Introduction

The third sector is recognised as the alternative supplier of public goods, complementing both the public and private sectors (Etzioni, 1973; Salamon & Sokolowski, 2016; Weisbrod, 1972). It plays a central role in the delivery of collective goods and services, promotion of civic action and policy development (Barman, 2013) and has a sizeable economic share in terms of employment and GDP in many developed countries (Casey, 2016; Sanders et al., 2008). The third sector embodies citizen empowerment, where individuals and groups take independent action to fill the gaps in the provision of social welfare goods and services. Voluntary, charitable and nonprofit nature of the sector suggest that it is driven by altruistic intentions, and not solely by financial rewards.

With some exceptions, the bulk of the literature conceptualising and describing the third sector focuses on organisations in high-income countries. The Malaysian third sector is not sufficiently described, and it lacks information on its structure, scope and activities. Only four English language papers on the Malaysian third sector were published between 2001 and 2013 (Hasan, 2015) and most papers published after 2012 mainly discussed accounting and financial disclosure aspects of Malaysian nonprofits. The dearth of scholarly discussion on the sector has led to a poor understanding of its role, contribution and relevance to the economy (Arshad & Haneef, 2016).

In order to understand the Malaysian third sector better, we ask: What are its constituents, their activities, and who do they serve? In doing so, we reflect on the extent to which Western-focussed theories of nonprofit organisation describe and explain the Malaysian third sector. The focus of our inquiry is therefore on the question of “what is” instead of exploring “what should be” the constituents of the Malaysian third sector. This is a definitional work, defined by Macmillan (2013) as an endeavour that “seeks to identify conceptually what holds objects together in a boundary and what distinguishes them from other entities—what is in and why, what is out and why”. We aim to establish the sector’s population profile, taking on a metatriangulation approach (Lewis & Grimes, 1999), emphasising on identifying the research interest, analysing data patterns and defining the identity of the phenomenon investigated. “Identification” is where constituents are ascertained, and exclusions are justified. “Analysis” refers to the process of classifying constituents based on activities and beneficiaries. “Defining” is where the findings are interpreted, formulating an identity that represents the prominent characteristic of the sector. In agreement with Never (2011), we argue for comprehensive mapping of the third sector as a means to facilitate research and assist decision-making through the identification of its constituents and their capabilities. The ability to identify, measure and study third sector components legitimises its existence (LePere-Schloop et al., 2021).

We begin by examining the various definitions of the third sector followed by a discussion on the sources of Malaysian third sector organisation (TSO) data and the mapping process. Next, the findings are presented and discussed. The paper concludes with a proposed description of the Malaysian third sector.

Defining the Malaysian Third Sector

We use Weisbrod’s government failure theory (1977) to conceptualise the Malaysian third sector as it focusses on how the interests of different groups in society might play a role in defining a nonprofit sector. Weisbrod posits that governments provide public goods at the level demanded by median voters, the group whose votes bring victory to the ruling government and less attention given to the voters further from the median made them turn to the third sector for their social welfare needs. Purchasing from the private sector may only provide poor substitutes in sub-optimal quantities of unknown quality which further necessitates the under-satisfied demanders to turn to the third sector for their supply of public goods and services. The theory, therefore, suggests that the relative size of the third sector depends on the heterogeneity and diversity of population demand, the less able the public sector to satisfy the demands of the population, the larger the expected size of the voluntary or third sector.

Non-distribution of surpluses was a primary criterion for the sector and its organisations (Hansmann, 1987; Salamon & Anheier, 1996), but the “non-distribution constraint” is latterly viewed as too restrictive; hence, profit-distributing cooperatives and social enterprises are proposed to be included into the definition (Salamon & Sokolowski, 2016). “Civil and solidarity-based economy” was suggested to be a suitable description, the solidarity and hybridisation of different economic principles as foundations of the sector and viewed from the European perspective, including social economy elements (Evers & Laville, 2004). Etzioni (1973) includes public–private partnerships into the scope of the third sector, but Lorentzen (2010) views extensive cooperation between the state and market blurs their borders, could eliminate the concept of an independent third sector and may give the impression of a residual category, housing organisations that do not fit into the other two so-called primary sectors (Corry, 2010).

