Economies of scale in the Korean district heating system: A variable cost function approach

Energy Policy 88 (2016) 197–203

Contents lists available at ScienceDirect

Energy Policy journal homepage: www.elsevier.com/locate/enpol

Economies of scale in the Korean district heating system: A variable cost function approach Sun-Young Park a,1, Kyoung-Sil Lee b,2, Seung-Hoon Yoo b,n a

Economic Research Institute, Korea University, 145 Anam-Ro, Seongbuk-Gu, Seoul 02841, Republic of Korea Department of Energy Policy, Graduate School of Energy & Environment, Seoul National University of Science & Technology, 232 Gongneung-Ro, NowonGu, Seoul 01811, Republic of Korea

b

H I G H L I G H T S








We examine economies of scale in the South Korean district heating sector. We focus on Korea District Heating Corporation (KDHC), a public utility. We estimate a translog cost function, using a variable cost function. We found economies of scale to be present and statistically significant. KDHC will enjoy cost efficiency and expanding its supply is socially efficient.

art ic l e i nf o

a b s t r a c t

Article history: Received 6 July 2015 Received in revised form 18 October 2015 Accepted 19 October 2015

This paper aims to investigate the cost efficiency of South Korea’s district heating (DH) system by using a variable cost function and cost-share equation. We employ a seemingly unrelated regression model, with quarterly time-series data from the Korea District Heating Corporation (KDHC)—a public utility that covers about 59% of the DH system market in South Korea—over the 1987–2011 period. The explanatory variables are price of labor, price of material, capital cost, and production level. The results indicate that economies of scale are present and statistically significant. Thus, expansion of its DH business would allow KDHC to obtain substantial economies of scale. According to our forecasts vis-à-vis scale economies, the KDHC will enjoy cost efficiency for some time yet. To ensure a socially efficient supply of DH, it is recommended that the KDHC expand its business proactively. With regard to informing policy or regulations, our empirical results could play a significant role in decision-making processes. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Scale economies District heating Variable cost function

1. Introduction While experiencing two energy crises in the 1970s, the South Korean government began to recognize that fundamental energysaving measures were required in the housing, industrial, and commercial sectors. As part of establishing those energy-saving measures, the government promoted district heating (DH), which uses high-efficiency equipment that leverage combined heat and power (CHP). The use of CHP equipment can improve energy use

Abbreviations: CHP, Combined heat and power; DH, District heating; DHS, District heating system; IHS, Individual heating system; KDHC, Korea District Heat Corporation; SCE, Scale economies; SUR, Seemingly unrelated regression n Corresponding author. Fax: þ 82 2 970 6800. E-mail addresses: [email protected] (S.-Y. Park), [email protected] (K.-S. Lee), [email protected] (S.-H. Yoo). 1 Fax: þ82 2 3290 2535. 2 Fax: þ82 2 970 6800. http://dx.doi.org/10.1016/j.enpol.2015.10.026 0301-4215/& 2015 Elsevier Ltd. All rights reserved.

efficiency by producing heat and power simultaneously. DH through CHP can use low-priced renewable energy, including waste heat from industrial processes and heat from waste incineration—other renewable energy sources include wood chips, landfill gases, solar heat, and sewage. DH has a number of advantages. First, the availability of a variety of energy sources for CHP facilitates reduction of oil dependency and better use of renewable energy. Second, the use of DH reduces production costs by using waste energy. Third, DH has been evaluated as a clean and eco-friendly energy form and the energy sources used in CHP help save fossil fuels and reduce air pollutants. Thus, the application of DH improves air quality. By utilizing CHP, DH produces only 39% of the air pollutants (e.g., nitrogen oxides, sulfur oxides, and dust) that individual heating systems (IHSs) do (e.g., Haichao et al., 2013; Ilic and Trygg, 2014). Fourth, DH can help resolve the problem of global warming, since DH can reduce greenhouse gas emissions in South Korea by 49.2%, compared to IHSs (Korean Ministry of Trade, Industry & Energy,

