19th International Conference of the ACEI
Valladolid, Spain, June 21-24, 2016
Analysis of the technical and allocative efficiency of publicly owned Spanish
museums
María José del BARRIO TELLADO (*)
Dpto. de Economía Financiera y Contabilidad. Universidad de Valladolid (Spain)
E-mail: [email protected]
Luis César HERRERO PRIETO
Dpto. de Economía Aplicada. Universidad de Valladolid (Spain)
E-mail: [email protected]
ABSTRACT
The aim of this work is to evaluate the efficiency and best practices of the Spanish national network
of museums. A non-parametric approach, Data Envelopment Analysis (DEA), is used to measure
relative efficiency in these institutions, and we employ a complex production function embracing a
number of inputs and outputs adapted to the various functions which museums fulfil: preservation,
research, communication, and exhibition. Most of the previous studies to have explored this topic
have mainly focused on determining technical efficiency, without taking into account the costs of
the factors, due to the difficulty involved in having information available concerning prices and costs
in this sector. We now aim to calculate overall levels of efficiency for the museums in the selected
sample, distinguishing them in terms of their components of allocative efficiency and technical
efficiency. To achieve this, we calculate the mean price of the factors used, taking as a guideline the
data contained in the annual accounts of a group of museums considered to be a reference sample
over the period analysed. Finally, we apply a two-stage approach to measure efficiency of museums
in order to analyse how and how much environmental variables might determine the level of
museum efficiency. In this case, we consider indicators such as reputation of the location, urban
size, accessibility, tourism capacity, and model of institution management as external variables
which might influence museum performance. The research findings may prove useful for
management of these cultural institutions, as well as for those responsible for public resource
allocation policies in the area of cultural heritage.
KEY WORDS: Economics of Museums, Technical Efficiency, Allocative Efficiency, Data Envelopment
Analysis, National Network of Museums in Spain
(*) Corresponding Author
Grupo de Investigación Reconocido en Economía de la Cultura
Facultad de Comercio. Universidad de Valladolid
Paseo Prado de la Magdalena s/n 47005-Valladolid
Tel. +34 983 184 642 // Fax: +34 983 423 056
www.emp.uva.es/giec
1- Introduction
1
The goal of our analysis is to delve further into the evaluation of cultural institutions’ results and
performance, taking as a case study the network of Spanish national museums. These institutions
conduct their activity in the public sector, and are funded through the taxes collected by the state as
well as by the donations made by private organisations or individuals albeit to a much smaller
degree. This implies a kind of tacit contract between the state and citizens, who agree to the
payment of taxes in exchange for a range of services which include cultural facilities. This by no
means clear contract has aroused an interest in gaining an understanding of the use which
museums make of the funds they are allocated, not only from the perspective of ensuring
management common sense vis-à-vis the use thereof but also with regard to ascertaining whether
the previously established objectives are achieved in an environment of dwindling resources and
ever-increasing competition for available resources. As a result, there is a desire to gauge
museums’ performance and to relate this to the resources used in an effort to ascertain whether
the museum is efficient in its work.
In the business domain, financial information proves an extremely useful tool for determining a
company’s achievements given the goal of maximising profit, which may be reflected through the
balance sheet. In the public sector, however, as occurs in the area of non-profits, the goals lie
outside what is a purely financial framework. The aim is to achieve goals of a social nature (in the
case of cultural institutions: preserving heritage, disseminating an artistic corpus, educating people,
etc.), objectives which are difficult to measure in monetary terms. The financial information made
publicly available in these institutions is seen as an indicator of their chances of survival or, as
pointed out by Carnegie and Wolnizer (1996), of their viability, although such information proves
insufficient vis-à-vis gauging their vitality, in other words, their performance.
The limited public information available has spurred the development of techniques aimed at
determining the performance of public institutions, and which may also be applied to non-profits.
Said techniques may be grouped into two categories. The first category proposes measuring
performance using tables of indicators applying varying sizes and degrees of complexity. In this line
of study and, particularly in the sector of cultural institutions, are to be found such works as Jackson
(1991), Weil (1995), Gilhespy (1999) and Turbide and Laurin (2009), which advocate the
evaluation of the various activities undertaken by institutions through devising a series of useful
indicators –whether for internal management purposes or external accreditation-, by enabling
comparisons to be established between the organisations in question for the various aspects being
evaluated. In a similar vein, Weinstein and Bukovinsky (2009) posit the use of a comprehensive
dashboard as a tool to evaluate cultural entities and which would allow for an internal link to be
established between the organisation’s strategies and its actions. Reid (2011) also points to the
advantages of applying a comprehensive dashboard for the running of academic libraries as an
instrument for gauging the quality of the services provided to users.
A second stream of studies aims to evaluate the performance of public institutions by relating the
inputs used in the products and services offered by the museums’ various activities by devising a
specific production function. In this case, the aim is to delimit a frontier of optimal behaviour in the
transformation process in order to determine how far away each unit lies from the optimum
established by the frontier. For this purpose, estimating the efficient frontier may be accomplished
by using parametric and non-parametric models (Fernández et al, 2013). The former require an
explicit definition of the production function and, together with their accuracy, have the added
advantage that said function is clearly defined. In the case of non-parametric models, the efficient
frontier is estimated through available data, and requires a prior definition of the production
function. This approach endows non-parametric models with greater flexibility, and allows them to
be applied to situations involving multiple outputs. One additional advantage of these models is
their functionality, since they provide information concerning each study unit which may be used to
establish action guidelines for improving efficiency. Of the existing non-parametric models, Data
Envelopment Analysis (DEA) is the most widely employed to study the efficiency of institutions, and
has been used on numerous occasions to examine cultural institutions.
2
Our work belongs to this latter category of studies; specifically, we apply the DEA (Data Envelopment
Analysis) model to evaluate the museums that form part of the Spanish national network of
museums, considering a function with numerous inputs and outputs. The inputs in the model are
defined in terms of the different resources used. By contrast, the outputs reflect what the institution
offers and are measured in terms of the activities available and public consumption. With regard to
the former, the manager of the institution may act by combining the available factors more
efficiently to produce the service. Nevertheless, final public consumption is not solely defined by
what the institution offers but is also shaped by contextual variables, which is why our work posits
calculating institutions’ technical efficiency using two different models. The former considers those
output variables controlled by the managers whereas the latter takes account of output variables
which are beyond managers’ control and which are shaped by certain contextual variables.
Subsequently, an analysis is made of how and to what extent environmental variables may impact
on a museum’s efficiency. Finally, a more restrictive measure of efficiency is posited by introducing
the notion of cost efficiency, incorporating information concerning the mean price of the factors
used.
The study we present is structured in four parts. First, we offer a review of the previous studies
carried out into museums’ performance. We then introduce the problem to be addressed, offering a
description of the sample of museums used as a reference and describing the methodology. Finally,
the results obtained are presented and analysed together with the main conclusions.
2- Efficiency analysis in museums: formulation and state of the art
Cultural institutions in general, and museums in particular, may conduct their activity within the
public or private sector. When operating in the public sector, the goal of maximising profit loses
relevance compared to the social goals, whereas in the private sector there is justification for
establishing performance related goals. Although there is a growing tendency towards museums
becoming self-financing1, the activities they engage in (acquisition, conservation, study, staging
exhibitions, dissemination) coupled with the aims which determine their very creation (study,
education, recreation) make it difficult to draw comparisons with the quest for financial
performance.
In the business world, those contributing funds know to what extent objectives have been
accomplished when they are paid, whether through performance or capital gains. By contrast, both
in the case of museums operating in the public sector as well as those operating in the private
sector there is a separation between those who provide the funds and those who benefit from the
service, thus necessitating a specific study of the institution’s results in order to gauge to what
extent the objectives initially set out have been achieved. For this reason, efficiency studies prove
particularly relevant in the domain of museums; in the case of those operating in the public sector,
because resources are acquired in a competitive environment –given the existing budgetary
restrictions- and it would seem logical to demand efficiency; and in the case of private entities,
because financial donors seek a return for their contributions through the entity’s social impact.
