The relationship between climate conditions and consumption of bottled water: A potential link between climate change and plastic pollution

Highlights

• Climate change and plastic pollution linked through effects of climate conditions on household consumption of bottled water.

• The decision of consuming bottled water has two stages: whetherhouseholds purchase it, and then the volume to be consumed.

• Temperature increases both the probability of consuming bottled water and the volume consumed.

• Temperature rises bottled water consumption more in rural areas where tap water is less reliable in quantity and quality.

Abstract

The solution to the challenges that society currently faces, for example climate change and global pollution, requires the understanding of the linkages between these problems. One of these linkages relates changes in climate conditions to consumption of bottled water, which constitutes a source of plastic pollution worldwide. Using information from the labour and households' environmental behaviour surveys and climate records of Ecuador, a geographically fragmented country, we determine how climate conditions affect the decision to consume bottled water and the volume consumed. We methodologically address the problem of selection when households make decisions about this type of consumption, and find that climate variables are important to explain whether and how much bottled water households consume. The temperature-elasticity of the demand for bottled water ranges from inelastic to elastic depending on the model specification. An increase of 1 °C in average temperature is associated, on average, with an increase of almost one-fifth of a water bottle. Temperature increases the consumption of bottled water even more in rural areas and among occupations exposed to climate conditions.

Introduction

Plastic pollution is one of the leading environmental problems that humanity currently faces (Eriksen et al., 2014). Plastic not only pollutes valuable resources, such as oceans and land, but also constitutes a financial burden and a problem for waste management agencies and cities (Chae and An, 2018). Consumption of bottled water, if plastic is not properly recycled, contributes to this type of pollution. The reduction of plastic consumption has become an environmental priority in different parts of the world, where some uses of plastic have even been completely banned (Godfrey, 2019; Clapp and Swanston, 2009). Concurrently, climate change is expected to increase the global average temperature and the frequency of heat waves. Although these two problems could seem not related, the potential link stems from an expected increased demand for water when the average temperature rises, especially for human consumption. Although sustainable alternatives to the consumption of bottled water exist worldwide (e.g., reusable water bottles and an increased supply of water fountains in public and private spaces), bottled water in many parts of the world can be perceived as the only safe source of water for human consumption (Grebitus et al., 2020; Ballantine et al., 2019; Díez et al., 2018). An increasing trend in the consumption of bottled water is being observed (Cohen and Ray, 2018), despite several studies questioning the quality of this type of water (Mason et al., 2018; Timilshina et al., 2012; Pip, 2000). However, higher temperatures and the incidence of heat waves can spur the consumption of bottled water worldwide, especially where heat waves are most expected.

This paper aims to provides evidence on the relationship between climate conditions, namely temperature, and bottled water consumption. We look at the effect of climate conditions on bottled water consumption in Ecuador, a country in the tropics with diverse local microclimates and expected to experience considerable impacts resulting from climate change. Ecuador is an ideal case to study this relationship because of its diversity of climate conditions (i.e., defined by temperature, precipitation and humidity levels), and the perceived quality of tap water. The combined effect of high temperature and humidity levels in some areas of the country reaches wet-bulb temperature¹ values that often go above 35 °C, considered the maximum heat above which human health is at risk (Pal and Eltahir, 2015; Epstein and Moran, 2006). Moreover, people in the country report that the quality of tap water needs improvement (i.e., one-fourth of the population reports that water quality is poor or very poor in our sample).