198

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

2014) Finally, CHP plants are usually sited in closer proximity to consumers than existing heating systems (e.g., thermal power plants, hydropower plants, and nuclear plants); this results in reduced transmission losses and lower transmission investment costs. For these environmental and economic reasons, the South Korean government has been interested in expanding the use of DH and has promoted the introduction of DH to local residents. Existing apartments have been retrofitted to use DH systems (DHSs), thus replacing IHSs; additionally, newly planned cities have constructed new DHS plants. South Korean residents prefer DHSs to IHSs for three reasons. First, the price rates of the former are lower than those of the latter. Second, the former does not require the use of an individual boiler. Finally, the average value of houses in DHS-serviced areas is higher than that of houses in IHSserviced areas, all else being equal. In summary, DHSs are expected to proliferate throughout South Korea to meet residents’ increasing demand for them, and there is a national plan in place to build more CHP plants. The DH industry is a network industry that requires pipelines to deliver energy to users. Some public goods suppliers—the so called public utilities in the economic literature such as electricity, railway, natural gas, water, and sewage companies—are also categorized as network industries. Generally, network firms construct delivery systems to connect plants to users. Large investments in the construction of plants and pipelines are required in the early stages of a project; this investment-based barrier can foster a natural or legal monopoly. The average cost of a monopolistic company reduces to low levels over time, on account of those investments initially made in the facilities. The monopolistic firm usually retains its monopolistic position and increases its production scale while enjoying profits. Although it seems efficient to have a single large firm generate DH, the disadvantages inherent in a monopoly could lead to several problems for energy users. For example, the average cost of the firm increases rather than decreases at some threshold scale; at that threshold value, production is not socially efficient. If a monopolistic firm produces DH in excess of that scale, granting permission to market entrants and dividing the monopolistic firm into several smaller firms would result in cost reductions and efficiency enhancements. In such a scenario, government officials may insist upon introducing a competition system to the DH industry as a part of liberalization. The issues of privatization and competition have arisen among network industries in South Korea. Indeed, in 1999, a monopolistic power generation public utility was divided into six generation companies, based on empirical results that concluded that returns to scale were decreasing. Additionally, a competition scheme was introduced to the country’s natural gas supply industry, whereupon the government allowed large-scale private consumers to import liquefied natural gas directly. Additionally, a national railway company was embroiled in a privatization controversy in 2013. As to the current study’s focus, however, the South Korean government established KDHC in 1985 to expand DHSs nationwide, while focusing specifically on new satellite cities in metropolitan areas. KDHC owns 12 CHP plants and buys heat from three CHP plants owned by other generation companies, and recently, private DHS suppliers have emerged in the market. In 2013, 10,895,352 Gcal was supplied to 1,248,846 households by KDHC, who enjoyed a 59% market share in the DHS market; thus, the DH industry was suspected of being monopolistic.3 Reducing the 3 There are 35 suppliers in the market of 2015 and about 70% of suppliers suffer from their deficit. Their total deficit is KRW 34,219 million (Korea Energy Agency, 2015). The second biggest supplier, GS power co., Ltd, has 12.7% of the market and SH corporation, which has a 8.8% market share, is in the third place.

government’s equity share of KDHC and/or strictly limiting the expansion of KDHC into the DH market have been considered. What is important in introducing a competitive system or denationalizing a network industry is whether doing so will increase, reduce, or maintain returns to scale. As noted earlier, network industries require large investments at their onset, and they usually experience a decreasing average cost for some time afterwards. Identifying and using a social efficiency scale would assist in removing the social loss that can occur when supplying resources or services. When an industry experiences decreasing returns to scale, the government would allow new entrants into its market. On the other hand, increasing returns to scale would imply that the presence of multiple suppliers would give rise to socially inefficient conditions; in such a case, the merging of firms or expansion into the supply area would be recommended, to create one or several firms that have healthy economies of scale. The objective of this study is to generate empirical data and determine the implications of economies of scale in KDHC. A competition issue always arises in any industry in which there appears to be a monopoly. The aforementioned industries have often been called network industries, all of which require initial infrastructure investments; like other network firms and industries, DH suppliers should have power plants and a network of pipelines in place. Given the nature of these essential facilities, DH has conventionally been considered a naturally monopolistic industry: these facilities are very large and expensive, and so only very large companies can afford them. Now would be an appropriate time to undertake an economic analysis of economies of scale for KDHC in South Korea: doing so could help inform policymakers as they make decisions vis-à-vis energy plans or policy restrictions. This paper is organized into four distinct parts. Section 2 provides a literature review and details the estimation model, variable cost function, and variable definitions used in this study. Section 3 presents our estimation results and our predictions of future scale economies; it also discusses policy implications with regard to KDHC, a South Korean DH utility, and makes suggestions for its efficient management. Section 4 concludes the paper.

2. Method 2.1. Literature review Many researchers have focused on economies of scale in a variety of industries, including electricity, water, railway transportation, telecommunications, airlines, and public transportation. The starting point is the study by Christensen and Greene (1976), which addresses the cost function of U.S. electric power generation by using ordinary least squares. The most popular focus in this body of literature has been on water utilities: since 2000, more than 40 papers have been published with respect to water and sewage (Bottasso and Conti, 2008; Kim and Lee, 1998; Martins et al., 2006; Renzetti, 1999; Saal et al., 2007; see a review in Saal et al., 2013). The railway transportation industry is an interesting research area with regard to privatization and liberalization (Caves et al., 1981; Ida and Suda, 2004; Loizides and Tsionas, 2002). Many studies focus on the telecommunications industry (Charnes et al., 1988; Evans and Heckman, 1988), the airline industry (Berry, 1992; Oum et al., 1993) and on public transport, including bus and metro transportation (Cambini et al., 2011; Di Giacomo and Ottoz, 2010; Farsi et al., 2007; Ottoz and Di Giacomo, 2012). However, little research has been done on economies of scale in the DH industry. With regard to cost, Sjödin and Henning (2004) calculated the marginal costs of a DH utility. While a few papers address cost efficiency by using data envelopment analysis, only

