|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 COMPETITING OBJECTIVES OF RESIDENTIAL WATER RATE DESIGN IN FLORIDA: AN EM I PRICAL STUDY By COLIN RAWLS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR T HE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2009
2 2009 Colin Rawls
3 To Katie
4 ACKNOWLEDGMENTS I would like to thank the members of my committee. Dr. Sanford Berg and Dr. Jeffery Burkhardt have provided excellent advice and guidance during the duration of this work. I would also like to thank Dr. Burcin Unel. She provided the data set and helped me interpret it. Dr. Tatiania Borisova, my chair, has provided generous feedback and support at every stage of the research process. Wi thout her patience and insight, this project would not have been possible. My parents, Sandra and Eric Rawls, have also provided support throughout my graduate school experience. I am grateful for their love and advice. Finally, I would like thank my wif e, Katherine Sherwood Rawls Her love support has also been crucial during my time as a graduate student.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...................................................................................................... 4 LIST OF TABL ES ................................................................................................................ 8 LIST OF FIGURES ............................................................................................................ 10 ABSTRACT ........................................................................................................................ 12 CHAPTER 1 INTRODUCTION ........................................................................................................ 14 Background ................................................................................................................. 14 General Research Problem: Competing Objectives of Water Rate Design ............. 18 Research Objectives ................................................................................................... 19 The Data ..................................................................................................................... 20 Organization of the Thesis ......................................................................................... 20 2 LITER ATURE REVIEW .............................................................................................. 24 Background ................................................................................................................. 24 Objective 1: Rate Structure and the Strength of Conservation Price Signals .......... 30 Objective 2: Rate Structure, Utility Revenue, and Revenue Variability ................... 35 Objective 3: Rate Structure, Equity, and Affordability ............................................... 3 8 Other Utility Objectives ............................................................................................... 40 3 EMPIRICAL DATA ...................................................................................................... 44 Overview ..................................................................................................................... 44 Description of Utilities ................................................................................................. 44 Relevant Variables ...................................................................................................... 45 Estimating Monthly Water Use ................................................................................... 46 Rate Structure Information Summary Tables ............................................................ 50 4 CONSERVATION PRICE SIGNALS .......................................................................... 60 Background ................................................................................................................. 60 Methods ...................................................................................................................... 61 Results ........................................................................................................................ 62 Average and Marginal Prices .............................................................................. 62 Analysis of Marginal Price ................................................................................... 64 Household Savings from Water Use Reduction ................................................. 65 Revenue Distribution ............................................................................................ 67 Summary ..................................................................................................................... 69
6 5 CONSERVATION RATES, REVENUE, AND REVENUE STABILITY ..................... 73 Background ................................................................................................................. 73 Hypotheses ................................................................................................................. 74 Evaluation Criteria ...................................................................................................... 74 Average Monthly Revenue .................................................................................. 74 Household Billing Variability ................................................................................ 75 Frequency Distribution of Household Bill ............................................................ 75 Regr ession Analysis for Household Bill Variability ............................................ 75 Overall Variability in Monthly Revenue ............................................................... 76 Methods ...................................................................................................................... 76 Average Monthly Revenue .................................................................................. 76 Frequency Distribution of Household Bill ............................................................ 76 Household Billing Variability ................................................................................ 76 Regression Analysis for Household Bill Variability ............................................. 77 Overall Variability in Monthly Revenue ............................................................... 78 Results ........................................................................................................................ 78 Average Monthly Revenue................................................................................. 78 Household Billing Variability ................................................................................ 80 Frequency Distribution of Household Bill ............................................................ 81 Regression Analysis for Household Bill Variability ............................................. 81 Overall Variability in Monthly Revenue ............................................................... 82 Conclusion .................................................................................................................. 82 Future Research Needs ............................................................................................. 83 6 CONSERVATION WATER RATES, AFFORDABILITY, AND FAIRNESS ............... 85 Background: Conflicting Opinions .............................................................................. 85 Methodology: Measuring Affordability And Fai rness Of Water Rate Structures ...... 86 Affordability ........................................................................................................... 86 Fairness ................................................................................................................ 90 Results ........................................................................................................................ 90 Affordability: Evaluating the Hardship Threshold Results .................................. 90 Rate Structures and Inflation ............................................................................... 92 Rate Structure and Geography of Utilities .......................................................... 93 Rate Structure and Demographics ...................................................................... 95 Fairness: Interpreting the Lorenz Curves and Gini Coefficients. ....................... 97 Conclusion .................................................................................................................. 98 Limitations of the Study .............................................................................................. 99 7 CONCLUSIONS AND IMPLICATIONS ................................................................... 101 Competing Objectives of Rate Design: The Big Picture.......................................... 101 Characteristics of Water Rates used by Florida Utilities ......................................... 102 Effectiveness of Water Rates in Stimulating Water Conservation .......................... 103 Revenue .................................................................................................................... 104 Equity and Affordability ............................................................................................. 104
7 Study Limitations ...................................................................................................... 104 Study Contribution .................................................................................................... 106 Recommendations for Future Research .................................................................. 107 REFERENCES ................................................................................................................ 110 BIOGRAPHICAL SKETCH .............................................................................................. 118
8 LIST OF TABLES Table page 1 -2 Trends in Florida Rate Design. Source: edited and expanded from Whitcomb (2005) ...................................................................................................................... 23 2 -1 Water Rate Structures for Residential Water Services: Results from Nation Wide Surveys. Source: AWWA and Raftelis (2006). ............................................ 43 3 -1 Description of Sample Utilities. .............................................................................. 51 3 -2 Customer Income Groups. ..................................................................................... 52 3 -3 Inflation During Study Period. Source: BLS 2009 ................................................. 52 3 -4 Esc ambia County (Uniform). ................................................................................. 53 3 -5 Hillsborough County (five blocks). ......................................................................... 53 3 -5 Indian River County (four blocks). ......................................................................... 54 3 -6 Lakeland (three blocks). ......................................................................................... 54 3 -7 Melbourne (Uniform). ............................................................................................. 54 3 -8 Miami Dade (five blocks). ....................................................................................... 55 3 -9 Ocoee (Multiple Rate Structures). ......................................................................... 57 3 -10 Palm Bay (Three Blocks). ...................................................................................... 57 3 -11 Pam Beach (Uniform). ............................................................................................ 57 3 -12 Sarasota (Three Blocks) ........................................................................................ 58 3 -13 Seminole (Five Blocks). ......................................................................................... 58 3 -14 Spring Hill (Uniform). .............................................................................................. 58 3 -15 St. Petersburg (Five Blocks). ................................................................................. 58 3 -16 Tallahassee (Uniform). ........................................................................................... 59 3 -17 Tampa (No Fixed Fees) ......................................................................................... 59 3 -18 Toho (Multiple Rate Structures) ............................................................................. 59 6 -1 Customer Income Groups. ................................................................................... 100
9 6 -2 Estimated Hardship Thresholds for Hypothetical Florida Household ................. 100 6 -3 Melbourne Rate Structure: 1998. ........................................................................ 100
10 LIST OF FIGURES Figure page 1 -1 Graphical Displays of Several Rate Structures. Source: Borisova and Rawls (2009) ...................................................................................................................... 22 3 -1 Geographic Distribution of Sample Utilities. Source: Whitcomb (2005) ............... 52 4 -1 Reduction in Bill after 40% Reduction in Use........................................................ 71
11 LIST OF ABBREVIATIONS AWE Alliance for Water Use Efficiency AWWA American Water Works Association DWR Department of Water Resources EFC Environmental Finance Center EPA Environmental Protection Agency FDEP Florida Department of Environmental Protection NRRI National Regulatory Research Institute SJFWMD Saint Johns Water Management District SWFWMD Southwest Florida Water Management District WMD Water Management District
12 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science COMPETITING OBJECTIVES OF RESIDENTIAL WATER RATE DESIGN IN FLORIDA: AN EMPRICAL STUDY By Colin Rawls December 2009 Chai r: Tatiana Borisova Major: Food and Resource Economics Economic theory suggests that price incentives can be used to encourage water conservation in resid ential consumers. Conservation water rates are designed to send price signals that encourage households to reduce discretionary water use in the long term. Often, these conservation rates use inclining block price structures, where progressively higher price blocks make the marginal price of water greater as water use increases. However, it is not always clear that conservation rates effectively provide meaningful incentives. Utilities themselves may also not have strong incentives to implement conserv ation rates. If conservation rates have a negative impact on revenue, or if they lead to increased revenue variability, utilities may have a disincentive to use them. In addition, block pricing structures may be inequitable, in the sense that the revenu e burden is borne disproportionately by some customer groups. This may make some rate structures unpopular with consumers and politically unfeasible. More research is needed about the tradeoffs associated with conservation water rate design. This study uses empirical evidence from Florida to explore the advantages and disadvantages associated with different types of water rate structures.
13 The broad objective of this study is to analyze the competing objectives of residential rate design in Florida. Ther e are several specific objectives that will contribute to this overall goal: To explore the relationship between type of rate structure and the strength of conservation price signals; To explore the relationship between type of rate structure and revenue and revenue stability; To examine the effect of conservation rate design on customer equity and water affordability. Data were provided by the 2005 Florida Water Management District report, Florida Water Rates Evaluation of Single Family Homes, by Dr. John Whitcomb. The report collected water use and price data from 3,000 Florida households from 19982003. The sample included sixteen utilities which used a total of 66 different rate designs. The empirical results suggest a number of conclusions. Fir st, there is evidence that stronger conservation price signals are indeed a key advantage of inclining structures with more than three price blocks. But, an important disadvantage of these multi block structures is that they tend to increase revenue variability. Consumer equity does not appear to be a significant disadvantage of any rate structure type. Low affordability of water for low income households is a disadvantage observed for some rates, but affordability does not seem to be correlated with rate structure type. Finally, a number of additional factors related to water price, including: sewer rates, taxes, inflation, and the ratio of fixed to volumetric charges are important for evaluating rate design.
14 CHAPTER 1 INTRODUCTION Background Economics has often been defined as the study of the allocation of scarce resources ( Mankiw 2004). Water is essential for human life and for all natural systems, and it is also a scarce resource. It is estimated that water scarcity affects nearly one in three p eople on the planet (WHO 2009). And, this proportion may be increasing. As the global population continues to increase, the worlds water needs have increased accordingly. Meeting these needs will not be easy. Fred Pearce, a journalist who has studied environmental issues for decades, refers to water as the defining crisis of the Twenty -First Century (Pearce 2008). Pearce goes on to explain that many countries, including the United States, are vulnerable to droughts and water shortages. Policy maker s, engineers, and specialists across the country are engaged in a discussion about how the water needs of future generations will be met. Economics as a discipline can offer much to this discussion. Economists have been exploring and debating natural resou rce issues for decades. For example, Harold Hotelling studied the optimal extraction of a nonrenewable resource (Hotelling 1931). Of course, water is usually thought of as a renewable resource. But, depending on how it is used, it may actually become no nrenewable. The term water mining was coined to describe the nonrenewable use of water (Cowen, 2008). In numerous cases, water mining has resulted in the loss of wetlands and the emptying of reservoirs. The body of literature related to water management and allocation has expanded dramatically in recent years (Agarwal, et al. 2000; Booker et al. 2005; Brown, Sharp, and Ashley 2005; Daigger 2009; Gonzales -
15 Gomez 2008; Kenny 2003). The state of Florida faces a number of specific water related issues, incl uding: droughts, saltwater intrusion, a growing population, the degradation of water resources due to point and non-point source pollution, and water demand that is highly seasonal. As the states population continues to grow, it is unclear how the water n eeds of the people will be met. A number of areas throughout the state have already experienced severe water shortages (SFBJ 2008; SFWMD 2009; US Water 2002). In the summer 2001, the state of Florida was suffering from a severe drought (FDEP 2008). Water shortages were plaguing numerous towns and cities, and wildfires were appearing with alarming frequency. In response to these grim circumstances, the Florida Department of Environmental Protection launched an initiative to aggressively explore ways to inc rease conservation and water use efficiency. The following year, key stakeholders, including utility companies and the states Water Management Districts (WMDs), signed an agreement, a joint statement of commitment (FEDEP 2003), that obligated them to a ttempt to implement the initiatives recommendations. These recommendations included the need to define and phase in residential demand management using several tools, including conservation water rate structures and well as drought rates. (FDEP 2002, p. 6163). In any community, residential water demand is a major component of total water usage. Therefore, residential demand management is a natural priority for regulatory institutions. There are two primary ways for policy makers to reduce residenti al demand. The most direct way is through water usage restrictions. These serve as a kind of command and control policy instrument, and can be quite effective.
16 Conservation pricing is another type of policy instrument. This strategy attempts to influ ence consumer behavior rather than mandate it. In contrast pricing allows consumers the flexibility to decide if, when and how, to reduce water use. Empirical research suggests that water usage responds to price changes (Whitcomb 2005, Dalhuisen 2003). T he idea behind conservation rates is that rather than charging each user a fixed fee, utilities can charge users based on their usage (i.e., volumetric fees). This gives users an incentive to use less water. In the past, the rates were usually implement ed through declining block rate structures. The more water a household used, the less it was charged per gallon. Customers were grouped into different price blocks based on usage. These rate structures were used due to the economies of scale associated with water distribution. Building a water distribution network is very capital intensive. As a result, the marginal cost of providing the first 1000 gallons of water is much higher than the marginal cost of providing the second or third. In recent y ears, as droughts have become more common in urban and suburban areas, regulators have looked for alternatives to declining block prices. One alternative is the simple uniform rate, where a uniform fee is charged for each gallon of water used. Another al ternative is the use of increasing block rates. These are the opposite of declining rates. The more water a household uses, the more it is charged per gallon. In addition the basic distinction between uniform and inclining block rate structure, there are a number of other components that can impact how conservation oriented a particular schedule of water rates is. For example, the uniform or block volumetric fees are only part of most rate structures. In Florida, there is also generally a
17 fixed mont hly fee that users pay regardless of water volume used. These fixed fees are used by utilities to stabilize revenue and help cover fixed costs. Sometimes, the fixed fees can constitute a large proportion of a households monthly bill. If this is the cas e, the conservation price signals may not be as strong as they first appear. Therefore the ratio between fixed fees and volumetric fees is an important indicator of the marginal price signal. Most residential consumers are billed for sewer fees at the sam e time as water. In fact, many sewer fees in Florida are volumetric in nature and based on wastewater use as estimated by water use during specified periods. If this is the case, then the nature of the sewer prices may also affect the overall impact of a conservation oriented rate. In Florida, dozens of utilities have implemented increasing block structures. Table 1 -1 demonstrates this trend. However, despite this trend toward using increasing block rates, there has been relatively little investigation into the success of price water demand management in Florida. In addition, there have been relatively few empirical studies about the implications of using different types of conservation rate. Conservation rate design involves selecting the general struct ural form of the rates (i.e., uniform rates or block rates), the number of blocks in the block structures (e.g., three block structures vs. six block structures), relative shares of fixed vs. volumetric charges, and relative proportion of sewer charges in household bills. As the state faces ever more severe water shortages, policy makers and utilities need more information about implications of specific rate design choices. The need for more empirical evidence is part of the impetus for this thesis.
