Carbon pricing for low carbon technology diffusion: A survey analysis of China's cement industry

Energy 106 (2016) 73e86

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Carbon pricing for low carbon technology diffusion: A survey analysis of China's cement industry Xianbing Liu a, *, Yongbin Fan b, Chen Li b a b

Kansai Research Centre, Institute for Global Environmental Strategies, Japan China Cement Association, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 November 2015 Received in revised form 4 March 2016 Accepted 10 March 2016

This study estimates the effect of using carbon pricing to promote the diffusion of low carbon technologies based on data collected from 78 cement companies in China. The analysis confirms that they are familiar with major energy saving and low carbon technologies in the sector and have made efforts in energy saving, but are lagging in terms of carbon management. An average payback time of 3.3 years is confirmed as the threshold for cement companies to determine technology investment. The adoptions of target technologies in this survey are at different stages; WHR (waste heat recovery power generation) systems have been largely diffused and the effect of carbon pricing is highly marginal for further adoption. On the other hand, levying a moderate carbon price, i.e., 60 Yuan/t-CO2, may accelerate the diffusion of EMOS (energy management and optimisation systems), recently introduced in China's cement industry. This research goes some way to clarifying the diffusion of low carbon technologies and provides implications for climate countermeasures for the target sector in China. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Low carbon technology Diffusion Energy saving Cement industry China

1. Introduction China is the largest GHG (greenhouse gases) emitter in the world and to realise its 20% reduction in energy intensity during the 11th FYP (five-year plan) period (2006e2010), it mainly relied on administrative approaches [45]. A similar policy framework continued into the 12th FYP period (2011e2015) to achieve a further 17% cut in carbon intensity. However, the country cannot rely on costly administrative measures to realise its pledge of a 40e45% reduction in carbon intensity by 2020 from 2005 levels or its commitment to cap GHG emissions by around 2030 under the joint climate statement between China and the U.S. announced on November 11, 2014. The importance of allowing the market to play a decisive role in resource allocation has been recognised by China's leadership, which calls for a ramp-up in use of MBIs (market-based instruments) as a complementary policy measure. In practice, the NDRC (National Development and Reform Commission) approved the pilot GHG ETS (GHG emissions trading

* Corresponding author. Kansai Research Centre, Institute for Global Environmental Strategies (IGES), Hitomirai Building 5F, 1-5-2, Wakinohama Kaigan Dori, Chuo-ku, Kobe, Hyogo, 651-0073, Japan. Tel.: þ81 78 262 6634; fax: þ81 78 262 6635. E-mail address: [email protected] (X. Liu). http://dx.doi.org/10.1016/j.energy.2016.03.044 0360-5442/© 2016 Elsevier Ltd. All rights reserved.

schemes) in late October 2011 for seven areas, including five metropolitan cities (Beijing, Shanghai, Tianjin, Chongqing and Shenzhen) and two provinces (Hubei and Guangdong) [38]. After preparations, Shenzhen began trading in June 2013 and the other pilot regions started between the end of 2013 and mid-2014. As a follow-up effort, NDRC announced that a nationwide carbon market would be established as early as 2016 [41], and the recently released interim measures for carbon emissions trading management was an additional step forward in this direction [40]. Meanwhile, over the last few years, experts at research institutes of related ministries have been discussing how to develop carbon tax policy in China, and they conclude that a carbon tax should be phased in gradually, and very conservatively, starting from 10 Yuan/ t-CO2 and increasing to 40 Yuan/t-CO2 some years later [28]. Strong resistance from industry has been identified as the most significant barrier for the implementation GHG ETS and carbon tax. Nevertheless, scant empirical research clarifying just what would be acceptable for business in terms of carbon pricing has been carried out. Aiming to close this gap, Liu et al. [28] discussed carbon price levels that in theory would be tolerated by companies in Northeast Asia and China in particular. Extending this line of enquiry, this study estimates the effects of using carbon pricing to enhance diffusion of LCT (low carbon technology) in China's cement industry. According to the literature review in Section 2, many

74

X. Liu et al. / Energy 106 (2016) 73e86

studies have attempted to empirically clarify the factors determining how business invests in LCT; however, scant literature concerning how policy functions in promoting LCT diffusion exists, especially for developing economies like China, and almost all the material concerns retrospective analyses of policies based on currently available historical datasets (e.g., [2,43]). This research differs in that it is a prospective policy analysis and uses fresh, and therefore up-to-date, information directly obtained from the companies concerned, and provides results that may contribute to understanding how Chinese companies might invest in LCT at different levels of carbon pricedwhich in turn informs on the technology and policy solutions needed in order to bring about low carbon development of the target industry in the future. Using a selection of main LCT according to their classification in the target industry, the information to be gathered by survey included: a) technology-specific data, e.g., historical adoption status, initial investment and energy saving potential; b) change in technology profitability with payback time as a proxy under assumed carbon prices; and c) change in business investment decision for technologies with different levels of profitability. China's cement industry was targeted in this research due to its high potential for energy saving and carbon mitigation both domestically and globally. China has held the No.1 spot in cement production since 1985, and in 2013 had a 60% share, at 2.42 billion tonnes. Demand is still rising within the country, however, and isn't predicted to peak until 2018 to 2020 [22]. The industry is characterised by its relatively low concentration: in 2012, the top 21 cement companies accounted for just over one billion tonnes of clinker capacity per year, representing about 58% of the country's total clinker capacity [17]. Uneconomical cement plants with a total capacity of around 250 million tonnes per year are due to be shut down between 2013 and 2015 [13]. Cement production is a major source of CO2 emissions in China owing to the large volume produced, thus estimations of its CO2 emissions have attracted much attention. However, different studies have delivered highly differing estimates. Ke et al. [20] reported that the cement industry accounted for 13%e14% of China's total fossil-fuel emissions during 2005e2007, whereas Wang [49] gives a figure of around 11% of national gross CO2 emissions in 2011. Although the domestically advanced values of comparable energy use are almost the same as those found abroad, the average energy efficiency of China's cement industry is about 10% lower, implying a large mitigation potential exists for LCT [33]. The remainder of this paper is structured as follows: Section 2 gives an overview of previous literature on LCT diffusions and describes the contributions of this research. Section 3 explains the research methodology, including the procedures and models applied for the simulations. Section 4 lists the major technologies for energy saving and carbon mitigation of China's cement sector and details three target technologies. Section 5 outlines the questionnaire, survey implementation and samples. As the main component, Section 6 discusses the results of this survey analysis. Lastly, Section 7 provides a summary and suggests topics for future research. 2. Literature review and contributions of this research Previous studies indicate that accelerated technology development may reduce the costs for achieving stringent climate goals [31]. The diffusion of energy saving technologies is an important part of energy and climate policy, and the diffusion of innovation is influenced by various factors [48]. One of these factors is the pricing of carbon emissions, which would induce profit-oriented businesses to adopt LCT; however, conventional economic theory alone may not perfectly explain the diffusion of LCT in reality [15,19]. The