The information which is compulsorily made available to the public is aimed at responding to the
need to inform concerning the efficiency of public institutions’ performance by including a report on
efficacy, efficiency and economic indicators that relate the resources used to the products obtained.
This line of approach includes the works of Ames (1994) and Jackson (1994) –who propose
drawing up tables of indicators to examine museums’ results- as well as the evaluations carried out
by the agencies responsible for issuing accreditation for museums2. This means of assessing
museums’ performance might prove useful in certain circumstances, yet it poses a restriction
stemming from the fact that an indicator reflects a partial aspect of each entity’s production,
1 By way of an example, in its action plan for 2013-2016 the Prado Museum envisages as one of its main lines of action in
terms of income “Enhancing the museum’s capacity to generate its own resources by implementing fresh initiatives”.
One example of museum accreditation programmes, specifically the one applied by the American
Association of Museums, may be found in Arroyo and Hart (2011).
2
3
whereas museums actually offer multiple products, in addition to which it proves complicated to
establish a hierarchy of institutions using this technique.
Another way of measuring museums’ efficiency is by calculating each museum’s distance from a
frontier made up of the best practices of a given group. In this case, it is necessary to define the
efficient frontier, for which parametric and non-parametric models may be used. Opting for one
model or another involves analysing the characteristics of the units to be evaluated. For this
purpose, it should be remembered that museums are institutions whose mission is quantified in
terms of achievements that prove difficult to measure, such as the conservation of heritage,
educating people or entertainment. Evaluating to what extent objectives are accomplished should
therefore be performed using proxy variables such as the number of temporary exhibitions staged,
the number of visitors, or the number of publications issued by the museum. Moreover, the lack of
any market which sets the prices of the products offered often leads to taking as a reference a
concept of efficiency which does not require valuations, such as technical efficiency, where
quantities of resources and products are linked without the need to allocate any value, as opposed
to the concept of allocative efficiency which seeks to pinpoint which combination of factors enables
the goal to be achieved at the lowest possible cost. Nevertheless, this latter concept of efficiency
provides valuable information by taking account of the cost at which units operate and not merely
the consumption of resources in real terms, thus also providing a more restrictive evaluation than
the notion of technical efficiency. On the other hand, the uncertainty surrounding the production
technology employed by these units makes the use of more flexible techniques seem more suited to
studying museums.
The inflexibility involved in defining the functional form of the frontier and the need to sum up in a
single output all the activities undertaken by these institutions has led to a somewhat limited use of
parametric models for studying museums. However, mention should be made of the works of
Jackson (1988) for a sample of North-American museums, and Bishop and Brand (2003) who
measure the efficiency of a group of British museums. In this instance, a simple production function
is estimated and leads to the conclusion that the amount of public funding received and volunteer
participation have a negative impact on mean efficiency in terms of visitor numbers.
Non-parametric techniques, particularly DEA and its offshoots, have been the most widely used to
study the performance of museums given the flexibility they afford when defining the frontier, and
because they adapt to the situation of museums by allowing the evaluation of a group of institutions
that employ different inputs to produce a series of outputs. Studies which opt for non-parametric
methods to evaluate museums include Paulus (1995), who examines the technical efficiency of a
group comprising 64 French museums. Analysis of the technical efficiency of a group of Belgian
museums is also the subject of the works by Mairesse (1997) and Mairesse and Vanden Eeckaut
(2002) where the technique applied is FDH (Free Disposal Hull), a variant of DEA that removes the
need for convexity in the frontier. The work of Taalas (1998) introduces the notion of allocative
efficiency for a sample of Finnish museums using the FDH model. Other works based on nonparametric models include Pignataro (2002) and Basso and Funari (2004). The former measures
the efficiency and technical change of Sicilian museums, whilst the latter introduces variable
returns to scale, decomposing the measurement of technical efficiency into a purely technical
component, and another of returns to scale for a sample of museums located in three major tourist
cities (Bologna, Florence and Venice). Finally, del Barrio, et al (2009) explore the efficiency of a
sample of museums in Spain based on a prior classification thereof using multivariate techniques,
and del Barrio and Herrero (2013) introduce a complex production function with multiple outputs to
evaluate the performance of museums belonging to an institutional network of museums in Spain.
3- Evaluation of state-run museums
3.1- Data
According to Spanish Ministry of Culture data through the Spanish Directory of Museums and
Collections, there are currently over 1500 museums in the country of which 1076 are publicly
4
owned compared to 481 that are privately owned. The group of publicly owned museums
encompasses institutions that cover a wide range of themes and that differ considerably in size. In
an effort to select a representative and even group that will allow us to apply the method proposed,
we have chosen a sample of museums run by the Ministry of Education, Culture and Sport. This is a
group of 89 museums involved in areas such as archaeology and fine arts, and which merges socalled national and provincial museums set up on the basis of royal collections as well as goods
seized from the church during disentailment, together with subsequent archaeological and art
collections. The national museums are run directly by the Ministry of Education, Culture and Sport
or, as is the case of the Prado, through a state-run independent body, whilst the provincial
museums, which house the main art collections and archaeological pieces from each area, are run
by the competent regional government through the corresponding regional Ministry of Culture.
TABLE 1: MUSEUMS IN THE SAMPLE
Museum of the National Library of Spain
Prado Museum
National Museum of Ceramics and Sumptuary Arts
National Sculpture Museum
Sephardic Museum
Lázaro Galdiano Museum
Álava Museum of Fine Arts
Museum of Almeria
Badajoz Provincial Museum of Archaeology
Burgos Museum
Caceres Museum
Cuenca Museum
Casa de los Tiros Museum in Granada
Úbeda Archaeological Museum
León Museum
Murcia Museum of Fine Arts
Palencia Museum
La Rioja Museum
Seville Museum of Fine Arts
Soria Numancia Museum
Tarragona National Archaeological Museum
Santa Cruz Museum
Valladolid Museum
Zamora Museum
From an initial group of 89 museums, those which had closed at the time the study was carried out
were removed from the sample. In addition, those which operated as a dependent section of
another museum were grouped into a single unit. So-called “house museums” were also removed
from the study since they were felt to constitute a separate study unit. These accounted for a group
of 50 museums. Based on this group of museums, useful data were finally obtained for a study of
24 museums, six of which were national and the rest provincial as can be seen in Table 1. In an
effort to avoid measurement errors as well as others resulting from extreme data, and in an attempt
to enhance the robustness of the subsequent statistical and econometric analysis, we posited a
dynamic efficiency analysis over four years so as to delimit a database panel in which each museum
is considered to be a decision unit each year. We finally obtained a total of 96 observations of 24
museums, over four years.
3.2- Methodology
Based on the homogeneous group of museums in the sample, we posited efficiency analysis using
DEA, which provides an individual efficiency index for each of the museums observed, by solving a
mathematical optimisation programme. For this purpose, we can define the index to measure
5
technical efficiency3 of museums such as the ratio of the weighted sum of the outputs obtained (y),
divided by the weighted sum of the inputs used (x). Considering that s inputs are used to obtain m
outputs, the efficiency index for museum j will be defined by the following expression:
hi =
u1iy1i + u2iy2i + ……………………. + usiysi
v1ix1i + v2iy2i + ……………………. + vmixmi
where u and v represent the weightings allocated to each input and output involved in the process.
DEA allows the most favourable group of weightings to be allocated to each DMU, in other words,
the one which allows the highest index for the value to be achieved, creating an efficient frontier.
Referring to the weighting vectors for the outputs and inputs as V and U, respectively, the group of
values allocated to each DMU is obtained by solving the input oriented mathematical program4
proposed by Charnes et al. (1978) (CCR model) as follows:
max U,V (U yi/Vxi)
subject to:
U yj/Vxi ≤1
ur, vi ≥ ξ
j = 1,..., n
r = 1,,…,s
i = 1,…,m
The model posits that no DMU may obtain a score of over 1 for the efficiency index with the
weightings obtained. If the value of the index is equal to 1, the unit analysed is deemed to be
efficient and is located at the frontier, whereas if the result is below 1, the unit analysed is not at
the efficient frontier, since there is another DMU able to achieve the same level of outputs using
fewer inputs. The second restriction posits the non-negativity of the weightings, or more strictly the
positivity of the weightings, with ξ being a non-Archimedean infinitesimal5.