Salamon and Anheier (1996) define the sector as a structured and independent nonprofit distributing group of organisations with significant voluntary participation, occupying an autonomous social space outside both the government and private sectors. Their structural-operational definition, however, defines the sector mainly from a North American perspective. Critics argued that it does not take into account the organisational diversity in particular those which are historically key components of the sector (Morris, 2000) and ignored organisations’ purpose by not taking into account the objectives and intended beneficiaries of some nonprofit organisations (Mohan, 2011).

Despite the many attempts to define the sector’s scope, there is still no clear agreement about what it included, or what it should be called, making it one of the most perplexing concepts in modern political and social discourse (Salamon & Sokolowski, 2018). It is described by Kendall and Knapp (1995) as a “loose and baggy monster” due to the myriad of organisations that constitute the sector. Nevertheless, the common characteristic agreed by a majority of researchers in the third sector lies outside the full control of the public and private sectors.

For the purposes of this paper, the scope of the Malaysian third sector is based on the boundaries proposed by Perai (2021) which was guided by Salamon and Anheier’s (1996) structural-operational definition. It includes registered non-profit distributing organisations which are institutionally separate from the government whose operations involve meaningful degree of voluntarism in personnel and finances. We focus on formal organisations registered with any of the following Malaysian regulatory bodies:

  1. 1.

    Registrar of Societies (ROS),

  2. 2.

    Companies Limited by Guarantee (CLBG) registered with the Companies Commission of Malaysia (SSM),

  3. 3.

    Office of the Sports Commissioner (SCO),

  4. 4.

    Registrar of Youth Societies (ROY),

  5. 5.

    Trusts and foundations registered with the Legal Affairs Division of the Prime Minister’s Department (BHEUU),

  6. 6.

    Charitable foundations registered with the Labuan Financial Services Authority (LFSA),

  7. 7.

    Department for Trade Union Affairs (JHEKS) and

  8. 8.

    TSOs governed by individual Acts of Parliament and state ordinances.

The pursuit of a well-described, specific set of organisations allows for a tight and distinct boundary thus avoiding the problem of the sector having “fuzzy and permeable boundaries” (Macmillan, 2013). Inclusion of faith-based organisations, political parties and labour unions ensures the diversity and impact of the sector are accounted for. Organisations governed by each regulator also meet the service provision, mutual aid and pressure group functions outlined by Brenton (1985) and Hall (1987). Following this definition, profit-distributing cooperative societies and social enterprises are excluded from the scope of the Malaysian third sector as they contravene the non-distribution condition. Hybrid organisations are placed in the sector that controls its direction and objectives, in-line with Salamon and Anheier’s (1996) criteria of a nonprofit organisation. While informal groups are increasingly recognised as an active participant in the third sector, their absence from any official database prevents their inclusion in a quantitative exercise such as this.

Mapping the Third Sector

Mapping is a form of classification exercise, a systematic arrangement of the components where objects are sorted based on criteria representing those objects, differentiating them from each other (Appe, 2011; Niknazar & Bourgault, 2017). It follows a conceptual or descriptive typology where categories are created and classified, from which concepts are formed and refined, a process that accumulates knowledge which leads to the development of theories (Kwasnik, 1999; Collier et al., 2012). Maps present a “statistical portrait” of the sector’s characteristics and resources (Kane & Mohan, 2010). Their purpose and structure are determined by the mappers based on its aims (Appe, 2012; Niknazar & Bourgault, 2017). Mapping and identifying its components ensure third sector organisations, and their services do not remain the “invisible continent on the social landscape” (Salamon, 2010). Maps present organisational data in a given geographical area, enabling its characteristics to be extracted, and allow identification of the pattern vis-a-vis the demographics and economic characteristics of the community they are in (Mohan, 2011; Nickel & Eikenberry, 2016).

