Asia Pacific Journal of Education

ISSN: 0218-8791 (Print) 1742-6855 (Online) Journal homepage: https://www.tandfonline.com/loi/cape20

Cost efficiency – one size fits all? A university-

level analysis of economies of scale and scope in
Vietnamese higher education

Carolyn-Dung Thi Thanh Tran

To cite this article: Carolyn-Dung Thi Thanh Tran (2020): Cost efficiency – one size fits all? A
university-level analysis of economies of scale and scope in Vietnamese higher education, Asia
Pacific Journal of Education, DOI: 10.1080/02188791.2020.1735996

To link to this article: https://doi.org/10.1080/02188791.2020.1735996

Published online: 10 Mar 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=cape20
ASIA PACIFIC JOURNAL OF EDUCATION
https://doi.org/10.1080/02188791.2020.1735996

Cost efficiency – one size fits all? A university-level analysis of

economies of scale and scope in Vietnamese higher education
Carolyn-Dung Thi Thanh Tran
Postgraduate Department, International College of Management, Sydney, Australia; UNE Business School,
University of New England, Armidale, Australia

ABSTRACT ARTICLE HISTORY

Like other nations, Vietnam annually subsidies higher education via state Received 4 July 2019
budget to maintain academic operations of public institutions. However, Accepted 24 February 2020
the question of how to fairly and appropriately distribute state funding to KEYWORDS
public institutions is still challenging, therefore unfair distribution to some Cost efficiency; economies of
extent influence their cost efficiency. Using the quadratic cost function scale; economies of scope;
with random effects for a panel structure of 2011/12–2013/14, this paper public university; urban; rural
aims to estimate economies of scale and scope of public universities and
to seek for explaining whether the even application of standardized “one
size fits all” policy in distributing public budget works or not when the
cost analysis is stratified into rural and urban universities. The findings
show that public universities are cost-inefficient in their performance and
there are no economies of scale and scope in public universities. Even
distribution of delivering state funding is more likely to be ineffective for
public universities due to their different characteristics. Further, integrated
teaching at various levels of higher education could be advantageous for
rural universities but might not to be the case for their urban counterparts.
In the end, policy implications for ameliorating the performance of public
institutions are suggested.

Introduction
Higher education plays a crucial role in a nation’s education system by virtue of its contribution to
the workforce market and socio-economic development. Although governments continue to pro-
vide public funding to universities to cover teaching and research expenditure, they have an interest
on efficiency of all aspects of higher education, especially in the context of limited budget. As
a consequence, public higher education has faced pressure in improving their cost efficiency to
move their performances forward (Agasisti, Catalano, Landoni, & Verganti, 2012; Tran and Villano,
2017; Johnes & Johnes, 2016).
Viet Nam, a Southeast Asian nation, has witnessed an economic miracle during the last three
decades under its economic reform policy (“Doi moi”) starting from 1986. Vietnam is now one of the
stars of the emerging markets universe (Vanham, 2018). Vietnam’s economic growth is estimated on
average 6–7% in comparison with its neighbour China (International Monetary Fund, 2019; Vietnam
News, 2019a). This growth miracle of Vietnam did not go unnoticed in international ranking. After
a decade, the ranking of Vietnam rose from 77th place in 2006 to 55th in 2017 in World Economic
Forum’s Global Competitiveness Report and from 104th place in 2007 to 68th place in 2017 in the
World Bank’s ease of doing business (Vanham, 2018). In terms of such a rapid economic growth,

CONTACT Carolyn-Dung Thi Thanh Tran [email protected] International College of Management, Sydney 2095,
Australia
© 2020 National Institute of Education, Singapore
2 C.-D. T. T. TRAN

Vietnam has become a comfortably middle-income country from one of the poorest nations in the
world and then has been recognized as part of a group of economies having done particularly well at
making their growth process more inclusive and sustainable (Vanham, 2018; World Economic Forum,
2018). In 2020, Vietnam again takes over the Association of South East Asian Nations (ASEAN)
chairmanship with a heavy responsibility, an important role and a source of pride to the ASEAN
community (VietNam News, 2019b).
Whilst Vietnam’s economy has obtained remarkable achievements, its higher education seems
not to go in the same direction albeit the government has implemented several higher education
reform policies, namely the 2005 Higher Education Reform Agenda (the 2005 HERA), Decree 49/
2010/ND–CP of the government on tuition revenue, the 2012 Education Law, and Circular 38/2013/
TT–BGDDT of Ministry of Education and Training on the procedure and periods of quality assurance
of curricula of universities and colleges. Specifically, in the 2005 HERA the government has planned
that by 2020, at least one university of Vietnam is ranked in the world’s top 200 universities (Tran and
Villano, 2017; Pham, 2014). However, to date, no universities are ranked in the world’s top 1,000
universities based on well-known world university rankings (Temmerman, 2019). This would raise the
importance of investigating operational efficiency of Vietnamese universities to design more appro-
priate policies to improve their performance on the way reaching the world’ higher education
standards.
Like other countries, Vietnam annually subsidies the national education system via state budget
to maintain the performance of education institutions at different levels. For example, in 2016,
Vietnamese government subsidies approximately 20% of total national budget to the national
education system, in which around 10% was for public tertiary institutions for teaching (including
tuition fees) and research (Duc; Vietnam News, 2018; Viet, 2018). However, it is claimed that this
funding is not sufficient enough for academic operations of public institutions because (1) distribu-
tion mechanism is complicated and unclear, implying equalitarianism, or the even application of
standardized “one-size fits all” policy rather than concern about the quality of education; (2)
following Decree 86/2015/ND-CP on tuition fees of the government, tuition fees that public institu-
tions are allowed to charge students are still low, thus not enough to cover expenditures for the
performance of public institutions; and (3) autonomy in using the state budget is still within the
regulation framework, therefore lessens operational efficiency of public institutions (Duc; Finance
Magazine, 2017; Truong, 2019; Viet, 2018). These claims are indeed worthy of concern, but till date,
there is little evidence on cost efficiency of public universities and whether they have economies of
scale and scope in producing their academic outputs.
Regarding financial autonomy in higher education, a university-industry linkage model in public
universities has been recently encouraged in Asian universities to diversify sources of funding for their
performances (Intarakumnerd & Schiller, 2009; Kuriakose & Iyer, 2016a; Lee & Gopinathan, 2003). In the
study of Kuriakose and Iyer (2016b), the university spin-off has emerged in India; however, the linkage
between the academic and commercial sphere is not strong enough because of the institutional
incapacity in the intellectual property right (IPR) system that inhibits knowledge transfer and the
absence of an effective entrepreneurial system that encourages commercialization. In Vietnam, the
linkage between university and industry has been undertaken, but at the nascent level, with a main
focus on the support for students’ scholarships and internships, rather than emphasizing collaborative
activities in research and technology transfer (Hoc & Trong, 2019). Thus, public universities in Vietnam
heavily rely on the government funding for maintaining their academic activities.
The government in some Asian countries, namely Hong Kong and Singapore, have recently been
supportive of privatization of publicly funded universities to reduce the financial burden imposed on
the government and encouraging building a world class university under private foreign collabora-
tion (Lee & Gopinathan, 2003). In Thailand, although public universities’ budget is most supported by
the government, but they are expected to yield more income from other sources via conducting
research and development activities that are relevant to the industry (Intarakumnerd & Schiller,
2009). On the other hand, a recent study of Hoang (2019) showed that public universities in Vietnam
 ASIA PACIFIC JOURNAL OF EDUCATION 3

made a way to increase their non-state budget via part-time enrolments. However, the part-time
enrolments could not increase without limitation, because in principle, annual enrolments are
confined by the regulations regarding the ratio of students to teaching staff and academic area
per student (Tran and Villano, 2018b).
Research on the performance of Vietnamese public universities have been conducted in recent
years such as Nguyen, Thenet, and Nguyen (2015), Tran and Villano (,2017, 2018a). However, these
studies only focus on the generic performance of public universities without indicating whether they
are cost-efficient and have economies of scale and scope in the process of producing different types
of outputs. Therefore, the comparative (dis)advantages of various outputs (e.g., postgraduate,
undergraduate, and associate undergraduate) have been not perused thoroughly to provide insights
to educational managers to improve their performance. Our paper aims to address this problem by
examining scale and scope economies of public universities including cost efficiency index for the
period of 2011/12-13/14. Our analysis is also stratified into rural and urban public universities to
investigate whether the “one-size fits all” policy works in Vietnam or not.
The organization of this paper is as follows. Section 2 presents a brief review on economies of
scale and scope in higher education. Section 3 introduces a synoptic description of public univer-
sities in Vietnam. The methodology is shown in Section 4, followed by data, variables, and the
empirical model of analysis in Section 5. Section 6 demonstrates empirical results and discussion.
Section 7 ends with the conclusion and policy implications.