18 Ge neral Research Problem: Competing Objectives of Water Rate Design Many utilities in Florida have adopted some form of increasing block rate structure. However, there is wide variation in the exact nature of these rates and the way that they are implemented. There is also very limited empirical evidence in Florida about the advantages and disadvantages of using different types of rate structures. For example, there may be significant tradeoffs associated with the addition of new price blocks to a rate str ucture. It is plausible that additional blocks increase the incentive for residents to conserve water while making utility revenue streams more volatile. Other potential tradeoffs exist in conservation rate implementation, but relevant empirical research is limited. The purpose of this study is to provide quantitative evidence about the nature of these tradeoffs. Providing an incentive to conserve water and/or use it more efficiently is clearly a major objective of many utilities and regulatory agencies. However, there are several concerns associated with using price as a water conservation policy tool. First, the comparative effect of various conservation water rate structures on residential water consumption has not been adequately examined. This paper will attempt to address this issue by analyzing relevant empirical evidence. Second, utilities are in the business of selling water. Any reduction in sales (due to the implementation of conservation programs or other reasons) is likely to negativ ely affect their revenues, so long as marginal price is less than marginal cost. Further, even if conservation rates have no negative impact on overall utility revenue, they may affect revenue variability. When household bills are determined by both water use volumes and the rate applied to specific blocks, revenue from month to month is likely to be less predictable. And, revenue uncertainty can impose real costs on utilities. (Beecher et al.
19 1994). Uncertainty in the revenue stream can impose administr ative costs by making planning more difficult, and it can increase the financial cost of borrowing. (Chestnutt et al.1993). Revenue and revenue stability are two very important utility objectives that may conflict with the conservation objectives. This paper attempts to examine both of them and how they interact with the conservation objective. Consumer equity is another important rate design objective. Rates that place large, disproportionate price burdens on particular customer groups are likely to be v iewed as unfair by customers and general public. And, while fairness is somewhat intangible, customer acceptance of water rates is very important for the perceived legitimacy of utilities providing water service. Therefore, when it comes to planning rates, utilities generally seek a process and an outcome that will be viewed as equitable. One goal of this paper is to determine if certain rate structure types are more or less equitable than others. Further, it is widely accepted that for residential consumers, there is a base amount of water use that is nondiscretionary, as in drinking, washing, cleaning, bathing, and other activities required for healthy living. Conservation rates that make non-discretionary water use unaffordable for low income households are likely to be politically unacceptable. This paper attempts to evaluate the implications of different rate structures on water affordability for low income households. Research Objectives The broad objective of this study is to analyze t he competing objectives of residential rate design in Florida. There are several specific objectives that will contribute to this overall goal.
20 Objective 1: To explore the relationship between type of rate structure and the strength of conservation pr ice signals. Quantitative analysis will be used to answer the following question: how do the size and number of price blocks in a water rate structure affect the household decisions to conserve water? Objective 2: To explore the relationship between typ e of rate structure and revenue and revenue stability. Statistical analysis will be used to examine the effects of conservation oriented rates on overall utility revenue and on the variability of the revenue stream. Objective 3: To examine the effect of conservation rate design on customer equity and water affordability. A possible disadvantage of conservation rates is that they may cause inequity among consumers or make water less affordable for low income consumers. From a policy perspective, neith er of these outcomes is desirable. This study will explore this issue by measuring the equity and affordability associated with different rate structures. The Data Empirical analysis of these issues is based on the extensive data set prepared by Dr John Whitcomb in his 2005 study titled Florida Water Rates Evaluation. The study was planned and funded by five Florida Water Management Districts, under the leadership of the Southwest Florida Water Management District. The study collected billing and water use data as well as socio-demographic and attitudinal data from 7200 households being served by 16 different Florida utilities from 1998 -2003. During the study period, a wide variety of rate structures were used, including uniform and inclining b lock with anywhere from two to six price blocks. One utility, the Tampa Bay Water department, used a water rate structure based entirely on volumetric charges, while other utilities employed both fixed and volumetric charges. Organization of the Thesis Chapter 2 presents a review of the relevant literature. The empirical results are analyzed in chapters 3, 4, and 5. Each of those chapters is distinct in purpose and methodology, but they are similar in terms of organization. They each include a
21 backgro und section, where additional relevant literature is reviewed and hypotheses are discussed. They also include a section about methods and a section summarizing and discussing the results. Chapter 3 provides a more detailed description of the empirica l data used in this study. It describes some of the methodology that is relevant to several different sections of the thesis, including the estimation of monthly household water use. Chapter 4 focuses on the issue of effectiveness of conservation rates in changing residential water usage. The basic hypothesis is that inclining block rate structures will be more conservation oriented, in terms of sending stronger conservation price signals. In addition, it is expected that the more price blocks a rate st ructure has, the stronger the conservation incentive will be. The effectiveness of various rate structures is estimated by comparing average and marginal prices faced by an average customer, as well as comparative shares of utility revenues generated through fixed and volumetric fees Chapter 5 focuses on utility revenue and revenue stability. The basic hypothesis is that rate structure does not have a negative impact on utility revenue. It is expected that average monthly revenue will not be lower for periods with inclining blocks than for periods with uniform rates. And, it is expected that that inclining block structures do not lead to increased revenue variability. Chapter 6 focuses on customer equity and water affordability. There are two main hy potheses. The first is that inclining block structures are more inequitable than uniform rates, in the sense that particular groups of customers contribute disproportionate share to the utility revenue. The second is that there is no correlation between type of rate
22 structure and affordability for low income / low use customers. Analysis of equity in Chapter 5 focuses on customer groups as differentiated by income and water usage. Chapter 7 summarizes the study results, conclusions, and implications. I t includes an additional qualitative discussion about the issue of economic efficiency. Limitations of the study and future research needs are also discussed. Figure 11. Graphical Displays of Several Rate Structures. Source: Borisova and Rawls (2009) $0.00 $0.50 $1.00 $1.50 $2.00 $2.50 $3.00 $3.50 $4.00 $4.50 $5.00 0 10 20 30 40Unit Charge, $/1000 gallonsThousand gallons / month Inclining Block Uniform Seasonal Summer Seasonal Winter
2 3 Table 1 2. Trends in Florida Rate Design. Source: edited and expanded from Whitcomb (2005) Utility 1998 Rate Structure 2003 Rate Structure 2009 Rate Structure Escambia County Uniform Uniform Uniform City of Tallahassee Uniform Uniform Increas ing with 3 blocks City of Melbourne Uniform Uniform Uniform City of Ocoee Uniform Increasing with 6 blocks Increasing with 6 blocks City of Palm Coast Uniform Uniform Increasing with 4 blocks City Spring Hill Uniform Uniform Increasing with 5 block s Palm Beach County Increasing with 3 blocks Increasing with 3 blocks Increasing with 4 blocks City of Lakeland Increasing with 3 blocks Increasing with 3 blocks Increasing with 4 blocks Miami Dade Increasing with 5 blocks Increasing with 5 blocks Incre asing with 4 blocks Indian River County Increasing with 4 blocks Increasing with 4 blocks Increasing with 4 blocks Hillsborough County Increasing with 5 blocks Increasing with 5 blocks Increasing with 4 blocks City of St. Petersburg Increasing with 4 bl ocks Increasing with 4 blocks Increasing with 4 blocks Toho Water (Osceola County) Increasing with 5 blocks Increasing with 5 blocks Increasing with 5 blocks Sarasota County Increasing with 5 blocks Increasing with 5 blocks Increasing with 5 blocks City of Tampa Increasing with 3 blocks Increasing with 3 blocks Increasing with 4 blocks Seminole County Increasing with 5 blocks Increasing with 5 blocks Increasing with 6 blocks
24 CHAPTER 2 LITERATURE REVIEW Background This chapter provides a general over view of some of the literature that is relevant to the subject of conservation water rates and utility objectives. Defining Conservation Rates Generally, any rate structure that provides an economic incentive to conserve water is considered a conservation rate structure. On a more technical level, studies present different requirements to conservation rate structures, focusing on the following main characteristics: (1) the structural form of the volumetric water rates; (2) the proportion of volumetric ch arge in the total customer bill; (3) the proportion of utility revenues recovered through fixed fees versus volumetric rates; (4) effective communication of the price signal through consumer billing (see, for example, AWE 2008, Beecher et al. 1994, and Minnesota DNR 2008). Conservation rates are usually associated with uniform, inclining block, and seasonal volumetric rate structures (often referred to as uniform, inclining block, or seasonal rates). With a uniform rate, the user pays a set charge for each unit of water used. Uniform rates have the advantage of being relatively simple to administer and easy for the customers to understand. Uniform rates also send customers a usage price signal, since the total water bill increases with increase in water c onsumption (AWWA 2000). With Declining block rates (also known as descending or decreasing block rates), the charge paid per unit of water decreases at certain usage thresholds. For example, the rate may be $3.00 per 1000 gallons for the first 10,000 gal lons used, and only $2.00
25 per 1000 gallons for all additional usage. Usage volumes up to such threshold form a price block (also referred to as tier). The declining block rate volumetric structures are often used by utilities that need to develop a s ingle rate schedule for various customer classes served. Such structure can allow utilities take into account the different costs and usage characteristics of all customers while remaining equitable to all of them. For example, an initial block can be des igned to recover costs associated with the volumetric use of residential and small commercial customers, and subsequent blocks can be selected to encompass the water use and associated demand costs of industrial customer class (AWWA 2000). Declining block rates can be used when utilities costs decline with increasing water usage (due to economies of scale), and when it is important to provide price incentives to encourage largevolume customers to remain on the system (instead of developing their own sourc e of supply by drilling a well, for example) (AWWA 2000). Declining block rate may also be used by utilities that need to encourage economic development (Childs and Kramer 2008). With an inclining block volumetric rate structure (also known as increasing, inverted, ascending block rates), price for additional units of water increases at certain water use thresholds. For example, the rate may be $2.00 per 1000 gallons for the first 10,000 gallons used, and $3.00 per 1000 gallons for all additional usage. Inclining block volumetric rate structures provide stronger disincentive for a customer to use large quantities of water (compared with uniform and declining block rate structures), and as a result, this structural form is most commonly presented as a conservation rate structure. However, inclining block rate structures are difficult to design and administer, since they require analysis of the water volumes sold per price
26 block and demand responses to price differentials between the blocks (AWWA 2000). U tility wide application of inclining block rate structures can also result in cost of service inequities, especially to commercial and industrial customers these customers may not impose costs on a water system proportional to the costs implied by i ncreasing block rates (AWWA 2000, p. 99100). Of course, if the opportunity cost of water itself is not included in cost recovery rates, inclining rates could, for a customer, capture the long-run consequences of high consumption. Furthermore, if signi ficant cost -recovery depends on those consuming in the higher blocks, changes in demand (due to unusual weather patterns, changes in population demographics, or changes in income) can lead to revenue shortfalls. Advantages and disadvantages of conservation water rates will be discussed in more details in the following sections. Economic incentives to conserve water created for the customers by an inclining block rate structure depend on the size and the number of the price blocks. The literature provides limited recommendations for the design of inclining block rate structures. Chestnutt and Beecher (1998), focusing on efficiency as the focus of rate design, recommend selecting a rate structure in such a way that the price of the last unit of the water c onsumed is equal to the additional (i.e. marginal) costs of new supplies. The Minnesota Department of Natural Resources (2008) recommends the increase in price between the price blocks to be 25% or more, with 50% increase between the last two blocks. The Alliance for Water Efficiency (2008) recommends selecting the first price block such that minimum water usage is provided to a typical household at a minimum reasonable price, and setting the price increase between the blocks to be greater than 50%. Furth er, an effective rate design will have more than half of
27 residential customers exceeding the first tier when the new rate structure is first implemented, and at least 30% and 10% of customers using water in the 3rd or 4th tiers respectively (at least duri ng seasonal peak demand) (AWE 2008). However, Nida and Eskaf (2009) examined the rate structures used by North Carolina utilities and showed that for majority of utilities, the first price block exceeds typical residential use. That is, the rates are eff ectively uniform for the water usage below 15,000 gallons per month, and majority of customers are unaffected by the higher price blocks. Similarly, in Georgia, the Environmental Finance Center (2007) reports that a customer that reduces their consumption by 40% from 10,000 to 6,000 gallons/month is likely to receive the same reward, both in terms of total bill reduction and percent bill reduction, whether they are being charged increasing block or uniform rates (p. 3). With respect to the number of price blocks, The Alliance for Water Efficiency (2008) suggests that 3 to 4 blocks are adequate for an effective residential rate design, and a nation wide survey of water utilities by AWWA and Raftelis (2006) shows that for the surveyed utilities that use incr easing block rate structures for residential water supply, the average number of blocks is 3.8. Nation wide surveys of water utilities indicate the drop in the use of declining block rate structures for residential water services, and an increase in the u se of inclining and especially uniform rates (Table 2). In Florida, out of 16 utilities surveyed by Whitcomb (2005), 6 utilities used uniform and 10 used inclining block rates for residential customers in 1998 (Whitcomb 2005). In 2008, from the same sample of utilities, 3 used uniform, and 13 used inclining block rate structures (see table 3 in Appendix).
