Management control, culture and ethnicity

during childhood and then examines how their interaction with the Javanese culture of pribumi employees, ethnic ten-sions between employers and employees, and organisational and economic factors aected management control. Con-

Introduction

There has been considerable interest in whethernational cultures produce dierent control systems(see Bhimani, 1999; Harrison & McKinnon, 1999).

eserved.

* Corresponding author. Address: Division of Accountingand Finance, Manchester Business School, University ofManchester, Crawford House, Booth Street East, ManchesterM13 9PL, UK. Tel.: +44 161 275 4014; fax: +44 161 275 4023.

E-mail addresses: s_eerin@ubaya.ac.id (S. Eerin), Tre-vor.Hopper@mbs.ac.uk (T. Hopper).

Accounting, Organizations and Soci0361-3682/$ - see front matter 2006 Elsevier Ltd. All rights rsistent with previous cultural contingency research the Chinese owners preferences resided with controlling behaviourthrough personnel and behavioural controls, low budget participation, centralisation, subjective rather than objectivecontrols, and tentatively, few rewards tied to results and the use of group rewards. Whether Chinese managers exhibitedlonger term orientations concerning planning and rewards could not be ascertained. However, ethnic tensions and com-mercial considerations mitigated the owners ability to control according to cultural preferences. Based upon these nd-ings reections on past research and suggestions for further developments are made with respect to methods,methodology, and incorporating a broader range of theories and issues, especially ethnicity, politics, and history. 2006 Elsevier Ltd. All rights reserved.in a Chinese Indonesian company

Sujoko Eerin a, Trevor Hopper b,c,*

a Universitas Surabaya, Surabaya, Indonesiab Manchester Business School, University of Manchester, Manchester M13 9PL, UK

c Stockholm School of Economics, Sweden and Victoria University, Wellington, New Zealand

Abstract

This study explores socio-cultural aspects of management control in a Chinese Indonesian manufacturing company.Ethnographic data collection methods were combined with grounded theory data analysis to explore how cultures, eth-nic dierences, history, politics, and commercial considerations shaped management controls. A combination of emicand etic methods were used to generate grounded comparisons with nomethetic research on culture and control in acultural contingency tradition.

Chinese Indonesians own most Indonesian private domestic capital despite being an ethnic minority (34% of pop-ulation) and having suered extensive discrimination. The case links the Chinese businessmens values to socialisationdoi:10.1016/j.aos.2006.03.009www.elsevier.com/locate/aos

ety 32 (2007) 223262

However, the research is beset with theoretical andmethodological controversy (Baskerville, 2003;Baskerville-Morley, 2005; Hofstede, 2003) andinconsistent and problematical results (Harrison& McKinnon, 1999). Bhimani (1999) argues thetwo main conceptual approaches used to date,structural contingency and culture based idea-

224 S. Eerin, T. Hopper / Accounting, Organiztional theories, need supplementing with theoriesof societal eects, new institutional sociology,and new accounting history, despite their dier-ent assumptions about what homogenises controlsystems, the nature of controls, appropriateresearch methodologies, and their focus of analy-sis. This paper illustrates Bhimanis claim that the-ories can be complementary and progress lies intheoretical triangulation rather than continuoussniping over the barricades of the accountingmethodology science wars.1

There has been particular interest in how Chi-nese culture impinges on management control.2

Our original primary research interest was similar it lay in exploring whether the cultural beliefs ofChinese owners of an Indonesian manufacturingcompany were consistent with Confucianism thefoundation of Chinese culture and how, if at all,these impinged on the rms management controlsystem (MCS). However, the Chinese owner-man-agers operated in a multi-cultural milieu for theiremployees were predominately pribumi (mainlyJavanese see glossary for denition of Chineseand Indonesian terms). Thus, if Chinese culture isimportant for MCSs then Javanese culture andinter-cultural interactions should be too.

Multi-culturalism is often associated with eth-nic dierentiation. Ethnicity is a source of groupidentity: it not only attributes characteristics(whether founded or imaginary) to members focalgroup but also to other ethnic groups. Ethnicitydenes the self in relation to others and can be a

1 An adaption of a point made by David Cooper in a plenaryat the Global Management Accounting Research Conference,Michigan State University, 2004.2 Studies of Chinese culture and MCSs have often been on

non-mainland Chinese operating in multi-cultural environ-ments. For ease of argument we equate the minority IndonesianChinese culture with that of mainland Chinese, i.e. a national

culture.source of action and meaning. This is so in Indone-sia, which has many ethnic groups. However, inthe political arena the primary ethnic distinctionhas lain between pribumi and Chinese. Politiciansaccusations that Indonesian Chinese contaminateor resist Indonesian national culture led to politi-cal suppression of Chinese culture and inuence.Chinese Indonesian business practices may drawfrom Confucianism but they must also co-opt sup-port in hostile local environments marked byresentment of Chinese businessmen, mutual ethnicsuspicions, and a history of state discrimination(Ong, 1999; Redding, 1993; Redding & Whitley,1990; Yeung, 1999). As Bhimani (1999), Wickra-masinghe and Hopper (2005), Wickramasinghe,Hopper, and Rathnasiri (2004) argue, addressingmulti-culturalism required addressing history, pol-itics, and ethnic conict.

Cultural contingency research based on surveys,often using cultural constructs from Hofstede(1980), has neglected issues of ethnicity andmulti-culturalism. For example, in multi-culturalsocieties, who is Chinese can be problematic, beingan issue of subjective denition rather than ances-try. Cultural contingency research gives littleexplanation of what Chinese values are, why andhow they emerged, why they matter, and how theyinuence controls. Hence it is unsurprising thatit has proven inconclusive (Baskerville, 2003;Chenhall, 2003; Harrison & McKinnon, 1999).Moreover, culture is neither totalising nor deter-ministic: managers have choices and can act con-trary to cultural beliefs. They must considereconomic considerations, competition, and factorssuch as organisational size and technology whenexercising control. As the researchers realised inthe course of eldwork, cultural beliefs may bemarginal, consequently the central research ques-tion broadened to:

How does the socio-cultural environment ofChinese Indonesian businessmen inuencethe design and operation of their companysMCS. Is it a consequence of Confucian val-ues, Javanese values, ethnic dierences, statethreats, or best business practice?

The authors believed ethnographic eldwork could

ations and Society 32 (2007) 223262inform this by building theory bottom up, avoid-

ganizing unduly predetermined categories and causalchains, being longitudinal, and embracing abroader set of issues.

However, the researchers had a dilemma.Understanding the social construction and mean-ing of an MCS requires emic analysis (whichdescribes indigenous values of a particular societyusing semiotic ethnography based on thickdescription from eldwork) whereas mostaccounting research on Chinese culture and con-trol is etic (it applies broader predetermined theo-retical models across several societies). An eticview is how outsiders see and interpret a phenom-enon, whereas an emic view is how insiders/partic-ipants interpret the phenomenon as part of theirworld (Marshall, 1998).

The researchers inclinations to emic researchwere tempered by a desire to engage with prior eticresearch, build knowledge cumulatively, and usetheories in a complementary, pluralistic manner(Bhimani, 1999). Hence grounded data was ana-lysed using etic categories from prior researchwhilst using emic analysis to create new categoriesand concepts when etic categories did not suce.Whether to concentrate on building rich descrip-tions of social groups or make cross-cultural com-parisons lies at the heart of methodologicaldebates within ethnography between semioticand behavioural approaches. The emic and eticdistinction was developed by the linguist KennethPike and adapted by Goodenough to developtypologies for cross-cultural comparison derivedfrom eld data (Sanday, 1979) to reconcile semi-otic and behavioural approaches. The researchersdo not claim combined emic and etic methodsare superior, or that emic analysis precludes com-parative analysis (though it is beyond many emicresearchers remit, and its creation of overlappingbut dierent categories hinders comparison),rather they claim that these methods were eectivefor their research aims. As Sanday (1979, p. 34)remarks, The main dierences are whether theprimary focus is on the whole, the meaning, orthe behaviour and the degree to which the analyticgoal is diagnosis or explanation. Which mode oneadopts is a matter of taste not dogma. . . . Whatcounts in the long run is not how the facts are con-

S. Eerin, T. Hopper / Accounting, Orstructed but whether they make sense.Management control systems and cultural

contingencies

Features of a MCS

The researchers denition of an MCS wasbroad namely a system within social, cultural,political, and economic environments used bymanagement to align employee behaviour withorganisational objectives and to manage internalinterdependencies (such as managementworkerand inter-departmental relations), and externalrelationships (with the state, society, customers,and suppliers) (Euske & Riccaboni, 1999).

Merchants model of MCSs (1998) was used foretic classication as it embraces a wide range offormal and social controls, is operationally welldened, has informed or is consistent with MCSmodels in previous research, and is not connedto large organisations (Davila, 2005). Manage-ment control is notoriously dicult to dene. Mer-chants model was used not because it is the bestbut because it provided eective etic categories toThe paper initially outlines the model of MCSand features of Chinese culture that provided eticcategories of analysis. It outlines propositions ofanticipated relationships between culture andMCSs derived from previous research for furtherexamination. The paper then discusses factorsnot captured by the etic analysis of eld data sub-sequently analysed in an emic fashion. These werethe Javanese culture of the rms workers and thelocal context especially the history of ethnic dif-ferentiation and discrimination against ChineseIndonesians. Then the research methods areexplained and justied in greater detail. The empir-ics initially describe the companys key actors,history, structure, membership, technology, andmarkets (also signicant for understanding theMCS). It then analyses the Chinese owners values,their sources, and how these factors especiallycultures, perceived best business practice, and eth-nic tensions; shaped result, action, and culturalcontrols. The conclusions summarise the ndingsand their implications for future research.

ations and Society 32 (2007) 223262 225analyse emic ndings, which enabled results to be

compared with previous research and be linked tointernal and external factors that emerged as sig-nicant.3 Merchant distinguishes three types of

226 S. Eerin, T. Hopper / Accounting, Organizcontrol: result, action, and personnel. They arenot mutually exclusive; rather MCSs dier accord-ing to emphases on each.

