Consumers’ choices for crucial firewood in Saudi Arabia and banning the use of local types

2.1 Theoretical framework of welfare measures

We follow Mitchell and Carson (2013) to explain the conceptual framework for welfare measures. A given policy assessment would incorporate the Pareto improvement criterion, implying that improving some individuals does not harm others. Realistically, we encounter policies with mixed outcomes for better-off and worse-off individuals. Banning the use of local firewood in Saudi Arabia is no exception. The Pareto improvement criterion is still applicable as long as it allows the beneficiaries to compensate for the harmed individuals. Fig. 1 indicates the underlined assumptions of both gainers and losers. The gainers are assumed to be environmentally cautious and voluntarily choose to buy imported firewood. The losers are those with preferences for local firewood exceeding that for environmental benefits and involuntarily buy imported firewood.

Close

Fig. 1

Type of Consumers and the Underlined Assumptions.

Export to PPT

Environmental quality is a public good; individuals do not select the amount consumed, and all enjoy its fixed level. Improving the environment aims to prohibit local firewood. Accordingly, the framework considers the change in welfare as environmental quality improves due to banning local firewood.

Following Mitchell and Carson (2013), the price for the public good is set at zero, and the compensating surplus measures the change in economic welfare responding to the environmental quality change. Fig. 2 depicts the welfare changes of the presumed gainers (the left panel) and losers (the right panel). The downward sloping curves are the Hicksian demand holding the utility level constant, also known as the compensated curves.

Close

Fig. 2

The Change in the Compensating Surplus.

Export to PPT

The magnitude of the Hicksian demand shift for the beneficiaries is larger than for losers. For gainers, an increase in environmental quality from $q_0$ to $q_1$ shifts the Hicksian demand upwards from $h_0$ to $h_1$ providing a higher utility level. For the worsen-off consumers, the Hicksian demand moves downwards from $h_0^{\prime }$ to $h_1^{\prime }$ , reducing the utility level. The latter group is assumed to benefit from environmental quality improvement, but their preferences for local firewood are rigid.

The area a defines the maximum amount the consumer is willing to pay to enjoy the quantity increase and maintain the initial utility level. However, the area c + d indicates the minimum amount the consumer is willing to accept the utility reduction. In mathematical notation, consider the inverse Hicksian demand function,

2.2 Econometric model

This paper aims to determine the effect of the influencing factor on consumers’ choices for imported firewood, contributing to designing better and more effective measures and policies to sufficiently align consumers’ preferences for commanding the control policy to ban local firewood. From the random utility model (McFadden 1974), two firewood alternatives exist, $j=1and2$ for imported and local firewood, respectively. For $N$ consumers, the utility of the individual $i$ corresponding to the alternative $j$ is $U_{\mathit{ij}}$ , $i=1,2,\cdots ,N$ . $U_{\mathit{ij}}$ is a function of observable, $V_{\mathit{ij}}$ , and unobservable, ${\epsilon }_{\mathit{ij}}$ , factors written as $U_{\mathit{ij}}=V_{\mathit{ij}}+{\epsilon }_{\mathit{ij}}$ . The individual $i$ selects the choice $j$ providing the highest utility. Its mathematical expression is $U_{\mathit{ij}}=\max (U_{i1},U_{i2}).$

The stated preference contingent valuation method involves surveying individuals to state their preferences. We asked them to select between local and imported firewood in our case. The response outcome variable, $Y$ , was coded as a binary variable equaling 1 if the consumer favored imported firewood and 0 otherwise. Given the influencing variables, we used Logistic regression (Greene 2012) to estimate the probability of selecting imported firewood. The model follows:

2.3 Survey design

The designed questionnaire directs individuals with alternatives to state the highest utility and to obtain observable characteristics. The questionnaire was written in Arabic and pretested before participants' recruitment.

The sample was collected through days of relatively lower temperatures and higher firewood consumption (from January 30 to February 22, 2021). As the weather cools down, many individuals do recreational activities involving spending time outdoors in deserted areas, thus increasing the demand for firewood.