28 In the 2006 survey conducted by AWWA and Raftelis (2006), over 73% of the responding water utilities indicated that they have the same volumetric structu re for residential and non-residential customers. The authors note that even under the same volumetric rate structure, the exact rates are not necessarily identical for residential and non-residential customers. Utilities that reported different volumetri c rate structures generally shift from an increasing block rate structure for residential customers to a uniform or declining block rate structure for non -residential customers. For example, 36% of utilities have an increasing block rate structure for resi dential customers, but only 23% have an increasing block rate structure for non-residential customers (AWWA and Raftelis 2006). Wang et al (2005) describe two utilities that experiment with conservation pricing for their non -residential customers. In Cleveland (OH), inclining block rate structure is used for industrial consumers, nearly doubling the price from the first block to the next. Louisville Water Company (KY) uses a pyramid block structure that includes a low rate per thousand gallons for both relatively small customers and very large customers, while intermittent heavy users (such as restaurants) face higher water rates. It is likely not accurate, however, to consider pyramid block rates to be water conservationoriented rates, as they result in the highest consumers within the commercial class paying less per unit that those who use less (Wang, Smith and Byrne, 2005). Some utilities increase their rates or implement a new rate structure during specific seasons of peak use (seasonal rates) or times of droughts (drought tares), to provide additional incentives for water conservation. Of the 231 utilities surveyed by AWWA and
29 Raftelis (2006), 36 reported that they use seasonal rates. More in -depth discussion of drought rates is presented in sect ion 4.7. Some utilities are also experimenting with rate structures based on individual household water budgets. Water budget based rate structures are also very effective in promoting conservation, though more difficult to implement. In this design, eac h residence has an inclining block rate structure designed according to its individual needs. The tiers are usually set based upon the quantity of occupants and the square footage of landscape; known to be the two most significant factors in residential w ater use. The prices of the tiers increase significantly (greater than 50%) after going beyond the base usage tier. This rate system requires a robust billing system to accommodate the quantity of individual rate structures (possibly equal to the number of customers); and the system requires a formal process to establish each homes base water usage, and respond to the many customers likely to appeal their base tier allotment (AWE 2008). One example of a more robust billing procedure is the use of sepa rate meters for irrigation. In some cases, the utilities have separate, unique rate structure specifically for these irrigation meters. In addition to volumetric rates, almost all utilities charge fixed fees (also referred to as base, minimum, monthly, or meter fee or charge) that are the same each billing period regardless of usage. The fixed fee is almost always based on meter reading, billing, and collection costs. Many utilities also include meter repair and replacement costs and a capacity charge in their fixed fee (AWWA and Raftelis 2006). The Alliance for Water Efficiency suggests that conservation rates should be designed so that a large portion (two -thirds or more) of the water charges are based on the quantity of water the
30 customer consumes (AWE 2008). According to the data from the nationwide survey by AWWA and Raftelis (2006), the monthly fixed fee for the median customer ($5.84) comprises 29.3% of the total water bill (1000 cubic feet or 7.5 thousand gallon of water usage). In addition to water charges, many utilities include wastewater charges in the total customer bill. Wastewater charges are typically based on a percentage of a customers monthly water use (AWWA and Raftelis 2006). As a result, wastewater charges make the customers pay mor e for the non discretionary water uses in comparison with the discretionary uses (effectively converting an inclining block into a declining block rate structure), and distort the economic incentives to conserve water created by conservation water rates. For example, Gainesville Regional Utilities (Florida) assess residential customers a wastewater charge of $4.94 per thousand gallons, based on average monthly water usage or winter maximum water use, whichever is lower. Consider a hypothetical household t hat uses six thousand gallons per month in winter. This household would pay $6.53 per thousand gallons for their first six thousand gallons ($4.94 of wastewater charge plus $1.59 of water charge), and only $1.59 per thousand gallons for any additional wate r use. If this households consumption exceeds nine thousand gallons, the water rate would still be only $3.11 per thousand gallons (up to twenty -five thousand gallons), much below the rate for the nondiscretionary water use (GRU 2008a, GRU 2008b). Objec tive 1: Rate Structure and the Strength of Conservation Price Signals To influence water demand, the conservation pricing must be understood by customers. Households should be able to estimate changes in their water bills corresponding to increases (or dec rease) in water usage. The estimates of the effects of
31 information campaign on consumer response to price signal varies from study to study. For example, Gaudin (2006) reports an increase in consumer responsiveness to price signals by up to 30%. In contras t, Carter and Milon (2005) found that the knowledge of the rates for additional units of water (i.e., marginal price) result in the increase in monthly water consumption. The authors hypothesize that the households tend to over estimate their marginal water rates, and hence, they increase water consumption in response to the knowledge of the accurate marginal rates. The survey of customers of sixteen Florida utilities conducted by Whitcomb (2005) showed that 39% of respondents are not knowledgeable about water rate structures (i.e., number, size, and prices of the blocks). At the time of the survey, only five of the sixteen participating utilities printed their water rates on their bills; this practice partially explains this lack of customer knowledge (W hitcomb 2005). Analysis of 1997 survey of customers of three North-Central Florida utilities by Carter and Milon (2005) shows that higher monthly income, larger household size, home ownership, larger lawn area, and awareness of nonprice conservation programs increase the likelihood of knowing the marginal price. Households facing block rate structures, however, are less likely to know the marginal price. In 2007, the California Urban Water Conservation Council established specific guidelines for what consti tutes a conservation rate (McLarty and Heaney 2008). To meet Californias conservation rate criteria, at least 70% of monthly utility revenue must come from volumetric rates (McLarty and Heaney 2008). Responsiveness of the water use to rate is measured through the price elasticity of demand. Price elasticity is defined as the percent change in water consumption in
32 response to the certain percent change in price (rate). The most likely price elasticity range for long -term overall (indoor and outdoor) residential demand is -0.10 to 0.30, with price elasticity coefficients for longterm industrial and commercial demand ranging up to 0.80 (AWWA 2000, p. 158). This means that for residential customers, a 10% increase in rate (given current rate level) wil l most likely result in reductions in water usage within the range of 1% 3%. Price elasticity depends on a variety of factors, such as: the value of subsidies available to consumers; the size of wastewater and fixed charges in customer bills; percent of total income spent on water; price of water from alternative water sources (such as private wells); length of time over which rates and water demands are evaluated; climate and weather events; initial water rates against which the elasticity is measured; customer class (e.g., residential, commercial, or industrial); type of water use (indoor vs. outdoor); season and time of the day (peak vs. off -peak periods); geographical region; customers knowledge of their water rates; presence of other conservation pr ograms; and customer education programs (AWWA 2000; Carter and Milon 2005; Cavanaugh, et al 2002, Dalhuisen et al 2003, Espey et al 1997; Howe 2002; Howe and Goemans 2002, Wang et al. 2005). Price elasticity appears to rise with an increase in rate levels (AWWA 2000). Also, water use is more responsive to the change in real prices (adjusted for inflation), than in nominal prices (not adjusted for inflation) (AWWA 2000). The rates for the additional unit of water (i.e., marginal water price) are low in U S (Cavanagh et al. (2002) cite the marginal price of $0.50 to $5.00 per thousand gallons). The average monthly bill for an average US customer (with about 7500 gallons of monthly usage) is $20.24 (AWWA and Raftelis 2006), which is a
33 small portion of aver age US household income. Such low rates explain (at least partially) the small response in household water consumption to price increase (Cavanagh et al. 2002). However, price levels sufficient to induce significant water savings are politically and socially controversial (Cavanagh et al. 2002, p. 6). Generally, water demand for outdoor discretionary uses (such as lawn watering, car washing, and swimming pools) is more elastic than the demand for non discretionary indoor water uses. In Florida, there is l ess outdoor discretionary water use during late fall and winter when water use for irrigation decreases and the demand for water may be less responsive to price changes during that season. Further, peak usage is more price -sensitive than off -peak usage ( AWWA 2000, p. 159). Price elasticity is greater when the response is measured over the long period of timemore than 3 to 5 years. ( Carter and Milon 2005). The presence of marginal price information on the bill next to quantity consumed increases price elasticity (by a factor of 1.4, according to Gaudin 2006). Further, when water restrictions are implemented, consumers can be less responsive to rate changes (Kenney et al 2008). High water users are generally more responsive to price than low water user s (Kenney et al. 2008). Low income households are significantly more price responsive in comparison to the relatively wealthy households reflecting the larger share of water bills in the low income household budget (Agthe and Billings 1987, Dalhuisen et al. 2003, Renwick and Archibald 1998). Based on the analysis of 64 studies and 314 price elasticities, Dalhuisen et al. (2003) shows that price elasticity estimates vary depending on geographical regions of US, so that price elasticities are greater in absolute value in the
34 arid West (p. 306), which may be related to more significant water use for discretionary purposes (irrigation). Rate structures themselves can affect consumer responsiveness to rate changes (Kenny et al. 2008). Cavanagh et al (2002) and Nieswiadomy and Molina (1989, cited by Nauges and Thomas 2000) found that price elasticity among households facing uniform marginal prices appears to be significantly smaller than among households facing block structure. If a household knows that hig her levels of use result in higher prices, it will be more sensitive to price (Cavanagh et al. 2002 p. 27). The differences in price elasticity estimates reported in existing studies can also be partially explained by the differences in the methodologies employed by the authors, specifically, the spatial and temporal level of data aggregation, period of time over which the elasticity is evaluated, price of water considered (average or marginal), and specific econometric estimation procedures employed (Cav anagh et al. 2002, Dalhuisen et al. 2003; Espey et al. 1997; Michelsen, et al. 1998). In addition, it may be difficult for consumers to distinguish the actual water rate from wastewater and fixed charges included in the water bills, which complicates the e stimation of price elasticity of water demand. Based on the responses to the nationwide survey of utilities conducted by Wang et al. (2005), many utilities do not consider elasticities in designing water rates. An exception is Tucson (AZ), where utilitie s believe that for some customers, a 10% increase in price will result in 4% reduction in water usage (elasticity = -0.4), but more common response is 2% decrease in usage (elasticity = -0.2). Further, San Antonio (TX) responded that it is difficult to is olate impacts of individual conservation programs
35 (focused on the Edwards Aquifer); however, it believes that water conservation rates had the main impact on the 25% reduction in per capita consumption between the mid 1980s and 1998. El Paso (TX) reported that its water conservation rates, along with other conservation programs, led to the drop in per capita consumption from 220 gallons to 165 gallons per day. This number would be even lower if non-residential consumers would have been excluded from estimat ions. Corpus Christi (TX) reported low amount of per capita consumption (130 gal per person per day) and attributed this record to education, planning, ordinances, aggressively pursuing irrigation leaks, and conservation rates (Wang et al. 2005). Clunie (2004) reports the results of two case studies from Hawaii. In Kauai County, average monthly single-family residential water use (normalized for weather) dropped by 3.7% in the year following the change in the water rate structure from uniform to inclini ng 3-block (with an average 32% rate increase). In contrast, in Hawaii County, average monthly single-family residential water use (normalized for weather) increased by 3.7% in the year following the change in the water rate structure from inclining 3blo ck to inclining 4 -block structure (with an average 29% rate increase). The author suggests that increase in the number of blocks and the steepness of rate blocks may have impacted relatively few customers. Also, customers with long-standing inverted block rates may have already changed their water use patterns (Clunie 2004, p. 23), which may have reduced their ability to react to higher water prices. Objective 2: Rate Structure, Utility Revenue, and Revenue Variability Economists recognized the apparent tradeoff between conservation and utility revenue at least as early as 1974. Jones (1974) discussed the issue and explored a few basic strategies for dealing with this dilemma. An early empirical investigation was
36 performed Bhatt and Cole in 1985. They used a single utility as a case study and estimated what would happen if demand fell by 20%. They concluded that revenue would fall by about 16%, while operating expenses would only fall by about 2%. This relatively simple exercise demonstrates the potential seriousness of the problem. However, it is important to note that the above results assumed that there was no change in rate structure. In other words, the massive drop in water usage was presented as a hypothetical consequence of something other t han a rate design change. If a steep increase in rates had the caused a drop in demand, the estimated effect on revenue would have probably been much less severe. In addition to revenue losses, conservation pricing can also cause another problem for util ities: revenue instability. By adopting a rate structure that relies more on volumetric (rather than fixed) fess, utilities expose their revenue streams to more risk. This can be problematic, even if average revenue over time does not change. Chesnutt e t al. (1993) explore this issue in depth. They explain that increased revenue variability inflicts direct and very real costs on utility companies, most notably in the form of greater financial borrowing expenses and planning/informational expenses. The authors strongly recommended that regulatory institutions explore this issue further in order to make the risks and trade offs of increased conservation explicitly clear to utility companies. The National Regulatory Research Institute (1994) did just that the following year. Its paper presents a broad overview of the relationship between conservation pricing and revenue. It suggests that, while the trade off can be very real in practice, it need not derail conservation programs or pricing schemes. One k ey reason for this has to do
37 with the demand elasticity for water. Because demand is inelastic, increased rates actually have the potential to increase revenue. In other words, overall water usage can fall without harming utility revenue. Of course, even if this is true, conservation rates may still make revenue streams more volatile. Among the 23 utilities nationwide responded to the survey by Wang et al (2005), 9% of utilities responded that conservation rates increased their revenues, while 26% rep orted that revenues decreased, 30% considered conservation rates to be revenueneutral, and 35% did not know or gave no response. The literature also discusses the potential for conservation rates to increase revenue variability (AWWA 200 0, Chestnutt 1993). This revenue volatility is because an increasing block rate anticipates recovering a proportionately greater percentage of the customer classs revenue requirement at higher levels of consumption. These higher levels of consumption tend to be more subject to variations in seasonal weather and, when coupled with a higher unit pricing, customers tend to curtail consumption in these higher consumption blocks (AWWA 2000, p. 100). Generally, revenue streams from inclining block structures are more variable than revenue streams from declining block structures (AWWA, 2000, p. 100). Smaller utilities may be more affected by revenue variability than larger utilities. In a survey of North Carolina utilities, Nida and Eskaf (2009) observed larger fixed fees in smaller utilities and hypothesized that smaller utilities may, on average, have less stable customer consumption and therefore decide to shift greater proportion of their operating costs into the base charge. (p. 5). The previously cited paper by McLarty and Heany (2008) describes the volatility issue in clear terms. The authors note that utility companies generally have very high