Since this is a problem of fractional programming, it must be converted to a full form by introducing
a normalisation restriction for the inputs, with the problem being set out as follows:
max μ,ν μyj
subject to:
νx i = 1
μyj – νxj ≤ 0
μ, ν ≥ ξ
j = 1,..., n
Set out thus, the model measures the efficiency derived from the management or technical
efficiency, ignoring the possible existence of economies of scale that can affect the performance of
the DMUs. Nevertheless, it is assumed that an organisation’s efficiency is also conditioned by how it
is run, by the scale in which it operates. For this reason, we consider a two-fold technological
3 Technical efficiency entails obtaining a given level of outputs using as few inputs as possible or, viewed from
the other perspective, obtaining the maximum levels of outputs given a certain consumption of inputs.
Technical efficiency does not introduce cost data. Allocative efficiency takes account of the market prices of
the factors, seeking a combination of inputs that allows a target level of outputs to be reached at the lowest
possible cost.
4 Input oriented models seek the highest proportional reduction in the input vector for a given level of outputs.
By contrast, output oriented models posit the highest possible increase in outputs with a given level of inputs.
We opted for input orientation, which is the most commonly chosen when applying DEA.
5 Even though the initial CCR model posits the non-negativity of the weightings, we here consider a
subsequent modification put forward by Charnes et al. (1979) which demands strictly positive values so as to
prevent the solution from failing to take account of all the inputs and outputs when calculating the efficiency
index.
6
hypothesis: constant returns to scale and variable returns to scale. In this latter case, the CCR
model, posited assuming constant returns to scale, may be modified to embrace the possibility of
inefficiencies due to each unit’s different scales, by introducing a new restriction in the model:
n
∑ λj = 1
j=1
This new restriction introduced by Banker et al. (1984) (BCC model) may be expressed in the
following terms: e λ = 1, with e being a row vector made up of units.
To analyse the results obtained by applying this technique, the most common procedure is to use
the dual or enveloping formulation of this programme, expressed as follows:
min θ, λ θ
subject to:
-yi + Y λ ≥ 0
θ xi – X λ ≥ 0
eλ=1
λj ≥ 0,
j = 1,...,n
As conjectured in the model, a museum will be technically efficient when it is able to obtain a given
level of outputs with the lowest possible amount of inputs. If information is also available on the
prices of the factors, it will prove possible to formulate the measure of efficiency in terms of costs. A
unit will be cost-efficient if it is able to achieve a given level of outputs at the lowest possible cost. In
this case, the following DEA model of cost minimisation can be solved in order to determine the
cost-efficiency of the units studied:
min λ, x, i Wi Xi*
subject to:
-yi + Y λ ≥ 0
Xi* – X λ ≥ 0
eλ=1
λj ≥ 0,
j = 1,...,n
where Wi represents the given price vector for inputs and Xi* the vector of the amounts of inputs
that allow costs to be minimised with a given level of outputs yi. Efficiency in costs or overall
efficiency for museum i might be calculated using the ratio linking minimum cost and observed cost:
CE = Wi xi* / Wi xi
Total efficiency or cost efficiency (CE) merges the notion of technical efficiency (TE) with allocative
efficiency (AE). This latter term reflects a museum’s capacity to use inputs in an optimum proportion
for prices of given factors. An organisation might be technically efficient yet might not be efficient in
allocative terms due to its not using a combination of inputs that minimises costs given a price
vector. Allocative efficiency may be determined by equality:
CE = TE x AE
Studies into museum efficiency usually resort to the concept of technical efficiency given the
difficulty of having data available concerning the price of the factors. Nevertheless, there are
various alternatives for dealing with problem, and which range from the possibility of using mean
prices to introducing limits in the variations of the weightings allocated to the inputs in the optimum.
In our case, we opted to use mean estimated prices for the inputs, taking as a reference the works
7
of Morey et al (1990) and Bymes and Valmanis (1994) in the domain of hospitals. We chose to use
this method as a result of having available data for the accounts of a sample of these institutions
that can be linked to the costs of the inputs chosen in our model: staff expenses, spending on
outside services or repayment on equipment. We also observed that the inputs of these institutions
were not subject to major price fluctuations over the period analysed.
In order to put our study into practice, we selected the variables for the model. As measurements of
the inputs, we considered three variables that are representative of the resources used by the
museum: labour, museum size, and equipment. As for the variables that were representative of
output, we chose the number of exhibitions staged, publications, outreach and dissemination
activities undertaken by the institution and the number of visitors. It should be noted that whilst
managers can take decisions which affect the value of the output variables that measure the
museum’s various activities, the number of visitors variable is also determined by several external
variables such as the cultural background of the target population, which is beyond the manager’s
control. Furthermore, other contextual variables which might determine museums’ activities should
be taken into account such as the size of the population, accessibility, or the location’s cultural
potential. In this regard, a museum located in an area which has a strong appeal for tourists is likely
to attract more visitors than another located off the beaten tourist track, and over which the
manager is clearly powerless. As shown in Figure 1, and based on this consideration, we first
conducted a complex methodological approach to measure the technical efficiency of museums in
three stages. We first posited an input oriented DEA model (Managerial Model) in which three input
variables are considered (labour, museum size, and equipment) so as to obtain some of a
museum’s own outputs (exhibitions, publications and other activities such as lectures, concerts,
etc.). In this initial model, both the input and the output are under the manager’s control. The
efficiency index allocated to each museum at this stage would be a true reflection of that part of the
service production process for which each manager can really be deemed responsible, thus
enabling an even comparison to be drawn between the various museums analysed. At the second
stage, we consider that managers can influence the number of visitors by scheduling an appealing
programme of activities. We thus propose a second input oriented DEA model (Outcomes Model)
where we take as intermediate inputs the outputs from the previous stage that would make up what
the institution offers in cultural terms. This then allows us to obtain a single output measured
through the number of visitors to the museum. Finally, we consider that in order to secure an
accurate evaluation of the efficiency of the units analysed these need to be harmonised in terms of
inputs, outputs and the circumstances in which they operate. Were this not the case, a negative
evaluation of a DMU might be caused by the differing conditions in which it operates. In order to
gauge how and to what degree the contextual variables might affect the performance of the various
institutions, taking the results from the previous stages, at a third stage we measure the impact
which the contextual variables have on museums’ technical efficiency using a regression analysis in
which we take the previously calculated efficiency indexes as dependent variables, and the
contextual variables as independent variables. The regression residuals would reflect that part of
the efficiency which is not explained by the contextual variables and which, as a result, might be
attributed to how well each museum is run.
To conclude, we may consider that the ultimate goal these institutions pursue focuses on achieving
the greatest possible impact, measured in terms of visitors, and for which the cost at which they
operate proves relevant in the present. In this sense –as can be seen in Figure 1- we construct a
model to measure the overall efficiency of museums, distinguishing between technical efficiency
and allocative efficiency. The input variables taken in this latter part of the work are labour, size of
the museum, and equipment, as opposed to a single output variable which measures the number of
visitors to the museum. The prices of the factors have been calculated based on the available
accounting information in the museums contained in the sample for the study period as well as
other publicly available data. In the case of the labour variable, we calculated the mean cost per
employee based on the staffing costs data of the statements of accounts for the period analysed.
As regards the size of the museum, we calculated the mean square metre cost taking as a
reference the expenditure on services outside the accounts. Finally, to determine the mean cost of
the equipment, we used a report issued by the Federation of Municipalities and Provinces which
8
provides information concerning the usual equipment found in museums together with the annual
maintenance cost6.
FIGURE 1: METHODOLOGICAL FORMULATION
Table 2 shows the descriptive statistics of the group of variables considered in the study, both those
involved in the museum’s production function at each of the two previously mentioned stages and
the contextual variables used at the final stage to evaluate the efficiency of the museums in the
sample.