The sample comprises consumers of firewood, aged at least 18 years old, and said to use firewood within a year. Our questionnaire is web-based, and the snowball method has chosen consumers through emails and varying social media platforms. An online survey would provide higher attendance promptly compared to other survey methods; specifically during the pandemic of Covid19, because health recommendations included taking preventive measures of wearing a mask and limiting gatherings in crowded closed areas, and unnecessary contacts with other people. Although a snowball survey may not provide a representative sample, it is prevalent in consumer behavior studies (Testa et al. 2019; Palmieri et al. 2020a, 2020b).

Local firewood is supposedly absent due to the ban on its extraction, transportation, promotion, and sale. The environmental police actively enforce this ban. Some of the enforcement measures are in Table 1. A revealed preference approach is unavailable to test the factors affecting consumers’ choices for imported vs local firewood due to the absence of observational data. Instead, a stated preference method allows for variations of local and imported firewood choices by utilizing a hypothetical scenario. What follows are the details of the factors influencing choices and the hypothetical scenario.

2.4 Factor influencing choices of imported firewood

The influencing factors fall into four categories. The first group considers environmental attitudes and knowledge. Following (Loureiro et al., 2002), we used a tradeoff scenario to measure attitudes ranging from saving the environment to gaining economic returns and asked the participants to position themselves on a scale in the range of 1–7, where 1 indicated preferring financial returns irrespective of saving the environment, and 7 implied otherwise. This point of reference assured variations within the environmental attitudes metrics (Loureiro et al., 2002). To determine the environmental knowledge, we developed a dummy variable, taking the value of 1 if the consumer considered firewood extraction to pose a risk to the environment and 0 otherwise.³

With newly approved environmental laws assuming less comprehension of the regulations among consumers is sensible. On July 10, 2020, the Royal Decree (M/165) approved the new Environmental Laws, and on February 2, 2021 (the fourth day of our data collection), the Ministry of Environment, Water, and Agriculture approved the new Implanting Regulations for Firewood Extraction (Umm Al-Qura Newspaper, 2021). Instead of directly asking about the familiarity with regulations, we used a proxy as a dummy variable that equaled 1 if participants have seen environmental police efforts to compact local firewood and 0 otherwise. This measure better reflects familiarity with the new regulations because the enforcement news is widespread on various media platforms such as daily press, official governmental websites, and official governmental accounts on social media.

We assumed that environmental attitudes and knowledge would positively affect the demand for imported firewood. So would familiarity with regulations, indicated by the proxy of the environmental police. Although the environmental police are conceptually endogenous (Levitt 2002), they crucially affect consumers’ choices in the short run due to the momentum of the force. The endogeneity is still of great concern, but perhaps in the long run, beyond the course of our investigation.

The second category determines the firewood attributes of aroma, flame color, burning duration, and country of origin, employing a scale ranging from 1 to 3, where 1, 2, and 3, respectively means, not important, less important, and important. The third category refers to consumers’ behavioral habits about consumption purpose, supply method, and primary source of firewood. We used a dummy variable to investigate whether the primary goal was recreational vs personal (for heating or cocking). The recreational purpose is expected to affect the demand for imported firewood positively. Arguably, leisure time in open places and deserted areas in a country like Saudi Arabia having limited forest coverage is more likely to inflate the marginal value for environmental conservation, thus increasing the demand for imported firewood.

The primary sources of purchasing firewood in the country are traditional firewood markets, stores of recreational supplies, supermarkets, and other sources, including street vendors in recreational areas. The traditional markets are mostly located at or adjacent to livestock markets. Before the prohibition, these markets were associated with local firewood compared to supermarkets and stores of recreational supplies. Accordingly, buying logs from traditional markets is expected to negatively influence imported firewood. Contrarily, buying logs from supermarkets and stores of recreational supplies is expected to positively affect imported firewood.

The supply method is also critical for consumers’ choices (Palmieri et al. 2020a). We asked consumers in what form (strapped, bagged, or full pickup truck-load) they buy firewood. The strapped and bagged forms supposedly affect imported firewood positively, but the opposite is true when purchasing a full pickup truck-load of firewood.

