be defined according to the question Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

20 Scenario C: all variables _F, 2nd digit
1 Filled directly from proxy interview/respondent interview/administrative source 2 Gross/net conversion 3 Filled by construction (deductive/logical imputation, excluding gross/net conversion) 4 Filled by statistical imputation 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

21 Scenario C: income variables _I, 1st digit – recorded net/gross
Net of tax on income at source and social contributions 2 Net of tax on income at source 3 Net of social contributions 4 Gross 5 Unknown 6 Mix Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

22 On the revision of the EU-SILC legal basis
Scenario C: income variables _I, 2nd digit and next – imputation factor Imputation factor = collected value / recorded value (in percent with no decimal) No imputation = 100 Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

23 Comments/questions on scenario C
_F, 1st digit – which value should be used for SR countries when SR responded; or when the household respondent answered May be more burdensome, especially the 1st time main source - implementation is very difficult and complex in terms of data processing and programming, and not applicable to all situations (derived variables might not always be sums of individual components) Proxy flag- it is not possible to control for each variable who actually answered to the question Task-force On the revision of the EU-SILC legal basis 24-25 September 2018

24 Other comments/ all scenarios (1)
Researchers' needs: keep the imputation factor for modelling information originates from (i) administrative registers, (ii) the survey interview, (iii) combines both in some way or (iv) has been imputed income was collected gross or net meta-data on the source of each variable, where this is the same across observations (individuals/households) in the data for a particular country and wave/year; variations should be in the flag (this also retrospectively if source changed for previous years) Task-force On the revision of the EU-SILC legal basis 24-25 September

25 Other comments/ all scenarios (2)
Explain more clearly which flag to choose for skipping due to routing "mixed" flag should be possible for all flag variables/digit No reason for having the proxy flag on individual variable level Why have negative imputation flag? It will bring confusion Drop imputation flag for all variables where its use is not documented 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

26 Other comments/ all scenarios (3)
Due to resource pressure, it is better to start with current scenario and introduce gradually new flags on voluntary basis A new scenario could be combining A for non-income variables with C for income variables Better to have separate flags instead of concatenation: they are built separately and are too complex to read Dropping some digits for less important variables would be misleading 24-25 September 2018 Task-force On the revision of the EU-SILC legal basis

27 On the revision of the EU-SILC legal basis
Comments, questions… 24-25 October 2017 Task-force On the revision of the EU-SILC legal basis


Download ppt "SILC implementing regulation - flags"

Similar presentations


Ads by Google