The People’s Republic of China (PRC) in February 2005 made one of largest state-sponsored commitments toward renewable energy, as the country adopted the Renewable Energy Law which encompassed directives addressed to the management of solid wastes. By 2013, the PRC was operating 166 WtE plants, converting over 30% of the nation’s Municipal Solid Waste (MSW) to energy [11].

Under Thailand 4.0 [12], renewable energy has been stated to be a key foundation in the quest for the use of innovation in the reduction of imported fossil fuels according to the government’s 2015 Power Development Plan (PDP 2015). Government policy statements have also indicated that fossil fuels are not only economically and ecologically unsustainable, they also expose the Kingdom to the unpredictability of global commodity markets [13].

As a component of PDP 2015, the Thai government also released the Alternative Energy Development Plan 2015 (AEDP 2015), which prioritized power generation from waste, biomass and biogas, increasing alternative energy capacity from 7,279 MW in 2014 to 19,635 MW in 2036 [13].

This is consistent with a World Bank study in which it was stated that because solid waste management is highly visible and affects residents’ perception of government functionality, government and its political representatives are also stakeholders [14]. Research from Nigeria supports this as it was stated that the establishment of a WtE facility was overwhelming (92.8%) viewed as a benefit to the community, when compared to the existing burning of waste in open landfills. The researchers also stated that community acceptability is additionally conditional on the promotion of it through community education, advocacy, and social marketing [15].

This was consistent from a United Kingdom (UK) study in which Pidgeon et al. [16] went on to explain the difficulties of science communication challenges involved when designing and conducting public deliberation processes on future energy system issues of national importance, although resource intensive, national-level deliberation is possible and can produce useful insights both for participants and for science policy.

Park [17] examined renewable energy related regulations, programs and financial incentives in 48 US states existing between 2001 and 2010. From this, it was stated that authoritative approaches are more likely to be effective in the governmental intervention toward a pre-existing market, although information instruments and citizen participation became important in the power industry in the 2000s.

Yi and Feiock [18] also indicated that renewable energy development in the US is influenced by regulatory institutions, party affiliations, legislative professionalism and the policies they implement.

MacArthur [19], however, took a more positive view on Canadian and Danish citizen engagement in policymaking and indicated that it represents an increasingly popular mechanism for both civic rejuvenation and environmental policy innovation. The research stated that it empowered the public and led to the design and implementation of more effective solutions to complex social and environmental problems.

From the above theories and scholars’ concepts of GP, the following three items were therefore placed into the research framework, which included:

• Policy Formulation (PF);

• Policy Implementation (PI);

• Troubleshooting (TR).

From this, the following hypotheses were developed:

• H1: GP positively influences WtE Power Plants;

• H2: GP positively influences CP;

• H3: GP positively influences Creating Shared Value (CSV).

According to Thailand’s Department of Alternative Energy Development and Efficiency [3], in order to promote energy-from-waste production, the government needs to run community campaigns to promote community participation, waste sorting activities, and knowledge sharing with municipals, communities, the general public, and students to enhance understanding of waste management for environment and energy purposes.

This is consistent with research from Sadeghi et al. [5], which determined that one important parameter in increasing incineration plants efficiency is waste sorting at the source, which requires increasing the community’s awareness and change in their attitudes concerning the environment.

In research concerning solid waste disposal in Uganda, it was established that because of the lack of public participation in solid waste management, the best way to start dealing with the problem was for the local government to educate the people concerning the value of proper waste disposal, while also involving them in the initial planning process [20].

From the above theories and scholars’ concepts of CP, the following four items were therefore placed into the research framework, which included:

• Information (IN);

• Listening (LI);

• CP;

• Community Empowerment (CE):

• H4: CP positively influences WtE Power Plants;

• H5: CP positively influences CSV.

Energy recovery from the combustion of MSW is a key part of the non-hazardous waste management hierarchy, which ranks various management strategies from most to least environmentally preferred [21].