Cost efficiency and economies of scale and scope in higher education: a brief review
Numerous studies on economies of scale and cost efficiency have been well documented in
literature of higher education, especially in advanced countries. Cost efficiency is defined as the
ratio of minimum cost to observed costs for each decision-marking unit or individual firm (Coelli, Rao,
O’Donnell, & Battese, 2005, p. 184). Research on cost efficiency, economies of scale and scope have
been widely implemented in higher education across nations, namely, UK, US, Australia, Italy, Turkey,
Japan, Taiwan and among others. Studies in UK showed that in general there were economies of
scale and scope in higher education in the UK (Izadi, Johnes, Oskrochi, & Crouchley, 2002; Koshal &
Koshal, 1995, 1999, 2000; Lenton, 2008), and the results vary according to the selection of estimation
method and different periods (Izadi et al., 2002; Johnes, 1997; Johnes, Johnes, & Thanassoulis, 2008;
Thanassoulis, Kortelainen, Johnes, & Johnes, 2011). In the same vein, American higher education
were found to have economies of scale (de Groot, McMahon, & Volkwein, 1991; Dundar & Lewis,
1995; Sav, 2004, 2011) and economies of scope (Agasisti & Johnes, 2015; Sav, 2011) at the different
levels. In Australia, universities were significantly found to have economies of scope in under-
graduate teaching, distance education and for overseas students (Worthington & Higgs, 2011;
Zhang & Worthington, 2017; Zhang, Worthington, & Hu, 2017) and when economies of scope
exist, their magnitude is small (Lloyd, Morgan, & Williams, 1993). As for European countries, it is
found that Italian universities have exhausted scale and scope economies, and experienced dis-
economies according to their size (Agasisti, 2016; Agasisti & Dal Bianco, 2007; Agasisti & Johnes,
2010). On the other hand, Lewis and Dundar (1995) revealed that Turkish universities can obtain cost
efficiency if they expand their teaching and research output and that economies of scale can differ by
the group of field and level of instruction. However, on the other hand, just a few studies were
conducted in Asian countries. For example, economies of scale were found in Japanese universities
(Hashimoto and Cohn, 1997; Nemoto & Furumatsu, 2014). However, Taiwanese higher education was
subject to diseconomies of scale (Fu, Huang, & Yang, 2011). Therefore, there is still room for studies in
other Asian nations to contribute diverse contexts to the cost literature of higher education.
Some methods of analysis commonly used to estimate economies of scale and scope among
these studies are constant elasticity of substitution, hybrid translog, and quadratic cost functions
(Zhang & Worthington, 2018). In addition, the ownership (private versus public institutions) was used
in the analysis of cost structure by decomposing the sample into two types (e.g., Chavas, Barham,
4 C.-D. T. T. TRAN

Foltz, & Kim, 2012; Cohn, Rhine, & Santos, 1989; Laband & Lentz, 2003; Sav, 2004) or using dummy
variable to differentiate the sample (Johnes et al., 2008; Johnes & Salas-Velasco, 2007). This variable
was of interest to investigate whether the funding of public institutions restricts their enrolments
relative to private institutions (Zhang & Worthington, 2018). Moreover, the managerial efficiency is
an important factor for consideration in cost structure of institutions using stochastic frontier such as
Izadi et al. (2002), Maripani (2007), Johnes and Johnes (2009, 2016).
Teaching outputs play a crucial role in estimating scale and scope economies in the literature of
higher education. The most common choice of teaching outputs was enrolments in recent studies
because of their data availability. On the other hand, the completions can be considered as an output
in the estimation of scale and scope economies (Agasisti, 2016; Worthington & Higgs, 2011; Zhang &
Worthington, 2017). Research output (publications, funding income, etc) are added in the model to
capture changes in cost pattern of universities (Johnes & Johnes, 2016; Zhang & Worthington, 2017).
However, there was no agreement on the choice of outputs in analysing the performance of
higher education due to specific objectives of research contexts (Tran and Villano 2017, 2018b,
Zhang & Worthington, 2018).
Finally, most studies in the literature used a panel dataset to estimate cost pattern with fixed
effects, that is assuming that characteristics of firms are constant across years according to specific
research context (Johnes & Johnes, 2016; Zhang & Worthington, 2017). However, very few studies
aim to test random effects that allow firms to change their features over the surveyed period. This
would be an open avenue for our study. Using the random effect model, our paper can capture
changes in characteristics of universities over time, thus generate more robust findings in estimating
cost efficiency of universities. In addition, testing separate institutions in different locations, namely
urban and rural areas, should also be an interesting factor for consideration. Our paper aims to
bridge these gaps in the literature by applying the cost function approach in a specific context of
Vietnamese public universities.

State universities in Vietnam

The Vietnamese higher education system consists of numerous universities and colleges. Among 454
higher education institutions (HEIs), universities account for 52% including 170 public and 65 private
institutions. Table 1 shows that the number of public universities increase considerably by 70% for the
period of 2005/06–2017/18, nearly three times higher than the number of private institutions. The
enrolments of public universities contribute approximately 85% of total university enrolments. This
implies the important role of the public university sector in the Vietnamese national education system.
Unlike private universities who use their own investment capitals for their academic activities,
public institutions are granted government funding annually to maintain their academic operations.
This funding is provided to public institutions on the basis of the difference between estimated total
budget approved for their yearly annual expenditures and real total turnover obtained from tuition

Table 1. Universities, enrolments and completions over years.

2005/06 2015/16 2016/17 2017/18
Institutions 125 223 235 235
Public 100 163 170 170
Private 25 60 65 65
Enrolments 1,046,291 1,753,174 1,767,879 1,707,025
Public 933,352 1,520,807 1,523,904 1,439,495
Private 112,939 232,367 243,975 267,530
Share of total enrolment
Public 89.2% 86.7% 86.2% 84.3%
Private 10.8% 13.3% 13.8% 15.7%
Completions 143,017 352,789 305,601 320,578
Source: MOET (2019).
 ASIA PACIFIC JOURNAL OF EDUCATION 5

fees of students (Joint Circular 20/20141). That is, if state universities earn more revenues from tuition
fees due to an increase in the number of students, the government will grant less funds to them, and
vice versa (Tran and Villano, 2018a). However, such a distribution mechanism is imputed to be
unclear and ineffective because this mechanism much depends on the number of enrolments and
emphasizes even distribution in public funding without considering education quality (Truong,
2019). This inevitably leads to less incentives among public universities to ameliorate their cost
efficiency (Vietnam News, 2018). In addition, innovation in the tuition policy which has been
conducted following Decree 49/2010/ND-CP for 2010/11-2014/15 and Decree 86/2015/ND-CP for
2015/16-2020/212 has still faced some challenges. For example, tuition fees are still low, no differ-
ence in tuition fees are given to different fields of education; therefore, public universities could not
cover real costs used for teaching and research (Hayden & Lam, 2015; Truong, 2019; Vietnam News,
2018). As argued by Kent (2005), the government budget is a useful incentive for policy undertaking
in higher education. However, this incentive needs to be evaluated to see if it can help in improving
the performance of public universities. Recent research of Tran and Villano (2018a) showed that
Vietnamese public universities were financially inefficient enough in the surveyed period and their
financial efficiency contributed quite weakly to the overall performance of public institutions.
However, authors did not show whether public universities have economies of scale and scope in
their production process, thus an open question for further studies. Our paper comes to bridge this
gap by investigating scale and scope economies on teaching outputs and by decomposing public
universities into rural and urban groups to examine the difference in cost structure of these groups.

