Bio-based and Applied Economics 7(1): 19-38, 2018 ISSN 2280-6180 (print) © Firenze University Press ISSN 2280-6172 (online) www.fupress.com/bae Full Research Article DOI: 10.13128/BAE-24046 SAM multipliers and subsystems: structural analysis of the Basilicata’s agri-food sector Mauro Viccaro1,*, Benedetto rocchi2, Mario cozzi1, SeVerino roMano1 1 School of Agricultural, Forestry, Food and Environmental Sciences, University of Basilicata, Potenza, Italy. E-mail: mauro.viccaro@unibas.it; mario.cozzi@unibas.it; severino.romano@unibas.it 2 Department of Economics and Management, University of Florence, Firenze, Italy. E-mail: benedetto.rocchi@unifi.it Date of submission: 2017 14th, February; accepted 2017 24th, November Abstract. Local agri-food products are conceived as a form of cultural capital, repre- senting potentially fruitful resources for rural development. Italy and its regions offer a rich and diverse agricultural and food heritage that has led to the creation of numer- ous quality agri-food systems. Despite their ability to absorb disturbances and main- tain their functions, it is important to develop economic models targeted to analyse the relationships among the components of food systems, in order to identify their strengths and weaknesses and drive the implementation of sectoral policies. In view of the new Rural Development Programme (2014-2020), the aim of this work is to ana- lyse the structure of the Basilicata’s agri-food system using a multi-sector model based on a two-region SAM, specifically developed for Basilicata, an Italian region charac- terised by a highly specialised agri-food sector. Results show that the availability of a highly disaggregate multi-sector model of the regional economy may be a valuable supporting tool to design regional policies for innovation and for the development of rural areas, laying the foundation for further analysis. Keywords. Rural development, agri-food systems, multi-sector model, SAM multi- pliers, subsystem approach. JEL codes. E16, R15. 1. Introduction Agri-food is considered as one of the most crucial sectors for economic growth, as it is the main source of livelihood that makes growth possible (Dethier and Effenberger, 2012; Schultz, 1964). Over the last few years, different studies have emphasised the con- tribution that agri-food can make for rural development, on the regional scale (Ilbery and Kneafsey, 2000; Kneafsey et al., 2001; Marsden et al., 2000; Murdoch et al., 2000; Parrott et al., 2002; Tregear et al., 2007). *Corresponding author: mauro.viccaro@unibas.it 20 Mauro Viccaro et alii Local agri-food products are conceived as a form of cultural capital; according to the principles of the endogenous development theory (Ray, 1998; Terluin, 2003), they repre- sent potentially fruitful resources for development, as they can incorporate and add value to many local resources with features that are peculiar to a specific area (Brunori and Ros- si, 2000; Marsden et al., 2000). This awareness, combined with the growing consumers’ demand for healthy and safe food, has induced producers to explore new ways of doing business, through initiatives that take over the idea that localised agri-food systems could provide not only economic but also social and environmental benefits, thereby combining marketing with political, socioeconomic and cultural activities that improve collective well-being (Volpentesta and Ammirato, 2012). In Basilicata region (southern Italy), agri-food makes a significant contribution to the regional economy in terms of output, employment and exports. The sector contrib- uted €771 million in value added to the Basilicata economy in 2014 (ISTAT, 2016a). While its contribution to the regional economy is small relative to other sectors, the Basilicata agri-food sector is of strategic importance for the sustainable development of rural com- munities. The spatial distribution of the production activities, as well as the presence of potential food and wine tourism products with geographical designations (Bencivenga et al., 2016), provide it with a key role in sustaining remote rural areas through the genera- tion of income and jobs. Different elements show the presence of a valuable “quality” potential in the region- al agri-food sector that may be exploited to increase the competitiveness of the regional economy. In the near future, the new 2014-2020 programming period of Rural Develop- ment Programme (RDP) will play a major role in enhancing innovation for agriculture, forestry and food industry providing financial support to business choices directed to improve economic and environmental performances, and promoting the organization of competitive food chains. The design and the implementation of regional policy, however, should be increasingly oriented by relevant knowledge on the structure of economy. There are different approaches applied in the analysis of agribusiness1. Cook and Chad- ded (2000) describe their evolution linked to the agribusiness, agri-food system and sub-sec- tor (filière) concepts. However, it is important to point out that any strategy of sectoral devel- opment should be based on top-down multi-sector approaches, which take into account the dynamics of the agri-food production activities within the wider regional economic system. By the mid-1950s, Davis and Goldberg (1976) developed the “agribusiness” concept valuating the extent and amount of agricultural and industrial relationships, by the use of the Leonetief input-output (I-O) model. Since Davis and Goldberg’s work, the input- 1 According to David and Goldberg’s definition (Davis and Goldberg, 1976), agribusiness includes three compo- nents: 1. the farm supplies aggregate including all intermediate consumptions of agriculture (backwards linkages of agriculture); 2. the farming aggregate composed by all operations of crop cultivation and livestock breeding at the farm level; 3. the processing and distribution aggregate composed by all operations of storage, process- ing and marketing of agricultural products both for food and non food uses. The processing component can be divide into two further components, the fiber processing and the food processing. This further decomposition of agribusiness allows to separate two subsystems: 1. the agri-food block, including a part of the farming aggregate (agricultural products for food production) and the food processing and marketing component ;  2. the agri- industry block composed by the remaining part of the farming aggregate (non food products) and fiber process- ing and marketing component. 