Econometric Model for Analysing the Structural Funds Absorption at Regional Level Sectoral

ECONOMETRIC MODEL FOR ANALYSING THE STRUCTURAL FUNDS ABSORPTION AT REGIONAL LEVEL – SECTORAL OPERATIONAL PROGRAMME INCREASING ECONOMIC COMPETITIVENESS

P. Rinderu and O. Gherghinescu

University of Craiova

Abstract - An econometric model represents an important tool for simulating the principal mechanisms of economic systems. This could be applied at different scales, namely regional, national and international. When approaching this research field it should be kept in mind, permanently, that macroeconomic theory represents a dynamic environment, with a large diversity of (sub)theories, each of them claiming as being the most relevant.

There is a large variety of such econometric models, but the basic principles of conceiving them are mostly the same. The present paper proposes an ARCH like model in order to analyse the absorption of structural funds within the Sectoral Operational Programme -Increasing Economic Competitiveness Management, at regional level. There are made considerations on the convergence of the model and on the applied statistical tests. There is also emphasized the role that such a model could play in improving future programming exercises.

Key-words: Model Construction and Estimation (C51), Model Evaluation and Testing (C52), Regional Development Policy (R58)

1. Introduction

The Sectoral Operational Programme Increasing Economic Competitiveness (SOP IEC) was elaborated taking into account the Community provisions in this field, according to the Council Regulation (EC) No. 1083/2006 laying down general provisions on the European Regional Development Fund, the European Social Fund and the Cohesion Fund, the Regulation (EC) No. 1081/2006 of the European Parliament and of the Council on the European Social Fund, the Commission Regulation No. 1828/2006 setting out rules for the implementation of Council Regulation (EC) No. 1083/2006 laying down general provisions on the European Regional Development Fund, the European Social Fund and the Cohesion Fund, ERDF Regulation 1080/2006, Cohesion Fund Regulation 1084/2006, EUROPEAN Grouping of Territorial Cooperation 1082/2006.

The SOP IEC is one of the seven instruments (Operational Programmes - OPs), under the Convergence objective, for achieving the priorities of the National Strategic Reference Framework (NSRF) derived from the National Development Plan 2007 – 2013 (NDP), which aims to strengthen the strategic focus of the Economic and Social Cohesion policies across Romania, and to make the correct and appropriate linkages to the European policies and the Lisbon Strategy for growth and job creation.

SOP IEC directly addresses the first priority of NDP “Increase of economic competitiveness and development of knowledge-based economy” and the second priority of NSRF i.e. “Increasing the Long Term Competitiveness of the Romanian Economy” and contributes, to different extents, to the implementation of all NSRF priorities.

SOP IEC was elaborated under the coordination of the Managing Authority for SOP IEC - Ministry of Economy and Finance (MEF), and is the result of the partnership consultations both with the strategic partners (ACIS within MEF, other MAs–ministries coordinating other OPs, institutions designated as Intermediate Bodies, other line ministries, and agencies, social partners, civil society organizations, potential beneficiaries, other stakeholders involved in this field).

The implementation of the programme is under the responsibility of the Managing Authority for SOP IEC within MEF. In order to carry out the programme efficiently, the Directorate for SF Management within the Ministry for SMEs, Trade, Tourism and Liberal Professions (former National Agency for SMEs and Cooperatives), Ministry of Education, Research and Youth – National Authority for Scientific Research, Ministry of Communications and Information Technology, and Ministry of Economy and Finance-Energy Policy General Directorate were designated as Intermediate Bodies (IBs) for SOP IEC.

The starting point for SOP IEC is the analysis of the current situation of entrepreneurship and innovation, with special emphasis on the small and medium-sized enterprises sector (SMEs), on resources for RDI sphere, on ICT sector, and on energy efficiency and environment protection issues in the energy and industry sectors.

It is followed by the SWOT analysis, on which the development strategy is built. The SOP IEC also contains a description of the priority axes, key areas of intervention and proposed operations, as well as financial tables, implementation provisions, partnership arrangements.

