2010 Bianchi System dynamics

& ResearchPaper

Applying System DynOrganizational Chang

th

Ca

Un2Ma3UP4Un

phenomena. To this end, three real cases are analysed to show how SD can facilitate a

,

l

taken into account when applying SD to enhance

SystemsResearch andBehavioral ScienceSyst. Res.27, 395^420 (2010)Published online inWiley InterScience(www.interscience.wiley.com)DOI:10.1002/sres.1042

Associate Professor in Economics and General Manager.organizational change programs in the publicsector, that is the:

Controller.kPhD Candidate in Public Management.#Full Professor in Public Management.INTRODUCTION

Applying the System Dynamics (SD) method tofoster organizational change, accountability and

to support effective decisionmaking in the publicsector implies a number of challenges which areinterrelated, not only because of the specificcomplexity of the public sector itself (Bianchi2010), but also as a consequence of the idiosyn-crasies that characterise a particular geographicaarea.Such issues can be referred to four main

interdependent sub-systems, that should be

*Correspondence to: Carmine Bianchi, c/o CED4 System DynamicsGroup, Via Mazzini, 59-90100 Palermo, Italy.E-mail: [email protected] Professor in Business and Public Management.zAssistant Professor in Business and Public Management.xbetter understanding of the relationships between the political and the organisationalsystem in the public sector, and how to promptly attain efficiency and improve outcome,given the constraints that the institutional and political systems constitute. Copyright#2010 John Wiley & Sons, Ltd.

Keywords public management; accountability; performance measurement; balanced scorecard;system dynamicsCopiversity of Palermo, Palermo, Italyssimo Theatre, Palermo, ItalyMC-Ismett Hospital, Palermo, Italyiversity of Benevento, Benevento, Italy

The goal of this paper is to discuss the role System Dynamics (SD) can play to enhanceperformance improvement in the public sector. It is remarked how SD can help decisionmakers to properly perceive the boundaries of the relevant system underlying observedPietro Ferrara3, Tiziana Landi4k an1rmine Bianchi1*,y, Enzo Bivona1z, Antonio Cognata1,2x,d Paolo Ricci4#and Performance inA Case-Based Italianyright # 2010 John Wiley & Sons, Ltd.amics to Fostere, Accountabilitye Public Sector:PerspectiveReceived July 2009Accepted April 2010

deals with relationships between the publicadministration system and external actors (e.g.citS

anactwhplarulincomsyspromaT

RESEARCHPAPER Syst. Res.themobje

relorimwefor

cocopoimprtivbe

Co

39may cause a failure in reaching thectives stated. Such a failure may arise fromtraditional organizational intervention, ignoring

may be considered outside the scope of a

the1. Socio-economic system;2. Political system;3. Institutional system;4. Organization system.

As Figure 1 suggests, the implementation ofprograms aimed at fostering change, account-ability and the improvement of service to users,and to thewider community, could be addressedby predominantly focussing on the organizationsystem (e.g. on processes, roles and respon-sibilities, planning and reporting, career andreward/incentive schemes), and at the sametime be hampered by a lack of attention to thewider socio-economic, political, and institu-tional systems with whom an organization in

public sector interacts. Although such factors

Figure 1 Four main sub-systems affecting organizationalchange interdependently in the public sector (adapted from

Borgonovi, 1996, p. 105)ying on the hypothesis that improving theganization system itself will, cause anprovement in the overall performance asll. However, this hypothesis may not be valida number of reasons.

First, a public sector organization is tightlynnected to other independent (looselyupled) public sector organizations, whoselicies may affect (positively or adversely) theplementation of its organizational changeograms (Bianchi, 2010). The rules and legisla-e schemes which define such relationships canreferred to as the institutional system that also

pyright 2010 JohnWiley & Sons,Ltd.

6leading) culture and values arising from theunderlying wider socio-economic systems, anypublic sector organizational change programought to take into account what opportunitiesthere exist to affect, in the long term, such aculture and values, possibly with the aim ofchanging the context (i.e. both the institutionaland political system) in which the managementoperates.When applied to the Italian context, the above

analysis is relevant to a wide scope of challengesassociated with an effective implementation ofpublic sector organization change programs, andthe use of SD to support them.In particular, this paper aims at highlighting

two challenges, that is the need to:

1. Properly address the relationships betweenthe political and the organizational (manage-rial) system. On this concern, primary researchquestions are the following:1.a How can SD facilitate a better under-

standing of the (often hidden) relation-ships between the two systems?

1.b How can SD foster a better communi-cation and strategic dialogue betweenpoliticians and managers, with the intentto improve policy making and imple-mentation?

2. Promptly attain efficiency and improve out-come at the same time, given the currentconstraints that the institutional and political

Syst. Res.27, 395^420 (2010)systinformal political system defines how power isually exerted and how the different rolesich are ruled by the institutional system areyed. This is affected by the hidden (social)es of conduct leading to set political prioritiesthe allocation of resources, and to outliningmunity goals. Ignoring such a politicaltem may significantly tackle any changegram which only takes into account thenagerial view of a public sector organization.hird, since both the institutional and politicalems are an expression of the emerging (orded by the formal institutional system; in fact,coizens and private sector organisations).1

econd, the above rules of conduct are not onlyDOI:10.1002/sres

Carmine Bianchiet al.

systems constitute. In this regard, primary

role SD can play to analyse and diagnose acoins(i.e20capotivsy

FOSTERING ACCOUNTABILITY INPUBLIC UTILITIES: THE ACQUA SPACASE-STUDY2

Introduction

Co

Fo

Syst. Res. RESEARCHPAPERmplex problem by combining a macro (i.e. thetitutional and political systems) with a micro. the organisation system) view (Bianchi,10). The third case emphasises the role SDn play in supporting the implementation oflicies aimed to recover efficiency and effec-eness at a micro level (i.e. the organisationstem).

pyright 2010 JohnWiley & Sons,Ltd.sterresearch questions are the following:2.a How can SD support management to keep

under control the key performance vari-ables driving to the success or failure oftheir organizations?

2.b How can SD support management toimprove social reporting on the outcomesgenerated by implemented strategies?

Based on a case analysis conducted by theauthors, in the next sections of this paper, we willprovide insights into the challenges listed above.While the three following cases are from quitedifferent contexts (i.e. water public utilities,health care, opera production and staging), theiranalyses suggest the existence of a number ofcommonalities that may offer possible answers tothe above research questions.In particular, the first case-study provides an

example of a Public Utility (PU) whose lack ofefficiency and effectiveness is due to hugepolitical interferences into the organisationsystem. The second case illustrates financialand social dysfunctions generated into the healthcare organisation system by setting surgerytariffs at political level disregarding the clinicalcontext. The third case shows how keeping undercontrol the key performance variables in amunicipal theatre may allow managers to attainefficiency and improve outcome.Though the analysis of all the three cases tries

to provide insights on both the above sets ofchallenges, the first two cases primarily addressthe research questions 1.a, 1.b and 2.b. The thirdcase is more focused on the research question 2.a.The first two case-studies also differ from the

third, if we consider the perspective throughwhich they are framed. In fact, they remark theOrganizational Change,Accountabilityand PeParticularly in the last decade, the debate on thecauses of the failure of PU performance improve-ment programs seems to be focused on two mainaspects, that is the intervention field and thereform of legal issues ruling the status of PUs.Under the first viewpoint, it has been shown

(World Bank, 2004) that the core of suchinterventions has been exclusively on the organ-ization system, rather than also on the widercontext in which such companies operate. Thisnarrow focus implies that, to improve PUperformance, it is sufficient to undertake areorganization process aimed to improve out-comes, efficiency, effectiveness and financialbalance.Under the second viewpoint, reforms have

been oriented to separate the responsibilities ofthe three main involved actors, that is: owners,Board and management (Baietti et al., 2006). Suchgoal has been primarily pursued by changing thelegal status of organisations, from a statutorybody to a limited corporation totally owned bylocal institutions. This policy, borrowed from theprivate sector, has proven ineffective since it doesnot address the public sectors peculiar complex-ity. In particular, PU ownership has not beenseparated from external political interventionsaimed to satisfy the interest of restricted groupsof stakeholders with the intent to gain consensusleading to re-election (Shapiro and Willig, 1990;Shleifer and Vishny, 1994).To improve PUs performance, it is not

sufficient to act into the PU organization perse. On the contrary, it is necessary to outline therelationships between the institutional, politicaland socio-economic systems, providing thecontext in which PUs operate. At least in Italy,such relationships often generate unsatisfactoryresults, in terms of both financial and serviceperformance.