The fourth category was about demographics and we added dummy variables for males, with a college degree or above, living in the Riyadh province, and having a high income (monthly household income above (SAR 18,000 [US$ 4,800]),⁴and two more dummies for age between 26 and 40 years old, and age of 41 years or older. Our reference groups had females, education levels below college, low income, and consumers younger than 26 years old. Table 2 depicts the definitions and descriptive statistics of our influencing factors.

2.5 Hypothetical scenario

A well-designed survey produces reliable information (Arrow et al., 1993), but the ban on local firewood could stimulate bias. A hypothetical scenario facing consumers to express their choices between imported and local logs may lead to biased results if responses are on the perception of amending the recently approved jurisdictions. To avoid the potential bias the hypothetical scenario could trigger, we leveraged the published government materials explaining the rationale for banning local firewood and encouraging the use of imported firewood. Doing so would rule out the expectation that the analysis is an ex-ante examination of regulations and align the analysis with our aim to provide policymakers with insights to encourage the use of imported firewood to reinforce the command and control approach.

After perusing the comparisons between local and imported firewood, participants were asked to answer a question on firewood consumption objectively while considering income and other consumption expenditures. The question was the following:

If you bought firewood with the availability of local and imported types at a price of SAR15 (US$4) per bundle, what type of firewood would you buy?⁵

Participants could select among three options: local, imported, or indifferent. We excluded indifferent consumers (98 observations) to center the analysis on those favoring local or imported firewood, which decreased the sample size to 375 observations.

3.1 Descriptive statistics

Table 2 presents descriptive statistics of the variables in the analysis. The sample revealed that the number of consumers choosing local or imported firewood was almost the same. About the environmental attitudes measure, its average is equal to 5.41, suggesting positive attitudes toward the environment. Consumers exhibited a comparatively high ecological knowledge and familiarity with environmental regulations. Seventy-two percent of participants were considering that cutting down trees for firewood posed a risk to the environment. The proxy of environmental police suggested familiarity with firewood regulations, in which 80 % of individuals noted that environmental police tried to prevent extracting, transporting, and selling local firewood.

Save for flame color; all the remaining firewood traits are crucial for consumers. The proportions of those choosing number 3 (important) on our 1 to 3 Likert scale concerning the importance of aroma, flame color, burning duration, and country of origin are 57.60 %, 19.73 %, 65.60 %, and 37.87 %, respectively. Twenty-four percent of consumers purchased logs from stores of recreational supplies, whereas 13 % bought them from supermarkets. About demographics of the surveyed sample, 30 % were high-income, 78 % obtained a college degree or above, 66 % were male, 42 % were aged at least 41 years old, and 82 % lived in Riyadh.

As suggested by the survey earlier, consumers’ environmental attitudes and knowledge, and familiarity with regulations are real and addressable. Nevertheless, about 50 % of consumers still choose local firewood over imported firewood, underlining this study’s indispensability in evaluating the factors, affecting consumers’ choices for imported firewood. These equal splits of preferences among imported and local firewood are also supporting the applicability of developed hypothetical scenario. With the restrictions on local firewood, consumers were selecting between imported firewood and local firewood.

3.2 Econometrics results

Table 3 presents the Logistic regression model results. The model explained 32 % of the variations (Nagelkerke R-square = 0.322). The Nagelkerke R-square was calculated using the Stata command Spsot 13 of Long and Freese (2014). The model correctly predicted 71 % of the probability values, and the overall model fit was statistically significant [ ${\chi }^2\left(17\right)=103.65,P<0.01$ ].

Environmental attitudes $(P<0.01)$ and knowledge $(P<0.01)$ positively influenced the likelihood of choosing imported firewood. However, the importance of aroma $(P<0.05)$ and country of origin $(P<0.01)$ negatively affected the possibility of preferring imported firewood. Shopping from recreational supplies $(P<0.05)$ and supermarket stores $(P<0.05)$ positively increased the chance of selecting imported firewood. Instead of purchasing a full pickup truck-load of logs, buying bagged logs boosted the probability of preferring imported firewood $(P<0.01)$ . Regarding demographics, education exerted a positive but marginal effect on choosing imported firewood $(P<0.1)$ , while living in Riyadh reduced such likelihood $(P<0.05)$ .