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

Wibe (2001) concludes that Swedish DH production does not remain cost-efficient when economies of scale exist.4 It could be true that for Swedish production at that time, some DH supplies were quite restricted. Other than Wibe (2001), some other studies have examined the efficiency of DHSs and undertaken environmental analyses (Agrell and Bogetoft, 2005; Lazzarin and Noro, 2006; Lygnerud and Peltola-Ojala, 2010; Munksgaard et al., 2005; Torchio et al., 2009). The current study employs a seemingly unrelated regression (SUR) model (Zellner, 1962) as most studies have used it. The data we apply to estimate the cost function usually causes a high probability of error in the variables; at the same time, the error term in the cost function would be serial-correlated. This method estimates multivariate regression equations simultaneously— namely, a cost-function equation and its cost-share equations. By adopting a SUR model, we derive efficient estimates and an additional degree of freedom by adding cost-share equations rather than solely estimating the cost function.5,6

ln VC = α0 +

199

∑ αi ln Pi + 0.5∑ ∑ αuj ln Pi ln Pj + βQ i

i

+ 0.5βQQ (ln Q )2 +

∑ βQ i ln Q

ln Pi + γK ln K

i

+ 0.5γKK (ln K )2 +

∑ γKi ln K ln Pi + γKQ

+ 0.5δ TT (T )2 +

∑ δ TiT ln Pi + δ TKT ln K + δ TQ T ln Q i

VC (PL, PM ; K , Q ) = Min(PL⋅L + PM⋅M )subject to Q = f (L, M , K ) L, M

(1)

where PL and PM are the prices of labor and materials, respectively. As mentioned, the translog variable cost function is applied in our analysis. The empirical formula is expressed in Eq. (2). 4 Since Wibe (2001) was written in Swedish, we obtained the main idea of Wibe (2001) from Söderholm and Wårell (2011). 5 By adding share function, we can corporate that the error terms are assumed to be correlated across the equations, which leads to increase degree of freedom. This method, called SUR, makes it possible to obtain more efficient estimator than least squares (e.g., Russell and MacKinnon, 1993, p. 306; Sparks, 2004, p. 17). 6 An anonymous referee suggested an application of a dynamic model to our data. We think SUR model is better than any dynamic model in that a company decides its level of production with regard to input prices at each time and crosssectional analysis explains company’s behavior well. 7 In our analysis, capital is treated as quasi-fixed input as is common in the literature. The assumption that fixed cost is given does not mean that capital level is fixed over time. Compared with the inputs such as labor and material, capital is not immediately determined. For example, it takes three to five years to construct a CHP plant. That is why we treat capital given and employ a variable cost function. 8 The variable cost function could be affected by other factors such as the size of plant, a governmental policy promoting DH use, etc. For instance, the size of plant was included in capital level. The effect from a governmental policy could be contained technical change, because technical change could be recognized as time effect or event effect that happened at that time. We made a model simplified and choose essential factors among many other factors that affect the cost.

(2)

where T denotes technical change. We impose several restrictions, such as homogeneity of degree one in input prices and symmetry.

∑ αi = 1, ∑ αij=∑ αij=0, ∑ βQ i = 0, ∑ δ Ti = 0, ∑ γKi = 0. i

i

j

i

i

i

(3)

The cost-share equation within this equation is obtained by differentiating the log of the variable cost by the log of the input prices. Its functional form is as follows.

∑ αij ln Pj + βQ

ln Q + γK ln K + δ TiT

j

A variable cost function is necessary due to a quasi-fixed factor to produce DH. As noted above, DH suppliers need large initial capital investments; they could not minimize their cost with respect to all factors including capital. A quasi-fixed factor was decided with a long run plan and expectation. This decision process does not satisfy a minimization of total costs. A variable cost at each quasi-fixed input level explains producers' decision.7 For example, a decision maker increases or decreases production level by adjusting the levels of inputs at a given production capacity. As other literatures, we applied variable cost function for analyzing economies of scale in DH (Bottasso and Conti, 2008; McGeehan, 1993). We select a translog cost function to calculate the economies of scale in DH with electricity. KDHC is assumed to make use of capital ( K ), labor ( L ), and materials ( M ) in its operations.8 KDHC is willing to minimize its variable cost at a given production level ( Q ), based on the fixed capital level. The variable cost takes the form:

ln K ln Q + δ T T

i

∂ ln VC /∂ ln Pi = αi + 2.2. Variable cost function

ln Q

j

i

(4)

where i = L, M . Note that the use of Eq. (3) would give rise to a singularity problem. To resolve this, we drop an arbitrary cost-share equation. Indirect parameter estimates in the omitted cost-share equation— which is to say, the material equation—could be obtained from the calculation derived from Eq. (3). We define the cost efficiency as the percentage change in the variable cost when the output is increased by 1%, holding all input prices and capital fixed. Thus, this can be expressed (Nelson, 1989, p. 282):

ECQ = (∂ ln VC /∂ ln Q )/(1 − ∂ ln VC /∂ ln K ).