21SAM multipliers and subsystems output analysis has largely been used to study the structure of an economy both at the regional and national level. With reference to Italy, among others, Chang Thing Fa (1981) and Chang Ting Fa et al. (2013) use a triangulation method of I-O tables to analyse the structural change of the agribusiness at Italian and European level, respectively. Belletti (1992), using an I-O subsystem approach, analyses the agri-food sector in more detail studying the delocalization of the agri-food supply chain in Tuscany region. However, there are some limitations in using Leotief input-output approach. In the Leotief model, “…the omission of the general equilibrium links relating output to factorial income and final consumption may be of critical relevance both in aggregate terms (lost gross output) and in the rank ordering of sectors (hierarchy shifting)” (Cardenete and Sancho, 2006: 322). These limitations can be overcome by extending the conventional I-O methodology in a Social Accounting Matrix (SAM) (Miller and Blair, 2009; Rocchi et al., 2015; Viccaro et al., 2015). In addition to the inter-industry transactions specific to input-output tables, a SAM include balanced accounts for factors, institutions (such as producers, consumers, government) and foreign sectors, closing the cycle of the income distribution and spending. There are numerous examples of the use of SAM models in the context of agricul- tural and food analysis. Caskie et al. (1999) use a regional SAM to analyse the impact (in term of output, income and employment) of a reduction in the final demand in beef on Northern Ireland’s economy. Psaltopoulos et al. (2006) evaluate the impact of CAP meas- ures on rural development in Archanes area (Crete, Greece) using an inter-regional SAM. The possibility to decompose through multiregional model the multiplier effects in inter- regional and intraregional effects is helpful to understand in depth the differences in econ- omy structure among regions, in this case between rural and urban areas. Other examples of SAM-based studies are reported in Rocchi (2009), Vega et al. (2014), Cardenete et al. (2014) e more recently Campoy-Muñoz et al. (2017). The cited work by Cardenete and colleagues shows the effectiveness of multiplier analysis by using SAM models in order to avoid the “missing linkages” typical of the Leontief models. Based on that, the objective of this work is to analyse the structure of Basilicata’s agri- food sector using a multi-sector linear model based on a two-region SAM, specially devel- oped for Basilicata. Through the SAM multiplier matrix (Miller and Blair, 2009), we will evaluate the con- tribution of each production sector in the agri-food sector to the regional economy and after, following the sub-system approach proposed by Momigliano and Siniscalco (1982), we will analyse the shares of the production sectors represented in the model that are directly and indirectly committed to satisfy the final demand towards different categories of food. In the light of results of the analysis, a set of policy implication will be discussed. In the following section the potential development linked to the “quality” of agri-food production in Basilicata is shortly discussed. Section 3 presents the SAM and describes the model used in the analysis. The main results are provided and discussed in section 4 while some final remarks are proposed to reader in section 5. 2. The “quality” potential of the agri-food sector in Basilicata Basilicata’s agri-food sector plays a major role in the regional economy, due to the sig- nificant weight of employment in agriculture and the wide range of typical and quality 22 Mauro Viccaro et alii agri-food products (eight products with a PDO/PGI), besides the number of farms and research bodies involved in the sector. The share of the agriculture’s contribution to the total value added is definitely the high- est in Italy (5.4% against about 2% at the national level) (ISTAT, 2016a). Considering Basilica- ta’s entrepreneurial community, the agriculture has the highest number of firms operating in the region (Table 1), which reflects, however, the small average size of agricultural holdings. As to the food industry, its contribution to the overall value added is lower compared to agriculture (2.3%) achieving, however, one of the highest levels on the national scale: this is due to the fact that a large part of the sector still focuses on the production and trade of low value added agri-food products. The importance of the food industry with- in the regional economy can be clearly seen when considering the manufacturing sector only: food industry ranks just after the automotive sector, both in terms of value added produced (20% vs. 35%) and number of labour units employed (just over 3,200 against about 4,500) (ISTAT, 2016a); food industry is the first in terms of number of operating firms (Table 2), which are mainly small and medium enterprises. The analysis of foreign trade enables an outline of the structure and trends of Basili- cata’s agri-food sector. The most interesting aspect is the reduction in imports recorded just after the economic crisis of 2007 till now, both for the agricultural products (-18%) and for food and beverages (-39%), combined with a significant increase in exports, equal to +24.5% for agriculture and +49% for the food sector (ISTAT, 2016b). Basilicata’s food exports in 2015 were above EUR 36 million, mostly represented by bakery products accounting for 55% of exports (about EUR 20 million), followed by vegetable fats and oils, mainly olive oil (14%) and beverages (10%), basically related to wine exports. Another important factor is the trend of food export observed over the last few years. Compared to the trend of the other manufacturing sectors (Figure 1), food is the only industry that has not been affected by the negative impact of the economic crisis of 2007, but rather continued to grow. Table 1. Number and % of Basili- cata’s operating firms by econom- ic sectors (2015). Sector n° % Agriculture 17,500 34% Construction 6,161 12% Other services 11,987 23% Trade 12,428 24% Manufacturing 3,818 7% Total 51,894 100% Source: own calculations on ISTAT data (ISTAT, 2016a). Table 2. Number and % of Basilicata’s manufacturing firms (2015). Sector n° % Food and Beverage 895 23% Textiles, wearing, leather and related products 300 8% Wood and of products of wood 553 14% Chemicals and plastic products 112 3% Non-metallic mineral products 308 8% Basic metals and metal products 778 20% Machinery and electrical equipment 266 7% Motor vehicles, trailers and semi-trailers 39 1% Others manufacturing 567 15% Total 3,818 100% Source: own calculations on ISTAT data (ISTAT, 2016a). 