38 up front fixed costs. As a consequence of this, they often must dedicate a fairly large share of revenue s ervicing debt they may have. Whether publicly or privately owned, they have a vested interest in revenue stability. By charging large flat fee, they can cover the fixed costs. A smaller volumetric fee can be used to cover variable costs. Unfortunately, rate designs with large fixed fees do not provide a strong incentive for water conservation. A revenue stabilization fund can be used to balance the need for conservation and the need for revenue stability (AWWA, 2000, p. 100). A certain percentage of surplus revenue can be allocated to the fund each month with surplus revenue; and the funds can be withdrawn from the fund when revenues fall below projections. A number of utilities in Florida, including Gainesville Regional Utilities, have adopted th is strategy of revenue stabilization (GRU 2008). Excess revenues can also be used to retire bonds in order to keep future rates low, to improve infrastructure, or to educate the public about water rates and water conservation. Deficit in revenues can also be addressed through increase in rates or taxes, through issuing bonds, inclusion of a risk margin in the calculation of revenue requirements, and developing a mechanism for more frequent rate adjustments (Wang et al. 2005). Objective 3: Rate Structure, Equity, and Affordability Poorly designed conservation rate structures can potentially lead to an inequitable billing of different customer groups (AWWA 2000). Renwick and Archibald (1998) find that water use of low income customers is more responsive t o price increase than the water use of high income customers. These results suggest that price policy will achieve a larger reduction in residential demand in a lower income community than in a higher income community, all other factors held constant. Res ults also suggest that
39 if price policy is the primary DSM (demand side management) instrument in a particular locale, lower income households will bear a larger share of the conservation burden (p. 357). However, Agthe and Billings (1987) demonstrate that with proper design of the inclining block rate structures, steeper price blocks will actually lead to greater distributional equity. The authors show that by making price blocks steeper, a utility could increase the incentive to conserve without adding any price burden to low income users. This conclusion is important; it means that when equity is a high priority of rate design, steeper price blocks can be a better option than increased fixed or uniform rates. To address the impact of conservation rates on low -income / low use customers, several utilities surveyed by Wang et al. (2005) charge minimum rates for the minimum amount of water necessary to meet basic needs (lifeline rate), which often constitute the fist block in the inclining block rate str uctures. The Kentucky Public Service Commission and several of the largest utilities in Texas support a lifeline rate of 2,000 gal per household per month. San Antonio, TX, uses the lifeline rate of 7,000 gal per month (Wang et al. 2000). Utilities focus on keeping the rates for the lifeline rate low to avoid setting excessive burden on low income customers. Some utilities forgive service charges to low -income customers, offer fixing water leaks for free, distribute free water efficient home appliances, of fer 50% discounts on the bills, or do not charge for water consumption within the first price block (Wang et al. 2005). To achieve utilities financial objectives, social tariffs (i.e., low rates) for low income / low use customers are often subsidized by other customer groups (e.g., by
40 customers from other regions, or by customers with other water use levels and/or higher income). Discussions of affordability and social tariffs should be open to all the stakeholders. Also, social tariffs should be based on precise definition of affordability and on reliable data on income distribution and water use. In the absence of such objective bases, there is a risk that the process be driven by political affordability (OECD 2009, p. 86). Fairness is somewhat intangible and imprecise, because it is related to public perception. An inequitable rate structure will probably be viewed as unfair by the public. Also, rate changes should be instituted in a proactive way, rather than in a way that could be viewed as punitive or reactionary. For example, utilities can be proactive by making rate changes in anticipation of future droughts rather than after a drought (GRU 2008). Other Utility Objectives Although this study only focuses on several objectives, of r ate design, the literature suggests that utilities often have additional objectives. One of these is efficiency. Economists have recommended that water prices should reflect the marginal cost of providing water, i.e. the cost of providing the next additio nal unit of water (AWWA 2000). Economic costs of water include utilitys operation and maintenance costs, costs of additional water supply to meet growing demands, and the social and environmental opportunity costs of losing other benefits that the water can provide (such as ecological and recreational values of water pumped for consumption from river basins) (Western Resource Advocates et al. 2004). Economic efficiency requires setting rates to each customer according to the customers specific marginal costs, and adjusting rates as the opportunity costs or the
41 water infrastructure use change (OECD 2009). Even if true marginal cost pricing is impossible, the literature strongly suggests that water rates should reflect costs of water provision (Griffin 200 1). When water rates are used for alternative purposes, economic inefficiency and inequity are the likely result, including underpricing (requiring a transfer from the governing body), overpricing (providing a transfer to the governing body), or subsidiz ing some customers at the expense of others (Griffin 2001, NRRI, 1994). In other words, water rates should be used to recover costs, and not as a tool to redistribute wealth (like a progressive income tax), subsidize development, or as source of addition al city revenue. In 1997, Tsur and Dinar explored the efficiency implications of various rate structures used for pricing irrigation water. They argue that efficiency in water pricing means that the marginal benefits of water use should be equal for al l users. With efficiency defined in this way, they find that economic efficiency can be achieved for irrigation rates using several different rate designs, including completely volumetric pricing and a two part tariff. A number of factors influence how and whether irrigation rates can be efficient, so rate design strategy should vary locally. One particularly important factor is what the authors call implementation cost. Implementation cost, as they define it, includes administrative costs, inform ational costs, enforcement costs, etc. The authors also find that irrigation rate design can have major impact on the choice of how crops are planted. T ransparency and accountability are also important criteria for successful rate design. In Florida there is no mandated rate design methodology. As a result, each utility has the authority to decide which rate structure to use based on its own criteria.
42 This allows great flexibility in rate design across the state. But, it also can present a public relations challenge. If the public does not consider the rate design process to be transparent and accountable, rate hikes could lead to resentment among customers. A rate process is transparent if the public understands why a rate change is necessar y before the change is implemented. Information about rate changes should be made available to the public via meetings, workshops, websites, or other means. Further, public involvement in rate design enhance public acceptance of the rates (AWWA 2000, Cuthbert and Lemoine 1996, Saarinen 1993). Also, rate changes should be coordinated with other demand management efforts and any supply expansion efforts. In several states surveyed by Wang et al (2005), conservation rates are supplemented by outreach programs such as conservation displays in schools, demonstration of low flow water use landscaping, and public service announcements. For example, in San Antonio (TX), the fee set to the fourth of the four residential rate blocks is used to fund provision for l ow flow toilets and rebates for efficient washing machines, free repairs of leaks for low income customers, and outreach efforts (Wang et al 2005). Finally, policy coordination is an important objective. Rate changes should be coordinated with other demand management efforts and any supply expansion efforts. In several states surveyed by Wang et al (2005), conservation rates are supplemented by outreach programs such as conservation displays in schools, demonstration of low flow water use lands caping, and public service announcements. For example, in San Antonio (TX), the fee set to the fourth of the four residential rate blocks is used to fund
43 provision for low flow toilets and rebates for efficient washing machines, free repairs of leaks for low income customers, and outreach efforts (Wang et al. 2005). Table 2 1. Water Rate Structures for Residential Water Services: Results from NationWide Surveys. S ource: AWWA and Raftelis (2006). 1996 1998 2000 2002 2004 2006 Declining block 3 6% 35% 35% 31% 25% 24% Uniform 32% 34% 36% 37% 39% 40% Increasing block 32% 31% 29% 32% 36% 36%
44 CHAPTER 3 EMPIRICAL DATA Overview This study relies on descriptive analysis of secondary data. The data were collected by Dr. Whitcomb, in his 2005 study, Florida Water Rates Evaluation Single Family Homes. The sample includes 16 utilities and 3,538 individual households. The data was collected from 1998-2003. It includes rate information, water use data, tax assessor data about property characteristics, and results from household surveys. The stated purpose of the study was to examine and quantify the functional relationship between water consumption and water prices for single -family residential customers in Florida. (p. 1). The project was funded b y the Floridas four water management districts, with the Southwest Florida Water Management District taking a lead role in planning. The study used econometric modeling to estimate elasticity of water demand and analyze the importance of various nonpric e variables. The data collected by Dr. Whitcomb were edited and organized by Dr. Burcin Unel, currently at the University of Bogazici in Turkey. Description of Utilities The Whitcomb study required a sample of utilities with the following characteristics : a variety of different rate structures, including uniform and inclining block rates large enough service areas that a sufficient number of households could be studied rate structures with large price ranges geographic variability, the study needed uti lities from as many of the water management districts as possible the ability and willingness to provide historical water use data for the selected homes.
45 Once these criteria were selected, a preliminary rate survey of 100 Florida utilities was conducted in order to provide basic information about possible options. Then a Project Advisory Committee was consulted in order to help choose the sixteen utilities that would best serve the purposes of the study. The sixteen utilities selected are summarized in T able 3 1. These utilities vary widely in terms of location or rate structure used. It is important to note that the type of rate structure used in 2003 was not necessarily the same as the one used in 1998. For example, Ocoee was using an inclining struct ure with six blocks in 2003. But, for most of the study period it used a uniform structure. The geographic distribution of the utilities is summarized in the following figure: Most of the utilities were either in the Southwest or the St. Johns Water Man agement District. None of the utilities were selected from The Suwannee River Water Management District jurisdiction. Relevant Variables As a part of Dr. Whitcombs study, 7200 surveys were sent out to individual households served by the sixteen utilities The survey focused on household economic and socio-demographic characteristics, as well as the knowledge of their water rates. Three thousand five hundred thirty eight households completed the survey (50.4% response rate). The survey results were supplemented with tax assessor data, which provided more information about each household, including property value, house size, and whether or not the house had a pool. Specific household water use information was then collected for the years 1998 through 20 03 from water utilities. The monthly water use for each of the 3,538 households was recorded for each billing period during the study period.
46 Rate structure information was also provided by each of the utilities for the study period. This rate information included fixed fees, volumetric fees, sewer fees, and any relevant taxes. In many cases, the rate structure changed during the course of the study. Each period of time where a distinct rate structure is used is referred to in this study as a rate structure period, not to be confused with billing period. Overall, between the sixteen utilities over the six years, there were 66 individual rate structure periods. Sewer Rates All utilities used sewer rates in order to raise additional revenue from househ olds. Sewer prices are used in chapter 5 (Revenue) and chapter 6 (equity and affordability) of this study. Generally, sewer rates had a fixed and variable component. The variable component was usually capped at a specified maximum sewer charge. The sewer rates are reported in the rate summary tables at the end of this chapter. Taxes Several of the utilities levied taxes on water consumption during the years of the study. This tax information is used in Chapter 6 (equity and affordability). The t axes generally were recorded as a variable charge that was calculated as a percentage on water use charges. Taxes are included in the rate summary tables at the end of this chapter. Household IncomeHousehold income information is also used in the stu dy of equity and affordability (Chapter 6). Even though precise household income figures are not reported in the Whitcomb data, the household surveys did ask a question about income. The results of this question allow each household to be placed in one o f the 6 income categories (Table 3-2). Relatively few households reported an income range in group one or six. The highest frequency of households were reported in groups 3 and 4. Estimating Monthly Water Use Monthly household water usage is estimated from household water usage reported for each billing period. The billing periods do not correspond perfectly with months, and each household in the sample has somewhat different billing periods.
47 The water use for each household in each month was estimat ed by taking a weighted average of the two billing periods that include days from a given month. For example, January 1998 water consumption for the household with survey ID 6301, USE6301Jan1998, can be estimated as: Equation 31 USE6301Jan1998= ( gallons in period 1 ) x ( January Days in period 1 ) Days in billing period 1 + ( gallons in period 2 ) x ( January days in period 2 ) ] Days in billing period 2 Similar calculations were conducted for every household in the sample and every month. Total monthly supply of the water utility to the sample households is a sum of monthly water use for all households in the sample: Equation 32 USEJan1998= USEID Jan 1998 where ID refers to household survey ID. Once the m onthly supply is estimated for all months, total water supply for specific utility and each rate structure period, z can be calculated: Equation 33 USERatePeriod=z= where subscript month refers to specific month of the rate period z Further, the average monthly water supply can be calculated for each rate structure period.
48 Equation 34 AVG USERatePeriod=z= USERatePeriod=z / n where n equals the number of months in the rate period z These monthly water supply results are utilized in chapters 4, 5, and 6. Correcting for Inflation The data were collected between 1998-2003, and in several chapters of this thesis, inflation is taken into account in order to identify real prices rathe r than nominal prices. The annual Consumer Price Index, as estimated by the Bureau of Labor Statistics, was used to estimate inflation. The rate of inflation during the study year is summarized in table 33. In most cases water prices tended to increase as utilities changed rate structures. But, in some cases, rates did not change during the study period. For these utilities, real prices declined over time. Inflation is taken into account in Chapter 4 and Chapter 6. Estimating Monthly Revenue In C hapter 5, it is necessary to estimate monthly revenue proxy for each utility. This subsection describes the estimation process. The utility revenue proxy in any given month is simply the summation of all household bills for the month. And, any household bill has several components. The monthly bill calculation is relatively simple when the rate structure is uniform. Below is the formula for the monthly revenue proxy given uniform rates: Equation 35:
49 Revenuem = f [( FixedW x NumH ) + ( VolW x Qm) + ( Fi xedS x NumH) + VolS x Qm ] where: m = month FixedW = fixed water fee NumH = number of households in sample VolW = volumetric water fee Qm = total quantity of water used by all households FixedS = the fixed Sewer fee VolS = the volumetric sewer fee When the rate structure is inclining block, the formula is slightly more complex. Below is the general formula for monthly revenue when there are three price blocks: Equation 36: Revenuem = f [( FixedW x NumH ) + (VolW1i) + ( VolW2i) + (VolW3i) + ( FixedS x NumH) + VolShi)] where: m = month FixedW = fixed water fee NumH = number of households in sample (VolWh1i) = the sum of the first block water fees paid by all households during the month (VolWh1i) = the sum of the second block water fees paid by all households during the month (VolWh1i) = the sum of the third block water fees paid by all households during the month FixedS = the fixed Sewer fee VolS = the volumetric sewer fee In rate structure periods with more than three price blocks, the same formula is used, except that additional price blocks are added. Once the monthly revenue proxy is calculated for each month, the average monthly revenue is calculated for each rate structure period. This calculation is carried out by adding the revenue proxies for every month in the period, and then dividing by the number of months in the period (equation 5 -3).
50 Equation 37 Average Monthly Revenue = m (Revenuem) / N where: Revenuem1 = the estimated revenue proxy for a given month N = the total number of months included in the rate structure period Rate Structure Information Summary Tables This section of the chapter summarizes the rate st ructure information of all sixteen utilities. The rate tables can be used as a reference in other chapters.