(5)

In this study, we use an index that represents economies of scale. Degree of scale economies (SCE) is derived by subtracting the cost elasticity from one. It can be written thus9:

SCE = 1 − ECQ .

(6)

When SCE has a positive value, an expansion of scale would cause the company's marginal cost to decrease, which gives them more profits. Expanding the scale with negative SCE leads the marginal cost to grow. It is recommended for a firm to reduce its supply size in terms of the average cost. If SCE is equal to 1, neither benefits nor losses exist as a result of scale expansion. 2.3. Definitions of variables We use six variables to estimate the scale economies of KDHC. First, VC (in millions of KRW) is the variable cost, and it is the sum of labor cost and material cost.10 To remove the price change over time, we adjust VC by using the producer price index. Second, PL is the price of labor; it is calculated as total labor expenses divided by the number of employees. Total labor expenses contain salaries and bonuses, but a reserve for retirement allowance is excluded. To make it a constant price, we use a gross domestic product (GDP) deflator; a constant labor price is derived by dividing the labor price by the GDP deflator. Third, PM is the material price; it is 9 We can judge scale economies through the cost efficiency, ECQ , by itself. ECQ > 1 indicates that a marginal cost increases the average cost, which means diseconomies of scale. The firm enjoys economies of scale when ECQ < 1. We used SCE as our index for convenience. 10 An anonymous referee pointed out that fuel is crucial input for DH production. There are two reasons why we did not separate fuel cost from material cost. A source of fuel can change over time and a complicated consideration of fuel price was required. We tried to analyze again with various versions that fuel price is exclusive, and find out that this separation does not change the overall conclusion.

200

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

obtained by dividing the total material cost by the scale of the power plant. In this calculation, the scale of the power plant is a proxy for total material size. The inflation adjustment is applied to PM with the producer price index. The next variable is the capital cost, K (in millions of KRW), which we define as a borrowed capital cost. The capital cost is determined by multiplying the fixed assets by the market interest rate; we use this method, as infrastructure data were difficult to obtain. Q (Gcal) indicates the heating, cooling, and electricity generation quantity of KDHC; the electricity that KDHC produces is measured in MWh, and 1 MWh is converted to 0.86 Gcal. Cooling means district cooling and electricity comes from CHP. Finally, we use T to denote technology innovation, including time trend. 2.4. Data description To quantify the economies of scale of South Korean DHSs, we need data vis-à-vis the production, labor, material, and capital costs of all DH suppliers, as well as data on their output. However, the period during which most DH firms have generated DH is relatively short—too short, in fact, to analyze economies of scale. In addition, these data are frequently either private or clandestine, making them very difficult to obtain. Rather than use data from all DH companies, we instead use 100 quarters of KDHC data from a 25-year period.11 Besides being the longest-running supplier of DH in South Korea, KDHC currently holds the country’s highest market share. Therefore, despite these data being limited, it does suffice in understanding overall trends. The KDHC data comes from three sources: financial statements, income statements, and cost reports. We collected the fixed assets from financial statements of KDHC. It will be capital cost in our data. District heating, cooling, and electricity generation of KDHC are obtained from income statements. Labor and material costs are acquired from their cost reports. Financial officers in KDHC helped us to secure these documents or data files. Trends in the changes that occurred between 1987 and 2011 in KDHC’s annual production facilities are depicted in Fig. 1. Changes in production at the facilities were very small between 1987 and 1991, but in 1992, production more than doubled relative to the 387 Gcal/h value seen in 1991. While that number slightly decreased during the 1999–2001 period, an increasing tendency was seen in the 10-year period ending in 2011. KDHC's production presents a shape similar to that seen among annual production capacities. As mentioned, production includes that of heating, cooling, and electricity. Fig. 1 shows an overall increase in production as that of the annual production capacities increases. A dramatic increase occurred in 1992, from 1 million Gcal to about 1.63 million Gcal. Between 1992 and 2000, the volumes of DH produced continued to grow. Despite transient drops in 2001 and 2006, annual production has grown since 2003; in fact, three outputs in 2011 reached a total of 19,609,000 Gcal.