2This case-study has been written by Carmine Bianchi, Enzo Bivona,Tiziana Landi and Paolo Ricci.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres

rformance in the Public Sector 397

main external and internal factors affecting

Table 1 General characteristics of ACQUA

2005 2006 2007

4520

6017

N5,8NN0

39,37,s3

RESEARCHPAPER Syst. Res.A major cause of this phenomenon can beattributed to the high influence of politicians inappointing the PUs Board of Directors to satisfytheir own parties interest and to promote socialconsensus aimed to re-election.A related cause of poor PU performance can be

also attributed to the wider socio-economicsystem where such institutions operate. In fact,a political system in which people culture and

General characteristics of ACQUA

Population servedConnectionsNumbers of employeesWorking ratioStaff per 1000 connectionsStaff per 1000 population servedAccounts receivable as a share ofannual revenue, expressed in months salesWater supplyWater supply turnover (m3) 3Unaccounted for water (m3)Unaccounted for water (%)Average domestic tariff (s/m3)Revenues sOperating costs sNet profit (loss)dominant values are oriented to the search ofpersonnel favour and influence peddling mayprovide a fertile field for unethical behaviourleading to inequality, inefficiency and pooroutcomes for the citizen. Therefore, even thebest change-oriented programs focused on theorganisation system and the best legislativeinnovations aimed to regulate the institutionalsystem (in particular, concerning the Municipal-ity-PU relationship) are destined to fail if asystems approach is not adopted.On the basis of the above remarks, the next

section will illustrate a public water utility case-study operating in Southern Italy and showpreliminary findings about main factors affectingPU performance. The case-study analysis hasbeen conducted based on the conceptual frame-work previously used in an internationalresearch (Baietti et al., 2006) aimed at outlining

Copyright 2010 JohnWiley & Sons,Ltd.

398public water utilities performance.By the light of the case-study analysis, some

vicious circles leading to PU performancedeterioration due to political interferences havebeen outlined and alternative interventionpolicies explored.0,000 450,000 451,2902,748 205,462 208,474409 410 4064% 64% 64%.80 0.81 0.82.77 1.77 1.786% 89% 88%

/A 69,783,898 67,482,42836,590 36,149,555 36,376,354/A 33,634,343 31,106,074/A 48 46.63 0.63 0.63709,699 s37,422,815 s35,514,684801,388 s38,615,585 s44,862,99161.854 s27.864,00 s (9.480.079,00)The ACQUA SpA Case-Study

Introduction and General DescriptionACQUA SpA3 is a joint stock company estab-lished in 2003. It is owned by a large number ofsmall municipalities (about 128) located inSouthern Italy, and provides water supply andsewerage services to more than 450,000 users.Table 1 reported below summarises main com-pany data.In order tomake explicit main decisionmaking

processes affecting PU performance, the case-study analysis has been carried out by taking intoaccount two complimentary perspectives. Thefirst refers to the institutional environment inwhich the PU operates. The second identifies themain internal factors affecting PU performance.

3The name of the PU has been intentionally disguised.

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


The Institutional EnvironmentThe Institutional environment can be associatedwith two main issues: the external autonomy andthe external accountability. The first relates to thedegree of independence of the PUmanagers fromexternal interferences in decision-making. Thesecond issue refers to the extent the PU respondsto external stakeholders on precise performanceindicators, based on which incentives are pro-vided.In terms of external autonomy, as ACQUA SpA

In order to recover such a negative perform-

Syst. Res. RESEARCHPAPERis a joint stock company, it has a strong formalindependent status. The utility is governed by aboard of directors consisting of five persons, whoare appointed by municipalities. However, sinceACQUA SpA is owned by a large number ofsmall municipalities, this often tackles the abilityto achieve mutually agreed decisions on Boardnominations. Therefore, political parties oftentake a leading role in designating the members ofsuch boards. This turns the PU external auto-nomy into a formal state only.Members of the Board are often appointed on

the basis of their political affiliation4and rarelyon their professional experience. This phenom-enon is consistent with the research conducted byGutierrez and Menozzi (2008) on 114 Italianpublic utilities in a 10-year period. In particular,the study demonstrates that unnecessary staff isoften hired, through politicians influences on theBoard, and such decisions have a negative impacton PU performance.Such phenomenon can be empirically found in

the ACQUA SpA, and explains its significantfinancial losses recorded in 2007. In particular, bycomparing 2006 and 2007 results, it is possible todetect a 5% decrease in revenues and, at the sametime, a 14% increase in the total costs.A significantpart of such an increase is due to personnel costs.5

4According to Gutierrez and Menozzi, and the prevailing literature(Faccio, 2005), a director has a political affiliation when he/she hasalternatively one of the following roles: (a) an active position (or hasbeen active) in the recent past in the political arena, (b) an officialpolitical position (for instance, deputy of the national Parliament,member of Regional, Provincial or Municipal governments), (c) been(or is) candidate for election, (d) a membership in a political party or afriendship with people associated with a political party.5This phenomenon can not be explained by looking at the operationalside of the company, as the population served is stable. On thecontrary, the increase in (unnecessary) human resources could beinterpreted as a result of political pressures on the PU.

Copyright 2010 JohnWiley & Sons,Ltd.Foster Organizational Change,Accountabilityand Peance, ACQUA SpA can not autonomously definethe water tariff. In fact, water distribution tariff isdetermined by a national law, based on theinvestment plan adopted by the PU, and shouldallow the PU to earn a 7% annual return oncapital invested.Concerning the external accountability, the

current PU targets (in terms of revenues, waterproduction, distribution, consumption, unac-counted water and new connections), that aremade explicit in the annual budget and sub-mitted by the board of directors to shareholders,are quite easily to achieve in comparison to theavailable resources. For instance, there is not asignificant evidence of the effort done by thecompany to recover the very high percentage(about the 50%) of the annual unaccountedwater.The above remarks also suggest that perform-

ance targets are not perceived by shareholders asa way of pressuring the utility to improve itsresults. Rather, as long as ACQUA SpA accom-plishes its main task of providing reliable waterservices, shareholders do not intervene.

The PU Organization SettingAmong the internal factors that are likely to affectPU performance, the degrees of internal autonomyand accountability, and customer orientation can betaken into account.The decision making process in ACQUA SpA

is highly centralised. Although the company isorganised in different departments, each head ofthese units has a limited decision power and verybounded autonomy. In fact, most decisions aremade by the managing director and approved bythe board of directors. The board mainly focuseson strategic decision making and the managingdirector focuses on only operational decisions.The decisions that are typically made by themanaging director, and authorised by the Board,are those related to the approval of the procure-ment of goods and services and the hiring ofemployees. As a consequence, the above de-cisions can be influenced by those politicianswhoappointed the board members.The high level of influence of political parties

through the Board on the PU strategic decisionscan be also explained by the low level of

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


company internal accountability and transpar-

reduces performance. A lower performance,increases debts and the degree of externaldependence, and provides the basis for furtherillegal political interventions on company activi-ties.

Concluding Remarks on the ACQUA SpACase

Based on the above analysis, it is possible tooutline intervention policies to allow the PU torecover sustainable competitiveness and financialequilibrium, at least in the long run. The possiblelevers onwhich to act should be related to the sub-

Syst. Res.27, 395^420 (2010)

RESEARCHPAPER Syst. Res.ency. For instance, the company does not adoptany system to evaluate employees performance.Although some managers undertake employeesevaluations at their own initiative, these evalu-ations are not structured. The formal respect ofprocedures is perceived as more important thanmeeting performance targets.Furthermore, ACQUA SpA has a weak orien-

tation towards customers. This can be detectedby several factors. First, there is not a formalprocedure on how to deal with customerscomplaints and response times to complaintsare not available. Second, benchmarking is in itsinfant stage. It focuses on tariffs, water quality,level of services and efficiency. However, itsoutcomes are not used in company strategic oroperational plans. Finally, although the companyintroduced a customer charter, it does notconduct any market survey aimed at detectingthe perception of the level of customer satisfac-tion.