The general objective of SOP is the increase of Romanian companies’ productivity, in compliance with the principle of sustainable development, and reducing the disparities compared to the average productivity of EU. The target is an average annual growth of GDP per employed person by about 5.5%. This will allow Romania to reach approx. 55% of the EU average productivity by 2015.

The specific objectives are:

• Consolidation and environment-friendly development of the Romanian productive sector
• Establishment of a favourable environment for sustainable enterprises’ development
• Increase of the R&D capacity, stimulation of the cooperation between RDI institutionsand enterprises, and increase of enterprises’ access to RDI
• Valorisation of the ICT potential and its application in the public (administration) andprivate sector (enterprises, citizens)
• Increased energy efficiency and sustainable development of the energy sector

Taking into account both the identified possibilities for improvement of the competitive position of Romanian enterprises to cope with the challenge and to be able to use the opportunities arising from operating on the European Single Market and the areas eligible for the ERDF support, the following Priority axes have been identified in the SOP IEC:

• Priority Axis 1: An innovative and eco-efficient productive system
• Priority Axis 2: Research, Technological Development and Innovation for competitiveness
• Priority Axis 3: ICT for private and public sectors
• Priority Axis 4: Increasing energy efficiency and security of supply, in the context of combating climate change
• Priority Axis 5: Technical Assistance

Technical Assistance (TA) will assist in the implementation and monitoring of the programme. The priority axes of SOP IEC are in full compliance with the lines of action of the Commission’sproposal regarding the framework for Competitiveness and Innovation 2007-2013, and take intoaccount the guidelines put forward by the EU Council for the cohesion policy for 2007-2013.

The ERDF contribution to SOP IEC budget for the 2007-2013 programming period is 2,554 million Euro, which represents 13.3% of the Community contribution to the NSRF.

The basic idea of the present model, due to the lack of consistent time-series for the structural funds absorption process, is to use a specific model with a mix input. This mix input takes into consideration data related to the pre-accession period and to the first monitoring exercise of structural funds absorption. Under this approach the time interval for the combined process raises from 3 to 10 years.

2. Research Methodology and Paper Review

Autoregressive Conditional Heteroskedasticity (ARCH) models are specifically designed to model and forecast conditional variances. The variance of the dependent variable is mod- eled as a function of past values of the dependent variable and independent, or exogenous variables.

ARCH models were introduced by Engle (1982) and generalized as GARCH (Generalized ARCH) by Bollerslev (1986). These models are widely used in various branches of econo- metrics, especially in financial time series analysis. See Bollerslev, Chou, and Kroner (1992) and Bollerslev, Engle, and Nelson (1994) for recent surveys.

In order to perform the Analysis we will use such a statistical model applied to the structural model presented in Fig. 1.

The following set of variables has been considered:

POS_CCE_P_xx- time serie with payments in SOP-HRD

POS_CCE_V_xx- time serie with contracted amounts in SOP-HRD

INFRA_PRE_xx- time series for infrastructure pre-accession funds at regional level

HRD_PRE_xx- time series for HRD pre-accession funds at regional level

IRU (x,y)- HRD data at regional level

POP_REG_xx- population at regional level

PIB_REG_xx- gross domestic product/capita at regional level

CD_PRE_xx- time series human resources in RD sector at regional level

Fig. 1 The structural model and the position of the current analysis into it

3. Results and Conclusions

After running the model, next results have been obtained, in the case of all 8 development regions:

a) South Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:25
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(1,1))-(C(1)*LOG(POS_CCE_V_X(1,1)
*POP_REG(1,1)*PIB_REG(1,1))+C(2)*LOG(INFRA_PRE_S(-1)
*CD_PRE_S(-1)) +C(3)*LOG(HRD_PRE_S(-1)*PIB_REG(1,1)
*POP_REG(1,1)*IRU(1,1)*IRU(1,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.414975 / 0.014682 / 28.26517 / 0.0000
C(2) / -9.20E-12 / 9.06E-05 / -1.02E-07 / 1.0000
C(3) / 2.24E-11 / 0.000426 / 5.25E-08 / 1.0000
C(4) / 2.21E-24 / 8.17E-05 / 2.70E-20 / 1.0000
C(5) / 0.150000 / 4.132536 / 0.036297 / 0.9710
C(6) / 0.600000 / 36.04581 / 0.016645 / 0.9867
Akaike info criterion / -49.00384 / Sum squared resid / 2.04E-23
Schwarz criterion / -49.21208 / Log likelihood / 153.0115
Durbin-Watson stat / 3.124492