According to the USEPA model, from the most preferred method to the least preferred are as follows:

• Source reduction and reuse;

• Recycling/composting;

• Energy recovery;

• Treatment and disposal.

The European Commission [22] modified this model somewhat and discussed:

• Prevention;

• Preparing for re-use;

• Recycling;

• Other recovery;

• Disposal.

It is important to note that for level 3 (recycling), anaerobic digestion of organic waste is suggested where the digestate is recycled as a fertilizer. At level 4 (other recovery), waste incineration and co-incineration operations with a high level of energy recovery along with the reprocessing of waste into materials that are to be used as solid, liquid or gaseous fuels is suggested. At the bottom of their upside-down pyramid, is level 5 (disposal), which consists of waste incineration and co- incineration operations with limited energy recovery along with the utilization of captured landfill gas.

WtE facilities can also generate a renewable energy source and reduce carbon emissions by offsetting the need for energy from fossil sources and reduce methane generation from landfills [1], [13]. In the European Union (EU), however, the EU Waste Acceptance Criteria (WAC) introduced strict limits on leaching that are difficult to achieve [23], but in spite of this, waste incineration has grown steadily and since 1995, the amount of municipal waste incinerated in the EU-27 has risen by 32 million tons or 100%, which in 2015 accounted for 64 million tons. Thus, municipal waste incineration has thus risen from 67 kg per capita to 127 kg per capita [24].

Where WtE processes are opted for, there is a need to ensure that the most efficient techniques are used, which thus far include [22]:

• Co-incineration in combustion plants;

• Co-incineration in cement and lime production;

• Waste incineration in dedicated facilities;

• Anaerobic digestion.

• This process includes the upgrading of the biogas into bio-methane for further distribution and use (e.g. injection into the gas grid and transport fuel).

From the above theories and scholars’ concepts of WtE Power Plants, the following four items were therefore placed into the research framework, which included:

• Waste Incinerator Pollution Control (WIPC);

• Ash and Dust Handling (ADH);

• Noise Pollution Control (NPC);

• Waste Water Quality (WWQ).

From this, the following hypothesis was developed:

• H6: WtE Power Plants positively influence CSV.

According to Wójcik [25], CSV is a conceptual response to deficiencies in Corporate Social Responsibility (CSR), with CSV proponents seeing business activity through the value creation in both the economic and social dimensions. Porter and Kramer [26], would agree with this and further stated that instead of companies putting a wedge between their business and society, they could instead create “shared value” by generating economic value in a way that also produces value for society by addressing its challenges. Specifically, firms can do this in 3 ways: by reconceiving products and markets, redefining productivity in the value chain and building supportive industry clusters near their locations.

Therefore, energy poverty is one of the most obvious issues in CSV where energy companies can add value to society. And according to the World Energy Outlook report [8], an estimated 1.2 billion people, or 16% of the global population, do not have access to electricity.

In a report from the Singapore Environment Institute, components of CSV were discussed as the “5 Rs”. These “Rs” are [27]:

• Refuse - Avoid buying unnecessary waste;

• Return - Return packaging materials to suppliers;

• Reduce - Reduce waste at the source;

• Reuse - Reuse everything that is possible;

• Recycle - Recycle any remaining waste streams.

This is consistent with European legislation in which 60% of all packaging put on the market had to be recovered. As a result, the landfilling rate compared with municipal waste generation, in the EU-27 dropped from 63.8% in 1995 to 25.3% in 2015 [24].

From the above theories and scholars’ concepts of CSV, the following three items were therefore placed into the research framework, which included:

• New Product Invention (NPI);

• New Production Norms (NPN);

• Cooperative Groups Development (CGD).

Based on the above hypotheses and review of the literature, the researchers have developed Figure 1 conceptual framework which includes the causal relationships between GP, CP, WtE and CSV.

Figure 1.