Vicious Circles Leading to ACQUA SpA Perfor-mance Deterioration due to Political InterferencesA feedback analysis can be helpful in high-lighting the vicious loops that have been affectingthe company performance. On this regard,staffing and procurement processes will beanalysed.Figure 2 shows how overstaffing, that is the

hiring of unnecessary (and often unqualified)staff due political interferences into the companymanagement, decreases staff motivation andskills, the level of service provided to customersand PU performance. In turn, the poor PUperformance raises municipality subsidies tofinancially sustain the organisation. At the sametime, such public fund intervention empowersmunicipal political representatives to intervenein the PU decision making process, for instance,by recommending the desired people to hire(see reinforcing loop R1 in Figure 2).The above interference in the managerial

decision making process is also likely to producea decline in the level of PU managementautonomy and, hence motivation, leading to afurther drop in PU performance (see reinforcingloop R2 in Figure 2).

Copyright 2010 JohnWiley & Sons,Ltd.400Finally, overstaffing also produces an increasein personnel costs, without a significant contri-bution in terms of efficiency or effectiveness. Thisphenomenon boosts operating costs, whichcauses a decay in PU results, and reinforcespoliticians interferences (see reinforcing loop R3in Figure 2).Another area of political interferences in the

PU autonomy refers to suppliers selection. Thisis likely to generate two other vicious circlesleading to PU performance deterioration (seereinforcing loop R4 and R5 in Figure 3). In fact, anintervention in the managerial decision processreduces PU managers autonomy, leading to adrastic drop in their motivation. Moreover, thisphenomenon increases operating costs (e.g. dueto low quality raw materials), and therefore

Figure 2 Vicious circles leading to PU performancedeterioration due to political pressure to hire desired (unqua-

lified) staffDOI:10.1002/sres


systems described in Figure 1, that is: theinstitutional, political and public utility systems.

survival of the community. As the above B1 loopshows, the positive effects produced by changesin rules determine a reduction in politicalinterferences.Concerning the second subsystem (i.e. the

political one), another balancing loop could befostered by introducing new frameworks implyingsocial/political accountability. The adoption of suchframeworks would allow the community toevaluate the effectiveness of political actionand related outcome. In order to achieve thisgoal, however, it is also necessary that thecommunity develops a culture oriented towardsthe proper role of the PUs in society (e.g. to satisfycitizens needs), rather than perceiving suchcompanies as a means to opportunistically obtainfavours from politicians through influence ped-dling. As the B2 loop in Figure 4 shows, such a

Figure 3 Vicious circles leading to PU performancedeterioration due to political pressure into the procurement

processes

Syst. Res. RESEARCHPAPERFigure 4 shows that, in order to counterbalancethe vicious effects generated by the reinforcingloop R1, a first balancing loop could be fosteredby introducing consistent changes in the system ofrules, which discipline the roles and responsi-bilities of various actors operating not onlyinside, but also outside the PU. As the loop B1in Figure 4 shows, the need for such policyshould be directly related to the level of politicalinterferences. On this regard, therefore, it iscritical not only the quality of adopted changes inthe system of rules, but also the time and theextent to which the phenomenon of politicalinterferences is perceived as a threat to theFigure 4 Intervention policies to recover PU sustai


Foster Organizational Change,Accountabilityand Pepolicy can be fostered by a proper and timelyperception of poor PU performance and itsnegative effects on the social costs of poorservices. This would determine a drop of socialconsensus for re-election, leading to a reductionin political interferences.Concerning the third subsystem (i.e. the

organisation sub-system), another balancing loopcould be fostered by introducing new mechan-isms inside the PU, aimed to improve the qualityof management systems. Also in this case, such apolicy should be fostered by the perception ofunsatisfactory PU performance levels. The adop-tion of proper systems to increase the quality of

nable competitiveness and financial equilibrium

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


managers, together with reduced political influ-ences leading to higher managerial autonomy(and possibly also supported by the hiring ofskilled employees), would determine an increase

regional health departments, are responsible

made through a system called DRGDiagnosis-

a groupecalled IC

acations orSince pa

aa

both hosp

Hospital P

as a medical or a surgical type

RESEARCHPAPER Syst. Res.for ensuring the delivery of a benefit packagethrough a network of population-based healthmanagement Organizations (local health units)and public and private accredited hospitals.Hospital carehas always represented the largest

share of health care expenditure and has oftenbeen a source of major concern to the centralgovernment. In the 1990s the Italian governmenthas established that the hospital payments will be

6This case-study has been written by Carmine Bianchi and PietroFerrara.

Copyright 2010 JohnWiley & Sons,Ltd.in PU performance (see loop B3 in Figure 4). Infact, a PU performance lower than a desired level,would determine the need of a higher quality inmanagement systems (e.g. in terms of hiringqualified personnel, improving organisationstructure and processes, information, planningand control systems). A higher quality ofmanage-ment systems would contribute to increase PUperformance.

LINKING POLITICAL WITHMANAGERIAL AND MEDICALPERSPECTIVES IN DRG SETTING: FROMAN OUTPUT TO A PATIENT-OUTCOMEVIEWPOINT. THE ISMETT CASE-STUDY6

Introduction

In Italy the majority of healthcare is provided bythe State, funded from general taxation. Theprovider is called SSN, that is Servizio SanitarioNazionale (National Healthcare System). Italyshealth care system is a regionally based nationalhealth service that provides universal coveragefree of charge at the point of service.The system is organized at a national and

regional level. The national level is responsiblefor ensuring the general objectives and funda-mental principles of the national health caresystem. Regional governments, through the402ALOS The average length of stay indicates theduration the patient stay for eachclassification

Weight Each of the DRGs has assigned a weightthat measures the amount of resources ana average basis that are required for thatpatient stay

Fee The DRG fee is the amount of revenuethat the hospital will receive for thatspecific classification

Most of the standard performance indicatorsthat are DRG-related provide quantity typemeasurements such as volumes, surgical percen-tages, patient days, occupancy rates, etc.There are even some quality measurements;

probably the most important and most utilized isthe DRG weight. This is called even a complexity

Syst. Res.27, 395^420 (2010)The DRG system provides a large amount ofinformation and many quantity measurementsand a quality measurements can be provided.The most frequent information/indicators uti-lized are:

Volumes Patients discharged: inpatient and dayhospital/day surgery

Type Each of the approx 500 DRGs is classifiederformance Measurementsformancel resources, this DRG system is used forital payments and for hospital per-measurements.clinicallyof hospitco-morbidities.tients within each category are similarnd are expected to use the same levelcedures,r software based on a coding systemD-9-CM, indicating diagnoses, pro-ge, sex, and the presence of compli-related group (Hsiao et al., 1986).The DRG system has been developed and used

in USA since 1983, with the purpose to determinehow much the Government (Medicare) pays thehospital. It is a system to classify hospital casesinto one of approximately 500 groups (also eachcalled simply DRGs), expected to have similarhospital resource use. The DRGs are assigned byDOI:10.1002/sres


index that qualifies the level of complexity of a

cation.

RGTechnological P

and

prevention of recurrent variceal

effectiveness of TIPS. In fact, since the onlydifference between the two treatments was thenumber of reinterventions for maintaining theshunt patent, it is conceivable that overcomingthis problem would result not only in a markedreduction of the overall treatment cost, but also inan appreciable improvement of the quality of life.In fact, the patient would reduce the times ofhospitalization.From a DRG stand point, which sometimes

becomes the hospital perspective, the resultscould be interpreted significantly differently. The

7ISMETT is an hospital based in Palermo and related to the PittsburghUniversity, USA. ISMETT is an excellent centre for transplants in theSouthern Italy.

Syst. Res.27, 395^420 (2010)

Syst. Res. RESEARCHPAPERbleeding

Beside any consideration of clinical advantage,which is not the scope of this paper, the interestlays on the alternatives of medical suppliesutilized for this procedure and their effect onthe results for the patients outcome.In the beginning, for the TIPS procedure was

used a bare stent that presented the disadvantageof requiring a strict surveillance program and

Copyright 2010 JohnWiley & Sons,Ltd.the above issues. To what extent does the DRGfoster or hinder technological progress in medi-cine?Such question is relevant to the subject of this

paper, since it refers to the relationships betweenthe political and organisation system, in thewider public sector. In this case, the politicalsystem settles the DRG policy, while theorganisation system (i.e. hospitals) managesservice provision and related technologicalimplications.To address the above question, a case example

will be considered: the trans-jugular intra-hepatic portal systemic shunt (TIPS).