51 Table 3 1. Description of Sample Utilities District Utility Utility Abbreviation Rate Structure Type # of Rate Blocks in 2003 NWFWMD Escambia County Utilities Authority Escambia Uniform 1 NWFWMD City of Tallahassee Tallahassee Uniform 1 SJRWMD City of Melbourne Melbourne Uniform 1 SJRWMD City of Ocoee Ocoee Inclining 6 SJRWMD City of Palm Coast Palm Coast Uniform 1 SWFWMD Hernando County U tility Spring Hill Uniform 1 SFWMD Palm Beach County Palm Beach Inclining 3 SWFWMD City of Lakeland Lakeland Inclining 3 SFWMD Miami Dade Water and Sewer Department Miami Dade Inclining 4 SJRWMD Indian River County Utilities Indian River Inclining 4 S WFWMD Hillsborough County Water Department Hillsborough Inclining 4 SWFWMD City of St. Petersburg St. Petersburg Inclining 4 SFWMD Toho Water Authority Toho Inclining 5 SWFWMD Sarasota County Utilities Sarasota Inclining 5 SWFWMD City of Tampa Tampa I nclining 5 SJRWMD Seminole County Utilities Seminole Inclining 6
52 Figure 31. Geographic Distribution of Sample Utilities. Source: Whitcomb (2005) Table 3 2. Customer Income Groups. Name Income range Income group 1 <$15,000 per year Income group 2 $ 15,000 $29,999 per year Income group 3 $30,000 $49,999 per year Income group 4 $50,000 79,000 per year Income group 5 $80,000 $100,000 per year Income group 6 >$100,000 per year Table 3 3. Inflation During Study Period. Source: BLS 2009 Year Percent Inflation Consumer Price Index with 1998 as Base Year 1999 2.2 1.022 2000 3.4 1.056 2001 2.8 1.084 2002 1.6 1.100 2003 2.3 1.123
53 Table 3 4. Escambia County (Uniform). Water Sewer Start End Per TG Fixed Charge 0 2 TG after 2 TG Fixed Charg e January 98 September 98 1.22 5.97 0 14.92 6.87 October 98 September 99 1.23 6.03 0 15.08 6.94 October 99 September 00 1.25 6.15 0 15.4 7.08 October 00 September 01 1.28 6.27 0 15.72 7.22 October 01 Septembe r 02 1.31 6.4 0 15.88 7.29 October 02 September 03 1.34 6.53 0 16.2 7.44 October 03 March 04 1.36 6.63 0 16.44 7.56 Table 3 5. Hillsborough County (five blocks). Water (per TG) Sewer Start End 1 8 9 15 16 30 31 50 51+ Fixed Charge Fixed 0 8 TG January 98 March 99 2.2 5 2 3.5 5 4.1 4.7 4 7.45 5.35 April 99 June 00 2.2 5 2 3.5 5 4.1 4.7 5.9 9.2 4.9 July 00 March 01 2.3 5 2 3.7 4.3 4.9 7.05 11.05 4.4 April 01 September 01 2.3 5 2 3.7 4.3 4.9 7.9 12.75 4.1 October 01 Septemb er 02 2.5 7 3 3.9 2 4.5 2 5.1 2 7.9 12.75 4.1 October 02 May 03 2.8 5 3 4.2 4.8 5.4 7.9 12.75 4.1 June 03 September 03 2.3 5 4 4.7 6.2 6.2 7.9 12.75 4.1 October 03 December 03 2.4 3 4 4.7 8 6.2 8 6.2 8 7.9 12.75 4.1
54 Table 3 5. Indian River County (four blocks). Water (per TG) Sewer Start End 1 3 4 7 8 13 14+ Fixe d 0 10 TG Fixed Charge January 98 September 99 1.75 2.15 2.55 4.8 5 11.2 3.3 5 15.50 October 99 December 03 2.20 2.42 3.85 7.7 0 9.05 2.8 6 15.87 Tabl e 36. Lakeland (three blocks). Wa ter (per TG) Sewer Start End 1 10 11 15 16 Fixed Fixed 0 12 TG January 98 September 98 1.01 1.20 1.42 3.64 7.95 1.7 October 98 September 99 1.01 1.20 1.42 3.64 7.95 1.7 October 99 September 01 1.06 1.26 1.49 3.80 8.12 1.7 October 01 September 02 1.1 9 1.45 1.79 4.30 8.12 1.74 October 02 September 03 1.22 1.48 1.83 4.40 8.31 1.78 October 03 March 04 1.28 1.57 1.96 5.20 9.00 1.92 Table 3 7. Melbourne (Uniform ). Water Sewer Start End Per TG Fixed Charge per TG Fixed Charge January 98 September 98 2.07 3.28 3.34 4.63 October 98 September 99 2.17 3.45 3.5 4.86 October 99 September 00 2.28 3.62 3.68 5.11 October 00 September 01 2.39 3.8 3.86 5.36 October 01 September 02 2.51 3.99 4.06 5.63 October 02 September 03 2.64 4.19 4.26 5.91 October 03 March 04 2.77 4.4 4.47 6.21
55 Table 3 8. Miami Dade (five blocks). Water (per TG) Sewer Start End 1 4 5 8 9 11 12 13 14 Fixed Char ge 0 4 TG 5 1 4 T G 15+ TG Fixed Charge Janu ary 98 March -98 1. 6 8 1.85 2. 07 2.36 2.7 0 1.923 8 2. 8 9 7 7 2.8977 4.04 April -9 8 Octob er -98 1. 4 6 1.61 1. 8 2.05 2.4 0 1.923 8 2. 8 9 7 7 2.8977 4.04 Nov emb er 98 April 99 1. 6 8 1.85 2. 07 2.36 2.7 0 1.923 8 2. 8 9 7 7 2.8977 4.04 May -99 Octob er -99 1. 4 6 1.61 2 0 5 2 4 1 9 2 3 8 2 8 9 7 7 4.0 4 Nov emb er 99 April 00 1. 6 8 1.85 2. 07 2.36 2.7 0 1.923 8 2. 8 9 7 7 2.8977 4.04 May -00 Octob er -00 1. 4 6 1.61 1. 8 2.05 2.4 0 1.923 8 2. 8 9 7 7 2.8977 4.04
56 Table 3 8. Continued. Nov emb er 00 April 01 1. 6 8 1.85 2. 07 2.36 2.7 0 1.923 8 2. 8 9 7 7 2.8977 4.04 May -01 Octob er -01 1. 4 6 1.61 1. 8 2.05 2.4 0 1.923 8 2. 8 9 7 7 2. 8977 4.04 Nov emb er 01 April 02 1. 2 4 1.85 2. 07 2.36 2.7 0 1.423 6 2. 8 9 7 7 2.8977 4.04 May -02 Septe mber 02 1. 0 8 1.61 1. 8 2.05 2.4 0 1.423 6 2. 8 9 7 7 2.8977 4.0 4 Octo ber 02 Septe mber 03 0. 5 1.5 2. 05 2.05 2.8 3 1.7 2. 7 5 3.4 3 Octo ber 03 Dece mber 03 0. 5 1.6 2. 2 2.2 3.1 3.2 1.85 2. 9 3.6 3.25
57 Table 3 9. Ocoee (Multiple Rate Structures ). Water (per TG) Sewer Start End 1 15 16 25 26 Fixed Fixed 0 12 TG January 98 March -98 0.5 1 0.51 0.51 7.64 13.81 1.47 April 98 December 98 0.6 0.92 1.25 7.64 1 3.81 1.47 January 99 September 01 0.5 1 0.51 0.51 7.64 13.81 1.47 October 01 September 02 0.5 4 0.54 0.54 7.64 13.81 1.56 October 02 September 03 0.5 7 0.57 0.57 7.64 13.81 1.65 1 6 7 12 13 18 18 24 25 30 31+ October 03 March 04 0.8 4 1.05 1 .31 3.2 8 4.9 2 5.9 8 7.64 13.81 1.98 Table 3 10 Palm Bay (Three Blocks ). Water (per TG) Sewer Start End 1 4 5 10 11 Fixed Fixed 1 4 TG 5 10 TG Jan 98 3 Dec 0.75 1.6 3.8 7.65 7.8 1.00 2.20 Table 3 10. Continued. Table 3 11 Pam Beach ( Uniform ). Water Sewer Start End Per TG Fixed Charge 0 8 TG Fixed Charge January 98 December 98 3.92 11.49 3.36 12.75 January 99 November 00 3.84 11.26 3.29 12.5 December 00 February 01 3.41 12.31 2.86 10.33 March 01 December 01 3.36 12.12 2.82 10.17 January 02 3.33 12 2.79 10.07
58 Tabl e 312. Sarasota (Three Blocks) Water (per TG) Sewer Start End 1 4 5 8 9 12 13 18 19 Fixed Charge Fixed 0 10 TG January 98 June 98 1.78 2.58 4.35 7 10 13.92 14.34 6.41 July 02 December 02 2.58 2.58 4.35 7 10 13.92 14.34 6.41 January 03 2.66 2.66 4.48 7.22 10.31 14.3 13.09 6.61 Table 3 13. Seminole (Five Blocks) Water (per TG) Sewer Start End 1 10 11 15 16 20 21 30 31 Fixed Charge Fixed 0 15 TG January 98 September 03 0.65 0.95 1.25 1.5 1.75 6.6 11.35 2.59 October 03 0.65 1 1.75 2.5 3.5 6.6 11.5 2.63 Table 3 14. Spring Hill (Uniform) Water Sewer Start End Per TG Fixed Charge 0 6 TG Fixed Charge January 98 December 98 0.97 4.22 2.3 10.72 January -99 April -01 1.12 4.85 2.6 4 12.33 May 01 1 4.35 2.6 4 12.33 Tab le 315. St. Petersburg (Five Blocks ). Water (per TG) Sewer Start End 1 6 7 8 9 15 16 Fixed Charge Fixed Per TG January 98 November 99 1.49 1.87 2.53 3.36 4.11 6.04 2.31 December 99 April 01 1.61 2.02 2.73 3.63 4.44 6.37 2.44 May 01 November 01 1.69 2.12 2.89 3.81 4.66 6.72 2.57 December 01 October 02 1.87 2.34 3.17 4.21 5.15 7.16 2.74 November 02 October 03 2.14 2.68 3.63 4.82 5.90 7.87 3.01 November 03 2.44 3.05 4.14 5.16 6.72 8.34 3.19
59 Table 3 16. Tallahassee (Uniform) Water Sewer Start End Per TG Fixed Charge per TG Fixed Charge Jan 98 3 Dec 1.22 5.1 8.04 8.4 Ta ble 3 17. Tampa (No Fixed Fees) Water (per TG) Sewer Start End 1 4 5 10 11 per TG January 98 September 01 1.39 1.39 2.25 11.14973 October 0 1 1.39 1.6 2.7 12.83422 Table 3 18. Toho (Multiple Rate Structures) Water (per TG) Sewer Start End 1 2 3 -6 7 15 1624 25Fixed Charge Fixed 0 2 T G 3+ TG January 98 September 03 0 0.9 5 0.9 5 0.95 0.9 5 1.91 6.64 0 3.2 2 October 03 0. 5 1 1.1 1. 5 2.5 1.5 6.64 0 3.2 2
60 CHAPTER 4 CONSERVATION PRICE SIGNALS Background The concept of conservation rate structures is to compel the water customer to implement cost effective water conservation measures and practices (AWE 2009). Therefore any rate structure evaluation should include an evaluation of consumer water use and/or water conservation decisions made given different rate structures. As Whitcomb (2005) noted, estimating consumer response from the data set is complicated by a number of other variables influencing consumer behavior, including: precipitation patterns and sociodemographic factors. This study does not attempt to estimate price elasticities in different rate structure perods. Rather, it examines the price signals of th e different rate structures. The following indicators will be used. Each will be calculated for all of the rate structure periods in the sample. Average and Marginal Price: The Environmental Finance Center, of the University of North Carolina at Chap el Hill, conducted an extensive study (EFC 2008) of price signals in the state of Georgia. The study notes that there is no consensus within the literature about what type of signals residential consumers respond to. On one side of the argument is the id ea that consumers respond to marginal price, as they would with most goods. Others have argued that over time they actually respond to average prices, because they are not aware of the marginal prices. There is also the possibility that they respond to s ome combination of the two. Therefore both are calculated and reported in the Georgia study, and both are calculated here. Estimated Bill Savings of a 40% reduction in total water use: The Environmental Finance Center also used this measure as w ay to quantify price signal strength. The measure calculates the absolute and percent reduction in water bill a household would experience after reducing water use by 40%. This provides a simple, but effective way to evaluate how strong the conservation incentives actually are in a given rate structure period. Household Bill Distribution: I t is possible to examine price signals by dividing a household bill into its various components. For example, in a given rate structure period, 30% of the total bill might come from volumetric water and sewer charges, with 70% coming from fixed charges. In general, the greater percentage of a bill
61 is volumetric, the stronger the conservation price signal. A uniform rate structure may actually send stronger price signals than an inclining structure if it has low fixed fees and the inclining structure has high fixed fees. Methods Before any of these indicators could be calculated, the mean monthly household water use had to be estimated for the entire sample, over all utilities in all months. This overall average is approximately 9,360 gallons per month. This number is used to make the following calculations. Average and marginal prices: The average and marginal price estimates were based on an assumed, c onstant consumption level: 9,360 gallons per month. Both average and marginal prices are calculated in terms of price per thousand gallons. Estimated Bill Savings: These estimates are based on a starting point of 9,360 gallons, followed by a 40% reduction in use to 5616 gallons. The actual bill savings is calculated with the very simple formula: Equation 41 Actual Savings = TotBill9360 TotBill5616 The percent savings is calculated with the following formula: Equation 42 1 (TotBill5616) /(TotB ill9360) Utility Revenue Distribution: To calculate the total component of the bill that is fixed in a rate structure period, the fixed fee must be multiplied by the total number of billing
62 periods. In the example, the fixed fee was $21.45 each month for each household. So, the formula for fixed component of total bill is as follows: Equation 43 FIXEDBILLRatePeriod1 = (Number of Households) X (number of months) X 21.45 The following formula can be used to calculate the volumetric comp onent of the total bill, where h equals each household and m indicies months, $1.47/gal is the volumetric rates for the first volumetric rate blocks, $0.51/gal is the rate increase between the first and the second blocks, and 1,200 gallon is the water cons umption volume at which the second rate block starts. Equation 44 VOLBILLRatePeriod1 = h,m [ USEh,m 1.47 [ IF(USEh,m>1200,(USEh,m 1200),0) 1.47 ] + USEh,m 0.51 ] In this case, USEh,m refers to the total water used by household h i n month m Once these two components are calculated, it is possible to investigate the basic price signals being sent by each rate structure. As previously explained, the ratio of fixed to volumetric fees is a strong indicator of how conservation oriented a rate structure is. Results Average and Marginal Prices All average and marginal price estimates incorporated inflation, but they did not incorporate taxes or sewer rates. The average price estimates yielded unexpected results. Chart A 17 shows the twelve rate structure periods with the six highest and six lowest estimated average prices. It was expected that highest average prices would be observed for periods with block rates and five price blocks in particular, while the lowest
63 would be observed for uniform rates. In fact, the four highest average prices were observed from Palm Coast, which had a uniform structure. All of the lowest observed average prices occurred in periods with inclining block rates. Five of six lowest periods had rate struct ures with five blocks. Palm Coast, with the highest average prices, had the following rate structure: In the case of Toho, the volumetric water charge in even the fifth price block is less than the uniform volumetric charge used by Palm Coast. The fixed charges are also much less for Toho than they are for Palm Coast. It appears that average prices are much more related to the absolute magnitude of the fixed and volumetric charges than to the type of rate structure being used. The rate structure for Sem inole County, where the next lowest average prices were observed, tells a similar story. Obviously, in this case, the fixed charges are larger. But, all charges are still relatively small compared to Palm Coast. It is unclear why rates are so much higher for Palm Coast and Sarasota than they are for Seminole County and Toho. The five block structure used by Seminole and Toho indicate that these utilities consider incentivizing water conservation to be an important priority. But, perhaps due to higher rev enue requirements in other areas, most the other utilities had higher average prices. If consumers respond to average prices, then these results suggest that type of rate structure is relatively unimportant in sending conservation price signals. What m atters is the overall size of volumetric and fixed fees. If Bill is correct, utilities interested in promoting conservation should consider raising overall prices rather than adding additional price blocks.