The values obtained when determining mean prices were mean annual cost per worker: 41; mean annual
cost per square metre: 0,2; mean annual cost of each piece of equipment 23 (in thousands of Euros). Given
the small variation observed in these, the same mean value for the four years was taken.
6
9
The data evidence a major difference with regard to the number of employees, with a standard
deviation of 108.59 out of an average number of 52 workers. The difference in size of the museum
measured in square metres is also significant, with a standard deviation of 12 896.25 for a mean
size of 6 855.88 m2. Less notable are the differences concerning the museum’s cultural offer
expressed in the number of exhibitions (mean 4.01 and standard deviation 3.24), publications
(mean 7.8 and standard deviation 11.48) and other activities (mean 141.71 and standard
deviation 228.49). It should be pointed out that the main differences between museums emerge
when counting the number of visitors: with a mean number of visitors per museum of 194 368.12
and a standard deviation of 539 606.67. This might support the idea that the number of visitors to
a museum is not only determined by what the museum offers in terms of culture but also –defined
by the activities carried out therein-, by another series of factors beyond the manager’s control and
which we take into account at the third stage of our analysis in the following terms: number of
protected heritage goods per inhabitant, number of hotels per inhabitant, whether the museum’s
management depends on the central or regional government, kilometres of motorway per
inhabitant, number of hotel places per inhabitant, available labour force and unemployment rate. All
of them seek to ascertain whether the level of cultural and tourist reputation of the various
museums’ location, the degree of urbanisation or accessibility or kind of management impact on
these institutions’ efficiency.
3.3- Results
The results from the first stage provide information concerning the technical efficiency of the
museums analysed from a range of different perspectives:
a) Stage 1: Managerial Model
The first stage of the efficiency analysis seeks to determine the efficient units in terms of
management. Applying the model to the available sample yields the results shown in Table 3. We
considered constant returns (CCR model) and variable returns to scale (BCC model), although our
conclusions were taken from the data obtained in the constant returns to scale model, since this
was felt to provide a measure of the total efficiency of museums with regard to achieving their
goals, added to which it is also the most commonly used guideline in efficiency analysis (Guccio et
al., 2014). Choosing a variable returns to scale model might yield a more favourable evaluation of
units located at the extremes as regards the scale of operations.
10
TABLE 3. MUSEUMS’ EFFICIENCY SCORES. MANAGERIAL MODEL
CCR POOL MODEL
National Library of Spain Museum
Prado Museum
González Martí National Museum of
Ceramics and Sumptuary Arts
National Sculpture Museum
Sephardic Museum
Lázaro Galdiano Museum
Álava Museum of Fine Arts
Almeria Museum
Badajoz Provincial Archaeological
Museum
Burgos Museum
Caceres Museum
Cuenca Museum
Casa de los Tiros Museum in Granada
Úbeda Archaeological Museum
Leon Museum
Murcia Museum of Fine Arts
Palencia Museum
La Rioja Museum
Seville Museum of Fine Arts
Soria Numancia Museum
Tarragona National Archaeological
Museum
Santa Cruz Museum
Valladolid Museum
Zamora Museum
Mean Eff.
No Mus Eff.
BCC POOL MODEL
2008
33.42
14.88
2009 2010 2011
27.01 37.33 50.66
9.34 73.17 84.35
Mean
37.11
45.44
2008 2009 2010 2011
42.55 38.69 46.14 51.08
23.96 12.66 86.65 100
Mean
44.62
55.82
100
51.59
5.2
25.03
69.67
21.31
100
30.73
11.12
21.37
56.96
42.17
100
43.78
18.8
31.27
44.59
28.36
52.79
47.88
20.88
35.56
69.75
44.2
88.20
43.50
14.00
28.31
60.24
34.01
100
61.5
68.75
29.67
97.87
33.66
92.08
59.55
65.84
33.01
92.83
41.91
16.83
10.6
89.08
100
81.63
100
43.47
76.83
3.29
87.72
100
3.55
16.13
30.99
71.14
100
100
100
43.4
100
23.92
85.78
64.09
7.11
15.19
5.05
45.47
16.94
50.44
100
36.23
100
42.93
98.68
86.81
5.75
13.26 15.35 66.04 65.48 65.37 64.71 65.40
4.57 12.80 41.51 49.48 41.86 41.65 43.63
33.85 59.89
100 71.63 50.83 47.35 67.45
48.35 66.32
100
100 71.23 87.2 89.61
71.57 75.91 88.02 100 65.16 79.61 83.20
100 100.00 100
100
100
100 100.00
62.5 46.40 60.96 60.42 58.19 78.24 64.45
100
94.21 77.94 100
100
100
94.49
57.87 32.00 57.99 57.89 68.52 80.16 66.14
98.68 92.72 93.98 98.33 100
100
98.08
39.71 72.65
100 65.41 86.85 43.45 73.93
17.88 8.57 44.57 46.07 41.24 42.15 43.51
89.96
85.53
33.55
95.74
55.8
4
85.34 100 79.54
42.77 63.88 55.9
94.99 100
100
81.71 99.32 100
56.1 56.0 57.9
5
4
4
88.71
62.02
82.14
94.19
56.4
1
100
100
68.35
96.22
73.1
7
100
30.98
69.81
28.72
91.37
45.69
100
48.4
63.56
35.29
84.01
39.69
68.33
97.32
61.24
38.37
98.08
48.6
86.49 100 79.57
55.93 67.98 62.79
98.36 100
100
87.23 99.67 100
69.2 71.7 73.7
5
6
6
91.52
71.68
91.68
95.78
71.9
1
The data show a mean level of technical efficiency of around 57% for the museums analysed, with a
small variation in range among the years considered. Broadly speaking, this would indicate that, on
average, the museums in the sample could achieve similar results if reducing the amount of
resources used by around 43%. It can be seen that only 20% of the museums in the sample
evidence efficiency in each of the years considered. Nevertheless, it should be highlighted that
there are significant differences with regard to efficiency in management in the museums analysed,
with the best results being obtained in the provincial museums run by regional governments and
located in middle sized cities. In sum, these are museums with sufficient inputs and which are
efficient in the outputs that are under the manager’s control, whereas the less efficient could be
considered oversized in terms of resources with regard to the activities they organise.
b) Stage 2: Outcomes model
At the second stage, an input oriented DEA model is again applied, adding the data obtained at the
previous stage as the input variables (number of exhibitions, publications and other activities), and
the number of visitors to the museum as the only output. The aim is to evaluate museums’ capacity
to attract visitors depending on the activities they organise. Table 4 shows the results of applying a
DEA model under constant returns and variable returns to scale, even though, again, we take the
results with constant returns to scale as a reference in order to avoid favouring the units that
operate at the extremes of the scales.
11
TABLE 4. MUSEUMS’ EFFICIENCY SCORES. OUTCOMES MODEL
CCR POOL MODEL
National Library of Spain Museum
Prado Museum
González Martí National Museum of
Ceramics and Sumptuary Arts
National Sculpture Museum
Sephardic Museum
Lázaro Galdiano Museum
Álava Museum of Fine Arts
Almeria Museum
Badajoz Provincial Archaeological
Museum
Burgos Museum
Caceres Museum
Cuenca Museum
Casa de los Tiros Museum in Granada
Úbeda Archaeological Museum
Leon Museum
Murcia Museum of Fine Arts
Palencia Museum
La Rioja Museum
Seville Museum of Fine Arts
Soria Numancia Museum
Tarragona National Archaeological
Museum
Santa Cruz Museum
Valladolid Museum
Zamora Museum
Mean Eff.
No Mus Eff.
