at SciVerse ScienceDirect

Energy 49 (2013) 444e458

Contents lists available

Energy

journal homepage: www.elsevier .com/locate/energy

Empirical investigation of energy efficiency barriers in Italian manufacturingSMEs

Andrea Trianni a,*, Enrico Cagno a, Ernst Worrell b, Giacomo Pugliese a

aDepartment of Management, Economics & Industrial Engineering, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milan, ItalybCopernicus Institute of Sustainable Development, Utrecht University, 3584 CS Utrecht, The Netherlands

a r t i c l e i n f o

Article history:Received 11 June 2012Received in revised form2 October 2012Accepted 7 October 2012Available online 11 November 2012

Keywords:Industrial energy efficiencyBarriersSmall and medium-sized enterprises

* Corresponding author. Fax: þ39 (0)2 2399 4067.E-mail address: andrea.trianni@polimi.it (A. Triann

0360-5442/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.energy.2012.10.012

a b s t r a c t

The paper identifies and evaluates barriers to industrial energy efficiency through the investigation of 48manufacturing Small and Medium-sized Enterprises (SMEs) in Northern Italy. The research providesinteresting suggestions both for enterprises and energy policy-makers. Firstly, economic and informationbarriers are perceived as the major obstacles to the adoption of energy-efficient technologies, whilstbehavioural barriers do not seem to affect enterprises very much. Nonetheless, despite what declared,the most relevant barriers are the lack of interest in energy efficiency and the existence of other prior-ities, thus showing that decision-makers tend to downgrade energy efficiency to a marginal issue.Furthermore, perceived barriers do not take place exclusively in implementing energy-efficient tech-nologies, but, with comparable importance, also in generating the interest and knowledge of theopportunities. Moreover, the study highlights that relevant differences can be appreciated for bothperceived and real barriers even among SMEs, that thus should not be bundled together. In addition tothat, other factors affect barriers, stimulating future research: indeed, lower real barriers can be observedwith higher complexity of the production, high variability of the demand and strong competitors.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Despite the strong effort of energy policies in Europe, the targetof reducing the energy consumption of 20% by year 2020 seemsunachievable. Indeed, recent estimates revealed that, with currenttrends, only 10% of the reduction will be obtained, on the one sideshowing the existence of barriers towards the diffusion of energy-efficient technologies, on the other side moving the EuropeanCommission towards a new set of measures to increase energyefficiency [1].

The industrial sector plays amajor role, as, according to themostrecent estimates [2], covers more than 50% of the total energydelivered. In this regard, reducing energy consumption in thissector should be seen as strategic by policy-makers to achieveenergy efficiency targets.

Within the industrial sector, according to a recent study by theEuropean Commission [3], when looking at Small and Medium-sized Enterprises (SMEs), “the picture is surprisingly unfav-ourable: close to two thirds of SMEs operating in the EuropeanUnion (EU) do not even have simple rules or devices for energy

i).

All rights reserved.

saving (63%). Less than three out of 10 SMEs (29%) have institutedsome measures to preserve energy and resources at their enter-prise. Only 4% of EU SMEs have a comprehensive system in place forenergy efficiency”. The Italian picture is even worse, with 74% ofSMEs neglecting to adopt measures for energy saving. This acquireseven greater importance if compared with the same statistics forLarge Enterprises: indeed, 29.8% of them do not have simple rulesfor energy saving, and 18.9% of them adopt a comprehensive systemfor energy efficiency. Nonetheless, SMEs are crucial also from thepoint of view of industrial energy consumption: indeed, as anexample, according to the most recent estimations in Italy, SMEscover about 60% of the Italian industrial energy consumption [4].Moreover, as reported by a recent investigation in Europe, SMEshave generated in the last decade about 85% of new job opportu-nities and about 60% of the European Gross Domestic Product [5],thus revealing to be really crucial for the future of the Europeanindustrial sector as a whole.

Hence, SMEs are strategic since they represent the largemajority of the enterprises, cover a major share of the domesticindustrial consumption, and are quite inefficient [6e9]. One reasonfor the low adoption of energy-efficient technologies within SMEsis represented by the lack of proper means to address their barrierstowards energy efficiency.

Table 1The taxonomy adopted in this study, with a clear distinction of the origin (external,or internal, with respect to the firm), and the actors affected by the barriers. Source:Cagno et al. [17].

Origin Actor/Area Barriers

External Market Energy prices distortionLow diffusion of technologiesLow diffusion of informationMarket risksDifficulty in Gathering External Skills

Government/politics Lack of proper regulationDistortion in fiscal policies

Technology/servicessuppliers

Lack of interest in energy efficiencyTechnology suppliers not updatedScarce communication skills

Designers andmanufacturers

Technical characteristics not adequateHigh initial costs

Energy suppliers Scarce communication skillsDistortion in energy policiesLack of interest in energy efficiency

Capital suppliers Cost for investing capital availabilityDifficulty in identifying the quality ofthe investments

Internal Economic Low capital availabilityHidden costsIntervention-related risks

Behavioural Lack of interest in energy efficiencyinterventionsOther prioritiesInertiaImperfect evaluation criteriaLack of sharing the objectives

Organisational Low status of energy efficiencyDivergent interestsComplex decision chainLack of timeLack of internal control

Barriers related tocompetences

Identifying the inefficienciesImplementing the interventions

Awareness Lack of awareness or Ignorance

1 The Lombardy Region has almost 10 million inhabitants and is located atNorthern Italy, 800 thousand enterprises and 1.2 million employees in themanufacturing activities. The Gross Product for the region is 296 billion euros, rep-resenting about 25% of the total Gross National Product, and being þ29% and þ37%higher of respectively the National and EU-25 per capita gross product [20].

A. Trianni et al. / Energy 49 (2013) 444e458 445

At the moment we can find very few contributions in the liter-ature addressing barriers to energy efficiency specifically inindustrial SMEs. Nonetheless, understanding the barriers to energyefficiency seems to be crucial to propose the most effective meansto overcome them. In order to do this, a holistic approach to barriersto industrial energy efficiency is needed, i.e. having a taxonomyable to understand and classify the wide spectrum of issues thatenterprises have to deal with when deciding to perform energyefficiency investments.

In the literature it is possible to find a large number of contri-butions, providing different perspectives on the taxonomies onbarriers to industrial energy efficiency, and showing that the debateis still open [10e16]. Nonetheless, most of the studies tend to lookat the barriers more from a theoretical viewpoint, rather than froman empirical one, i.e. by the enterprises’ perspective.

In recent studies Cagno et al. [17,18] have proposed an innova-tive taxonomy encompassing the major contributions in theprevious literature, but, at the same time, providing useful insightsto empirically investigate the barriers to industrial energy effi-ciency. Indeed, the new taxonomy aims at contributing to fulfil theneed of having a tool e in terms of theories and practices e usefulboth for enterprises and policy makers in order to clearly point outwhere the difficulties are rooted. In this study we adopt theirapproach investigating the new taxonomy in manufacturing SMEslocated in Northern Italy.

The structure of the paper is as follows: in Section 2 we describethe theoretical framework that represents the starting point for thisresearch. This sectionwill performabrief reviewof the characteristicsof the taxonomy adopted, pointing out the issues emerging for theapplication to SMEs and considering the transferability of thetaxonomy to SMEs. In Section 3 we describe the methodologyadopted to empirically investigate the barriers. Section 4 and5will bedevoted, respectively, to the presentation and discussion of results.

2. Theoretical approach

Cagno et al. [17] propose a taxonomy in which the barriers areclassified according to the responsible actor in which they origi-nate, as reported in Table 1.

Considering the need to empirically investigate the barriersamong enterprises, Cagno et al. have developed a taxonomy forfield investigation [17], as reported in Table 2, in which they pointout the origin of the barrier, that, with respect to the enterprise,might be either internal or external. In fact, it is apparent that theenterprise is not able to evaluate the real value of the externalbarriers, e.g. is not able to evaluate the value of the Lack of Interestfor Energy Efficiency for technology suppliers, but rather theenterprise will provide how this barrier reflects on itself, i.e. theperception that energy-efficient technologies are not available.

Moreover, in the studies of Cagno et al. [17], the taxonomyproposed is capable of evaluating barriers on the decision-makingprocess of the enterprise, attributing each of them to one or moreactions of the process, respectively the Generation of Interest(Action 1), Knowledge of inefficiencies and opportunities (Action2), Investment Analysis and Intervention Implementation (Action3), as reported in Table 2.

A third feature of the proposed taxonomy is the classification ofthe barriers with respect to their effect on the enterprise (the so-called spectrum of influence on the enterprise), i.e. affecting anyinvestment, only energy efficiency investments, or even dependingon the specific investment in energy efficiency under consideration.This classification allows to understand that some barriers varyaccording to the considered technology: as a consequence, it willnot be possible to obtain a real value, limiting the research toobtaining a general perception of those barriers.