Result control denes outputs expected fromemployees. Their achievement is often reinforcedby rewards. It is most useful if what constituteseective performance is known, employees caninuence results, outputs are measurable, andemployees perceive managers authority as legiti-mate (Merchant, 1998). Action control monitorsmeans (behaviour) rather than ends (results) byprohibiting undesirable acts (behavioural con-straints), deriving desired employee behaviourfrom plans (pre-action review), and monitoringbehaviour by direct observation or formal controls(action accountability). Personnel/cultural controls(hereafter called cultural control) focus uponrecruitment, training, job design, and promotingshared norms and values to induce employee self-control (Merchant, 1998, p. 121).

Dimensions of Chinese culture

The research also needed to determine etic cate-gories of Chinese culture. The researchers deni-tion of culture is, a shared way of life thatincludes values, beliefs, and norms transmittedwithin a particular society from generation to gen-eration (Scupin, 1998, p. 36). It provides . . . thebases for choice, by a social group, of particularends and of particular means by which these endsare to be accomplished (Lachman, Nedd, & Hin-ings, 1994, p. 41), i.e. it denes what is good or bad,holy or unholy, beautiful or ugly. Individuals judgeorganisations and respond according to stable corecultural values and less enduring peripheral valueswhose breach brings less severe sanctions.

Confucianism, the foundation of Chinese cul-tural values (Suryadinata, 1978), espouses morallaws, tao, that emphasise social order, harmony

3 The authors do not deny the value of more emic inclinedresearch on national cultures and control such as Ahrens(1997), Ansari and Bell (1991), and Euske and Riccaboni

(1999). Indeed they commend it.(Redding, 1993), and social relations based onjen that embraces chung and shu (Suryadinata,1978). Jen is like the Christian doctrine of lovingothers as oneself but extends to justice and equal-ity similar to secular concepts of law and order.In Confucianism humanity depends upon commu-nity status derived from human reciprocity. Chunginvolves sincerity and honesty whereas shu empha-sises altruism. Both emphasise doing to others asyou would wish to be done to yourself.

Confucian values of vertical and horizontalsocial order inuence Chinese business culture(Suryadinata, 1978). Vertical order covers hierar-chical social relationships, such as husband andwife, parents and children, masters and servants,rulers and subjects. Each social position hasascribed responsibilities and duties known as li,the most important being lial piety (hsiao). Par-ents are expected to educate their children anddirect them towards a correct life, whereas childrenshould respect parents, care for them, and protecttheir dignity. Chinese society is marked by strongfamily collectivism (Redding, 1993): it assumespeople who neglect family responsibilities willnot honour obligations to outsiders. The need tofull individual responsibility (li) underlies allother values. Hsiao extends to societal responsibil-ities, notably respecting, obeying, and demonstrat-ing loyalty to superiors, being wise rulers, andcaring for subordinates. Failure to do this willembarrass not just the perpetrators but also theirparents.

Self-identity and social order in Chinese cultureis embedded in horizontal social relationships(Redding, 1993). Individual face and reputationstems from social group membership. Chinesesociety consists of concentric circles. The core isthe family, surrounded by a lineage group orextended family. Family resources should be pro-tected and enhanced to foster self-suciency.Relations between each circle are cemented bymutual trust (Fei, 1939; quoted in Redding,1993, p. 58). Friends and acquaintances are trustedaccording to established mutual dependencies inguanxi ethnic based networks that, inter alia,protect family resources. Everyone has a sense ofwhom to trust and co-operate with, for what pur-

ations and Society 32 (2007) 223262poses. Outside these circles polite but guarded rela-

and employees, especially if both are Chinese,making it relatively easy to establish disciplineand stable hierarchies. Patrimony may fosterorganisational adaptability, goal congruence, com-

ganizations and Society 32 (2007) 223262 227tions prevail. Socialisation by parents and schools(Redding, 1993) and Chinese Indonesians harshhistorical experiences have reinforced and shapedtheir values over generations (Yeung, 1999), pro-moting guanxi networks as a defensive form of sol-idarity, co-operation, and self-reliance withinfamilies.

Personal trust derives from meeting personalobligations, reputation and face. Neutral, imper-sonal relationships are dicult as they make deter-mining trustworthiness dicult. Good socialrelations and mutual trust are valued who youknow is as important as what you know (Redd-ing, 1993). Hierarchical decisions such as ring orhiring, rewards, discipline, and evaluating employ-ees tend to be based on friendship, trust, emotionalties and loyalty rather than results and eciency.Cultural controls are legitimate in Chinese organ-isations, are easily established, and tend to super-sede bureaucratic controls (Redding, 1993).

Major values derived from li and hsiao are qinjian, ke ji, qin fen, and jing shen (Cleary, 1992;Xu, 2000). Qin jian (frugality) emphasises saving,conserving resources, and displaying wealth mod-estly. Ke ji (asceticism) stresses controlling desiresfor wealth, power and pleasures and fulllingsocial responsibilities. Qin fen advocates diligenceand seeking knowledge and wisdom. Jing shenembraces prudence and foresight. According toConfucius (Cleary, 1992, p. 25), people who donot think far enough ahead inevitable have worriesnear at hand.

Confucian values have produced a patrimonialChinese business style (Redding, 1993, p. 155).Power stems from ownership: owners and employ-ees view companies as family property andmanagement as akin to regulating a family.Owner-managers instil values of personal trust,money consciousness and prudence on the familysbehalf. Critical positions tend to be assigned tofamily or trusted members of lineage groups.Leadership is autocratic but paternalistic. Employ-ers are responsible for employee welfare, allocatingjobs, stewardship of resources, helping the ine-cient, providing security for the old and showingunderstanding. In return employees should showunquestioning obedience and diligence. The values

S. Eerin, T. Hopper / Accounting, Orof li and hsiao are powerful norms for employerspliant and diligent subordinates, and stableorganisational membership. On the other hand itcan produce nepotism, organisational cliques,non-family members with insucient authority,information secrecy, subjective performance evalu-ation, restricted promotion opportunities, riskavoiding employees, and constraints on organisa-tional growth.

If delineations of eective MCS practices derivefrom the West, and Chinese culture diers fromWestern culture, then MCS practices may dierin Chinese organisations (OConnor, 1995; Tsui,2001). Merchant, Chow, and Wu (1995) conjecturehow this may occur using ve dimensions ofnational cultures delineated by Hofstede (1980)and Hofstede and Bond (1984): namely collectiv-ism, masculinity, power distance, uncertaintyavoidance, and Confucian dynamism.4 Ourresearch adapted Merchant et al.s (1995) reviewto generate seven propositions that guide ourempirical analysis and provide linkages to previousresearch. They are that Chinese managers will use:

1. Personnel and action controls rather than resultcontrols (Harrison, 1993; Merchant, 1998;Ouchi, 1979, 1980).

2. Little participation due to patrimony (Lau &Tan, 1998). Controls will be centralised due tolower individualism (Birnbaum & Wong, 1985;Harrison, McKinnon, Panchapakesan, &Leung, 1994; Lau, Low, & Eggleton, 1995;OConnor, 1995).

3. Subjective rather than objective controls due toan emphasis on trust and personal relations(Merchant et al., 1995; Redding, 1993).

4. Few rewards tied to results (outputs) to avoidloss of face and risk associated with uncertainty(Merchant et al., 1995).

4 Hofstede and related accounting work, and the relationshipof the ve dimensions of national cultures, especially Chineseculture, are not discussed for reasons of space. Fuller exposi-tions are in Lau and Tan (1998), Merchant et al. (1995), and

Tsui (2001).

ganiz5. Group-based rewards due to values of collectiv-ism and shared risk (Merchant et al., 1995).

6. Few long-term incentives as collectivist valuesdecrease short term gaining (Merchant et al.,1995).

7. Long-run time horizons when planning usingsubjective information garnered from guanxior social networks (Harrison et al., 1994).

There has been empirical support for somepropositions, e.g. participation and centralisationbut other studies have produced surprises. Forexample, Merchant et al.s (1995) study of (Tai-wanese) Chinese and USA managers found con-textual factors such as management educationand experience, beliefs about stock markets, busi-ness growth, type of industry, and labour forcemobility aected management controls more thancultural factors. A laboratory study by Awathsi,Chow, and Wu (1998) found USA subjects usedgroup controls in situations of interdependencyand made more self-sacrices than Chinese sub-jects, contrary to expectations. The failure to con-sistently corroborate links between Chinese cultureand controls have several possible reasons, includ-ing assuming that national cultures are unitarywhereas societies are often multi-cultural and besetby ethnic tensions, as in this study, where ChineseIndonesian businessmen employed pribumi, mostlyJavanese, workers. Thus there needs to be an elab-oration of Javanese culture and its possible conse-quences for control.

