Decision Supporting Procedure for Strategic Planning: DEA Implementation for Regional Economy Efficiency Estimation Karine Mesropyan1 1 Institute of Socio-Economic and Humanitarian Researches of Southern Scientific Center of RAS, 41 Chekhova prospect, 344006 Rostov-on-Don, Russia karineemesropyan@gmail.com Abstract. The algorithm of decision supporting procedure based on Data En- velopment Analysis (DEA) along with the Malmquist Productivity Index is suggested in the paper. The procedure’s core consists of evaluations complex for preliminary data processing and adjustment as well as creating of analytic materials in the field of regional strategy planning. The crucial study issue is to define boundaries of DEA applicability in this field and to eliminate DEA shortcoming, such as the scores dependence on a set of inputs and outputs. The efficiency scores of Russian regional agrarian sector are obtained in order to verify the procedure and add knowledge to current indicators’ systems of re- gional economic efficiency by improving approach objectiveness. It is shown how obtained results can be applied in the strategic planning to increase effec- tiveness of state regional policy activity. Keywords. Efficiency, Malmquist Index, Data Envelopment Analysis, Region, Procedure, Strategy Planning Key terms. DecisionMaking, MathematicalModel, Methodology, Develop- ment, Management 1 Introduction Theoretical model of this study is based on the Pareto-Koopmans concept (1951) [1]. System technology is efficient by Pareto-Koopmans if and only if the object does not have an opportunity to improve its resource (input) or product (output) without sacri- ficing some other input or output. Charnes et al. (1978) have proposed Data Envel- opment Analysis (DEA) based on this concept of efficiency that was combined opera- tional research tools within works of Koopmans (1951) and Farrell (1957) [2, 3]. DEA is a non-parametric frontier approach for comparative efficiency measure- ment in which a set of similar objects with multiple inputs and outputs is analyzed. 386 K. Mesropyan The aim of this study is to suggest the procedure for providing strategy planning by analytical reports based on DEA scores implementation. It is obvious that productivity analysis by DEA has at least three current issues. The first is to define a set of objects which will be compared in the study. The second is to formulate convenient conditions for concrete models’ modifications using. The third issue is to improve discrimination capability. Therefore, efficiency assessment procedure by DEA is primarily based on the following grounds: the formation of ob- jects’ set to be compared, identification of inputs and outputs, and model selection. Taking into account issues mentioned above, it is necessary to adapt basic DEA models and its implementation. Furthermore, DEA procedure is considered as a core of the evaluation of regional economy efficiency scores. This paper consists of five parts. We state the main issues in this, first, part. This study’s background is presented in the second part. Part 3 deals to description of sug- gested evaluation procedure. The applying of investigation of the Russian regions agrarian sector to management tasks by using the procedure is reported in the part 4. We make conclusions in the last part. 2 Theoretical and Methodological Background DEA application has a big number of advantages. First of all, a calculation of an inte- grated assessment is produced for each region reflecting the efficiency of input factors using for output products. Besides, the Pareto-optimal set of efficient regions in the multidimensional space of inputs and outputs is being obtained. Secondly, it is unnec- essary to attract an expert knowledge in a priori assignment of weights for variables corresponding to inputs and outputs. Despite of this, using of additional data on re- gion external factors is helpful for creating the right model. Thirdly, it is very impor- tant that there are no restrictions on the functional form of the relation between inputs and outputs. The study is carrying out with the hypothesis that DEA implementation needs the formal procedure in order to obtain stable scores and apply research results to analytic background of current regional strategic planning. The multilateral and penetrating analysis of DEA possibilities and its application’s restrictions are presented in Dyson et al. (2001), Cook and Seiford (2009) [4, 5]. Along with these works there are reviews of this method application, for instance, in papers of Avkiran and Parker (2010), Liu et al. (2012) [6, 7]. The common bases pro- ductivity measurement presented in Caves et al. (1982) [8]. The application possibilities of Malmquist Productivity Index in different in- tertemporal comparisons are described in the research of Färe and Grosskopf (1996) [9]. Tsuneyoshi et al. (2012) used Malmquist Index for the comparative analysis of 97 countries calculated by DEA models for period 1981-2004 [10]. Yamamura and Shin (2008) determined the nature of inequality impact on capital accumulation and growth performance by evaluation DEA indexes from 1965 to 1990 [11]. According to review presented in [12], although there are a DEA advantages, the general method’s shortcoming is considered as crucial because the scores signifi- Decision Supporting Procedure for Strategic Planning … 387 cantly depend on a set of inputs and outputs. This study suggests the special proce- dure for DEA implementation for the needs of regional strategic planning. It is a re- sult of attempting to eliminate the mentioned DEA drawback and provide the decision process of strategic planning by analytic materials. Golany and Roll (1989), Em- rouznejad and De Witte (2010) offered procedures of DEA application which are very useful for common case [13, 14]. This study based on results of these works. 