Nonetheless, when considering Small and Medium-sizedEnterprises (SMEs), previous studies have pointed out that thestructure of those enterprises, quite different from Large Enter-prises (LEs), also affect empirical investigation [6,9,18]. Indeed, inthe case of SMEs the organizational structure is very simple, andquite often the decision belongs exclusively to a single person, i.e.the entrepreneur him/herself. Therefore, with the exception of Lackof Time, organizational barriers like Divergent Interests (betweenwho decides on energy efficiency and who invests), ComplexDecision Chain, Lack of Internal Control or even Low Status ofEnergy Efficiency may fade.

3. Methodology

Due to the explorative nature of the study, the research has beencarried out as case-studies using semi-structured interviews andquestionnaires. The 48 Small andMedium-sized Enterprises (SMEs)investigated (according to the definition provided by the EuropeanCommission [19]) are located in the Lombardy Region,1 one of themost industrialized in Europe, had all participated in a project onfinancing energy-efficient technologies carried out in 2010, andcontributed voluntarily to this research.

Table 2Table of synthesis of the taxonomy adopted for empirical investigation highlighting the characteristics of the barriers investigated. Source: Cagno et al. [17].

Barriers for empirical investigation (a) Origin:internal (I) orexternal (E)

(b) Decision-makingstep

Spectrum of influence of the barriers

(c) To anyinvestment

(d) To energyefficiency: general(D) or intervention-dependent (D)

Technology-relatedbarriers

Technologies not adequate E 1, 3 DTechnologies not available E 1, 3 D

Information barriers Lack of information on costsand benefits

E 2 D

Information not clear bytechnology suppliers

E 2 D

Trustworthiness of theinformation source

E 2 D

Information issues onenergy contracts

E 2 D

Economic Low capital availability I 1,2,3 l GInvestment costs E 3 DHidden costs I/E 2.3 DIntervention-related risks I/E 3 DExternal risks E 1 GIntervention not sufficientlyprofitable

I/E 3 D

Behavioural Lack of interest in energyefficiency interventions

I 1 G

Other priorities I 1 GInertia I 1 l GImperfect evaluation criteria I 3 l GLack of sharing the objectives I 3 G

Organisational Low status of energy efficiency I 2.3 GDivergent interests I 1 GComplex decision chain I 2.3 GLack of time I 1.3 l GLack of internal control I 3 G

Barriers related tocompetences

Identifying the inefficiencies I 1.2 G/DIdentifying the opportunities I 1.2 G/DImplementing the interventions I 3 G/DDifficulty in gathering externalcompetences

E 2 G/D

Awareness Lack of awareness or ignorance I 1 G/D

(a) The barrier may have its origin within the firm (Internal, I), or outside the firm (External, E).(b) The barrier affects the Action “Generation of Interest” (1), the Action “Research of inefficiencies and opportunities” (2), or the Action “Investment analysis and interventionimplementation” (3). One barrier can affect multiple actions.(c) The barrier may affect any investment of the firm, i.e. not only those specific for the energy efficiency improvement.(d) The barrier, affecting exclusively energy efficiency, can be quantified in general (thus not depending on which action to be considered, G), or its value can vary according toa specific investment to be considered (intervention-dependent, D). One barrier can be both general and intervention-dependent.

A. Trianni et al. / Energy 49 (2013) 444e458446

A case-study has been structured as follows: in the first part, therespondents e all of them responsible for energy issues at theirsite e provided details about their firm’s characteristics and theirview about the type of market in which they operate and theirposition within that market. In the second part, the respondentshave been asked to fill out a guided questionnaire covering variousaspects of the energy efficiency topic, starting from providingdetailed data about their firm’s size, production, turnover, energyexpenditures, and delving into barriers to and practices for energyefficiency investments.

The investigation aimed at obtaining the perceived (Yi0) and thereal (Yi) e where available e values of the barriers (Bi), since theytogether contribute to determine the full picture of the barriers. Theperceived values can be obtained asking directly one or morequestions. For what concerns the Awareness barrier, consideringthat we are investigating SMEs, it does not seem to be reasonable toask for its perceived value, that would be unreliable and sensitive,since it would require a self-evaluation of the interviewed person.The real values are instead obtained through gathering several data(Di) about practices and behaviours on the energy-efficient tech-nologies investment processes. It is apparent that gathering the realvalues e as they do not refer to feelings, judgements or opinions bythe respondents e could not be so straightforward, requiring, for

a single barrier, several data and/or information, not alwaysobtainable and/or derivable.

In the following sections we will present separately how theperceived and real values of the barriers have been collected.

3.1. Perceived barriers: detailed analysis of the questions

Measuring the perception of a given barrier could be quitedifficult, since the respondent might be in the position ofanswering about his or her opinion or behaviour with respect to thetopic. Therefore we formulated the questions in the form of sen-tences to which they agree or disagree, and, in order to rank theresponses, to adopt a Likert scale score from 1 (“I completelydisagree with the sentence”) to 4 (“I completely agree with thesentence”).

In some cases e reported in the following e, the value ofa barrier depends on a combination of the responses to severalquestions, to reduce the risk of socially acceptable responses toa specific question.

The Hidden Costs barrier (Y3.30) takes into account all the costspre, during, or post the intervention. The pre-intervention hiddencosts encompass the research of energy inefficiencies and oppor-tunities for increasing energy efficiency (Q3.3.1). The hidden costs

A. Trianni et al. / Energy 49 (2013) 444e458 447

during the implementation of the investment consider thedisruption costs, and all the costs related to the modification of theproduction system (e.g. layout of the equipment) that is needed inorder to install the new energy-efficient technology (Q3.3.2).Considering the post-intervention hidden costs, they include thecosts for training the personnel on the proper use of the newtechnology, developing new procedures for maintenance, adaptingthem to the modified production system (Q3.3.3). As a conse-quence, the Hidden Costs perceived barrier will be deemedimportant if just one of the three will be considered as a significantbarrier (1):

Y3:20 ¼ maxðQ3:3:1;Q3:3:2;Q3:3:3Þ (1)

Analogously for the perceived value of the Intervention notSufficiently Profitable barrier: the firm might not decide to imple-ment an energy-efficient intervention since it considers either thatthey have a high pay-back time (Q3.6.1) or, independently, witha too low economic return (Q3.6.2). Therefore, the perceived valueof the barrier Intervention not Sufficiently Profitable (Y3.60) will behigh if just one of them will be deemed important (2):

Y3:60 ¼ maxðQ3:6:1;Q3:6:2Þ (2)

Moreover, for the perceived value of the Lack of Interest inEnergy Efficiency Interventions barrier (Y4.10): the firm might benot interested (3) either because it considers itself as already effi-cient (Q4.1.1), or, independently, if it considers its energy expen-ditures as so low that it is not worth reducing them (Q4.1.2).

Y4:10 ¼ maxðQ4:1:1;Q4:1:2Þ (3)

3.2. Real barriers: detailed analysis of the questions

The process of gathering the needed information to evaluatethe real barriers is not as straightforward as for the perceivedbarriers, due to the limitations already described in Section 2. Inthe following, we report the description of the real barriersinvestigated.

Low Capital Availability (Y3.1): the value of the Low CapitalAvailability barrier is measured as the distance between thecapital needed by the enterprise and the capital availability bythe enterprise to perform investments specifically for energyefficiency. Therefore, it is possible to propose an index repre-sented by the average ratio (over the last 5 years) of the capitalrequired by the enterprise to perform investments in energy-efficient technologies on the available budget. We acknowl-edge that, in some cases, the request of resources could beinfluenced by the available budget. Nonetheless, we have veri-fied that the values provided by the respondents would not beinfluenced by the budget. Adopting a Likert scale from 1 to 4 toevaluate this parameter, the threshold values proposed are 1, 1.2and 1.5. Of course, the first threshold is related to the existenceof the barrier. If the ratio is lower than 1, the capital needed forinvestments is lower than the devoted budget, therefore thebarrier does not exist. The other thresholds have been proposedbased on experience in similar enterprises.Lack of interest in energy efficiency (Y4.1): an enterprise mightnot be interested in energy-efficient technologies mainly fortwo reasons: on the one side, since the energy expenditures arenot relevant; on the other side, because the enterprise considersitself to be already energy-efficient. The evaluation of theeffective energy efficiency status of the enterprise could not beobtained without an energy assessment. Indeed, given thevariety of technologies and processes adopted, it is not possible