DRG N. 191and 192

Major procedure on pancreas, liverplus shunt with/withoutcomplication

Procedure TIPSDescription This is an interventional radiology

procedure called TIPS for theA puzzlingFoster Organizarogress in Medicine?

recurring problem is related toDoes the D Classification Capture thespecific hospital.This type of performance indicators does not

cope the need to include for each classificationsome process measurement indicators (e.g.,drugs, medical supplies, physician decisions,etc.). This need is originated by the convictionthat each patient in several cases is unique andhis/her treatment therefore could be signifi-cantly different inside the same DRG classifi-tional Change,Accountabilityand Perepeated reinterventions (several procedures andhospitalization for the patient) for preserving itspatency, with a marked increase of the overallcost over that of the initial procedure.Afterwards, the technological progress in

medicine has introduced new and advancedmaterials, in fact the rates of stent dysfunctionand reinterventions have been significantlyreduced by the utilization of these new coveredstents. The experience at ISMETT,7 with 106patients who underwent bare stent and 111covered stent TIPS, further confirms these results(see Figure 5).The use of covered stent (nowadays the

standard of use) would much improve the cost

Figure 5 Cumulative proportion of patients free of TIPSdysfunction according to the type of stent received (numbersbelow the abscissa are patients still free of TIPS dysfunctionper each observation period. Patients were treated at

ISMETT, Palermo)DOI:10.1002/sres


DRG revenue using either type of stents (bare or

sequently earn an extra DRG revenue for each

all hospitals mission.

highlighting the outcomes originated by regu-latory policies undertaken by public sectororganizations in order to foster efficiency andeffectiveness in health care systems.8

On this regard, the so called DRG-creepphenomenon has been emphasized by variousauthors. This implies that patients are placed inhigher priced DRGs than their actual state of

RESEARCHPAPER Syst. Res.On this regard, a number of questions forpolicy makers should be considered whiledrafting the national and regional health plan:is there a detachment in the objectives between thepolicy maker (regional government) and the processmaker (physician or hospital administrator)? Couldthe performance indicators contemplate the patientoutcome?

A Preliminary Analysis of the ProblemThrough the System Dynamics Approach

Unintended consequences associated to DRGfunding are not a new phenomenon. Particularlyin the last decade, several studies have been

Copyright 2010 JohnWiley & Sons,Ltd.time that patient repeats the procedure. This isthe second factor that could incentive the hospitalin using a certain protocol.It is possible that if the hospital would follow a

blind path towards the Profit and Loss statement,associated with the weaknesses of the DRGsystem to detect these differences, the patientoutcome would not be addressed.

Implications for Public Policy/Management,and System Dynamics Modelling

DRGs are a well-established provider paymentsystem. Because of their imminent potential ofcost reduction, they have beenwidely introducedby many governments.But we have seen that there could be potential

pitfall of the objectives between the DRG systemand patient outcome. The first is a mandatoryhospital measurement tool that provides ameasure of revenue (cost) and productivity.The second should be the fundamental part ofcoated) is the same, but the hospitals cost onlyfor the stent is much different (the bare cost 25%compared to the coated). This is the first potentialadvantage.Adopting the bare stent the hospital would

probably need to have the patient repeat theprocedure every 1218 months, and con-404health would warrant (Christensen et al., 2004).Other unintended consequences of DRG fund-

ing have been related to the tendency of somehospitals to increase the volume of patients, inorder to maximize the flow of revenues (beingDRGs related to the diagnosis that originates ahospital service), or to affect the case-mix (e.g. byencouraging the admission of those patientswhose DRG is higher than others) (Mikkola et al.,2001; Kamke, 1998; Sheigold, 1986). Early dis-charge rates have also been referred as commonpractices in the field (Cantu` et al., 2008).Related to this phenomenon, not only a typical

problem of goals avoidance driven by opportu-nistic behaviour following the introduction ofperformance measurement frameworks, but alsoa service quality phenomenon arises. In fact, afocus of performance measurement frameworkson only outputs, rather than also outcomes, atleast in the perspective of hospitals, has been aprimary cause of their inclination to dismisspatients too quick (and often also too sick). Forthis reason, the Swiss Cantons of Zurich andBerne have recently re-modulated their ownDRG system on an outcome, rather than on aninput and output-only basis (Draghic, 2006).Similar problems can be envisaged concerning

the previously described theme, that is: capturingmedical and technological progress in DRGframework systems. The case-study describedabove clearly illustrates how anchoring DRGs ononly a nominal classification of medical orsurgical type, rather than also on the technologyused in medical care, can be a major cause of

8More widely speaking, health care has proven a fertile field for SDmodelling. Though a detailed analysis of the SD literature is not thefocus of this study, we may here remind some of the most represen-tative studies covering both macro (e.g. regional health care systems;Vennix and Gubbels, 1994; Wolstenholme, 1999) and micro (e.g. hos-pitals; Lane and Huseman, 2008) perspectives of applications. Inparticular, on SD applied to DRG policies, see: Maliapen and Danger-field, 2009; Trailer and McLaughlin, 2007.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


of the flow of patients needing reinterventions.Such a phenomenon may determine a newstrong increase in the current regional treat-ment costs.The above described series of events generates

Syst. Res. RESEARCHPAPERunintended negative effects, leading to higherheath care system costs on a longer timeperspective.Figure 6 shows how according to the current

policy setting a given threshold in regionalhealth care costs implies a reduction in DRGs, ifthe current regional expenses for healthcare arehigher than the budget (i.e. the threshold). Thebalancing loop depicted in the figure shows theeffect that such a policy is meant to generate: thehigher costs, the higher the gap will be; a highergap will reduce DRGs (or will contain them); thiswould lead to lower costs, that would allow theRegional administration to meet the budgetthreshold.In the case here described, this phenomenon is

observed in the short term. In fact, such a policyonly aims to foster efficiency improvements inthe different organizations operating in aregional healthcare system.Nevertheless, such an indiscriminate cost-

Figure 6 Short and long term effects from myopic DRGsetting by Regional policy makers.reducing and (more specifically) revenuecutting policy naturally generates an incli-nation of hospitals to reduce the quality of theirservices. In such case, this implies a lowerinclination to adopt more sophisticated andinnovative surgical technologies, since theircosts would not permit hospitals to meet theDRG targets.Therefore, such a myopic DRG policy is

likely to generate a lower inclination toinnovate in hospital heath care. This would,in turn, lead after a delay to an increase inthe stock of patients needing treatment, because


Foster Organizational Change,Accountabilityand Pein the long run a reinforcing loop, linking costs toDRGs, and DRGs to the inclination to innovate,to patients needing retreatment and to costsagain.The same phenomena can be depicted through

a SD simulation model (Figure 7).9

The basic structure of this insight model,consisting of the two interconnected stocks, aimsto reproduce the same behaviour observed inFigure 5.It is hypothesized here that a stock of 100

patients needing treatment is drained by the flowof new treated patients, as an outflow of patientsreceiving surgical operations (the average recov-ery time is here an input parameter, equal to 12months). Such an outflow accumulates into thestock of treated patients, whose initial value hasbeen supposed also 100 people. This last stock isdrained by the outflow named Patients needingretreatment. Such variable is the ratio betweenthe Treated patients stock and the averageretreatment time associated to the adoptedtechnology: the bar stent technology implies a30 months, while the covered stent technologyimplies a 220 months period.10

The results portrayed by the simulation modelconfirm how the current policy (which is at leastimplicitly oriented to adopt the bar stenttechnology) tends to increase, over a 4-yearperiod, the level of patients needing treatment.Such phenomenon is due to the higher volume ofthe flow of patients needing retreatment over thesame time horizon.The model also shows that a higher stock of

patients needing treatment for the first policy,in respect to the second one originates a higherflow of new treated patients, that is a greaternumber of surgical operations. Therefore, suchoperations generate unintended higher costs forthe regional heath care system.

9Model Equations are available on request from the authors.10The above two parameters have been set in a way that given a stockequal to 100 treated patients the corresponding survival curves canapproximate to those in Figure 5.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


RESEARCHPAPER Syst. Res.This phenomenon is more evident if weinclude in the model another inflow into thestock of Patients needing treatment. Such aninflow corresponds to the new patients to treat,that is to the number of new surgery treatmentsto make for the first time on patients. The testdata for this inflow has been set to a constant of 4patients per month.On this regard, Figure 8 shows that while the

dynamics of such stock portrays an increasingpattern of behaviour in the first run, it ratherdeclines in the second run. The cause of thisphenomenon is associated to the higher stock oftreated patients, leading to higher volumes ofpatients needing retreatment over time. From thesecond graph of Figure 8 it is possible to observehow the higher value of this stock generates amuch higher volume of patients needing a new

Figure 7 A stock- and flow-diagram model showing short and lomakers: Effects of Regional


406surgical operation, if the first current policy (barstent) is adopted.Key:

Policy 1: Bar stent policy (solid line).Policy 2: Covered stent policy (dotted line).