3. Results and discussion 3.1. Estimation of cost function The coefficient estimates, as well as information on their statistical significance—both of which were obtained by estimating 11 DH use has a seasonal pattern: high in winter, low in summer, and moderate in spring and fall. DH use has a seasonal pattern: high in winter, low in summer, and moderate in spring and fall. Seasonality in supplying DH is an inherent feature. A cost-minimizing decision of KDHC incorporates its seasonality. Thus, quarterly data represents its characteristics better than yearly and monthly data formats.

Fig. 1. Annual production capacities and production at KDHC (1987–2011).

Table 1 Estimation results of the cost function. Coefficients

Variables

Estimates

t-values

α0 αL αM αML αLL αMM βQ

Intercept Labor price Materials price Materials price  Labor price Labor price2 Materials price2 Production

 3.533 0.518*** 0.483***  0.057*** 0.057*** 0.057*** 0.992***

 1.29 22.92 21.37  21.85 21.85 21.84 3.09

βLQ

Labor price  Production

 0.052***

 25.09

βMQ

Materials price  Production

0.052***

25.09

βQQ

Production2

0.068*

1.66

γK

Capital

0.205

0.41

γKQ

Capital  Production

 0.114***

γLK

Labor price  Capital

0.020

γMK

Materials price  Capital

γKK

Capital2

δT δTT δTL δTM δTK δTQ

Time Time Time Time Time Time

trend trend2 trend  Labor price trend  Materials price trend  Capital trend  Production

 0.020*** 0.123**  0.001  0.001 0.001***  0.001*** 0.004**  0.001

Log-likelihood

 3.44 12.13  12.13 2.09  0.06  0.77 14.24  14.24 2.43  0.63 392.05

Note: * **

indicate statistical significance at the 10%. indicate statistical significance at the 5%. indicate statistical significance at the 1% levels, respectively.

***

the cost function—are shown in Table 1. It was found that 13 of the 21 estimated coefficients are significant at the 1% level, and that that of Production2 is significant at the 10% level. The coefficient of Capital2 and of Time  Capital is significant at the 5% level. The Wald statistic is 821,886.23, and its p-value is less than 0.01. The result of the Wald test indicates that none of the coefficients of this cost function is zero; this means that the translog cost function is suitable for measuring the cost function of KDHC. Table 2 illustrates the value of SCE from 1987 to 2011; this value was determined by substituting the means of each variable. Since, 1987, SCE has maintained a value of 0.16 or more. Its highest value, in 1998, was close to 0.4; the minimum SCE value appeared in 1987, with a 1% significance level. SCE has maintained since 1994a value in the vicinity of 0.32. Since this value has exceeded zero in all years, the SCE values indicate that economies of scale have existed; more specifically, all the positive estimated SCE values imply that KDHC had economies of scale during the study period. In reading the data in Table 2, one can conclude that KDHC's current production level is smaller than the minimum efficient scale. Thus, KDHC would benefit from expanding its supply area.

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

Table 2 Annual economies of scale at KDHC (1987–2011). Year

SCE

t-values

1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

0.163 ** 0.247 ** 0.245 ** 0.239 ** 0.262 ** 0.292 ** 0.316 ** 0.344 ** 0.362 ** 0.371 ** 0.371 ** 0.385 ** 0.364 ** 0.363 ** 0.348 ** 0.346 ** 0.345 ** 0.346 ** 0.349 ** 0.353 ** 0.350 ** 0.339 ** 0.325 ** 0.333 ** 0.355 **

2.46 5.08 5.81 6.71 5.16 4.65 4.81 4.87 4.66 4.37 3.95 3.82 4.70 4.70 5.09 5.06 5.01 4.98 4.84 4.66 4.34 4.02 4.28 4.08 3.73

Note: **

indicates statistical significance at the 1%

level.

3.2. Forecasting economies of scale The “minimum efficient scale” refers to the maximum firm size at which there are still increasing returns to scale. If KDHC has a scale that exceeds the minimum efficient scale, it would experience diseconomies of scale, and increasing the scale would not be appropriate. To calculate the appropriate scale of heating, cooling, and electricity production, we use the average value from 2011. The heating, cooling, and electricity production level (Q*) at which economies of scale are lost can be calculated with the constant production level and labor and material prices of 2011, assuming that the business environment did not change significantly thereafter (save for the amount of heating, cooling, and electricity produced by KDHC). By undertaking examinations of how KDHC's economies of scale change as it increases its production of heating, cooling, and electricity while maintaining current production techniques, we can derive important policy implications. It was found that economies of scale for KDHC are lost when production volumes exceed 122,138,800 Gcal; this value is about 25 times the 4,902,430-Gcal average production volume of 2011. We can estimate and extrapolate future SCE values by leveraging data vis-à-vis KDHC's medium and long-term energy outlook sales for the 2013–2020 period. The results, shown in Table 3, indicate that the SCE values will remain roughly the same over that period. From this prediction, we expect that economies of scale will be maintained at KDHC for some time yet. SCE will supposedly decrease in some way and approach zero in the future, but the way in which it will decrease is not clear. Currently, it is difficult to predict how SCE could decrease after 2020. It may be possible to predict a pattern upon accumulating more data in the future. The intervals implying uncertainties in the future are presented in the last column of Table 3. The intervals include a lower bound and an upper bound. The small number is calculated with a 90% of the expected production amount each year, and, similarly, the