BCC POOL MODEL
2008 2009 2010 2011
16,68 15,08 16,87 15,56
100
100 97,02 100
Mean
16,05
99,26
2008
100
100
2009
100
100
2010
100
97,09
2011
100
100
Mean
100,00
99,27
100 62,86 57,34 100
11,13 9,41 13,97 15,44
100 91,27 73,28 78,49
6,19 6,27
6,9
9,07
4,67 5,54 5,11 7,79
12,95 16,97 12,03 10,61
80,05
12,49
85,76
7,11
5,78
13,14
100
41,38
100
31,62
98,52
46,07
93,54
34,29
100
32,92
95
58,52
83,34 100
27,91 28,16
84,21
80
42,7 41,34
91,72 88,67
49,24 47,4
94,22
32,94
91,05
37,15
93,48
50,31
8,46
6,18
32,75
3,86
12,44
25,86
13,01
10,25
3,76
7,14
60,84
96,65
13,35
7,17
30,76
3,88
15,88
30,11
15,77
22,22
4,81
9,71
36,83
46,37
10,71
6,51
31,23
3,65
14,38
29,14
13,23
16,04
3,34
8,92
47,12
73,46
100
64,26
44,56
100
69,47
100
61,29
52,46
100
65,8
60,96
100
99,44
62,96
41,61
100
71,43
100
62,74
50,64
66,79
73,48
47,15
100
100
100
42,22
100
64,37
100
63,64
38,7
78,05
64,84
45,9
100
100
100
50,43
100
66,11
100
80,95
48,44
66,29
63,9
45
76,3
99,86
81,81
44,71
100,00
67,85
100,00
67,16
47,56
77,78
67,01
49,75
94,08
18,62 19,09 18,49 18,59
100
100 81,29 76,54
7,16 11,97 11,12 10,33
8,06
8,7
4,71 7,68
31,9 30,2 26,4 28,6
4
3
0
2
18,70
89,46
10,15
7,29
29,3
0
39,02
100
97,23
77,27
77,1
8
49,23 43,46 56,76
100 89,74 100
80,6
100 94,46
77,27 73,91 77,27
74,9 74,2 75,5
7
6
8
47,12
97,44
93,07
76,43
75,4
3
9,47
6,4
30,61
3,33
17,01
27,05
12,43
14,16
2,41
8,64
46,08
100
11,57
6,29
30,8
3,51
12,17
33,53
11,69
17,54
2,37
10,19
44,74
50,82
The results evidence low levels of efficiency when it comes to attracting visitors, with a mean level
of around 30% for the years analysed. Only one museum achieve above a 90% mean efficiency for
the years evaluated. The data show that the museums which reach the highest values are national
museums located in Madrid and in cities with the greatest tourist projection, such as Toledo,
Valencia and Seville. The data confirm that taking the number of visitors as the measure of output,
museums’ efficiency is determined by where they are located. Efficiency values below 10% generally
correspond to museums in towns with a smaller population or which compete with other major
tourist attractions to draw visitors.
GRAPH 1. Efficiency rates of Management Models of Museums (First and Second Stage)
Ubeda
100,00
Numantino
Burgos
80,00
Sefardi
60,00
Badajoz
40,00
Lazaro G
20,00
Murcia
Zamora
La Rioja
Tarragona
Valencia
0,00
Palencia
Valladolid
Almeria
Granada
Bteca
Sevilla
MNE
Prado
Cuenca
Sta Cruz
Alava
León
CCR First stage
CCR Second stage
Caceres
12
Graph 1 illustrates the results obtained in the first and second stages. It can be seen that while
appreciable levels of efficiency are obtained regarding the use of available resources to organise
the museums’ activities (management model), the level of efficiency when attracting visitors is
substantially lower. It should also be pointed out that there are differences in terms of which
museums achieve notable efficiency rates at the first and second stages, reinforcing the idea that
there are outputs in the museums which managers may control and which concern what the
institution offers in cultural terms. In contrast, in addition to the available resources, the number of
visitors to the museums is determined by other factors beyond the manager’s control, such as the
museum being located in a place with a greater or lower capacity to attract tourists, competing with
major tourist attractions to attract visitors, or even some museums’ star-like status which, thanks to
their quality or reputation in art terms, already attract large numbers of visitors.
MUSEUM
TABLE 5. POSSIBLE IMPROVEMENTS IN ACHIEVEMENT OF INPUTS AND OUTPUTS BY MUSEUM
MUSEUMS
EXHIBITIONS
SIZE
EQUIPMENT
EMPLOYMENT
ACTIVITIES
Gain (%)
Gain (%)
Gain (%)
Gain (%)
Gain (%)
M.M.
O.M.
M.M.
O.M.
M.M.
O.M.
M.M.
O.M.
M.M.
O.M.
National Library of Spain
Museum
Prado Museum
National Museum of Ceramics
and Sumptuary Arts
National Sculpture Museum
Sephardic Museum
Lázaro Galdiano Museum
Álava Museum of Fine Arts
Almeria Museum
Badajoz Provincial
Archaeological Museum
Burgos Museum
Caceres Museum
Cuenca Museum
Casa de los Tiros Museum in
Granada
Úbeda Archaeological Museum
Leon Museum
Murcia Museum of Fine Arts
Palencia Museum
La Rioja Museum
Seville Museum of Fine Arts
Soria Numancia Museum
Tarragona National
Archaeological Museum
Museum de Santa Cruz
Valladolid Museum
Zamora Museum
PUBLICATIONS
Gain (%)
M.M.
O.M.
-3,1
-29,2
-60.0
0
-62,9
-54,6
-91.7
-0.8
-66,6
-81,1
-84.0
-2.1
0,0
22,9
-100.0
-2.6
0,0
0,0
-96.8
-16.2
-8,8
-17,3
-60,4
-50,4
-46,9
-76,6
-32.2
-89.9
-5.9
-91.6
-97.8
-90.1
-56,5
-86,0
-71,7
-67,4
-66,0
-87.5
-19.7
-93.0
-96.8
-93.4
-11,8
-56,5
-86,0
-71,7
-39,8
-66,0
-20.0
-87.5
-14.2
-92.9
-94.2
-86.9
0,0
109,4
-37.7
-92.1
0,0
0,0
0,0
-98.3
-100.0
-97.9
0,0
0,0
0,0
172,9
0,0
207,7
-20.0
-94.4
-58.4
-92.9
-99.0
-89.1
-88,6
-85,9
-29,2
-11,2
-93.4
-95.0
-84.3
-94.7
-84,6
-87,2
-40,1
-38,8
-92.6
-97.2
-68.8
-97.8
-84,6
-87,2
-40,1
-33,7
-89.3
-93.5
-68.8
-96.4
0,0
-98.1
0,0
-99.3
3,3
18,1
272,6
0,0
-91.0
-94.6
-68.8
-99.5
-19,3
0,0
-70,3
-5,3
-78,1
-13,0
-26,3
-88,7
-92.6
-33.0
-97
-94.6
-99.1
-97.2
-81.8
-40.6
-24,1
0,0
-59,1
-5,8
-78,2
-33,6
-27,3
-91,4
-85.6
-70.9
-91.0
-85.2
-98.5
-95.0
-52.9
-43.8
-24,1
0,0
-53,6
-5,8
-68,0
-7,3
-27,3
-91,4
-85.6
-70.9
-86.8
-84.0
-96.8
-91.1
-52.9
-40.2
0,0
0,0
0,0
1,5
-99.7
-98.7
-100.0
-99.9
0,0
6,1
-100.0
-87.3
0,7
41,5
33,5
3,5
66,2
358,3
0,0
0,0
-97.1
-71.4
-93.6
-95.7
-97.2
-92.7
-91.0
-27.2
0,0 -81.2
-61,8 -11.5
-14,5 -95.1
-12,5 -95.4
M. M. = Managerial Model; O. M. = Outcomes Model
-11,3
-38,0
-17,9
-8,5
-81.3
-16.9
-89.9
-96.7
-11,3
-38,0
-17,9
-5,8
-81.3
-10.5
-89.9
-92.7
57,0
0,0
0,0
0,0
-98.1
-24.8
-98.0
-100
0,0
-97.2
0,0
0,0
0,0
-98.6
-99.6
0,0
0,0
-0,7
0,0
0,0
0,0
115,7
-98,6
-44,5
-98,4
-98,8
0,0
0,0
-94,2
-98,1
0,0
26,1
917,6
0,0
-96,2
-98,6
-93,1
-99,5
252,1
0,0
-99,0
-55,1
-96,0
To conclude, Table 5 shows the improvement which each inefficient museum must undergo if it is to
become efficient from the standpoint of either reducing inputs or increasing outputs. The first three
columns show the percentage reduction which the inputs consumed must achieve in order to reach
the frontier where the technically efficient units are located, while the latter columns display the
increase required by the outputs in percentage terms if they are to reach the efficient frontier. In the
case of the Prado Museum, for example, no improvements would emerge from a change in size
measured in square metres, whereas a 2.1% reduction in the number of employees and a 0.7%
reduction in equipment would enable it to reach the frontier of efficient institutions.