Javanese culture

There are no studies of Javanese culture andMCSs to the best of our knowledge hence weturned to anthropological work for insight. Thepillars of Javanese culture are alus-kasar and lair-batin (Geertz, 1960). Spiritual excellence stemsfrom alus and kasar. Alus means pure, rened,polite, exquisite, ethereal, subtle, civilised, andsmooth. God and his mystical experience are alus.Human behaviour and actions are alus if they fol-low appropriate manners and etiquette. Kasarmeans the opposite: impolite, rough, and uncivi-lised. Everyone from peasant to king is ordered

228 S. Eerin, T. Hopper / Accounting, Orbetween these two poles. Lair, the outer realmof human behaviour refers to an individualsexternal actions, motions, postures, and speech,whereas batin the inner realm refers to theiremotional life: fuzzy, shifting private feelings intheir phenomenological immediacy (Geertz, 1960,p. 232). Correct combinations of alus-kasar andlair-batin enable even an uncultivated peasant toattain the level of a hyper-civilised divine king(Geertz, 1960, p. 233). The closer one is to alusand the more batin experiences one has, the moreideal one is. This produces three clusters of cul-tural values: social order involving andap-asorand bapakism, social harmony (rukun), and mysti-cism including rituals such as slametan andkenduri.

Social order is manifest in Javanese etiquette.This has rigid, formal rules of interaction withinsocial hierarchies, including linguistic forms andandap-asor, which means to humble oneselfpolitely and demonstrate correct behaviour.Everyone should know their position and that ofothers: behaving and choosing words correctlyaccording to who is being addressed is a mark ofrespect. Words have a status hierarchy rangingacross alus to kasar, and Javanese pattern speechon this axis according to the addressees statusand the conversations context. Failure to showandap-asor can cause shame (sungkan) for the reci-pient, especially if they cannot behave or reply aswell as the rst person (Geertz, 1960, 1961).Although modern education has brought moreegalitarian attitudes demonstrating andap-asor isstill valued.

Bapakism is a Javanese form of paternalism andpatronage (Geertz, 1961; Rademakers, 1998).Bapak literally means father but can also mean acharismatic gure that cares for community mem-bers. Bapaks demand respect, obedience and loy-alty from subordinates. They can claimprotection, gifts and help when needed but theymust listen to, empathise with and proer advice.Giving compliments is important: their proper andconstant delivery motivates subordinates and com-mands respect.

Rukun is the maintenance of social harmony(Geertz, 1961, p. 149). A society without overtexpression of divisive opinions and feelings is

ations and Society 32 (2007) 223262rukun. This is manifest through collective (musya-

ganizwarah) and unanimous (mufakat) decisions, andco-operation (gotong royong). Reaching harmoni-ous compromise without intense feelings or expres-sion of resentment is ideal. This can promoteevasion, covert disobedience and mutual avoid-ance in social relations: when conicts burst openthey can be traumatic and severe.

Javanese have strong mystical beliefs. Manybelieve ancestors, places, and spirits can communi-cate. Maintaining harmonious relations with spir-its is important for a tranquil life. This requiresfollowing rituals such as slametan/kenduri a com-munal feast on important occasions symbolisingthe mystic and social unity of participants (Geertz,1960). Though it incorporates some Islamicelements, most Javanese regard slametan as dis-tinctively Javanese (traditional spiritualism) andpre-Islamic or even Hindu in inspiration. Thegoals are psychological: the absence of aggressivefeelings, no disturbance, and building unitythrough social compromise (Beatty, 1999).

Prima facie, Javanese and Chinese cultures aresimilar in that both emphasise paternalism, hierar-chy, reputation, social harmony, and social orderthough Javanese culture has a distinctive mysti-cism and set of behavioural expectations. If so,there may be little cultural dissonance if MCSsreciprocate these values.

Problems of cultural contingency

Cultural contingency studies have come undergrowing attack. First, analyses may be too general-ised they presume what is Chinese is relativelyhomogenous within and across nations. They paylittle heed to cultural diversity in societies andinteractions between cultures (Baskerville-Morley,2005). Second, such studies tend to be static theyignore how and why cultures change (Bhimani,1999). Third, they ignore the impact of historicaland external organisational factors involving, interalia, political and economic institutions and strug-gles upon culture and control (Bhimani, 1999).Fourth, they assume who is Chinese is objectivelyascertainable when it may be subjectively dened.Fifth, they fail to establish how cultural valuesare imbued by socialisation in the family, educa-

S. Eerin, T. Hopper / Accounting, Ortion systems, and social experiences within andoutside the workplace (Bhimani, 1999) or demon-strate links between beliefs and action how andwhether culture is enacted within MCSs isneglected (Harrison & McKinnon, 1999): actionmay involve agency or responses to other commer-cial or social pressures. Sixth, there is undue reli-ance on problematical concepts and surveyresearch instruments derived from Hofstede, oftenin tandem with problematical ones from contin-gency theory and research on accounting perfor-mance measurement (Baskerville, 2003; Chenhall,2003; Harrison & McKinnon, 1999; McSweeney,2002) building concepts and factors bottom upthrough eld studies has been neglected.

Bhimani (1999) argues that theories of SocietalEects, New Institutional Sociology, and theNew (especially Foucauldian) accounting canaddress these deciencies despite their dierentassumptions. For example, whereas contingencytheory matches predetermined dimensions of cul-ture and controls, seeks universal results involvingfunctional relationships, assumes purposeful con-trols can be constructed, and uses nomotheticmethods (systematic studies, often statistical, toestablish general relationships) the other theoriesare more inclined to incorporate context, theemergence of systems, and ideographic methods.In contrast, the new history stresses uniqueness,complexity, and serendipity; how external institu-tions bear upon the emergence of controls;rationalities as products of historical power-knowledge regimes; and longitudinal ideographicmethods; new institutionalism focuses on subjec-tivity and contextualised conceptions of social real-ity; and societal eects has rigorous grounding ofcontextual factors expressed as theoretical proposi-tions suitable for deeper comparative analysis. Weleave it to others to demonstrate the potential ofthese alternative theories. Our research does notexplicitly incorporate the formal substance of anyof them. However, we share Bhimanis advocacyof more contextual, emergent, and ideographicresearch, and we believe our research methods thatcombine the emic and the etic, and incorporate thepolitics of ethnicity, multiculturalism, history,business pragmatism, socialisation, and actionover time to study culture and control, reect the

ations and Society 32 (2007) 223262 229spirit of what Bhimani intended.

Ethnicity, Indonesian politics, and Chinese tions such as traders, police, military, and bureau-crats for particular ethnic groups exacerbatedethnic dierentiation. Ethnicity rather than classmobilised competition for resources, a common

230 S. Eerin, T. Hopper / Accounting, Organizations and Society 32 (2007) 223262businesses

From the outset the researchers knew about thehistory of ethnic strife upon Chinese businessesand Indonesian politics. However, the originalfocus was upon Chinese culture and controls,hence the emphasis upon their etic categorisation.The signicance of ethnicity and politics was aminor surprise revealed by eldwork and subse-quent data analysis. For ease of exposition ethnic-ity and Indonesian politics are discussed here butthe reader should be aware that the processes ofdiscovery were, as is oft so in case study research,more iterative and messy than the papers chro-nology may suggest.

Indonesia, despite the similarity of Javaneseand Chinese culture has national integration prob-lems due to religious and ethnic divisions. Approx-imately 85% of Indonesians are Moslems. Theremainder include Christians, Hindus, Buddhists,Khonghuchus,5 and followers of Aliran. Moslemsdisagree whether the state should be secular or fol-low Islamic law. The largest ethnic groups areJavanese (45%) and Sundanese (15%). The other400 ethnic groups complain about Javanisationof politics and culture but the major ethnic dichot-omy is between the Chinese (3% of population)and others (the so-called pribumi) (Brown, 1994).

Ethnicity classies people according to allegedphysical and social identities (Fenton, 1999). Eth-nic group boundaries may be symbolic (language,ancestry, religion, kinship, or culture) and mayhave a material and class base. Ethnicity mayinvolve racial and religious stereotyping and beempirically unfounded (Eriksen, 1993) but it inu-ences social action. Culture is often sub-consciousand taken-for-granted whereas ethnicity more con-sciously denes the self and others. For example,groups use ethnic claims to dierentiate themselvesand stigmatise other groups to mobilise politicalprogrammes.

The state can shape ethnicity (Brown, 1994).Colonial rulers imported labour to Indonesia andadopted divide and rule policies. Reserving occupa-5 A derivation of Taoism, Buddhism and Confucianism.view being, Weve got less because theyve gotmore (Fenton, 1999, p. 47). This continues todayin Indonesia ethnicity rather than class is corre-lated with economic segregation, social ranking,and power. Despite only representing 68 millionof 200 million Indonesians, the Chinese share ofprivate domestic capital far exceeds that of anyother ethnic group (Leong & Lim, 1992; Robison,1986)6 yet they have suered political, social, andeconomic deprivation (Heryanto, 1997, 1998).

The earliest permanent Chinese settlements inIndonesia date from the late 13th century (Fryer& Jackson, 1977; Rickles, 1993). Until the late18th century, most migrants were males who inter-married with locals, adopted local lifestyles, andare now regarded as indigenous pribumi. However,immigration waves after World War I broughtnumerous Chinese women to Indonesia. Chinesechildren with Chinese parents raised in a Chineseculture rapidly increased. They became dierenti-ated as non-pribumi. However, denitions of ethnicChinese in Indonesia do not rest simply on biolog-ical criteria. For example, many considered as Chi-nese are by ancestry less than one-quarter so,whilst others with a greater Chinese lineage areconsidered by themselves and others to be indige-nous pribumi (Skinner, 1963). Who is Chinese isa social construction stemming from how individ-uals function within society and with whom theyidentify.