The insight SD model also embodies an analysisof the regional DRG setting policies. As shown inthe stock and flow model depicted in Figure 10here we have a threshold for regional treatmentcosts on the above TIP disease treatment, that is:s70.000 per month.11The corresponding current regional treatment

costs are calculated as a product between the flowof new treated patients and the average unitDRG.

ng term effects from myopic DRG setting by Regional policyDRG Policies on costs

11This is also a test rather than empirical value.

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


Syst. Res. RESEARCHPAPERBased on such threshold, the current regionalpolicy is to initially set the DRG at s8.400 pertreated patient.12 Such a parameter is related tothe Low DRG policy factor, which currentlydetermines a high percentage of cases treatedthrough the bar stent technology. The modelshows that such DRG is set according to the ratiobetween the regional treatment cost thresholdand the perceived current number of treatedpatients.13 Therefore, during the simulation, ifthe DRG is based on the current approach (Low

Figure 8 Results from two simulation ru

12Based on such DRG, current regional treatment costs per month, atthe beginning of the simulation, meet the budget, that is: s8.400 8.33patients/months70.000/month.13Due to the relative difficulty to reduce the DRG in the short time, theobservation delay for the New treated patients flow has been set to 5years.


Foster Organizational Change,Accountabilityand PeDRG policy), the average unit DRGwill decreaseif (given a threshold level for TIP diseasetreatment) the flow of new treated patients willincrease.According to an alternative hypothesis, the

DRG on TIP disease treatment recognized tohospitals wishing to adopt the covered stenttechnology, could be set to s16.000. Althoughsuch a substantial increase would not represent amultiple of 4 (i.e. the estimated cost ratio betweenthe two technologies), with respect to the DRGrecognized to the hospitals using the oldertechnology, this higher DRG might encourageinnovation policies in surgery operations. Simu-lation results will show now that an even higherDRGwould be unsustainable, in consideration ofthe budget threshold.

ns (second version of the SD model)

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


From Figure 7 it is also possible to observe thatthe policy maker may operate through a checkbutton and a slider bar, decisions on DRGpolicies. Precisely, the simulator policy switchenables one to use either the current policy

Threshold for regional treatment costs (small-

care would be a major driver.

RESEARCHPAPER Syst. Res.(implying a low and undifferentiated DRG) or ahigher DRG policy. If the player checks this lastoption, then the simulator will use the (higher)DRG, that is s16.000 per patient, as an inputparameter. Alternatively, the basic policy at thebeginning of the simulation is to set the DRG ats8.400 per patient.The simulation results portrayed in Figure 9

show the costs associated with the twopolicies.14

The two simulation runs here show that,although the higher DRG policy implies inthe short term much higher costs than the lowDRG policy, it becomesmore favourable after 2.5years, because of the lower total DRGs associatedto patients retreatment (see Gap in currentregional treatment costs dynamics).In the long run, such a benefit tends to decline,

so that at the end of the simulation bothalternative policy results portray the same gapin current regional treatment costs. Therefore,none of them is able by itself to meetthe target. While the higher DRG policy ismore expensive in the first period and moreeffective in the second part of the simulation, thelow DRG policy shows a just opposite beha-viour.However, it is worth remarking that the

declining increase rate of the gap in the currentregional treatment costs, related to the low DRGpolicy, is not due to a higher effectiveness of theoutcomes it generates. It only depends on thegradual reduction in the average unit DRGadopted by the regional policy makers becauseof the increasing volume of new treated patients in order to meet the budget threshold (see firstgraph in Figure 9).Key:

Policy 1: Bar stent policy (solid line).Policy 2: Covered stent policy (dotted line).

14Here, the simulation time horizon has been extended from 4 to 7years, in order to better appreciate the expected long-term effects ofadopted policies on costs.

Copyright 2010 JohnWiley & Sons,Ltd.408Therefore, in the long run, a higher capability ofthe system to increase the value of health (at properefficiency standards) should be a measure thatcould lead to a higher potential of a Region tomake the acquisition and allocation of financialresources to its health care system sustainable inthe long run (see R2 reinforcing loop inFigure 10).

Concluding Remarks on the ISMETT Case

The above analysis has shown that both simu-lations policy results show that the budgetthreshold is not met. This is not only, unlikely,the result of an insight model with just test data.It is, rather, one of the significant dynamiccomplex problems that policy makers now facein many Italian Regions.The second simulation as said seems to

perform better than the first one, at least in themedium long-run.However, also the second simulation run does

not portray any radical solution to the problem ofmeeting budget thresholds. In this case, resultscan be preferable only because the DRG policycan affect hospital behaviour not only on aninternal efficiency (i.e. resource consumption),

Syst. Res.27, 395^420 (2010)dotted line).

However, such a lower DRG could prove to befinancially unsustainable for hospitals, also in thecase of adoption of a more traditional and lessexpensive technology. Therefore though themodel does not embody such set of possibleevents this could determine further inefficien-cies (i.e. higher costs for the system) or furtherdrops in the levels of outcomes, which would inturn determine a lower generated value on aregional level.Such hypothesis opens up the boundaries of

our model to the analysis of factors impacting onthe outcomes of health care policies. Theseoutcomes can be considered as a measure ofthe capability of the regional health care systemto generate (i.e. to add) value to the community.To this end, the inclination to innovate in hospitalDOI:10.1002/sres


Figure 9 Results from two simulation runs

Copyright 2010 JohnWiley & Sons,Ltd. Syst. Res.27, 395^420 (2010)DOI:10.1002/sres

Foster Organizational Change,Accountabilityand Performance in the Public Sector 409

Syst. Res. RESEARCHPAPER

IMPROVING PERFORMANCE THROUGHA DYNAMIC BALANCED SCORECARDFOR OPERA HOUSES15

Introduction

RESEARCHPAPER Syst. Res.but also on an outcome perspective. This lastviewpoint, in turn, proves to be relevant also forthe efficiency of the regional system.SDmodelling and simulation can substantially

help decision makers on a political and manage-rial level to increase the level of synergy of theirown behaviour, with a view to improving theperformance of the overall relevant system, thatis in this case the regional health care system.This is possible through various alternative orcomplementary ways, which may imply, forinstance: assisted simulation sessions centred oninsight models, group model building sessions(Vennix, 1996), the development and use ofmodels for planning and the assessment of policy

Figure 10 Impact of value generated by health care serviceson the sustainability in the long run of DRG set by Regional

policy makersoutcomes.Obviously radical improvements in the sys-

tems performance e.g. in terms of bothmeetingbudget costs thresholds and outcome levels,could be pursued by deeper and wider scopeinterventions, for example aimed at fosteringpreventive care, or continuous home-care, so tobe able to decrease the two stocks of patientsdepicted by the model, and therefore the costs ofhospital care.Further applied research also implying a

direct involvement in modelling by decisionmakers operating in the political, managerial,and medical settings, will be needed in order tobetter explore such a dynamic complex andconsiderable problem for our society.