3 Efficiency Estimation Procedure Different levels of the regional economy scale and the return to scale effect are con- sidered as a reason of inequality between regional output performances. That is why the model with variable return to scale is suggested for this study. This model was introduced by Banker et al. (1984). Data for a research by DEA is presented by a number of indicators in form of the matrix of inputs Xt={xtij} and matrix of outputs Yt={ytkj}. The efficiency criterion for a multidimensional assessment of an object is to assign some input and output parame- ters for all objects, some weights and then to calculate and maximize the ratio for each object: (1) where: j – index of the estimated production facility, j=1…, n; xij – matrix of input parameters that reflect the system resources, i=1,…,m; ykj – matrix of outputs which reflect the products of system, k=1,…,s; ui /wk – weights for outputs/inputs. According to the DEA framework, this function should be maximized under re- strictions for all objects: (2) The Malmquist Productivity Index is calculated using such DEA efficiency scores for evaluation of total factor productivity change: (3) The suggested procedure for regional efficiency assessment has the complex of procedures for preliminary data processing and adjustment (fig.1). 388 K. Mesropyan Fig. 1. Five stages of Efficiency Estimation Procedure The dual linear program model for evaluation the criterion given above is: (4) where: η – comparative efficiency score of region j (j=1…, n); λj – dual model variables. If η<1 then the region belongs to inefficient set, otherwise (η = 1) it is a part of Pareto set. According to procedure carrying out, the result of all evaluations using (3), (4) is a set of regional types by dynamics character. Generally, the result of DEA application is the set of scores also which shows the ways of comparative efficiency improvement for each inefficient region. Neverthe- less, this issue is not treated in this procedure because it requires the special attention and investigation due to its complexity. 4 Evaluation by Using Procedure We examined the issue of inequality of regional economy performance for the period from 2008 to 2010. Federal State Statistics Data is used from www.gks.ru [15]. We Decision Supporting Procedure for Strategic Planning … 389 evolved a set of indicators that can be used in a broader context in order to identify factors influencing on a regional underdevelopment. Stage 1. The set of indicators for models consists of resources and results of re- gional agrarian sector performance. Indicators are reported in Table 1. Stage 2. Set of model’s variables defined this way: five resources are taken as in- puts while three results are taken as outputs. The volume of region population is con- sidered as the special variable for the set’s normalization. Table 1. The set of regional performance indicators Type Indicator Resources number of cattle, thousand heads (x1) organizations acreage under crops, ha (x2) average number of employees, thousand people (x3) power capacity, thousand horsepower (x4) equipment park (tractors), units (x5) Results gross grain yield, thousand tons (y1) production of milk, thousand tons (y2) production of livestock and poultry, thousand tons (y3) Variable for Nor- malization volume of region population, thousand people Stage 3. Next, the homogeneity of conditions was checked for all agrarian regional systems, and asymmetry of land’s quality founded out. The input called “organiza- tions acreage under crops” is adjusted by the coefficient of cadastral value of agricul- tural land. The rule of ratio of variables’ number and objects’ number is kept, there- fore, modeling is made for 53 quite similar agrarian Russian regions. Additional restriction to weights is used in order to improve the discrimination ca- pability of the model. Direct restriction on the ratio of the quantity of employees and the power capacity presented by the following ratios: (5) (6) (7) where: 390 K. Mesropyan price_x3 – regions’ average monthly salary; price_x4 – regions’ average energy price; x3, x4 – inputs which are taken in the normalized forms according to the second stage. Stage 4. According to (3) and (4), the calculation cycle is done, and obtained scores are insensitive regarding the model parameters changes. It was approved by decreasing of the set of analyzed objects. Besides, the Malmquist Indices values are similar to current expert opinion on the character of current tendencies of technologi- cal progress changes in the industry for analyzed period. Stage 5. The quantitative scores combined with qualitative evaluation of risks and conditions of regional development allow finding out the regions taxonomy by using obtained knowledge on type of efficiency dynamics. The procedure has conducted from the first to the fifth stage given in Fig.1. The most significant agrarian regions of Russia are located on the Southern terri- tory which consists of two state districts, namely Southern Federal District and North- ern Caucasus Federal District. There are two strategies for these regions: Southern Federal District Strategy