Column 3 in Table 3 reports the marginal effects at the means. Consumers are 8 % more likely to select imported firewood given a unit increase in our measure of environmental attitudes. Environmentally knowledgeable consumers are 21 % more likely to buy imported firewood. The chances of choosing imported firewood are likely to decrease by 11 % and 16 %, respectively and the importance of aroma and country of origin increase. Buying bagged logs is associated with 30 % higher odds of selecting imported firewood than purchasing a full pickup truck-load of firewood. Individuals living in Riyadh are 18 % less likely to buy imported firewood, while those with a college degree and beyond are 16 % more likely to choose imported firewood.

The results are in line with our hypothesis. Environmental knowledge and attitudes boost the likelihood of buying imported firewood. These findings also agree with studies suggesting the consumption of environmentally friendly firewood. Palmieri et al. (2020b) reported that ecologically conscious consumers were more likely to pay a higher premium for less polluting firewood. The message for the government as to encourage the use of imported firewood is thus increasing environmental awareness among consumers. This would also require novel means to increase the efficacy of the educational efforts in stimulating the use of imported firewood. One example is leveraging the inconsistency between the high levels of ecological knowledge and imported firewood selection to induce cognitive dissonance that could potentially shift behavior toward saving the environment, by buying imported firewood (Dickerson et al., 1992).

The negative sign for the country of origin’s coefficient shows the decreased probability of choosing imported firewood. One implication for the government to promote imported firewood is to establish the mentality to choose among different types of imported firewood, probably through mandating the provision of labels. Perhaps when brands inform about the export origin, it could form choices on the perception that local logs are no longer an option.

The aroma result suggests exciting implications. Although some attributes of imported firewood, such as burning duration, were better than those of local firewood, the aroma coefficient had a negative sign, reducing the likelihood of purchasing imported firewood. Consumers may not risk buying imported logs with unfamiliar aroma because the aroma attribute is not predetermined by physical sensations such as the attribute of dryness or by using labels. The solutions offered are through a market-based policy to compensate consumers for losing the aroma attributes or a marketing strategy to provide types of foreign firewood with similar aroma.

Obtaining firewood from the primary source presents the expected patterns. Purchasing from recreational supplies and supermarket stores positively increased the likelihood of choosing imported firewood compared to the reference group, comprising buying firewood from traditional firewood markets, street vendors, and others. Before the local firewood ban, the traditional markets and street vendors used to sell local firewood, essentially available in uneven and unorganized logs per unit. However, the imported firewood is mostly available in evenly put together pieces, enabling retailers to display them on their store shelves.

Instead of purchasing a full pickup truck-load of logs, buying bagged or strapped logs, though the latter is statistically insignificant, positively influenced the likelihood of imported firewood consumption. This result underlined the relevance of the supply method to stimulate the utilization of imported firewood. Likewise, using a multilevel Logistic regression model, Palmieri et al. (2020a) revealed that, the willingness to consume the eucalyptus firewood analysis by consumers caring about the supply method were 2.2 times more likely to purchase eucalyptus logs.

The negative coefficient of living in Riyadh lacks a solid justification. Although preferring a region in deterring activities of illegal firewood could form an externality in displacing choices to other areas (Gonzalez-Navarro, 2013; Latham et al., 2017). This locational-based pattern of Riyadh is also worth scrutinizing to reduce any negative spillovers across different locations (Romanach and Frederiks, 2021).

The results are intuitive and robust among different specifications. Column 2 shows that the estimated coefficients from the Logistic regression model with their robust standard errors are almost identical to that of column 1. However, the coefficients for education and buying from recreational supplies became marginally significant at 10 %.

This study is explanatory, and we are cautious about the generalization of our interpretations for two reasons. First, our data collection was during the prime time of the Covid19 pandemic. Second, we could not cross-validate the results with - previous research. As per our knowledge, our study was the first to center on the consumers’ choices of local vs imported firewood based on the new environmental rules in Saudi Arabia, giving insights to policymakers to encourage the use of imported firewood.