Conceptual framework

The research method used a 63-item instrument to assess the four constructs in the CSV model. All questionnaire items used a 7-point agreement scale response format [28]. The questionnaire was administered to 361 individuals who were community residents in five areas in which waste management and power generation facilities were operating. The questionnaire was developed from the literature review and related theory and was constructed as a tool to measure concept definition and practice (Table 1).

Five experts determined the reliability of the questionnaire so as to ensure that the responses collected through the instrument were reliable and consistent. The five experts included the:

• Managing Director of Pracharat Samakkhi Petchaburi (Social Enterprise) Limited;

• Dean of Faculty of Engineering, Southeast Asian University;

• Secretary-General, Association for the Prevention of Global Warming;

• Director, Office of Natural Resources and Environment, Phitsanulok Province;

• Senior Executive Vice President, SPCG Public Company Limited.

A trial assessment of 25 questionnaires was conducted prior to the actual survey to determine questionnaire reliability and consistency. The reliability value was calculated by using Cronbach’s Alpha ( α ) [29] to ensure internal consistency within the items. According to Best and Kahn [30], when interpreting α, it ranges from 0 to 1 and a value of ≥ 0.70 reflects good reliability. According to the pre-test, α averaged 0.836, indicating reasonable reliability [31].

Furthermore, the survey questionnaire was divided into two parts, with Part 1 consisting of four items concerning the community resident’s personal information (Table 2), while Part 2 consisted of the actual questionnaire concerning the resident’s views about the survey items. For this, Part 2 measured 59 items and was divided into four parts, with government policy consisting of 11 items, CP with 16 items, WtE with 12 items, and CSV with 20 items. Scale measurement made use of a 7-level Likert type agreement scale [28], with 1 indicating the resident strongly disagrees with the item’s statement, while 7 indicated the resident strongly agreed with the item’s statement. Therefore, from the seven levels of frequency, the interpretation of these responses was calculated by using the following formula:

which was calculated for seven levels of frequency detailed in Table 2.

To test the proposed research model, the researchers adopted the survey method for data collection, whose hypotheses were examined by use of Linear structural relations (Lisrel) 9.1 for the collected data [32]. Measurement and data collection implies an evaluation of the measurement model, which for the study included:

• The individual item reliabilities;

• The model’s convergent validity;

• Discriminant validity.

Individual item reliability was examined by looking at the loadings, or correlations, of each indicator on its respective construct. For reflective indicators, it is generally accepted that items must have a factorial load (λ) of 0.707 or above and all values have been statistically significant (|t| ≥ 1.96), representing convergent validity of scales. This threshold implies that there is more variance shared between the measures and their constructs than there is error variance. The initial analysis indicated that elimination of some items would enhance the fit indices, with standardized residuals indicating significant cross loadings for several items being deleted if they exceeded 2.0. Reliability for the derived scale scores was also measured via internal consistency coefficient α [29].

Sample size suggestion usually depends on the complexity of the specified model, but typically ranges between 5 to 20 questionnaires per observed variable, with overall sample size preferred to exceed n = 200 cases [33]. Therefore, from the above and other reviewed theory, a ratio of 20:1 was deemed to be reliable. Thus, the study’s sample size of 361 individuals for 14 observed variables (14 × 20 = 280) was deemed to be highly reliable. All surveys were conducted face-to-face from 09.00-20.00 at the resident’s home or local place of business. Deep interviews were also conducted with 10 executive level individuals from 3 April to 1 May 2016.

To access the measurement models, a Confirmatory Factor Analysis (CFA) is used followed by Structural Equation Modeling (SEM) to examine the general fit of the proposed model with data and to identify the overall relationships among these constructs [34]. Wong [35] also noted that for marketing research, a significance level of 5%, a statistical power of 80%, and R² values of at least 0.25 are considered normal. Standard modelling accepts the proposed model if the p value is higher than 0.05, and if the x²/df ratio is less than 2 [36]. This is consistent with Kline [37] and Ullman [38] which also indicated that the relative Chi-square (x²) should be less than two. Additionally, another common reported statistic is to use the Root Mean Square Error of Approximation (RMSEA), as a measure of goodness-of-fit in SEMs [39] and to measure the discrepancy per degree of freedom [40].