410Nonprofit art firms, especially performing artsfirms, have to deal with a difficult problem: overtime costs tend to rise more rapidly thenrevenues, thus generating a widening budgetgap. This is why most of the economic literature(Heilbrun and Gray, 2001) on the subject focuseson how demand can be increased in order to raiserevenue and on what performing arts firms cando to lower their production costs.Since Baumol and Bowen (1966) pioneering

study, economics offers an explanation to thisproblem, an explanation called cost disease. Theheart of the cost disease theory posits the inabilityof performing arts firms to enjoy productivityadvances relative to the rest of the economy. Theargument can be summarized as follows: It takes70 musicians as much playing time to perform aBeethoven symphony today as it did when thepiece was written. In other words, since theperformers labour is the output, there is no wayto increase output per unit of time. However, atthe same time, new productivity-enhancingtechnologies have created huge increases inlabour productivity in other sectors of theeconomy. Now, since wages tend to rise withproductivity all over the economy, the upwardpressure in the progressive sectors will deter-mine rising wages also in the stagnant sectors(the orchestra). But since musicians productivityhas not changed, these higher wages in orchestraare not compensated by higher revenues. Theresult is that, over time, costs rise relative torevenue. According to this explanation, the onlyway for a nonprofit performing art firm to dealwith a widening budget gap is to increaseexternal resources, for example actively seekingprivate grants and donations and/or asking formore government funds.Besides the cost disease, other arguments have

been invoked to justify public support to some

15This case-study has been written by Carmine Bianchi and AntonioCognata.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


forms of artistic activities. The oldest and most conductors, stage directors, designers and sing-ers. This has inevitably led to a great similarity inthe core product and standards of the operahouses. Therefore, the art-form itself dictates to agreat extent the resources and the organizational

16

the Massimo Theatre was to implement an

Syst. Res. RESEARCHPAPERoften invoked argument is that art, whatever itsform, is a public good (Frey and Pommerehne,1989; Throsby, 2001). According to this argu-ment, art benefits not only those who consumeand pay for it, but also all other citizens, who donot necessarily wish to contribute voluntarily toits production or to its preservation. For instance,museums, opera houses, orchestras and artfestivals attract tourists and visitors and generatespillover effects on hotels, restaurants and thelike. In this way, some people and firms gain abenefit for which they do not pay, and whichtherefore do not influence the production de-cisions on art markets.Art is also said to be a merit good, a means of

educating the public taste. Rather than acceptingconsumers preferences, political decisionmakers should decide production levels basedon the inherent worth of the artistic good. Anincome distribution argument posits that theconsumption of arts should be open to all classesof society and should not be reserved for the rich.Therefore, the government should support thearts in order to make consumption available tothose who are not able to pay for them.Of course, on the subject of government

support of the arts, there are also manycounter-arguments. Some economists are uncon-vinced of the relevance of the market failurearguments and of the importance of the sup-posed external effects. Public choice theorists(Grampp, 1989) argue that government supportis often used by some people for their ownbenefits. Empirical studies have shown that thecost disease is not as severe as argued; that lowprice elasticity of demand for performing artsmakes price increase possible; that technicalprogress (e.g. DVD or broadcasting) may enlargeaudiences. Additionally, careful distributionalanalysis has shown that in some instances high-income recipients are the major consumers ofcultural goods, and thus the main beneficiaries ofgovernment support.Opera is the Queen of live performing arts, and

in some way a conventional art form. Today allthe major opera houses around the world presentrelatively similar programmes, produced andperformed by a core group of international

Copyright 2010 JohnWiley & Sons,Ltd.Foster Organizational Change,Accountabilityand Peappropriate production plan with the aim of

16For example, to stage Verdis Aida requires a fairly specific numberof skilled performers, a certain type of sets and performance space,wherever it is performed, in order to comply with the conventions ofthe art form and with the standards of the world of opera.

Syst. Res.27, 395^420 (2010)structures needed to produce opera.

The Massimo Theatre in Palermo

With its budget of about s42 millions and morethan 400 employees, the Massimo Theatre is thethird largest public opera company of Italy. Morethan 90% of the revenue comes from publicsubsidies, the remaining 10% comes from the boxoffice and the sale of other services. Salaries forfull-time employees represent more than 70% ofproduction costs.In the fiscal year 2004 theMassimo Theatre was

in a very difficult financial situation: losses fors1.5 millions, short-term bank debts for abouts26 millions and production at its minimum. Anew management took over at the beginning of2005 with the aim of preparing and implement-ing a 3-year plan to restore viable financialconditions.As said, there is essentially one option

available for the opera management to deal witha structural earnings gap: reducing the gap byincreasing the share of earning income relative tototal operating costs. This strategy can take twoforms: increasing the earned income or reducingoperating costs. On the latter side, the two mostimportant moves have been to bargain a newcontract with the labour unions on a much moreflexible base, and to negotiate a bankmortgage tolower debt costs. On the former side, despite thebenefits of increasing earned income, there arelimits to how far opera companies can go in thisdirection. In fact, increasing the number ofproductions brings not only increased revenue,but also increased costs. The strategy followed atDOI:10.1002/sres


maximizing the number of performances, giventhe

three groups: at least 3 blockbuster titles (each of

the Balanced Scorecard (BSC), that is: custo-mer, internal processes, learning & growth,and financial.

2. Understand the key linkages between suchindicators and related performance drivers.

17

num

RESEARCHPAPER Syst. Res.which is staged for 812 performances), 34repertory titles (78 performances), and 12 rareand/or contemporary operas (46 perform-ances). Since at the Massimo Theatre patronssubscribe for the whole season, a mix ofrepertoire is charged for in the package and thiscan cross-subsidise less popular repertoire withpopular operas. Each season, of the wholenumber of staged operas, usually 3 are newlyproduced by the house, the remaining are rentedfrom other houses and/or re-used from theMassimo Theatre stock of previous productions.

Supporting Performance Improvement forthe Massimo Theatre Through a DynamicBalanced Scorecard

In order to support planning and performanceimprovement at the Massimo Theatre, a programimplying the development of a DynamicBalanced Scorecard (Bianchi and Montemag-giore, 2008; Kaplan and Norton, 1996; Linardet al., 2002; Ritchie-Dunham, 2001) has beenstarted by the authors.The analysis is being conducted with the

primary purpose to:

1. Define performance indicators associated to end-results, with regard to the four dimensions of

Copyright 2010 JohnWiley & Sons,Ltd.Theatre does not have enough facilities to beused for orchestra end stage rehearsals. Since youneed to use the venue (the theatre itself) also forthe rehearsals, this limits the number of days onecan uses it for performances. This is especiallytrue when production is organized according tothe Italian system of stagione (the season): acertain number of performances over severalweeks, followed by a rehearsal period for the nextproduction after which that opera is performedfor the next several weeks, and so on. The typicalstagione at the Massimo Theatre is usually madeby 1011 productions, 89 of which are operasand 2 ballets. Opera productions are divided inTperf412structural and organizational constraints.he first constraint in enlarging the number oformances is the fact that the Massimoof customers, funding from different institutions, revenues,ber of productions and performances).

Syst. Res.27, 395^420 (2010)Posreporberssible source to calibrate the model will from internal companyts provided by the management information system (e.g. num-pro

uence of the planning process for operaduction, which was described in the previousTseq3. Identify the strategic resources affecting per-formance drivers.

4. Analysing the factors impacting on theaccumulation and depletion flows affectingstrategic resource dynamics, and particularly:policy levers on which managers can operate,external constraints, and decision makersinvolved, in-and-outside the firm.

At present, the study has been only conductedwith the Director of the theatre. However, when afirst version of the dynamic balanced scorecard(DBSC) will be completed, it will be shared withthe other managers, with the goal to use it as aplanning and performance evaluation/improve-ment tool, on a current basis.Main feedback loops have been identified, in

order to depict the processes affecting thetheatres performance. Modelling has beeninitially conducted on a qualitative basis throughperiodical meetings with the Director. Qualita-tive modelling has been also supported by thegradual design of the DBSC structure, andtherefore by the SD simulation model portrayingthe BSC.In particular, the following issues have been

focused from the start of the project till now: (1)decision levers, that have been included in theDBSC control panel; (2) performance indicators (i.e.objectives, measures, and targets) related to theabove four dimensions, which have beenincluded in the scorecard window of the SDmodel; (3) reference behaviour modes for keyvariables. Such behaviours will be reproducedinto DBSC graph sections related to strategicresources, performance drivers and end-results,when the SD model will be calibrated.17

he simulator depicts the same logicalDOI:10.1002/sres


section. Therefore, a three-dimensional analysis

vathesimpa

and so they are enabled to review and discussthem several times. This contributes to their ownlearning processes and to the quality of com-munication. Furthermore, since simulationresults are framed by the DBSC model to showthe dynamics of performance, this learningprocess has also a significant impact on the

Syst. Res. RESEARCHPAPERstrategies concerning the above three issues,

Copyright 2010 JohnWiley & Sons,Ltd.areFonerating.Therefore, by using the simulator, managersable to adjust and better calibrate theirbasegee span the possible effects (i.e. the keyriables behaviour) of adopted decisions incoming years. This also allows managers toulate according a rolling plan mode howst decisions might require future adjustments,d on the outcomes they are currentlyneetimis done, that is:

1. The opera mix by type (blockbuster vs. repertoryvs. contemporary/rare) is planned, not only inpercentage but also in absolute terms.

2. For each type of opera, titles by source areplanned. For instance, it is planned howmanyof the blockbuster operas will be made up bynew, rented, and re-used titles; and thesame is for repertory and contemporary/rareoperas.