23SAM multipliers and subsystems These positive results seem to have been affected by the quality upgrading of export- ed products and the subsequent strengthening of those factors, such as certified quality, innovation (organic products) and originality, which constitute the established strengths of image abroad. A recent study on the commercial performance of regional agri-food sectors in Ita- ly between 1991 and 2012 confirms the good performance of Basilicata in foreign trade, mainly related to the level of specialization of its productions (Platania et al., 2015). Evidence of the main regional specializations, in economic but also cultural and social terms, derives from some experiences of specialized territorial clusters with a broad pro- duction base, such as the Distretto Agroindustriale del Vulture (6,489 businesses covering 15 municipalities), the Distretto Agroalimentare di Qualità del Metapontino (7,430 businesses covering 12 municipalities), which have been operating since 2004, and the most recent rural districts (since 2010), including the Distretto rurale di Pollino-Lagonegrese (27 munici- palities) and the Distretto delle Colline e delle Montagne Materane (19 municipalities). These clustering systems, involving the largest part of the region, focus on quality spe- cialized productions (wine, olive oil, mineral waters, dairy products, pork meat process- ing, fresh pasta, bakery products, fruit and vegetables and cereal production, preserved food, honey) and their promotion through tourism activities. Considering the potential integration that agri-food production activities have with the tourism sector, tourist flows can be an asset in the forthcoming years to promote the development of Basilicata agri-food sector, especially with reference to the opportunities offered by “Matera 2019”, event, when the town of “Sassi” will be European Capital of Cul- ture. Based on a direct interview to the agribusiness operators working in Basilicata’s agri- food districts, even though the event has not yet generated any impact on the sector, the Figure 1. Trend of export of the main sectors of Basilicata’s manufacturing industry1 (2005-2015) � �� �� �� �� ��� ��� ��� ��� ��� ��� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� � ��� �� �� � ����������� ����� � ������� ��������������������� ���������������� ����������� ������������� �������������������� �������� ������������� ���������� �������� �������� � ��������������� �­������� Source: own calculations on ISTAT data (ISTAT, 2016b). 1 The graph does not show the automotive sector, due to a purely graphical reason, given the signifi- cant amount of exports (over 2.2 billion of euro in 2015). Even this sector, however, shows a drop just after 2007, with a gradual recovery starting only from 2014. 24 Mauro Viccaro et alii shared perception of operators is that it can offer good development opportunities in the forthcoming years. A further aspect of interest that emerged from the interviews is the great attention the agribusiness operators attach to research, development, training and innovation in the sector. Indeed, based on the analysis conducted by the European Commission on the European regions’ innovation capacity in 2014, Basilicata was classified as “Moderate Innovator” (European Commission, 2016). Investments in innovation in order to improve the competitiveness of the sector (Contò et al., 2009) may be supported by the research institutions present in the region that are already operating in that direction. Other notable initiatives include the spin-offs coming out of research, recognized by the University of Basilicata since 2012 in the fol- lowing areas: environment, new agri-food products (donkey milk) and innovative services. Finally, the new 2014-2020 programming period of Rural Development Programme (RDP) represents an opportunity to increase the competitiveness of the regional economy, providing financial support to business choices directed to improve economic and envi- ronmental performances, and promoting the organization of competitive agri-food supply chains. In this context, the structural analysis of the Basilicata’s agri-food sector presented in the following paragraphs will provide a set results with interesting policy implications. 3. Materials and methods 3.1 A two-region SAM model In the present study, the structure of the regional agri-food sector has been analysed using a two-region SAM model (Basilicata vs. Rest of Italy) with a detailed disaggregation of accounts for agriculture and food industry production activities (see section 3.3 below). The SAM (Miller and Blair, 2009) is a two-entry matrix recording the flows occur- ring between all actors of an economic system, in a given place and for a given time peri- od (usually one year). Each row/column pair represents respectively the inflows and the outflows of a given account, so that by definition the matrix is balanced (the row totals must equal the column totals). A SAM may be considered as an expansion or a generali- zation of a Leontief input-output table. While in the latter, emphasis is laid on the produc- tion system, in the SAM the perspective is larger. The simultaneous representation of the accounts for production activities, production factors, institutions (households, firms, and public administration), capital formation and exchanges with the rest of the world makes it possible to follow the formation of value-added and its distribution and redistribution in the form of income to the institutions. SAMs are crucial databases for many quantitative models (e.g. SAM linear mod- els and Computable General Equilibrium models). Beside their statistical content, SAMs are a useful tool to evaluate policy interventions both at the national and regional level. Through the solution of a linear model (establishing an appropriate “closure rule”: Miller and Blair, 2009), it is possible to analyse the structural features of the economy and calcu- late the impact that exogenous changes on single components have on the whole econom- ic system. In our study, the closure rule considers as exogenous the accounts for national Government, capital formation and the rest of the world’ so that the resulting SAM multi- 25SAM multipliers and subsystems pliers take thus the value of Leontevian-Keynesian multipliers. Let us consider the matrix of SAM coefficients of a single region “r” (Miller and Blair, 2009: 515): [1] where A is the matrix of inter-industry technical coefficients, C is the matrix of endog- enous final expenditure coefficients, V is the matrix of endogenous value-added factors shares, Y is the matrix of endogenous coefficients distributing income to institutions and H is the matrix of endogenous coefficients for income re-distribution among institutions. The structure of the matrix of SAM direct coefficients of our two-region model, as in any two-region I-O model, is (Miller and Blair, 2009: 77-80): [2] where the blocks along the main diagonal account for flows within the two regions (b = Basilicata and i = rest of Italy) while the blocks along the other diagonal represent the (commodity and financial) flows between the two regions. By solving the linear system x = Sx + f (where x is the vector of totals of endogenous accounts and f is the vector of exogenous account flows) for x, we have: x = (I - S)-1f [3] where M = (I - S)-1 is the matrix of SAM multipliers. Each coefficient quantifies the total increase for each account i deriving from a unit exogenous shock on the account j. Note that since the SAM-based model endogenizes transactions that not included in the input-output interindustry models, the SAM multi- pliers will generally result larger than the input-output ones. The advantage of using a two-region disaggregation of accounts lies in the possibil- ity of considering the rest of Italy as being endogenous to the model; this makes it pos- sible to breakdown impacts on Basilicata’s economy calculating not only the total but also the intra- and interregional impacts (spillovers and feedbacks). If the calculated matrix of multipliers M enables the estimate of the total impact, the breakdown of the matrix of accounting coefficients S into intraregional and interregional elements enables to calculate the following2: Intraregional effects: [4] 2 For details of the multiplier decomposition for multi-region model see the chapter “Decompositions in an interregional context” (Miller and Blair, 2009: 286-288). 