Hooper et al. [41] indicated that items with low multiple R² values (≤ 0.20) should be removed from an analysis as this is an indication of very high levels of error. Hair et al. [31], used higher criteria and suggested that the R² values should be greater than 0.25.

Table 4 shows that the factors that affect CSV, which includes:

• WtE Power Plant;

• CP;

• GP.

Interpreted results from the 7-point survey ranged from 4.71–4.83 [28].

CFA analysis of the dependent and independent variables was built on the conceptual framework derived from the study of relevant documents and scholarly research (Figure 2 and Figure 3). By analysing the confirmatory components with the LISREL 9.1 program, x² was determined not to be statistically significant (p > 0.05), x²/df was ≤ 2.00, RMSEA ≤ 0.05 and Standardized Root Mean square Residual (SRMR) ≤ 0.05. The Goodness-of-Fit Index (GFI) was also indicated to be 0.995, which shows good fit as it is higher than the suggested 0.90 by Hooper et al. [41]. Also, the value for the Adjusted Goodness-of-Fit Index (AGFI) was 0.986, which indicates a well-fitting model as its value is also greater than 0.90.

Figure 2.

Confirmatory factor analysis of the external latent variable government policy x² = 0.00, df = 0, p value = 1.000, RMSEA = 0.000

Figure 3.

Confirmatory factor analysis of the internal latent variables x² = 9.92, df = 23,p value = 0.99175, RMSEA = 0.000, SRMR = 0.09, GFI = 0.995, AGFI = 0.986

From the LISREL 9.1 analysis of the data, and the measurement of the four constructs and their hypotheses, it was determined that there was a good model fit with the empirical research data. Also, to assess the validity of a test, convergent validity and discriminant validity were used. In structural equation modelling, CFA is usually used to access construct validity [32]. Hair et al. [31] and Byrne et al. [36] indicated that factor loadings or regression weight estimates of latent to observed variables should have values greater than 0.50, which indicates that all the constructs conform to the construct validity test and validity convergence.

Results in Table 5 show that the x² value was 34.04, which had 44 degrees of freedom (df). Therefore, the ratio between x² and the df was equal to 0.774 when tested, which showed statistical significance as it was ≥ 0.05, which confirms the model’s hypotheses are not different from the empirical data. Further confirmation was established as the results of the GFI equaled 0.987, and the AGFI equalled 0.969 [42]. The RMSEA was equal to 0.000. The SRMR was equal to 0.013. As SRMR is an absolute measure of fit, a value of zero indicates a perfect fit with a value of < 0.05 indicating a good fit [40].

The validated results are detailed in Table 6 and Table 7, as well as Figure 4.

SE = Standard Error, CR = Critical Ratio (t -value)

Figure 4.

SEM final model with values from estimates (n = 361), x² = 34.04, df = 44, p value = 0.860, RMSEA = 0.000, SRMR = 0.013, GFI = 0.987, AGFI = 0.969

Table 8 shows the Direct Effect (DE), Indirect Effect (IE) and Total Effect (TE) of each construct. CSV is influenced by the direct and positive contribution of government policy the greatest, due to the total effect value of 0.83.

Both Hooper et al. [41] and Hair et al. [31] discussed low R² values (≤ 0.20 and ≤ 0.25, respectively) and suggested that they be removed from the analysis as this is an indication of high error rates. The SEM results (Figure 4) showed that the model met the required criteria as the Chi-squared index was not statistically significant, p = 0.86, RMSEA = 0.00, GFI = 0.99, AGFI = 0.97 and SRMR = 0.01. All causal factors in the model were shown to have a positive influence on the shared value of the waste management power plant and the local community, with 68% of the variance of the factor affecting CSV (R²). Ranked in importance, the three latent variables were GP, WtE Power Plant and CP, with a total score of 0.83, 0.37 and 0.36, respectively.