ENERGY STAR Appliance Market Shares: Do They Respond to Electricity Prices, and Does It Matter?

AuthorSchwarz, Peter M.
  1. INTRODUCTION

    In 1992, the U.S. Environmental Protection Agency (EPA) introduced the ENERGY STAR (ES) program as a voluntary labeling program with the goal of identifying and promoting energy-efficient products "to reduce greenhouse gas emissions."(1) In 1996, EPA partnered with the U.S. Department of Energy (DOE) to provide information to consumers about appliances that save a designated amount of energy over otherwise comparable models. The ostensible goal of the program was to reduce carbon emissions by encouraging consumers to pay the higher up-front cost to purchase more energy-efficient appliances. Energy efficiency is often cited as the lowest-cost method of reducing carbon emissions. Yet its effectiveness depends upon the degree to which consumers are willing to adopt such measures.

    While it is difficult to quantify the benefits and the costs of the program, especially in the absence of a counterfactual, several studies, such as Datta and Gulati (2014) and Jacobsen (2015), investigate the market share of ES appliances across the various United States. Datta and Gulati investigate how rebates on ES appliances influence market shares using data from 2000-2006; Jacobsen uses data from 2000-2009 and extends their analysis to include electricity prices, natural gas prices, and personal income.

    Datta and Gulati find that rebates increase market shares of clothes washers, but not dishwashers or refrigerators; they are unable to examine air conditioners as rebates on this appliance were not available during their sample period. They do conclude that the clothes washer rebate is worthwhile by comparing the cost of $140 per ton of [CO.sub.2] emissions saved to the cost of energy efficiency savings in buildings, although it appears there are other efficiency measures available at a lower cost per ton. Jacobsen uses Datta and Gulati's rebate data and updates them through 2009. confirms their result regarding rebates, and using a fixed-effects model finds that electricity prices do not seem to correspond with changes in the market shares of ES appliances. Houde andAldy (2017) find that appliance rebates are not very effective in that approximately 70% of consumers who file a rebate are inframarginal and that an additional 15-20% of consumers simply shifted the timing of their purchase. These customers receive payments for purchases they would have made anyway, a type of free-rider problem.

    The lack of change in ES market shares in response to a change in the price of electricity is the focus of our study because if consumers do not respond to either implicit or explicit changes in the price of energy efficient appliances, ES may not be an effective way to encourage energy efficiency or carbon reductions. Such findings run counter to the law of demand and engineering estimates. According to the EIA, the residential sector in the U.S. accounts for 35% of end-use efficiency potential and will consume 29% of baseline energy in 2020 (McKinsey & Co., 2009). The McKinsey estimates use an engineering approach to calculate the potential for energy savings, and some critics point out that expected gains from investment in efficiency often fail to materialize because of improper installation, user error, unobserved costs, and rebound effects. Yet, the engineering estimates of the potential efficiency gains from reducing residential energy use are too large to ignore.

    If consumer purchases of energy-efficient appliances do not respond to electricity prices or rebates, in defiance of the law of demand, what are possible explanations? One explanation could be the energy efficiency gap, an example of which occurs when consumers apply an above-market discount rate to energy savings and therefore underinvest in energy efficiency. (2) Other explanations of the gap include market failures such as lack of information (Palmer, et al., 2012), principal-agent problems such as whether landlords or tenants are responsible for energy bills, uncertainty about future energy prices, and credit market inefficiencies. Behavioral economics explanations include consumer myopia, bounded rationality, limitations in self-control on the part of low-income consumers (Tsvetanov and Segerson, 2014), and the perception of not being able to capitalize energy efficiency into property values (Palmer, et al., 2012).