to characterize SMEs with energy consumed by unit of product.Therefore, the importance of energy expenditures, expressed asthe ratio between the energy expenditures and the total costs ofproduction, given the variety in the SMEs’ activities, is the soleidentified indicator. The three values for the thresholdsproposed are 1%, 3% and 5%. Based on the recent indicationsprovided by the US Department of Energy [21] and Rohdin andThollander [6] and the finding of a study by Ramirez et al. [22],we assume the value of 3% as the threshold between energyintensive and non-energy intensive enterprises.Other priorities (Y4.2): this barrier could be measured as therelative importance of investments in energy efficiency.Therefore, we propose an index represented by the ratiobetween the capital specifically invested for energy-efficienttechnologies, and the total amount of capital invested in thepast five years. From the latter it would be preferred to subtractthe amount of capital for investments for business continuity,and therefore are not optional for the decision-maker. Weacknowledge that with this index some possibilities existing inreal cases are not encompassed: nonetheless, as first approxi-mation, this index is often (as by experience) the first indicatorof the priorities. The three proposed values on a Likert scale are0%, 5% and 10%.Inertia (Y4.3): the Inertia barrier encompasses two differentissues, i.e. the resistance to change and, secondly, the resistanceto risk. To evaluate the resistance to change, we propose tomeasure the number of changes the enterprise has adopted inthe last 5 years, expressed in terms of product portfolio andtechnological processes. The indicators required are, respec-tively, the length of the average life-cycle of products, and theaverage life-cycle of processes (or modifications of productionprocesses). In general terms, it is quite difficult to set a uniquethreshold on the basis of market and sector characteristics.Nonetheless, for the type of enterprises considered, the value of1 reflects a continuous change in the portfolio, 2 years seems torepresent a reasonable threshold for measuring both changes inproducts and processes, whilst 5 years is considered as a verystable process. An average of the two indicators has been ob-tained as representative of the resistance to change.

By considering the aversion to risk, the indicator proposed isthe maximum pay-back time (PBT). Indeed, this parameterexpresses the caution with which the enterprises tends to protectitself from future uncertainties. In particular, the shorter the PBT,the more risk resistant the enterprise is. Another indicator is therisk-premium, i.e. the minimum amount of money to take intoaccount the uncertainty of an investment. The higher this indi-cator, the more risk averse the enterprise. Nonetheless, adoptinga risk-premium was not so common: in the few cases in which itwas possible to obtain both parameters, we have observed thatthey proved to be strongly correlated, therefore providing thesame indication about the risk aversion of the enterprise. Thethree threshold values are set at 1, 2 and 5 years. In particular, ifmaximum PBT is lower than one year, it means that the enterprisetends to consider investments uniquely referable to a given fiscalyear. Having both a measure of the resistance to change and risk, itis possible to estimate the inertia barrier. There is no reason toassume that one of the two components of inertia, either resis-tance to change or risk aversion, could have a greater impact.Therefore, the authors propose to give the same weight to the twoelements.

Imperfect evaluation criteria (Y4.4): this barrier occurs when thedecision-maker does not have the proper knowledge or criteriato evaluate the investments in energy-efficient technologies. On

Table 3eInvestigated barriers (perceived, Yi0 , and real, Yi, values) and proxies adopted.

Category Barrier Perceived barrier Real barrier

Technology-related barriers

Technologies not adequate Direct question e

Technologies not available Direct question e

Information barriers Lack of information oncosts and benefits

Direct question e

Unclear information bytechnology suppliers

Direct question e

Trustworthiness of theinformation source

Direct question e

Information issues onenergy contracts

Direct question e

Economic Low capital availability Direct question P5i¼1

annual investmens needed for energy efficiencyiannual budget for energy efficiencyi

5Investment costs Direct question e

Hidden costsmax

( costpre� intervention;costs during the implementation;

costs post� intervention

)e

Intervention-related risks Direct question e

External risks Direct question e

Interventions notsufficiently profitable max

(high PBT;low returns

)e

Behavioural Lack of interest in energyefficiency interventions max

(enterprise already efficient;

low energy expenditures ðnot worth to reduceÞ

)annual energy costs

annual production costs

Other priorities Direct question P5i¼1

total annual investmens for energy efficiencyiðtotal annual investments � total annual investments for business continuityÞi

5Inertia Direct question

average

8>>>><>>>>:

resistance to change

(average product life� cycleaverage process life� cycle

)

risk aversion

(PBTrisk premium

)9>>>>=>>>>;

Imperfect evaluation criteria Direct questionaverage

(risk considered when evaluating investment

decision under life� cycle perspective

)

Organizational Lack of time Direct question e

Barriers related tocompetences

Identifying the inefficiencies Direct questionaverage

( specific equipment for energy consumption measurmenttable for consumption evaluation

personnel for mapping energy consumption

)

Identifying the opportunities Direct questionaverage

(BAT=Ps identification evaluationexisting equipment evaluation

)

Implementing the interventions Direct question P3i¼1

intervention without external supportiinterventions implementedi

with i ¼( 1 recovery2 optimization of performance

3 innovation

)

Difficulty in gathering externalskills

Direct question e

A.Trianni

etal./

Energy49

(2013)444

e458

448

Table 4Evaluation and ranking of the perceived barriers for the total investigated sample.

Rank Range Value Perceived barrier

1 >3 3.04 Y3.20 Investment costs2 y3 2.98 Y2.40 Information issues on energy contracts3 y3 2.90 Y3.30 Hidden costs4 y3 2.88 Y3.60 Intervention not sufficiently profitable5 >2.5 2.73 Y6.40 Difficulty in gathering external skills6 >2.5 2.65 Y3.10 Low capital availability7 >2.5 2.63 Y1.10 Technologies not adequate8 >2.5 2.58 Y1.20 Technologies not available9 >2.5 2.56 Y2.20 Information not clear by technology suppliers10 >2.5 2.52 Y5.10 Lack of Time11 >2.5 2.52 Y6.30 Implementing the interventions12 y2.5 2.42 Y2.10 Lack of information on costs and benefits13 >2 2.35 Y3.40 Intervention-related risks14 >2 2.21 Y6.20 Identifying the opportunities15 >2 2.17 Y4.10 Lack of interest in energy efficiency interventions16 >2 2.13 Y4.20 Other priorities17 >2 2.08 Y4.40 Imperfect evaluation criteria18 y2 2.00 Y3.50 External risks19 y2 2.00 Y2.30 Trustworthiness of the information source20 <2 1.88 Y6.10 Identifying the inefficiencies21 <2 1.71 Y4.30 Inertia

A. Trianni et al. / Energy 49 (2013) 444e458 449

the one side, the criteria are approximated or downgraded toroutines. On the other side, the risks associated to differentinvestments are not taken into account. We measure therespondents’ capability for accounting the investment riskwhen evaluating its profitability, with respect to investmententity and priority, on a scale from 1 (“not able”) to 4 (“alwaysaccounting for risk”). For the adoption of approximated criteria,the operating costs of the technologies are usually not consid-ered as a driver for the decision. Therefore, in the same scale, wehave asked how frequently they make decisions basing theirevaluations on a life-cycle cost perspective.Identifying the inefficiencies (Y6.1): this barrier represents thelack of knowledge of the enterprise about the needed compe-tences to identify the inefficiencies in their production system.Therefore, in the interview, the respondents have been askedabout the existence of specific equipment to measure energyconsumption, the existence of tables to evaluate the trend ofconsumption of the equipment, and the presence of personnelable to develop a detailed map of the energy consumption.Based on these responses, we have developed an indicator ableto quantify the capability to identify the inefficiencies, measuredwith a Likert scale from 1 to 4.Identifying the opportunities (Y6.2): we have asked the respon-dents to rank the capabilities to evaluate energy efficiencyopportunities. In particular, we have used two discriminantfactors to evaluate this. On the one side, we have asked therespondents to identify, at least by areas (e.g. compressed airsystem) the Best Available Technologies [23]. On the other side,we have investigated the respondents’ knowledge about theirinstalled technologies and related energy efficiency opportuni-ties, asking, for each of them, their values of energy consump-tion, investment costs and operating costs. Based on theseresponses, we have developed an indicator to quantify thecapability to identify the opportunities, measured with a Likertscale from 1 to 4.Implementing the interventions (Y6.3): this barrier occurs whenthe enterprise is not able to perform an intervention for energyefficiency without external support. For simplicity, we havedivided the interventions into three broad categories:� specific energy efficiency interventions on the productiontechnology;

� light modifications to equipment having a moderate influ-ence on the production and/or on the ancillary systems;

� interventions of restoring and/or optimizing the operativeconditions of the equipment.

The evaluation of this barrier has been obtained as a weightedaverage of how frequently, in the last five years e with respect tothe total number of interventions performed by category e theenterprise has been able to perform the intervention, without anexternal support, always with a Likert scale from 1 to 4.