201

Table 3 Prospects of the degree of scale economies (SCE) over the period 2013–2020. Year

SCE estimates

t -values Intervals of SCE attained with 90% and 110% of production expected

2013 2014 2015 2016 2017 2018 2019 2020

0.3533 ** 0.3535 ** 0.3539 ** 0.3547 ** 0.3560 ** 0.3561 ** 0.3563 ** 0.3564 **

3.68 3.67 3.66 3.62 3.57 3.56 3.56 3.55

Note:

nn

0.3526–0.3539 0.3528–0.3541 0.3532–0.3545 0.3540–0.3553 0.3553–0.3566 0.3554–0.3567 0.3556–0.3568 0.3557–0.3569

indicates statistical significance at the 1% level.

large number is computed with a 110% of them. All numbers in the intervals are placed between 0.35 and 0.36, which represent the company holds scale economies. Consequently, the intervals indicate that scale economies will continue in the light of probable errors. We conclude that it is efficient to increase the production of heating, cooling, and electricity, in that KDHC would gain economies of scale as it expands its business. Although it is not possible to obtain data concerning district energy companies and we cannot derive the policy implications of the district energy market, it is recommended that KDHC increase its production level, regardless of the number or size of other DH companies. KDHC may merge with another company who has had business management challenges, and establish more branch offices. On the basis of the current environment, KDHC can derive benefits from reducing the average production costs as it produces each additional unit. As KDHC has the largest DH market share in South Korea, any increase it makes in DH production will lead to change in the district energy industry structure. 3.3. Discussion The findings of this paper are in the line with the fourth National Plan for DHS of Korea as follows. First, this plan is related with increasing production level of district cooling, it could drive a decrease in average costs of district cooling based on the results of the paper. From the decline of average costs, KDHC is able to enjoy the economies of scale. Second, the results of paper also related with a decrease in production costs of supplying a heat since one of the major purpose of the plan is to build heat networks. Constructing a huge networks of providing a heat could leading a lower level of average costs, thus KDHC also benefits from the economies of scale. We suggest five ways in which KDHC can manage itself in a reasonable and efficient manner. First, it is essential that unused or idle facilities be fully utilized. Although additional demand would be needed to consume the additional product, KDHC can create an extra DH supply in a highly efficient manner, on account of its decreasing average costs. Second, finding out extra heat demand is recommended although the Korean government has prevented KDHC from beginning a new business. For instance, building heat networks helps to increase its heat sales without a participation in new businesses. Heat networks connect one another among heat suppliers. KDHC would have an opportunity to sell its heat to another supplier. This suggestion might be well since KDHC is able to provide heat with low price based on the scale of economies, and buying heat from KDHC at low price could help heat suppliers having more benefits. Many of the heat suppliers in the market are in trouble with financial constraints from a huge investment paid or loss earnings. Thus, the suggestion could make heat suppliers happy for lowering