Methodology
This section presents the specification and method of analysis for estimating cost economies of scale
and scope for universities defined as decision-making firms (DMF) producing multiple academic
outputs (Cohn & Cooper, 2004; Zhang & Worthington, 2017). Assume that Y is a vector of r outputs
(y1, . . . yr) produced and X is a vector of s inputs (x1, . . ., xs) used by each DMF, given P, a vector of
p input prices (p1, . . ., ps) in its production process. The cost function for producing multiple outputs
is expressed as follows:
TC ðY; pÞ ¼ minfp:xjðx; yÞ 2 ψg (1)
X0

where TC ðY; pÞ depicts total cost used to produce Y outputs with input prices P. The equation on the
right-hand side demonstrates the condition for this cost function. Specifically, the cost function
should be minimized in combination of producing r types of output Y, given the input prices (p)
Ps 
under the multiple product technology set ψ; p:x is the internal product i¼1 pi xi added into the
observed total cost.
In this paper, we use four categories of cost economies that can be useful for universities’
decision-making on the allocation of outputs and output mix (Johnes & Johnes, 2016; Zhang &
Worthington, 2017). The formulae to compute these types are presented as follows:

(a) The ray economies of scale (RES)

The ray (overall) economies of scale are defined as the behaviour of costs when the quantities of the
products are increased proportionally (Hashimoto and Cohn, 1997, p. 108). In other words, “the ray
economies of scale determine cost advantages or disadvantages as output size is altered while
maintaining a constant output mix” (Sav, 2004, p. 608). From Equation (1), the ray economies of scale
(RES) can be estimated as
TCðy; pÞ
RES ¼ Pr @TC
(2)
i¼1 yi @yi
6 C.-D. T. T. TRAN

where TCðy; pÞ is the total cost used to produce all different outputs, given the mean prices; @TC
@yi is
the marginal cost of producing output yi . Ray economies of scales illustrate the effect of
a simultaneously proportional rise of all output scale along a ray in the output space, keeping the
structure of each university’s outputs unchanged (Zhang & Worthington, 2017). If RES > 1, we have
ray economies of scale, meaning that an increase in the output bundles leads to a decrease in total
costs. Otherwise, we have diseconomies of scale (Johnes & Johnes, 2016).

(b) The product-individual economies of scale (PIES)

The product-individual economies of scale refer to the level that a university’s operation costs could
vary by changing the quantity of one output, holding the quantities of other outputs constant (Cohn
& Cooper, 2004). Using Equation (1), we can estimate the product-individual economies of scale as
below

AICðyi Þ
PIES ¼ @TC
(3)
@yi

where AIC ðyi Þ is the average incremental cost of output i, measured by AICðyi Þ ¼
½TC ðyn Þ  TC ðyni Þ=yi , TC ðyn Þ is the total cost of producing all the outputs at the levels in vector Y;
TCðyni Þ is the total cost of producing all outputs at the levels in vector Y except for output i which is
set to zero; i = 1, . . ., n. If the PIES > 1, then there are product-individual economies of scale for output
i. Otherwise, no economies of scale are present. The estimates in (a) and (b) give universities
information to make decision how large a scale at which they should operate. The following
estimates in (c) and (d) further show the scope of universities’ operations.

(c) The global economies of scope (GES)

If cost efficiencies can be gained by joint production of multiple products rather than producing each
separately, economies of scope are registered, that is global and/or product-specific economies of
scope (Dundar & Lewis, 1995; Hashimoto and Cohn, 1997). In this paper, the global economies of
scope (GES) are measured to indicate the scope of operations of a university. The GES is referred to as
the proportion of rise in total costs from specialized production:
Pn
TC ðyi Þ  TC ðy; pÞ
GES ¼ i¼1
(4)
TC ðy; pÞ
P
where ni¼1 TC ðyi Þ is the cost of producing output i in n separate universities, and TCðy; pÞ is the cost
of producing all n outputs jointly within a university. If GES > 0, there are global economies of scope
from producing all the outputs together rather than each one in a separate university. Otherwise, we
have diseconomies of scope.

(d) The product-single economies of scope (PSES)

The product-single economies of scope (PSES) refer to cost advantages due to production of each
product jointly with the other outputs (Hashimoto and Cohn, 1997; Koshal & Koshal, 1999) and are
estimated based on the following formula:

½TCðyi Þ þ TCðyni Þ  TCðy; pÞ

PSES ¼ (5)
TC ðy; pÞ

where TC ðyi Þ is the cost of producing output i in a separate university at the same level as in the
output vector y. If PSES > 0 (PSES < 0), there are product-single economies (diseconomies) of scope
for output i, meaning that there are cost-savings (cost-dissavings) from producing this product along
with all of the others.
 ASIA PACIFIC JOURNAL OF EDUCATION 7

The abovementioned four measures provide universities and policy makers with facts of the
current cost pattern that support academic operations of universities including average incremental
costs, marginal cost, economies of scale and scope and cost efficiency index obtaining from their
activities. Until now, estimating economies of scale and scope in higher education has been
conducted in several different countries, but not in developing countries like Vietnam. In addition,
differentiation in geographical and institutional characteristics would affect the current cost struc-
ture that appears to be neglected in previous studies, and thus this motivates us to fill this gap in the
literature of cost efficiency in higher education.

Data, variables and the empirical model specification

Data and variables
In the Vietnamese higher education system, all universities are operating under the same education law
but under different line management systems, depending on the ministry or local authorities who they
belong to. On average 13% of the higher education institutions (HEIs) are under Ministry of Education
and Training whereas the remaining HEIs belong to 13 different ministries and local authorities (Tran
and Villano, 2017). Although public universities are probably located in different regions across the
country, they are classified as the same type in nature in terms of institutional structure and funding
category that they are supported by the government. In this paper, we classify public universities into
two groups, urban and rural, based on their geographical locations. Urban universities refer to
universities located in the large cities of Vietnam, namely Hanoi City, Hochiminh City and Danang
City whereas rural universities are universities located outside these large cities. The sample of public
universities taken is of the same institutional type, consisting of 82 public universities in Vietnam for the
period of 2011/12–2013/14. These data were extracted from the Ministry of Education and Training
(MOET) and the website of universities. The data report of MOET has technically provided generic
figures of the whole sector, not for individual institution, whereas individual reports of institutions
publicly posted in their websites are often insufficient for all institutions or not continuous enough for
multiple periods. Whilst collecting panel data for all public universities seems challenging,
a representative sample of 82 public universities who are well following regulations of MOET to publish
their annual reports on their websites for leaners’ reference are chosen for analysis in this paper. In
a statistical analysis, a sample of 82 public universities for a period of three years, equivalent to 246
observations, are more than enough to produce robust results in the non-linear regression models.
The inputs and outputs used in this paper follow a production function approach to modelling
university behaviour. That is to say that universities utilize both labour and non-labour factors to
produce multiple outputs in their production process. The outputs used in this paper comprise
postgraduate, undergraduate and associate undergraduate completions. It is noted that postgraduate
completions include master and PhD completions, and associate undergraduate completions are
graduates who completed the 12 + 3 bachelor degree (at the college level). The college degree training
has gradually been lessening in universities so that they only focus on their main tasks to educate and
train the undergraduate and postgraduate degrees. The research output (e.g., publication, conference
papers, book chapters, etc.) has not been surveyed due to shortage of available data. In addition, the
quality variables of teaching and learning have not been included because the single report of
universities has not been well reported these indicators. Further studies may add these details in future.
The main costs of university production process are the wages of academic and non-academic staff.
However, unfortunately, the data do not allow to separate academic staff for teaching and research and
nor to decompose non-academic staff into management, teaching and research-related support
services. Other inputs include non-labour expenditures that cover the physical support of academic
activities (e.g., teaching, research, learning facilities, other related support services). In Vietnam, the
non-labour expenditures are calculated based on the floor area for academic spaces, that is, universities
invest capitals in the academic spaces with full of learning, teaching and research facilities that ensure
8 C.-D. T. T. TRAN

the teaching and research process to run smoothly. For public universities, after the salaries for staff, the
rest of expenditures is extended for this target.
Accordingly, the input price of academic (non-academic) staff for a single university is determined
by dividing the expenditure on academic wages by the number of full-time equivalent academic
(non-academic) staff. The price of capital (non-labour) is proxied by dividing non-labour expenses by
the floor area for academic spaces. The total cost of each university every year is the sum of all labour
and non-labour expenditures.
Table 2 presents a statistics summary for annual costs, input prices and outputs for three years
period. On average, public universities spent 71 billion VND (equivalent to 3.07 million USD at the
2019 exchange rate). However, there was a vast variation within the surveyed sample depending on
the student numbers and campus size. As can be observed, the number of postgraduate and
undergraduate completions had a decreasing trend whereas the figure of associate undergraduate
completions witnessed an increase over years. Regarding input prices, it is interesting to see that the
academic price is less than its non-academic price counterpart. This can be explained by the fact that
the non-academic price includes the salaries of the different levels of management that may be high
relative to teaching staff’s salaries in the Vietnamese university context. This result is contrary to the
result found in Worthington and Higgs (2011) in respect of Australian universities in which the
academic price was greater than the non-academic price in Australian universities.
As indicated above, our analysis is limited to the locality of universities, i.e.,, urban or rural. The
findings from Table 2 show that the figures of total costs, outputs and input prices are higher in
urban universities than in rural universities. This implies that there is a difference in the cost structure
of two kinds of universities in Vietnam.