For simplicity and clarity for the reader, in Table 3 we report thebarriers investigated in terms of perceived and real values.

4. Results and discussion

4.1. Whole sample

The analysis has been firstly devoted to investigate the values ofthe perceived barriers for the whole sample (see Table 4). In thiscase the major barrier is represented by the Investment Costs forenergy-efficient technologies. This conclusion is in line with otherstudies [6,8,9,18,24e26], but was also expected. Indeed, the resultsreasonably reflect the status-quo of industries struck by the globalfinancial crisis. As a consequence, enterprises have such a dramatic

perception of their situation, often without sufficient capital to paythe suppliers, that seem to be paralysed, and assume energy-efficient technologies to be too expensive.

Nevertheless, too high Investments Costs come along withother economic issues to be addressed, i.e. the presence of barriersas Hidden Costs and Technologies not Sufficiently Profitableexpresses two sides of the same coin, since they represent, on theone side, higher costs than expected, on the other, insufficientreturns. Moreover, the Low Capital Availability is found to beimportant, with an average score of 2.65 and ranked sixth amongthe barriers.

Moreover, with high values (close to 3), we find that costs,although representing the major issue, are not alone. Indeed, it ispossible to find the Information issues, i.e. Issues on EnergyContracts (2nd in the rank), Information not Clear by Technologiesproviders (9th), and the Lack of information on Costs and Benefits(12th). Considering the literature, although the issues on energycontracts have not explicitly been pointed out by previous empir-ical studies, information issues emerged as a primary barrier onlyin the findings of Rohdin and Thollander [6] and Trianni andCagno [9].

Furthermore, we also find a group of barriers with a scoregreater or almost equal to 2.5, reflecting on the one side the issuesof technologies, perceived neither adequate nor available, on theother, but less important, the issues on competences (expressed byDifficulty in Gathering External Skills and Difficulty in Imple-menting the Interventions). We interpret these findings as follows:SMEs perceive to be forced to adopt not customized solutions, butrather technologies usually specifically developed for largercustomers, as revealed by a study of Kalantaridis [27], and findmany difficulties in adopting external consultants able to supportthem in implementing the interventions.

Likewise, it is interesting to find that the behavioural barriersare ranked (with the exception of lack of time) in the lowestpositions (ranked below the 15th position), with scores close to orlower than 2. Indeed, the interviewed companies perceive them-selves as quite proactive on energy efficiency.

For the effect of the barriers on decision-making, the analysisrevealed an almost equal importance of the three issues e Gener-ation of Interest, Knowledge of the Opportunities and InvestmentAnalysis and Intervention Implementation. Therefore, the impor-tance for the first decision-making action, i.e. Generation of

A. Trianni et al. / Energy 49 (2013) 444e458450

Interest, represents an interesting result for policy-makers, since itmight point out the need, beside financial support of energy-efficient technology, for adequate actions to increase interest inenergy efficiency among SMEs.

In order to identify a possible common trend among theperception of the barriers, we have performed a correlation anal-ysis. The results of this analysis (Table 5) provide additional value tothe taxonomy adopted [17] in the research. Indeed, reducing themto the lowest common denominator e as proposed by the adoptedtaxonomy e is proven by the usually very low correlation coeffi-cients. For example, the barrier Technologies not Adequate (Y1.10)and Interventions not Sufficiently Profitable (Y3.60), present a verylow correlation coefficient (only 0.08), showing that the perceptionof the barriers might be related to issues that are independent, andthus it is important to distinguish them. Instead, the two barriers,considering the most complete taxonomy provided by Sorrell et al.taken since now as reference [15], fall both indistinctly into theHeterogeneity barrier.

Table 5Correlation analysis for the perceived barriers in the total sample investigated.

Y1.10 Y1.20 Y2.10 Y2.20

Technologiesnot adequate

Technologiesnot available

Lack ofinformationon costsand benefits

Informnot cletechnosupplie

Y1.10 Technologiesnot adequate

e 0.02 0.00 0.02

Y1.20 Technologiesnot available

0.20 �0.16

Y2.10 Lack ofinformationon costs and

�0.54

Y2.20 Informationnot clear bytechnology

Y2.30 Trustworthinessof theinformation

Y2.40 Informationissueson energy

Y3.10 Low capitalavailability

Y3.20 Investmentcosts

Y3.30 Hidden costsY3.40 Intervention-

related risksY3.50 External risksY3.60 Intervention not

sufficientlyY4.10 Lack of

interest inenergyefficiency

Y4.20 Otherpriorities

Y4.30 InertiaY4.40 Imperfect

evaluationcriteria

Y5.40 Lack of timeY6.10 Identifying the

inefficienciesY6.20 Identifying the

opportunitiesY6.30 Implementing

the interventionsY6.40 Difficulty in

gatheringexternal

In few cases, we see fairly detectable correlations (coefficientsslightly higher than 0.6). This might be related to the existence ofpossible dynamics among barriers, as suggested by Cagno et al.[17]. For example, it is realistic to assume that enterprises thathave and perceive few competences to identify inefficiencies, willpresent similar perceptions to identify opportunities (r ¼ 0.62)and implement the interventions (0.61). Likewise, it is alsointeresting to note the correlation of these barriers with thebarrier Lack of Time. In fact, it seems reasonable to assume that,people with scarce competences, would suffer much more fromnot having sufficient time to investigate inefficiencies, opportu-nities, and to implement interventions. As one interviewed said:“Well, I barely know what to do, and I do not even have time to dothese things!”.

In addition to that, the analysis has been devoted also toinvestigate the values of the real barriers (Table 6) and point outpossible differences (Fig. 1), when available, with the perceivedvalues.

Y2.30 Y2.40 Y3.10 Y3.20 Y3.30

ationar bylogyrs

Trustworthinessof theinformationsource

Informationissues onenergycontracts

Low capitalavailability

Investmentcosts

Hiddencosts

�0.12 0.27 �0.17 �0.16 �0.07

�0.02 �0.32 0.15 �0.10 �0.10

0.59 0.17 0.19 0.43 0.35

�0.52 0.32 0.36 0.38 0.39

�0.10 0.28 0.31 0.33

�0.13 0.10 0.26

�0.58 0.12

�0.19

Table 6Evaluation and ranking of the real barriers for the total sample investigated.

Rank Range Value Barrier

1 >3 3.29 Y4.1 Lack of interest in energy efficiency interventions2 >3 3.13 Y4.2 Other priorities3 >2.5 2.57 Y3.1 Low capital availability4 >2.5 2.54 Y6.1 Identifying the inefficiencies5 >2.5 2.52 Y6.3 Implementing the interventions6 >2 2.41 Y4.4 Imperfect evaluation criteria7 >2 2.13 Y4.3 Inertia8 <2 1.27 Y6.2 Identifying the opportunities

A. Trianni et al. / Energy 49 (2013) 444e458 451

Fig. 1 shows that enterprises have an almost “perfect” pictureof their difficulties to obtain capital for energy efficiencyinvestments (average difference between perceived and real of0.08) and to implement interventions (in this case, bothperceived and real values are equal to 2.52). Nonetheless, theyhave serious interest and priorities problems when consideringenergy efficiency interventions (average difference respectivelyof 1.12 and 1.00 more than the perceived value). Therefore,enterprises perceive themselves to be interested and to considerenergy efficiency properly, but not in reality. This is proven by thefact that resources and capital devoted to energy efficiency aremarginal with respect to other issues. In addition to that, highervalues for the inertia and imperfect evaluation criteria barriers(average difference of 0.42 and 0.32) show that enterprises,although declaring to be proactive and to adopt correct criteria toconsider energy efficiency investments, are reluctant to change,and to risk.

Another distorted perception can be found for Identifying theOpportunities. From our results (average perceived value 0.94

Y3.40 Y3.50 Y3.60 Y4.10 Y4.20 Y4.30 Y4.40

Intervention-related risks

Externalrisks

Interventionnot sufficientlyprofitable

Lack of interestin energyefficiencyinterventions

Otherpriorities

Inertia Imperfevaluacriteria

0.11 �0.24 0.08 �0.04 0.15 0.16 �0.14

�0.02 0.34 0.13 0.01 0.17 0.00 �0.21

0.34 0.03 0.21 0.10 0.09 0.29 �0.25

0.29 0.06 0.28 0.07 0.15 0.15 0.02

0.24 0.25 0.27 0.12 0.17 0.41 �0.10

0.15 �0.34 0.22 0.01 0.03 �0.32 0.11

0.00 0.36 0.15 �0.08 0.45 0.06 0.05

0.19 0.18 0.06 �0.01 0.15 0.03 �0.04

�0.36 0.08 0.19 0.23 0.06 0.10 �0.05�0.12 0.45 0.20 �0.08 0.24 �0.20

e �0.07 �0.06 0.25 0.13 0.05�0.16 0.33 0.18 �0.33

�0.19 0.25 �0.28

�0.16 �0.17

e �0.12

greater than the real one), it seems that enterprises think thatenergy-efficient technologies should be something particularly“special”, therefore they have a perception of this barrier higherthan the real value.