b) South-West Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:26
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(2,1))-(C(1)*LOG(POS_CCE_V_X(2,1)
*POP_REG(2,1)*PIB_REG(2,1))+C(2)*LOG(INFRA_PRE_SV(-1)
*CD_PRE_SV(-1)) +C(3)*LOG(HRD_PRE_SV(-1)*PIB_REG(2,1)
*POP_REG(2,1)*IRU(2,1)*IRU(2,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.416016 / 0.014969 / 27.79101 / 0.0000
C(2) / 2.81E-11 / 0.000205 / 1.37E-07 / 1.0000
C(3) / -2.04E-11 / 0.000444 / -4.59E-08 / 1.0000
C(4) / 2.01E-24 / 0.000153 / 1.31E-20 / 1.0000
C(5) / 0.150000 / 65.58582 / 0.002287 / 0.9982
C(6) / 0.600000 / 37.79759 / 0.015874 / 0.9873
Akaike info criterion / -49.11302 / Sum squared resid / 1.85E-23
Schwarz criterion / -49.32126 / Log likelihood / 153.3390
Durbin-Watson stat / 2.110570

c) South-East Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:27
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(3,1))-(C(1)*LOG(POS_CCE_V_X(3,1)
*POP_REG(3,1)*PIB_REG(3,1))+C(2)*LOG(INFRA_PRE_SE(-1)
*CD_PRE_SE(-1)) +C(3)*LOG(HRD_PRE_SE(-1)*PIB_REG(3,1)
*POP_REG(3,1)*IRU(3,1)*IRU(3,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.411808 / 0.000972 / 423.7230 / 0.0000
C(2) / 4.10E-13 / 3.44E-06 / 1.19E-07 / 1.0000
C(3) / -9.30E-12 / 7.23E-05 / -1.29E-07 / 1.0000
C(4) / 9.16E-25 / 4.23E-05 / 2.17E-20 / 1.0000
C(5) / 0.150000 / 2.633311 / 0.056963 / 0.9546
C(6) / 0.600000 / 3.713266 / 0.161583 / 0.8716
Akaike info criterion / -49.86358 / Sum squared resid / 8.45E-24
Schwarz criterion / -50.07182 / Log likelihood / 155.5907
Durbin-Watson stat / 1.749123

d) West Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:28
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(4,1))-(C(1)*LOG(POS_CCE_V_X(4,1)
*POP_REG(4,1)*PIB_REG(4,1))+C(2)*LOG(INFRA_PRE_V(-1)
*CD_PRE_V(-1)) +C(3)*LOG(HRD_PRE_V(-1)*PIB_REG(4,1)
*POP_REG(4,1)*IRU(4,1)*IRU(4,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.417924 / 0.008448 / 49.46793 / 0.0000
C(2) / 1.59E-12 / 3.97E-05 / 4.00E-08 / 1.0000
C(3) / -2.33E-12 / 0.000378 / -6.16E-09 / 1.0000
C(4) / 1.16E-25 / 8.17E-05 / 1.42E-21 / 1.0000
C(5) / 0.150000 / 17.55780 / 0.008543 / 0.9932
C(6) / 0.600000 / 21.53193 / 0.027866 / 0.9778
Akaike info criterion / -51.95175 / Sum squared resid / 1.07E-24
Schwarz criterion / -52.15999 / Log likelihood / 161.8552
Durbin-Watson stat / 1.993097