Empirical model specification

In this section, we present the empirical model specification based on the generic cost function in
Equation (1) and variables expounded in Section 5.1. The principle assumption is that universities
seek to minimize their costs at the existing level of outputs subject to the relative prices of the factor
inputs (Worthington & Higgs, 2011). While alternative methods to research on costs of universities
comprise parametric and non-parametric cost frontier as indicated in review papers of Worthington
(2001) and Johnes (1996), a quadratic cost function is used in this paper. The advantage of the
quadratic cost function is a flexible specification which can be applied to multiple factor production.
This cost function also allows us to take into account of the linear, quadratic, and cross-product terms
for multiple factor inputs (Worthington & Higgs, 2011). Accordingly, the economies of scale can vary
corresponding with various levels of input prices and outputs.
The quadratic cost function can be expressed as follows:
X3 X3 X3
TCðy; pÞ ¼ β0 þ i¼1
βi yi þ 1=2 i¼1
βii yi2 þ β yy
i;j¼1 ij i j
þ εi (6)
iÞj

where β0 is the intercept, βi are the slope coefficients of the linear term, βii are the slope coefficients
of the quadratic term, βij are the slope coefficient of the cross-product term (iÞj) and TC is the total
operating costs of each university.
For the linear analysis, fixed or random effects can be used to capture heterogeneity of uni-
versities over time. In the Vietnamese context, we use the linear random effects for a panel structure
to reflect different characteristics in managing the cost structure of public universities which receive
the government funding under the same framework.
Following Coelli et al. (2005), the factor demands can be derived by differentiating the cost
function in Equation (6) with respect to the input prices, specifically
Table 2. Summary statistics of inputs and outputs (n = 82).
2011/12 2012/13 2013/14
Unit Mean Standard deviation Mean Standard deviation Mean Standard deviation
The entire industry
Total costs Billion VND 71 73 89 86 100 92
Postgraduate completions Student 129 292 122 333 90 363
Undergraduate completions Student 1667 1723 1646 1753 1494 1277
Associate undergraduate completions Students 308 670 308 556 1238 2525
Academic price Billion VND 0.088 0.061 0.104 0.051 0.113 0.054
Non-academic price Billion VND 0.139 0.095 0.169 0.108 0.183 0.114
Other price Billion VND 0.744 0.800 0.672 0.463 0.702 0.540
Urban
Total costs Billion VND 82 84 102 97 115 104
Postgraduate completions Student 181 340 187 404 105 423
Undergraduate completions Student 1852 1712 1799 1825 1620 1305
Associate undergraduate completions Students 216 704 217 608 1114 3078
Academic price Billion VND 0.099 0.071 0.116 0.058 0.124 0.059
Non-academic price Billion VND 0.150 0.097 0.189 0.122 0.206 0.126
Other price Billion VND 0.902 0.924 0.810 0.503 0.829 0.588
Rural
Total costs Billion VND 53 43 71 56 75 60
Postgraduate completions Student 39 134 37 42 64 228
Undergraduate completions Student 1347 1695 1327 1618 1275 1219
Associate undergraduate completions Students 468 569 960 415 1454 1032
Academic price Billion VND 0.068 0.027 0.090 0.028 0.094 0.039
Non-academic price Billion VND 0.121 0.087 0.138 0.067 0.142 0.075
Other price Billion VND 0.472 0.366 0.458 0.247 0.483 0.358
ASIA PACIFIC JOURNAL OF EDUCATION
9
10 C.-D. T. T. TRAN

@TC @TC pi x i pi
¼ ¼ ¼ FCSðpi Þ or
@pi @pi TC TC
@TC X3 (7)
FCSðpi Þ ¼ ¼ βi þ βii yi þ βy
i;j¼1 ij j
@pi
iÞj

where FCSðpi Þ is the ith factor cost share and xi is the quantity of the ith input and pi is the price of the
ith input (academic, non-academic labour and capital). The efficiency of the parameter estimates can
be improved by estimating Equation (6) with the cost share Equation (7). It is noted that the cost
share for the three input prices must be equal to one. The system of equations is estimated jointly by
omitting one of the cost share equation in (7). In this case, the price of academic staff is the omitted
input price.
Equation (6) can be rewritten under some transformation by taking logarithms including the
parameters of the input prices in Equation (7) as
0 1
B X 3 1 X 3 X 3 C X3
lnðTC Þ ¼ ln@β0 þ βy þ
i¼1 i i
βii yi2 þ i;j¼1
βij yi yj A þ γ lnðwi Þ þ εi
i¼1 i
(8)
2 i¼1
iÞj

From Equation (8), the cost-efficiency index of individual universities can be estimated by taking
difference between predicted costs and real costs. If this index is greater than 1, a single university is
not efficient in using their expenditures. Otherwise, they are efficient in cost utilization, meaning that
real costs are less than theoretical costs.

Empirical findings
The cost function estimates for the entire industry are presented in Table 3. This model shows
a goodness of fit with χ2 = 68.9 which is statistically significant at the 1% level of significance. The
rural and urban models are statistically significant at the 1% level of significance in terms of χ2 = 60.4
and 40, respectively. It is observed that although the urban model has less significant-estimated
coefficients than the entire industry and rural model, its goodness of fit χ2 = 40 is statistically
significant. This implies that the model specification is appropriate and other factors could poten-
tially influence expenditures of universities located in the urban areas. Further studies should
consider this to add more variables with data availability in the model to explain more variations

Table 3. Cost function estimates for public universities, 2011/12–2013/14.

The whole industry Urban Rural
Coefficient Standard error Coefficient Standard error Coefficient Standard error
β1 −0.0389 0.1024 −0.136 0.124 0.943*** 0.264
β2 −0.0362 0.1460 0.483 0.607 −0.322** 0.137
β3 −0.0466 0.0829 0.047 0.118 −0.074 0.124
β11 0.0823*** 0.0251 0.098*** 0.033 0.075** 0.035
β22 0.0648*** 0.0243 −0.020 0.093 0.135*** 0.022
β33 0.0362*** 0.0134 0.030 0.021 0.038** 0.016
β12 −0.0281** 0.0144 −0.022 0.018 −0.146*** 0.035
β13 0.0035 0.0052 0.001 0.007 0.007 0.009
β23 −0.0097 0.0099 −0.017 0.018 −0.006 0.012
γ1 0.7624 - 0.744 - 0.811 -
γ2 0.338*** 0.119 0.370** 0.155 0.114 0.151
γ3 −0.101 0.0742 −0.114 0.101 0.074 0.095
β0 2.825*** 0.546 1.386 1.972 2.718*** 0.624
Obsa 246 156 90
χ2 68.9*** 41*** 60.4***
a
refers to the number of observations for three years, *** and ** denote the levels of significance at 1% and 5%.
 ASIA PACIFIC JOURNAL OF EDUCATION 11