and Northern Caucasus Federal District Strategy [16, 17]. Obtained results can be part of analytic reports of these policy development docu- ments (tables 2-3). Development scores are presented for period 2008 - 2010 in the tables. The indicator is equal to «+»/«–» in the case of positive/negative dynamics. Table 2. Efficiency Scores for the Southern Federal District Strategy Region Development Region Score Type Republic of Adygeya ++ Stable Growth Republic of Kalmykiya ++ Stable Growth Krasnodar Region –+ Unstable Decline Astrakhan Region +– Unstable Growth Volgograd Region ++ Stable Growth Rostov Region +– Unstable Growth Table 3. Efficiency Scores for the Northern Caucasus Federal District Strategy Region Development Region Score Type Daghestan –– Stable Decline The Ingush Republic ++ Stable Growth Republic of Kabardino-Balkariya –– Stable Decline Republic of Karachaevo-Cherkesiya –– Stable Decline Republic of Northern Osetia Alaniya –+ Unstable Decline The Chechen Republic –– Stable Decline Stavropol Region ++ Stable Growth Decision Supporting Procedure for Strategic Planning … 391 Although analyzed period covers the crisis years, the agrarian production of the South of Russia shows the reserve of stability. Besides, it is brought out that the Southern regions belong to Pareto-Efficient set of Russian regions. Thus, only 4 regions among 13 of the South are estimated as having the stable de- cline. The economic development opportunities of this regions are significant, never- theless the considerable potential of regions is not using. Such indicators’ further analysis can be used for adjustment of scenario data tasks in the field of regional development foresight and strategic planning. The obtained results also can be suitable for equalization policy design in order to steady the level of regional efficiency during long-term period. In addition to this, it is important that risks, conditions and possible consequences of the policy should be assumed for each scenario of regional development. 5 Conclusions The verification of suggested procedure along with the DEA model demonstrates the positive results that approve the possibility of the procedure application to prospective studies in the field of production analysis as well as strategic management. As it was shown, the obtained results can be part of the quantitative investigations for the current strategies of development policy. In addition to this, the development scores can be used together with regional development risks and opportunities analy- sis, indicators of economical efficiency, such as gross domestic product per capita, enterprises profitability, etc. Thus, obtained scores will add knowledge to current indicators’ systems of regional economic efficiency and improve approach objective- ness and effectiveness of state regional policy activity. This article is conducted within Program of the Presidium of Russian Academy of Sciences № 32 “Fundamental Issues of Polyethnic Region Modernization in Terms of Tensions Growth”. References 1. Koopmans, T. C.: An Analysis of Production as an Efficient Combination of Activities. Activity Analysis of Production and Allocation. Cowless Comission for Research in Eco- nomics. Monograph No. 13, New York: Wiley, pp. 15–32 (1951) 2. Farrell, M. J.: The Measurement of Productive Efficiency. J. of the Royal Statistical Soci- ety, Series A (General), Part III. 120, 253–281 (1957) 3. Charnes, A., Cooper, W. W., Rhodes, E.: Measuring the Efficiency of Decision Making Units. European J. of Operational Research, 2, 429-444 (1978) 4. Cook, W. D., Seiford, L. M.: Data Envelopment Analysis (DEA) – Thirty Years On. Euro- pean J. of Operational Research, 192, 1–17 (2009) 5. Dyson, R. G., Allen, R., Camanho A. S., Podinovski V. V., Sarrico C. S., Shale E. A.: Pit- falls and Protocols in DEA. European J. of Operational Research, 132, 245–259 (2001) 6. Avkiran, N. K., Parker, B. R.: Pushing the DEA Research Envelope. Socio-Economic Planning Sciences, 44, 1–7 (2010) 392 K. Mesropyan 7. Liu, J. S., Lu, L. Y. Y., Lu, W.-M., Lin, B. J .Y.: Data Envelopment Analysis 1978–2010: A Citation-Based Literature Survey. Omega (2012) 8. Caves, D. W., Christensen, L.R., Diewert, W.E.: The Economic Theory of Index Numbers and the Measurement of Inputs, Outputs and Productivity. Econometrica, 50(6), 1393– 1414 (1982) 9. Färe, R., Grosskopf, S.: Intertemporal Production Frontiers: With Dynamic DEA. Kluwer Academic, Boston, MA (1996) 10. Tsuneyoshi, T., Hashimoto, A., Haneda, S.: Quantitative Evaluation of Nation Stability. J. of Policy Modeling, 34, 132–154 (2012) 11. Yamamura, E., Shin, I.: Effects of Income Inequality on Growth through Efficiency Im- provement and Capital Accumulation. MPRA Paper No. 10220, http://mpra.ub.uni- muenchen.de/10220/ (2008) 12. Mesropyan, K., Goryushina, E.: Economic Heterogeneity and Political Instability: Experi- ence and Prospects for Cross-Country Comparisons. The Region Economy: Problems, Findings, Prospects, Issue 13. OON RAN, ISERH SSC RAS, Volgograd, 58–66 (2012) (in Russian) 13. Emrouznejad, A., De Witte, K.: COOPER-Framework: A Unified Process for Non- Parametric Projects. European J. of Operational Research, 207(3), 1573–1586 (2010) 14. Golany, B., Roll, Y.: An Application Procedure for DEA. Omega, 1(3), 237–250 (1989) 15. Federal State Statistics Data, www.gks.ru (in Russian) 16. Southern Federal District Strategy, http://www.minregion.ru (in Russian) 17. Northern Caucasus Federal District Strategy, http://www.minregion.ru (in Russian)