202

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

costs and could lead a decrease of the average costs of KDHC with greater production level. Third, KDHC have to reinforce its competitiveness with lowpriced heat sources. When KDHC builds the new facilities or power plants, KDHC need to adapt low energy cost system. Ignored or non-popular energy sources such as biomass and waste will reduce its production cost, thus, its production would increase due to its economical price of heat. The price of energy can change over time and it tends to increase. A low price of energy might reduce these risks and increase earnings. Fourth, extending supply of district cooling could help KDHC for two ways: reducing seasonal instability and increasing total production level. Reducing seasonal instability would lower KDHC's business volatility and lower the average costs. Increasing total production level also benefits for KDHC because of a decrease in average costs. In sum, increasing supply of district cooling is cost effective way for KDHC. Fifth, the drive to reduce production costs will help promote total cost reductions. Based on our findings, KDHC currently operates under economies of scale, but there is room to improve its operational efficiencies. Finding those efficiencies will help KDHC maintain its economies of scale and the resulting benefits. Something to bear in mind is that we analyzed the presence of economies of scale while assuming that the business environment and technology seen and used today will be maintained into the future. Uncovering additional means of reducing KDHC’s production costs might enlarge the company’s minimum efficient scale in the future. On the other hand, when the company is run in an inefficient manner—either in whole or in part—the minimum efficient scale will be smaller than the current expected size. KDHC should find a way of reducing its costs, if it wishes to enjoy cost efficiencies farther into the future. The results suggest that DHS is more preferred than pre-existing system of individual heating using natural gas. This preference is based on the strong points of DHS, as noted earlier, DHS produces less air pollutants and CO2 and more energy-efficient than individual heating system. The preference to DHS will lead the growth of DHS in Korea from now on. The future of DHS is promising because DH suppliers try to provide heating, cooling, and electricity with less production costs. Decreasing the average cost will fasten the spread of DHS in Korea. Based on this study’s results, we can make some policy recommendations. First, the South Korean government could allow KDHC to expand its DH business.12 KDHC will be the supplier of choice as DH demand grows, and if KDHC were to expand its provision area, the average associated cost would be reduced as its business benefits from economies of scale. Secondly, it is not efficient to split up KDHC’s offices or branches, as the company benefits from their associated economies of scale. (This study does not speak of the existence of economies of scale in suppliers other than KDHC, though they may certainly exist.) Changing the operator of what is currently KDHC’s provision area would lower efficiencies; it is therefore recommended that KDHC continue to supply DH within its provision area, as KDHC’s unit cost would be lower than that of a new, incoming company. Third, we encourage KDHC to invest in infrastructure such as pipeline systems and plants, if it wishes to maintain its cost 12 Since 2008, Korean Ministry of Trade, Industry & Energy has regulated the market share of KDHC lower than 50% to make DH business market more competitive. Thus, KDHC cannot begin new DH business until its market share drops below the half. There are both pros and cons for the regulation. Some argue that it is necessary to loosen or abrogate the regulation for the consumers' benefit because the price of heat supplied by KDHC is lower than that by any other DH providers. Others assert that the regulation should be continued to make competitive market environment.

efficiency. Following the introduction of new operators to the South Korean DH sector, some have experienced management challenges or fallen into a deficit state. Efficient business operations demand appropriate investments in power plants or other facilities. The findings of this study indicate that KDHC is currently enjoying economies of scale and that its investment in plants is appropriate. As it increases the scale of its provisions, its production capacity will approach the minimum efficient scale, resulting in lower average costs. Therefore, enlarging or maintaining its investment in DH would be desirable for KDHC.

4. Conclusion and policy implications Cost efficiencies in South Korea's DH industry are highly evaluated. For a while, new entrants entered and new businesses were launched, in support of competition-oriented policy. Given the characteristics of DH plants, this change led to many financial problems among a number of companies; indeed, many operators are now “in the red” or have abandoned their supply plans. Many stakeholders have cast doubt on the assertion that the industry structure is a cost-efficient one. This study sought to determine whether KDHC has economies of scale; if economies of scale were found to be present, supplies of DH from KDHC would be considered more efficient than those from other new operators, especially as demand increases and new supplies are proposed. Because KDHC was deemed to have economies of scale, it does indeed have an advantage over other companies. We estimated economies of scale by using a variable cost function and data from KDHC, a South Korean supplier of DH. We found that KDHC is running its business with economies of scale: in other words, KDHC holds an excess capacity rather than an optimal capacity, and this will minimize its costs in the long term. The lowest point on the average cost curve is the facility level at which costs are minimized in the long term; however, the current KDHC production level is to the left of that point, and therefore in the cost reduction area. This means that KDHC produces at a scale smaller than that at which it could minimize its cost in the long run. KDHC could expand its business into regions where economies of scale are reserved and its average costs would not increase. We predicted the prospects of scale economies by using mid- and long-term DH projections. We found that growth in KDHC's provision would be socially efficient, and that the company's DH production would be at a minimally efficient scale. The implication is that KDHC should aggressively source new business areas, in order to improve its social efficiency. Expanding its business will affect its operational efficiency, in that it can alleviate or mitigate shocks relating to the industry or materials provision.

References Agrell, P.J., Bogetoft, P., 2005. Economic and environmental efficiency of district heating plants. Energy Policy 33, 1351–1362. Berry, S., 1992. Estimation of a model of entry in the airline industry. Econometrica 60, 889–917. Bottasso, A., Conti, M., 2008. Scale economies, technology and technical change in the water industry: evidence from the English water only sector. Reg. Sci. Urban Econ. 39, 138–147. Cambini, C., Filippini, M., Piacenza, M., Vannoni, D., 2011. The influence of corporatization process on the cost of publicly provided local utilities: some insights from bus transit systems. Rev. Law Econ. 7, 196–217. Caves, D.W., Christensen, L.R., Swanson, J.A., 1981. Productivity growth, scale economies, and capacity utilization in U.S. railroads, 1955–1974. Am. Econ. Rev. 71, 994–1002. Charnes, A., Cooper, W.W., Suetoshi, T., 1988. A goal programming/constrained regression review of the bell system breakup. Manag. Sci. 34, 1–26. Christensen, L.R., Greene, W.H., 1976. Economies of scale in U.S. electric power