Previous research has been motivated by a remarkable increase in households' bottled water consumption worldwide (Illsley, 2017; Rodwan Jr, 2017) despite its exorbitant price compared to the price of tap water (Graff Zivin et al., 2011; Kolokytha et al., 2002). These studies show that consumption of bottled water is mostly associated with both objective (Allaire et al., 2019; Wrenn et al., 2016; Graff Zivin et al., 2011) and subjective assessments of the quality and reliability of tap water (Xu and Lin, 2018; Qian, 2018; Huerta-Saenz et al., 2012; Saylor et al., 2011; Jakus et al., 2009; Kolokytha et al., 2002). However, households' socioeconomic conditions influence averting behaviour in response to perceived risks about water quality differently (Allaire et al., 2019). For instance, bottled water consumption is associated to its price (Ward et al., 2009), household income (Johnstone and Serret, 2012; Doria, 2006), taste preferences (Hu et al., 2011; Saylor et al., 2011; Ward et al., 2009), the convenience of bottle sizes and availability (Qian, 2018; Jakus et al., 2009; Ward et al., 2009), the volume of tap water consumed (Roche et al., 2012; Saylor et al., 2011), low trust in authorities (Parag and Timmons Roberts, 2009), personal habits among the youth (Qian, 2018), and the belief that recycling minimizes the environmental impact of plastic pollution (Saylor et al., 2011). Although not necessarily affecting the decision to purchase bottled water, family size and kids' presence at home explain households' adoption of water purification measures (Johnstone and Serret, 2012). Some authors have suggested that increased temperature levels will increase the demand for water for human consumption among daily laborers and farmers in developing countries, where climate change is expected to hit populations the hardest (Dally et al., 2020; Wagoner et al., 2020; Li et al., 2018; Geruso and Spears, 2018; Wesseling et al., 2016; Baez et al., 2017; García-Trabanino et al., 2015). An alternative to satisfy this increased demand for drinking water in areas with no access to tap water is bottled water.

Although the relationship between temperature and water consumption in the residential sector has been studied before (Zapata, 2015), our paper is the first attempt to identify the effect of climate conditions on bottled water consumption. Moreover, most studies on the determinants of bottled water consumption correspond to developed countries (Johnstone and Serret, 2012). To augment the scarce evidence from developing countries (Cohen and Ray, 2018), we study the case of Ecuador, a tropical country where extreme climate patterns will result from climate change. Drawing on the existing literature, we specifically test the following hypotheses: 1) temperature and bottled water consumption are positively related; 2) households consuming and not consuming bottled water have different characteristics; and 3) individuals directly exposed to climate conditions (i.e., living in rural areas and working in agriculture) consume more bottled water. Our work contributes to our understanding of the determinants of this type of consumption and helps anticipate an increase in the consumption of bottled water, and potentially in plastic pollution, due to higher average temperature levels worldwide in the context of climate change.

The estimation strategy employed to identify the effect of climate conditions on bottled water consumption acknowledges the problem of selection. This problem consists of households consuming bottled water being systematically different from households that decide otherwise (Wooldridge, 2010). We compare findings from selection (Two-Part and Heckman) and non-selection (Probit and Tobit) econometric models, and provide evidence that the problem of selection in household consumption of bottled water exists. An attractive feature of selection models is that the household's decision is separated into two steps: 1) the decision about whether to consume bottled water, and 2) the quantity of bottled water to be consumed. This approach allows explanatory variables in the regression model having differentiated effects on each of the decision steps.

Our findings suggest that temperature increases are associated with higher levels of consumption of bottled water. These relationship holds after controlling for factors related to water quality perceptions and measures to improve it at the household level. Moreover, the temperature level is an important variable explaining the household decision process (i.e., whether to consume bottled water and how much of this good to consume). Additionally, bottled water and tap water have a relationship of substitution, shown by a positive value of the cross-price elasticity of demand for bottled water. The results of the two-step estimation procedure also show that some variables explain why households purchase bottled water (i.e., household income and education), whereas others explain the volume consumed (i.e., family size).

This paper is organized as follows. Section 2 presents background information about our study case. Section 3 describes the sources of information, the dataset employed in this paper, and the descriptive statistics of the sample. Section 4 discusses the methodology adopted to study the relationship between climate conditions and consumption of bottled water. Section 5 presents the results, and the last section concludes.