Chinese Indonesians can be classied into totokand jiaosen. Totoks still practise and educate theirchildren in Confucian values, speak Mandarin orother Chinese dialects, and celebrate Chinese tra-ditional events. Most embrace Buddhism orKhong Hu Chu religions. In contrast, most jiaosenare Christians (some embrace Islam), follow fewChinese traditions in daily life, speak little or no

6 The share of Chinese capital in the Indonesian economy iscontroversial (Amir, 1997; Hadiz, 1997; Kwik, 1997; Ning,1997) and it is dicult to get reliable data. Estimates vary fromless than 50% to 70%. Nevertheless, it is widely accepted that

Chinese capital is the largest source of private domestic capital.

foreign investments, and establishing political sta-bility through authoritarian rule (Robison, 1986).The new political leaders saw Chinese Indonesians

ganizations and Society 32 (2007) 223262 231Chinese dialect, and adopt local customs. Theirculture is a mix of Western, local, and Chinese/Confucian inuences. Younger Chinese areincreasingly becoming jiaosen but many Chineseparents still teach Confucian values to their o-spring, and many jiaosen separate religion and cul-ture, conning religion to personal faith, whereasculture denes proper social conduct.

During Dutch colonial rule (pre-independence)the Chinese were used as economic bridges to otherethnic groups and many prospered. Followingindependence, the Soekarno regime (19451966)incorporated anti-foreign and wealth redistribu-tion sentiments (Chalmers, 1997) in economicprogrammes to encourage indigenous ownershipand ethnicity. The Benteng (Fortress) Programme(19501957) and the Economic Urgency Pro-gramme (1951mid-1950s) promoted pribumieconomic power relative to Dutch and Chinesecapital. The Benteng programme restricted import-ing to registered pribumimerchants. The EconomicUrgency Programme tried to create a strongpribumi business class by funding modernisationof small-scale pribumimanufacturers and restrictedsome markets to them. It failed because brokerswith political connections won licences and soldthem to more skilled capitalists, mostly ethnicChinese. In 1959 a ban on aliens retailing outsidedesignated urban centres undermined the Chineserole as middlemen in national marketing. Manyrural Chinese were forcibly transferred into largetowns. Refusal to obey evacuation orders enragedarmy personnel who used harsh physical measuresto enforce compliance (Skinner, 1963). This regula-tion only applied to Chinese without Indonesiancitizenship but it made all Chinese insecure formany families were a mix of Indonesian andChinese citizens. Following regulatory pressuresand attacks on their properties many Chinese leftIndonesia.

Soekarnos nationalistic economic policy pro-hibited international capital inows and its con-frontation with ethnic Chinese capital resulted ininadequate investment (Hill, 1996; Hill & Mackie,1994). The New Order State (19661998) assumedpower when the economy collapsed. Its focus andlegitimacy rested on economic development,

S. Eerin, T. Hopper / Accounting, Orthrough state, private (mainly ethnic Chinese) andas engines of economic development and lessthreatening to their power than pribumi (Robison,1986). Nevertheless the Chinese suered discrimi-nation, resentment, scapegoating, harassment,and extortion. They joined the West, Communism,and Fundamentalist Islam as the fourth majorOther inNewOrder political discourse (Heryanto,1998). The suppression of Chinese culture was jus-tied rhetorically as being alien to Indonesiannational culture. Public Chinese festivals werebanned, Chinese schools closed, entry quotasapplied to public universities and schools, and Chi-nese symbols removed (for example customs decla-ration forms categorised printed materials inChinese characters alongside pornography, arms,and narcotics). The othering of ethnic Chinesewas justied by references to their origins outsideIndonesia;7 Chinese colonial history; and allega-tions that they socially segregated themselves, over-stated their religious and cultural traditions,achieved economic dominance through economiccrimes, and their primary loyalties lay with main-land China and communism (Heryanto, 1998). Inshort, Chinese Indonesians were portrayed as lesserIndonesians who contaminated authentic Indone-sian identity.

There were frequent, mass attacks on Chineseproperty and life (Heryanto, 1997; Ning, 1997).Indonesian newspapers reinforced Chinese stig-matisation (Coppel, 1983, p. 158). Anti-Chineseattacks were often reported as natural and sponta-neous a populist search for justice provoked byChinese economic domination (Heryanto, 1998).The Chinese turned to the state for protection.This led Chinese businessmen to becoming pariahentrepreneurs having wealth without power(Wang Gungwu, quoted in Blusse, 1990) in anenvironment of conicts, contradictions, and co-operation. They relied upon powerful politico-bureaucrats for protection who exploited Chinese

7 Arabs, Indians, and Indo-Europeans also had externalorigins but did not suer discrimination being too few to

constitute a threat.

business skills for personal economic and politicalends (Mackie, 1992; Robison, 1986).

Soehartos fall changed policies. The newregime brought a more democratic state, civil notmilitary rule, and ethnic tolerance and pluralism.This improved the position of Chinese Indonesiansbut conicts between employers (mainly Chinese)and workers (mainly pribumi) persisted partlybecause wages deteriorated following high Indone-sian ination8 after the 19971998 Asian nancialcrises. According to Ministry of Labour statistics(Binawas, 2000), there were 145 strikes in Indone-

232 S. Eerin, T. Hopper / Accounting, OrganizResearch methods

Methodology

Most research on MCSs and Chinese culture isetic: its categories draw from concepts of bureau-cracy and legal-rational authority in Western soci-eties. Thus ethnographically inclined researchersface a dilemma. They can respect cultural relativismand use emic methods exclusively to generate ana-

8 Ination was 11.79%, 77.54%, 2.01%, 9.35%, 12.55% in1997, 1998, 1999, 2000, and 2001, respectively (Badan PusatStatistik Indonesia/Indonesian Oce of Statistics).9 Major examples include strikes of: 4000 workers from a

garment factory in Bogor-West Java that led to the destructionof the companys oce (Kompas, 15 December 2000); 500workers from a shoe factory in Bekasi-West Java who occupiedan inter-city highway for three hours (Kompas, 4 February2000); 2000 workers from four companies following thedismissal of a worker in Surabaya (Surabaya Post, 22 February1999), and thousands of workers following the dismissal ofseven workers accused of being provocateurs in an electricalappliances manufacturing company in Surabaya, which led tothe destruction of the companys marketing oce (Mediasia during 1999 involving 48,239 workers and915,105 working hours9 but the actual numberwas greater since many strikes went unreported.The institutionalisation of ethnicity over 350 yearspermeates economic activities and MCS practicesin Indonesia for it fosters ethnic stereotypes andsuspicions. For example, many pribumi regardChinese businessmen as exploitative and unscru-pulous whereas many Chinese believe pribumi areunreliable, lazy, and untrustworthy.Indonesia, 26 October, 2000).lytical categories from eld data and not generalisendings beyond the site. However, many ethnogra-phers wish to generalise which requires systematic,cumulative work, better suited to etic approaches.Consequently, anthropological research commonlyuses a mixture of emic and etic approaches (Mar-shall, 1998). This research pursued this path inorder to engage with prior MCS and Chinese cul-ture research using etic categorisations. Theresearch model, derived from the review of culture,ethnicity, andMCSs above, is summarised in Fig. 1.It provided etic categories for analysing the emicdata collected. The MCS propositions were origi-nally restricted to Chinese preferences for actionand cultural controls. The remainder were addedsubsequently to engage with prior cultural contin-gency research.

As Bhimani (1999) notes, conventional contin-gency theory is convergent: it implicitly assumesthat factors such as size, technology, competitionwill make MCSs across the world similar. Itsincorporation of national cultures is signicantfor it introduces ideational notions (Bhimani,1999). The presumption that people behave simi-larly according to their membership in a widersociety assumes cultural beliefs inuence humanaction and perceptions, thus it introduces subjec-tive issues more commonly studied by qualitativeresearch methods. However, most work onnational culture and MCSs, including that onoverseas Chinese, tends to follow nomotheticresearch methods that test hypothesised associa-tions of predetermined attributes of culture, oftenusing research instruments from Hofstede, withdimensions of MCSs measured by instrumentsfrom previous contingency work. As argued previ-ously, this approach is problematic.