3. The number of performances by opera type(blockbuster vs. repertory vs. contempor-ary/rare) is then planned.

The planning time horizon is 3 years. There-fore, each year the number of titles andperformances to stage for the coming and thenext 2 years is planned. Planning in advance ofthe next season is crucial to properly managerehearsals and the acquisition of resources;among them, human resources are the mostcrucial. In fact, particularly concerning blockbus-ter performances, it is critical to contract bestperformers (so called superstars) quite in advance,in order to pursue a well balanced quality/costratio in the delivered service. Therefore, themodel also supports a proper assessment of thefinancial sustainability of the planned seasons.This is not a trivial issue, if one considers theabove said financial criticalities (previouslydefined as cost disease) that in general, andmore specifically in Italy (particularly in theSouth) affect the financial equilibrium of a publicopera house.Though the planning time horizon is 3 years,

the simulation time is 6 years. This is due to thed to allow the user to figure out over a longerster Organizational Change,Accountabilityand Pequality of planning and control systems.Figure 11 shows three main stock-and-flow

diagrams embodied in the model, concerning thethree above issues which are relevant for operaplanning (i.e. opera mix by type; titles by source;performances by opera type). As it is possible tonotice from the model section here provided,18

the first two structures are depicted as agingchains: the final flow Blockbuster titles staged inthe current year depends on the ability of thefirm to implement its 3 years plan (Figure 11a).Lack of resources will have a final impact on sucha flow, both in volume and in quality ofperformances. Such quality will also be affectedby the opera mix by type. On this regard,Figure 11c shows how the same flow determinesthe number of performances by type.19

The three sequential stocks (and correspond-ing outflows) relate to the planned season for theyears (from left to right) n 2, n 1, n, where thelast year is the current (i.e. the coming year).The stock-and-flow structure depicted in

Figure 11a determines the number of titles bysource (i.e. new, rented, reused), depicted inFigure 11b.20 Here, the flow New titles to year 2in turn affects (together with the planned rentedand re-used titles) the flow of blockbuster titlesplanned to year 2.Therefore, the three subsystems are strictly

related each other, and depend on the decisionsmade by managers through the simulator controlpanel. As it is possible to see from Figure 12, suchpanel depicts the same aging chain structurepreviously described. Therefore, the planner isasked to define the number of titles per type andrelated mix (new, rented and reused). He/she is

18For reasons of synthesis, we have shown here only the Blockbustertitles stock.19For reasons of synthesis, we have shown here only the Blockbusterperformances stock.20Also in this case, for reasons of synthesis, we have shown here onlythe New titles stock.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


Input Blockbuster

Titles

Number of launched

NumbRente

T

lockbulanned

y

Reused Titles tocurrent year

(a)

RESEARCHPAPER Syst. Res.Blockbuster Titlesplanned year 1

Titles planned year1

Total BlockbusterTitles planned year 1

Total BlockbusterTitles planned year 2Blockbuster Titles

planned to year 2

New Titles to year 2

Rented Titles toyear 2

Reused Titles toyear 2

New Rented Titles

New Reused

New Titles

New BlockbusterTitles per year

Number of launchedReused Blockbuster

Titles per year

Bp

RejectedBlockbuster Titles

year 1

RejectedBlockbuster Titles

year 2also asked to define the number of performancesper each type of title in the next 2 years, and (ifapplicable) the change of the number of per-formances per title planned for the current year.Other subsystems of the simulator portray the

dynamics of the company human resources, thecustomer base, and financial resources. As said,such resources feedback into the opera pro-

New Titles plannedyear 1New Titles New Titles

current yea

Number of launchedNew Titles per year

Rejected New Titlesyear 1

New Titles plannedyear 2New Titles to year 2

Rejected New Titlesyear 2

Number of launchedNew BlockbusterTitles per year

Number of launchedNew Contemporary

Titles per year

Number of launchedNew RepertoryTitles per year

#Blockbuster Titles

staged in thecurrent year

Number ofPerformances perblockbuster title

current year

BlockbusterPerformancesstaged in thecurrent year

Total BlockbusPerformances sta

in the current y

(b)

(c)

Figure 11 (a) Planning opera mix by type. (b) Planning


414er of launchedd Blockbuster

itles per year

ster Titles to currentear

Total BlockbusterTitles available in the

current year

New Titles tocurrent year

Rented Titles tocurrent year

Time to stage Titles

Blockbuster Titlesstaged in thecurrent year

Total BlockbusterTitles staged in the

current year Blockbuster Titlesstaged in currentyear outflowduction and performance subsystems, since theymay either support or hinder the implementationof plans for opera staging.The DBSC chart embodied in the model

portrays main indicators related to the fourBSC perspectives. A section, concerning thecustomer and internal processes perspectives,is depicted in Figure 13a and b. Both figures

tor

New Titles availablein the current yearNew Titles staged

in the year

Time to stage Titles

New Titles staged inthe current year New Titles staged in

the current yearoutflow

terged

ear Total BlockbusterPerformancesstaged in the

current year outflow

titles by source. (c) Planning performances by type

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


ula

Syst. Res. RESEARCHPAPERembody according to the BSC framework main objectives, performance measures, relatedtargets as well as the current level achieved foreach of these targets. While performance targetsare input variables, and therefore the corre-

Figure 12 The simsponding value can be changed by the planner atthe beginning of the simulation, the currentsituation values are results of the model, and forthis reason have been emphasised in a box with adifferent colour from targets, since they can notbe changed. Comparing targets with currentresults at each simulation stage and for each set ofstrategic objectives supports decision makers tomake judgements on the way the systemresponds to the adopted policies. Therefore, itsupports to calibrate better decisions through thecontrol panel. This process also supports adeeper analysis focused on the feedback struc-ture of the relevant system underlying observedbehaviour from adopted policies. Such analysis,properly supported by a facilitator, is likely toenhance a double loop learning process (Ster-man, 1994).In another section, the simulator separately

portrays, through proper time graphs, thebehaviour of the strategic resources affecting


Foster Organizational Change,Accountabilityand Pethe above performance indicators. It also showsin separate sections how the above indicators canbe distinguished into performance drivers andend-results. Also these two other sections consistof time graphs.

tor control panelFigure 14 offers a synthetic representation ofhow planning opera production may signifi-cantly affect the future growth of the theatre.In particular, the positive loop R1 shows how

an increase in available resources may allow themanagement to plan for more titles and per-formances, leading (other conditions beingequal) to higher revenues, liquidity and otheravailable resources to foster future growth.If the loop R1 is the growth engine of the

theatre, the negative loop B1 is a potential limitto such growth. In fact, the more opera titles (andperformances) are planned, produced, andstaged, the more resources will be consumed.If we particularly consider that opera productionplanning is characterised by a 3-year timehorizon, this means that such phenomenon willreduce (other conditions being equal) theresource endowment to fund the next year plan.The loops B2 and B3 refer to the achievement

of targets related to planned operas and

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


RESEARCHPAPER Syst. Res.performances. In fact, the higher the gap from thecurrent to the target number of operas, the higherthe number the opera titles to plan will be. Thiswill, in turn, decrease the gap in opera titles. Thesame for the planning of performances.The R2 loop shows the effect of proper opera

planning on image and liquidity. A highernumber of planned and staged operas (andperformances) will increase image. This will helpthe theatre to attract financial resources fromprivate funders. This will increase liquidity and

Figure 13 (a) Indicators related to the customer perspective.


416the available resources to foster sustainablegrowth.The loop R1 portrayed in Figure 15 is similar

to the same previously commented aboutFigure 14. In this case, the effect of perceivedquality as a main driver of customer loyalty isremarked. In fact, a higher production of operatitles will lead (other conditions being equal) to ahigher perceived quality by customers. This willturn into a higher image, customer base, andliquidity. More financial resources will allow the

(b) Indicators related to the internal processes perspective

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


Syst. Res. RESEARCHPAPERcompany further growth. This will be possible byboth increasing the number of opera titles/performances and increasing the number ofcontracted superstars that will perform blockbus-ter and repertory operas. This last option will alsoallow the firm to strengthen growth (reinforcingloop R2).There is a number of balancing loops to take

into account, also on this regard. In fact, anincreasing number of operas planned in a seasonwill raise costs. This other conditions beingequal will reduce liquidity to plan for the next

Figure 13 (C


Foster Organizational Change,Accountabilityand Peseasons (loop B1). Likewise, an increase incontracted superstars will also raise operatingcosts and determine a second potential limit togrowth (loop B2).