26 Mauro Viccaro et alii where ; Interregional spillover effects: Mspill = I + S* [5] where ; Interregional feedback effects: Mfeed = [I - (S*)2]-1 [6] 3.2 The sub-system approach Starting from the concept of vertically integrated sector (Pasinetti, 1973), the struc- tural analysis of Basilicata’s agri-food sector has been integrated by the sub-system I-O approach that makes it possible to study an individual sector, or group of sectors, that is considered a subsystem which interacts with the rest of the productive system (Belletti, 1992; Llop and Tol, 2013; Momigliano and Siniscalco, 1982; Montresor and Marzetti, 2010). In particular, we use the approach proposed by Momigliano and Siniscalco (1982), extending Belletti’s work (Belletti, 1992) in a two-regional model. The input-output approach is based on the representation of the interdependencies existing between different economic sectors. In fact, the level of activation of different production processes in the sectors of the economy depends not only on the final demand directed to them, but indirectly, via the circular flow of the economy, on the final demand directed towards all sectors. In the subsystem approach the production system is divided into blocks, constituted by the shares of the production sectors represented in the original matrix that are directly and indirectly committed to satisfy the final demand towards dif- ferent categories of goods. Let be A the matrix of accounting coefficients representing only the interdependencies existing among production sectors (industries) in the economy. The re-classification of economic quantities from the production sectors to the different “blocks” of the economy working to meet the final demand towards different sectors, may be carried out using the following “B operator” (Momigliano and Siniscalco, 1982: 155): [7] where the symbol ^ indicates diagonalisation. A generic element bij of the matrix B is the share of activity of the i sector triggered by the final demand directed to the j sector. Thus the sum of all rows of matrix B is 1, since the level of activation of different production sectors is completely covered by the production required by the final demand towards the whole production system. Any economic quantity may be reclassified from sectors to “blocks” by multiplication, using the operator B. For instance, if l is the vector of employment in different sectors, the matrix L: [8] 27SAM multipliers and subsystems subdivides the labour employed in different sectors among different blocks (“subsystems”) of the economy. Through matrix L it will be possible not only to assess the relative impor- tance of different subsystems in terms of employment but also to characterise the compo- sition of different subsystems in terms of “shares” of the original sectors. The same opera- tion may be carried out using the vector of value added of different sectors. The reclassification by subsystems is based only on the matrix of coefficients repre- senting the interdependencies existing between different production sectors (the A subma- trix in the right side of equation 1), implicitly using a Leontevian-type multiplier that does not consider feedbacks through consumption as in the SAM multiplier decomposition proposed in the previous section. However, the application of the sub-system approach to a two-region model, like that used in this study, makes it possible to extend the analysis to the participation of each region’s sectors in the fulfilment of the demand addressed to the production sectors of the other region. 3.3 A social accounting matrix for the agri-food system analysis The SAM used in this study is a two-region (Basilicata vs. Rest of Italy) matrix refer- ring to 2011, produced in collaboration with the Regional Institute for Economic Planning of Tuscany (IRPET, Florence) with most recent statistical records available. The structure of the matrix includes a total of 347 accounts, concerning 51 production activities, 64 goods and 3 production factors (employment and self-employment, and capital), 3 types of institutions (households, businesses, public administration) in the two regions. The household sector is subdivided by income deciles into ten groups, whereas the public admin- istration is distinguished as local and central. There are of course also the accounts entitled to the capital formation and to real and financial flows with the rest of the world. In order to analyse the structure of Basilicata’s agri-food sector, the accounts concern- ing agriculture and the food industry have been broken down in some detail for both Basilicata and the rest of Italy, using the matrix of inter-industry technical coefficients derived from the national supply-use table produced by the Dipartimento di Scienze per l’Economia e l’Impresa of the University of Florence for the year 2009 (Rocchi et al., 2016). By combining it with the official statistics made available by ISTAT for the year 2011 (val- ue added, employment, import, export), the accounts concerning the food industry activi- ties have been broken down into ten sub-sectors and relative commodities: 1. Meat 2. Fish 3. Olive oil 4. Vegetable oils, sugar, pasta 5. Vegetables and fruits 6. Dairy products 7. Cereals 8. Animal feed 9. Wine 10. Water and other beverage The agricultural sector, conversely, has been subdivided using the data of RICA (Rete di Informazione Contabile Agricola) (CREA, 2016) as well as the available data of 28 Mauro Viccaro et alii FADN (Farm Accountancy Data Network) (European Commission, 2016b). By combin- ing the two databases, the agricultural sector has been initially broken down into 8 groups of businesses by type of farming for the rest of Italy, and into 5 production activities for Basilicata. In order to ensure a greater consistency of the analysis, the Italian agriculture has been subsequently regrouped in the following 5 subsectors: 1. Cereal grains 2. Horticulture 3. Permanent crops 4. Livestock 5. Mixed Agricultural commodities are grouped under the heading of agricultural products while the final demand is represented by consumption functions (bundle of commodities classified according to the COICOP classification). Discrepancies between row and col- umn totals of accounts after disaggregation were reconciled balancing the table according the Stone-Camperhown-Meade approach (Round, 2003). 