Stage 3: Analysis of conditioned efficiency
The final stage of the first part aims to gauge the impact of the contextual variables on museums’
technical efficiency. To achieve this, a regression is performed taking the previously obtained
13
efficiency indexes as dependent variables, and the contextual variables that make up the area
where the museum conducts its activities as the independent variables. To give an outline of the
framework in which the institutions studied conduct their activities, we consider four variables:
1- Cultural relevance, measured in terms of the number of heritage goods which enjoy special
protection from the authorities (cultural interest goods).
2- The location’s tourist capacity, measured through the number of hotels and hotel places.
3- Accessibility, measured through the number of kilometres of motorway.
4- Variables oriented towards measuring the size and dynamics of the city, such as the
population density, labour force participation and unemployment rate.
All the variables considered have been related to the number of inhabitants or the surface area of
the province where each museum is located.
TABLE 6. ANALYSING EXTERNAL FACTOR IMPACT ON MUSEUMS’ EFFICIENCY RATES. OLS REGRESSION
Variables
Heritage Sites
Hotels
Institution Management
Motorways
Beds-Hotels
Labour force
Unemployment
Constant
No. of Observations
R-Square
Coefficient
4.639.131
-2.898.026
1.586.976
-0.109240
-0.152085
0.956153
-405392.0
0.869534
Std. Error
1.754.065
9.687.677
8.897.630
1.266.175
0.087830
0.673301
256561.2
3.329.096
100
0.20
Coefficient
Std. Error
3155082**
1.450.740
-15.67685**
8.068.568
2219890*
8.454.645
3.478.697
9.398.487
100
0.14
Note: Statistically significant at 1% level; ** statistically significant at 5% level; *** statistically significant at
10% level
We proposed a regression analysis using ordinary least squares (OLS) for a combination of
transversal data and time series. In other words, in an effort to gain significance in the estimations,
we considered each museum for each year as a decision unit whose efficiency we are seeking to
evaluate, and we obtain a total of 96 observations for 24 museums over a four-year period. The
results of this analysis are shown in Table 6 and may be summed up as follows:
a) There is no relation between the size of the urban area and the museum’s level of efficiency
when attracting visitors.
b) The availability of hotel places in the urban nucleus where the museum is located is not
determinant of each institution’s capacity to attract visitors.
c) By contrast, there does exist a link between the cultural potential of the location and the
surrounding area, and vis-à-vis how efficient the museum is at attracting visitors. As regards
how the museum is run, this relation is once again highlighted particularly through
museums managed by the Ministry of Culture.
To conclude, we consider that a museum’s goal is to have a social impact in terms of improvements
in education, fostering a cultural identity, and designing an appealing offer that can attract tourists,
etc. One common way of gauging how successfully these goals have been accomplished is through
the number of visitors. In order to ascertain to what degree the funds dedicated to this purpose
have been used efficiently, we devised a model for evaluating cost-efficiency which takes the
number of employees, the size of the museum and the available equipment in each institution as
input variables, and the number of visitors as the output variable. We introduced a mean price
vector to pinpoint technical and allocative inefficiencies. Technically efficient units can use different
combinations of inputs to achieve a given level of output, although not all of them will do so at the
same cost, given the prices of the factors. Table 7 provides information concerning the results
obtained using this model. Data concerning overall efficiency or cost-efficiency show low mean
values (14.8%), with most of the inefficiency being due to technical reasons (26.2% mean efficiency
14
in the period studied) rather than inefficiency owing to service provision costs (66% mean efficiency
in the period studied).
Graph 2 also shows that, on the whole, provincial museums tend to be more efficient in allocative
terms than national museums. These results would seem to indicate that provincial museums are
able to reach a certain level of impact in their action (measured in the number of visitors) operating
at a lower cost than national museums. It should not be forgotten, however, that the proposed
method entails limitations, one of which concerns not taking into consideration data relating to the
quality of the services and facilities provided, and only taking account of the amounts offered.
Nevertheless, we could consider that visitors bear in mind qualitative factors when taking the
decision to visit the museum. In this case, attracting visitors is achieved at a generally lower cost in
provincial museums.
G RAP H 2. Cos t E f f i c i en c y rat es , Al l oc at i ve E f f i c i en c y Rat es an d
Thec n i c al E f f i c i en c y Rat es
Sefardi
150,00
Valencia
Cuenca
Sta Cruz
Bteca
Prado
Ubeda
100,00
Sevilla
Tarragona
Numantino
50,00
Palencia
0,00
León
Burgos
Zamora
Lazaro G
MNE
Caceres
Badajoz
Valladolid
Almeria
Granada
Alava
Murcia
La Rioja
Mean CCR ET
Mean CCR TT
Mean CCR AS
15
TABLE 7. COST-EFFICIENCY
TECHNICAL EFFICIENCY CCR POOL MODEL
National Library of Spain Museum
Prado Museum
National Museum of Ceramics and Sumptuary Arts
National Sculpture Museum
Sephardic Museum
Lázaro Galdiano Museum
Álava Museum of Fine Arts
Almeria Museum
Badajoz Provincial Archaeological Museum
Burgos Museum
Caceres Museum
Cuenca Museum
Casa de los Tiros Museum in Granada
Úbeda Archaeological Museum
Leon Museum
Murcia Museum of Fine Arts
Palencia Museum
La Rioja Museum
Seville Museum of Fine Arts
Soria Numancia Museum
Tarragona National Archaeological Museum
Santa Cruz Museum
Valladolid Museum
Zamora Museum
Mean Eff.