The ontological assumption here was that MCSpractices are a product of individual meanings,shared negotiations and are socially constructed(Hopper & Powell, 1985) hence ethnographicmethods to collect data. The epistemologicalassumption was that understanding derives fromclose observation of everyday interactions andactors explanations hence grounded theory toanalyse data to test prior categories and relationsand form new ones (Strauss & Corbin, 1998). We

ations and Society 32 (2007) 223262do not deny that social structures capable of gen-

tem

ls not r

y of Cian

conomoliticaociallytigmatarassmiscrimthnic g

y exteal enden

ganizManagement Control Sys

Personnel and action contro

EthnicitIndones

E P S S

hde

Confucian Values

Vertical Order o Ascribed duties (li)o Filial piety hsiao)

Horizontal Order o Family centrism o Collectivismo Guanxi o Trust o Reputation & face

Personal Qualities Frugality (qin jian) Diligence (qin fen) Prudence (jing shen) Asceticism (ke ji)

Compan& interninterdep

S. Eerin, T. Hopper / Accounting, Oreralisation exist but they are created and sustainedin social interaction (Scapens & Macintosh, 1990,1996). They are subjectively created and sustainedand are capable of exception, change, and contes-tation. They are not laws as nomethetic culturalcontingency research might imply. We questionBhimanis division of possible research methodol-ogies to study culture between nomothetic and ide-ational on philosophical and practical grounds forwhilst it may be useful for expositional purposes itis invalid for, as Bhimani points out, once contin-gency research embraces ideational notions associ-ated with culture it enters an ideational terrain.However, we also reject versions of grounded the-ory that restricts ethnography to rich descriptionsof observations on a particular site (see Alvesson& Skoldberg, 2000; Silverman, 2004). As Bour-dieu, Chamboredon, and Passeron (1991) note,Social science must create its own social conceptsformed in an entirely dierent purview than the

Little participation and centralisati Subjective not objective controls Few rewards tied to results Group-based rewards when reward Few long-term incentives Emphasis on subjective informa

networks Long-run time horizons

Fig. 1. Management control in a Cesult controls

hinese

ically powerful lly vulnerable alienatedisation,ent, extortion,

ination from other roups and state

Javanese Values

Javanese etiquette(alus-kasar)

Language (lair-batin) Correct behaviour

(andap-asor) Social harmony

(rukun) Rituals & mysticism

(slametan) Father figure

(bapakism)

rnal

cies

Other factors

ations and Society 32 (2007) 223262 233notions and frames of reference of the everydayworld (quoted in Alvesson & Skoldberg, 2000, p.33). Thus categories and relationships of culturalcontingency research may be useful for compara-tive analysis but they need corroboration fromeld data, relating to action, not be seen as abso-lute or necessarily persistent, and be open to newelements. Treated in such a manner cultural con-tingency work can form part of integrative eth-nography (Baszanger & Dodier, 2004) consistentwith anthropologists and sociologists argumentsthat understanding phenomena requires an itera-tion of etic and emic views (Scupin, 1998; Smith& Young, 1998; Wilk, 1996).

Data collection and analysis

Gaining access was not easy: overseas Chinesebusinessmen are usually guarded about their com-panies and the research questions are sensitive in

on

s used

tion from social

hinese Indonesian business.

Indonesia. To overcome this, trust and condencebetween the researcher and the respondents needed

234 S. Eerin, T. Hopper / Accounting, Organizestablishment and personal networks pursued.Access in Friends Company10 stemmed from theprincipal researchers previous involvement as aninformal advisor to the company and friendshipwith the owners since 1988. This made him aninsider with empathy and access to the ownersand managers, facilitated social interaction in nat-ural settings, and enabled him to participate inwork activities. It gave a deep understanding ofthe rms activities, personnel, and physical infra-structure, and revealed factors that pass unnoticedby other research methods.

Ethical considerations were important. Toensure participants willingness to participate andto protect them; condentiality and anonymitywere guaranteed; a tape recorder was used onlywith permission; and the researcher disclosed torespondents his true identity, research purposesand how the ndings would be disseminated.

The eld researchers Chinese Indonesian eth-nicity had strengths and weaknesses. It sensitisedand gave him access to emic views of ChineseIndonesian businessmen on sensitive matters.Good access is dicult without prior social rela-tions and reciprocity. Feedback on MCS systemswas an entry condition. The owners were con-scious of their lack of modern managerial knowl-edge and wished to use external advice whereappropriate. During the research a consultantwas employed to improve organisation structuresand systems of accounting but the owners wouldask the researcher, who had no specic assignedduties, for comments on this and other matters.The owners however decided whether to actaccordingly. Spradley (1980) identies four modesof participant observation ranging from non-par-ticipative passive to active involvement. The for-mer helps maintain research objectivity but limitsinsights from social engagement. The researchwas inclined to moderate participant observationrather than action research to balance participa-tion and observation, be an insider and an outsider

10 For reasons of condentiality the names of the company

and personnel have been changed.(Spradley, 1980); and prevent participation dilut-ing critical analysis (Hammersley & Atkinson,1995).

The owners and key employees fully supportedthe research but the researchers ethnicity andentry via the owners was a barrier with pribumiblue-collar employees, who were reluctant to beinterviewed for fear it might aect their jobs.Hence, data about them came mainly from casual,friendly conversations and interviews with externalJavanese gures and, to oset possible researcherbias, Javanese researchers were trained to followa rigorous research protocol involving the secondnon-Chinese researcher to collect data from prib-umi employees. However, events from a pribumiperspective are not presented. The principalresearchers position within Friends, his involve-ment in its aairs, and his ethnicity and valuesprobably skewed his behaviour and interpretationsof events towards the Chinese owners perspective(Hammersley & Atkinson, 1995). The study is notobjective in a positivistic sense. Its objectivity liesin its openness, willingness to listen and observe,and representing respondents views as accuratelyas possible by diligent eldwork and analysis(Strauss & Corbin, 1998).

The eldwork lasted 1 year. Data came frominterviews, documents, and participant observa-tions (Mason, 1996; Spradley, 1980). Repeatedsemi-structured interviews, i.e. conversations witha purpose (Burgess, 1984, in Mason, 1996) wereheld with the four owners and key employees,including the consultant. Further interviews wereheld with ve Chinese businessmen from othercompanies, four pribumi businessmen who dealwith Chinese businesses, and a pribumi Islamicintellectual. Initial interviews sensitised theresearcher to issues for exploration in subsequentinterviews. Interviews with the Chinese ownersexplored ethnicity, control, Chinese social identity,social vulnerability, solidarity, family welfare, whatorganisations mean to them, personal values, andtheir future. The management consultant and theChinese businessmen gave insights into manage-ment in Chinese companies. Trained pribumiresearch assistants conducted some interviews withpribumi businessmen and the Islamic intellectual to

ations and Society 32 (2007) 223262encourage frankness about pribumi perceptions.

tionships proved robust.

ganizQuestions during interviews and observationshelped open up lines of enquiry and direct theoret-ical sampling. All interviews, observations, anddocumentary analysis were in Indonesian andinterviews were tape-recorded and transcribed.Documents studied included charts of accountingsystems and procedures, associated reports, manu-als and documents, budget forms, the cash book,and nancial statements. The researcher had fullaccess to these and they provided an initial under-standing of work processes but analysis of inter-views and observations transpired to be moreimportant for formulating ndings. Field noteswritten during or immediately after each sessionabout the situation, discussions, and nativephrases and terms were cross-checked with inter-view data to promote balance and improve datareliability and validity.

Data analysis consisted of transcription andmicroanalysis (Strauss & Corbin, 1998, Chapter5). Microanalysis (see Appendix) helped generateconcepts and relationships at the outset, identiedand linked native terms and beliefs, guided furtherinterviews and observations, constructed explana-tory statements for comparison with data catego-ries and relations in previous research,systematically coded them into categories with acentral idea, and nally integrated them into amodel. Preliminary results formulated questionsfor subsequent interviews and the focus of subse-quent observations. Interrelations between datacategories and explanatory statements of relations(what, when, how and why) between them wereconstructed. Data categories were derived fromthe theoretical issues of the study (result control,action control, personnel/cultural controls, Chi-nese and Javanese cultural values, ethnicity, tech-nical aspects of organisational activities, historyof the company and its environment). Those pro-cedures were often conducted simultaneously withno rigid, clear-cut boundary between them, andcategories and linkages were constantly renedand modied to ensure the nal categorisationadequately reected data.

Concepts and their relationships were then sys-tematically coded. Similar concepts were groupedwithin a category with one central idea. Once iden-

S. Eerin, T. Hopper / Accounting, Ortied a category became easier to remember and toFriends company

Brief history

Friends Company is situated in an industrialpark in East Java. It is owned by four sharehold-ers: Mr. O, Mr. W, Mr. H, and Mr. A. It producesplastic sheets in rolls or forms such as plastic bags/sachets, inner layers of paper bags and appliancesfor industry and households. Customer require-ments vary according to materials used, sheetdimensions, colour, bundle, model (with or with-out handle) and weight. Selling prices are basedon weight and specications.

Friends was established in June 2000. Its previ-ous owners for 30 years, Mr. Ed and his father,had grown the company from 20 to 250 workersby 19941997. Mr. Ed is a 56-year-old totok Chi-nese Indonesian with no formal education beyondJunior High School. He used traditional Chinesefamily business practices, spurning modern man-agement techniques. There was a simple organisa-develop by breaking it down into its subcategories(when, where, why, how, and so on), and nallyintegrate the major categories into a model.Details of the coding and their explanation andjustication are given in the Appendix.

Theoretical sampling tested the internal consis-tency and completeness of the model. Constantlycomparing concepts against eld data revealedvariations among concepts, made categories den-ser, veried and conrmed links between the cate-gories/concepts being developed, and developednew lines of enquiry and theory evolution. Sam-pling did not necessarily entail gathering new data:returning to old transcriptions/memos was some-times sucient. Sampling continued until catego-ries were saturated, i.e. no new or relevant dataemerged, variation in categories dimensions andproperties were well demonstrated, and relation-ships between categories were established and val-idated. Reaching absolute theoretical saturationproved dicult as new data continually emergedbut the major grounded concepts and interrela-

ations and Society 32 (2007) 223262 235tional structure, a strong paternal culture, and

centralised decision making the owner subjec-tively assessed performance and determined sala-

236 S. Eerin, T. Hopper / Accounting, Organizries. In 19982000 the rm experienced cash owproblems: management could not cope withexpansion and a customer defaulted on a large bill.Employees decreased to around 150 and in 1999Mr. Ed oered his company for sale. An interme-diary introduced Mr. Ed to Mr. O, who conveneda meeting with three of his closest friends, Mr. W,Mr. A and Mr. H. In May 2000, the four friendspurchased the rm.