4. Results 4.1 The structure of Basilicata’s agri-food sector based on the matrix of multipliers According to the SAM, in 2011 the output value of the fifteen sectors of Basilicata’s agri-food system amounted to about EUR 1.2 billion, 7.9% of the regional total value. The agri-food share increases when considering the value added (8.3%) and, above all, employment (16.6%). High values are due, as previously mentioned, to the importance of agriculture, which is a typically a labour intensive sector, within Basilicata’s economy. Table 3 compares the output multipliers of seven macro-sectors making up Basilicata’s production system, as it is represented in the SAM. The first two rows indicate the increase in final output required to satisfy 1 EUR of additional demand addressed to each macro-sector. Since Basilicata is a region framed within a national economy, a significant share of activation is transmitted outside its regional boundaries. For example, a one-million additional demand addressed to Basili- Table 3. Output multipliers in Basilicata’s economy macro-sectors. Agriculture Other primary activities Food industry Other manufacturing Constructions Trade and services Public administration Bas’ output 2.813 2.745 3.017 2.922 3.027 2.829 2.803 RoI’ output 0.934 0.819 1.243 1.156 1.015 0.931 0.848 RoI/Bas* (%) 34.0% 32.0% 38.1% 37.5% 33.4% 33.8% 32.2% Labour** 64.3 0.8 8.7 5.7 19.3 18.4 26.8 *Bas: Basilicata; RoI: Rest of Italy. **Labour Unit for millions of euro. 29SAM multipliers and subsystems cata’s agriculture generates a 2.8 million increase in the output produced by Basilicata’s economy (mostly in the agricultural sector but also in all other sectors), and nearly one million Euros in the rest of the Italian economy. As a whole, the share of the output mul- tiplier operating outside regional boundaries is about 30 to 40%. Significantly, agriculture also shows the highest employment multiplier, whereas the food industry shows a multi- plier value that is basically in line (although slightly above) the average of the other manu- facturing activities. The analysis of multipliers is detailed in Table 4 that proposes data referred to the fif- teen sectors in which agri-food has been broken down. The table shows the results of the regional breakdown of multipliers. The three columns to the right of the total multiplier break down the multiplier effect (that is the output growth generated in addition to the initial stimulus) into three components: the regional effect, i.e. the additional output gen- erated by interdependencies (among industries and through final consumption) within the region; the interregional spillover, which is the impact transmitted outside the regional boundaries and generating an increase in the activity in different sectors in the rest of Ita- ly; and the interregional feedback, i.e. the additional increase in the regional output result- ing from the output growth in the rest of Italy. The share of each subsector within the two components of agri-food system (agri- culture and food industry) is shown in the second column. It can be noted that most of the regional agricultural output is produced by the farms “specialised” in arable crops and animal husbandry, and by the farms classified as “mixed”. To assess these data we should Table 4. Regional breakdown of output multipliers in the agri-food sectors. Macro- sectors’ output (%) Total multiplier Regional effects Interregional spillover Interregional feedback RoI/Bas* (%) Agriculture Cereal grains 13.2% 2.814 0.866 0.944 0.004 109.01% Horticulture 0.0% 2.473 0.705 0.765 0.003 108.51% Permanent crops 6.5% 3.060 1.048 1.008 0.003 96.18% Livestock 67.1% 2.641 0.744 0.893 0.004 120.03% Mixed 13.2% 3.068 0.963 1.101 0.004 114.33% Food industry Meat 6.0% 3.268 1.048 1.214 0.006 115.84% Fish 0.7% 2.774 0.864 0.907 0.003 104.98% Olive oil 6.0% 3.438 1.207 1.226 0.006 101.57% Vegetable oils, sugar, pasta 41.6% 3.154 0.967 1.182 0.005 122.23% Vegetables and fruits 5.9% 3.268 1.066 1.197 0.006 112.29% Dairy products 17.4% 3.277 0.987 1.284 0.006 130.09% Cereals 3.2% 3.331 1.115 1.210 0.006 108.52% Animal feed 1.1% 3.327 1.018 1.303 0.006 128.00% Wine 5.7% 3.311 1.038 1.268 0.005 122.16% Water and other beverage 12.5% 3.481 1.029 1.447 0.005 140.62% *Bas: Basilicata; RoI: Rest of Italy. 30 Mauro Viccaro et alii consider that the type of farming (based on which the agricultural sector is broken down) classifies farms (typically multi-product firms) based on the prevalence of certain produc- tion processes. A farm is classified as “specialised” in a given process if the latter repre- sents at least two thirds of the output value. Hence the output produced in the farms clas- sified as “livestock farms” is not consisting solely of livestock products but includes also a significant share of other products. Similarly the output produced by the other groups of farms consists for its part of a basket of goods. The output multipliers tend to be higher in the food industry than in the agricultural activities. In the first case the initial impact determined by the final demand addressed towards the sectors generates almost always a three times larger total increase of the out- put produced: in the case of olive oil and of the beverage industry, the overall growth of output is about three and half times the initial stimulus. This is a quite typical structural difference, because in the agricultural activities a lower ratio between intermediate con- sumption and output results in a lower impact on the activities supplying inputs. Considering the regional breakdown of multipliers, since Basilicata is a small region- al economy open to the rest of the Italian production system, it is not surprising that the “return” feedback towards Basilicata’s production activities is negligible. Conversely, spillovers towards the rest of Italy are significant, thereby certifying how Basilicata’s production activi- ties depend on imports from the rest of Italy. Spillovers tend to be higher for industrial activi- ties. The last column of the table sums up these results showing the percentage ratio between the share of the multiplier effect remaining in the region (regional effect plus feedback) and the spillover component. Only in the case of farms specialised in permanent crops, the addi- tional growth of the output that remains within Basilicata is higher than that “transmitted” to the rest of the Italian economy. The industrial activities most open to the rest of Italy are the dairy and the beverage industries. These are two important sectors within Basilicata’s food industry, accounting for about 30% of the output value: especially in the case of dairy indus- try, this figure could suggest interesting spaces for an additional integration with the region- al agriculture. On the other hand, the olive oil industry shows a higher integration with the regional production system, with an internal multiplier effect that is equal to the produced spillovers: although this is a minor sector in terms of output value, it proves that the commit- ment to quality can have positive impacts on the rest of the regional economy. 