64 Analysis of Marginal Price The results for mar ginal prices were mixed. To a certain extent, they supported the hypothesis that more price blocks would results in higher marginal prices. But, this relationship was imperfect. Chart A 18 demonstrates this. Four of the six periods with the highest obs erved marginal prices did have block rates, while four of the six periods with lowest observed marginal prices had uniform rates. But, two of the six lowest were from Seminole County, which had five price blocks. If consumers respond to marginal prices, then these results suggest that inclining price blocks may be superior to uniform rates in sending conservation price signals. But, there does not appear to be a correlation between the number of price blocks and the marginal prices. In other words, incli ning blocks may be better than uniform rates, but five block structures do not appear to be better than three block structures. If consumers respond to both average and marginal prices, then these results suggest that the type of rate structure cannot g uarantee strong conservation price signals. Every rate structure is unique. The implementation of a rate structure with more price blocks will not necessarily lead to higher marginal and average prices. The trends in average and marginal prices suggest that, over time, rate increases often do not keep up with inflation. Of the sixteen utilities, five seemed to show a downward trend in average and marginal prices. Escambia (Chart A 1), Miami Dade (A 6), Palm Coast (A 9), Sarasota (A 10), and Spring Hill (A7) seemed to show a general downward trend in both average and marginal prices. Therefore, in real terms, water gradually become less expensive for customers of those utilities during the study period. There were also some utilities that demonstrated t he opposite trend. St. Petersburg (A -13), Tampa (A -15) and Toho (A -16) all showed clear upward trends in
65 average and marginal price. So, for consumers in these areas, the real price of water increased during the study period. These rises in price may hav e been related to capacity expansion, alternative supply development, or changes in management. Another utility with interesting results was the Indian River County utility (Chart A 3). It only had two rate structure periods. In the second period, the average price decreased but the marginal price increased. From rate period one to rate period two, the fixed fee decreased from $11.2 per thousand gallons to $9.05 per thousand gallons. Meanwhile, all of the volumetric rates increased. The Tampa Bay Wat er Department was expected to send the strongest price signals. With three price blocks and no fixed fees, the incentive to conserve would appear to be stronger in this case. But, the results for Tampa yielded average and marginal prices that were not among the highest or lowest observed (Chart 315). Even with no fixed fees, the volumetric fees for Tampa were not particularly high in any of the three price blocks. Household Savings from Water Use Reduction Tables A -4, 35, and A -6, and Charts A 19 and A-20 summarize the results for this section. These estimates do not include inflation, taxes, or sewer fees. The reduction in household bill is estimated based on a 40% reduction in household use at the 9,360 gallon level (i.e. to the level of 5,616 gall on). These results contradict the results of the previous section. While, the marginal and average prices suggested that rate structure may be relatively unimportant, the incentive results suggest otherwise. Specifically, the results indicate that in clining blocks rate structures with more than three price blocks send stronger conservation
66 price signals than uniform structures or inclining structures with three or less blocks. Chart A 19 illustrates this graphically. The chart is a scatterplot with each point representing the total savings (in dollars) and percentage savings in water bills when a customer reduces use by 40% (for different rate structure periods). The blue line represents the 40% mark. Every point on this line represents a 40% reduc tion in water bill. In this scatterplot, the upper right quadrant represents the area where actual and percent reductions in water bill are the highest, and hence, the price incentive to reduce consumption is highest. And, the analysis of water rate struc tures show that all the points in the quadrant or on the edge of it represent inclining block rate structure periods with more than three price blocks. Interestingly, the potential bill savings associated with uniform structures appear to be stronger than those associated with structures that have three price blocks or less. The conclusion is supported by Chart 320, which averages the results summarized in the scatterplot. On average, the actual and percentage savings were greater for those periods with more than three price blocks. There appears to be little or no difference between uniform periods and periods with three or less blocks. If anything, the uniform periods may actually be sending stronger price signals. The actual reduction in price was h igher on average for uniform periods than for periods with three or less blocks. This section, in opposition to section on average marginal prices, suggests that rate structure is, in fact, very important in determining conservation rate effectiveness. Th e results indicate that only inclining structures with more than three price blocks are likely to provide a greater incentive to conserve water. Further empirical research could provide additional information about this potential issue.
67 Revenue Distributi on Charts A -21 through A -34 represent the results for this section. Revenue is estimated based on household bill for every household in each rate structure period. Several utilities start at period 3 or 4 because no usage data were available for early p eriods. Miami Dade was the only utility to utilize seasonal rates based on winter and summer months. To a certain extent the price incentive results are supported by the revenue distribution results. Once again, the evidence in this section supports the argument that rate structure is important and relevant to conservation rate effectiveness. Chart A 21 shows the top and bottom six rate structure periods in terms of revenue generated from volumetric fees. All of the top six periods have inclining block rates, and four of the six have five price blocks. On the other end of the spectrum, five of the six periods have uniform rates. The percent of revenue from volumetric charges is an important indicator of how conservation oriented a rate structure is. U sing the 70% criterion, 70% utility revenue in a given period must come from volumetric sources to be considered a conservation rate. In this sample, more than half of the sixteen utilities generated more than 70% of estimated revenue from volumetric sour ces during at least part of the duration of the study. Another conclusion that can be drawn from these results is that changes in rates rarely have large effects on revenue distribution unless the type of rate structure changes. In most of the cases, t he pie charts changed very little from one period to the next. One exception is Ocoee (Charts A 27 A,B,C).
68 During the four periods where Ocoee used uniform rates, about 59% of revenue came from fixed fees. But during the two periods where it used three b locks and six blocks, it generated 48%, and 44%, respectively, from fixed fees. At least in terms of revenue distribution, the inclining block structures were significantly more conservation oriented than the uniform structures. Another interesting re sult is that utilities often generate a large proportion of residential revenue from sewer fees. Tampa, for example, generated more than 70% of its revenue from sewer fees in both of its rate structure periods. Tampa had no fixed fees. But, some other utilities also generated large amounts of revenue from sewer fees. Tallahassee, for example, generated about 56% of its revenue from volumetric sewer fees alone. Miami Dade also generated more than 50% of its revenue from volumetric sewer fees in each of its rate structure periods. These results confirm that sewer rates are an important component of residential water bills. Much revenue is generated from sewer fees, so it is important that they also are incorporated into a conservation pricing strategy. If percent of revenue from volumetric sources is a good indicator of how conservation oriented a rate structure is, then the results of this section support several key conclusions. First, rate structure does matter. Inclining block rates, especially with more than three price blocks have a tendency to generate a revenue stream that largely comes from volumetric sources. Second, minor changes in rate structure rarely have a major impact on price signals. A change in the type of rate structure (like th e ones implemented by the Ocoee utility) may be necessary to affect revenue distribution. Third, analysis of revenue distribution results do not necessarily yield the same results
69 as other types of price signal analysis, especially when compared to average and marginal prices. Summary The results of Chapter 3 do not tell one, simple story. Rather, they suggest different conclusions and interpretations depending on the factors being studied. In the case of average and marginal prices, there is no strong ev idence that any one type of rate structure sends stronger price signals than any other type. The Price Incentive results (i.e. the analysis of household savings due to water use reduction), on the other hand, suggest that what really matters is the number of price blocks. Two and three block inclining structures may not send stronger price signals, but five block structures do. Finally, the revenue distribution results support the conclusion that rate structure does matter for sending price signals. Blo ck structures of any kind seem to send stronger signals than uniform structures, while more price blocks seem to make this signal even stronger. Overall, the individual elements of any rate structure define the strength of its price signals. High fixed fees can, in some cases, increase average prices. But, they may reduce the price incentive to conserve water. Also, adding price blocks may increase the incentive, but they may affect only a very small number of households. In Florida, individual utilities have the authority to decide what a conservation rate structure effective. Hopefully, in the future, a consensus will emerge about how effectiveness should be defined. The analysis in the chapter focus on Conservation Rate effectiveness, as defined by the strength of conservation price signals. An entirely separate question is, do consumers respond to these signals? Answering this question is beyond the scope of
70 this study. Some empirical evidence, including the Whitcomb (2005) study suggests that c onsumers do respond. But, ambiguous billing procedures, inelastic demand for water, and other factors surely have an impact. More research is clearly needed in this area.
71 Figure 41. Reduction in Bill after 40% Reduction in Use 0 5 10 15 20 25 30 35 40 45 50 0 2 4 6 8 10 12 14 16Percentage Reduction in BillActual Reduction in Bill Uniform Three Blocks or less >Three Blocks
72 Table 6 1. Customer Income Groups. Name Income range Income group 1 <$15,000 per year Income group 2 $15,000 $29,999 per year Income group 3 $30,000 $49,999 per year Income group 4 $50,000 79,000 per year Income group 5 $80,000 $100,000 per year Income group 6 >$100,000 per year Table 6 2. Estimated Hardship Thresholds for Hypothetical Florida Household Year Monthly Househol d Income on the Poverty Line Hardship Threshold 1998 $12,138 $40.46 1999 $12,367 $41.22 2000 $12,584 $41.95 2001 $12,999 $43.33 2002 $13,357 $44.357 2003 $13,564 $45.21 Table 6 3. Melbourne Rate Structure: 1998. Water Sewer Taxes Per TG Fixed Charge per TG Fixed Charge per TG 2.07 3.28 3.34 4.63 None Table 6 4. Seminole County rate Structure: 2003 Water (per TG) Sewer Taxes 1 10 11 15 16 20 21 30 31 Fixed Charge Fixed 0 15 per TG 0.65 0.95 1.25 1.5 1.75 6.6 11.35 2.59 4% on water
73 CHAPTER 5 CONSERVATION RATES, REVENUE, AND REVENUE STABILITY Background This chapter focuses on the relationship between rate structure and utility revenue. One potential pitfall of using conservation rates is reduction of utility revenue, as discussed, for example, by Jones (1974) and Bhatt and Cole (1985). A contrary opinion is expressed by the National Regulatory Research Institute (1994), which argues that because water demand is inelastic, increased rates actually have the potential to increase rev enue. In addition to the potential effects on average revenue, conservation pricing can also cause another problem for utilities: revenue instability. McLarty and Heany (2008) note that utility companies generally have very high up front fixed (inve stment) costs that are financed by borrowed capital. As a consequence of this, utilities must dedicate a fairly large share of revenue to service the debt that they have. As a result, utilities have a vested interest in revenue stability. By charging a large fixed fee, utilities can cover their fixed costs. A smaller volumetric fee can be used to cover variable costs. Unfortunately, rate designs with large fixed fees do not provide a strong incentive for water conservation. Similarly, Chesnutt et al. ( 1993) explain that increased revenue variability leads to the possibility of increased borrowing expenses and greater planning/informational expenses. Utilities are concerned about monthto month revenue variability and about the unpredictability of the bills paid by each household. With uniform rates, month-to month variability reflects only monthly changes in use. With inclining block rates,
74 changes in use can cause households to move between different price blocks. This can generate, over time, more variability in household bills. Hypotheses Two hypotheses are tested in this chapter. First, the study tests whether conservation rate structures have a negative impact on overall revenue, as defined by average monthly revenue. Secondly, the study investigates whether rate structure affects revenue variability. Variability in revenues collected from each household (i.e., household bills), and changes in month-to month utility revenue are examined. Evaluation Criteria Several different indicators are used to test these hypotheses. Average Monthly Revenue To test the first hypothesis, average monthly utilities revenues are estimated based on the households in the data sample. Specifically, using water rates, sewer rates, and water use, a bill i s calculated for each household in each month. The sum of all household bills is used as a proxy for monthly utility revenue. Further, average monthly utility revenue in each rate structure period is calculated as a sum of utility monthly revenues divided by the number of months (see chapter 3 for esimtiation details). The average monthly utility revenues are compared across rate structure periods and across utilities. A few periods are omitted due to lack of data. Specifically, the first two rate struc ture periods for Lakeland, the first four periods for Melbourne, and the first period for Palm Coast did not have any water use data recorded. In reality, utility revenue includes commercial, institutional, and other sources of revenue. In this study, r evenue is estimated in each month based on only the
75 households used in the sample. Therefore, these estimates only capture changes in price or use. They do not reflect changes in the number of households. Household Billing Variability One argument against conservation pricing is that it may cause revenue variability. To test this, the standard deviations of monthly household bills are calculated and plotted for each utility and each rate structure period. The changes in the standard deviations are e xamined for each utility over time. Frequency Distribution of Household Bill Variability in household bills is analyzed graphically, using frequency distributions. Each household bill in each month is used as individual observation to be plotted. Freq uency distributions are examined for three types of rates: uniform, inclining block with three or less blocks, and inclining with more than three blocks. Intuitively, these distributions are expected to be skewed left, with several outlier households using much more water (and therefore having much higher bills) than the average household. The frequency distribution for three or less block structures is expected to be more skewed than the distribution for uniform structures, while the distribution for rate structures with four or more block structures is expected to be the most skewed. This is because, at high levels of water use, the marginal price of water is typically much higher for structures with many price blocks. Regression Analysis for Household B ill Variability Once standard deviations in household bills are estimated for each month and each utility, statistical analysis is used to test for significant differences. Specifically, two OLS regressions are used to determine if there is any statisti cal evidence that the use of inclining block structures results in greater household bill variability.