They introduced modern management, aggres-sive marketing, recruited new employees, andimmediately hired a management consultant, Mr.L, an old friend of Mr. A, to establish a newaccounting system and organisation structure.They invested in machine repairs, maintenance,some new machines, factory enlargement, andredesigned layouts. Maximum production capacitywas 46.7 tons/month but in June 2001 usablecapacity was only 22 tons. Monthly sales werecirca Rp 190,000,000 (15,417). Production islabour-intensive and there are three 8 h shifts. Pro-duction costs were direct materials 50%, directlabour 40%, and factory overheads 10%.Friends total assets in the June 2001 report arevalued at Rp 5.5 billion (458,300). There were190 factory workers and 25 sta.

Fig. 2 details Friends organisational structure.It has three main responsibility centres: Market-ing, Manufacturing, and Finance and Accounting(F&A) Divisions.11 Manufacturing is a cost centre.Mr. H is the director responsible for production,delivery, scheduling, maintaining machines andfactory facilities, inventories, and recruiting manu-facturing employees. Marketing is a revenue andexpense centre. Mr. A is the director. It determinesselling prices, sales terms and conditions, market-ing strategies, sales training, sales incentives, andliaises with Manufacturing over delivery times.F&A is a discretionary expense centre. Mr. W,the President Director, is its director. It supportsthe line with mainly non-nancial data, adminis-ters petty cash, processes cash receipts and pay-

11 Division is used by Friends instead of department and is used

accordingly here.ments, provides nancial information to theowners, and protects company assets againstfraud. Its responsibilities are not quantiable soperformance is judged qualitatively.

Production processes are simple and well under-stood. In contingency theory parlance there is lowtask uncertainty. All production is job order: cus-tomers requirements determine product character-istics variations need dierent inputs andcompletion times. Extruders process plastic oresinto plastic rolls despatched to buyers or processedfurther. Scrap and defects, which can also be pur-chased externally, are recycled into lower qualityplastic ores and re-used, which enables selling pricesto be modied. Mixing new and cheaper recycledplastic ores reduces costs more than minimisingwaste but excessive defects and scrap indicate wasteof new ores or over-use of recycled materials. Pro-duction volume per shift inuences unit costs asworkers wages are time not output based.

The market for Friends products is large forplastic sheets have diverse uses in households andindustry. About 90% of Friends sales are to indus-try: the remainder is to independent wholesalers.Five large companies (including Friends) in EastJava compete for large orders but many smallrms compete for small orders. Friends marketis characterised by a concentrated and longstand-ing customer base (mainly 10 big companies); rev-enue determined by order size not frequency; andfew competitors. Bargaining power with customersis weak: losing one signicantly reduces revenue.

Social backgrounds of organisational members

The four owners (messrs. O, W, H and A) andtwo senior employees (Mr. U and Cik K) playkey roles. Mr. O is a 35-year-old bachelor from arich jiaosen Catholic Chinese Indonesian businessfamily with an MBA in nance from a USA uni-versity. He is not formally involved in Friendsmanagement but attends meetings involving strat-egy: the other shareholders hold him in highesteem as he is an experienced manager. Mr. W(28 years) and Mr. H (32 years) are brothers froma rich, totok Catholic Chinese business family.Their father has several businesses and both wereraised with strong Chinese values. Mr. A is a 28-

ations and Society 32 (2007) 223262year-old bachelor from a prosperous jiaosen Cath-

Direct

r W

ng Dire

H

the Depuction

r U

ganizPresident

M

Manufacturi

MrF&A Director

Mr W

AccountingSupervisor Head of the Dept

of Personnel and General Affairs

Head of of Prod

M

PurchasingOfficer

S. Eerin, T. Hopper / Accounting, Orolic Chinese family who did not nish his under-graduate degree. The four owners have been closefriends from high school or university whencemany business relationships among Chinese Indo-nesians stem. Knowing a persons character is aprerequisite for developing trust and business rela-tions, as Mr. H and Mr. W commented:

All partnerships begin from friendship . . .We must understand the personalities ofour intended partners. If we are not assured,probably the relationship will be limited tomere friendship. (Mr. H)We have known each other for more than10 years. . . . Our characters match, we areall cengli, and we can give in to each other.

Petty Cashier

Accounting Staffs Staffs

Security Officers

Maintenancecoordinator QC staff a

TechniciansHeads of

ProductionShifts

Foremen

Factory Workers

Fig. 2. Friends companyoor

ctor

Marketing Director

Mr A

t

Head of the Deptof Warehousing

Assistant for Manufacturing

Director

Sales Supervisor

Cik L

Secretary

ations and Society 32 (2007) 223262 237We want to be very prudent in choosingpartners. Trust must be developed rstbefore all else. Only after getting close, canwe decide to do business together. In fact,we met with many other people during thatperiod of 10 years. Yet, we feel that they donot match our criteria. (Mr. W)

The other two key players, Cik K and Mr. U,joined Friends in the Mr. Ed era. Cik K (the seniormarketing supervisor) is about 40 years, unmarried,and from a middle-class but not wealthy totok Chi-nese family. She is a senior high school graduate andhas worked in Friends for 17 years. Her nickname,Cik K, denotes employees recognition of herseniority, authority, and social rank. She is well

Production dministration PPC staff

Ware housing staffs Salesman

CustomerServices

Sales Admin.

rganisation structure.

known to important Chinese competitors, buyers,and suppliers, and respected in their guanxi, some-

238 S. Eerin, T. Hopper / Accounting, Organiztimes being almost a family member. She remainsan important business contact. Cik K played a vitalrole in saving Mr Eds business when it was nearbankruptcy. She spoke to suppliers on behalf ofMr. Ed and guaranteed payment. The supplierstrusted her and continued to supply materials,which enabled Mr. Ed to repay his debts.

This reveals four important points. First, theimportance Chinese, external and internal toFriends, attach to integrity, trustworthiness, loy-alty and expertise: without this Cik Ks guaranteewould not have been valued, nor would the newowners have taken pains to retain her. Second,the importance Cik K plays in the guanxi of thisbusiness, which will not be forgotten. Third, thehigh value placed on trust by Chinese business-men: when Cik K gave her guarantee she gaveno money, written assurance, or valuables as apledge. Her word was enough. Fourth, how in aChinese company personal and business relation-ships are inseparable, especially if the employermaintains kinship-like relations. Cik K regardedMr. Ed not only as her formal superior andemployer but also as a senior family memberwhom she must respect and obey, consistent withConfucian values of hsiao. For Cik K, workingin Friends not only meant pursuing a career butalso respecting her long relationship with Mr. Ed.

Mr. U, the head of the production department(circa 38 years) is a nominal Moslem from an eco-nomically comfortable BugisBanjar12 family. Hehas an MBA from a Jakarta university and hasworked 6 years for Friends. Previously he was amanager in a bank, which acquainted himwith Chi-nese businessmen. Mr. Ed was his customer for4 years and a formal business relationship changedinto friendship. The Indonesian government liqui-dated the bank and, following a chance encounter,Mr. Ed recruited Mr. U to handle production andworker aairs. The company was beset by strikesas workers went unpaid due to nancial diculties.Mr. U calmed the situation and got Mr. Ed to hon-

12 Bugis and Banjar are predominantly Moslem pribumi ethnicgroups. The Bugis were originally from Southern Sulawesi and

the Banjar from Kalimantan.ourworkers rights. Subsequently no signicant dis-putes or strikes have occurred.

Pribumi workers considered Mr. U as theirbapak. Although Mr. U is not Javanese, employeesfelt stronger emotional and social ties with himthan the owners because they believed he was prib-umi and thus would guarantee fair treatment fromthe Chinese owners. The owners needed Mr. U tosecure pribumi workers loyalty and eort, as Mr.O recognised:

We need Mr. U because he is our mediatorin dealing with the workers. He looks intelli-gent and cengli. Yet this means that he hasthe capability to provoke them for his owninterests. . . . In the future, we have to be ableto control him so he will not turn againstus.

Thus the owners saw Mr. U as an important butpotentially dangerous employee. Given volatileIndonesian politics and the history of agent provoc-ateurs in Chinese companies he could exploit hisinuence over workers to mobilise ethnic clashes.The owners co-operated withMr. U whilst simulta-neously trying to reduce their dependency on him.

The organisation structure (see Fig. 2) revealsfour management tiers: top management consist-ing of the President Director and divisional Direc-tors; the middle-level consisting of ve supervisors;the lower-level consisting of sta in the F&A,Manufacturing, and Marketing Divisions; andthe bottom-level comprising security ocers andfactory workers. For convenience, bottom-levelemployees are called blue-collar and the remainderwhite-collar.

Dierences in ethnic identity, gender, education,and religion permeate divisional membership. Onlythree Manufacturing employees are Chinese allsenior managers with a degree. There are 46 maleand 155 female employees but all 11 senior manag-ers are men. Only four pribumi employees allsenior managers are educated beyond school.The women have particularly low education junior high or below. All employees are Moslemsapart from seven Christians, none of whom areshop oor workers. Most Manufacturing employ-ees are pribumi (mainly Javanese), female, not edu-

ations and Society 32 (2007) 223262cated beyond junior high school, and Moslems.

dhist, an unknown, and a Moslem. In the F&A

ganizDivision there are three Chinese and three pribumi all Christians with tertiary education and, exceptfor the Director, women. White-collar employeeshave similar religions, minority identities, school-ing, educational attainments, and attitudes, whichmade communication easier and reduced ethnictensions.