factors determining the pace of LCT diffusion can be classified into twodone is associated with the uncertainty of climate policies and the other is the factors influencing technology growth in the presence of favourable climate policies. Uncertainty in climate policies may induce a certain ‘opportunity value’ of postponing the technology adoption [18]. In particular, climate sensitivity, international commitments and the stability of carbon prices influence the behaviours of risk-averse and risk-neutral investors [7]. For example, producers facing market uncertainty over CO2 prices might invest in carbon-saving technology earlier than if the actual price path were known beforehand. However, uncertainty over governmental policy would prompt them to adopt a wait-and-see approach, since the government's future commitment to climate policy is unknown [8]. The lack of appropriate regulatory framework for the transport and storage of CO2 would likely impede the commercialisation of CCS (carbon capture and storage) [10]. Further, the high transaction cost of intellectual property and weak patent protection in developing economies hinder the transfer and diffusion of LCT from the developed world [4]. Besides policy uncertainty, lack of financial resources was confirmed to be an important barrier to the adoption of LCT, which usually incurs high upfront costs [4]. Cagno and Trianni [5] conducted a survey on small and medium-sized companies in northern Italy and identified access to capital as one of the chief barriers. Technology diffusion is also influenced by the preference of riskaverse stakeholders [21]dfor instance, CCS might experience public opposition due to concern over CO2 transportation and injection [24]. Characteristics of the industry such as market structure and internal information flow may also constrain the diffusion of LCT; it may take some time for the potential users to obtain information about the technology and adapt it to their own circumstances [19]. The business alliances influence the technology diffusion as means for knowledge transfer and information dissemination, and may reduce the risk of technology adoption [3]. The characteristics of individual companies also affect the diffusion of technology; whether or not a company adopts LCT depends on its capacity and whether the timing is appropriate in respect of other business cycles [35]. Prindle [44] distributed a questionnaire survey to a number of U.S. companies and identified a lack of funding, personnel with appropriate skills and technical information as common barriers. Several studies have empirically investigated the historical growth of LCT, and some focus on the technologies in the energy sector. Kramer and Haigh [23] postulated on the growth of energy technologies over the last century and found that once an energy technology materialises, e.g., accounts for around 1% of total global energy, the growth changes to linear. They also found that most energy technology transitions occur at low rates. Lund [30] showed that the takeover time from a 1% to 50% share of market potential varies from less than 10 to 70 years for different technologies and that short takeover times below 25 years are mainly associated with energy end-use technologies. Other studies are related to energy efficiency technologies. Pizer et al. [43] examined the adoption of energy saving technologies by U.S. manufacturing companies, and found that once an energy saving technology has diffused to 10% of the companies, the remaining potential users would adopt within an average of about nine years, regardless of industry. It was also revealed that even dramatic changes in energy prices could only generate modest changes in energy efficiencyda policy that immediately raises energy prices without allowing companies to anticipate such change may result in reduced technology adoption, due to the reduced financial health of such companies [43]. Anderson and Newell [2] confirmed that companies adopted around half of the energy-saving projects recommended by the U.S. energy audit programme and that adoption rates were

X. Liu et al. / Energy 106 (2016) 73e86

systematic theory [46]. The theory has basically remained unchanged and states that when the number of users of a technology is plotted versus time, the resulting curve is typically an S-shaped distribution. Geroski [9] surveyed the literature and focused on alternative explanations for this predominant factdthe most commonly found one is the so-called ‘epidemic’ model, which is premised on the speed of usage being limited by the lack of available information on a certain technology. The leading alternate is the called ‘probit’ model, following the premise that different firms, with different goals and abilities, are likely to adopt a technology at different times. The learning, or epidemic model widely applied in the marketing and sociological literature on technology diffusion, is simple, transparent and able to capture the main features of the technology diffusion process and thus was applied in this analysis [11]. Despite the appeal of the potential insights, the conventional diffusion analysis measures have a key drawback: the required computations involve micro-level data sets, which are hard to assemble. Referencing Lund [30]; which provides an empirical study of the market penetration of energy technologies, this research simulates the diffusion of target technologies in China's cement industry by fitting the survey data to an epidemic model. Defining the number of users of a technology at time t to be V(t), the remaining potential for the technology will be V*  V(t), where V* is the maximum adoption potential of the technology, then,

higher for projects with shorter payback periods, lower initial costs and greater energy savings. Companies are more responsive to the initial costs than annual savings, suggesting that financial subsidies may be more effective than energy price increases for promoting energy-efficiency technologies. Diverging from the retrospective policy studies reviewed above, this research is an attempt at a prospective policy analysis to estimate the change in LCT diffusion in response to certain assumed carbon price levels. In terms of method this study extends the application of the MBDC (multi-bounded discrete choice) formatda method frequently applied to measure individual willingness-to-pay for environmental improvementdto businesses. Using MBDC this research measured the likelihood of companies of investing in technologies according to level of profitability. The effects of certain policies are reflected in the changes in the profitability of target technologies, as observed from the technology investment possibility curves. This makes it possible to predict the technology diffusion changes as consequences of the assumed policy conditions. In spite of the inherent difficulties in surveying the companies, this research managed to obtain responses from China's cement sector to the extent possible in order to procure the statistics and simulations in this research. 3. Methodology The analytical frame for this study is outlined in Fig. 1 and consists of three steps. Step 1 is to simulate the diffusion of target technologies in the case of BAU (business-as-usual) using the collected data of technology historical adoptions. Step 2 measures the investment possibility of companies for technologies with various payback times. The last step is the calculation of technology diffusion changes under the assumed policy scenarios in comparison with the BAU case. The methods are described in detail as follows.

dV ¼ b

f ðtÞ ¼

Adoption time and scale of technology by samples

Investment possibility at various payback times

(Survey data)

(Survey data)

f (t), t = t0,

1 1 þ a*expð  b*ðt  t0 ÞÞ

Data compilation

Sample shares by investment possibility at various payback times

, now

Step 3 Energy use ratios and prices of samples (Survey data)

Energy emission factors and technology payback time (Survey and literature data) Calculation

Regression

Investment possibility curves by payback time

Regression

Technology payback times under assumed carbon prices Calculation

BAU: f (t), t = t0, (Simulated diffusion ratio)

(3)

where f(t) is the share of technology adoption potential realised; t is the year; t0 is the starting year for technology adoption; and a and b are the coefficients to be estimated. Clearly, f(t0) > 0, which implies

Step 2

(Accumulated diffusion ratio)

(2)

Assuming b to be constant, Eq. (2) can be solved by integration to yield the well-known logistic time curve,

Step 1

Calculation

(1)

df ¼ bf ð1  f Þdt

Much literature in various fields has been published on the diffusion of innovation, most of it since Rogers' creation of a

Number and scale of technology adoption by year

 V
* V  V dt V*

Denoting f(t) ¼ V(t)/V*, Eq. (1) can be written as,

3.1. Empirical model estimating the diffusion of low carbon technologies

Data compilation

75

Calculation

Investment possibility changes from BAU

Policy scenarios: f (t), t = Now, (Calculated diffusion ratio)

Fig. 1. Analytical frame and procedures for this study.

76

X. Liu et al. / Energy 106 (2016) 73e86

that this model can only explain the technology diffusion after some early users have already entered into the market. The factor b describes the diffusion pace and can be defined as the penetration rate of the technology. This coefficient determines the slope and steepness of the diffusion curve and a high value is preferred for the technology to achieve sizeable diffusion within a period. 3.2. Method of measuring the possibility of companies to invest in technologies As a method developed from contingent valuation, the MBDC format allows respondents to vote on a wide range of referendums and express voting certainty for each, which reinforces the quantity and quality of the data. Considering the advantage of this method, the MBDC format was applied in this study to measure the possibility of companies to invest in technologies with different economic profitability, using the technology payback time as a proxy. With reference to Liu et al. [26]; the question and format prepared for this survey and an example response from a cement company are shown in Table 1. A total of 15 thresholds of payback time are listed for the companies to circle their investment possibility, and multiple choice optionsd‘very high’, ‘high’, ‘moderate’, ‘low’ and ‘very low’ possibilitydwere offered. Given a payback time threshold of PBij, the probability for the companies to invest in a technology can be written as,



 Pij ¼ Pr Vi > PBij ¼ 1  F PBij

(4)

Once Pij, the probability for company i to invest under the jth payback time PBij, is known by assigning numerical values to the verbal MBDC answers, Eq. (4) can be estimated for each company. Assuming a specific function for F(PBij), e.g., a normal accumulative distribution with a mean of mi and a standard variance of si, the estimation model can be written as,