COST-EFFICIENCY CCR POOL MODEL
ALLOCATIVE EFFICIENCY CCR POOL MODEL
2008
2009
2010
2011
Mean
2008
2009
2010
2011
Mean
2008
2009
2010
10.51
99.21
100
11.13
100
6.19
4.67
8.3
6.22
3.4
32.73
3.86
12.44
25.86
13.01
10.25
1.82
7.11
60.78
48.22
18.62
78.05
7.16
4.37
28.1
9.5
98.04
62.03
9.41
91.27
6.26
5.54
8.29
9.47
3.47
30.61
3.33
17.01
27.05
12.43
14.16
2.41
8.64
46.07
99.8
19.09
74.21
11.97
4.71
28.1
10.62
92.63
57.05
13.97
73.28
6.28
5.11
7.73
8.5
3.41
30.79
3.51
12.17
33.53
11.69
17.54
2.37
10.19
44.73
40.15
18.49
58.93
11.12
4.12
24.1
9.8
100
56.73
15.44
78.49
8.56
7.79
6.96
8.68
3.88
30.75
3.88
15.88
30.08
14.32
22.22
4.81
9.71
36.83
40.98
18.59
50.88
10.33
4.16
24.6
10.11
97.47
68.95
12.49
85.76
6.82
5.78
7.82
8.22
3.54
31.22
3.65
14.38
29.13
12.86
16.04
2.85
8.91
47.10
57.29
18.70
65.52
10.15
4.34
26.2
10
33.45
14.37
6.35
100
4.64
3.26
5.21
3.69
2.8
22.59
3.79
9.04
22.08
7.74
6.65
1.07
4.71
25.41
25.55
15.73
22.29
4.32
3.42
14.9
9.04
33.01
13.61
5.38
91.27
4.69
3.94
5.17
5.63
2.85
22.46
3.27
12.36
23.1
7.4
9.28
1.39
5.36
20.24
52.87
16.13
21.2
7.22
3.69
15.9
10.11
30.62
12.6
8.19
74.55
4.75
3.57
4.95
5.05
2.8
22.83
3.3
8.45
28.63
6.78
13.43
1.32
6.74
19.51
23.25
15.62
18.67
6.71
3.18
14.0
9.32
33.57
12.53
9.81
78.28
6.41
5.44
4.43
5.16
3.19
23.04
3.64
11.19
25.69
7.59
15.6
2.68
6.26
16.1
23.96
15.7
16.12
6.23
3.17
14.4
9.62
32.66
13.28
7.43
86.03
5.12
4.05
4.94
4.88
2.91
22.73
3.50
10.26
24.88
7.38
11.24
1.62
5.77
20.32
31.41
15.80
19.57
6.12
3.37
14.8
95.14
33.72
14.37
57.05
100
74.92
69.84
62.76
59.44
82.23
69.04
98.29
72.68
85.39
59.51
64.84
58.98
66.17
41.81
52.98
84.48
28.56
60.33
78.2
65.4
95.14
33.67
21.94
57.15
100
74.92
71.04
62.35
59.44
82.2
73.37
98.29
72.68
85.39
59.51
65.54
57.68
62.01
43.94
52.98
84.48
28.56
60.33
78.2
65.9
95.14
33.06
22.09
58.63
101.73
75.54
69.84
64.04
59.44
82.2
74.14
94.04
69.46
85.39
57.96
76.6
55.74
66.17
43.61
57.91
84.48
31.68
60.33
77.2
66.5
2011
95.14
33.57
22.09
63.52
99.72
74.92
69.84
63.6
59.44
82.2
74.94
94.04
70.47
85.39
53
70.23
55.74
64.53
43.71
58.45
84.48
31.68
60.33
76.22
66.1
16
Mean
95.14
33.51
20.12
59.09
100.36
75.08
70.14
63.19
59.44
82.21
72.87
96.17
71.32
85.39
57.50
69.30
57.04
64.72
43.27
55.58
84.48
30.12
60.33
77.46
66.0
4- Conclusions
In this work, we present a methodological approach to analyse the efficiency of a
homogeneous group of museums that use a series of inputs (labour, equipment, size) to
achieve a set of outputs (exhibitions, publications, visitors). Given the multiplicity of inputs
and outputs, and the difficulty in defining the production function in these institutions, we
opted to use non-parametric DEA models to evaluate their performance. In addition, the
methodological formulation has taken into account that there are outputs which are under
the manager’s control –such as the number of exhibitions and publications- as opposed to
others which are determined by contextual variables –as is the case of the number of
visitors-. Said reflection has led us to posit a twin model to study museums’ technical
efficiency. At the first stage, data on exhibitions, publications and other external activities
organised by the museum have been taken as outputs, which in turn act as input variables
for the following stage where performance is measured in terms of visitors to the museum.
We have also sought to ascertain to what extent the results achieved are influenced by
contextual variables. Finally, we put forward a model to calculate museums’ cost-efficiency,
taking as values the estimated mean cost prices of the factors obtained from the data in
the museums’ annual accounts.
The main results to emerge from our research are as follows. First, most museums achieve
a significant level of technical efficiency in the model which aims to measure performance
in terms of management. In this case, it is particularly the provincial museums which reach
the highest levels of efficient management. The second model indicates that few museums
achieve high levels of efficiency when it comes to attracting visitors, with national museums
proving more efficient in this case, since they are able to draw a larger number of visitors
than provincial museums. It should be taken into account that national museums tend to
be located in cities that display a high capacity to attract tourists. In this regard, we find
contextual variables which can determine the levels of efficiency attained and which are
related to the tourist attractions of the area in which the museum is located. As regards
cost-efficiency, it can be said that provincial museums attain higher allocative efficiency
levels than national museums since they are able to draw visitors using fewer resources
than national museums.
To conclude, it can be said that there is a group of museums whose efforts to attract
visitors are restricted by the available resources. Although it is true that they arouse only
limited public interest, these museums operate at a proportionally lower cost when
attracting visitors. In contrast to this group, there are other museums which might be
termed “stars” and which are able to draw large numbers of people, yet, which consume a
larger amount of resources. In this latter instance, it would prove interesting to see what
part of the resources consumed by these museums are generated by the institution itself
and what part correspond to contributions from the state through subsidies and support,
since said capacity to attract should be reflected in an increased ability to generate selffunding.
BIBLIOGRAFÍA
Ames, P., 1994. Measuring museums merits. In K Moore (ed): Museum management.
Routledge, London.
Arroyo, L., and Hart, J., 2011. Museum Accreditation Program. Taylor & Francis. USA.
Banker, R.D., A. Charners, and Cooper, W., 1984. Some Models for Estimating Technical
and Scale Inefficiencies in Data Envelopment Analysis, Management Science, 30 (9),
1078-1092.
Banker, R.D. and Maindiratta, A., 1988. Nonparametric analysis of technical and allocative
efficiencies in production. Econometrica, 56-6, pp 1315-1332.
Basso, A. and Funari, S., 2004. A Quantitative Approach to Evaluate the Relative Efficiency
of Museums, Journal of Cultural Economics, 28 (3), 195-216.
17
Bishop, P. and Brand, S., 2003. The Efficiency of Museums: A Stochastic Frontier
Production Function Approach, Applied Economics, 35 (17), 1853-1858.
Boyle, S., 2007. Impact of changes in organisational structure on selected key performance
indicators for cultural organisations, International Journal of Cultural Policy, 13 (3), 319334.
Byrnes, P. and Valdmanis, V., 1994. Analyzing Technical and Allocative Efficiency of
Hospitals. In Charnes A, et al. Data Envelopment Analysis: Theory, Methodology and
Application. Kluwer Academic, pp 129-44.
Camanho, A.S. and Dyson, R.G., 2003. Cost efficiency measurement with price uncertainty:
a DEA application to bank branch assessments. European Journal of Operational
Research, 161 (2005), pp 432-446.
Carnegie, G.D. and Wolnizer, P. W., 1996. Enabling accountability in museums. Accounting,
Auditing & Accountability Journal 9:5 , 84-99.
Castro, M.F. Rizzo, I., 2009. Performance measurement of heritage conservation. Activity in
Sicilia. International Journalof Arts Management; Winter 11, 29-40.
Charners, A., W. Cooper and Rhodes, E., 1978. Measuring the Efficiency of Decision Making
Units, European Journal of Operational Research, 2, 429-444.
Cooper, W.W., Seiford, L.M. and Tone, K., 2007. Data Envelopment Analysis. A
comprehensive text with models, applications and references and DEA-Solver Software.
Springer. USA.
Cooper, W.W., Thompson, R.G. and Thrall, R.M., 1996. Introduction: Extensions and new
developments in DEA. Annals of Operations Research, 66(1996), pp 3-45.
Cuccia, T., Guccio, C., Rizzo, I. 2013. Does Unesco inscription affect the performance of
tourism destinations? A regional perspective, ACEI working paper series, AWP-04-2013.
Daraio, C., Simar, L. 2005. Introducing Environmental Variables in Nonparametric Frontier
Models: A Probabilistic Approach, Journal of Productivity Analysis, 24, (1), pp 93-121.
De Witte, K. and Geys, B., 2011. Evaluating Efficient Public Good Provision: Theory and
Evidence from a Generalised Conditional Efficiency Model for Public Libraries, Journal of
Urban Economics, 69 (3), 319-327.
Del Barrio, M. J., Herrero, L.C. and Sanz, J. A., 2009. Measuring the Efficiency of Heritage
Institutions: A Case Study of a Regional System of Museums in Spain, Journal of Cultural
Heritage, 10 (2), 258-268.
Del Barrio, M. J. and Herrero, L. C., 2013, ‘Evaluating the efficiency of museums using
multiple outputs: evidence from a regional system of museums in Spain’, International
Journal of Cultural Policy, http://dx.doi.org/10.1080/ 10286632.2013.764290.