In summary, there are few Chinese and Chris-tian employees (5% and 7.5%, respectively) buttop management is their preserve: none areemployed in bottom level jobs. Males and tertiaryeducated personnel are a minority (25.7% and8.9%, respectively) but they dominate top andmiddle management positions and most supervi-sory ones. In contrast, blue-collar employees arepribumi, mainly women, almost all Moslems, withmodest education. There is a high correspondencebetween ethnic identity, religion, education, hier-archical level, and formal authority. In Friends,Chinese means being Christian, educated, seniormanagement, and well rewarded.

Most pribumi workers had negative stereotypesof Chinese employers as wealthy but unscrupulousand cunning opportunists prepared to do anythingfor prot. They claimed ethnic exploitation wascommon in Indonesia and working hard only ben-eted the Cino. On the other hand, the ownersworried that pribumi employees would extort themif not properly controlled. Ethnic antagonism wasexacerbated by the Indonesian economic crisis.Friends paid wages above government rates andhonoured workers legal rights but wages weredeteriorating in real terms, partly because marketprices for plastic sheets had risen less than ina-tion. This increased workers resentment, createdmotivational problems, and increased Friendsexposure to strikes. Workers suspected the ownersused economic excuses for ethnic exploitation.

The companys MCS

Data on control was analysed using etic catego-In the Marketing Division there are ve Chineseand two pribumi. All have tertiary educationexcept Cik K. There are four Christians, a Bud-

S. Eerin, T. Hopper / Accounting, Orries of result, action, and cultural controls andrelated to Chinese and Javanese cultural values(see Fig. 1), ethnic tensions, modern management,and business pragmatism.

Result control

Result controls existed in three areas: budgetsetting, sales targets, and production defect andscrap rates.

Setting budgets

Meetings of the four owners, occasionally Mr.L (the management consultant), and Mr. W andMr. Hs father, established budgets and targetsby consensus. The importance of budgets for con-trolling operations varied. The sales budget wascrucial as it determined overall performance.Other budgets were used to predict working capi-tal requirements: the emphasis lay on nancingand planning rather than management control.The father advised on Indonesian politics, eco-nomic developments, workers, technical matters,and their likely eect upon plans. The owners tookheed because of his experience and their respect forlial piety (li) within family-centric businesses. Theowners would use conventional, modern, MCSpractices but how was inuenced by culture, ethnicproblems, and business pragmatism.

Politics and economics are intertwined in Indo-nesia (Coppel, 1983; Heryanto, 1997, 1998; Robi-son, 1986). Government stability aected riotsagainst Chinese businesses, exchange rates, andination all critical to Friends sales. For exam-ple, in 2000 exchange rates uctuated from Rp700011,000: $US1 according to presidential suc-cession conicts, legal actions against membersof the previous regime, and religious and ethnicriots. A depreciating Rupiah increases many priceswith demand consequences. During the NewOrder era an authoritarian state guarded businessstability and business owed from good relation-ships with it, but Soehartos fall changed this dra-matically. The current political situation, whilstless discriminatory, remains volatile. Reformistsdemands for New Order members to be tried forpolitical and economic crimes can provoke massriots. Given the history of anti-Chinese attacks

ations and Society 32 (2007) 223262 239there was anxiety that dissidents would hire

ganizprovocateurs to provoke strikes or attack Chineseproperties to fuel national chaos.

The owners network or guanxi, often based ondiscussions in pubs, restaurants, or golf clubs, pro-vided crucial information on Indonesian politicsand the economy. Friends owners checked, dis-cussed, and reconrmed predictions with friendsprior to budget meetings. The hostile environmentinvoked solidarity amongst Chinese Indonesians,fostered the guanxi as a medium of informationexchange, and reinforced the owners predilectionto act according to personal trust.

Both totok and jiaosen Chinese used the guanxifor self-protection, albeit dierently. A Chinesesenior management consultant noted:

There are two types of Chinese business:traditional Chinese and modern Chinese.The former emphasises family networkswhereas the latter emphasises social net-works from church or other associations.

Nevertheless, every business and family group hadits own guanxi: when linked to friends guanxisthey were extensive. Becoming a member is noteasy: it requires at least one members personalrecommendation and a record of trustworthinessand reliability. Once accepted, a member enjoysfavourable treatment and support but loss of trustmakes transacting with other Chinese dicult.Reputation must be protected and promisesbe honoured. A Chinese businessman illustratedthis:

Once, we delivered goods to our buyer. Hechecked everything and received them butwhen he wanted to use them the goodsstarted discolouring. We cannot say, Youhave checked and signed the delivery order,although legally we can do so. Finally, wenegotiated how we would share the loss. Thiswas the etiquette adopted to create trust sowe get repeat orders from them.

The guanxi helped nd credible suppliers and buy-ers, nance, trustworthy employees, and businesspartners, provided information and help, and re-duced exposure to the state.

Once political trends were determined, their

240 S. Eerin, T. Hopper / Accounting, Orbudget ramications were translated into technicalconsiderations such as: types of orders to pursue,likely competition, which ores to use, capacity,likely performance, and production costs. Outputtargets were derived from four budgets: sales, pro-duction, direct material purchases, and marketingand general expenses. Key performance indicatorswere sales growth, product costs, productivity, andquality. Budgets incorporated three scenarios optimistic, medium, and pessimistic. The salesbudget was broken down by product and customertype over time and by machines used (new or old).The production budget was more problematic:standard costing was impractical due to the multi-ple product dimensions, though labour and over-heads were relatively xed. Hence the emphasison scrap and defect rates as a surrogate for mate-rial usage control.

All directors discussed each budget and deci-sions required unanimity. There was no divisionaccording to divisional roles, no hint of budgetaryslack, or any manipulation of divisional perfor-mance results because budgets were about familymoney and chung (sincerity and honesty) wastaken for granted. However, such data did notpass beyond the inner group. Behaviour was con-sistent with collective values, qin jian (frugality)and jing shen (prudence and thinking ahead). Vari-ances were investigated in a spirit enunciated as,We are all new in this business, so lets learntogether and support each other. Goal congru-ence, personal trust, informality, and family collec-tivism were important. Friendship prior to beingbusiness partners was essential and simplied bud-geting. Budgets were only used for delegated con-trol with respect to sale and scrap rates.

Sales targets and commissions

The sales target, derived from the sales budget,was important. Mr. A gave each salesman amonthly sales target derived from the mediumlevel budget. Every Saturday he discussed weeklyresults and problems with Marketing employees.A monthly formal performance review was nor-mally symbolic because problems were picked uppreviously. More than 50% of the salesmensincome came from commission based on individ-

ations and Society 32 (2007) 223262ual sales to motivate order maximisation. Mr. A

tor (Mr. H), the production head (Mr. U) andworkers responsible for the defects. Mr. H and

ganizations and Society 32 (2007) 223262 241gave lower monthly commission based on aggre-gate sales to other Marketing employees to pro-mote co-operation, sales support, and clientservice but xed salaries formed most of theirincome. Only salesmen were punished for notreaching targets. If monthly targets were unmet,salesmen could recover by meeting their 3-monthtargets or, failing this, their 6-month targets. If thiswas not achieved the formal rule was dismissal,though this was the owners prerogative. Ethnicantagonisms between employees and owners werelow in Marketing and there was minimal conictover commissions.

Production defect and scrap rates

Manufacturing costs were critical they consti-tuted the bulk of costs. The owners scrutinisedproduction volumes, the direct material purchasesbudget, material price and mix variances, and pro-duction defect and scrap rates but only productiondefect and scrap rates were passed down as opera-tional targets. Delegated output targets and associ-ated rewards were technically dicult due to thelack of a standard product.

Scrap was pieces of plastic sheets not deliveredto customers. The standard rate (0.5% per batch)was derived directly from the budget. Scrap rateswere checked at the end of each batch. Excessscrap indicated worker carelessness in setting-upmachines or faulty machines. No rewards or sanc-tions were tied to scrap rates: they merely helpedprevent workers stealing scrap.

Product quality, essential for customer satisfac-tion and repeat orders, was measured by produc-tion defects. Mr. H transmitted defect ratetargets set by the owners independently of the bud-get to the production head (Mr. U) who relayedthem to foremen and workers. Constant deviationfrom normal rates indicated problems with work-ers, machines, or material quality (the purchasingocers responsibility). Quality control sta esti-mated defects by end of batch sampling and, ifabove 2%, reworked batches could be demanded.Weekly and monthly defect rate reports went tothe President Director (Mr. W) and Mr. H.Defects above standard brought a penalty of Rp

S. Eerin, T. Hopper / Accounting, Or2000 per kilo levied on the Manufacturing Direc-Mr. U were penalised to show divisional memberswere treated equally but workers carried the bulkof the penalty.

Penalties needed sensitive handling to preventblue-collar employees perceiving them as wagereductions and ethnic discrimination. Mr. U, beingthe workers bapak, became a guarantor of man-agements good intentions, as he elaborated:

A nancial penalty is eective since itdirectly aects their daily income. The onlything we have to be careful about is theamount and the way we explain it. They willnot resist if they make mistakes. We need toteach them how to be disciplined andaccountable.

Mr. U explained to workers how improved prod-uct quality would improve Friends future, jobsand wages, and he promised penalties would notbe extortionate. The owners reliance upon Mr.U13 contravened their preference for appointingclose, trusted colleagues, often family, known tobe loyal, sincere and honest. Their dependenceon him meant they had to sacrice values of patri-monial, autocratic management, and sharplydemarcated employer and employee responsibili-ties. Similarly, applying nancial penalties to Mr.H because of his functional responsibility violatedthe owners values of collectivism, shared responsi-bility, and vertical order (li and hsiao). Chinesevalues had to be modied in the face of ethnictensions.