  PBij  mi þ li Pij ¼ 1  F

(5)

si

Table 1 The question and an example response of MBDC format in this survey. Question: The adoption of energy saving low carbon technologies can reduce energy use and carbon emissions of companies. Accordingly, energy and carbon emissions costs of companies can be reduced. The initial investments of technologies may be recouped within a certain period. The payback time of technologies are different due to various initial investments, operation costs, energy saving potentials and life spans, etc. We would like to know your company opinion on the decision making of investment in energy saving low carbon technologies with different payback times. Please tick the possibility of your company to invest in the technologies under various payback times. Payback time (year)

Your company's possibility to invest Very high High Moderate Low

Very low

0.25 (3 months) 0.5 (half a year) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 5.0 6.0 7.0 8.0 9.0 10.0

Ⓐ Ⓐ A A A A A A A A A A A A A

E E E E E E E Ⓔ Ⓔ Ⓔ Ⓔ Ⓔ Ⓔ Ⓔ Ⓔ

B B Ⓑ Ⓑ B B B B B B B B B B B

C C C C Ⓒ Ⓒ C C C C C C C C C

D D D D D D Ⓓ D D D D D D D D

where Pij is the probability for company i to invest; PBij is the threshold of technology payback time; mi and si is the mean and standard variance of the distribution; and li is an error term. 3.3. Calculation of policy effects for low carbon technology diffusion Referring to Liu et al. [26]; the relationship between energy price increase of companies on average (MEANEnergy) and the imposed carbon price can be expressed as Eq. (6), where i indicates the energy type.

P

Emission factori  Energy ratioi MEANEnergy ¼ Carbon price  iP Energy pricei  Energy ratioi i

(6) Using energy use ratios and the energy prices collected from the companies, as well as emissions factors of energies from the literature, the assumed carbon prices can be converted to average energy price increases for the samples. Assuming the initial investment, operation cost and energy saving of a technology to all be constant, the decrease in technology payback time under assumed carbon prices from the BAU case would be equal to the price increase ratios of the energies saved. The increased possibility of technology investment from the BAU can be obtained from the information achieved by the approach in Section 3.2. Applying the increased investment possibility and the relative market growth rate of the technology in the BAU case, which may be expressed as [f(t)  f(t  1)]/[1  f(t  1)] for the year t, technology diffusion rates under the assumed carbon prices can be quantified by repeating the calculations. 4. Energy saving and low carbon technologies targeted in this survey There are four levers of technology measures for carbon reductions in the cement industry: a) Alternative raw materials and fuels, i.e., use of less carbon-intensive raw materials and fossil fuels, and more biomass in cement production; b) Energy efficiency improvements, e.g., deployment of state of the art technologies and retrofits of energy-efficient equipment; c) Clinker substitution, such as replacing clinker by other materials with similar properties but lower carbon content; and, d) CCS (carbon capture and storage). Such technology levers all interact in different ways to realise the potential for carbon reductions [16]. Typical alternative fuels include pre-treated industrial and municipal wastes, discarded tires, plastics, textiles and paper residues and biomass. Fuel-related CO2 emissions account for about 40% of total emissions from cement manufacturing and the reduction potential from alternative fuels could be significant. To improve energy efficiency, new companies usually install the most advanced technologies, and for cement production this is the NSP (new suspension pre-heater) kiln. Developments have also been made in burner technology. Fourth-generation clinker coolers are also available [47]. Some minerals, notably blast furnace slag, fly ash and natural volcanic materials, can be used to partially substitute clinker in cement and reduce the process and fuel-related CO2 emissions from clinker production. CCS is an end-of-pipe technology with promising potential but still unproven on the industrial scale in cement production. In practice, the industry can only handle energy efficiency improvements at this time [16]. Major energy saving and low carbon technologies in the cement industry are depicted in Fig. 2. A total of 14 technologies are listed and individually abbreviated as T01 to T14. In previous studies (e.g., [14,50]), a much longer list of energy efficiency measures were

X. Liu et al. / Energy 106 (2016) 73e86

considered. Hasanbeigi et al. [14], for example, surveyed 16 cement plants in China's Shandong Province in order to quantify their potential for energy saving potential with a total of 34 technologies. The survey used in the present research focuses more on the technologies with relatively high energy saving and carbon mitigation potentialsdthose listed in the national technology promotion cataloguedand all were listed in the questionnaire in order to monitor a company's subjective awareness thereof. As explained above, alternative fuels and alternative raw materials are influenced by many factors outside China's cement industry management control and CCS is still not at the stage of commercial application. Only three energy efficiency technologies, T09 to T11 in Fig. 2, are selected for the survey analysis of their historical adoption and company investment changes under assumed policy conditions. The survey only targeted a very limited number of technologies, mainly due to concern over the presumed reluctance, on the part of companies, to divulge detailed private information, but also because they were sufficiently representative as regards the extent of their diffusion. T09 (Waste heat recovery power generation) has been widely applied, while the other two are at the early stage of application, which the authors thought would allow the policy effects to be compared in terms of the different features of technology diffusion. The three technologies are individually described in more detail below. 4.1. Waste heat recovery power generation Waste heat recovery power generation (Hereinafter shortened as WHR system) utilises the heat of exhaust gases from cement kilns to produce low-pressure steam for power generation, and can be applied for medium- to large-sized kilns. The waste heat contained in the exhaust gases represents about 30% of the energy consumed in a clinker burning system, thus this technology is essential for reducing energy consumption per unit of cement product [32]. For a kiln of 5000 t-clinker/day capacity, up to 9 MW of power can be extracted using the facilitated WHR system. Power generation may range from 37 to 42 kWh/t-clinker, which equates to over 1/3 of the electricity consumed by the cement production

77

lines. The related retrofit costs around 56 million Yuan, provides an energy saving of 22,000 tce per year, and has a technology payback time of around 3.0 years [37]. In 2008, China's cement industry had over 60 WHR facilities in operation and over 100 at the design and construction stage, with a diffusion rate of this technology of around 8.5% [37]. By the end of 2008, there were around 935 NSP lines, of which about 700 had a production capacity over 2000 t-clinker per day and the potential for adopting the WHR system. Installation of WHR systems hit its peak in 2009 when 181 new systems went into operation, and after that the number steadily dropped. As of 2012, 60% of modern kiln lines in China had them. In terms of share of the global market, China had 739 WHR systems, or 85.4% at that time [12]. 4.2. Energy saving monitoring and optimisation system in cement kilns Cement kilns are often operated based on the experience of workers and it is difficult to precisely control the temperature and other parameters, which affects the clinker quality and leads to excessive energy use. The advanced energy saving monitoring and optimisation system (hereinafter abbreviated as EMOS technology) involves sampling the exhaust gases in cement kilns and estimating the state of combustion and energy consumption according to the gas composition and provides guidance on operational optimisation. This information and communication technology is utilised to construct large-scale monitoring networks in order to enable energy saving. Technical and management staff can also access the latest data in order to tweak production parameters [39]. EMOS technology has already been put to use in some cement companies in China. Using EMOS, heat consumption may be reduced by 70 kcal/kg-clinker for NSP kilns with production capacities above 2500 t/d [39], energy savings of about 1.4 million tce per year [39] can be expected, and 10% of NSP lines look set to adopt it by 2015. The MIIT (Ministry of Industry and Information Technology) has implemented more than 30 EMOS pilots in cement and glass companies since 2009, and actual operation confirms an improvement of 3e5% in overall energy efficiency of the

Fig. 2. Energy saving and low carbon technologies in cement industry.