76 Overall Variability in Monthly Revenue Finally, to examine the effect of the rate structure on month-to -month variability of utilities revenues, monthly revenues are compared for three groups of rate structures: uniform, inclining block with three or less blocks, and inclining block with more than three blocks. Methods Average Monthly Revenue The average monthly revenue numbers cannot be compared between utilities because each the number of households included into the data sample was different for each utility. In other words, while Equation 3-7 does account for the length of each rate structure period (by dividing by the number of months) it does not account for the number of households in each sample. Nevertheless, the average revenue figures are useful for comparisons between the rate structure periods of a given utility. They allow for an examination of revenue over time, and they make it po ssible to examine the revenue impacts for a given utility whenever it changes rate structure. Frequency Distribution of Household Bill In order to examine the frequency distributions graphically, three histograms are used: 1) monthly household bill for every household in every month given uniform rate structures; 2) monthly household bills for every household bill in every month that used block rate structures with three or less price blocks ; and 3) monthly household bills for every household bill in every month that used block rates with four or more price blocks. Household Billing Variability In order to measure billing variability in each rate structure period, the standard deviation of household bills was calculated:
77 Equation 54 = [1 1 xix )2] where N = number of observations (i.e. the number of households times the number of months in rate structure period) xi = total bill paid by a given household in a given month x = the mean household monthly bill during th e rate structure period Standard deviations are compared between rate structure periods of a given utility, and between utilities. Regression Analysis for Household Bill Variability Two OLS regressions are used to examine the effect of rate structure on standard deviation of household monthly bills. In the first case, the explanatory variable is the number of price blocks in a rate structure: Equation 55 Yi = 1 + 1Xi + where Yi = the standard deviation in monthly household bills for rate structure period i Xi = the number of price blocks in the period (for uniform periods, the value of this variable is 1) 1 = the coefficient for X, as estimated by OLS regression The statistical hypotheses will be as follows: H0 : 1 = 0 (the number of price blocks has no impact on the standard deviation of household bills.) Ha 1 on the standard deviation of household bills) A second regression was used to test for the significance of the type of rate structure, where rate structure is a binary, categorical variable. The same general form
78 of the model is used as above. The variable Xi is binary with the value of zero for uniform rate structure period, and the value of one for inclining block structure periods. Overall Variability in Monthly Revenue To estimate variability in overall revenue for each of the three categories of rates (uniform, three blocks or less, and four or more blocks), the standard deviation of monthly utility revenue is estimated using the following formula: Deviation of monthly utility revenue is estimated using the following formula: = [1 1 xix )2] N = number of observations (# of months in every rate structure period that used this type of rate structure) xi = total revenue generated by a given utility in a given month x = the mean revenue during the rate structure period Once calculated, these three standard deviation figures make it possible to compare overall variability during periods where each the three types of rate structure were used by utilities. Results Average Monthly Revenue Chart B 33 shows the six h ighest and the six lowest recorded figures for monthly revenue. The bars in the charts are color coded for consistency. Blue indicates a period with uniform rates, yellow indicates an inclining structure with three or less price blocks, and red indicates a period with four or more blocks. The results suggest that there is no clear correlation between type of rate structure and average monthly revenue. Chart B -33 shows that four of the six periods with highest average monthly revenue, and five of the six periods with the lowest average monthly revenue had inclining block rate structures. There does not appear to
79 be trend related to the number of price blocks and the average monthly revenue. In addition, it is difficult to draw any conclusion because the monthly revenue figures largely reflect the number of households in the sample group of each utility. For example, the Toho Water Utility only had a sample of 61 households, by far the fewest number. Not surprisingly, its two rate structure periods h ad the lowest average monthly revenue. Despite this limitation, the average revenue figures are useful for exploring the dynamics of revenue within each utility, where the number of households is constant. The most interesting case is Ocoee, where the typ e of rate structure changed several times (Chart 4 13). Most of Ocoees periods had uniform rates. Period 2 and Period 6 were the exceptions. In Period 2, Ocoee had a three block structure, and in Period 6, it had a six block structure. The two periods with block rates were two highest periods of average monthly revenue. Period 6 had average revenue of $17,621 which was significantly higher than the highest uniform period, period 3 ($11,628). So, in this case, the implementation of inclining block pri cing did not have an adverse effect on utility revenue. If anything, revenue increased with block pricing. The only other utility which changed the type of rate structure used was Toho. It started out with a twoblock structure, and changed to a fiveblo ck structure. Once again, this change seems to be correlated with an increase in average monthly revenue ($1,465 to $2,271). And, this is despite the fact that the fixed fee for water actually decreased, from $1.91 per month to $1.50 per month. These tw o utilities provide some evidence that adding price blocks does not have an adverse effect on overall utility revenue, and may actually have a positive effect.
80 Household Billing Variability The standard deviation estimates for every utility and every r ate structure periods provide some support to the argument that inclining block pricing may lead to increased variance in individual household bills. Chart B -36 shows the six highest and the six lowest standard deviation figures. All six of the highest w ere reordered during periods of inclining block rates. Moreover, four of the six were in periods with more than three price blocks. On the other side of the chart, four of the six lowest standard deviations were recorded for periods with uniform rates. In addition, the two periods with by far the highest recorded variability were Tampas two rate structure periods (134.96 and 137.14). Tampa was the one utility with no fixed fees of any kind. As discussed in chapter 3, this 100% volumetric rate structur e sends very strong price signals by providing a large incentive to reduce consumption. But, the results in this section suggest that there may, in fact, be a tradeoff associated with a rate structure of this kind. Fixed fees are designed to add stabili ty to the revenue stream. In the case, the lack of fixed fees probably contributes to a more variable stream of revenue. Other results support this conclusion. In the case of Ocoee (Chart 414), the two rate structure periods with the highest standard deviations were the periods with inclining block rates. Period 6, with six blocks was substantially higher than any of the others. So, in this case there once again appears to be a tradeoff. By adding block rates, the Ocoee utility department made its rat e structure more conservation oriented. It also increased its average monthly revenue. But, these changes came at a cost: the bill paid by any given household from month to month became more variable. The results for the Toho Water Authority (Chart B 3 2) also support this conclusion. After moving from two blocks to five blocks, the standard deviation in
81 revenue increased from 9.78 to 11.42. This may not be a huge jump, but it is significant nonetheless. Frequency Distribution of Household Bill The histogram of household bills for rate structure periods with three or more price blocks (Chart B 38) clearly shows a greater variation in household bills for rate structure periods with more than three price blocks. More surprisingly, the histogram of household bills for rate structure periods with three or less price blocks (Chart B 37) appears to show the narrowest and least variable distribution of the three. The histogram for uniform periods (Chart B -36) is somewhere between the two. It is not clear why the household bills are more variable for three or less price blocks than they are for uniform rates. Regression Analysis for Household Bill Variability Descriptive statistics for monthly household bill given uniform and inclining block r ate structures is presented in Tables B -1 and B -2. One interesting result is that, despite differences in overall sample size and utility characteristics, the mean household bills were fairly similar given different rate structures. Unexpectedly, the me an for the uniform periods was slightly higher than the mean for the inclining block periods ($58.73 compared to $47.55). The variance and standard deviation was higher for the block periods than for the uniform periods (48.73 compared to 41.23). This ag ain suggests that the inclining block structures tend to result a more variable payments (bill) for utility customers than uniform rate structures, which may contribute to uncertainty or unpredictability of utility revenue. However, it is important note th at the r2 was quite low (about 0.065). So, this basic model has little explanatory power. The
82 number of price blocks alone is not enough to predict how variable a revenue or billing distribution might be. In the second OLS regression, the explanatory v ariable is a dummy variable representing inclining block rate structures. The variable value is equal to zero for the months with uniform rates, and one for the periods with block rates. Once again, the standard deviation of revenue was the dependent var iable (Table B 4). With a pvalue of 0.19, there is fairly strong evidence that the type of rate structure affects variability. The coefficient (16.26) was positive, indicating that inclining block rate (with unspecified number of price blocks) tend to g enerate revenue streams that are more variable than uniform structures. Also, once again, the r2 was low, about 0.096. This suggests that type of rate structure is not by itself enough to predict how variable the revenue stream might be. Overall Vari ability in Monthly Revenue Chart B 35 displays the standard deviations of monthly utility revenues given a) uniform rates, b) inclining block rates with three blocks or less, and c) inclining blocks rates with more than three blocks. The smallest standard deviation was estimated for the months with uniform rates, the second for three or less price blocks, and the highest standard in monthly revenue was recorded for periods with four or more price blocks. Conclusion The results of the chapter suggest two conclusions. First, no evidence was found that adding price blocks decreases overall utility revenue. In fact, it is possible that inclining block structures actually tend to generate more revenue than uniform rates. That is, it does not appear that th e objective of cost recovery is more difficult for utilities that use inclining blocks.
83 However, there is some evidence that the tradeoff between price blocks and revenue variability does exist. Since revenue risk does impose costs of its own, policy ma kers and utility managers should consider revenue variability as a potential downside of conservation oriented water rates. With multiple price blocks, utility revenue is estimated to be more variable, especially at the household level. Household water u sage seems to fluctuate from one price block to another, making the revenue stream more difficult to forecast. In terms of monthto month revenue variability, there is also some evidence that inclining block rates (especially with more than three blocks) lead to revenue streams that vary more from month to month. Future Research Needs The results suggest several areas of analysis where further research would be particularly useful. First, more advanced statistical procedures or econometric modeling wou ld help make a stronger case for or against the conclusions presented here. In the future additional factors should be examined, such as weather and demographic variables. Simple linear regressions may not be sufficient to prove that a relationship betwe en rate structure and revenue does or does not exist. Also, more research is needed to estimate the precise cost of variability in household billing and in overall revenue. Uncertainty can be costly, but more information is needed about the exact nature o f this cost. Policy makers cannot adequately evaluate the advantages and disadvantages of conservation rates until more evidence is available about the costs of uncertainty. Time series studies of individual utilities may provide more information about r evenue and revenue stability. Cases like Ocoee, where changes occur in the type of rate
84 structure, present a kind of natural experiment to examine the effect of rate structure on utility revenue. If the sample of households is held constant over time, as it was in this case, large changes in revenue or revenue stability can be attributed to changes in rate structure.
85 CHA PTER 6 CONSERVATION WATER R ATES, AFFORDABILITY, AND FAIRNESS Background: Conflicting Opinions Another potential rate design issue is the tradeoff that may exist between water conservation and affordability. It is possible that the implementation of conse rvation rates could make water too expensive for low income consumers to afford. As more stringent environmental requirements are adopted, affordability problems become heightened. (EPA 2002). A related issue is that of rate structure fairness. Some pricing may be viewed as unfair by the public if they result in a transfer of wealth from one group to another. Since privatization, water pricing policy has been gradually moving toward economic equity the principle that users of utility service sh ould pay, as near as possible, the costs they impose on the system. (Bakker 2001, p. 147). This concept of economic equity in rate design is sometimes referred to more generally as fairness. In any case, affordability and fairness are important criteria for effective rate design. Poorly designed conservation rate structures can potentially lead to inequitable billing of different customer groups (AWWA 2000). Renwick and Archibald (1998) find that water use of low income customers is more responsive to price increase than the water use of high income customers. These results suggest that price policy will achieve a larger reduction in residential demand in a lower income community than in a higher income community, all other factors held constant. Results also suggest that if price policy is the primary DSM [demand side management] instrument in a particular locale, lower income households will bear a larger share of the conservation burden (p. 357). In contrast, Agthe and Billings (1987) argued that i nclining block rate structures
86 can lead to greater distributional equity. The authors show that by making price blocks steeper, a utility could increase the incentive to conserve without adding any price burden to low income users. Fairness is somewhat sub jective, because it is related to public perception. Some argue that conservation water rates that rely on inclining block structures are less fair than uniform rate structures. Uniform rates make every user, regardless of income or water use, pay the exact same volumetric charge per unit of water. With inclining block rates, high volume users pay much higher volumetric charge per unit of water they use (Heaney 1998). Morgan (1987) adopts Prices (1980) approach to measuring fairness of a water rate struc ture by comparing the share of the goods received by different users relative to the share of the total costs paid by the same users (p. 144). Analysis of equity and fairness is made possible by the income information included in the Whitcomb (2005) data set. In the data set, there are six household income groups (Table 61). Even though precise household income figures are not reported in the Whitcomb data, these income brackets make it possible to compare the equity and fairness of the different rate structures in the sample. Methodology: Measuring Affordability And Fairness Of Water Rate Structures Affordability To examine the impact of different water rate structures in low income customers, the share of income that the low -income customers spend on water is analyzed given different water rate structures. A number of studies (EPA 2002, Saunders et al. 1998) have explored life -line water rates as a policy tool to protect low income consumers and to make basic (non-
87 discretionary) indoor water use (drinking, washing, sanitation, etc.) affordable. However, none of the utilities in the Whitcomb sample used low -income discounts or lifeline rates during the years of the study. So, as an alternative, this section of the thesis explores the idea that water prices can generate hardship for low income households. To do this, a basic assumption must be made about threshold of water affordability. Miniaci R et al (2008) cites several studies which suggest that water bills are affordable when they account for anywhere from 1% to 3% of total household income. Saunders et al (1998), concludes that water is affordable to a household as long as the water and wastewater bill combined do not exceed 4% of total household income. And, some experts have estimated tha t, based on this 4% guideline, as many as 18% of US households may face water and sewer fees that exceed hardship levels (WIN 2000). This part of this thesis will use the 4% guideline for water/wastewater affordability. It will be assumed that any combined water and wastewater bill constituting more than 4% of household income will exceed the households water price hardship threshold. Water hardship levels are estimated for a hypothetical Florida household with an income at the Federal poverty line. T he purpose of this procedure is to compare how much water a hypothetical household could buy (in each rate period) without exceeding the hardship level. A three-step process is required to calculate hardship thresholds for each year. First, the Federal Po verty line must be recorded for the year in question. The Federal Poverty line is estimated by the Department of Health and Human Services based on the official poverty line guidelines (DHHS 2008). The official poverty guidelines take into account annual inflation.