in expenditures of urban universities. Based on estimated χ2 values, the estimated models are valid
and can be used for predicted estimates.
The coefficients of quadratic terms have significantly positive signs showing that total costs of
universities decrease when the number of completions increases. The cost compensation is sig-
nificantly found between postgraduate and undergraduate completions. That is, cost for postgrad-
uate and undergraduate education can be lessened in combination of teaching. This is opposite to
the results found in the study of Worthington and Higgs (2011) that was the cost complementarity of
undergraduate and postgraduate education. In Australia, the qualification of faculty in teaching is
highly demanding. The Australian Higher Education Standards Framework 2015 (Tertiary Education
Quality and Standards Agency [TEQSA], 2019) indicates that academic teaching staff must be
qualified to at least one level of qualification (Australian Qualification Framework (AQF) level or
equivalent) higher than the course of study being taught, for example, a faculty who teaches the
postgraduate programs needs to have a doctoral degree, thus would cost more for universities. In
addition, learning and teaching facilities and services in Australia often cost more for the postgrad-
uate programs, given the quality of facilities and services. By contrast, these factors seem not to be
critical in the case of Vietnam where the quality of learning and teaching is on the way of higher
education reform. In addition, the price elasticity of non-academics (γ2 = 0.388) contributes sig-
nificantly to an increase in the total costs of universities.
To gain further analysis, we stratified the entire industry into rural and urban models. These
models fit very well with χ2 values being statistically significant at the 1% level of significance. The
findings in Table 3 reveal that in the urban model, the elasticity coefficient of the second term is
significantly positive (p < 0.01), meaning that the cost of urban universities reduces at the increasing
degree when the number of postgraduate completions goes up. Similar to the entire industry model,
the price elasticity of non-academics (γ2 = 0.37) is significantly positive at the 5% level of significance.
However, this coefficient is not significant in the rural model. This implies that the price for manage-
ment and professional staff plays a crucial role in the cost structure of urban universities rather than
that of their rural counterparts.
In a closer view, the elasticity coefficient of postgraduate (β1 = 0.943) is significantly positive in the
rural model, suggesting that this output makes total costs increase at the rising level (β11 = 0.075).
However, the undergraduate coefficient (β2 = −0.322) gives an opposite effect on total cost at the
increasing degree (β22 = 0.135) at the 1% level of significance. It is also found that there is the cost
compensation between the undergraduate and associate undergraduate completions (β12 = −0.146)
at the 5% level of significance. That is, the integrated training of undergraduate and the lower level
can be cost saving for universities.
The quadratic cost function estimates in Table 3 are used to estimate average incremental cost
(AIC) and marginal costs (MC) for each of the three university products (postgraduate, undergraduate
and associate undergraduate completions) for the levels of mean output ranging from 50% to 200%
(e.g., 100% is equal to the mean output; 200% is double of the mean output). The mean values of
three outputs are 129 postgraduate completions, 1,667 undergraduate completions and 308 associ-
ate undergraduate completions. Table 4 presents the results of AIC for the whole system, urban and
rural universities. We can see that for the whole industry, the AIC per undergraduate completion has
a decreasing trend when the output mean increases over 100% and that the AIC per undergraduate
completion is the highest at 100% of the mean output. However, the AIC per associate under-
graduate completion has an opposite effect, an increasing trend is going on until the mean output
reaches to the level of 200%, then it starts declining. Interestingly, there is no increase in the AIC per
postgraduate completion.
Having a closer look at the urban and rural models, the AIC per undergraduate completion has the
same pattern as happening in the entire industry. Specifically, the AICs per undergraduate and
associate undergraduate completions increase significantly with an increase in the mean output and
saving costs if they can increase postgraduate completions. However, the rural analysis witnesses an
opposite cost pattern. The AICs per postgraduate and associate undergraduate completions increase
12 C.-D. T. T. TRAN

Table 4. Average incremental costs (AIC) simulated by the mean of outputs.

Postgraduate Undergraduate Associate undergraduate
Mean completions completions completions
The entire industry
0.5 −1,073,330 2,449,047 582,590
0.75 −1,386,119 2,881,970 705,162
1 −1,591,168 3,029,682 779,373
1.25 −1,712,394 2,996,943 816,761
1.5 −1,769,146 2,855,160 826,518
1.75 −1,777,009 2,652,441 815,934
2 −1,748,485 2,420,685 790,757
Urban
0.5 −777,741 2,580,306 351,284
0.75 −1,519,600 4,039,770 678,640
1 −2,639,213 5,751,895 1,165,400
1.25 −4,297,280 7,838,518 1,876,241
1.5 −6,717,181 10,445,438 2,899,882
1.75 −10,208,195 13,752,699 4,357,568
2 −15,197,039 17,987,086 6,414,457
Rural
0.5 778,842 −384,445 660,223
0.75 928,055 −464,535 776,407
1 982,983 −498,970 811,642
1.25 976,089 −502,490 795,500
1.5 930,473 −485,820 748,541
1.75 862,351 −456,681 684,834
2 782,906 −420,553 613,803
Unit measurement: VND deflated at the price of 2011/12.

Table 5. Marginal costs simulated by the mean of outputs.

Postgraduate Undergraduate Associate undergraduate
Mean completions completions completions
The entire industry
0.5 8,792,171 887,673 1,695,919
0.75 8,961,780 1,078,001 1,777,378
1 9,134,661 1,309,139 1,862,749
1.25 9,310,876 1,589,835 1,952,222
1.5 9,490,491 1,930,715 2,045,991
1.75 9,673,571 2,344,685 2,144,265
2 9,860,182 2,847,415 2,247,259
Urban
0.5 6,243,613 768,086 2,057,493
0.75 6,624,751 700,972 2,070,147
1 7,029,155 639,723 2,082,878
1.25 7,458,245 583,825 2,095,688
1.5 7,913,529 532,811 2,108,577
1.75 8,396,606 486,255 2,121,544
2 8,909,171 443,767 2,134,592
Rural
0.5 47,275,394 1,137,799 1,406,129
0.75 39,407,653 1,646,620 1,543,705
1 32,849,289 2,382,983 1,694,740
1.25 27,382,392 3,448,647 1,860,553
1.5 22,825,316 4,990,872 2,042,589
1.75 19,026,645 7,222,776 2,242,435
2 15,860,162 10,452,781 2,461,834
Unit measurement: VND deflated at the price of 2011/12.

with the mean outputs whereas the AIC per undergraduate completion is cost saving, around
VND498,970 (around $US223) per completion at the 100% of the mean output.
Regarding the marginal cost of outputs, Table 5 indicates that for the whole industry, the
marginal cost of all outputs increases over the different levels of output. For example, the MC of
 ASIA PACIFIC JOURNAL OF EDUCATION 13

undergraduate completion rises from VND887,673 to VND2,847,415 (around $US40 to $US129).

Meanwhile, the MC of postgraduate completion increases from VND 8,792,171 to VND9,860,182
(around $US399 to $US450), being ten times higher than the MC of each undergraduate completion.
This result is in accordant with the study of Worthington and Higgs (2011) in respect of Australian
universities in which the marginal cost of postgraduate completions increased with the outputs. In
addition, Lewis and Dundar (1995) found that PhD completions are the costliest output at the margin
for universities whereas undergraduate completions are least costly.
Our further findings show that there is a difference in the MC of postgraduate and undergraduate
completions between rural and urban universities. The MC of postgraduate completion rises slightly
in urban universities over various levels of the mean outputs whereas the MC of postgraduate
completion declines considerably in rural universities. It is noted that the MC of postgraduate
completion in rural universities is much greater than the MC of postgraduate completion in their
urban counterparts. For example, at the 100% level of the mean output, rural universities spent
VND32,849,289 (around $US1,416) for each postgraduate completion whereas urban universities
spent VND7,029,155 (around $US303) for each postgraduate completion.
However, on the other hand, the MC of undergraduate completions in urban universities has
a decreasing trend while the MC of undergraduate completion in rural universities increases over
different levels of outputs. Interestingly, a slight rise in the MC of associate undergraduate comple-
tion is present in rural universities whereas it seems as if there is not much change in the MC of
associate undergraduate completion in urban universities. From these results, it can be observed
that output completions in rural universities are costlier than those in urban universities. This clearly
depends on the student inputs that affect the output completions afterwards.
The ray and product-individual economies of scales are shown in Table 6. As defined earlier, if the
point estimate is greater than 1, then ray or product-individual economies of scales exist across the
output set. Otherwise we do not have economies of scales. The findings from Table 6 reveal that
there are no ray and product-individual economies of scale in the surveyed sample. This suggests
that the whole system is currently experiencing diseconomies of scales, and that expanding outputs
would not generate efficiency for universities in the reported period, keeping the quality of

Table 6. Economies of scales simulated by the mean of outputs.