78

X. Liu et al. / Energy 106 (2016) 73e86

companies. According to MIIT plans, 200 EMOS are set to be established in the building materials industry before 2020, of which about 140 will be set up in cement plants [34]. 4.3. Oxygen-enriched combustion Oxygen-enriched combustion uses air of a higher oxygen content than usual to improve thermal and heat efficiency in cement kilns. This technology includes producing the oxygen and delivering oxygen-rich air into the combustion equipment. Oxygen is usually produced by the membrane method. Compared with normal air, oxygen-enriched combustion technology can reduce coal consumption by 8e10% and reduce conventional pollutants like SO2 and NOx. No adoption of this technology took place in China's cement sector in the 11th FYP period (2006e2010), but grew to about 5% by 2015 [36]. 5. Outline of the survey and samples The data was collected by a survey from November 2014 to February 2015. Based on an understanding of China's cement industry, a questionnaire was prepared with the main purpose of measuring the awareness of companies on energy saving and low carbon technologies; clarifying the adoption status of the three target technologies; estimating the possibility for companies to invest in technologies with various payback times; and identifying the determinant factors influencing a company's technology adoptions. The questionnaire format consists of four components: a) Basic information of the company; b) Status of company energy saving management and adoption of target technologies; c) Factors determining the investment in LCT; and, d) Policies promoting technology diffusion of the companies. The survey was coordinated by CCA (China Cement Association) and carried out in two phases. The first phase was an experimental survey carried out in November 2014. A draft format was sent to six cement companies to test the easiness of answering questions, and based on the responses and comments therefrom, the questionnaire was slightly modified, particularly the question on WHR system adoption status. In the second phase, the finalised format was sent to a total of 270 cement companies across the country using email and postal mail, and directed to the energy and environment managers, who were asked to answer the questions on behalf of their companies via consultation with the appropriate staff, i.e., the financial managers. Replies deemed useful for analysis were received from 78 companies, giving a response rate of 28.9%. The samples came from a total of 18 provinces and municipalities directly under the central government, of which 35 respondents were based in coastal regions. The number of samples from the Central, Northeast and Northwest China was 21, 10 and 7 individually, and the remaining 5 samples were from Southwest China and municipalities such as Beijing and Tianjin. Overall, the samples were representative in terms of geographical location. The distribution of samples by the number of employees and company ownership is listed in Table 2. Despite the small sample size, the 78 companies control 200 cement kiln lines, with a production capacity of 2500 to 7000 tclinker/d and are sufficiently representative of the overall level of production processes in China's cement industry. According to CCA statistics, 1715 NSP lines were in operation in 2013 with a production capacity of between 2000 and 8000 t-clinker/d, representing 90% of the total domestic capacity in the same year [22]. By number of employees, more than half the samples (43, with a share of 55.1%) had staffs of 300 to 1000 and 32.1% respondents had 100 to 300 employees. Nine samples (11.5%), had staffs of over 1000. By ownership, 53 companies (67.9% of the total) are state owned, with

9 privately owned (11.5%). Foreign-funded companies (fully foreign-funded or joint-venture) account for 18.0% of the total. By sales amount in 2013, three of the samples (3.9%) had sales of between 3 and 20 million Yuan. Companies with sales from 20 to 400, and above 400 million Yuan are 43 and 32, respectively accounting for 55.1% and 41.0%. Based on the classification criteria of the NBSC (National Bureau of Statistics of China) (large companies: over 2000 employees, annual turnover of over 300 million Yuan and registered capital of over 400 million Yuan), most samples in this survey can be categorised as SMEs (small and medium-sized enterprises). 6. Results and discussions 6.1. Energy consumption status of the samples Fig. 3 depicts the distribution of samples by total energy use. All 78 respondents checked the scale of their energy use in 2013, and 22 (28.2%) had an energy use below 100 thousand tce. The numbers of samples with energy use amounts of 100e150 and 270 or more thousand tce nearly equate, for a share of nearly 25% each. Meanwhile, around 10% of the samples used 150 to 200 thousand tce in 2013 and the number of companies with energy use of 200e270 thousand tce is 11 (14.1%). This confirms that the samples are large energy-consuming companies and distributed evenly by energy use amount in 2013. The cement companies were requested to indicate the types of energies and ratio thereof in terms of total energy use. The 66 responses received provided detailed information and show that coal is the dominant energy source for cement companies in this survey, with an average total energy use share of 82.4%. The second source is electricity, with a share of 16.9%, and oil had a share of 0.7% as a minor source. These results agree with the energy use structure of China's cement industry as a whole. In 2012, the total energy consumption of China's cement industry was 207 million tce (tonnes of standard coal equivalent), of which coal consumption was 208 million tonnes and electricity use was 168 billion kWh (with a share of 96.6% for both) [22]. This confirms the samples are sufficiently representative of the sector. Fig. 4, which shows the distribution of samples by share of energy cost in total sales, reveals that energy cost is high for the surveyed companies. Only 2.6% have energy cost ratios below 20%, presumably because they obtain coal at lower than average coal prices. Less than 10% of the samples had energy cost ratios between 20 and 30% or 30e40%, whereas over half of the samples (57.7%) had ratios of over 50%. The 40e50% range is accounted for by 23.1% of the companies. Overall, 80% of the sampled companies had an energy cost ratio of over 40%, which shows that the energy intensity of China's cement industry is high, a result similar to that of Ref. [51]; who documents that coal and electricity cost usually accounts for over 50% of the production cost of cement companies. As a key question, the companies were asked to provide prices for their major energy sources. Statistical results are listed in Table 3, based on prices acquired from nearly all the respondents for electricity and coal use. The average electricity price is 0.614 Yuan/kWh, the lowest is 0.38 Yuan/kWh and the highest is 1.00 Yuan/kWh. The average price of coal is nearly 520 Yuan/t. Oil is a minor energy source for cement companies and prices obtained from 51 samples gave a mean of nearly 7000 Yuan/t. The surveyed companies were requested to give subjective evaluations on current energy prices. All 78 samples provided answers. Of these, three (3.8%) indicated it as very high, 56 (71.8%) as high, and 17 (21.8%) as moderate. In summary, cement companies feel highly pressured by energy prices, which may mean that they are more likely to adopt energy saving technologies.

X. Liu et al. / Energy 106 (2016) 73e86

79

Table 2 Distribution of the respondents from cement industry. Number of employees

Ownership

Category

Number

Percentage

Category

Number

Percentage

Below 100 100e300 300e1000 Above 1000 In total

1 25 43 9 78

1.3 32.1 55.1 11.5 100.0

State-owned Domestically private Joint-venture Fully foreign-funded Others In total

53 9 13 1 2 78

67.9 11.5 16.7 1.3 2.6 100.0

Fig. 3. Distribution of the samples by total energy use in 2013 (N ¼ 78, unit: 10,000 tce).

Fig. 4. Distribution of the samples by share of energy cost in total sales (N ¼ 78).

Similarly to Liu et al. [25] who targeted SMEs in Taicang, China, the surveyed cement companies in the present research had gone to great lengths to save energy via managerial practices. Further, most samples had set up energy saving targets. Specifically, 50 (or 64.1% of the total), had annual energy saving targets; 42 (53.8%) had short-term targets of 3e5 years; and 16 (20.6%) had medium and long term targets of 5e10 years. As regards internal monitoring and statistics, 75.6% of the samples confirmed they had established a relatively comprehensive energy use statistics system, with major energy-consuming equipment and processes under measurement. Nearly 80% of the companies had established a specific division and assigned staff for internal energy management. Fig. 5 depicts the distribution of samples by GHG emissions in 2013. As mentioned in Section 4, around 40% of CO2 emissions of cement production is due to fuel use, and over 50% is due to the decomposition of raw materials like limestone. Therefore, the ranges of GHG emissions provided for the samples to choose from are similar to the energy use ranges in Fig. 3. All 78 samples provided answers, which showed that 32 (41.0%) are not aware of their actual emissions due to the lack of measurement and statistics of GHG emissions within their companies. Overall, the surveyed companies were confirmed to be large GHG emitters: 18 (23.1%) had emissions of less than 0.7 million t-CO2 in 2013, with emissions figures of 0.7e0.9, 0.9e1.5, 1.5e2.0 and over 2.0 million t-CO2 for 10 (12.8%), 6 (7.7%), 3 (3.8%) and 9 (11.5%) of the companies. Excluding unclear responses for emissions, the distribution of samples by GHG emissions is roughly consistent with the result by total energy use, particularly for the samples with emissions of over 0.7 million t-CO2. This goes some way to corroborating the consistency of the data collected in this survey. In stark contrast with the target setting for energy saving demonstrated by the samples, most had no targets for carbon mitigation. In detail, 15 companies (19.2%) had no target or plans for the target setting; 36 (46.2%) had no targets but are planning to set

11.5% 3.8% Unclear

Table 3 Statistics of energy prices of the surveyed cement companies.