A well-described sector facilitates policy development, improves resource allocation and increases its overall effectiveness (Banks & Brockington, 2019; Barman, 2013; Kane & Mohan, 2010; Never, 2011). Maps can be used by third sector actors, policymakers, research communities, donors and beneficiaries to develop strategies and policies and allocate resources to where they are needed most (Never, 2011). Data from maps assist regulation, provide information for comparative and policy development purposes and identify potential partnership or collaboration opportunities (Appe, 2011; Banks & Brockington, 2019). In addition, knowledge of the sector’s diversity and density would help in understanding societal structures and relationships (Salamon & Sokolowski, 2018) and data from mapping permit such studies. In producing a map of the Malaysian third sector, we can critically reflect on the fit and suitability of the theoretical explanation and categorisations of sector activity.

Douglas and Pedersen (1973) presented a detailed overview of the Malaysian third sector, but it has never been presented in totality since, as many recent studies did not account for all possible organisations that make up the sector. Kasim et al. (2006) published a report on philanthropy and the third sector in Malaysia but despite providing an in-depth description, the information presented requires updating as it no longer reflects the present third sector landscape. The recent literature did not consider organisations registered with the Registrar of Youth Societies, the Office of the Sports Commissioner, Labuan Financial Services Authority and the Department of Trade Union Affairs, to be part of the sector. They were excluded because the scope of the sector was not sufficiently defined, resulting in the sector being under-described.

Maps can be presented from the supply side to show the distribution, activities and size of TSOs or from the demand side to show geographical regions and the demand for third sector products and services in each. This paper focuses on the former—the map of the Malaysian third sector is presented showing the geographical distribution of TSOs, their activities and beneficiaries. Organisational data are arranged based on the International Classification of Non-profit Organisations (ICNPO), enabling the TSOs to be categorised for descriptive and comparative purposes. Grouping TSOs according to the ICNPO enables classification in an internationally recognised system and facilitates methodical identification of the sector’s constituents. The ICNPO groups TSOs based on its major economic activity, and in cases where an organisation has multiple activities, its “major economic activity” is the one that consumes the largest share of the operating budget (Salamon & Anheier, 1996: 3). It consists of 11 groups and 27 subgroups of specific activity or beneficiary plus one group for organisations not covered by the other groups. The INCPO classification is detailed in Table 1.

Table 1 International classification of nonprofit organisations (Salamon & Anheier, 1996)
Full size table

The ICNPO can be tweaked to present a more detailed and representative classification of TSOs. Additional subgroups may be necessary to classify groups with sizeable number of organisations in order to present the sector better (Smith, 1996). This has been done to differentiate medical research sponsors from medical researchers (Mohan & Barnard, 2013), to separate kindergartens and preschools from primary and secondary schools, to differentiate grant making from fundraising organisations, and to separate business and professional associations from labour unions (Sanders et al., 2008) and to distinguish between Jewish and non-Jewish religions TSOs, faith-based financial and in-kind aid organisations, and a subgroup for memorial organisations (Gidron et al., 2003: 183–186).

We added three new sub-groups to the ICNPO: 2500 groups alumni and student associations, 7400 for organisations promoting the rights of specific ethnic groups, and 10200 for burial associations. A large number of organisations in the Malaysian third sector would be better represented by having their own subgroup as they are not well captured within the standard ICNPO classification.

Sources of Organisational Data

Data on ROS registered organisations were obtained directly from the regulator, but only basic organisational information was provided. Data on sports associations and youth associations were obtained from the Office of the Sports Commissioner and the Regulator for Youth Societies, respectively. There are fewer than 400 trusts and foundations registered, and their details were freely available on the BHEUU website at http://www.bheuu.gov.my/. Organisational information on CLBGs was obtained from Malaysia’s open data portal (www.data.gov.my) but only their names and registration numbers. Additional information was extracted via web scraping from the websites of two Malaysian private credit reporting companies, CTOS Data Systems and RAM Credit Information, whose data are sourced directly from the SSM to build businesses’ credit profiles. A disadvantage of obtaining data from the internet is the difficulty to verify data quality. However, the two companies have significant reputation and standing which provided assurances that data contained in their websites are of reasonable quality and are assumed to be free from major inaccuracies.

Communication with The Labuan Financial Services Authority’s customer relations unit via email confirmed there are 17 foundations registered with them, but requests for more details were unsuccessful. Request for data from the Department for Trade Union Affairs was also not successful; therefore, Malaysian trade unions data had to be sourced from the government open data portal (www.data.gov.my). However, they were not extensive, and only included a summary of the total number of trade unions and their affiliations. As a consequence, we were not able to present any descriptive statistics on the distribution of trade unions based on the year/period of establishment or the state where there are registered.