Concluding Remarks on the MassimoTheatre DBSC Case

The above case has shown how SD can supportmanagement to keep under control the keyperformance variables driving to the success

ontinued)

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


Figure 14 Effects of opera production on

RESEARCHPAPER Syst. Res.(or failure) of an Opera house, with specificreference to the Massimo theatre case. Moreparticularly, the development of a DBSCis likely to focus management attention onthe trade-offs between performance measuresrelated to financial, internal processes,customer and learning and growth perspect-ives. It also helps managers to properly perceivetrade-offs between short and long term effects ofFigure 15 Effects of opera production on perceived quality,customers and financial results


418adopted policies, as well as different companysub-systems. Therefore, likewise any otherDBSC, such tool is likely to foster a learningoriented-approach to strategic planning andperformance management.However, more particularly in this case, this

tool supports managers to perceive and framespecific complexity factors of the public sector,namely: public funding issues (in terms of

financial results and company imagepossible sources, volumes and delays), pricepolicies and stakeholders strategies (municipal-ity, banks, patrons).Next steps in the development of the

DBSC will be gathering actual data to feedthe model, conducting simulation runs andcalibrating the model on key variables refer-ence behaviour modes, and involving in usingthe model manages from different companysectors.

CONCLUDING REMARKS

The three case-studies analysed in this paper,though they might seem so different each other,present a number of common factors, whichsupport the thesis outlined in the introductorysection, that is: applying SD to foster organiz-

Syst. Res.27, 395^420 (2010)DOI:10.1002/sres


ational change, accountability and to supporteffective decision making in the public sector

institutional and political systems determine, has

Bianchi C, Montemaggiore G. 2008. Buildingdynamic balanced scorecards to enhance strategy

Syst. Res. RESEARCHPAPERbeen the focus of the Massimo Theatre case, butcan be to some extent recognised in the othertwo cases.Further research will be needed to validate the

research hypotheses and results here outlined.On this concern, the analysis and comparison ofmore cases and the development of SD simu-lation models will be necessary.However, we believe that the findings originat-

ing from the three case-study analysis canprovide the reader and different involved actorsin the public sector, useful insights on thepowerful role SD can play to foster a betterunderstanding of the causes of poor perform-ance, and to critically reconsider and refocusmental models.

REFERENCES

Baietti A, KingdomW, Van GinnekenM. 2006. Charac-teristics of well performing Public Water Utilities,World Bank Group, Water Supply & SanitationWorking Notes, Note No. 9, May 2006.

Baumol WJ, Bowen WG. 1966. Performing ArtsThe Economic Dilemma. Twentieth Century Fund:Cambridge, MA.

Bianchi C. 2010. Improving performance and fosteringaccountability in the Public Sector through SystemDynamics modelling. From external to an internalperspective. In Public Sector Applications of theSystem Dynamics Approach, Systems Research andBehavioral Science 27(4): 361384.

Copyright 2010 JohnWiley & Sons,Ltd.implies a number of challenges which arecontingent, not only because of the specificcomplexity of the public sector itself, but alsoas a consequence of specific issues that can befound on a territorial base.In particular, the role that SD modelling can

play to properly deal with the relationshipbetween the political and the organizationalsystem has been discussed both in the publicutility and the health care case-studies.The role of SD to support the management of a

public sector organization to promptly recoverefficiency and improve outcomes at the sametime, based on the current constraints that theFoster Organizational Change,Accountabilityand Pedesign and planning in public utilities: key-findingsfrom a project in a city water company. SystemDynamics Review 24(2): 175213.

Borgonovi E. 1996. Principi e sistemi aziendali per leAmministrazioni Pubbliche (Management Principlesand systems for Public Administrations). Egea: Milano.

Cantu` E, Carbone C, Anessi-Pessina E. 2008. Do ItalianRegions Effectively use DRG Funding to Steer Pro-vider Behaviour? Paper presented at the 4th TADTransatlantic Dialogue Conference, Milan (http://www.4tad.org/ws/paper_wks6_anessi-cantu-carbone.pdf).

Christensen T, Laegreid P, Stigen I. 2004. PerformanceManagement and Public Sector Reform: the Norwe-gianHospital Reform. Stein RokkanCentre for SocialStudies, Working Paper n. 17; 22 (http://www.ub.uib.no/elpub/rokkan/N/N08-05.pdf).

Draghic T. 2006. Il Progetto dIntroduzione del SistemaAP-DRG nella Clinica Luganese S.A., MasterThesis Work (http://www.chuv.ch/bdfm/cdsp/59562.pdf).

Faccio M, 2005. Politically connected firms. The Amer-ican Economic Review 96(1): 369386

Frey BS, Pommerehne WW. 1989. Muses & Markets.Explorations in the Economics of the Arts. London:Blackwell.

Grampp WD. 1989. Rent-seeking in arts policy. PublicChoice 60(2): 113121.

Gutierrez UM. Menozzi A. 2008. Board compositionand performance in state-owned enterprises: evi-dence from the Italian public utilities sector, paperpresented at the European Financial ManagementAssociation Annual Meeting, June 25- 28, Athens,Greece.

Heilbrun J, Gray CM. 2001. The Economics of Art andCulture. Cambridge University Press: Cambridge,UK.

Hsiao WC, Sapolsky HM, Dunn DL, Weiner SL. 1986.Lessons of the New Jersey DRG payment system.Health Affairs 5(2): 3245. (Summer 1986)

Kamke K. 1998. The German health care system andhealth care reform. Health Policy 43: 171194.

Kaplan R, Norton D. 1996. The balanced scorecard: trans-lating strategy into action. Harvard Business SchoolPress: Boston.

Lane D, Huseman E. 2008. System dynamics mappingof acute patient flows. Journal of the OperationalResearch Society 59: 213224.

Linard K, Fleming C, Dvorsky L. 2002. SystemDynamics as the Link between Corporate Visionand Key Performance Indicators. Presented at the20th System Dynamics International Conference,Palermo, July.

Maliapen M, Dangerfield BC. 2009. A systemdynamics-based simulation study for managingclinical governance and pathways in a hospital.

Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


Journal of the Operational Research Society 61(2): 255264. (February 2010)

Mikkola H, Keskimaki I, Hakkinen U. 2001. DRG-related prices applied in a public health care sys-temCan Finland learn fromNorway and Sweden?Health Policy 59: 3751.

Ritchie-Dunham JL. 2001. Informing Mental Modelsfor Strategic Decision Making with ERPs and theBalanced Scorecard: a Simulation-Based Exper-iment. In Proceedings of the 19th System DynamicsInternational Conference, Atlanta.

Shapiro C, Willig R. 1990. Economic rationales for thescope of privatization. InThe Political Economy of PublicSector Reform and Privatization. Suleiman EN,Waterbury J. (eds.). Westview Press: London; 5587.

Sheigold S. 1986. Unintended Results of MedicaresNational Prospective Payment Rates. In HealthAffairs, Winter; 521.

Shleifer A, Vishny RW. 1994. Politicians and firms.Quarterly Journal of Economics CIX: 9951025.

Sterman J. 1994. Learning in and about complex sys-tems. System Dynamics Review 10(23): 291330.

Throsby D. 2001. Economics and Culture. CambridgeUniversity Press: Cambridge, UK.

Trailer JW, McLaughlin MB. 2007. The problem ofdelayed discharge in labor and delivery, Proceed-ings of the 25th International Conference of theSystem Dynamics Society, Boston, July 29August2, 2007.

Vennix JAM. 1996. Group Model-Building: FacilitatingTeam Learning Using System Dynamics. John Wileyand Sons: Chichester.

Vennix JAM, Gubbels J. 1994. Knowledge elicitation inconceptual model Building: a case-study inmodelinga regional Dutch health care system. In Modeling forLearning Organization, Morecroft J, Sternam J. (eds.).Productivity Press: Portland, Oregon; 121145.

Wolstenholme E. 1999. A patient flow perspective ofU.K. health services: exploring the case for newintermediate care initiatives. System DynamicsReview 15(3): 253271.

World Bank. 2004. Reforming Infrastructure: Privatiza-tion, Regulation, and Competition. Oxford UniversityPress: Washington, DC, World Bank: New York.

RESEARCHPAPER Syst. Res.Copyright 2010 JohnWiley & Sons,Ltd.420Syst. Res.27, 395^420 (2010)

DOI:10.1002/sres


Documents

2010 Bianchi System dynamics