Evans, P., 1997. Local Authorities and the Arts in the United Kingdom: The Development of
Policy-Specific Evaluative Indicators, International Journal of Cultural Policy, 4 (1), 161181.
Farrell, M. J., 1957. The Measurement of Productive Efficiency, Journal of the Royal
Statistical Society. Series A (General), 120 (3), 253-290.
Fazioli, R. and Filippini, M., 1997. Cost Structure and Product Mix of Local Public Theatres,
Journal of Cultural Economics, 21 (1), 77-86.
Fernández-Blanco, V., L. C. Herrero and Prieto-Rodríguez, J., 2013. Performance of Cultural
Heritage Institutions, In: I. Rizzo and A. Mignosa, eds. Handbook on Economics of
Cultural Heritage, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing
Ltd.
Finocchiaro Castro, M., C. Guccio and Rizzo, I., 2011. Public Intervention on Heritage
Conservation and Determinants of Heritage Authorities Performance: A semi-Parametric
Analysis, International Tax and Public Finance, 18 (1), 1-16.
Finocchiaro Castro, M. and Rizzo, I., 2009. Performance Measurement of Heritage
Conservation Activity in Sicily, International Journal of Arts Management, 11 (2), 29-41.
Frey, B., 1998. Superstar Museums: An Economics Analysis, Journal of Cultural Economics,
22 (2/3), 113-125.
Ganley, J. A. and Cubbin, J., 1992. Public Sector Efficiency Measurement: Applications of
Data Envelopment Analysis, Elsevier Science Publishers.
18
Gapinski, J. H., 1980. The Production of Culture, The Review of Economics and Statistics,
62 (4), 578-586.
Gilhespy, I. 1999, Measuring the performance of cultural organizations: A model.
International Journal of Arts Management, 2 (1), pp 38-52.
Goñi Legaz, S., 1998. El análisis envolvente de datos como sistema de evaluación de la
eficiencia técnica de las organizaciones del sector público: aplicación en los equipos de
atención primaria, Revista Española de Financiación y Contabilidad, 27 (97), 979-1004.
Guccio, C., Pignataro, G., Rizzo, I., 2014. Evaluating the performance of public procurement
contracts for cultural heritage conservation works in Italy, Journal of Cultural Economics,
38, (1), pp 43-70.
Jackson, J.M., 1994. Performance indicators: Promises and pitfalls. In K Moore (ed):
Museum management. Routledge, London.
Jackson, R., 1988. A Museum Cost Function, Journal of Cultural Economics, 12 (1), 41-50.
Lange, M., J. Bullard, W. Luksetich and Jacobs, P., 1985. Cost Functions for Symphony
Orchestras, Journal of Cultural Economics, 9 (2), 71-85.
Last, A. K. and Wetzel, H., 2010. The Efficiency of German Public Theaters: A Stochastic
Frontier Analysis Approach, Journal of Cultural Economics, 34 (2), 89-110.
Leibenstein, H., 1966. Allocative Efficiency vs. X-Efficiency. The American Economic Review,
56(3), pp 392-415.
Mairesse, F. and Vanden Eeckaut, P., 2002. Museum Assessment and Fdh Technology:
Towards a Global Approach, Journal of Cultural Economics, 26 (4), 261-286.
Mairesse, P., 1997. Les Limites D'analyse D'efficacité Dans Le Secterur Des Musées, Paper
presented at the Fourth AIMAC Conference, Conference, San Francisco.
Morey, R.C., Fine, D.J. and Loree, S.W., 1990. Comparing the allocative efficiencies of
hospitals. Omega, 1990, 18(1), pp 71-83.
Park, K.S. and Cho, J-W., 2011. Pro-efficiency: data speak more than technical efficiency.
European Journal of Operational Research. 215 (2011), pp 301-308.
Paulus, O., 1995. Museums’ Efficiency, Paper presented at the Fourth European Workshop
on Efficiency and Productivity Analysis, Conference, CORE, Université Catholique de
Louvain.
Paulus, O., 2009. Measuring museums performance: A study of museums in France and
the United States. International Journal of Arts and Management. 6:1, 50-63.
Peacock, A. T., 2003. Indicatori Di Performance E Politica Cultural, Economia della Cultura,
13 (4), 513-524.
Pignataro, G., 2002. Measuring the Efficiency of Museums: A Case Study in Sicily, In: I.
Rizzo and R. Towse, eds., The Economics of Heritage. A Study in the Political Economy of
Culture in Sicily, Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing
Ltd, pp. 65-78.
Pignataro, G., 2003. Performance Indicators, In R. Towse ed. A Handbook of Cultural
Economics, Cheltenham, UK and Northampton, MA, USA: Edward Elgar, pp. 366-372.
Puig-Junoy, J., 2000. Partitioning input cost efficiency into allocative and technical
components: an empirical DEA application to hospitals. Socio-Economic Planning
Sciences, 34 (2000), pp 199-218.
Reid, M. M., 2011. Is the balanced scorecard right for academic libraries? The Bottom Line:
Managing Library Finances, 24(2), pp 85-95.
Roll, Y., Cook, W.D. and Golany, B, 1991. Controlling factor Weights in Data Envelopment
Analysis. IIE Transactions, 23-1, pp 2-9.
Roll, Y. and Golany, B. 1993. Alternate methods of treating factor weights in DEA. Omega,
21-1, pp 99-109.
Seiford, L.M. and Thrall, R.M. 1990. Recent Developments in DEA. The mathematical
programming approach to frontier analysis. Journal of Econometrics, 46(1990), pp 7-38.
Simar, L., Wilson, P. W., 2007. Estimation and inference in two-stage, semiparametric
models of production processes, Journal of Econometrics, 136, pp 31-64.
Taalas, M., 1997. Generalised Cost Functions for Producers of Performing Arts – Allocative
Inefficiencies and Scale Economies in Theatres, Journal of Cultural Economics, 21 (4),
335-353.
19
Taalas, M., 1998. Efficiency of Finnish Museums - Free Disposal Hull Method to Measure
Cost Efficiency, Paper presented at the Tenth International Conference on Cultural
Economics, Conference, Barcelona, Spain.
Thanassoulis, E., 2001. Introduction to the theory and Application of Data Envelopment
Analysis. Kluwer Academic Publishers. USA.
Thompson, R.G., Dharmapala; P.S., Humphrey, D.B., Taylor, W.M. and Thrall, R.M., 1996.
Computing DEA/AR efficiency and profit ratio measures with an illustrative bank
application. Annals of Operations Research, 68(1996), pp 303-327.
Thompson, R.G., Langemier, L.N., Lee, C-T., Lee, E. and Thrall, R.M. 1990. The role of
multiplier bounds in efficiency analysis with application to Kansas farming. Journal of
Econometrics, 46 (1990), pp 93-108.
Turbide, J. and Laurin, C., 2009. Performance Measurement in the Arts Sector: the Case of
the Performing Arts, International Journal of Arts management, 11 (2), 56-95.
Uri, N.D. 2001. Technical efficiency, allocative efficiency, and the implementation of a price
cap plan in telecommunications in the United States. Journal of Applied Economics, IV-1,
pp 163-186.
Vitaliano, D. F., 1998. Assessing Public Library Efficiency Using Data Envelopment Analysis,
Annals of Public and Cooperative Economics, 69 (1), 107-122.
Weil, S. E., 1995. Performance Indicators for Museums: Progress Report from Wintergreen,
The Journal of Arts Management, Law, and Society, 23 (4), 341-351.
Weinstein, L. and Bukovinsky, D., 2009. Use of the Balanced Scorecard and Performance
Metrics to Achieve Operational and Strategic Alignment in Arts and Culture Not-forProfits, International Journal of Arts Management, 11 (2), 42-56.
Zieba, M., 2011. An Analysis of Technical Efficiency and Efficiency Factors for Austrian and
Swiss Non-Profit Theatres, Swiss Journal of Economics and Statistics, 147 (2), 233-274.
20