Classification Process

Organisation names can be a rich source of information on their objectives, activities and beneficiaries. Inspection of the data showed that organisation names contained many descriptive keywords that could be used for classification. However, there are limitations in classifying based on organisation name: names can be non-specific, or describe only part of an organisation’s focus, or be ambiguous. For the Malaysian third sector, richer data on organisation aims, objectives and activities are not available; therefore, the classification process was guided by values and understanding of the local landscape which in turn influenced its outcome, similar to the work of LePere-Schloop et al. (2021). In our classifications, we try to mitigate these limitations through use of other data fields where available regulatory classifications and manual-checking of ambiguous terms.

Due to the unstructured nature of textual data, machine-assisted classification was required to assign organisations into the relevant ICNPO group based on keywords in their names. All seven regulator datasets were merged into one central dataset and uploaded into the statistical software “Stata” where a coding algorithm was used to organise and analyse the data.

Keyword Identification

The ICNPO classifies TSOs according to the activity or service that consumes the largest portion of its budget. However, as the Malaysian Third Sector Dataset does not contain precise information on major economic activity or finances, classification was based on combinations of words contained in their names that identifies their major economic activity. Some organisations have more than one identifier in their names, and these were manually checked and assigned based on the objective indicated by their full name. Organisations were assigned to only one primary category to avoid double-counting.

Keywords used to assign TSOs to the respective ICNPO groups were identified in two stages. The first involves listing common terms used, both in Malay and English, to describe an organisation such as “association”, “organisation”, “foundation” “society”, “club”, “movement”, “body”, “chamber”, “guild” and “council”. This was followed by adding common activities or beneficiaries to the identified words, for example, care home or orphanage, nursing homes, chambers of commerce, professions, youth associations, sport, house of worship, religion, or ethnic group. Also included are: residents’ associations, alumni associations, employee welfare organisations, arts and culture organisations, social and recreational clubs and names of organisations identified by ICNPO such as “Kiwanis” and “Young Women/Men Christian Associations. Selection of common terms was based entirely on observation of organisation names in the Malaysian Third Sector Dataset, and classification was made based on their most common application and meaning.

First Stage: Stata Assisted Classification Based on Identified Keywords and Phrases

Similar to the approach undertaken by Litofcenko et al. (2020), our classification uses an algorithm to assign the inputs to the required output based on an “IF—THEN” rule. In this instance, the input is the keyword in the organisation name, and the output is the relevant ICNPO category. “IF” the keyword is found in the organisation name, “THEN” Stata assigns it to the ICNPO group identified for that keyword. However, some organisations were not assigned to any group, as none of the keywords were found in their names.

Second Stage: Classification of Regulator and ICNPO Defined Organisations

The next step classifies organisations whose categories are “regulator defined” into the correct ICNPO group; keywords were not used to identify and assign the organisations, but instead, classification was based on which regulator the organisation is registered with. Sports bodies registered with the SCO were classified under ICNPO 1200 because they are deemed to be involved in the provision of amateur sport, physical fitness and sporting events by virtue of being registered with the sports regulator. Youth organisations registered with the ROY were classified under 7100 because they are deemed to promote and serve the interests of specific group of people, i.e. youths. Organisations classified under the “politics” category by the ROS were assigned to ICNPO 7300, and due to the small number of organisations involved, manual checking to ensure accuracy was possible. Also assigned at this stage were organisations whose categories were determined by ICNPO, as well as burial and funeral associations which were not assigned to ICNPO 10200 during the first stage.

Third Stage: Classifying Organisations with Multiple ICNPO Matches

Next, organisations matched to two and three ICNPO groups were assigned to the correct subgroup. Each organisation assigned to two or more ICNPO subgroups was manually examined to determine the correct primary subgroup from their full names and re-assigned.

Fourth Stage: Assign Organisations in ICNPO 12100 to the Correct ICNPO Group

After all organisations tagged to more than one ICNPO subgroup have been assigned, organisations which were machine classified into ICNPO 12100 were scrutinised and re-assigned to the relevant subgroup. The variable in the Malaysian Third Sector Dataset that describes the activity of ROS’ registered organisations and the variable describing the activities of CLBGs were used to assist in this process.