Moreover, the correlation analysis between the real barriersinvestigated did not show any significant link, with coefficients

Y5.40 Y6.10 Y6.20 Y6.30 Y6.40

ecttion

Lack oftime

Identifying theinefficiencies

Identifying theopportunities

Implementing theinterventions

Difficulty ingatheringexternalcompetences

�0.05 �0.07 0.10 0.12 0.25

�0.06 �0.15 �0.19 �0.26 �0.23

0.39 0.27 0.58 0.49 0.48

0.35 0.51 0.61 0.49 0.28

0.39 0.34 0.56 0.47 0.56

0.15 0.14 0.23 0.24 0.33

0.32 0.39 0.30 0.18 0.11

0.30 0.49 0.36 0.37 0.23

0.22 0.33 0.20 0.14 0.260.05 0.05 0.02 0.01 0.12

e �0.07 �0.06 e �0.250.30 0.23 0.14 0.16 0.01

0.05 0.10 �0.05 �0.03 0.04

0.41 0.27 0.23 0.21 �0.01

0.04 0.08 0.03 0.10 0.25�0.07 0.09 �0.07 �0.03 0.00

e 0.63 0.57 0.57 0.47e 0.62 0.61 0.36

e 0.83 0.45

e �0.42

e

Fig. 1. Total sample e evaluation of the difference between the perceived and real values of the investigated barriers.

A. Trianni et al. / Energy 49 (2013) 444e458452

never greater than 0.35, thus again showing the capability todistinguish the effects.

In conclusion, we find that, considering all barriers investigated,energy efficiency is effectively a marginal issue by the managementof the firm. The findings seem to show a possible path to reducethese barriers. To increase the priority of investments in energy-efficient technologies, more attention should be provided toclearly point out the benefits (direct and indirect) from adoptingthose technologies, as the approach followed by Worrell et al. [28].

In the following, we have tried to analyse the perceived and realvalue for clusters of enterprises, demonstrating the role of thefollowing factors:

Table 7Evaluation of perceived and real barriers by firm’s size.

Perceived barriers

Y1.10 Technologies not adequateY1.20 Technologies not availableY2.10 Lack of information on costs and benefitsY2.20 Information not clear by technology suppliersY2.30 Trustworthiness of the information sourceY2.40 Information issues on energy contractsY3.10 Low capital availabilityY3.20 Investment costsY3.30 Hidden costsY3.40 Intervention-related risksY3.50 External risksY3.60 Intervention not sufficiently profitableY4.10 Lack of interest in energy efficiency interventionsY4.20 Other prioritiesY4.30 InertiaY4.40 Imperfect evaluation criteriaY.5.10 Lack of timeY6.10 Identifying the inefficienciesY6.20 Identifying the opportunitiesY6.30 Implementing the interventionsY6.40 Difficulty in gathering external competences

Real Barriers

Y3.1 Low capital availabilityY4.1 Lack of interest in energy efficiency interventionsY4.2 Other prioritiesY4.3 InertiaY4.4 Imperfect evaluation criteriaY6.1 Identifying the inefficienciesY6.2 Identifying the opportunitiesY6.3 Implementing the interventions

a The column reports the level of significance of the comparison.b The average delta between the sub-samples (SEs vs. MEs) is 20% greater than the avc The average delta between the sub-samples (SEs vs. MEs) is 15% greater than the av

� firm’s size;� energy expenditures;� complexity of the production;� variability of the demand;� strength of competitors.

As the first two are the most traditional factors in evaluatingbarriers to energy efficiency, due to their direct link with energyuse, we have evaluated the other three since are linked to factorsrelated to the complexity of the context in which the enterpriseoperates [29]. In particular, if the complexity of the production isa proxy for the internal complexity, the demand variability and

Average Questionnaire (32 SEs-16 MEs)

Total D ¼ SEs-MEs Significancea

2.63 0.372.58 �0.342.42 �0.41 c

2.56 �0.382.00 �0.192.98 0.162.65 �0.163.04 �0.132.90 �0.162.35 �0.032.00 �0.192.88 �0.47 c

2.17 0.53 b

2.13 0.091.71 0.132.08 �0.162.52 �0.251.88 �0.192.21 �0.162.52 �0.162.73 0.06

Average

Total D ¼ SEs-MEs Significancea

2.57 0.48 c

3.29 0.313.13 �0.032.13 �0.092.41 0.38 c

2.54 �0.131.27 0.032.52 �0.53 b

erage value for the response.erage value for the response.

A. Trianni et al. / Energy 49 (2013) 444e458 453

strength of competitors offer a view of the complexity with respectto clients and market.

In particular, due to the limited number of responses, it is notpossible to perform a proper statistical analysis: nonetheless, wepoint out the differences between the sub-samples whether theaverage differ respectively at least of 20% (**) or 15% (*) from theaverage for the whole sample.

4.2. Firm’s size

Small and Medium-sized Enterprises (SMEs) are usuallybundled together, as shown in previous research [9,30]. In thisstudy we adopt the classification within SMEs proposed by theEuropean Commission [19], that combines the number ofemployees and the annual turnover (expressed in million V),creating a further class within medium-size enterprises, as follows:

� Small Enterprises (SEs), from 10 to 49 employees and an annualturnover up to 10 million V;

� Medium Enterprises (MEs), from 50 to 99 employees and anannual turnover up to 20 million V;

� Medium-Large enterprises (MLEs), from 100 to 249 employeesand an annual turnover up to 50 million V.

In order to analyse the typical characteristics of smaller enter-prises, following the results obtained by Trianni and Cagno [9] andMicheli and Cagno [30], we focus on SEs and MEs. The investigatedsample is composed by 32 SEs ad 16 MEs. From an analysis of theperceived barriers (see Table 7), we identify several differences.

Table 8Evaluation of perceived and real barriers by energy expenditures.

Perceived barriers Average

Total

Y1.10 Technologies not adequate 2.63Y1.20 Technologies not available 2.58Y2.10 Lack of information on costs and benefits 2.42Y2.20 Information not clear by technology suppliers 2.56Y2.30 Trustworthiness of the information source 2.00Y2.40 Information issues on energy contracts 2.98Y3.10 Low capital availability 2.65Y3.20 Investment costs 3.04Y3.30 Hidden costs 2.90Y3.40 Intervention-related risks 2.35Y3.50 External risks 2.00Y3.60 Intervention not sufficiently profitable 2.88Y4.10 Lack of interest in energy efficiency interventions 2.17Y4.20 Other priorities 2.13Y4.30 Inertia 1.71Y4.40 Imperfect evaluation criteria 2.08Y.5.10 Lack of time 2.52Y6.10 Identifying the inefficiencies 1.88Y6.20 Identifying the opportunities 2.21Y6.30 Implementing the interventions 2.52Y6.40 Difficulty in gathering external competences 2.73

Real Barriers

Y3.1 Low capital availabilityY4.1 Lack of interest in energy efficiency interventionsY4.2 Other prioritiesY4.3 InertiaY4.4 Imperfect evaluation criteriaY6.1 Identifying the inefficienciesY6.2 Identifying the opportunitiesY6.3 Implementing the interventions

a The column reports the level of significance of the comparison.b The average delta between the sub-samples (Low Energy Expenditures vs. Higher Enc The average delta between the sub-samples (Low Energy Expenditures vs. Higher En

Firstly, looking at the barrier Lack of Interest in Energy EfficiencyInterventions, the perceived value is obtained evaluating the twoeffects of either having considered the enterprise as already effi-cient, or with such low energy expenditures that it is not worthinvesting in energy efficiency.

Looking further to the differences in the responses for the twoquestions, the greatest difference between SEs and MEs occurs forthe perception of having very low energy costs. Indeed, if MEs donot consider energy costs as marginal, SEs do it. On the contrary,when investigating the perception of being already efficient, bothclusters declare to disagree. This result was expected since theenterprises voluntarily participated in the study on energy effi-ciency, thus showing a tendency of not perceiving themselves asalready efficient. As a conclusion, the difference in the lack ofinterest between SEs and MEs is explained only by a differentperception of the importance of their energy costs.

Secondly, the barrier Intervention not Sufficiently Profitable isobtained considering the two effects, i.e. if the investment in energy-efficient technologies presents either toohigh PBTor too low returns.From a separate analysis of the two questions, we see that in bothquestions the highest values occur forMEs. Indeed, it is reasonable toassume thatMEsmighthave awiderportfolio of investments, havingaccess to investments with higher rate of returns, thus tending todowngrade investments in energy-efficient technologies.