4.2 Sub-system analysis The structural analysis of Basilicata’s agri-food system can be enhanced through in- depth studies using the results of the subsystem-analysis. Figures 2 and 3 show the contri- bution of each subsector of Basilicata’s agri-food sector to four different subsystems of the Italian economy (Basilicata and Rest of Italy) satisfying certain “blocks” of final demand. Since the four blocks represent the total final demand in the SAM, the total of the per cent values of each row is always 100. Comparing the data classified as “sector” in the two figures, it is possible to see how industrial activities tend to meet the sector-specific demand more than agricultural activ- ities (with the only exception of those of the farms specialised in horticulture that rep- resent, however, a negligible component of Basilicata’s agriculture). The participation in the subsystems associated with the demand of the other sectors of Basilicata’s agri-food 31SAM multipliers and subsystems industry is quite variable in the case of the regional food industry, ranging from 15% of the activity of grains and starch products to 2.2% of the meat-processing industry. The value is more homogeneous in the case of agricultural sectors, which commit, on average, around 11% of their activity to the participation in other subsystems. The datum shows the degree of integration among the components of the agri-food system and could be an interesting indicator of possible areas for innovation to promote its competitiveness. Considering the participation in subsystems oriented to the final demand of the other regions (Rest of Italy), the higher percentage in agricultural subsectors may be assessed as Figure 2. Share of Basilicata’s agricultural sectors to the various “blocks” of the production area. 0% 20% 40% 60% 80% 100% Cereal grains Horticulture Permanent crops Livestock Mixed Sector Other agrifood sectors Other sectors Rest of Italy Figure 3. Share of Basilicata’s food sectors to the various “blocks” of the production area. 0% 20% 40% 60% 80% 100% Meat Fish Olive oil Vegetable oils, sugar, pasta Vegetables and fruits Dairy products Cereals Animal feed Wine Water and other beverage Sector Other agrifood sectors Other sectors Rest of Italy 32 Mauro Viccaro et alii an interesting opportunity for the economy in so far as it expresses the capacity of “attract- ing” (either directly or indirectly) a higher share of final demand towards the regional production system. However, since Basilicata is a small economy open to the rest of the national economy, it would be important to assess the stability of this participation to the subsystems of the rest of the Italian economy. Much depends on the upgrading of agricul- tural products directed towards the rest of the national production system: in the case of commodities without a specific quality differentiation (as may be the case of cereal produc- tion), they would suffer pressures from regional (or international) potential competitors. Tables 5 and 6 show the composition of the subsystems satisfying the final regional demand (broken down into macro-sectors of activity). The composition is expressed both in terms of employment and value added produced. The analysis of composition is repeat- ed using two different breakdowns of the final demand, in order to fully exploit both sec- toral and regional breakdown of the model. Interdependencies between agriculture and food industry obviously appear especial- ly in the subsystem concerning the food industry, which includes Basilicata’s agricultural activities accounting for 26.9% of employment and 11.6% of the value added produced. To satisfy the demand of food industry products, however, the relevant subsystem also activates the agriculture of the other regions (12.5% of subsystem employment and 9% of the value added), in addition to a far more important component in services’ activities. This is again an indicator of interesting areas of integration that might be boosted up in the regional system. The participation of other sectors of Basilicata’s economy in the food industry subsystem could probably be increased, resulting lower than that in other subsys- tems of regional manufacturing. The same analysis of Tables 5 and 6 is proposed again in Table 7, except that subsys- tems are referred to the final demand oriented towards the single sectors of Basilicata’s Table 5. Composition of subsystems in terms of employment: macro-sectors. Subsystems Final regional demand Agriculture Food industry Other manufacturing and constructions Trade and services Public administration Regional agriculture 92.3% 26.9% 0.4% 1.1% 0.2% Regional food industry 0.2% 30.6% 0.1% 0.4% 0.1% Rest of Italy’ agriculture 1.8% 12.5% 0.9% 1.1% 0.2% Other manufacturing and constructions 2.1% 7.2% 67.9% 5.1% 4.7% Trade and services 3.2% 21.3% 28.9% 90.9% 9.8% Public administration 0.4% 1.5% 2.0% 1.4% 85.1% Regional agri-food sector 92.5% 57.5% 0.4% 1.5% 0.2% Rest of regional economy 2.7% 13.9% 71.3% 85.4% 94.0% Rest of Italian economy 4.8% 28.6% 28.3% 13.1% 5.8% Labour (LU x 1,000) 15.2 10.7 56.6 55.1 67.1 33SAM multipliers and subsystems agri-food production activities. For the sake of convenience, the presentation of data has been transposed, with the subsystem components into columns (totals of rows equal to 100). Table 7 shows the subsystem composition in terms of regional location of activities. Table 6. Composition of subsystems in terms of value-added: macro-sectors. Subsystems Final regional demand Agriculture Food industry Other manufacturing and constructions Trade and services Public administration Regional agriculture 79.6% 11.6% 0.1% 0.3% 0.1% Regional food industry 0.5% 36.5% 0.1% 0.3% 0.1% Rest of Italy’ agriculture 3.3% 9.0% 6.0% 1.6% 0.9% Other manufacturing and constructions 6.4% 11.6% 63.6% 5.3% 5.6% Trade and services 9.7% 30.0% 29.0% 91.6% 12.1% Public administration 0.5% 1.3% 1.2% 0.8% 81.3% Regional agri-food sector 80.1% 48.0% 0.2% 0.6% 0.1% Rest of regional economy 7.4% 18.0% 67.1% 84.3% 91.3% Rest of Italian economy 12.5% 34.0% 32.7% 15.1% 8.6% Value-added (M€) 320 456 3 347 3 450 3 334 Table 7. Subsystem composition: sectors of Basilicata’s agri-food system. Regional subsystems Employment Value-added Regional agri-food sector Rest of regional economy Rest of Italian economy Regional agri-food sector Rest of regional economy Rest of Italian economy Cereal grains 97.0% 1.1% 1.9% 73.2% 9.9% 16.9% Horticulture 100.0% 0.0% 0.0% 96.2% 0.9% 2.9% Woody 96.4% 2.3% 1.3% 57.5% 24.5% 18.0% Livestock 83.5% 4.9% 11.6% 85.2% 4.8% 10.0% Mixed 82.2% 6.4% 11.4% 68.1% 11.6% 20.2% Meat 62.7% 10.2% 27.1% 51.6% 14.9% 33.5% Fish 54.9% 17.9% 27.2% 48.4% 18.6% 33.0% Olive oil 56.2% 15.7% 28.1% 32.3% 26.9% 40.8% Vegetable oils, sugar, pasta 58.6% 15.3% 26.1% 52.1% 18.1% 29.9% Vegetables and fruits 61.6% 11.0% 27.4% 48.9% 16.5% 34.6% Dairy products 60.6% 9.2% 30.2% 52.4% 12.1% 35.5% Cereals 64.3% 8.8% 26.8% 50.5% 14.5% 34.9% Animal feed 55.0% 12.1% 32.9% 44.6% 16.1% 39.2% Wine 53.7% 16.8% 29.5% 39.3% 22.6% 38.1% Water and other beverage 42.2% 20.7% 37.1%   34.3% 23.7% 42.0% 34 Mauro Viccaro et alii Except for the case of the few farms classified as specialized in horticultural produc- tion and showing a sub-system virtually “self-contained” at the regional level (notably in terms of employment generated both directly and indirectly), activities involving animal husbandry (both specialized and mixed) are, among the agricultural ones, those with the highest participation of non-regional components to the satisfaction of final demand. In the case of food industry products, the level of participation in the subsystems by non- regional production activities is quite homogenous, especially in terms of employment. The comparison of the “parallel” subsystems referred to the final demand in the two regions (Basilicata and Rest of Italy) can provide relevant additional indications. Table 8 shows the case of the two regional subsystems devoted to fulfill the final demand for accommodation and food services. These are production activities that might benefit from a regionally-based integration with the agri-food sector, in particular with a view to quali- tative differentiation and promotion of typical regional products. The level of participation of both regional agriculture and food industry is lower in the case of the Basilicata’s subsystem as compared with “the average” of the rest of Italy. This means that the final demand towards Basilicata’s activities supplying restaurant and accommodation services is less able to activate production and employment within the regional borders than those operating in the rest of Italy. Such results suggest the existence of an unexploited space for the integration of tourism activities with the local agri-food system. 5. Conclusions This study has proposed a structural analysis of Basilicata’s agri-food system, based on a two-region SAM model, appropriately broken down. The objective of the analysis was to make available helpful information for defining sectoral regional policies, associated with the implementation of the new programming period of Rural Development policies. The 2014-2020 Rural Development Programme actually offers a major opportunity to increase the regional system competitiveness, by providing public funds equal to EUR 680 million. Agri-food is an important component of Basilicata’s economy, not only in terms of value produced and employment created, but especially for its quality production and local production systems in which all steps of supply chains (agricultural, industrial, and Table 8. Final demand composition of the restaurant and accommodation services. Subsystems Employment Value-added Basilicata Rest of Italy Basilicata Rest of Italy Regional agriculture 3.0% 7.3% 1.4% 3.7% Regional food industry 1.7% 2.8% 2.3% 3.4% Other primary activities 3.8% 0.3% 3.7% 0.4% Other manufacturing and constructions 4.4% 3.3% 8.2% 5.2% Trade and services 86.3% 85.6% 83.5% 86.4% Public administration 0.8% 0.7% 0.8% 0.9% 35SAM multipliers and subsystems marketing steps) find their coordination. The structural analysis of interdependencies between the various components of the agri-food system, via the two-region model, has highlighted important areas of further integration that could drive innovation processes. To enhance the positive impact of agri-food production activities on the regional eco- nomic development, two basic strategies may be followed. The first consists in attracting increasing shares of non-regional demand towards Basilicata’s products. In this respect, the growth of exports in challenging times, like the recent ones, indicates that first steps in that direction have been made. But the scope for improvement and strengthening in this broad area is still large, encompassing the trade with the other Italian regions. Basil- icata’s agri-food system (in particular its agricultural component) invests a relevant part of its activities for the direct or indirect fulfillment of the final agri-food demand of the other regions. This is a segment of activity requiring a specific strategy to consolidate the comparative advantages to base them mostly on unique features of the regional system, including the quality of the environment, the specificities of the varieties produced, and the knowledge of the context related to production traditions. If the participation in the market of agricultural commodities (like in the case of cereals for the pasta industry) is an important business segment in Basilicata’s agri-food system, it can and must be made sta- ble by innovation processes aimed at increasing product qualitative differentiation. The second strategy could be described as the strengthening of interdependencies within the regional production system aimed at increasing the share of the multiplier effect remaining within the regional economic system. The subsystem analysis has demon- strated that there are large areas for increasing integration within regional food chains, in particular between agricultural production and industrial processing. In this sense, rural development policies, especially the measures aimed to promote coordinated actions at the district scale, may be a good basis for creating local supply chains and tighter links among regional production activities. This process, however, should again be driven by qualitative differentiation. If, on the one hand, “shorter” food chains can increase the regional multiplier effect through an enhanced integration between agriculture and food industry, they could also represent an important factor to upgrade (and hence add val- ue to) production, with the possibility of increasing “downstream” integration with other regional sectors. The analysis has shown that in Basilicata the integration with food and restaurant and accommodation services is lower than in other Italian regions. But there also is much room for intensifying the interdependencies of the regional agri-food system with elements of the public administration (such as, for instance, public providing activ- ities in school canteens or in hospitals). These market opportunities would be useful to improve final consumers’ awareness of regional production peculiarities and might have long-term additional effects on the growth of demand addressed towards the regional pro- duction system. Results show that the availability of a highly disaggregate multi-sector model of the regional economy is a valuable tool in supporting the design of regional policies for inno- vation and for the development of rural areas. The structural analysis described in this paper could be further extended at the level of each single chain, with the characteriza- tion of the main forward and backward linkages and the interaction with the rest of the national economy. 36 Mauro Viccaro et alii Aknowledgments The paper presents part of the results of the Mauro Viccaro’s PhD Thesis on “Sam- based Models For The Impact Analysis Of Regional Policies On Growth, Poverty And Inequality” developed within the research project “Models for the analysis of the develop- ment processes at regional scale: the social accounting matrices (SAM) as a tool for the analysis of the effects of regional policies on growth, poverty and inequality’ funded by Shell Italia E&P References Belletti, G. (1992). Delocalizzazione dei processi produttivi agroalimentari e analisi di fil- iera. Una applicazione alla Toscana mediante l’approccio per subsistemi. La Ques- tione Agraria 42: 165-208. Bencivenga, A., Vollaro, P.D., Forte, F., Giampietro, A.M. and Percoco, A. (2016). Food and Winetourism in Basilicata. Agriculture and Agricultural Science Procedia 8: 176-185. Brunori, G. and Rossi, A. (2000). Synergy and Coherence through Collective Action: Some Insights from Wine Routes in Tuscany. Sociologia Ruralis 40(4): 409-423. Campoy-Muñoz, P., Cardenete, M.A. and Delgado, M.C. (2017). Economic Impact Assess- ment of Foodwaste Reduction on European Countries through Social Accounting Matrices. Resources, Conservation and Recycling 122: 202-209. Cardenete, M.A. and Sancho, F. (2006). Missing Links in Key Sector Analysis. Economic Systems Research 18(3): 319-325. Cardenete, M. A., Boulanger, P., Del Carmen Delgado, M., Ferrari, E. and M’Barek, R. (2014). Agri-Food and Bio-Based Analysis in the Spanish Economy Using a Key Sector Approach. Review of Urban & Regional Development Studies 26(2): 112-134. Caskie, P., Davis, J. and Moss, J.E. (1999). The Economic Impact of BSE: A Regional Per- spective. Applied Economics 31(12): 1623-1630. Chang, T.F.M. (1981). L’Ipotesi Agribusiness: Una Verifica sulle Matrici Intersettoriali dell’Economia Italiana 1965-1973. Rivista di Economia Agraria 2: 291-344. Chang, T.F.M., Piccinini, L.C., Iseppi, L. and Lepellere, M.A. (2013). The Black Box of Economic Interdependence in the Process of Structural Change. EU and EA on the Stage. Italian Journal of Pure and Applied Mathematics 31: 285-306. Conto, F., La Sala, P. and Papapietro, P. (2009). The Metapontum Agro-Food District of Quality: A Case Study OF Knowledge, Innovation and Improvement of Human Capital in Territorial Rural Development. Paper presented at 113th Seminar Europe- an Association of Agricultural Economists, Belgrade, Serbia (No. 57477), December 9-11, 2009. Cook, M.L. and Chaddad, F.R. (2000). Agroindustrialization of the Global Agrifood Econ- omy: Bridging Development Economics and Agribusiness Research. Agricultural Economics 23(3): 207-218. CREA (2016). Rete di Informazione Contabile Agricola. http://rica.crea.gov.it/public/it/ index.php. Accessed 30 January 2017 Davis, J.H. and Goldberg, R. A. (1957). A Concept of Agribusiness. Division of Research. Graduate School of Business Administration. Boston: Harvard University. 37SAM multipliers and subsystems Dethier, J.J. and Effenberger, A. (2012). Agriculture and Development: A Brief Review of the Literature. Economic Systems 36(2): 175-205. European Commission (2016). Regional Innovation Scoreboard 2016. European Union. http://ec.europa.eu/DocsRoom/documents/17824/attachments/1/translations/en/ renditions/native. Accessed 30 November 2016. European Commission (2016b), The Farm Accountancy Data Network. Agricultural and Rural Development. European Union. http://ec.europa.eu/agriculture/rica/. Accessed 30 November 2016. Ilbery, B. and Kneafsey, M. (2000). Producer Constructions of Quality in Regional Spe- ciality Food Production: A Case Study from South West England. Journal of Rural Studies 16: 217-230. ISTAT (2016a). I.stat: il tuo accesso diretto alla statistica italiana. http://dati.istat.it/Index. aspx. Accessed 24 October 2016. ISTAT (2016b). Coeweb – Statistiche del commercio estero. https://www.coeweb.istat.it. Accessed 24 October 2016. Kneafsey, M., Ilbery, B. and Jenkins, T. (2001). Exploring the Dimensions of Culture Economies in Rural West Wales. Sociologia Ruralis 41(3): 296-310. Llop, M. and Tol, R. S. (2013). Decomposition of Sectoral Greenhouse Gas Emissions: A Subsystem Input-Output Model for the Republic of Ireland. Journal of Environmen- tal Planning and Management 56(9): 1316-1331. Marsden, T., Banks, J. and Bristow, G. (2000). Food Supply Chain Approaches: Exploring Their Role in Rural Development. Sociologia Ruralis 40(4): 424–439. Miller, R.E. and Blair, P.D. (2009). Input-Output Analysis, Foundations and Extensions. New York: Cambridge University Press. Momigliano, F. and Siniscalco, D. (1982). Note in tema di terziarizzazione e deindustrial- izzazione. Moneta e Credito 35(138): 143-182. Montresor, S. and Vittucci Marzetti, G. (2010). The Deindustrialisation/Tertiarisation Hypothesis Reconsidered: A Subsystem Application to the OECD7. Cambridge Jour- nal of Economics 35(2): 401-421. Murdoch, J., Marsden, T. and Banks, J. (2000). Quality, Nature and Embeddedness: Some Theoretical Considerations in the Context of the Food Sector. Economic Geography 76(2): 107–125. Parrott, N., Wilson, N. and Murdoch, J. (2002). Spatializing Quality: Regional Protection and the Alternative Geography of Food. European Urban and Regional Studies 9(3): 241–261. Pasinetti, L.L. (1973). The Notion of Vertical Integration in Economic Analysis. Metroeco- nomica 25(1): 1–29. Platania, M., Rapisarda P. and Rizzo, M. (2015). Italian Trade Specialization: Persistence and Change in Regional Agri-Food Production. AGRIS on-line Papers in Economics and Informatics 7(4): 101. Ray, C. (1998). Culture, Intellectual Property and Territorial Rural Development. Sociolo- gia Ruralis 38(1): 3– 20. Rocchi, B. (2009). The CAP Reform between Targeting and Equity: A Structural Policy Analysis for Italy. European Review of Agricultural Economics 36(2): 175-201. Rocchi, B., Landi, C., Stefani, G., Romano, S. and Cozzi, M. (2015). Escaping the resource 38 Mauro Viccaro et alii curse in regional development: a case study on the allocation of oil royalties. Inter- national Journal of Sustainable Development 18(1/2): 115-138. Rocchi, B., Romano, D., Sadiddin, A., Stefani, G., Zucaro, R. and Manganiello, V. (2016). An Assessment of Agro-Food Frauds in the Italian Economy: A SAM-based Approach. Final Report of the Research project on “Analysis of Fraud Risks in the Agro-Food Sector”, Dipartimento di Scienze per l’Economia e l’Impresa, Università degli Studi di Firenze. Florence (IT), December 2016, mimeo. Round, J.I. (2003). Constructing SAMs for Development Policy Analysis: Lessons Learned and Challenges Ahead. Economic Systems Research 15(2): 161-183. Schultz, T.W. (1964). Transforming Traditional Agriculture. New Haven, CT: Yale Univer- sity Press. Sraffa, P. (1960), Production of Commodities by Means of Commodities: Prelude to a Cri- tique of Economic Theory. Cambridge: Cambridge University Press. Terluin, I. (2003). Differences in Economic Development in Rural Regions of Advanced Countries: An Overview and Critical Analysis of Theories. Journal of Rural Studies 19: 327–344. Tregear, A., Arfini, F., Belletti, G. and Marescotti, A. (2007). Regional Foods and Rural Development: The Role of Product Qualification. Journal of Rural studies 23(1): 12-22. Viccaro, M., Rocchi, B., Cozzi, M. and Romano, S. (2015). The sustainability of non- renewable resources use at regional level: a case study on allocation of oil royalties. In Vastola, A. (ed.), The Sustainability of Agro-Food and Natural Resource Systems in the Mediterranean Basin, Springer, Cham, 225-241. Vega, A., Miller, A.C. and O’Donoghue, C. (2014). Economic Impacts of Seafood Produc- tion Growth Targets in Ireland. Marine Policy 47, 39-45. Volpentesta, A.P. and Ammirato, S. (2013). Alternative Agrifood Networks in a Region- al Area: A Case Study. International Journal of Computer Integrated Manufacturing 26(1-2): 55-66.