Postgraduate Undergraduate Associate undergraduate
Mean completions completions completions RAY
The entire industry
0.5 −0.0011 0.0017 0.0006 0.0041
0.75 −0.0014 0.0017 0.0006 0.0030
1 −0.0015 0.0014 0.0007 0.0022
1.25 −0.0016 0.0012 0.0007 0.0016
1.5 −0.0016 0.0009 0.0007 0.0012
1.75 −0.0016 0.0007 0.0006 0.0008
2 −0.0016 0.0005 0.0006 0.0006
Urban
0.5 −0.0008 0.0019 0.0003 0.0022
0.75 −0.0015 0.0033 0.0006 0.0030
1 −0.0024 0.0051 0.0011 0.0042
1.25 −0.0037 0.0076 0.0017 0.0058
1.5 −0.0054 0.0112 0.0027 0.0079
1.75 −0.0077 0.0161 0.0040 0.0108
2 −0.0108 0.0231 0.0058 0.0148
Rural
0.5 0.00044 −0.00025 0.00059 0.00321
0.75 0.00062 −0.00021 0.00063 0.00192
1 0.00079 −0.00016 0.00060 0.00113
1.25 0.00094 −0.00011 0.00054 0.00065
1.5 0.00108 −0.00007 0.00046 0.00037
1.75 0.00120 −0.00005 0.00038 0.00021
2 0.00131 −0.00003 0.00031 0.00012
14 C.-D. T. T. TRAN

education constant. This result is in line with the findings of Tran and Villano (2018b) in which the
overall efficiency of public universities decreased in 2013/14. These authors showed that during
the year 2013 the government suggested universities to decrease the number of new enrolments in
some specific fields of education such as business and education because of a high unemployment
rate of these fields in the workforce market. The economies of scales are also not found in the rural
and urban models. As noted by Worthington and Higgs (2011), diseconomies of scales can coincide
with increasing returns to scale4 since increases in factor prices have compensated any physical
gains.
Like almost university system in the world, the academic production of universities in Vietnam
involves in joint training of different levels of programs, thus it is unlikely to separate costs of
production precisely into their single uses. Table 7 indicates that the global economies of scope of
the entire system are negative, meaning that public universities have no economies of scope from
producing all the outputs together. This result provides useful information for university managers to
seek better solutions to enhance their economies of scope.
To be more specific, in the rural model, the economies of scope for individual outputs show that
there are cost-savings from producing this output with all the others. However, in the urban model,
the economies of scope of postgraduate completions reveals that producing this output with other
outputs can generate the cost-savings. However, this ability is not applicable for associate under-
graduate completions with the mean output over 100%, and not applicable for undergraduate
completions with the mean output over 50%.
In the final analysis, we look at the cost efficiency index of public universities over the sample period. As
indicated earlier, cost efficiency index greater (less) than 1 shows that actual costs are used more (less)
than the theoretical/expected costs. That is to say that the university is cost-inefficient (-efficient) in terms
of the industry benchmark if the index value is greater (less) than unity. It is noted that cost efficiency
comprises technical and allocative efficiency. From Table 8, it can be seen that the whole sector is not cost-
efficient for the period of three years. Overall, 41 public universities achieve cost efficiency (i.e., their actual
costs are less than the theoretical costs), accounting for 50% of total public universities in the sample.

Table 7. Economies of scope stimulated by the mean of outputs.

Postgraduate Undergraduate Associate undergraduate
Mean completions completions completions Global
The entire industry
0.5 0.527 0.444 0.506 −1.189
0.75 0.387 0.231 0.351 −1.572
1 0.288 0.050 0.235 −1.865
1.25 0.218 −0.104 0.147 −2.090
1.5 0.169 −0.235 0.079 −2.264
1.75 0.135 −0.346 0.027 −2.400
2 0.112 −0.440 −0.015 −2.507
Urban
0.5 0.254 0.062 0.220 −1.782
0.75 0.142 −0.182 0.082 −2.132
1 0.092 −0.352 0.008 −2.342
1.25 0.073 −0.475 −0.036 −2.478
1.5 0.069 −0.567 −0.064 −2.573
1.75 0.072 −0.638 −0.084 −2.644
2 0.078 −0.695 −0.100 −2.700
Rural
0.5 0.604 0.628 0.585 −1.179
0.75 0.468 0.516 0.431 −1.581
1 0.362 0.440 0.303 −1.893
1.25 0.279 0.389 0.196 −2.136
1.5 0.214 0.359 0.106 −2.326
1.75 0.163 0.344 0.030 −2.473
2 0.123 0.341 −0.035 −2.588
 ASIA PACIFIC JOURNAL OF EDUCATION 15

Table 8. Cost efficiency index over years 2011/12–2013/14.

2011/12 2012/13 2013/14 Overall
The entire industry (n = 82)
Mean 1.096 1.025 1.039 1.053
Standard deviation 0.614 0.217 0.295 0.412
Min 0.131 0.589 0.251 0.131
Max 5.035 1.555 2.928 5.035
χ2 test a 0.21
Efficient universities 41
Urban (n = 52)
Mean 1.090 1.074 1.036 1.067
Standard deviation 0.501 0.577 0.289 0.469
Min 0.132 0.264 0.282 0.132
Max 3.740 4.810 2.626 4.810
χ2 test 0.38
Efficient universities 29
Rural (n = 30)
Mean 1.018 1.025 1.021 1.021
Standard deviation 0.192 0.228 0.206 0.207
Min 0.631 0.593 0.473 0.473
Max 1.434 1.539 1.608 1.608
χ2 test 0.058
Efficient universities 14
Kruskal–Wallis equality-of-populations rank test.

Figure 1. Kernel density distribution of cost efficiency index for the entire industry over 2011/12–2013/14.
16 C.-D. T. T. TRAN

To gain further analysis, we classify the whole sector into rural and urban universities. The findings
indicate that both types of universities are not cost-efficient in the reported period. The cost
efficiency of the urban group slightly declines over three years albeit this decline is not significant.
On the other hand, cost efficiency of the rural group minimally varies across three years. The overall
findings in Table 8 show that approximately 56% of urban universities and 47% of rural universities
are cost-efficient. Figure 1 demonstrates the kernel distribution of the overall cost efficiency of the
whole sector over three years. As we can see, the distribution of cost efficiency varies across three
years but better fits normal distribution with the average of cost efficiency for these years.
To have a clearer picture on whether location, where universities are located affects the cost
efficiency of the entire sector, we conduct the Spearman’s correlation test between the overall cost
efficiency index, the urban cost efficiency index and the rural cost efficiency index. The result from
Table 9 shows that the urban cost efficiency index is significantly and positively correlated to the
overall cost efficiency index at 0.521. On the other hand, the overall cost efficiency is negatively and
significantly related to the rural cost efficiency at a lower level of 0.215. This suggests that cost
efficiency of urban universities has driven cost efficiency of the whole system.
Figure 2 represents the matrix cost efficiency indices which support this evidence, indicating that
the distribution of the urban cost efficiency is densely and positively correlated to the cost efficiency
of the whole sector whereas the distribution of the rural cost efficiency is closely and negatively
related to the overall cost efficiency.

Table 9. Spearman correlation test among cost efficiency indices (n = 82).

Overall cost efficiency index (OCEI) Urban cost efficiency index (UCEI) Rural cost efficiency index (RCEI)
OCEI 1.000
UCEI 0.521*** 1.000
RCEI –0.215** –0.064 1.000
***denotes at the 1% level of significance.

Figure 2. Scatter plot matrix among cost efficiency indices of the entire industry, urban and rural regions.
 ASIA PACIFIC JOURNAL OF EDUCATION 17

Discussion
Public universities in Vietnam are facing challenges to ensure their income adequately for academic
operations. Increasing the number of enrolments is one of the common ways to increase turnover,
for example, increasing part-time enrolments (Hoang, 2019). However, this would lead to concern
about education quality (Truong, 2019). The empirical results in this paper show that the more the
students, the greater the marginal cost for extra students in the reported years. In addition, urban
universities have greater marginal costs than their rural counterparts. Therefore, public universities
need to be aware of this to have an appropriate plan for annual fresh enrolments.
Moreover, the whole system, and each type of rural and urban universities do not have economies
of scale and scope in their operations. This implies that increasing the quantities of students does not
help in reducing the total costs of teaching activities and even integrating different types of teaching
outputs is not advantageous for public universities. Together with this, public universities do not
illustrate cost efficiency, implying that they need to reduce their current expenditures to obtain the
full efficiency. This is in line with the conclusion of Tran and Villano 2018a) that public universities are
in need of improving their financial efficiency. Accordingly, regardless of complications in the system
of the state budget distribution (Finance Magazine, 2017; Vietnam News, 2018), our results would
suggest that public universities first need to reconsider their ability of using input resources to
reduce total annual expenditures, then profit could potentially be obtained based on their input
savings.
On the top of that, we would argue that the “one size fits all” policy of the government bounds to
fail in the process of distributing the state budget to public universities as recently claimed in public
discussion in Vietnam (Duc; Finance Magazine, 2017; Truong, 2019; Viet, 2018). This is because the
costs for training an additional student in rural regions is much more expensive than the costs for
training an extra student in urban regions. Further, integrating the teaching process at different
levels of education, namely postgraduate, undergraduate and associate undergraduate contribute
significantly to cost efficiency of rural universities but not to cost efficiency of urban institutions.
Clearly, geographical location plays a dominant role in assessing cost efficiency of public institutions.
Last but not least, cost inefficiency cannot be imputed to managerial inefficiency of single univer-
sities, instead the tuition fees policy (e.g., Decree 186/2015/ND-CP) needs to be reconsidered
because the low level of tuition fees makes revenues of institutions not likely to be increased at
the same pace to the rising level of expenditure. In other words, a better strategy for public funding
distribution, e.g., increasing the autonomy level of managing institutional expenditure should be
appropriately redesigned for the development of the sector.

Conclusion
This paper investigated cost efficiency, scale and scope economies of public universities using the
quadratic cost function. The further analysis was conducted by stratifying the sample into rural and
urban universities to examine whether there was a difference in cost efficiency, scale and scope
economies between two university groups or not. By using this method of analysis, our contributions
to the literature are twofold. In the first place, we investigated economies of scales and scopes of
public universities in Vietnam that have not been conducted before. Secondly, we examined
whether the “one size fits all” policy can be applicable to public universities via the geographical
location. This provides a novel finding that would help policy makers to have better strategies of the
state funding delivery for the sustainable development of public universities.
Our empirical results show that public universities do not have economies of scales and scopes and
marginal cost for extra students would increase together with a rise in enrolment numbers, in which
rural universities spend more costs than their urban counterparts to produce the same output, keeping
the quality of education constant. In addition to this, they are not cost-efficient in the surveyed years.
These findings led us to noteworthy conclusions that (1) the even application of standardized “one size
18 C.-D. T. T. TRAN

fits all” policy in distributing the state funding to universities in different regions is not effective, thus
need is for reconsideration of a better strategy; (2) individual universities need to ameliorate their
performance to more efficiently use their expenditure; and (3) the tuition fees policy should also be
revisited to assist public universities to increase revenues for the aim of maintaining their performance.
Our empirical research has provided interesting results regarding cost efficiency and economies
of scale and scope of public universities to policy makers and institutional managers. It is important
that policymakers should possess a sound knowledge of specific characteristics of the different sub-
categories of public universities which are located either in urban or rural area while formulating
policy. Moreover, policy design must pay attention to the specific constraints faced by different kinds
of public universities (e.g., rural versus rural) and avoid a ‘one-size-fits-all’ approach. This will better
assist public universities in achieving their “best-practice” in education service provision.
Our findings have partly addressed recent concern of the public and stakeholders about the
government policy and institutional performance. However, the importance of some factors should
be highlighted for further consideration. Specifically, adding private universities in the cost analysis is
preferable to provide cost efficiency of the entire higher education sector. In addition, some external
factors should be tested in the cost function to capture their influences on the cost structure of
institutions, namely the committed leadership, the quality of academics’ intellectual engagement
(Eldridge, Larry, Baird, & Kavanamur, 2018; Oleksiyenko & Ros, 2019). Finally, research outputs should
be investigated in length and breadth to provide more insights of the performance of institutions.

Notes
1. Joint Circular 20/2014/TTLT–BGDDT–BTC–BLDTBXH on implementation of the tuition fee policy.
2. Decree 49/2010/ND-CP on tuition fees for higher education from 2010/11 to 2014/15 and Decree 86/2015/ND-
CP on tuition fees for higher education from 2015/16 to 2020/21.
3. The exchange rate at 24/3/2019, $US1 = VND23,205.
4. Increasing returns to scale occurs when output increases by a larger proportion than the increase in inputs
during the production process.

Disclosure Statement
No potential conflict of interest was reported by the author.

Notes on contributor
Carolyn-Dung Thi Thanh Tran is Postgraduate Program Manager (Business) and Director of Scholarship of Learning and
Teaching at International College of Management, Sydney, New South Wales, Australia. She is also an Adjunct Research
Fellow in Business School, University of New England. She has written extensively on higher education and local
government, especially on operational efficiency, municipality economies of scale and productivity. Recent publications
consist of Local Government Studies, Public Administration Quarterly, Australian Journal of Public Administration,
Public Finance and Management, Applied Economics, Asian Economic Journal, and International Transaction in
Operational Journal.

ORCID
Carolyn-Dung Thi Thanh Tran http://orcid.org/0000-0002-5798-0543

References
Agasisti, T. (2016). Cost structure, productivity and efficiency of the Italian public higher education industry 2001–2011.
International Review of Applied Economics, 30(1), 48–68.
Agasisti, T., Catalano, G., Landoni, P., & Verganti, R. (2012). Evaluating the performance of academic departments: An
analysis of research-related output efficiency. Research Evaluation, 21, 2–14.
 ASIA PACIFIC JOURNAL OF EDUCATION 19

Agasisti, T., & Dal Bianco, A. (2007). Cost structure of Italian public universities: An empirical analysis. Higher Education in
Europe, 32(2–3), 261–275.
Agasisti, T., & Johnes, G. (2010). Heterogeneity and the evaluation of efficiency: The case of Italian universities. Applied
Economics, 42(11), 1365–1375.
Agasisti, T., & Johnes, G. (2015). Efficiency, costs, rankings and heterogeneity: The case of US higher education. Studies in
Higher Education, 40(1), 60–82.
Chavas, J., Barham, B., Foltz, J., & Kim, K. (2012). Analysis and decomposition of scope economies: RandD at US research
universities. Applied Economics, 44(10–12), 1387–1404.
Coelli, T.J., Rao, D.S.P., O’Donnell, C.J., & Battese, G.E. (2005). An introduction to efficiency and productivity analysis. USA:
Springer Science and Business Media.
Cohn, E., & Cooper, S.T. (2004). Multi-product cost functions for universities: Economies of scale and scope. In G. Johnes & J. Johnes
(Eds.), International handbook on the economics of education (pp. 579–612). Cheltenham, UK: Edward Elgar Publishing Ltd.
Cohn, E., Rhine, S., & Santos, M. (1989). Institutions of higher education as multi-product firms: Economies of scale and
scope. The Review of Economics and Statistics, 71(2), 284–290.
de Groot, H., McMahon, W.W., & Volkwein, J.F. (1991). The cost structure of American research universities. The Review of
Economics and Statistics, 73, 424–431.
Dundar, H., & Lewis, D.R. (1995). Departmental productivity in American universities: Economies of scale and scope.
Economics of Education Review, 14, 119–144.
Eldridge, K., Larry, L., Baird, J., & Kavanamur, D. (2018). A collaborative governance approach to improving tertiary
education in Papua New Guinea. Asia Pacific Journal of Education, 38(1), 78–90.
Finance Magazine (2017). Giải pháp tự chủ tài chính tại các trường đại học công lập [Solutions for financial autonomy in
Vietnamese public universities]. Retrieved from http://tapchitaichinh.vn/tai-chinh-kinh-doanh/giai-phap-tu-chu-tai-
chinh-tai-cac-truong-dai-hoc-cong-lap-116662.html
Fu, T.T., Huang, C.J., & Yang, Y.L. (2011). Quality and economies of scale in higher education: A semiparametric smooth
coefficient estimation. Contemporary Economic Policy, 29(1), 138–149.
Hashimoto, K., & Cohn, E. (1997). Economies of scale and scope in Japanese private universities. Education Economics, 5
(2), 107–115.
Hayden, M., & Lam, Q.T. (2015). A 2020 vision for higher education in Vietnam. International Higher Education, 44, 11–13.
Hoang, L. (2019). Twin privatization in Vietnam higher education: The emergence of private higher education and partial
privatization of public universities. Higher Education Policy, 32(3), 359–380.
Hoc, L.H., & Trong, N.D. (2019). University–industry linkages in promoting technology transfer: A study of Vietnamese
technical and engineering universities. Science, Technology and Society, 24(1), 73–100.
Intarakumnerd, P., & Schiller, D. (2009). University-industry linkages in Thailand: Successes, failures, and lessons learned
for other developing countries. Seoul Journal of Economics, 22(4), 551–589.
International Monetary Fund. (2019). Five charts explain Vietnam’s economic outlook Retrieved from https://www.imf.
org/en/News/Articles/2019/07/11/na071619-five-charts-explain-vietnams-economic-outlook
Izadi, H., Johnes, G., Oskrochi, R., & Crouchley, R. (2002). Stochastic frontier estimation of a CES cost function: The case of
higher education in Britain. Economics of Education Review, 21(1), 63–71.
Johnes, G. (1997). Costs and industrial structure in contemporary British higher education. The Economic Journal, 107,
727–737.
Johnes, G., & Johnes, J. (2009). Higher education institutions’ costs and efficiency: Taking the decomposition a further
step. Economics of Education Review, 28(1), 107–113.
Johnes, G., & Johnes, J. (2016). Costs, efficiency, and economies of scale and scope in the English higher education
sector. Oxford Review of Economic Policy, 32(4), 596–614.
Johnes, G., Johnes, J., & Thanassoulis, E. (2008). An analysis of costs in institutions of higher education in England. Studies
in Higher Education, 33, 527–549.
Johnes, G., & Salas-Velasco, M. (2007). The determinants of costs and efficiencies where producers are heterogeneous:
The case of Spanish universities. Economics Bulletin, 4(15), 1–9.
Johnes, J. (1996). Performance assessment in higher education in Britain. European Journal of Operational Research, 89, 18–33.
Kent, R. (2005). The changing role of the state in Mexican higher education: From the crisis of ineffectual populism to
new forms of system coordination. In A. Gornitzka, M. Kogan, & A. Amaral (Eds.), Chapter 11, Reform and changes in
higher education (pp. 187–205). Dordrecht, the Netherlands: Springer.
Koshal, R.K., & Koshal, M. (1995). Quality and economies of scale in higher education. Applied Economics, 27(8), 773–778.
Koshal, R.K., & Koshal, M. (1999). Economies of scale and scope in higher education: A case of comprehensive
universities. Economics of Education Review, 18(2), 269–277.
Koshal, R.K., & Koshal, M. (2000). Do liberal arts colleges exhibit economies of scale and scope? Education Economics, 8(3),
209–220.
Kuriakose, F., & Iyer, D.K. (2016a). India of ideas: Mapping the status of higher education in India and mobilizing
discourse towards a quest for equity and excellence. Higher Education for the Future, 3(2), 213–226.
Kuriakose, F., & Iyer, D.K. (2016b). Exploring university-industry technology transfer in India: Two models. Journal of
Research Innovation and Management Science, 2(4), 56–62.
20 C.-D. T. T. TRAN