7.7%

41.0%

Below 70

Energy type

Unit

Obs.

Mean

Std. dev.

Min.

Max.

70-90

Electricity Coal Oil

Yuan/kWh Yuan/t Yuan/t

77 74 51

0.614 519.7 6970.3

0.107 114.5 1843.3

0.38 120 4000

1.00 730 17,000

90-150 150-200

12.8%

Above 200

6.2. Energy saving and low carbon management of the companies Several questions were listed in the questionnaire in order to gauge the current status of energy saving and low carbon management of the companies.

23.1% Fig. 5. Distribution of the samples by GHG emissions in 2013 (N ¼ 78, unit: 10,000 tCO2).

80

X. Liu et al. / Energy 106 (2016) 73e86

them; 23 (29.5%) actually had targets; and 18 (23.1%) had targets in absolute emissions reductions. China's cement companies do not view LCT innovation as a strategic business chance and are not active in the R&D (research and development) of the related technologies. This survey indicates that 35.9% of respondents only allocate less than 5% of their R&D budget for LCT innovation; 21 (26.9%) allocate 5e10%. In total, over 60% of the samples utilise less than 10% of their R&D budget for the development of LCT. 6.3. Awareness of companies of energy saving and low carbon technologies As mentioned earlier, 14 items of energy saving and low carbon technologies in the cement industry, as indicated in Fig. 2, were listed in the questionnaire to qualitatively measure to what extent the sampled companies are aware of them. A five-point scale was applied, with ‘5’ ¼ very aware; ‘4’ ¼ aware; ‘3’ ¼ moderate awareness; ‘2’ ¼ low awareness; and, ‘1’ ¼ not know at all. The average scores for company technology awareness are depicted in Fig. 6. Overall, the respondents indicate good awareness of the major energy saving and low carbon technologies in their sector; an average of over 4.0 was given for over half of the 14 listed technologies. They are individually T01 (Alternative fuels use), T02 (Alternative raw materials), T03 (Vertical roller mills for raw materials, coal and cement), T04 (Roller press grinding system), T05 (Energy efficient cement clinker calcination), T07 (Combustion system improvement), T08 (High efficiency grate cooler) and T09 (Waste heat recovery power generation). Of these, T03 and T09 achieved the highest means of 4.62. T02, T04 and T07 gave a mean of around 4.50. In summary, this reveals that the surveyed companies are highly converse with the energy efficient mill and

grinding system to save electricity, and the WHR system to recover waste heat from kiln exhaust gases. This may be partly attributed to the wide application of these energy saving technologies in China's cement industry [12]. Four technologiesdT06 (Preheater-precalciner technology with high ratio of solid to gas), T10 (Energy saving monitoring and optimisation system), T11 (Oxygenenriched combustion) and T12 (Co-processing of wastes, e.g., sludge and waste)dachieved a mean of over 3.50. In comparison, T13 (Production of high belite cement) and T14 (Carbon capture and storage) received much lower scores for technology awareness, with a mean of below 3.00. Fortunately, the surveyed companies were highly aware of the three target technologies, T09, T10 and T11. This enabled them to provide detailed information on the adoption of these technologies and made this analysis possible. 6.4. Historical adoptions of the target technologies and the diffusion simulation results Three technologies, WHR system, EMOS technology and Oxygen-enriched combustion, were targeted in this survey. As several WHR facilities may be installed in a single cement company, information on each WHR system was requested, including adoption time, power generation capacity, initial investment, annual power generation amount and operation cost. For the other two technologies the companies were asked to solely provide information on the adoption time, initial investment and annual energy saving. As explained in Section 3, the collected data on technology adoption was compiled, and the technology adopted by all the samples in the same year was amassed to achieve figures for chronological adoption. Technology diffusion rates were calculated using the accumulated adoption figures and the maximum potential. Each company may establish only one EMOS and the total number of samples is used as the maximum potential for this

2.27

T14: Carbon capture and storage (CCS)

2.73

T13: Production of high belite cement T12: Co-processing of wastes, e.g., sludge and waste

3.67

T11: Oxygen-enriched combustion

3.68

T10: Energy saving monitoring and optimization system

3.97 4.62

T09: Waste heat recovery power generation T08: High-efficiency grate cooler (The fourth generation)

4.29

T07: Combustion system improvement (i.e., multitunnel burner)

4.50

T06: Preheater-precalciner technology with high ratio of solid to gas

3.56 4.09

T05: Energy efficient cement clinker calcination T04: Roller press grinding system (For raw materials and cement)

4.45

T03: Vertical roller mills (For raw materials, coal and cement)

4.62

T02: Alternative raw materials

4.54

T01: Alternative fuels use

4.05

2.0

2.5

3.0

3.5

4.0

4.5

Average Score (5 point scale) Fig. 6. Awareness of cement companies on energy saving and low carbon technologies (N ¼ 78).

5.0

X. Liu et al. / Energy 106 (2016) 73e86

technology. Conversely, a cement company may have several kilns and one or more WHR system within the company, therefore the largest number of WHR units installed in the sampled companies is defined as the maximum adoption potential. This calculation is based on an assumption that the target technologies would diffuse evenly throughout the companies at different scales. Using the data obtained and simulation model in Eq. (3), diffusion curves of the technologies were produced. Fig. 7 depicts the results for WHR system and EMOS technology. Oxygen-enriched combustion was omitted since its adoption by the companies in this survey was nearly zero, thus the amount of collected data was too low to enable simulation. As shown in Fig. 7, the R-squared of the regressions for the WHR system and EMOS technology are respectively 0.9862 and 0.9869, indicating a good fit between the observed data and regression curves. As described in Section 4, these two technologies were adopted by the surveyed cement companies in different years. The WHR system was firstly applied in the samples in 2001, accelerated rapidly after 2007 and had reached 80% by 2014. This result is backed up by similar statistics for WHR systems adopted within China's cement industry as a whole [17], confirming the data in this survey is sufficiently representative and reliable. On the other hand, the EMOS technology installed in the surveyed companies took place in 2009, and the diffusion rate thereof in 2014 was 11.5%, implying a large potential remained for its future adoption in the cement industry. Fig. 7 shows that the market for the WHR system is predicted to saturate soon before 2020, while full diffusion of EMOS in the cement industry occurs after 2025 if the adoption of the technology were to increase at its historical pace. This result is consistent with previous studies such as Pizer et al. [43]; confirming that full diffusion of industrial energy saving technologies would be realised 10e20 years after the slow introduction phase. 6.5. Possibility of companies to invest in energy saving and low carbon technologies The possibility of companies to invest in technologies with various levels of profitability was measured by the MBDC format in