S.-Y. Park et al. / Energy Policy 88 (2016) 197–203

generation. J. Polit. Econ. 84, 655–676. Di Giacomo, M., Ottoz, E., 2010. The relevance of scale and scope economies in the provision of urban and intercity bus transport. J. Transp. Econ. Policy 11, 161–187. Evans, D.S., Heckman, J.J., 1988. Natural monopoly and the bell system: response to charnes cooper Suetoshi. Manag. Sci. 34, 27–38. Farsi, M., Fetz, A., Filippini, M., 2007. Economies of scale and scope in local public transportation. J. Transp. Econ. Policy 41, 345–361. Haichao, W., Wenling, J., Lahdelma, R., Pinghua, Z., Shuhui, Z., 2013. Atmospheric environmental impact assessment of a combined district heating system. Build. Environ. 64, 200–212. Ida, T., Suda, M., 2004. The cost structure of the Japanese railway industry: the economies of network density and of scope and the cost gap between Japan's regional railways after privatization. Int. J. Transp. Econ. 31, 23–27. Ilic, D.D., Trygg, L., 2014. Economic and environmental benefits of converting industrial processes to district heating. Energy Convers. Manag. 87, 305–314. Kim, E., Lee, H., 1998. Spatial integration of urban water services and economies of scale. Rev. Urban Reg. Dev. Stud. 10, 3–18. Korea Energy Agency, 2015. 2014 Community Energy Service Handbook. Korea Energy Agency, Yongin. Korean Ministry of Trade, Industry & Energy, 2014. The fourth framework for the district energy supply, Sejong-si (in Korean). Lazzarin, R., Noro, M., 2006. Local or district heating by natural gas: which is better from energetic, environmental and economic point of views? Appl. Therm. Eng. 26, 244–250. Loizides, J., Tsionas, E.G., 2002. Producticity growth in european railways: a new approach. Transp. Res. Part A Policy Pract. 36, 633–644. Lygnerud, K., Peltola-Ojala, P., 2010. Factors impacting district heating companies' decision to provide small house customers with heat. Appl. Energy 87, 185–190. Martins, R., Fortunato, A., Coelho, F., 2006. Cost Structure of the Portuguese Water Industry: A Cubic Cost Function Application. University of Coimbra, Coimbra, Working Paper No. 9. McGeehan, H., 1993. Railway costs and productivity growth: The case of the republic of Ireland, 1973–1983. J. Transp. Econ. Policy 27, 19–32.

203

Munksgaard, J., Pade, L.L., Fristrup, P., 2005. Efficiency gains in Danish district heating. Is there anything to learn from benchmarking? Energy Policy 33, 1986–1997. Nelson, R.A., 1989. On the measurement of capacity utilization. J. Ind. Econ. 37, 273–286. Ottoz, E., Di Giacomo, M., 2012. Diversification strategies and scope economies: evidence from a sample of Italian regional bus transport providers. Appl. Econ. 44, 2867–2880. Oum, T.H., Zhang, A., Zhang, Y., 1993. Inter-firm rivalry and firm-specific price elasticities in deregulated airline markets. J. Transp. Econ. Policy 27, 1–22. Renzetti, S., 1999. Municipal water supply and sewage treatment: costs, prices and distortions. Can. J. Econ. 32, 688–704. Russell, D., MacKinnon, J.G., 1993. Estimation and Inference in Econometrics. Oxford University Press, New York, p. 874. Saal, D., Parker, D., Weyman-Jones, T., 2007. Determining the contribution of technical change, efficiency change and scale change to productivity growth in the privatized English and Welsh water and sewerage industry: 1985–2000. J. Prod. Anal. 28, 127–139. Saal, D., Arocena, P., Maziotis, A., Triebs, T., 2013. Scale and scope economies and the efficient vertical and horizontal configuration of the water industry: a survey of the literature. Rev. Netw. Econ. 12, 93–129. Sjödin, J., Henning, D., 2004. Calculating the marginal costs of a district-heating utility. Appl. Energy 78, 1–18. Söderholm, P., Wårell, L., 2011. Market opening and third party access in district heating networks. Energy Policy 39, 742–752. Sparks, R., 2004. SUR models applied to an environmental situation with missing data and censored values. J. Appl. Math. Decis. Sci. 8, 15–32. Torchio, M.F., Genon, G., Poggio, A., Poggio, M., 2009. Merging of energy and environmental analyses for district heating systems. Energy 34, 220–227. Wibe, S., 2001. Substitution och teknologisk förändring inom den svenskafjärrvärmesektorn, Report 124. Department of Forest Economics, Swedish University of Agricultural Sciences, Umeå, Sweden. Zellner, A., 1962. An efficient method for estimating seemingly unrelated regressions and tests for aggregation bias. J. Am. Stat. Assoc. 57, 348–368.