    Is the energy efficiency gap large enough to lead to no price response? Do consumers not connect the electricity prices to the consumption of energy through their household appliances? Are electricity price changes sufficiently small that consumers do not find it in their best interest to purchase a potentially more expensive ES appliance?(3) Or are annual data too aggregate to capture demand responses to electricity price changes? Alternatively, might these results be generated by a specific econometric framework and the exclusion of additional relevant variables? We examine the effect of electricity prices on the market share of ES appliances including additional variables such as percentage of owner-occupied housing and education levels and using a random-effects model that estimates both within-state and between-state effects. (4)

    We have three goals in this paper. The first is to replicate in a narrow sense (see Duvendack, et al., 2017; Pesaran, 2003), the main findings of Jacobsen. We confirm that when using a fixed-effects estimator there is no statistically significant response of market shares to within-state changes in electricity prices. This presents the apparent paradox: ES appliances violate the intuition that consumers respond to higher electricity prices by increasing their demand for energy efficiency.

    The second goal is to investigate another source of variation in ES appliance market shares: between-state variation. We do this in the context of Mundlak's (1978) estimator, recently reframed by Bell et al. (2019), which simultaneously estimates the within-state and between-state effects of explanatory variables using a random-effects estimator. We find that within-state effects of electricity prices remain insignificant but that market shares are positively related to between-state differences in electricity prices for three of the four ES appliances investigated.

    The third goal is a calculation of the impact on carbon emissions via the market share of ES appliances of a four-cent increase in electricity prices through a $ 100 per ton carbon tax. Resources for the Future (2010) estimates that an increase in residential electricity rates resulting from such a tax would be on the order of 4 cents/kWh. (5) We find that the impact of the ENERGY STAR program following such a price change would be approximately 2.1 million megawatt hours per year, equivalent to removing approximately 0.11% of all U.S. vehicles from the nation's roads. The marginal effect of the carbon tax is 0.2 million megawatt hours and the equivalent of removing 0.01% of U.S. vehicles from the nation's roads.

    The remainder of the paper is as follows. Section 2 describes the data and replication efforts. Section 3 describes our extension of previous studies to resolve the apparent paradox between ES market shares and electricity prices. Section 4 uses our new econometric results to simulate what might happen to US carbon emissions through ES appliances after a $100/ton carbon tax. Section 5 offers concluding remarks.

  2. DATA DESCRIPTION AND REPLICATION

    The data used by both Datta and Gulati and Jacobsen include publicly available market shares reported by the U.S. Energy Information Administration (EIA) from 2000 through 2009 for four ES appliances: room air conditioners (AC), dishwashers (DW), clothes washers (CW), and refrigerators (RF). (6) Residential electricity prices were obtained from the U.S. Energy Information Administration and represent the mean residential price for the total electric industry. (7) The parameter on this variable is of the most interest for answering the primary question of whether variation in electricity rates among states explains variation in market share of ENERGY STAR appliances. We also gathered state-level per-capita income from the Bureau of Economic Analysis, and the residential price of natural gas from the EIA. (8) We obtained the unweighted state-level annual average rebates for each appliance from Datta and Gulati.

    Table 1 reports the descriptive statistics of the variables utilized in the replications alongside those of Jacobsen; our sample descriptive statistics are essentially the same as those used in Jacobsen and Datta and Gulati. The market shares for ES appliances average 28.88 for clothes washers, 34.54 for room air conditioners. 29.02 for refrigerators, and 59.32 for dishwashers. In states where rebates were offered, state rebates average $51.19 for air conditioners, $53.61 for clothes washers, $36.45 for dishwashers, and $38.29 for refrigerators. Per-capita income averaged $37,670, owner-occupied households averaged 70.22%, and the percent of a state's population with a bachelor's degree averaged 26.25%.

    Our replication focuses on the pooled sample results reported in Table 3 of Jacobsen. Our sample period is the same as his and we use the same variables and specifications as he does. Table 2 reports our replication efforts. We can replicate his results using our sample, including the result that within-state changes in electricity prices have no impact on ES appliance market shares. (9)

  3. RESOLVING THE PARADOX OF ELECTRICITY PRICES AND ES MARKET SHARES

    First, we expand the set of explanatory variables by including the percentage of a state's population with a bachelor's degree, under the assumption that a college education might make an individual more sensitive to problems associated with carbon emissions and that a college education might make one more sensitive to intertemporal comparisons of savings associated with an ES appliance. We also include the percentage of residences that are...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT