HYBRID INTERREGIONAL INPUT-OUTPUT
CONSTRUCTION METHODS. PHASE ONE: ESTIMATION
OF RPCS
Fernando Escobedo-Cardeñoso1 & Jan Oosterhaven2
Tuesday, February 16, 2016
Abstract
This paper searches for an optimal combination of non-survey methods when constructing a
Spanish interregional input-output table for the region of Madrid and the five provinces of the
region of Castilla-La Mancha (CLM), given thirteen Spanish regional input-output tables for
around 2000 and 2005. Hence, we develop a regression analysis to obtain the intra-regional
trade matrices for the provinces of CLM. This regression analysis is based on statistical data
on the road transport of goods, on input-output interpolation and extrapolation techniques to
calculate the regional purchase coefficients ,and uses a new method to add cell-specific
information to the aggregate regressions.
Keywords: Interregional input-output analysis, Non-survey methods, Regression analysis,
Commodity trade flows, Spanish regions.
1. Introduction
The debate about the pros and cons of different methods to construct regional and
interregional input-output tables (IOTs) is predominantly phrased in terms of “either/or”. It
evolves around such research questions as whether non-survey methods are acceptable or not
and even whether using non-survey sector multipliers without IOTs are acceptable. At a more
complex level questions arise about which type of coefficients, such as location quotients
(LQs) or cross industry quotients, are best when estimating a non-survey IOT (Schaffer &
Chu, 1969; Round, 1978), and whether using national coefficients and adapting them with
RAS to the totals of the region at hand is acceptable or not (Hewings, 1969) or whether using
IOT of different regions and adapting them with RAS is acceptable (Thuman, 1978). Even
more subtle issues are the choice of survey strategies. Is it better to ask firms for their
sales/exports behaviour or for their purchase/import behaviour, and how useful is it to oversample wholesale and transport sectors in order to get information on the export or the import
coefficients of the products traded or shipped? (Boomsma & Oosterhaven, 1992). Also there
is the issue whether the accuracy of non-survey methods should be evaluated at the level of
the cells of the IO table (partitive accuracy) or better at the level of IO sector multipliers
(holistic accuracy) (Jensen, 1980). Finally, specifically at the interregional level, at the edge
between IO table construction and IO model building, there is the choice between the
interregional IO versus multiregional IO versus gravity IO methods (Isard, 1951; Chenery,
1953; Moses, 1955; Leontief & Strout, 1963).
1
School of Civil Engineering and Planning, University of Castilla-La Mancha, Avenida de Camilo Jose Cela 2,
13071 Ciudad Real, Spain. Email: fernando.escobedo@uclm.es.
2
Faculty of Economics and Business, University of Groningen, Postbus 800, 9700 AV Groningen, The
Netherlands. Email: j.oosterhaven@rug.nl.
In many practical cases, however, the issue is not so much about which method is better or
worse, but which method is more appropriate in which data situations or in which model
applications, and in which it less. Thus, it may occur that both Beemiller (1990) and Bourque
(1990) are right. Bourque in claiming that the RIMS non-survey IOTs based on LQs produces
too large errors to be acceptable, and Beemiller in claiming that combining survey
information for the exogenous impulse with a non-survey IOT produces acceptable impact
estimates. The same conclusion was drawn from a tourism impact study in the Netherlands.
Multipliers for the tourism-related sectors that were survey-based were very close to those of
a comparable survey IOT, whereas the non-survey multipliers for the other sectors were quite
off the mark, notwithstanding that they did not influence the impact variables much (Spijker,
1985). Consequently, West (1990) and Lahr (1993) conclude that the future is for hybrid IO
tables and models. When constructing a series of intercountry IOT for the EU Van der Linden
and Oosterhaven (1995) use a – given the available data – optimal combination of the
interregional IO model, the multiregional IO model, and RAS to re-price the import matrices
from ex customs prices to producers’ prices.
In this paper, we present the first construction phase of an interregional IOT for the Spanish
regions of Madrid and Castilla-La Mancha (CLM), given the available data. This first phase
has two main steps. The first step has as its main goal to approximate the Regional Technical
Coefficients (RTCs). Finally in the second and last step of this first phase the Regional
Purchase Coefficients (RPCs) are estimated, and in this way we are able to calculate the intraregional transactions matrices for the five CLM provinces.
For this purpose we use the national survey IOT and thirteen survey type regional,3 domestic
imports and foreign imports IOTs with between 60 and 120 sectors, for around the period
2000-2005. Besides, we develop new methods to cope with the richness of survey data about
RPCs, based on the IO interpolation and extrapolation techniques from Oosterhaven (2005).
2. Construction problem and solution strategy
The accounting structure of the ideal interregional input-output table (IOT) for a certain
nation (E) is given in Table 1. Each of the larger rectangular submatrices has the size I*(I+Q),
where I = number of industries and Q = number of domestic final demand categories. The
number of these interregional submatrices equals (R+1)*R, where R = number of regions and
1 = foreign imports. The other submatrices relate to remaining final demand F (foreign
exports and changes in stocks) and gross value added at market prices V (product taxes minus
subsidies, labour cost and operating surplus). Note that their overall totals equal total imports,
exports and value added4. Thus, Table 1 depicts the sectoral disaggregation of both the regional
and the national macro economic accounting identities (Y = C + I + G + E – M)5.
The construction of any interregional IOT, within general accounting principles, essentially
depends on the availability of survey data. Details on the data available for the construction of
an interregional IOT for the Region Castilla-La Mancha (CLM) and the Comunidad de
Madrid are given in the following pages. The location of the two regions and the constituent
five Provinces of CLM are given in Figure 1. Their relative locations underscore the
3
Spain is a country constituted by seventeen regions, i.e. Andalucía, Aragón, Asturias, Baleares, Canarias,
Cantabria, Castilla La Mancha, Castilla León, Cataluña, Extremadura, Galicia, Madrid, Murcia, Navarra, Rioja,
Comunidad Valenciana and País Vasco, and two autonomous cities, Ceuta and Melilla.
4
The zeros indicate that transit trade is assumed to be equal to zero.
5
In Table 1, taking cell totals of the vectors and matrices mentioned, the national macro economic accounting
identity equals vE = ∑r yr + fE – mE, while the regional macro identity equals Vr +Vyr = yr + ∑sr (Zrs + Yrs) + Fr –
∑sr (Zsr + Ysr) – Mr – Myr, with yr = Cr + Ir + Gr.
2
economic logic of combining the region of Madrid (M) with that of Castilla-La Mancha (C)
into one IOT. The five CLM-provinces are: Albacete (A), Ciudad Real (D), Cuenca (U),
Guadalajara (G) and Toledo (T).
Spain (E), Madrid (M), and Castilla-La Mancha have detailed industry-by-industry IOTs with
73-59-68 sectors respectively for 2005, although finally as we face the different sector
classification of the three tables we have a 50 sector classification for the model, as it can be
seen in Table 2. In the case of Spain a full matrix with imports from the Rest of the World
(RoW) is available, whereas for exports only columns for the Rest of the EU (RoEU) and the
Remainder of the RoW (RRoW) are available. Additional to that information, the comparable
IO tables for Madrid and CLM also have a full import matrix for the Rest of Spain (RoS),
separate import matrices for the RoEU and the RRoW, and an export column for the RoS. For
more details see Table 3, where we may check that we have eleven symmetric regional IO
tables (Andalucía, Aragón, Asturias, Baleares, Castilla-La Mancha, Castilla León, Comunidad
Valenciana, Galicia, Madrid, Navarra and País Vasco) and other two supply and use tables
(Canarias and Cataluña).
Besides this, we have detailed information on the size of the sectors in the five Provinces of
CLM and in the Region of CLM for 2005 related to the employment6 of 60 sectors7, the two
digit sectors NACE Classification8. Besides, at the level of 6 sectors for the provinces and 29
sectors for the regions, aggregate data about sectoral gross values added is available.
Furthermore, detailed foreign export data are available for all provinces as well as detailed
road transport data.The latter, however, are only considered reliable when aggregated to
regions. Thus, hardly more information is available for the Region of CLM than for the
Provinces of CLM. Consequently, we have chosen to construct a seven region IOT for Spain,
with Madrid, the five CLM provinces and the RoS instead of a three region table with Madrid,
CLM and the RoS. In this way, we are able to reach the NUTS 3 level in the regional
statistical classification of the European Union9. This seven-region IOT will be estimated
such that it is consistent with the national IOT for Spain, which will thus function as the
double-entry control total for the interregional IOT (see Boomsma & Oosterhaven, 1992, for
the optimal use of double-entry bi-regional book-keeping).
The location of the two regions and the constituent five Provinces of CLM are given in Figure
1. Their relative locations underscore the economic logic of combining the region of Madrid
(M) with that of Castilla-La Mancha (C) into one IOT. The region Rest of Spain (R) is Spain
less Madrid less CLM, that is, the total of Andalucía, Aragón, Asturias, Baleares, Canarias,
Cantabria, Castilla León, Cataluña, Comunidad Valenciana, Extremadura, Galicia, Murcia,
Navarra, País Vasco, Rioja and the autonomous cities of Ceuta and Melilla. Finally the region
Rest of the World (RoW) is the World less Spain.
Another source of information is the Spanish census of houses and people, “Censo de población y viviendas”,
of 2001. There, we can find basic demographic data, place of residence, academic profile, employment data
(inter alia. hours worked and economic sector) and commuting data.
7
These data come from the “Tesorería General de la Seguridad Social”, an agency of the Spanish Ministry of
Labour and Inmigration, which controls the employees affiliated to the national social security system. It does
not include the public servants affiliated to the specific security system “Muface”, the soldiers affiliated to
another specific security system “Isfas” and other professionals affiliated to private security systems.
8
The “Nomenclature generale des Activites economiques dans les Communautes Europeennes” refers to the
industrial classification as defined in Revision 1 used by Eurostat. It has 17 Sections (letters A to Q), 31
Subsections (2-character alphabetical codes), 60 Divisions (2-digit codes), 222 Groups (3-digit codes) and 503
Classes (4-digit codes).
9 The five provinces of CLM are NUTS 3 regions according to Regulation No 1059/2003 on the “Nomenclature
of Territorial Units for Statistics” (NUTS).
6
3
Given the geographical characteristics and the data availability, the core of the construction
problem and the core of its solution are two-fold. First, regional technical coefficients and
regional final expenditure coefficients have to be estimated. Second, regional purchase
coefficients have to be estimated. With the these two sets of estimates, the intra-regional
intermediate and final purchases ZRR and FRR can be estimated.
2.1 Technical and expenditure coefficients per sector and category
To get a precise description of the procedure we define the following parameters:
 n , number of sectors of the economic structure of Spain and its different regions.
 wiP , employment of sector i of region P .
 xiP , total output of sector i of region P .
 viP , total value added of sector i of region P .
 mclm, number of provinces of the global region Madrid + CLM.
Besides, we have to define net taxes in CLM provinces:
(1)
niP  wiP
niCLM
, P  A, D,U , G, T ; i  1...n,
wiCLM
In this step next to the given matrices with total intermediate requirements and total final
requirements for Madrid (i.e. the sum of the domestic and foreign origin matrices),
comparable technical requirement matrices have to be estimated for the five CLM provinces.
After that, it has to be checked that the total requirements by purchasing sector and region
plus value added 1st step, are equal for total output.
We may define RTCijP , the Regional Technical Coefficient of sector j of region P with
respect to sector i , as the total purchases of sector j of region P to the sector i of Spain and
the rest of the world. In the case of Castilla-La Mancha, with mijCLM indicating the imports of
sector j of region CLM from sector i in RoW, the Regional Technical Coefficient is:
(2)
RTC ijCLM 
zijCLM  zijrS ,CLM  mijCLM
x CLMj vCLM
j
If we define rijP as total requirements of sector j of region P from sector i , regardless its
location, we estimate it as follows:
(3)


rijP  zijP  zijrS ,P  mijP  RTCijCLM x Pj  v Pj , P  A, C,U , G, T ; i, j  1,..., n
Obviously (3) secures, as requires, that the total of all requirements equals total use, that is
n
(4)
r
i 1
P
ij
 x Pj  v Pj .
2.2 Estimation of regional purchase coefficients
In relation to the given intra-regional transactions matrix for Madrid, comparable matrices
with intra-regional intermediate and final transactions have to be estimated for the five CLM
provinces.
4
The most common way to estimate these matrices is by multiplying the total requirement
matrices from step 1 with provincial self-sufficiency ratios, also labelled as regional purchase
coefficients (RPCs, Stevens & Trainer, 1980). The Spanish IO data are unique in the sense
that 50x54=2700 cell-specific RPCs can be calculated for at least eleven Spanish regions. To
keep the construction methodology tractable, we have chosen to explain the (row) aggregate
RPCs for the 50 supplying sectors by means of regression analysis with sectoral and regional
fixed effects, and to use the average row pattern of the cell-specific RPCs to differentiate the
aggregate RPCs by purchasing industry and purchasing category of final demand (see
Ralston, Hastings & Brucker, 1986, for the necessity to do this). To explain the aggregate
sectoral RPCs from the thirteen regional IOTs, we will use the aggregate sectoral RPCs from
the regional trucking survey, regional sectoral employment shares and geographical
indicators, such as land surface share and the regional sectoral employment densities (cf.
Oosterhaven, 2005). This will give a maximum estimate for the aggregate RPCs for the five
CLM provinces.
We define ARPC iP , the Aggregate Regional Purchase Coefficient of sector i of region P , as
the ratio between the total purchases of region P from the sector i in region P to the total
purchases of region P from the sector i in the whole world:
n
(5)
ARPCiP 
z
j 1
n
n
z
j 1
PP
ij
PP
ij
n
  zijE  P , P   mijP
j 1
j 1
Next, we undertake a regression analysis (RA) and use the trucking survey and other socioeconomic variables to check the past results (Wilson, 2000). The regression analysis has the
general form indicated in Equation 6:
(6)
Being:
ARPCiP   i   P   1
tsiPP
wiP
LP
wiP






  iP , i
2
3 E
4
tsiPP  tsiE P ,P  tsiFP
wP
L
LP
 i , fixed effects constants related to the economic sectors.
  P , fixed effects constants related to the regions.
  1 ,  2 ,  3 . 4 , coefficients of the variables.
 tsiRS , trucking survey economic value of weight of good i transported from region
R to region S in the same year of the regional IO table of region S
 LR , land surface of the region R .
 i , type of commodity transported, i  1...s
Before reasonable results are reached in the regression analysis in terms of t-Statistic and Rsquared values, we have to go through some stages improving each time the results of the
previous stage.
Next, we define RPCijP , the Regional Purchase Coefficient of sector j in region P with
respect to commodity i , as the ratio of the total purchases of sector j of region P to the
commodities i from region P to the total purchases of sector j of region P to the sector i
of the whole world:
5
zijPP
RPCijP 
(7)

rijP
zijPP
zijPP  zijrS ,P  mijP
Then, we assume that the ratio between the individual regional purchase coefficient and the
aggregate regional purchase coefficient for the five CLM provinces,
aS
z
RPCijP
(8)
ARPC

P
i
PP
ij
z
PP
ij

mclmr
z
a 1
aP
ij
 mijP
,
aS
n
z
j 1
n
z
PP
ij
j 1
PP
ij

mclmr n
 z
a 1
j 1
n
aP
ij
  mijP
j 1
is the same than the Castilla-La Mancha:
RPCijP
(9)
ARPCiP RA

RPCijCLM
ARPCiCLM
If we take into account the former results we will have:
(10)
RPC  ARPC
P
ij
P  RA
i
RPC ijCLM
ARPCiCLM
; P  M , A, C ,U , G, T , R ; i  1...n
Finally we will be able to approximate the intra-regional IO table for the five provinces of
CLM, as follows:
(11)
zijPP  ARPCiP RA
RPCijCLM
ARPC
CLM
i
z
PP
ij

 zijrS ,P  mijP ; P  A, C ,U , G, T ; i, j
And then, for the provinces, a maximum estimate is:
n
(12)
z
j 1
CC
ij
n

  zijAA  zijDD  zijUU  zijGG  zijTT

j 1
3. Results and discussion
For the regression which is the core of this paper we need the following data: Intermediate
demand and final demand of the IO tables, the statistics with respect to the carriage of goods
by road, employment data and land areas data. The two last data are easily obtained from the
Spanish Statistics National Institute.
As it has been said above, the intermediate and final demand is taken from the thirteen
Spanish regional IO tables, whose economic sector classification are related to the NACE 93
rev. 1 Classification10 or to the CPA 96 Classification 11. The trucking survey is undertaken
10
Nomenclature generale des Activites economiques dans les Communautes europeennes (NACE) refers to the
industrial classification as defined in Revision 1 which is used by Eurostat. It has 17 Sections (letters A to Q), 31
Subsections (2-character alphabetical codes), 60 Divisions (2-digit codes), 222 Groups (3-digit codes) and 503
Classes (4-digit codes). It is regulated in the “Council Regulation (EEC) No 3037/90 of 9 October 1990 on the
statistical classification of economic activities in the European Community”, whose purpose is to establish a
common statistical classification of economic activities within the European Community in order to ensure
comparability between national and Community classifications and hence national and Community statistics.
11
Established in the Council Regulation (EEC) No 3696/93 of 29 October 1993 on the statistical classification of
products by activity (CPA) in the European Economic Community. The statistical Classification of Economic
Activities (NACE) and the statistical Classification of Products by Activity (CPA) in the European Community
6
every year in all the countries of the European Union12. In the case of Spain, for example, the
survey carried out in 2000 had 227,931 data where the most important variables are the type
of goods, according to the groups in the NSTR/67 classification, the weight of the goods
(gross weight in 100 kg), and the region or province of loading and unloading of the goods
(country in the case of imports or exports).
The first and main difficulty we have to overcome is the different product classifications
between the input-output tables and the trucking survey. The two digits CPA 96 classification
has 60 economic sectors and the NSRT/67 classification has 52 two digits sectors. When we
face both classifications, CPA and NSTR, we realize that all the NSRT sectors only
correspond to the first half of the CPA sectors, as the second half of the CPA sectors are
related to services, not to products, so for the services sectors we do a regression between the
total trades between two territories.
After this step we undertake the regression analysis but in a first stage we mix the fixed
effects i and  P into one only constant. As we may see in Table 4, the results are quite
good in terms of t-statistics, as all the variables are significant, that is, are above 1.96, 95% of
reliability. Nevertheless the area share is not as good as it has 87.6% of reliability. With
respect to the R-squared this is not good as it is 0.21 when the acceptable values are clearly
above 0.6 or 0.7, so it is necessary to carry out more regressions but in this occasion with
fixed sectoral and regional effects.
Another analysis we have to do is the multicolinearity or correlation between the variables.
This analysis can be seen in Table 5, where we can check that there is a strong correlation
between the variables associated to the employment, that is, employment share and
employment density. As this last one has a worse t-statistic value it would be adequate not to
use it in the following regression when we are to use fixed effects.
4. Conclusions
In this paper Phase 1 of the construction of an interregional input-output table for Spain is
presented and it is mainly based on the statistical data with respect to the carriage of goods by
road that is undertaken every year in the different countries of the European Union. These
statistical data provide us with abundant data (more than 200,000 surveys in Spain in 2000)
for the economic sectors that produce goods like agriculture, fish, mining, and manufacturing
but do not contribute with data in the services sector, so we have to assume that the economic
transactions in this sector are proportional to the totals in the sectors that interchange goods.
The key point of the model with respect to obtaining the intra-regional input-output matrices
is a regression analysis between input-output tables of Spanish regions which have detailed
regional, domestic imports and foreign imports and the statistical returns in respect of the
carriage of goods by road, besides other parameters like employment and land areas. This
regression analysis can be solved with a statistical software package and it has to focused the
attention in three main points: t-statistics, R-squared and correlation between the variables.
Then we have to select the number of variables and create and optimal mixture of fixed
sectoral and regional effects to obtain the best results in terms of the statistical parameters
mentioned above.
are part of the integrated system of statistical classifications. They have the same structure, their difference is
that NACE is for industries and CPA is for products.
12
This is the subject of the “Council Regulation (EC) No 1172/98 of 25 May 1998 on statistical returns in
respect of the carriage of goods by road”. This regulation states that Member States shall compile statistical data
relating to the following areas: a) vehicle-related data; b) journey-related data and c) goods-related data.
7
Acknowledgements
Yves Mahieu and other colleagues from Eurostat provided us with useful correspondence
tables between different economic classifications used in this paper.
References
Beemiller, R. M. (1990) Improving accuracy by combining primary data with RIMS:
Comment on Bourque, International Regional Science Review, 13, 1-2, pp. 99-101.
Boomsma, P. & Oosterhaven J. (1992) A double-entry method for the construction of biregional input-output tables, Journal of Regional Science, 32, 3, pp. 269-84.
Bourque, P.J. (1990) Regional multipliers: WAIO vs. RIMS, International Regional Science
Review, 13, 1-2, pp. 87-98.
Chenery, H.B. (1953) Regional analysis, in: H.B. Chenery, P.G. Clark & V.C. Vera (eds.) The
structure and growth of the Italian economy (Rome, U.S. Mutual Security Agency).
Hewings, G. J. D. (1969) Regional input-output models using national data: The structure of
the West Midlands economy, The Annals of Regional Science, 3, 1, pp. 179–191.
Isard, W. (1951) Interregional and regional input-output analysis, a model of the space
economy, Review of Economics and Statistics, 33, pp. 318-28.
Jensen, R.C. (1980) The concept of accuracy in regional input-output models, International
Regional Science Review, 5, 2, pp. 139-54.
Lahr, M. (1993) A review of the literature supporting the hybrid approach to constructing
regional input-output models, Economic Systems Research, 5, 3, pp. 277-93.
Leontief, W. W. & Strout A. (1963) Multiregional Input-Output Analysis, in T. Barna (ed.)
Structural Interdependence and Economic Development (New York, St. Martin’s Press).
Linden, J. A. van der & Oosterhaven J. (1995) European Community Intercountry InputOutput Relations: Construction Method and Main Results for 1965-1985, Economic Systems
Research, 7, 3, pp. 249-69.
Moses, L. N. (1955) The stability of interregional trading pattern and input-output analysis,
American Economic Review, 45, 5, pp. 803-32.
Oosterhaven, J. (2005) Spatial Interpolation and Disaggregation of Multipliers, Geographical
Analysis, 37, 1, pp. 69-84.
Ralston, S.N., S.E. Hastings & S.M. Brucker (1986) Improving regional I-O models:
Evidence against uniform purchase coefficients across rows. Annals of Regional Science 20,
1: 65-80.
Round, J. I. (1978) An interregional input-output approach to the evaluation of non-survey
method, Journal of Regional Science, 18, 2, pp. 179-94.
8
Schaffer, W. A. & Chu K. (1969) Non-survey techniques for constructing regional
interindustry models, Papers of the Regional Science Association, 23, pp. 83-101.
Spijker, J. (1985) Een combinatie van directe en indirecte methoden, in: J. Oosterhaven en
B.B.A. Drewes (eds.), Constructie en actualisering van regionale en interregionale inputoutput tabellen, (Arnhem, Economisch-Technologische Instituten).
Stevens, B. H. & Trainer G. A. (1980) Error generation in regional input-output analysis and
its implications for non-survey models, in S. Pleeter (ed.) Economic impact analysis:
methodology and applications (Boston, Martinus Nijhoff).
Thuman, R.G. (1978) A comment on “Evaluating the possibility for exchanging regional
input-output coefficients”. Environment and Planning A 10: 321-5.
West, G. R. (1990) Regional trade estimation: A hybrid approach, International Regional
Science Review, 13, 1-2, pp. 103-18.
Wilson, A. G. (2000) Complex Spatial Systems: The Modelling Foundations of Urban and
Regional Analysis. Devon, England, Northcote House Pub Ltd.
9
TABLE 1. Layout of an ideal interregional input-output table.
Region 1
…
Region R
1
Z11 Y11
...
Z1R Y1R
F1
x1
:
:
:
:
:
:
R
ZR1 YR1
...
ZRR YRR
FR
xR
M1 My1
...
MR MyR
0
mE
V1 Vy1
...
VR VyR
0
vE
(x1 y1)’
...
(xR yR)’
(fE)’
∑
10
∑
TABLE 2. The classification of the Spanish interregional input-output model.
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
Name
Agriculture, hunting, forestry, logging, fishing, fish farming and related service activities
Mining of coal and lignite; extraction of peat; extraction of crude petroleum and natural gas; service activities
incidental to oil and gas extraction, excluding surveying; mining of uranium and thorium ores; manufacture
of coke, refined petroleum products and nuclear fuel
Mining of metal ores and other mining and quarrying
Production, processing and preserving of meat and meat products
Manufacture of dairy products
Processing and preserving of fish, fish products, fruit and vegetables. Manufacture of vegetable and animal
oils and fats, grain mill products, starches and starch products, prepared animal feeds and other food products
Manufacture of beverages and tobacco products
Manufacture of textiles
Manufacture of wearing apparel; dressing and dyeing of fur
Tanning and dressing of leather; manufacture of luggage, handbags, saddlery, harness and footwear
Manufacture of wood and of products of wood and cork, except furniture; manufacture of articles of straw
and plaiting materials
Manufacture of pulp, paper and paper products
Publishing, printing and reproduction of recorded media
Manufacture of chemicals and chemical products
Manufacture of rubber and plastic products
Manufacture of glass and glass products
Manufacture of other nonmetallic mineral products, except glass
Manufacture of basic metals
Manufacture of fabricated metal products, except machinery and equipment
Manufacture of machinery and equipments n.e.c.
Manufacture of office machinery, computers, medical, precision and optical instruments, watches and clocks
Manufacture of electrical machinery and apparatus n.e.c.
Manufacture of radio, television and communication equipment and apparatus
Manufacture of motor vehicles, trailers and semi-trailers
Manufacture of other transport equipment
Manufacture of furniture; manufacturing n.e.c. and recycling
Electricity, gas, steam and hot water supply and collection, purification and distribution of water
Construction
Sale, maintenance and repair of motor vehicles and motorcycles; retail of automotive fuel
Wholesale trade and commission trade, except of motor vehicles and motorcycles
Retail trade, except of motor vehicles and motorcycles; repair of personal and household goods
Hotels and restaurants
Transport via railways
Other land transport; transport via pipelines
Water and air transport
Supporting and auxiliary transport activities of travel agencies
Post and telecommunications
Financial intermediation, except insurance and pension funding
Insurance and pension funding, except compulsory social security and activities auxiliary to financial
intermediation
Real estate activities and renting of machinery and equipment without operator and of personal and
household goods
Computer and related activities
Research and development
Other business activities
Public administration and defence; compulsory social security
Education
Health and social work
Sewage and refuse disposal, sanitation and similar activities
Activities of membership organisations n.e.c. and recreational, cultural and sporting activities
Other service activities
Activities of households
11
NACE groups
01, 02, 05
10, 11, 12, 23
13, 14
15.1
15.5
15.2-15.4, 15.6-15.8
15.9, 16
17
18
19
20
21
22
24
25
26.1
26.2-26.8
27
28
29
30, 33
31
32
34
35
36, 37
40, 41
45
50
51
52
55
60.1
60.2-60.3
61, 62
63
64
65
66, 67
70, 71
72
73
74
75
80
85
90
91, 92
93
95
Symmetric Input-Output table
Use table
Supply table
TABLE 3. Details of the survey regional input tables in Spain.
Region or Country and
Andalucía
Aragón
year
2000 (IOT
1999 (IOT
2005
2005 in
available)
process)
Prices
Number of products
Adjustment items
Total output categories
Number of industries
Imports columns
Transformation to
purchasers’ prices
categories
Prices
Number of products
Imports matrices
Adjustment items
Value added
Number of industries
Final consumption
expenditure
Gross capital formation
Exports columns
Prices
Number of industries or
products
Imports matrices
Adjustment items
Value added
Final consumption
expenditure
Gross capital formation
Exports columns
Asturias
2000 (IOT
2005
available)
Baleares
2004
Canarias
2002 (IOT
2005 in
mid 2010)
Castilla-La
Mancha
2005
Castilla y
León 2000
(IOT 2005
in the end
of 2009)
Basic
99
d, f
Cataluña
2001 (IOT
2005 in
mid 2009)
Comunidad
Valenciana
2000 (IOT
2005 in
process)
Basic
98
f
s, t, u
84
a, b, c
w, v, z, x,
y, r
Galicia
1998 (IOT
2005 in the
end of
2009)
Basic
109
Madrid
2000 (IOT
2005 in
mid 2009)
Navarra
2005
País Vasco
2000 (IOT
2005
available)
España
2000 (IOT
2005
available)
Basic
90
f
s, t, u
76
a, b, c
w, v, x-y
Basic
100
d, f
48
a, b, c
w, v, x-y
Basic
101
d, f
s, t, u
87
a, b, c
w-v, x-y
Basic
118
d, f
s, t, u
75
b, c
w, v, x-y
Basic
90
a, b, c
e, f, g
h, i, j
76
k, l, m-n
Basic
100
a, b-c
d, e, f, g
h, i, j
48
k, l, m-n
Basic
101
a, b-c
e, f, g
h, i, j
87
k, m-n
Basic
118
b-c
d, e, f, g
h, i, j
75
k, l, m-n
Basic
86
f
s, t, u
86
a, b, c
v, w, x-y
Basic
90
d, f
s, t, u
68
a, b, c
w, v, x-y, r
Basic
77
f
s, t, u
65
a-b-c
w, v, x-y
Basic
62
f
s, t, u
57
a, b, c
w, v, x-y
No
Supply
table
Basic
73
d, f
s, t, u
69
a, b, c
w, v, y, g, r
Basic
86
a, b, c
e, f, g
h, i, j
86
k, l-m, n
Basic
90
a-b-c*
g, r
Basic
77
a, b, c
e, f, g
h, i, j
65
k, l-m, n
Basic
62
a, b, c
e, f, g
h, i, j
57
k, l, m-n
Basic
61
a, b-c
e, f, g
h, i, j
61
k, l-m-n
Basic
73
a, b, c
d, e, f, g
h, i, j
69
k, l, m-n
Basic
99
a, b, c
d, e, f, g
h, i, j
60
k, l, m-n
Basic
122
a, b-c
e, f, g
h, i, j
122
k, l-m-n
Basic
98
a, b, c
e, f, g
h, i, j
84
k, l, m-n
Purchasers’
109
a, b, c
p, q
a, b, c
Basic
68
industries
p, q
a, b, c
Basic
58
Products by
58
industries
a, b, c
d, e, f, g
h, i, j
p, q
a, b-c
p, q
a, b, c
Basic
84
industries
p, q
a, b, c
Basic
65
industries
p, q
a, b, c
Basic
74 products
by 74
industries
p, q
a, b, c
Basic
48
industries
p, q
a, b, c
Basic
87
industries
p, q
b, c
Basic
73
industries
a-b-c*
IO
a-b-c*
e, f, g
h, i, j
a, b, c
e, f, g
h, i, j
a, b-c
d, e, f, g
h, i, j
a, b-c
e, f, g
h, i, j
b-c
d, e, f, g, r
h, i, j
k, l, m-n
table
k, l, m-n
k, l, m-n
k, l, m-n
k, m-n
k, l, m-n
p, q
a, b, c
p, q
a, b, c
p, q
a, b, c
p, q
b, c
68
k, l, m, n
p, q
a, b, c
Basic
86
industries
p, q
a, b, c
Basic
68 products
by 68
industries
p, q
a, b, c
Basic
65
industries
p, q
a, b, c
Basic
62
products
p, q
a, b-c
a, b, c
e, f, g
h, i, j
a-b-c*
d, e-f, g, r
h, i, j
a, b, c
e, f, g
h, i, j
a, b, c
e, f, g
h, i, j
symmetric
IO
a, b, c
d, e, f, g
h, i, j
k, l-m, n
k, l, m, n
k, l-m, n
k, l, m-n
table
k, l, m-n
p, q
a, b, c
p, q
a, b, c
p, q
a, b, c
p, q
a, b, c
No
p, q
a, b, c
60
a, b, c
w, v, x-y
p, q
a, b, c
No
Supply
table
No
symmetric
p, q
a, b, c
66
a, b, c
v-w, r, x, y
h, i, j
67
k, l-m, n
d-e-f-g-h-ij
k-l-m-n-pq-a-b-c
a: Rest of Spain. b: Rest of the European Union. c: Rest of the world.
d: Cif/ fob adjustments on exports. e: Purchases on the domestic territory by non-residents. f: Direct purchases abroad by residents. g: Net taxes on products. r: Value added tax.
h: Compensation of employees. i: Other net taxes on production. j: Gross operating surplus and mixed income.
k: By households. l: By Non-profit institutions serving households (NPISH). m: By government, individual services. n: By government, collective services.
p: Gross fixed capital formation and valuables. q: Changes in inventories.
s: Market output. t: For own final use. u: Other non-market output.
v: Transport margins. w: Trade margins. x: Taxes on products. y: Subsidies on products. z: taxes on imports.
-: Means categories combined in one row, in one column or in one matrix.
*: The row totals are disaggregated in three values: domestic imports, EU imports and rest of the world imports.
TABLE 4. Results of the different regression.
Variable
Constant
Trucking Survey
Employment Share
Area Share
Employment Density
R-squared
Coefficient
Std. Error
t-Statistic
Prob.
0.195259
0.037668
5.183736
0.000000
0.378000
0.052061
7.260639
0.000000
4.699469
0.827184
5.681285
0.000000
0.315400
0.204679
1.540953
0.123899
18.793769
7.915307
2.374358
0.017919
Mean Dependent Variable
0.514682
0.210302
Standard Deviation
Adjusted R-squared
0.204600
Dependent Variable
0.319870
0.285277 Akaike Information Criterion
0.338192
Standard Error
Regression
Sum Squared
Residuals
45.086157
Schwarz Criterion
0.376887
Log likelihood
-89.524648
Hannan-Quinn Criterion
0.353303
Statistic
1.110429
F-Statistic
36.883407
Prob (F-statistic)
0.000000
Durbin-Watson
TABLE 5. Correlation between the variables.
Trucking Employment
Employment
Survey
Share
Area Share
Density
Trucking Survey
1.00000
0.06568
-0.11687
-0.03135
Employment Share
0.06568
1.00000
-0.01215
0.60135
Area Share
-0.11687
-0.01215
1.00000
-0.24065
Employment Density
-0.03135
0.60135
-0.24065
1.00000
FIGURE 1. Seven regions of the Spanish interregional input-output table.
R
M
T
G
U
D
A
M: Madrid.
A: Albacete.
D: Ciudad Real.
U: Cuenca.
G: Guadalajara.
T: Toledo.
R: Rest of Spain.
14