81

Table 1. The internal consistency of this measurement was checked via Cronbach's alpha calculation, which gave an alpha of the sample answers of 0.9287, which is over 0.70, the threshold recommended by Nunnally and Bernstein [42]; confirming the scale reliability of this data construct. 6.5.1. Statistics of investment possibility of the surveyed companies Table 4 lists the statistics of possibility for the companies to invest in the technologies with various payback periods. In this analysis, possibilities of ‘very low’, ‘low’, ‘moderate’, ‘high’ and ‘very high’ refer to investment likelihoods of 0.01%, 25%, 50%, 75% and 99.9%, respectively. The number of usable samples is 62. With the shortest threshold of 0.25 years (3 months), 95.2% of the respondents indicated ‘very high’ and the remaining 4.8% selected ‘high’. The share of samples with selections of ‘very high’ dropped to 64.5% for a payback time of 1.0 year, and then dropped sharply to 21.0% for 2.0 years. Accordingly, the share of samples indicating ‘high’ increased to 50.0% at this threshold. Only 11.3% of the companies believed that they would highly possibly invest in a technology with a payback time of 3.5 years. Investment possibility continues to drop with increased technology payback time. Nearly half the samples thought a payback time of 4.0 years to be too long and selected ‘low’ and ‘very low’. A total of 79.0% of the samples would not invest in a technology with a payback time of 5.0 years. Fig. 8 depicts the aggregated data listed in Table 4 and the corresponding simulation curves. Two groups of data, very high and high possibility, and moderate possibility and over, are shown since they are meaningful for observing the range of payback time for the samples to decide on technology investment. A reversed cumulative normal distribution model was applied for the regressions using shares of the samples as dependent variables and the payback time as the independent variable. The R-squared for regressions of the two sets of data is 0.9995 and 0.9981, confirming a good fit between the observed data and regression curves. Payback time on the part of 50% of the samples corresponds to 2.5 and 3.9 years on the two curves, which implies that the mean payback time of technologies for the samples to invest in is within this range. The payback time threshold calculated in this survey is slightly longer

Fig. 7. Historical adoptions and regression curves of the target technologies.

82

X. Liu et al. / Energy 106 (2016) 73e86

Table 4 Statistics of investment possibility responses of the surveyed companies (N ¼ 62). Payback time (year)

Very low (%)

Low (%)

Moderate (%)

High (%)

Very high (%)

Total (%)

0.25 (3 months) 0.5 (half a year) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 5.0 6.0 7.0 8.0 9.0 10.0

0.0 1.6 1.6 3.2 11.3 16.1 17.7 29.0 29.0 40.3 50.0 61.3 82.3 91.9 91.9

0.0 0.0 0.0 3.2 6.4 8.1 16.1 9.7 17.7 38.7 40.3 33.9 16.1 8.1 8.1

0.0 0.0 4.9 11.3 11.3 27.4 37.1 50.0 46.8 19.4 9.7 4.8 1.6 0.0 0.0

4.8 17.7 29.0 48.4 50.0 38.7 21.0 11.3 6.5 1.6 0.0 0.0 0.0 0.0 0.0

95.2 80.7 64.5 33.9 21.0 9.7 8.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Fig. 8. Possibility of the samples to invest in technologies with different payback time (N ¼ 62).

than that for small and medium-sized plants in the U.S. to invest in energy efficiency projects, as confirmed by Anderson and Newell [2]: over 98% of the plants have estimated payback thresholds of less than 5.0 years, about 79% have payback thresholds less than 2.0 years, and the mean payback threshold is 1.4 years. On the other hand, the mean payback time in this research is shorter than the threshold applied in the Dutch voluntary agreements on energy efficiency, which require to plan and implement all the measures with a payback period of less than 5.0 years [1].

calculating infinity. Table 5 lists the mean and percentiles of payback time for the samples to make investment decisions. The resulting mean payback time is 3.3 years, based on a range of 2.5e3.9 years, as observed from Fig. 8. The samples' standard variance of technology payback time is 1.8 years and the medium value for the companies is 3.5 years.

6.5.2. Estimations of payback time for individual companies to invest The mean and standard variance of payback time for each company to invest in energy saving and low carbon technologies were estimated by Eq. (5). Numerical likelihood values of investment were assigned to the verbal expressions in the MBDC format. A ‘very low’ answer was given a probability of 0.1%, as a value of zero would generate infinity in the estimation. A response of ‘low’ possibility was given a value of 25%, ‘moderate’ 50% and ‘high’ 75%. ‘Very high’ gave a value of 99.9% for the same reasondto avoid

To clarify the factors for cement companies to invest in energy saving and low carbon technologies, a question in this survey requested the respondents to indicate to what degree certain factors prevented them from the corresponding investments. As with Liu et al. [27]; 10 factors were listed to enable subjective evaluations thereof by the companies, which are classified into three categories: general factors related to decision-making of companies in investment; financial constraints, and the factors related to uncertainties of technology quality, price change, and policy requirement changes. A five-point scale was applied for the factor influence evaluations, with

6.6. Factors influencing investment in energy saving and low carbon technologies

X. Liu et al. / Energy 106 (2016) 73e86 Table 5 Distribution of estimated payback time for individual companies to invest (N ¼ 62). Variable

Percentile

Centile

95% confidence interval

Mean of payback time: 3.3 years The std. dev. of payback time: 1.8 years

10 20 30 40 50 60 70 80 90

1.4 1.9 2.3 3.0 3.5 3.8 3.9 4.2 5.0

1.0 1.4 1.8 2.2 2.9 3.3 3.8 3.9 4.4

1.8 2.4 3.1 3.6 3.8 4.0 4.3 4.9 5.8

‘5’ ¼ very high influence; ‘4’ ¼ high influence; ‘3’ ¼ moderate influence; ‘2’ ¼ low influence; and, ‘1’ ¼ no influence at all. The average scores of the listed 10 factors are depicted in Fig. 9. In general, cement companies presented low to moderate scores to all 10 factors. This result echoes the finding from a previous survey targeting SMEs from energy intensive industries in China [27]. Comparatively, uncertainty in the quality of technologies was identified by the surveyed cement companies as a factor with higher importance in hindering their technology investment. Accordingly, F07 (Uncertainty in the quality and reliability of new technologies) gave the highest mean of 3.68. The samples also confirmed that certain other factors also moderately affected them in determining whether they would invest in energy saving and low carbon technologiesdsuch as F01 (Other investments are more important), F03 (Current installations are efficient enough), F05 (Internal constraint of the budget), F06 (Difficult for external financing) and F10 (Technology to invest may not satisfy future requirements on energy efficiency). These five factors gave average scores of around 3.50. The other factors were similarly evaluated by cement companies as minor determinants: F02 (Energy cost and efficiency has low priority), F04 (Difficulty due to internal

83

management), F08 (Technology will become cheaper) and F09 (Better to wait for governmental subsidies) achieved low scores between 2.50 and 3.00 on average. As shown in Fig. 9, the sampled cement companies had made efforts in energy saving, as confirmed in Section 6.2, and thus had attained a level of competence in terms of energy efficiency of their current facilities. Some of them may have difficulties in seeking financial resources, internally and externally, for further investments in energy saving and low carbon technologies. The companies also appear unsure about future trends in China's policies but also appear to be anticipating stricter requirements in energy efficiency and carbon mitigation being imposed on the cement industry, which may be confirmed by the answers of other related questions in the survey. Of the total 78 samples, 51 (65.4%) believe they will be covered by nationwide GHG ETS to be introduced in the near future, which implies that providing financial subsidies may be supportive in leading the investments of cement companies towards better energy efficiency [27]. The dissemination of technology-related information may be useful and necessary in order to overcome any concerns cement companies may have as regards technology uncertainty and thus promote investment. The industrial association (CCA), as the sector representative, therefore needs to play a bigger role in this regard [25]. 6.7. Calculation results of carbon pricing for promoting the diffusion of target technologies The core purpose of this study is to estimate the effects of attaching a price to carbon in order to promote diffusion of the target technologies in China's cement industry. As described in the introduction, experts at China's ministry research institutes have suggested initiating a carbon tax policy, with a price from 10 Yuan/ t-CO2 and rising to 40 Yuan/t-CO2 some years later [28]. Liu et al. [29] suggested that 10e30 Yuan/t-CO2, with tax relief to energy-

UNCERTAINTY

F10: Technology to invest may not satisfy future requirements on energy efficiency

3.45

F09: Better to wait for government subsidies

3.10

F08: Technology will become cheaper

2.97

FINANCIAL

F07: Uncertainty in the quality and reliability of new technologies

3.68

F06: Difficult for external financing

3.41

F05: Internal constraint of the budget

3.49

GENERAL

F04: Difficult due to internal management

2.49

F03: Current installations are efficient enough

3.54

F02: Energy cost and efficiency has low priority

2.36

F01: Other investments are more important

3.46

2.0

2.5

3.0

3.5

Average score (5 point scale) Fig. 9. Determinant factors for company's investment in energy saving technologies (N ¼ 78).