Fifth Stage: Assign Remaining Unmatched Organisations to Most Relevant ICNPO Subgroup

The final stage was to assign the remaining unmatched organisations to a subgroup that best reflects their activity or beneficiary, assigned according to their categorisation in the ROS dataset. This final step completes the classification of Malaysian third sector organisations into the ICNPO groups (Fig. 1).

Fig. 1
figure 1

Workflow to classify Malaysian TSOs into ICNPO groups

Full size image

A machine-assisted classification strategy is crucial due to the large number of organisations in The Malaysian Third Sector Dataset making it too large to be manually processed. Replicability is a strength of the algorithm used in this study as any text dataset that requires classification can apply the algorithm with minor linguistic changes to the keywords and/or categories. The codes can also be further refined with more data on TSOs’ activities or objectives. The algorithm is not restricted to classifying TSOs; being a keyword-based classification code, it can be used to classify any text dataset. The codes are made available through an online project management and code versioning system, which is also used as the repository for the raw data used in this paper.

Findings

A descriptive analysis is presented to describe the main features of the Malaysian Third Sector Dataset, to understand its characteristics and to identify patterns within the data. Figure 2 shows the distribution of Malaysian TSOs by regulator. Almost three quarters of registered Malaysian TSOs are under the purview of the ROS.

Fig. 2
figure 2

Malaysian third sector organisations by regulator

Full size image

Figure 3 shows registered organisations by state where they were registered. One third of Malaysian TSOs are registered in the Klang Valley where a quarter of the Malaysian population reside.

Fig. 3
figure 3

Registered organisations by state. Note: Information on trade unions not available

Full size image

Being registered in a particular state does not prevent a TSO from organising activities or providing services in other states, having a large number of registered TSOs does not suggest the state or region has an active third sector. Therefore, the number of registered TSOs may not be a good indication of the prevalence of third sector activity in a geographical location or region. Getting accurate measure of third sector activity requires more data, such as, the frequency of activities or amount of money spent on such endeavours in each region. Focussing on the number organisations registered in each state may lead to a “registered office bias”, where conclusions on the size of sector may be misled by the number of registered organisations in a given jurisdiction.

Figure 4 presents the distribution of Malaysian TSOs by ICNPO group. More than 30% of TSOs in Malaysia are classified as cultural and recreational, and the three largest groups make up 60% of the Malaysian third sector population.

Fig. 4
figure 4

Malaysian third sector organisations by ICNPO group

Full size image

Collectively, subgroups in ICNPO Group 1 make up the largest proportion of Malaysian TSOs, but individually, ICNPO 10100—Religious Congregations and Associations, has the largest number of registered organisations. 15.4% of registered TSOs are classified as involved in administering religious beliefs and rituals. However, most organisations listed in ICNPO 10100 are non-Muslim, indicating that their houses of worship are mainly privately funded, unlike mosques, which are largely state funded. Searching for the keyword “masjid” (mosque) only found 16 entries but 12 of them are classified as what Smith (1996) terms as “non-integral congregation groups”. These organisations have significant recreational or social component but are not the mosque itself. There are 6,506 mosques in Malaysia (Jabatan Kemajuan Islam Malaysia, 2021), and a sizeable portion of building and maintenance costs is borne by the government. Each state also has a government department assigned to matters relating the welfare and development of Muslims, a privilege not accorded to the other faiths, and hence, there is less need for the Muslim community to turn to the third sector for assistance on religious matters. Services and support provided by the government to the majority, Muslim community suggest that they are the median voters and this explains the presence of high number of non-Muslim places of worship in ICNPO 10100. Associations providing funeral assistance and cemeteries are also commonly registered as third sector organisations. However, these organisations are evenly distributed across the different religions which suggests that while the Muslim community depends on the government for their houses of worship, their funeral assistance is provided by the third sector (Fig. 5).