) Regiunea Sud_

e) North-West Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:29
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(5,1))-(C(1)*LOG(POS_CCE_V_X(5,1)
*POP_REG(5,1)*PIB_REG(5,1))+C(2)*LOG(INFRA_PRE_NV(-1)
*CD_PRE_NV(-1)) +C(3)*LOG(HRD_PRE_NV(-1)*PIB_REG(5,1)
*POP_REG(5,1)*IRU(5,1)*IRU(5,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.413227 / 0.002588 / 159.6975 / 0.0000
C(2) / 3.37E-13 / 6.31E-05 / 5.34E-09 / 1.0000
C(3) / 5.40E-12 / 7.34E-05 / 7.36E-08 / 1.0000
C(4) / 2.66E-25 / 4.53E-05 / 5.88E-21 / 1.0000
C(5) / 0.150000 / 6.494221 / 0.023097 / 0.9816
C(6) / 0.600000 / 6.460936 / 0.092866 / 0.9260
Akaike info criterion / -51.19588 / Sum squared resid / 2.46E-24
Schwarz criterion / -51.40412 / Log likelihood / 159.5876
Durbin-Watson stat / 2.033292

f) North-East Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:31
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(6,1))-(C(1)*LOG(POS_CCE_V_X(6,1)
*POP_REG(6,1)*PIB_REG(6,1))+C(2)*LOG(INFRA_PRE_NE(-1)
*CD_PRE_NE(-1)) +C(3)*LOG(HRD_PRE_NE(-1)*PIB_REG(6,1)
*POP_REG(6,1)*IRU(6,1)*IRU(6,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.405669 / 0.001854 / 218.8113 / 0.0000
C(2) / -4.09E-12 / 3.05E-05 / -1.34E-07 / 1.0000
C(3) / 4.80E-12 / 5.85E-05 / 8.22E-08 / 1.0000
C(4) / 2.40E-25 / 1.72E-05 / 1.40E-20 / 1.0000
C(5) / 0.150000 / 4.123622 / 0.036376 / 0.9710
C(6) / 0.600000 / 5.967284 / 0.100548 / 0.9199
Akaike info criterion / -51.31174 / Sum squared resid / 2.21E-24
Schwarz criterion / -51.51998 / Log likelihood / 159.9352
Durbin-Watson stat / 2.186409

) Regiunea Cent

g) Center Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:34
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(7,1))-(C(1)*LOG(POS_CCE_V_X(7,1)
*POP_REG(7,1)*PIB_REG(7,1))+C(2)*LOG(INFRA_PRE_C(-1)
*CD_PRE_C(-1)) +C(3)*LOG(HRD_PRE_C(-1)*PIB_REG(7,1)
*POP_REG(7,1)*IRU(7,1)*IRU(7,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.400926 / 0.002701 / 148.4521 / 0.0000
C(2) / 9.57E-13 / 7.99E-05 / 1.20E-08 / 1.0000
C(3) / 0.000000 / 0.000139 / 0.000000 / 1.0000
C(4) / 2.36E-26 / 3.55E-05 / 6.65E-22 / 1.0000
C(5) / 0.150000 / 8.051192 / 0.018631 / 0.9851
C(6) / 0.600000 / 15.86596 / 0.037817 / 0.9698
Akaike info criterion / -53.54317 / Sum squared resid / 2.18E-25
Schwarz criterion / -53.75141 / Log likelihood / 166.6295
Durbin-Watson stat / 0.989340

) R
) Regiunea Ce

) Regiunea
) R
) Regiuneh) Bucharest-Ilfov Region

Dependent Variable: Implicit Equation Estimated by GMM
Method: ML - ARCH
Date: 09/11/10 Time: 21:35
Sample(adjusted): 2 7
Included observations: 6 after adjusting endpoints
Convergence achieved after 1 iterations
LOG(POS_CCE_P_X(8,1))-(C(1)*LOG(POS_CCE_V_X(8,1)
*POP_REG(8,1)*PIB_REG(8,1))+C(2)*LOG(INFRA_PRE_BIF(-1)
*CD_PRE_BIF(-1)) +C(3)*LOG(HRD_PRE_BIF(-1)*PIB_REG(8,1)
*POP_REG(8,1)*IRU(8,1)*IRU(8,9)))
Coefficient / Std. Error / z-Statistic / Prob.
C(1) / 0.403396 / 0.000969 / 416.3396 / 0.0000
C(2) / -3.86E-13 / 4.30E-05 / -8.97E-09 / 1.0000
C(3) / 8.22E-13 / 2.67E-05 / 3.08E-08 / 1.0000
C(4) / 4.05E-26 / 1.96E-05 / 2.06E-21 / 1.0000
C(5) / 0.150000 / 15.65835 / 0.009580 / 0.9924
C(6) / 0.600000 / 7.139587 / 0.084038 / 0.9330
Akaike info criterion / -53.12744 / Sum squared resid / 3.74E-25
Schwarz criterion / -53.33568 / Log likelihood / 165.3823
Durbin-Watson stat / 1.211053