Laband, D.N., & Lentz, B.F. (2003). New estimates of economies of scale and scope in higher education. Southern
Economic Journal, 70(1), 172–183.
Lee, M.H., & Gopinathan, S. (2003). Reforming university education in Hong Kong and Singapore. Higher Education
Research & Development, 22(2), 167–182.
Lenton, P. (2008). The cost structure of higher education in further education colleges in England. Economics of
Education Review, 27(4), 471–482.
Lewis, D.R., & Dundar, H. (1995). Economies of scale and scope in Turkish universities. Education Economics, 3(2),
133–157.
Lloyd, P.J., Morgan, M.H., & Williams, R.A. (1993). Amalgamations of universities: Are there economics of scale or scope?
Applied Economics, 25(8), 1081–1092.
Maripani, J.F. (2007). Efficiency in higher education: An analysis of the “traditional universities” in Chile (Unpublished
doctoral dissertation). University of Connecticut, Ann Arbor, MI.
Nemoto, J., & Furumatsu, N. (2014). Scale and scope economies of Japanese private universities revisited with an input
distance function approach. Journal of Productivity Analysis, 41(2), 213–226.
Nguyen, T.T.H., Thenet, G., & Nguyen, K.M. (2015). Applying DEA sensitivity analysis to efficiency measurement of
Vietnamese universities. Management Science Letters, 5, 983–992.
Oleksiyenko, A., & Ros, V. (2019). Cambodian lecturers’ pursuit of academic excellence: Expectations vs. reality. Asia
Pacific Journal of Education, 39(2), 222–236.
Pham, H. (2014). Aiming for at least one world-class university by 2020. University World News, 304, 9–11.
Sav, G.T. (2004). Higher education costs and scale and scope economies. Applied Economics, 36(6), 607–614.
Sav, G.T. (2011). Panel data estimates of public higher education scale and scope economies. Atlantic Economic Journal,
39(2), 143–153.
Temmerman, N. (2019). Transforming higher education in Vietnam. University World News. Retrieved from https://www.
universityworldnews.com/post.php?story=20190129142655883
Tertiary Education Quality and Standards Agency. (TEQSA, 2019). Higher education standards framework 2015 Retrieved
from https://www.teqsa.gov.au/higher-education-standards-framework-2015
Thanassoulis, E., Kortelainen, M., Johnes, G., & Johnes, J. (2011). Costs and efficiency of higher education institutions in
England: A DEA analysis. Journal of the Operational Research Society, 62(7), 1282–1297.
Tran, C. D T, & Villano, R. A. (2018a). Measuring efficiency of Vietnamese public colleges: an application of the DEA-based
dynamic network approach. International Transactions in Operational Research, 25(2), 683–703.
Tran, C-D.T.T, & Villano, R.A. (2017). An empirical analysis of the academic performance : the case of Vietnamese higher
education institutions. Journal of Further and Higher Education,41(4), 530–544.
Tran, C-D.T.T, & Villano, R.A. (2018b). Financial efficiencies of Vietnamese public universities: a second-stage dynamic
network data envelopment analysis approach. The Singapore Economic Review. doi: 10.1142/S0217590818500133
Truong, T.Q. (2019). Tự chủ trong các trường đại học công lập ở Việt Nam: Thách thức và giải pháp [Financial autonomy in
Vietnamese public universities: Challenges and solutions]. Retrieved from https://baomoi.com/tu-chu-trong-cac-
truong-dai-hoc-cong-lap-o-viet-nam-thach-thuc-va-giai-phap/c/29594845.epi
Vanham, P. (2018). The story of Vietnam’s economic miracle. World Economic Forum. Retrieved from https://www.
weforum.org/agenda/2018/09/how-vietnam-became-an-economic-miracle/
Viet, D. (2018). Can doi moi mo hinh phan bo ngan sach cho giao duc dai hoc [Need to innovate the mechanism of state
budget for higher education]. Retrieved from http://thoibaotaichinhvietnam.vn/pages/nhip-song-tai-chinh/2018-12-
15/can-doi-moi-mo-hinh-phan-bo-ngan-sach-cho-giao-duc-dai-hoc-65487.aspx
Vietnam News. (2018). Universities should enjoy more autonomy. Retrieved from https://vietnamnews.vn/society/
464097/universities-should-enjoy-more-autonomy.html#T2AwudwuAVbSLy5y.97
Vietnam News. (2019a). Vietnam institute for economic and policy research forecasts Vietnam’s 2019 growth at
7.05 per cent. Retrieved from https://vietnamnews.vn/economy/536739/vepr-forecasts-vns-2019-growth-at-705-per-
cent.html#j1TsHLjD6usxj63L.97
VietNam News. (2019b). Vietnam takes over the ASEAN chairmanship in November. Retrieved from https://vietnamnews.
vn/politics-laws/537128/viet-nam-to-take-over-asean-chairmanship-in-november.html#vy3BmRcAxexU6D6o.97
World Economic Forum. (2018). The inclusive development index: Summary and data highlights. Retrieved from http://
www3.weforum.org/docs/WEF_Forum_IncGrwth_2018.pdf
Worthington, A. (2001). An empirical survey of frontier efficiency measurement techniques in education. Education
Economics, 9(3), 245–268.
Worthington, A.C., & Higgs, H. (2011). Economies of scale and scope in Australian higher education. Higher Education, 61
(4), 387–414.
Zhang, L.C., & Worthington, A.C. (2017). Scale and scope economies of distance education in Australian universities.
Studies in Higher Education, 42(9), 1785–1799.
Zhang, L.C., & Worthington, A.C. (2018). Explaining estimated economies of scale and scope in higher education: A
meta-regression analysis. Research in Higher Education, 59(2), 156–173.
Zhang, L.C., Worthington, A.C., & Hu, M. (2017). Cost economies in the provision of higher education for international
students: Australian evidence. Higher Education, 74(4), 717–734.