Fig. 5
figure 5

Malaysian third sector organisations by ICNPO sub-groups. Note: 14 other sub-groups make up 1.82% of all registered TSOs

Full size image

Almost 30% of the Malaysian TSO population is comprised of social recreational associations and sports clubs. TSOs in ICNPO 1200 are open to everyone except ethnic-based sports associations and paralympic sports bodies. TSOs in ICNPO 1300 are mainly members clubs, and most are social recreational or family of private companies or government agencies and departments. Member serving organisations can be defined by their degree of accessibility (Gordon & Babchuk 1959); highly accessible organisations are those with minimal membership qualifying criteria, while low accessibility organisations restrict entry by imposing qualifying criteria such as ethnic, religious, academic or vocational background. We found two types of member-serving organisations in the Malaysian Third Sector Dataset. Based on Gordon and Babchuk’s definition, we term the first as “voluntary membership organisations” whose membership is open to anyone. The second type is termed as “restricted membership organisations” which are only open to those meeting specified criteria imposed by the organisation.

Trade unions, chambers of commerce, and professional associations are classified as restricted membership organisations, and they make up 9% of the Malaysian TSO population. Almost all of TSOs in ICNPO 6100 are residents' associations, which are also restricted membership organisations. They are included in 6100 because their objectives do not meet ICNPO 6200’s definition of housing association or housing assistance. 6200 is the only subgroup that is not relevant to the Malaysian third sector as residents' associations in Malaysia promote and protect the welfare and interests of residents of a particular neighbourhood, they do not function in the same way as housing associations whose object is to provide public housing assistance and related legal services.

Another subgroup in Group 6, “Employment and Training” also has small number of organisations. This could be due to the Human Resources Development Fund (HRDF) of the Ministry of Human Resources providing support for high-skilled training and certification. Availability of employee development programmes provided by a well-funded state-backed agency means there is less space or need for the third sector to offer similar service.

Two ICNPO subgroups with the lowest entries are 3100 (Hospitals and Rehabilitation) and 2200 (Higher Education). Only eight organisations in the Malaysian Third Sector Dataset are registered as nonprofit hospitals. Medical services in Malaysia are mainly provided by either the public or private sectors and small number of nonprofit hospitals suggest the third sector is not relied upon to provide formal medical care as it is seen to be sufficiently provided by the public and private sectors.

Similarly, higher education is also mostly provided by the government and private sectors; there are 21 public universities and 71 private universities and university colleges including branch campuses of foreign universities operating in Malaysia (www.data.gov.my). Like private hospitals, a number of institutions of higher learning are listed on the Malaysian stock exchange which suggest that both health and education are lucrative businesses. As with higher education, primary and secondary educational facilities are mostly provided by the government and increasingly, the private sector. Registered nonprofit schools are very few and mostly are independent Chinese schools or Islamic religious schools. Organisational data therefore suggest that the Malaysian third sector plays only a small role in the provision of formal education. Twelve of the 29 ICNPO subgroups make up 95% of the Malaysian TSO population. The three additional ICNPO subgroups introduced in this paper were aimed at providing a better picture of the Malaysian third sector.

Exclusive, Member Serving Organisations

“Exclusivity” is defined in this paper as services or activities offered only to a select group where members or beneficiaries are based on various criteria. “Exclusive” is used in this context to highlight the criteria for membership. It is not limited to the universally common ones such as ethnicity and religion, but also include education institution attended, employer, profession and residential address. Normally, charities define beneficiaries in terms of the services, such as for the disabled, single mothers, refugees, orphans, homeless, without additional criteria such as ethnicity, religion, school attended, residential address or profession.

Organisations in ICNPO subgroups 2500, 6100, 7400, 10100, 10200 and 11100 are classified as exclusive or “restricted membership organisations” because membership is restricted to education institution attended (2500), residence (6100), ethnicity (7400), religious belief (10100 and 10200), or profession or trade (11100). Some organisations in ICNPO 1300 are also exclusive, particularly social and recreational clubs of corporations and government department and ministries. These TSOs are mainly sporting, recreational and welfare clubs catering to their employees and their families. Parent–Teacher Associations and alumni associations based on ethnicity or religion were found among Malaysian TSOs, and this can be considered a unique case. Exclusive organisations make up 43.15% of all Malaysian TSOs and this sizeable percentage of registered TSOs that serve only selected beneficiaries can be considered to be on the high side which suggests that exclusivity is a feature of the Malaysian third sector. Table 2 presents the number of exclusive TSOs.