Thirdly, considering the barrier Lack of Information on Costs andBenefits, MEs perceive this barrier stronger than SEs. It is possiblethat SEs might have a closer and more direct contact with theinformation source. Moreover, it should be observed that thetrustworthiness in SEs is perceived as a less important barrier than

Questionnaire (39 < 500,000 V/year e 9 > 500,000 V/year)

D ¼ Low�high D ¼ Low�high

0.08�0.53 b

�0.31�0.54 b

�0.27�0.160.110.19�0.26�0.38 c

�0.14�0.43�0.21�0.12�0.22�0.030.23�0.020.120.37�0.06

Average

D ¼ Low�high D ¼ Low�high Significancea

2.57 0.133.29 0.72 b

3.13 0.022.13 �0.182.41 0.232.54 0.261.27 0.062.52 �0.59 b

ergy Expenditures) is 20% greater than the average value for the response.ergy Expenditures) is 15% greater than the average value for the response.

A. Trianni et al. / Energy 49 (2013) 444e458454

inMEs. Taking into account both results, we can say that SEs tend totrust to their information source, thus perceiving the availableinformation as sufficient.

Moreover, although with an average difference not larger than15% of the average value, it seems remarkable to note that SEs seemto perceive the barriers related to the technologies more impor-tantly than MEs. This result seems reasonable, due to the very largevariety of processes and technologies that are typical of SEs. Indeed,as emerged in one interviews, a respondent said: “I know they haveit, but it does not fully work for us”.

Finally, when looking at the decision-making process, we cannotdistinguish any particular difference, with exception of a smalllower perception of the barriers for SEs with respect to MEs in allthe three actions.

The analysis of correlations shows that different behaviour ofSEs and MEs can be observed: in particular, we have observeda higher correlation between the information barriers and thebarriers related to competences. This can be interpreted as follows:a very strong and direct contact with the information source, witha clearer form of the information about the technologies (in termsof costs, benefits, etc.), could reduce the perception of competence-related barriers, enabling the information about energy-efficienttechnologies to be closer to the sensitivity of entrepreneurs.

Instead, for MEs, we can see closer links between the informa-tion and the economic barriers, alsowith those related to risks, thusshowing that a perception of insufficient or unclear informationcould be connected to greater perceived uncertainties, possiblyrepresenting economic barriers to investments.

Table 9Evaluation of perceived and real barriers by complexity of the production.

Perceived Barriers Av

To

Y1.10 Technologies not adequate 2.6Y1.20 Technologies not available 2.5Y2.10 Lack of information on costs and benefits 2.4Y2.20 Information not clear by technology suppliers 2.5Y2.30 Trustworthiness of the information source 2.0Y2.40 Information issues on energy contracts 2.9Y3.10 Low capital availability 2.6Y3.20 Investment costs 3.0Y3.30 Hidden costs 2.9Y3.40 Intervention-related risks 2.3Y3.50 External risks 2.0Y3.60 Intervention not sufficiently profitable 2.8Y4.10 Lack of interest in energy efficiency interventions 2.1Y4.20 Other priorities 2.1Y4.30 Inertia 1.7Y4.40 Imperfect evaluation criteria 2.0Y.5.10 Lack of time 2.5Y6.10 Identifying the inefficiencies 1.8Y6.20 Identifying the opportunities 2.2Y6.30 Implementing the interventions 2.5Y6.40 Difficulty in gathering external competences 2.7

Real Barriers

Y3.1 Low capital availabilityY4.1 Lack of interest in energy efficiency interventionsY4.2 Other prioritiesY4.3 InertiaY4.4 Imperfect evaluation criteriaY6.1 Identifying the inefficienciesY6.2 Identifying the opportunitiesY6.3 Implementing the interventions

a The column reports the level of significance of the comparison.b The average delta between the sub-samples (Low Complexity of the production vs.

response.c The average delta between the sub-samples (Low Complexity of the production vs.

response.

Considering the real barriers, the picture is surprisingly clear:firstly, SEs have, effectively, greater difficulties in affording theinvestments in energy-efficient technologies, as shown by the LowCapital Availability barrier. As shown by previous studies [7,9,11],small business enterprises face strong difficulties to access credit.This might be due also to the fact that financial institutions do nothave programmes specifically addressed for SEs. Therefore, whenan SE requires capital to invest in energy-efficient technologies,greater difficulties for evaluating the investment (and the afford-ability by the firm) arise. This barrier clearly points out to need todevelop a credit system able to support smaller enterprises toperform energy efficiency investments.

Secondly, in SEs it is quite hard to have structured systems toevaluate investments. Therefore, the “rule of thumb” approach,pointed out by the Imperfect Evaluation Criteria, is more diffused.

Thirdly, MEs suffer from the implementation of the interven-tions more than SEs (respectively of 2.34 for SEs, 2.88 for MEs). Thismight be due to the fact that SEs are simpler and leaner, therefore inMEs the impact of the intervention on the existing functions isgreater, in terms of time (and costs) of the production disruption,organizing the operations, etc.

4.3. Effects of other important factors on the results

Energy expenditures: we have performed a separate analysis forthe factor “energy expenditures”, since it is widely assumed tobe an important factor. Indeed, creating two classes (annualenergy expenditures below or greater than 500 thousandsV per

erage Questionnaire (24 Low e 24 High or very high)

tal D ¼ Low�high D ¼ Low�high

3 0.138 0.54 b

2 0.176 0.38

0.33 c

8 �0.215 0.46 c

4 e

0 0.295 0.210 0.178 0.337 0.083 0.50 b

1 0.258 0.132 0.54 b

8 0.171 0.38 c

2 0.333 �0.21

Average

D ¼ Low�high D ¼ Low�high Significancea

2.57 0.183.29 �0.083.13 0.102.13 �0.032.41 �0.232.54 0.251.27 0.29 c

2.52 �0.13

High Complexity of the production) is 20% greater than the average value for the

High Complexity of the production) is 15% greater than the average value for the

A. Trianni et al. / Energy 49 (2013) 444e458 455

year), we can appreciate, as reported in Table 8, that enterpriseswith higher energy costs generally have a higher perception ofthe barriers than those with low energy costs.

Moreover, although not as evident as for the barriers listedabove, large differences exist for the barriers Interventions notsufficiently profitable and Difficulty in Implementing theInterventions.

In general, when assessing the values of real barriers, an inter-esting trend emerges: with few exceptions, the factor “energyexpenditures” has an impact similar to the firm’s size, which wasexpected (the greater the size, the greater the energy expendi-tures). Nonetheless, we see that higher energy expendituresincrease the effect of the size factor for perceived barriers. Indeed, itis a multiplier of the perception of the barrier, adding worries thatdo not seem to effectively exist.

The conclusions have been confirmed by the correlation analysisof the perceived barriers in the two sub-samples. The results arequite in line with the analysis by firm’s size.

Complexity of the production: we evaluated the possible effect ofthe complexity of the production on barriers to industrialenergy efficiency. The factor has been obtained as a combinationof variety of production and production volumes.

Considering the perceived barriers, as reported in Table 9, andthe previous analysis by firm’s size, we see that, in addition to thebarriers already identified, in general enterprises that considertheir production as not complex tend to perceive the barriers as

Table 10Evaluation of perceived and real barriers by Demand variability.

Perceived Barriers Av

To

Y1.10 Technologies not adequate 2.6Y1.20 Technologies not available 2.5Y2.10 Lack of information on costs and benefits 2.4Y2.20 Information not clear by technology suppliers 2.5Y2.30 Trustworthiness of the information source 2.0Y2.40 Information issues on energy contracts 2.9Y3.10 Low capital availability 2.6Y3.20 Investment costs 3.0Y3.30 Hidden costs 2.9Y3.40 Intervention-related risks 2.3Y3.50 External risks 2.0Y3.60 Intervention not sufficiently profitable 2.8Y4.10 Lack of interest in energy efficiency interventions 2.1Y4.20 Other priorities 2.1Y4.30 Inertia 1.7Y4.40 Imperfect evaluation criteria 2.0Y.5.10 Lack of time 2.5Y6.10 Identifying the inefficiencies 1.8Y6.20 Identifying the opportunities 2.2Y6.30 Implementing the interventions 2.5Y6.40 Difficulty in gathering external competences 2.7

Real Barriers

Y3.1 Low capital availabilityY4.1 Lack of interest in energy efficiency interventionsY4.2 Other prioritiesY4.3 InertiaY4.4 Imperfect evaluation criteriaY6.1 Identifying the inefficienciesY6.2 Identifying the opportunitiesY6.3 Implementing the interventions

a The column reports the level of significance of the comparison.b The average delta between the sub-samples (Low Demand Variability vs. High Dema

more critical compared to other enterprises. This is particularlyevident for the barriers Technologies not Available and OtherPriorities.

For the same reason, enterprises with low complexity presentgreater difficulties than those with very high complexity in the firststage of the decision-making process, i.e. the Generation of Interestfor energy efficiency. In addition, the correlation analysis hasshown that enterprises with low complexity present an interestingtrend, tending to relate information barriers to competence-relatedbarriers. Therefore, this factor may point out the existence of otherfactors able to relate the complexity of the production to the pro-activity of enterprises with respect to energy issues.

When looking at the real barriers, indeed, enterprises withdeclared simple production suffer from the barriers to identifyinefficiencies and opportunities more than others. This might bedue to the fact that, with a greater complexity of the production,greater attention is given to the processes, increasing the knowl-edge on the performance of equipment, and on possible solutions.

Demand variability: in Table 10 we report the results of theanalysis of the perceived and real barriers for demand variabilityof the enterprises.

For the perceived barriers, we observe that, with the exceptionof two cases, there are not relevant differences, even for thedecision-making actions and the analysis of correlations. Theperceived barriers with a deviation of more than 15% of the averagevalue are Low Capital Availability and External Risks, which aregreater for enterprises with low demand variability. The result

erage Questionnaire (13 low e 35 high or very high)

tal D ¼ Low�high D ¼ Low�high

3 0.028 0.072 0.056 �0.050 0.018 �0.115 0.52 b

4 0.170 0.015 �0.260 0.37 b

8 �0.257 0.073 0.201 �0.038 �0.032 �0.098 �0.131 �0.102 �0.143 �0.03

Average

D ¼ Low�high D ¼ Low�high Significancea

2.57 0.113.29 0.173.13 0.042.13 0.33 b

2.41 0.38 b

2.54 �0.031.27 0.042.52 0.09

nd Variability) is 15% greater than the average value for the response.

Table 11Evaluation of perceived and real barriers by strength of competitors.

Perceived Barriers Average Questionnaire (30 weak e 18 strong)

Total D ¼ Weak�strong D ¼ Weak�strong

Y1.10 Technologies not adequate 2.63 0.26Y1.20 Technologies not available 2.58 �0.03Y2.10 Lack of information on costs and benefits 2.42 �0.22Y2.20 Information not clear by technology suppliers 2.56 �0.34Y2.30 Trustworthiness of the information source 2.00 �0.09Y2.40 Information issues on energy contracts 2.98 0.14Y3.10 Low capital availability 2.65 �0.21Y3.20 Investment costs 3.04 �0.11Y3.30 Hidden costs 2.90 0.19Y3.40 Intervention-related risks 2.35 �0.14Y3.50 External risks 2.00 �0.27Y3.60 Intervention not sufficiently profitable 2.88 �0.29Y4.10 Lack of interest in energy efficiency interventions 2.17 0.18Y4.20 Other priorities 2.13 �0.42 c

Y4.30 Inertia 1.71 �0.11Y4.40 Imperfect evaluation criteria 2.08 0.40 c

Y.5.10 Lack of time 2.52 �0.14Y6.10 Identifying the inefficiencies 1.88 �0.11Y6.20 Identifying the opportunities 2.21 �0.02Y6.30 Implementing the interventions 2.52 �0.14Y6.40 Difficulty in gathering external competences 2.73 0.19

Real Barriers Average

D ¼ Weak�strong D ¼ Weak�strong Significancea

Y3.1 Low capital availability 2.57 �0.02Y4.1 Lack of interest in energy efficiency interventions 3.29 �0.04Y4.2 Other priorities 3.13 �0.13Y4.3 Inertia 2.13 �0.10Y4.4 Imperfect evaluation criteria 2.41 0.43 c

Y6.1 Identifying the inefficiencies 2.54 0.16Y6.2 Identifying the opportunities 1.27 0.26 b

Y6.3 Implementing the interventions 2.52 �0.23

a The column reports the level of significance of the comparison.b The average delta between the sub-samples (Low Strength of Competitors vs. High Strength of Competitors) is 20% greater than the average value for the response.c The average delta between the sub-samples (Low Strength of Competitors vs. High Strength of Competitors) is 15% greater than the average value for the response.

A. Trianni et al. / Energy 49 (2013) 444e458456

seems to be realistic, since it is reasonable to assume that enter-prises with higher demand variability are more used to facecontinuous changes.

Nonetheless, more interesting findings can be observed for thereal barriers. In this case, enterprises with a low demand variabilitypresent almost everywhere greater barriers. This is particular evidentin three cases, i.e. Lack of Interest in Energy Efficiency Interventions,Inertia, and Imperfect Evaluation Criteria. Considering that sampleeexclusively composed by Small and Medium-sized Enterprises(SMEs)e, wenote that greaterdifferences are not found for economicbarriers, rather for behavioural ones. Within the limited sample, itseems remarkable that the attention of policy-makers should bedevoted to increase the interest towards energy efficiency, especiallyfor very simple SMEs with low demand variability.

Strength of competitors: As reported in Table 11, we have ana-lysed the barriers in enterprises that operate into a market withweak or strong competitors. The results on the perceivedbarriers do not show evident trends, with the exception of twobarriers, i.e. Other Priorities and Imperfect Evaluation Criteria.Moreover, we do not see clear differences on the decision-making process.

In the first case, enterprises with strong competitors havea higher perception of this barrier, reasonably worried of losing,due to the implementation of new technologies and consequentproduction disruption, their clients. In the second case, it isreasonable to assume that enterprises with weak competitors couldnot be so familiar with proper or thorough investment analyses,

therefore perceiving higher values for Imperfect EvaluationCriteria.

Looking at the analysis of correlations, we notice the highcorrelation (in some cases even greater than 0.7) between thebarriers Lack of Information on Costs and Benefits, Information notClear by Technology Suppliers, and Trustworthiness of the Infor-mation Source. This, coupled with the lower absolute values of thebarriers (with respect to enterprises with strong competitors),means that enterprises with weak competitors hardly perceive theinformation issues as barriers.

For the real barriers, the analysis shows that the presence ofstrong competitors plays an important role. Therefore, enter-prises working in a very competitive market present lowervalues of barriers like Imperfect Evaluation Criteria and Identi-fying the inefficiencies. This means that, thanks to the externalpressures (competitors), some internal factors (within the firm)have moved the attention of enterprises towards energy-efficienttechnologies. Moreover, although not strictly “significant”,enterprises with strong competitors suffer more from the diffi-culties in implementing the interventions, reasonably due tofearing a production disruption when realizing energy-efficientinvestments.

5. Conclusions

The most recent inputs of the European Commission on theneed to increase energy efficiency renewed the interest towards theexistence of the energy efficiency gap due to barriers that inhibitinvestments in technologies that are both energy-efficient and

A. Trianni et al. / Energy 49 (2013) 444e458 457

(apparently) economically efficient. This is critical for Small andMedium-sized enterprises (SMEs), in which simple rules anddevices for energy saving have not been widely adopted yet.

Through the investigation of 48 manufacturing SMEs in theLombardy region it has been possible to obtain a clearer picture ofthe perceived and real barriers to industrial energy efficiency,evaluating the importance of the perceived barriers on thedecision-making process and their correlations. The analyses havebeen conducted accounting for characteristics of the investigatedsample, e.g. firm’s size, energy expenditures, complexity of theproduction, demand variability, and strength of the competitors.

By considering the whole sample, the major perceived barriersemerged from our study are represented by economic barriers (interms of high Investment Costs, Hidden Costs and Intervention notSufficiently Profitable) and Information barriers (i.e. InformationIssues on Energy Contracts, Information not clear by TechnologySuppliers and Lack of Information on Costs and Benefits). Moreover,we find that behavioural barriers are ranked in the lowest positions,thus showing that the enterprises perceive themselves as pro-active with respect to the topic. As the nature of the sampleinvestigated, that declares to be particularly aware with respect toenergy efficiency issues, and has recently participated withina project related to energy efficiency, this result has to be consid-ered with caution, and within its limitation. Indeed, uniquelythrough an extensive campaign of research it would be possible togather the full necessary information and build the knowledge tomake statistically-based and more robust conclusions.

For the analysis of the effects on the decision-making process, theresult seems to show an almost equal importance of Generation ofInterest, Knowledge of the Opportunities and Investment Analysisand Intervention Implementation. This finding is particularly impor-tant for policy-making, since this reflects theneed, for the enterprises,to increase the interest towards energy-efficient technologies.

