Presentation

advertisement
Monthly Estimates of GDP from
Quarterly National Accounts and
Social Security Register
Javier Aramburu, Haritz Olaeta and Marta Salvador
Helsinki, 4th May 2010
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
1. Introduction
Purpose of QNA
 Provide coherent picture of economic situation
 Basis for business cycle analisys and modelling purposes
 Be more timely than Annual National Accounts
Objection
 Demand of prompt information has increased enormously
 QNA is available within 6 to 9 weeks after the quarter due to
the lack of quarterly indicators
Purpose of this work
Obtain monthly estimations of GDP from administrative
sources (Social Security Register) consistent with QNA
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
2. Social Security data in Spain
The Social Security System includes:
- persons who carry out professional activity
- persons who fulfil non-contributive modalities requirements
Person
with work
activity
Affiliation
number
Recorded in
Soc. Sec.
Register
Evolution
of
affiliates
Evolution of
economical
activity
Evolution of
GDP
2. Social Security data in Spain
Monthly available information
Monthly
means
2-3 days
Basque
Country
provinces
NACE Rev.2
aggregations
Last-dayof-month
data
NACE Rev.2 incorporates significant changes from NACE Rev.1
which has caused additional difficulty when estimating GDP by
kind of activity, HERE NOT CONSIDERED
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
3. Relationship between affiliation - GDP
1.000.000
980.000
960.000
940.000
920.000
900.000
880.000
Gross data
08
-I
08
-II
08
-II
I
08
-IV
09
-I
09
-II
09
-II
I
09
-IV
07
-I
07
-II
07
-II
I
07
-IV
06
-I
06
-II
06
-II
I
06
-IV
05
-I
05
-II
05
-II
I
05
-IV
04
-I
04
-II
04
-II
I
04
-IV
860.000
Seasonally adjusted data
Figure 1. Affiliates. Gross and Seasonally adjusted data.
Clear seasonal component in the data
USE SEASONALLY ADJUSTED DATA
To analyse correlation between social security data and GDP:
6,0
5,0
4,0
3,0
2,0
1,0
0,0
-1,0
2005- 2005- 2005- 2005- 2006- 2006- 2006- 2006- 2007- 2007- 2007- 2007- 2008- 2008- 2008- 2008- 2009- 2009- 2009- 2009I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
I
II
III
IV
-2,0
-3,0
-4,0
-5,0
GDP yty
SS gross yty
SS seasonally adjusted yty
Figure 2. GDP. Gross and Seasonally Adjusted Affiliates. Year-to-year growth
R2 = 0.97
high correlation between GDP and Social Security data
To analyse the residuals:
600.000
500.000
400.000
300.000
200.000
100.000
0
-100.000
04-I 04-II 04- 04- 05-I 05-II 05- 05- 06-I 06-II 06III
IV
III
IV
III
06- 07-I 07-II 07IV
III
07- 08-I 08-II 08IV
III
08- 09-I 09-II 09IV
III
09IV
-200.000
Figure 3. Residuals
worst adjustment in 2009
probably due to the need of introducing other relevant indicators
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
4. Affiliation series to estimate GDP
Estimation of GDP
ECOTRIM
univariate methods of time
disaggregations
Basque QNA
(EUSTAT)
quarterly
monthly
Chow and Lin (1971)
Fernández (1981)
Litterman (1983)
Due to lack of overall
cointegration
affiliation - GDP
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
5. Preliminary results
2,0
1,0
0,0
-1,0
08-VII
08-VIII
08-IX
08-X
08-XI
08-XII
09-I
09-II
09-III
-2,0
-3,0
-4,0
Monthly estimation
Quarterly estimation
EUSTAT QNA
2,0
1,0
0,0
-1,0
08-X
08-XI
08-XII
09-I
09-II
09-III
09-IV
09-V
09-VI
-2,0
-3,0
-4,0
Figure 4. Monthly disaggregated GDP and quarterly GDP. Year-to-year growth
5. Preliminary results
-1,0
-2,0
-3,0
-4,0
-5,0
09-I
09-II
09-III
09-IV
09-V
09-VI
09-VII
09-VIII
09-IX
-6,0
Mo nthly estimatio n
Quarterly estimatio n
EUSTAT QNA
-1,0
-2,0
-3,0
-4,0
-5,0
09-IV
09-V
09-VI
09-VII
09-VIII
09-IX
09-X
09-XI
09-XII
-6,0
Figure 5. Monthly disaggregated GDP and quarterly GDP. Year-to-year growth
5. Preliminary results
Table 1. Quarterly GDP. Year-to-year and quarter-to-quarter growth
2009-I
GDP FROM
AFFILIATION
15.658.696
T-1
T-4
-0,7
-1,2
GDP
EUSTAT-QNA
15.462.462
2009-II
15.398.731
-0,4
-3,1
15.322.562
-0,7
-3,7
0,3
0,6
2009-III
15.281.079
-0,3
-4,2
15.309.650
-0,1
-4,0
-0,2
-0,2
2009-IV
15.333.916
0,2
-2,6
15.347.648
0,3
-2,5
-0,2
-0,1
T-1
T-4
DIF (T-1)
DIF (T-4)
-2,0
-2,5
1,2
1,2
 Errors observed in third and fourth quarters appear to be practically
insignificant (0.2 and 0.1 absolute error respectively).
 First and second quarters show significant deviations (1.2 and 0.6
respectively in y-t-y growth; 1.2 and 0.3 in q-t-q growth rates).
 Structural change shown between the Affiliates and observed quarterly
GDP since late 2008, more remarkably in early 2009.
Outline
1. Introduction
2. Social Security data in Spain
3. Relationship between affiliation data and quarterly GDP
4. Use of affiliation series to estimate monthly GDP
5. Preliminary results
6. Conclusions and future work
6. Conclusions and future work
Conclusions
 This results are encouraging as THE MONTHLY ESTIMATE
SEEMS TO MAKE SENSE. The quarterly flash estimates
(obtained summing up the monthly estimates) have shown
to be, quite a good predictions of the real quarterly GDP.
 The economic crisis has had an impact on the relationship
between these two series. Therefore, is DESIRABLE TO
INCLUDE MORE AVAILABLE INFORMATION to get an
indicator with a higher cointegration with quarterly GDP.
 The DATA used is AVAILABLE THE FIRST WEEK AFTER
THE REFERENCE QUARTER. This means an advance of 5
to 8 weeks compared to the GDP provided by QNA.
6. Conclusions and future work
Future work
The Social Security Register offers Affiliates data of the Basque
Country at PROVINCE level and for different ECONOMIC
SECTORS. Therefore, the next extensions of this work will be:
 Obtention of monthly ESTIMATES OF THE COMPONENTS OF
GDP from the production approach
 Disaggregation of the monthly GDP for each of the Basque
PROVINCES
Thank you !!!
Eskerrik asko !!!
Kiitoksia !!!
Download