Table 2 Exclusive, member serving organisations in the Malaysian third sector
Full size table

This is made more complex with some TSOs having two-level exclusivity, for example, ethnicity and trade, such as “Chinese Chamber of Commerce”, or religion and welfare services, such as “Muslim Residents” Association”. This “two-level” exclusivity adds to the sector’s peculiarity which is, based on our research, not observed elsewhere.

Exclusivity could also be attributed to Malaysian society being fractionalised along ethnic and religious lines. 7.5% of registered TSOs are ethnic based and 18.4% are religious, a quarter of Malaysian TSOs cater exclusively to members of a specific ethnic, or religious group. The government’s affirmative action policies focussing on the welfare and development of native Malay and Bumiputera population are likely explanations for non-Malay/non-Bumiputera organisations making up 70% of TSOs with ethnic keywords in their names. This supports the applicability of Weisbrod’s (1977) median-voter-focussed government failure theory on the Malaysian third sector, analysis of the sector’s constituencies vis-à-vis state policies, in particular affirmative action programmes, suggests that state policies favouring median voters may have influenced the composition of the sector.

Ethnic and Religious Fractionalisation in the Malaysian Third Sector

Malaysia is a plural nation and the origins of Malaysian third sector organisations coincided with the arrival of immigrants from China and India during the late nineteenth and early twentieth centuries. Support systems that help immigrants adapt to a new environment are often in the shape of social groupings, mostly in the form of clan, cultural, language-based or religious associations. It has been suggested that the nonprofit or third sector is larger in countries with ethnic and religious diversity (James, 1987; Weisbrod, 1977). Lu (2020), however, cautions the blanket application Weisbrod’s theory, suggesting organisational characteristics, population age and education also be included as measures of nonprofit size.

Fractionalisation in the Malaysian third sector is seen from the prevalence of ethnic and religious-based organisations. Ethnic and religious terms in organisation names reflect their objective and beneficiaries, identifying the target group of its activities or services. Ethnic- and religious-based organisation are also common in many countries; 9.2% of charities in the United Kingdom (NCVO Almanac) is faith-based, 18% of registered charities in New Zealand conduct religious activities (Charities Services New Zealand) and the figure is 11.4% in Ireland (Charities Regulator Ireland), 9.3% in the United States (Guidestar), 46.2% in Singapore (Commissioner of Charities) and 47% in Japan. In Australia, a “Basic Religious Charity” is a restricted Australian Charities and Not-for-profits Commission category for faith-based charities, one which cannot include any other charitable activity, and they make up a mere 0.6% of total charities (Australian Charities and Not-for-profits Commission). At 18.4%, the proportion of faith-based TSOs in Malaysia is lower than Singapore and Japan but is higher than the UK, the US, Australia and Ireland. Ethnic-based TSOs make up 4% in the UK, and in Australia, only 102 (0.02%) charities are exclusively ethnic-serving. New Zealand, Ireland, the US, and Singapore did not identify any registered TSO as ethnic-based. At 7.5%, Malaysia has a much higher proportion of ethnic-based TSOs.

Table 3 shows TSOs with ethnic terms in their names but were not classified into ICNPO 7400 because they are not primarily focussed on promoting ethnic rights. Many social, recreational, welfare and trade organisations also cater only to specific ethnic or religious group. This observation is consistent with Douglas and Pedersen’s (1973) conclusion that religion and ethnicity are drivers of Malaysian associational life.

Table 3 TSOs with ethnic terms in their names but not classified under ICNPO 7400
Full size table

Religion also characterises the third sector in Thailand (Pongsapich, 1997) and the Philippines (Carino and Staff, 2001). However, unlike in Thailand and the Philippines, majority of religious organisations in the Malaysian third sector are not from the dominant religion. This further reinforces the relevance of the government failure theory to the Malaysian third sector where non-median voters turn to the third sector for their religious needs. A small number of TSOs with religious names were not included in ICNPO 10100 or 10200 due to them not being involved in religious rituals, religious propagation or funeral assistance. Data show a striking characteristic of the Malaysian third sector where Muslims are more inclined than other religious groups to organise trade, professional, rights or residents' group along the lines of religion. In this case, the group that makes up 61.3% of the population is more inclined to turn to the third sector to protect and/or promote their commercial or social interests (Table 4).