After running the models for each region, the following set of conclusions has been depicted:

• Due to the differences in magnitude order of several variables it was considered a logarithmic scale in order to facilitate the convergence process. A very peculiar task was to slightly modify the values of time-series in cases when the same value for two consecutive years appeared, hence to eliminate the overflow errors.
• All models converge, but present a quite high degree of volatility. This is explained both by the limited number of observations and by the impossibility of modelling some external factors (e.g. political factors, economic crisis, external loans etc.).
• All applied statistical tests (Akaike, Schwarz, Durbin-Watson) and the corresponding correlograms present normal values and shapes.
• It is very much sensitive to asses the quality of the absorption process at regional level. However, as an example, if using the Akaike criterion, it ranges between -49.00 (South Region) down to -53.54 (Center Region). A ranking, under these assumptions, in terms of efficiency of absorption the funds via SOP IEC, is: Region S-SW-SE-NW-NE-W-BIF-C.
• The model might be used for future analyses concerning the absorption of structural funds in Romania.
• The model could be refined by introducing supplementary variables and could be also serve as a powerful instrument in developing future strategies for absorbing the structural funds in Romania, to have better programming exercises in the future.

Acknowledgment: This work was supported by the strategic grant POSDRU/89/1.5/S/61968, Project ID61968 (2009), co-financed by the European Social Fundwithin the Sectoral Operational Programme Human Resources Development 2007 – 2013.

References

1. *** Institutul Naţional de Statistică, Anuarele 2000-2010
2. Bollerslev, T. (1986) Generalized Autoregressive Conditional Heteroskedasticity, Journal of Econometrics, 31, 307–327.
3. Engle, R.F. (1982) Autoregressive Conditional Heteroskedasticity with Estimates of the Variance of U.K. Inflation,” Econometrica, 50, 987–1008.
4. Bollerslev T., Chou Y. R, Kroner F. K., (1992) Arch Modeling in Finance: AReview of the Theory and Empirical Evidence, Editura Journal of Econometrics, 52, 5-59.
5. Bradley John, Janusz Zaleski, (2003) Modelling EU accession and Structural Fund impacts using the new Polish HERMIN model - Report.
6. Capello, R. (2007) Regional Economics, Editura Routledge, Marea Britanie.
7. Engle, R.F., Lilien, M., and Robins, P. (1987) Estimating Time Varying Risk Premia in the Term Structure: The ARCH-M Model, Econometrica, 55, 391–407.
8. Gherghinescu, O., Rinderu, P. (2010), Economic model for analysing the structural funds absorption at regional level – the regional operational programme,”Finance – Challenges of the future” Journal, 12 , 147-158
9. Gherghinescu, O. (2010), Econometric model for analysing the structural fundsabsorption at regional level – sectoral operationalprogramme human resources development, Annales Universitatis Apulensis Series Oeconomica, 12(2), 550-562
10. Gherghinescu, O., Iova, C., Rinderu, P., (2009) Financing human resources development at regional level. Comparison between HRD pre-accession funds and European Social Fund absorption, Metalurgia International, Vol.XIV, Special issue no.13, ISSN: 1582-2214
11. Gherghinescu, O., Rinderu, P., Spulbar, C. (2009) Analysis of structural and cohesion funds absorption in Romania, ”Finance – Challenges of the future” Journal, 9, 223-230