4.0

84

X. Liu et al. / Energy 106 (2016) 73e86

intensive or energy efficient sectors, earmarking the revenues for climate change countermeasures and starting as early as the 13th FYP period (2016e2020), would be more preferable for Chinese businesses. The tax rate could be raised to 30e50 Yuan/t-CO2 in a few years but would not approach 100 Yuan/t-CO2 since this level would likely be unacceptable for China's industry. The GHG ETS pilots in seven regions of China indicate that the market price of carbon emissions allowances ranges from around 20 Yuan/t-CO2 in Hubei province to nearly 70 Yuan/t-CO2 in Shenzhen [6]. Referring to the above, three policy scenarios were assumed, with the price of carbon emissions respectively set at 20, 60 and 100 Yuan/t-CO2, abbreviated as S1, S2 and S3. The price level of S1 is low and similar to the carbon tax rate suggested by the experts and would be acceptable for the industry. S2 is a moderate price but close to the upper limit of prices in the pilot markets of GHG ETS in China. Using Eq. (6), energy use ratios and energy prices of the respondents in this survey, and emission factors by energy type, the assumed carbon prices were converted to average energy price increases for cement companies. The changes in payback time of target technologies under the assumed policy scenarios from the BAU case were then calculated. The BAU payback time of the WHR system was obtained from official documents issued by NDRC. In the appendix of the ‘Promotion Catalogue of National Key Energysaving Technologies (1st Batch)’, the technical report provided information on the investments, benefits and payback periods of typical application cases of WHR in the cement industry, and thus was referred to [37]. The payback period of EMOS technology in the BAU case is calculated by the data on the initial cost, operation cost and annual energy saving amount gathered in this survey. Based on the payback times in the BAU and the policy scenarios, the investment possibility increases of target technologies in different carbon prices from the BAU were estimated according to the information from Fig. 8. The calculation results are listed in Table 6. The calculation results of WHR system and EMOS technology diffusions under assumed carbon prices are depicted in Figs. 10 and 11, respectively. As listed in Table 6, the payback time for WHR in the BAU case is 3.0 years, which corresponds to an investment possibility of 28.4% in Fig. 8. The payback time of this technology would drop to 2.6 years if the highest carbon price of 100 Yuan/tCO2 were levied. The increased possibility of technology investment from the BAU would be 14.5%. Since the WHR system has almost reached saturation point within China's cement industry, the diffusion rates of this technology would only slightly change under the three policy scenarios. This reveals the only marginal effect of levying carbon prices for promoting the diffusion of technologies that have been largely diffused, like the WHR system in this study. Nevertheless, carbon pricing indicates some efficacy in promoting the diffusion of EMOS technology in this analysis: levying a

Table 6 Changes of technology payback time and investment possibility in various carbon prices. Technology

WHR system EMOS technology

Payback time (year)

Investment possibility increase from the BAU (%)

BAUa

Carbon price (Yuan/t-CO2)

3.0 1.7

Carbon price (Yuan/t-CO2) 20

60

100

20

60

100

S1

S2

S3

S1

S2

S3

2.9 1.6

2.8 1.4

2.6 1.3

2.9 2.9

8.8 7.0

14.5 10.1

a BAU payback time of WHR is from Ref. [37]; EMOS payback time in the BAU case is calculated by the survey data.

carbon price of 20 Yuan/t-CO2 would generate a 2.6% increase in technology diffusion rate from the BAU level in 2015, a 6.3% increase in S2 (with a price of 60 Yuan/t-CO2) and 9.2% under S3 (with a price of 100 Yuan/t-CO2) in the same year. Over the years, the diffusion rates under the three policy scenarios from the BAU case would at first increase and then decrease. By 2020, the increased technology diffusion rates would be individually 7.4%, 15.9% and 21.1%, which implies that EMOS technology would be fully adopted by around 2030 under the BAU, and that the pricing of carbon emissions could bring about the full diffusion of this technology some time earlier. In particular, levying a moderate carbon price at 60 Yuan/t-CO2 would generate an effect similar to a price as high as 100 Yuan/t-CO2 in order to bring about market saturation of EMOS technology in China's cement industry by around 2025. 7. Conclusions This study targets the cement industry of China and estimates the effects of carbon pricing in promoting the diffusion of the selected technologies. The analysis confirms that the sampled cement companies have a good awareness of major energy saving and low carbon technologies in the sector and have made efforts in energy saving, but are somewhat lagging in carbon managementdthe setting of carbon mitigation targets, keeping carbon emissions records and in the innovation of related technologies. Applying the MBDC format, the possibility of companies investing in the technologies was measured. On average a payback time of 3.3 years would be the decision threshold at which the samples decide to invest in the technologies. In terms of actual practice, the surveyed companies would adopt the target technologies at different stages, and the pricing of carbon emissions would be ineffective in promoting further adoption of technologies that are already well diffused, like the WHR system in this study. Conversely, for EMOS technology, which is still in its early stages in China's cement industry, applying a moderate carbon price may generate comparatively significant effects in its diffusion, which has implications in the sphere of policymaking. This analysis suggests that systematic, sector-specific approaches would be more effective in diffusing energy saving and low carbon technologies, and, based on China's focus on use of administrative mechanisms, advocates the early introduction of carbon pricing policies in order to redirect business efforts into low carbon management. It was further noted that companies behave differently for technologies with different characteristics. Future policies therefore need to focus on specific technologies, and for those still at the initial stages of diffusion, increased demonstration and application are required, for which public finances could be used. There exist several limitations in this study. The questionnaire survey relied on self-reporting from company managers, which could introduce bias associated with the data collection; due to difficulties in obtaining cooperation from individual companies only three technologies were focused on, and the number of samples collected for analysis was limited; and bias may be introduced as regards generalisation of the results due to the limited technology scope and small sample size. Future studies would close any gaps by expanding the analysis to a wider range of technologies and samples as the policy effects may vary for different technologies and companies. A further direction research could take is in applying the results of technology diffusions for quantitative estimations of carbon mitigation and the investment required under the assumed policy scenarios. Such efforts may lead to a better understanding of policy and technology solutions in order to realise the climate goals of the target sectors.

X. Liu et al. / Energy 106 (2016) 73e86

85

Fig. 10. Diffusion of WHR system in various policy scenarios.

Fig. 11. Diffusion of EMOS technology in various policy scenarios.

Acknowledgements The authors wish to thank Hyogo Prefectural Government of Japan for providing financial support for this research. We also appreciate the comments of Prof. Hidefumi Imura and Dr. Mark Elder at the Institute for Global Environmental Strategies (IGES) during the preparation of this paper. References [1] Abeelen C, Harmsen R, Worrel E. Implementation of energy efficiency projects by Dutch industry. Energy Policy 2013;63:408e18. [2] Anderson ST, Newell RG. Information programs for technology adoption: the case of energy-efficiency audits. Resour Energy Econ 2004;26:27e50.

[3] Barreto L, Kemp R. Inclusion of technology diffusion in energy-systems models: some gaps and needs. J Clean Prod 2008;16(1-S1):S95e101. [4] Brown MA, Chandler J, Lapsa MV, Sovacool BK. Carbon lock-in: barriers to deploying climate change mitigation technologies. ORNL/TM-2007/124. Oak Ridge TN: Oak Ridge National Laboratory; 2008. [5] Cagno E, Trianni A. Evaluation of the existing barriers for energy efficiency in industrial operations: a study within the north-Italian manufacturing sector. In: Proceedings of the 16th Annual International Sustainable development Research Conference, May 30eJune 1, 2010, Hong Kong; 2010. [6] Fan Y, Wang X, Zhu L. Transaction cost and the cost effectiveness of ETS pilot in China. Presentation at Singapore-China Energy Forum, National University of Singapore, October 2e3, Singapore. 2014. [7] Fuss S, Szolgayov a J, Khabarov N, Obersteiner M. Renewables and climate change mitigation: irreversible energy investment under uncertainty and portfolio effects. Energy Policy 2012;40:59e68. [8] Fuss S, Szolgayova J, Obersteiner M, Gusti M. Investment under market and climate policy uncertainty. Appl Energy 2008;85(8):708e21.