The analysis of the real barriers allows to draw an interestingpicture: indeed, the enterprises have an almost perfect knowledgeof their difficulties to obtain needed capital, but present muchhigher real barriers for the Lack of Interest and Other Priorities.Actually, they presume to be interested in energy efficiency, but inreality are not, marginalizing energy efficiency. This result isinteresting for future research, since it points out the need toexplore this misalignment between perceived and real barriers, inorder to get a more detailed and precise picture of the barriers thatrepresents the basis for future energy policies.

The differences within SMEs in this study emerge for severalperceived barriers, as the Lack of Interest for energy efficiency,Intervention not Sufficiently Profitable, and Lack of Information onCosts and Benefits. Considering the real barriers, small enterprises(SEs) seem to suffer more than medium enterprises (MEs) from theLow Capital Availability and Imperfect Evaluation Criteria barriers,thus having greater difficulties to afford the investments in energy-efficient technologies and adopting more frequently rules of thumbor routines to evaluate investments. A greater difficulty suffered byMEs with respect to SEs is the implementation of the interventions,possibly due to the greater impact of disruption on enterprises’production. The results are confirmed in the analysis of the sampleby classes of energy expenditures. In conclusion, it seems necessaryto avoid bundling together SEs, MEs and MLEs in a broad definitionof SMEs to investigate energy efficiency, but to distinguish these.When considering energy expenditures, despite higher energycosts usually bring a higher perception of barriers, this factor seemsto lead to generally lower real barriers, in particular for whatconcerns the interest towards the topic.

Interesting findings for the study come from the analysis ofother factors, i.e. complexity of the production, demand variabilityand strength of competitors. Although with several differences for

the perceived barriers and their effect on the decision-makingprocess, the real barriers seems to be lower with high or veryhigh complexity of the production, high variability of the demandand strong competitors. The trend represents an interesting andchallenging issue for future research, since it might imply theexistence of other distorting factors (not only in terms of externalpressures, but also within the firm) moving enterprises towardsenergy-efficient technologies.

Finally, we can provide some additional hints for futureresearch. Firstly, it seems interesting to deepen the analysis ofenterprises’ behaviour combining energy efficiency and otherprocesses, such as innovation and entrepreneurship. Secondly, theinvestigation of direct and indirect benefits for energy efficiency,already performed for large energy-intensive enterprises, has notyet considered SMEs. Thirdly, this empirical investigation onbarriers to industrial energy efficiency calls for future research inthe world of drivers to energy efficiency with much greater effortson empirical studies [6,25,31]. Indeed, the real behavioural barriersemerged open the research on understanding which means mightbe promoted in order to overcome those barriers. Other barriers, asthe perceived Lack of Information on Energy Contracts, has pointedout that amissed uptake of an energy-efficient technologymight bedue also to perceived problems related to external actors. Inconclusion, it is clear the need to explore which drivers, eitherexternal or internal with respect to an enterprise, might beexploited to promote industrial energy efficiency. Future researchin this direction is really needed, in order to provide policy-makersa wider perspective on the most promising means to promoteenergy-efficient technologies.

Acknowledgements

The authors thank the research agency CESTEC, Center forTechnology Development, Energy and Competitiveness of theLombardy Region, for the support in contacting the enterprises, andgathering the necessary data for the research.

References

[1] [EC] European Council. Proposal for a directive on energy efficiency andrepealing directives 2004/8/EC and 2006/32/EC, COM. Official Journal of theEuropean Union 2011;370. 22nd June 2011.

[2] [EIA] Energy Information Administration. International energy outlook 2010.Washington, D.C: U.S. Department of Energy, Office of Integrated Analysis andForecasting; 2010.

[3] Eurobarometer Team e European Commission (EC), Observatory of EuropeanSMEs e analytical report, 2007.

[4] [ISTAT] Istituto Nazionale di Statistica. 8� Censimento generale dell’industria edei servizi. Italy: Roma. Available at: www.istat.it; 2004.

[5] EIM Business & Policy Research. Do SMEs create more and better jobs?. AAZoetermeer, The Netherlands; November 2011.

[6] Rohdin P, Thollander P. Barriers to and driving forces for energy efficiency inthe non-energy intensive manufacturing industry in Sweden. Energy 2006;31(12):1836e44.

[7] Cagno E, Trucco P, Trianni A, Sala G. Quick-E-Scan: a methodology for theenergy scan of SMEs. Energy 2010;35(5):1916e26.

[8] Nagesha N, Balachandra P. Barriers to energy efficiency in small industryclusters: multi-criteria-based prioritization using the analytic hierarchyprocess. Energy 2006;31(12):1969e83.

[9] Trianni A, Cagno E. Dealing with barriers to energy efficiency and SMEs: someempirical evidences. Energy 2012;37(1):494e504.

[10] Blumstein B, Krieg B, Schipper L, York C. Overcoming social and institutionalbarriers to energy conservation. Energy 1980;5:355e71.

[11] Intergovernmental Panel for Climate Change. Barriers, opportunities, andmarket potential of technologies and practices (chapter 5) in “climate change2001, mitigation” contribution of working group III to the third assessmentreport of the intergovernmental panel on climate change. Cambridge, UK;2001.

[12] Painuly JP, Reddy BS. Electricity conservation programs: barriers to theirimplementation. Energy Sources 1996;18(3):257e67.

[13] de Almeida AT, Fonseca P, Falkner H, Bertoldi P. Market transformation ofenergy-efficient motor technologies in the EU. Energy Policy 2003;31(6):563e75.

A. Trianni et al. / Energy 49 (2013) 444e458458

[14] Weber L. Some reflections on barriers to the efficient use of energy. EnergyPolicy 1997;25(10):833e935.

[15] Sorrell S, Schleich J, Scott S, O’Malley E, Trace F, Boede U, et al. Reducingbarriers to energy efficiency in public and private organizations. Final Report.SPRU. Available at:, http://www.sussex.ac.uk/Units/spru/publications/reports/barriers/final.html; 2000.

[16] Sorrell S, Mallett A, Nye S. Barriers to industrial energy efficiency: a literaturereview, background study for the UNIDO industrial development report (IDR)industrial energy efficiency pays, why is it not happening? Brighton: SPRU,University of Sussex; 2010.

[17] CagnoE,Worrell E,TrianniA, PuglieseG.Anovelapproach forbarriers to industrialenergy efficiency. Renewable and Sustainable Energy Reviews, in press.

[18] TrianniA, ThollanderP,BacklundS,CagnoE.Barriers to industrialenergyefficiencyin foundries: an European comparison. Journal of Cleaner Production, in press.

[19] European Commission. Commission recommendation of 6 May 2003 con-cerning the definition of micro-small and medium-size enterprises, 2003/361/EC. Official Journal of the European Union 6th May 2003.

[20] Lombardy Region. Nel Nord dell’Italia, nel cuore dell’Europa. Available at:http://www.regione.lombardia.it/; 2012 [in Italian].

[21] [EIA] Energy Information Administration. Annual energy outlook 2007. DOE/EIA-0383. Washington, D.C: U.S. Department of Energy, Office of IntegratedAnalysis and Forecasting; 2007.

[22] Ramirez CA, Patel M, Blok K. The non-energy intensive manufacturing sector.An analysis relating to the Netherlands. Energy 2005;30(5):749e67.

[23] Cagno E, Trianni A. Evaluation of benefits and costs for the most effectiveenergy efficiency interventions in manufacturing SMEs: a study on primarymetals, plastic and textile sectors. ASME Journal of Energy Resources Tech-nology 2012;134(2) (June), 021005 (9 pages).

[24] Reddy BS, Shrestha RM. Barriers to the adoption of efficient electricity tech-nologies: a case study of India. International Journal of Energy Research 1998;22:257e70.

[25] Rohdin P, Thollander P, Solding P. Barriers to and drivers for energy efficiencyin the Swedish foundry industry. Energy Policy 2007;35(1):672e7.

[26] Sardianou E. Barriers to industrial energy efficiency investments in Greece.Journal of Cleaner Production 2008;16(13):1416e23.

[27] Kalantaridis K. Processes of innovation among manufacturing SMEs: theexperience of Bedfordshire. Entrepreneurship & Regional Development: AnInternational Journal 1999;11(1):57e78.

[28] Worrell E, Laitner JA, Ruth M, Hodayah F. Productivity benefits of industrialenergy efficiency measures. Energy 2003;28(11):1081e98.

[29] Bozarth CC, Warsing DP, Flynn BB, Flynn EJ. The impact of supply chaincomplexity on manufacturing plant performance. Journal of OperationsManagement 2009;27(1):78e93.

[30] Micheli GJL, Cagno E. Dealing with SMEs as a whole in OHS issues: warningsfrom empirical evidence. Safety Science 2010;48:729e33.

[31] Cagno E, Trianni A. Exploring drivers for energy efficiency within small- andmedium-sized enterprises: first evidences from Italian manufacturing enter-prises. Applied Energy, in press.