Table 4 Organisations with religious term in their names but not classified in ICNPO 10100 or 10200
Full size table

Analysis of ethnic and religious fractionalisation on the distribution of Malaysian TSOs showed almost half of registered chambers of commerce are ethnic based, while six percent are religious-based. Youth associations, welfare bodies and residents’ associations were also found to have notable ethnic or religious bias. Table 5 shows the distribution of commercial, social and charitable organisations based on their ethnic and religious affiliation.

Table 5 Distribution of ethnic- and religious-based social, charitable and development organisations. Percent in parentheses
Full size table

Nevertheless, despite the anomaly observed in the mostly exclusive commercial and residents’ associations, fractionalisation in the Malaysian third sector could still be explained by the government’s focus on the median voter, who are also the majority ethnic and religious group. Weisbrod’s government failure theory explains Malaysia’s fractionalised nature of the third sector where focus on the median voters compels other ethnic and religious groups to rely on the third sector for their social, welfare and religious needs, establishing exclusive TSOs to cater to these needs.

Conclusions

Much focus on the measurement and categorisation of the third sector has been on more affluent, Western nations, where data are more easily available (Casey, 2016; Salamon & Anheier, 1997). Theory flows from data, and a large amount of third sector data is focussed in Western third sectors. Therefore, prevailing theory largely describes those sectors and more study of sectors in other contexts is essential. Our work contributes in a number of areas. First, we presented a machine-assisted framework to classify large amounts of unstructured data based on available information. Secondly, we discovered that member-serving organisations are evident in the Malaysian third sector, making criteria-guided exclusivity its distinguishing feature. Thirdly, we learnt that ethnic serving organisations are not well-captured by the Western-derived classifications.

Exclusivity and fractionalisation could lead to duplicate services where multiple TSOs serve different groups within the same community or geographical area. While cultural and/or religious differences may require identical services to be provided in different ways, duplication could also indicate that resources are not being efficiently employed. Exclusive TSOs could also cause certain segments of society not to get adequate support from the sector. While it complements the state’s provision of social welfare services, excessive reliance on the third sector by certain ethnic or religious group would indicate inequalities in policy development. Eliminating duplicate programmes or activities would enable TSOs to reallocate resources and increase their geographical coverage. Malaysian third sector actors could reflect on the findings by reviewing their operational model to evaluate if exclusivity maximises the efficiency and effectiveness of their activities and services. Policymakers should consider measures to address fractionalisation by developing policies that ensure the sector promotes inclusiveness and provides everyone in the community access to their services. Nevertheless, there are many TSOs which are inclusive, serving everyone regardless of background or affiliation and this map can help identify areas which are only served by exclusive TSOs, so they can extend their support to those communities.

Our map highlights some limitations of the predominant sector classifications in the Malaysian context. The significance of ethnic identities in the make-up and distribution of the TSOs is not well-captured by the Western-focussed classifications, but forms a significant determinant of the types of organisations and activity observed in Malaysia. While the ICNPO classification does allow for bespoke customisation to account for local differences, this risks losing the benefits of standardisation if repeated in different ways in different contexts. Nevertheless, we have demonstrated that the ICNPO classification has the potential to highlight activities and beneficiaries and also able to capture ethnic and religious differences in the make-up of the sector.

This paper uncovered the shape, scope and organisational size of the Malaysian third sector but not its economic and human resources size due to unavailability of financial, employee and volunteer data. It is proposed that further research on the human resource, finances and socio-economic contribution of the sector be undertaken to enhance these findings, which can also be used as basis for regional comparison and development of a southern theory of the third sector. A limitation of this paper is not being able to account for informal groups due to them being absent from any official database. They nevertheless are recognised as an increasingly integral component and form part of the conceptual definition of the third sector. Another shortcoming is the possibility of less accurate classification due to the reliance on organisation names as identifiers. To remedy this, it is hoped that the exercise to map the Malaysian third sector is continued with more support from data custodians.