86

X. Liu et al. / Energy 106 (2016) 73e86

[9] Geroski PA. Models of technology diffusion. Res Policy 2000;29(4e5):603e25. [10] Gibbins J, Chalmers H. Carbon capture and storage. Energy Policy 2008;36(12):4317e22. [11] Hall BH, Khan B. Adoption of new technology. New economy handbook. November, 2002. p. 1e37. [12] Harder J. Latest waste heat utilisation trends in cement plants. In: 2nd Global Cement Electrical Power Conference and Exhibition, London, June 4e5, 2013; 2013. [13] Harder J. Developments in the global cement markets. Cem Int 2014;2014(1): 34e41. [14] Hasanbeigi A, Price L, Lu HY, Wang L. Analysis of energy-efficiency opportunities for the cement industry in Shandong Province, China: a case study of 16 cement plants. Energy 2010;35(8):3461e73. [15] Hultman NE, Malone EL, Runci P, Carlock G, Anderson KL. Factors in lowcarbon energy transformations: comparing nuclear and bioenergy in Brazil, Sweden, and the United States. Energy Policy 2012;40:131e46. [16] IEA (International Energy Agency), WBCSD (World Business Council for Sustainable Development). Cement technology roadmap 2009-carbon emissions reductions up to 2050. IEA and WBCSD; October, 2010. [17] IIP (Institute for Industrial Productivity), IFC (International Finance Corporation). Waste heat recovery for the cement sector: market and supplier analysis. June, 2014. p. 23. [18] Jaffe AB, Newell RG, Stavins RN. Environmental policy and technological change. Environ Resour Econ 2002;22(1e2):41e70. [19] Jaffe AB, Newell RG, Stavins RN. A tale of two market failures: technology and environmental policy. Ecol Econ 2005;54(2e3):164e74. [20] Ke J, McNeil M, Price L, Zheng N, Zhou N. Estimation of CO2 emissions from China's cement production: methodologies and uncertainties. Energy Policy 2013;57:172e81. [21] Kemp R, Volpi M. The diffusion of clean technologies: a review with suggestions for future diffusion analysis. J Clean Prod 2008;16(1-S1):S14e21. [22] Kong XZ. Decrease cement industry coal total consumption and reduce carbon dioxide emission. In: Proceeding of China International Cement Conference, China Cement; 2015. p. 17e29 (in Chinese). [23] Kramer GJ, Haigh M. No quick switch to low-carbon energy. Nature 2009;462: 568e9. [24] Lilliestam J, Bielicki JM, Patt AG. Comparing carbon capture and storage (CCS) with concentrating solar power (CSP): potentials, costs, risks, and barriers. Energy Policy 2012;47:447e55. [25] Liu XB, Niu DJ, Bao CK, Suk SH, Shishime T. A survey study of energy saving activities of industrial companies in Taicang, China. J Clean Prod 2012;26: 79e89. [26] Liu XB, Wang C, Zhang WS, Suk SH, Sudo K. Company's affordability of increased energy costs due to climate policies: a survey by sector in China. Energy Econ 2013a;36:419e30. [27] Liu XB, Wang C, Zhang WS, Suk SH, Sudo K. Awareness and acceptability of Chinese companies on market-based instruments for energy saving: a survey analysis by sectors. Energy Sustain Dev 2013b;17:228e39. [28] Liu XB, Suk SH, Yamamoto R. The feasibility of pricing of carbon emissions in three Northeast Asian countries: Japan, China and the Republic of Korea. Policy Brief, Number 29. Institute for Global Environmental Strategies (IGES); April, 2014. [29] Liu XB, Wang C, Niu DJ, Suk SH, Bao CK. An analysis of company choice preference to carbon tax policy in China. J Clean Prod 2015;103:393e400. [30] Lund P. Market penetration rates of new energy technologies. Energy Policy 2006;34:3317e26.

[31] McJeon HC, Clarke L, Kyle P, Wise M, Hackbarth A, Bryant BP, et al. Technology interactions among low-carbon energy technologies: what can we learn from a large number of scenarios? Energy Econ 2011;33(4):619e31. [32] MIIT (Ministry of Industry and Information Technology). Implementation plan for the promotion of low-temperature waste heat recovery power generation technology in new dry process cement kilns (in Chinese). MIIT; January 20, 2010. [33] MIIT (Ministry of Industry and Information Technology). National industrial energy efficiency guideline (2014 Version, in Chinese). MIIT; December, 2014. [34] MIIT (Ministry of Industry and Information Technology). Implementation plan for the construction of energy management centres in building materials companies (in Chinese). MIIT; January 11, 2015. [35] Montalvo C. General wisdom concerning the factors affecting the adoption of cleaner technologies: a survey 1990-2007. J Clean Prod 2008;16(1-S1):S7e13. [36] MOST (Ministry of Science and Technology). Technical report on the transformation and dissemination list of energy-saving pollution reduction and low carbon technologies (1st Batch) (in Chinese). MOST, http://www.most.gov.cn/mostinfo/xinxifenlei/fgzc/gfxwj/gfxwj2014/201403/t20140320_ 112354.htm; January 6, 2014. [37] NDRC (National Development and Reform Commission). The promotion catalogue of national key energy-saving technologies (1st Batch) (in Chinese). NDRC; May, 2008. [38] NDRC (National Development and Reform Commission). A circular on launching pilot carbon emissions trading (in Chinese). October 29, 2011. [39] NDRC (National Development and Reform Commission). The promotion catalogue of national key energy-saving technologies (5th Batch) (in Chinese). NDRC; December, 2012. [40] NDRC (National Development and Reform Commission). The interim administrative measures for emissions trading (in Chinese). December 10, 2014. [41] NDRC (National Development and Reform Commission). On the basic conditions and working plan to promote the establishment of a nationwide carbon emissions trading market (in Chinese) China Econ Trade Her 2015;(1):15e6. [42] Nunnally JC, Bernstein IH. Psychometric theory. 3rd ed. New York: McGrawHill; 1994. [43] Pizer WA, Harrington W, Kopp RJ, Morgenstern RD, Shih JS. Technology adoption and aggregate energy efficiency. Discussion paper 02e52. Resource for the Future; 2002. [44] Prindle WR. From shop floor to top floor: best business practices in energy efficiency. 2010. Available at: http://www.pewclimate.org/energy-efficiency/ corporate-energy-efficiency-report. [45] Qi Y. Annual review of low-carbon development in China 2011-2012. Beijing: Social Science Academic Press; 2011. [46] Rogers EM. Diffusion of innovations. 5th ed. New York: Free Press; 2003. [47] Schneider M. Technology developments in the cement industry. Cem Int 2014;2014(1):42e7. [48] Tran M. Technology-behavioural modelling of energy innovation diffusion in the UK. Appl Energy 2012;95:1e11. [49] Wang L. Study on CO2 emission of China cement industry. In: Proceeding of China International Cement Conference, China Cement; 2015. p. 48e56 (in Chinese). [50] Zhang SH, Worrell E, Crijns-Graus W. Evaluating co-benefits of energy efficiency and air pollution abatement in China's cement industry. Appl Energy 2015;147:192e213. [51] Zhou HJ. Review of cement price situation of this year (in Chinese). 2008. Available at:, http://www.jszj.com.cn/zaojia/showinfo/showinfo.aspx? infoid¼b8125ca1-478b-44e2-a6cc-48e97191a5fa.