Thus the owners used result controls for criticalvariables especially in sales and production. Aca-demics are highly respected in Indonesia and theowners, who believed their knowledge of modernmanagerial techniques was limited, often soughtto discuss such issues with the researcher. They fre-quently mentioned the need for professional man-agement using modern techniques such as budgets,performance-based rewards, and reliable account-ing information.

13 After the research Mr. U left Friends to work again for Mr.Ed. Since then a pribumi personnel ocer has struggled to

maintain control of production workers.

ganizResult control in production was dicult as astandard output could not be ascertained andmaterial purchases were subject to currency uctu-ations and some customers supplying their ownraw materials. Nevertheless, the owners did notextend budget participation beyond their inner cir-cle and their guanxi due to values of patrimonyand familial control. Subjective political consider-ations initially predominated budgeting thoughthe conclusions were translated into detailed bud-gets. However, ethnic tensions aected employeesperceived legitimacy of controls hence the own-ers had to work through the bapak, Mr. U, andsubject Mr. H to penalties for production defects,in violation of their beliefs that owners responsi-bilities were collective. The owners were not averseto implementing delegated result controls but theyhad to accommodate a complex mix of interactingfactors. Pragmatic and commercial reasons couldcoincide with cultural values but enacting the lat-ter could be inconsistent, for example delegatedresult control may violate patrimonial values butbe congruent with frugality (qin jian) and seekingknowledge and wisdom (qin fen). Control practicesneeded adapting to commercial considerations,technologies, customers, the legacy of employeerelations left by Mr. Ed, local politics, ethnic ten-sions, and desires to incorporate best internationalbusiness practice as well as cultural beliefs.

Action control

In Friends action control was enacted throughthe accounting system (behavioural constraintand pre-action review), authority limits for deter-mining selling prices (pre-action review), andaction accountability.

Accounting system

Accounting systems prohibited certainemployee behaviour. Formal procedures for pro-duction orders, production controls, billing, issuesof cash receipts, payroll, purchasing and receiving,and bank and petty cash disbursements weredesigned to prevent fraud. As Mr. W observed:We cannot just rely upon employees but if wedont believe in them and delegate our authority

242 S. Eerin, T. Hopper / Accounting, Orwe cannot work as well. Hence, we need a systemto secure us. Standardised procedures and docu-ment ows across major transaction cycles (selling,purchasing, cash receipts, and cash disbursement)separated functions (custody, recording, checking,and approval) across the three owner/managers.

The accounting system programmed behaviourthrough rules and procedures but nancial controlwas not delegated. Rather than leaving nancialstewardship to unknown professionals, the own-ers monopolised critical nancial functions suchas approving bills and purchases to personallyprotect family assets and cement ties of personaltrust and protect their reputation with customersand their guanxi. However, outside inner circlesthe accounting system was designed to protectresources in a climate of distrust. Mr. L, the con-sultant who designed Friends accounting systemcommented: The most important thing for themis to avoid being cheated by their employees. Theyare so worried about this. The accounting systemevolved in a context of ethnic prejudice and suspi-cion, as Mr. H noted: Dealing with (blue-collaremployees) is tricky. I dont want to be over-sensi-tive but this system enables us to secure our inter-ests and avoid unnecessary confrontation. Mr.Ed had relied on spasmodic personal surveillancefor internal control, which resulted in a majorfraud. The current owners wished to avoid thisby adopting accounting procedures that sup-pressed ethnic distrust by appearing neutral, objec-tive, and administered discipline and tidinessregardless of rank or ethnicity. Such a bureau-cratic control system was culturally legitimate forowners and employees alike. The owners realisedthey could not rely on traditional, Chinese familybusiness practices like trust and informal transact-ing exclusively. Their nancial caution combinedChinese values of frugality and money conscious-ness (qin jian), and prudence/thinking ahead (jingshen) with business pragmatism.

Authority limits in determining selling prices

Each salesman had limited discretion to negoti-ate selling prices. If a customers wanted a priceless than the salesmans discretionary limit heneeded approval from Cik K the marketingsupervisor. If the price was below her discretionary

ations and Society 32 (2007) 223262limit (which was higher) the transaction went to

S. Eerin, T. Hopper / Accounting, Organizations and Society 32 (2007) 223262 243Mr. A who had no limit. If he was unsure, whichoccurred frequently with large orders, he secureda consensus decision from the owners. Thus,bureaucratic rules provided limited delegationwithin Chinese cultural preferences for formalmanagerial hierarchy and collective decision-mak-ing by owners. Formal limits of authority onemployees reconciled distrust of outsiders withpragmatic recognition of needs for constraineddelegation. Individual authority and responsibil-ity, and performance evaluation of owners actingas divisional managers was irrelevant due to theowners collectivist values and mutual personaltrust.

Action accountability

Elsewhere denitions of acceptable action lay inunwritten, shared understandings of Friendsmembers. Employees were expected to obey own-ers orders, be loyal, i.e. show lial piety (hsiao);be diligent (qin fen), helpful, and above all, trustthe wisdom and prudence of the owners (jing shen),demonstrate honesty by being truthful and sincere(chung), and elevate company interests above per-sonal goals and accomplish assigned duties (li).These expectations had cultural legitimacy withowners and employees alike.

For the owners this derived from Chinese valuesof respecting vertical order (li and hsiao). Mr. Wcommented: If there is a highly trustworthy,loyal, committed and capable person, he will beour right-hand and be given more authority. Suchan employee is ideal. Chinese employees sharedexpectations of lial piety (hsiao) due to their Con-fucian parental education. For example, Cik Kcommented:

One has to serve ones master as best as onecan. In the past, my mother worked for acompany and she also did her best. WhenMr. Ed had problems with the workers inthe past he was taken to the Bakorstanas. Iwas there with him from 11 a.m. till 2 a.m.the next day. He was the only one interro-gated by the ocer but I was there to helphim as best as I could. . . . Employees skillis important. However, I always observe

which one of them is ready to lend a handto his colleagues. . . . Even if a duty is not for-mally mine, I dont mind helping them out-side working hours.

According to traditional Chinese managementvalues authority stems from personal ownership not written, binding agreements. Constrainingbehaviour by codied rules contradicts both Chi-nese and pribumi employees belief that delegatingauthority and responsibility is the employers priv-ilege, so employees never sought written jobdescriptions. Given the ethnic antagonism ofblue-collar employees the owners were wary ofdoing so. Mr. U commented: They dont needtoo many formalities such as a written job descrip-tion . . . [employees] may use it as a weapon toneglect their responsibilities. . . . Instead of pro-moting order you may get more trouble and con-ict. The symmetry between Javanese values ofbapakism emphasising loyalty and obedience ofservants and Chinese values of vertical ordermeant owners and pribumi employees had com-mon cultural expectations regarding performance,written job descriptions, and rules.

The owners direct monitoring of employeebehaviour was largely conned to white-collaremployees observing blue-collar employees wasimpractical due to their number and ethnic ten-sions. Mr. U and worker informants were a crucialsource of information on Manufacturing employ-ees behaviour: they were the eyes and ears of own-ers in the factory. They reported which employeeswere obedient, loyal, honest, and skilled, and theyrelayed ethnic tensions promptly so precautionscould be taken quickly. Workers trusted Mr. Uas their bapak and relayed their aspirationsand complaints to him. Mr. H also cultivated per-sonal relationships with informants experiencedblue-collar employees with proven loyalty to theprevious owner, who often gleaned sensitive infor-mation. Often Mr. H relied more on their observa-tions than supervisors reports. The ownersreliance on Mr. U and worker informants contra-vened their preference for relying on trusted peoplebut, given mutual distrust between pribumi work-ers and the owners, and the practical constraintson personally exerting action control, they were

pragmatic in sacricing core values.

During Mr. Eds ownership there was no stan-dard incentive scheme. His arbitrary determination

244 S. Eerin, T. Hopper / Accounting, Organizof wages produced discrepancies between new andold employees salaries that compounded employeeperceptions of ethnic discrimination and little con-nection between eort and rewards. Both blue-and white-collar employees suspected that Mr. Edfavoured Chinese employees. Mr. U conrmed thiswhen explaining why so many strikes occurred:

During the era of Mr. Ed the ethnic dis-crimination was very apparent, especially interms of salary. . . . There was a Chineseemployee in the production department. Hereceived Rp 1,000,000 for his rst salarywhile my rst salary was only Rp 400,000.Many employees asked me: Look at him,do you accept this? The ethnic gap is muchless now and this is more eective for con-trolling employees.

The new owners systematised salaries accordingto position, length of service, dependants andachievement to promote perceptions of fairness.Workers salaries now consist of the main salarybased on government UMR policy, a night shiftpremium, length of service, and a skill allowance,14

paid on daily attendance. Wages of other blue-col-lar employees (heads of shift, technicians, andforemen) comprise the main salary (standardisedbelow junior management salaries) and allowancesfor length of service, dependants, and skills.

White-collar employees received their main sal-ary plus allowances for length of service, depen-dants, and achievement. Standard wages betweenDivisions varied and, except for achievementallowances, across management levels. Achieve-ment allowances were based solely upon the own-ers subjective judgment. Salesmen also received atransport allowance and, like other Marketingsta, sales commission. In addition, employeesreceived an annual incentive called Tunjangan HariRaya (THR) on their main religious days. Moslememployees and factory workers receive this beforethe Eid-Al-tre/Idul Fitri celebration (usually the

14 The skill allowance is for the few workers able to operate

extruders an