88 The second step involves estimating the precise poverty level income of an average hypothetical household in Florida. The Federal Poverty line in a given year starts out with a base figure for a one person household. An added income allotment is given for each additional member of the household. For the purposes of this study, it was assumed that the average hypothetical Florida household had 2.46 people (US Census 2000). For example, in 1998 the base figure was $11,942/year. The additi onal person allotment was $2,800 per person per year. Hence, the poverty line for the hypothetical Florida household for 1998 equals an annual household income of $12,138. The income of a hypothetical household is converted to a monthly income by dividing the annual figure by 12: Monthly Average household Poverty Line income = [(Base Income) + (1.46) x (additional person allotment)] / 12 The third and final step involves using the assumption that a household water bill (including sewage charges and taxes) exceeds the hardship level if it accounts for more than 4% of total income. Therefore, the hardship thresholds are calculated by taking 4% of the monthly average household poverty line income (Table 2). Next, the maximum quantity of water the hypothetical household could use without exceeding the hardship level is calculated, based on water rates, sewer rates, and taxes charged by water utilities. Two examples illustrate the method. The first, from Melbourne in 1998, is a uniform structure wi th no taxes. The rates are as follows: In, 1998, the hardship threshold for a hypothetical 2.46 person household was $40.46. In Melbourne, the water and sewer fixed fees account for $6.62 per month,
89 leaving $33.84 for volumetric charges. The total volumetric charges (for water and sewage) were $5.41 per thousand gallons. Dividing the remaining monthly income ($33.84) by the total volumetric charges ($5.41) provides a figure representing the maximum use without exceeding the hardship level. In this cas e, the division equals 6.26. In other words, in 1998, the hypothetical 2.46 person household on the poverty line could purchase 6.26 thousand gallons of water without exceeding the hardship level. A second example, from Seminole County in 2003, illustrates a more complex, inclining block rate structure: In addition to these fixed and volumetric charges, Seminole County added a 4% sales tax on water only. The hardship threshold in 2003 was $45.21. The fixed fee for sewage was $11.35, and the fixed fee f or water, with the sales tax included, was $6.86. Hence, the total monthly fixed fee was $18.21, leaving $27.00 (=$45.21$18.21) for volumetric charges. The first price block was $0.65 per thousand gallons, for the first 10,000 gallons used. The volumet ric sewage charge was $2.59 per thousand gallons for the first 15,000 gallons of water used. The total volumetric charges were $3.24. Dividing the remaining monthly income ($27.00) by the total volumetric charges ($3.24) provides a figure representing th e maximum use. In this case, the division equals 8.33. So, in 2003, the hypothetical 2.46 person household in Seminole County could consume 8.33 thousand gallons of water per month without exceeding the hardship threshold. It is important to note that i f this number (8.33) had exceeded 10, then an additional calculation would have been necessary to account for the higher fee in the second price block.
90 These two results allow for a basic comparison. From the perspective of low income households (househ olds on or near the poverty line), the rate structure for Seminole County in 2003 was preferable to the structure for Melbourne in 1998. Accounting for inflation, the hypothetical household on the poverty line could purchase 8.33 thousand gallons per mont h in 2003 Seminole County, and 6.26 thousand gallons in 1998 Melbourne, without exceeding the hardship threshold. Fairness Following Morgan (1987), the Gini coefficient estimates and Lorenz curves are used to measure the fairness implications of var ious rate structures. These measures are often used in development economics to examine income inequality. Morgan (1987) demonstrated that they can be used to examine water rate equity A Lorenz curve for a specific water rate structure maps the proport ion of water use by different customer income groups against the proportion of utility revenue collected from these income groups. Each Lorenz curve is compared with the perfect equity line, where each customer income group contributes equal shares to the total utility water supply and the total utility revenue. In turn, the Gini coefficient is defined graphically as the ratio of the area between the Lorenz curve and perfect equity line and the total area under the perfect equity line. All Gini coefficient values are numbers between zero and one (Morgan 1987). The lower the coefficient, the fairer a rate structure is. Results Affordability: Evaluating the Hardship Threshold Results Wide variation was observed between the low income affordability of the various rate structures. The lowest maximum use (without exceeding the hardship level) was 1.72 thousand gallons per month, for Tallahassee in 1998 (Chart C -14). The highest
91 observed was 11.38 thousand gallons per month, for Lakeland in 2001 (Chart C 4). A closer look at these two rate structures reveals an interesting picture. Tallahassee had a uniform structure: At first glance, these charges seem fairly low. However, Tallahassee had a 37.4 sales tax on overall water use, including sewage. This was by far the highest tax rate, and was an important part of the reason that low income consumers in Tallahassee could buy so little water without exceeding hardship levels. Lakeland in 2001 had an inclining, three block structure: Both the fixed and volumetr ic fees were relatively low, and there were no taxes on water. This explains the high purchasing power results. The volumetric water fee does increase in the second price block, but almost all consumption for low income consumers occurs in the first pric e block. The comparison of maximum potential water consumption without exceeding the hardship level purchasing power shows that several utilities had rate structures that resulted in consistently high maximum use (Chart C 17). Toho and Lakeland, in part icular, stand out. It is worth noting that both of these utilities used inclining block structures throughout the study period. On the other hand, the rate structure with the lowest estimated maximum use was Tallahassee. It started out with 1.09 and end ed 1.72. It used the same uniform structure the entire study period. Fixed fees and high taxes (a 37.5% sales tax) explain the results. Another interesting case is Tampa (Chart C 15). Tampa used an inclining block structure without any fixed fees during the entire duration of the study. Without fixed fees, the maximum use for low -income households was expected to be relatively high.
92 However, this expected result was not actually observed. In fact, the estimated maximum uses, 3.42 and 3.46 thousand gallons at the beginning and the end of the study period, was relatively low compared to most of the other utilities. This was probably due to the fact that the volumetric sewer fees were quite high (11.15 per thousand gallons) as were the taxes (10% sales tax on water). Tampa illustrates that the purchasing power of low -income households may be relatively low even when there are no fixed fees. Overall, these results suggest that inclining block rate structures are not necessarily worse for low income households that uniform rate structures. In this sample, several of the utilities with the highest observed maximum uses had inclining block rates, while the utility with the lowest, Tallahassee, had uniform rates. The effect of a rate structure on low -inc ome households clearly depends on a number of factors, including: the relative size of fixed fees, the volumetric charges for wastewater, and taxation. Rate Structures and Inflation The results also provide a good picture of how in inflation can affe ct the affordability of water. For several utilities, the rate increases seemed to keep up with inflation. For example, the maximum use of low -income households in Hillsborough County (Chart C -2) started out at 3.91 thousand gallons per month and ended at about 3.94 thousand gallons per month. The estimated maximum use did fluctuate throughout the various rate changes, but never by more than a thousand gallons per month. Ocoee (Chart C 7) was a similar example. It had six different rate structure periods, but the estimated maximum never changed by more than a thousand gallons. It started out at 9.6 thousand gallons per month and ended at 8.82 thousand per month.
93 Some utilities clearly increased rates by amounts that exceeded inflation, at least from the perspective of residential consumers at the poverty line. Melbourne (Chart C 5) and St. Petersburg (Chart C -13) are the two utilities that show the clearest downward trend in maximum use. Melbourne started at 6.02 and ended at 4.93 thousand gallons per month. St. Petersburg started at 10.77 and ended at 8.22 thousand gallons per month. Of course, these fluctuations were fairly modest. Both utilities had rates that provided a higher maximum use for low income households than some of the other utilit ies. There were some utilities where the maximum use for low income consumers increased as a consequence of rates that did not keep up with inflation over time. Miami Dade (Chart C 6), for example, showed a general upward trend in observed values throughout its 12 different rate changes. For Miami, the maximum use started at 4.63 and ended at 7.24. Palm Coast, Spring Hill, and several others also demonstrated this trend. Two utilities (Palm Beach and Tallahassee) did not change rate structures at all d uring the entire period. These utilities, as expected, showed an increase in maximum use over time because the Federal poverty line increased with inflation while the rates stayed the same. Rate Structure and Geography of Utilities Figure ES 1 shows the location of each utility in the sample. The utilities were not distributed evenly between the water management districts. The Northwest WMD had two, the Suwannee River WMD had zero, the St. Johns River WMD had six, the Southwest WMD had six, and the South Florida WMD had two. The SWFWMD was interesting. It had the utility with the highest maximum use, Lakeland, and a utility with one of the lowest, Sarasota. The other utility with a particularly high maximum use for low income customers was in Toho (Ch art C -16). It
94 was in the St. Johns Water Management District. But, this district is also where Palm Coast and Indian River County are locatedtwo utilities that had fairly low maximum uses. The lowest purchasing power was in Tallahassee, in the Nort hwest District. The only other utility in this District, Escambia, also had fairly low observed maximum use (Chart C 1); it fluctuated between 3.25 thousand gallons per month and 3.46 thousand per month. In general, the observations do not seem to be corr elated with which Water Management District the utility is in. But, the Northwest district may the one exception. Both of its utilities had consistently low maximum use observations. However, with only two utilities, it is difficult to draw any conclusi ons about how government policy may have affected rate structure in the Norwest district. In general, the results do not seem to be correlated with which Water Management District the utility is in. In other words, the regulatory differences between the W MDs do not affect the equity of utility rate structures. WMDs do not have the legal power to regulate water pricing, which may explain this result. Another geographic factor that might be important is coastal location. It is possible that utilities locat ed on the coast are more concerned with conservation because of saltwater intrusion. If this is the case, the coastal utilities may exhibit different patterns than inland utilities. There are fewer inland utilities (five) than coastal utilities (eleven) i n the sample. A close look at them does not reveal any trend that might be related to concerns about saltwater intrusion. For example, two of the inland utilities (Escambia and Tallahassee) had some of the lowest hardship threshold results. But, two of the others (Ocoee and
95 Toho) had some of the highest estimated thresholds. Therefore, geographic location once again does not appear to help explain the results. Rate Structure and Demographics In addition to geographic factors, it is possible that dem ographic variables affect rate structure equity. One might hypothesize that the utilities with observed maximum use for low income consumers would be located in counties with relatively high median income and a relatively low proportion of households below the poverty line. The United States Census estimates median household income (Chart 18) and number of households below the poverty line (Chart 19). The utility with the lowest maximum use, Tallahassee, had a relatively high percentage of households below the poverty line: 12.1 (Census 2000). Also, its median household income ($38,791/year) was in the middle of the household income values for other counties (Census 2000). So, the income distribution in Leon County (where Tallahassee is located) does not seem to explain why its rate structure provides such a low purchasing power to low income households. In the case of Tallahassee, high taxes are a major part of the picture, but it is not clear why the sales tax on water is so high. Sarasota had the sec ond lowest observed maximum use, and the income distribution in Sarasota County may help explain why. The county has a relatively high median income ($42,265) and it had the lowest percentage of households below the poverty line (7.3%). The highest maximum use was observed in Lakeland, which had a relatively low median income ($35,438) and a relatively high percentage of households below the poverty line (13.2). The other utility with particularly high purchasing power was Toho,
96 which, surprisingly, had a median income of the middle of observed counties ($38,309/year) and a relatively low percentage of people in poverty (9.5). Lifeline Rates and Affordability An important question related to customer equity is the so -called life line amount of water a household needs for basic needs: washing, cooking, bathing, etc. One study, Mizgalewicz (1991) suggested that in Arizona, the lifeline amount is about 6 ccf, or 4,448 gallons per person per month. In our hypothetical household with 2.46 people, that wou ld be equal to 10,942 gallons per month. This is a high estimate, and it would suggest that all of the utilities in this sample, with the exception of Toho and Lakeland, have rate structures that force low income consumers to exceed the hardship threshold. Another study, Stallworth (2000) suggests that the lifeline amount consists of non-discretionary water use, which is about 8,000 gallons per person per month. In our hypothetical household, this would amount to 19,680 gallons per month. If this is an a ccurate estimate of nondiscretionary usage, than every rate structure in this study would force low -income households to exceed the hardship threshold. However, it should be noted that the 8,000 gallon per person estimate is considerably higher than other estimates. Mizgalewicz (1991), Gleick (1996), and California DWR (2009) suggest that less water can meet nondiscretionary residential needs. Gleick (1996, p. 84) estimates that urban users can get by on 100 liters a day or less, which translates to less than 1000 gallons per month. California DWR (2009) estimates that residential users need a minimum of 70 gallons per person per day, which translates to about 2100
97 gallons per person per month. This estimate explicitly excludes outdoor irrigation needs. The results of this part of the study depend greatly on assumptions made about the affordability of water. More research is needed about when water rates force consumers to exceed hardship thresholds. In the case of Florida, lifeline water rates th at make nondiscretionary use affordable may be justified. Fairness: Interpreting the Lorenz Curves and Gini Coefficients. Lorenz curves (Charts D2 D 5 ) and Gini coefficients (Table D 2) were examined for each of the sixteen water utilities and each of t he water rate structures they used. This analysis included the fixed and volumetric rate components. Water taxes levied by local authorities (municipal or county) were also included. However, wastewater charges were not taken into account. All of the rate structures are found to be fairly equitable. That is, in none of them do one or more of the income groups grossly underpay or overpay for water used. Table D -2 shows that all of the Gini coefficients are calculated to be less than 0.2. The highest es timated so far, Tallahassee, is 0.17. This result is somewhat unexpected. Due to the large differences between rate structures, it would seem reasonable to hypothesize that there be would fairly wide variations in the Gini coefficients. Instead, it appears to be the case that changes in rate structure translate to very small differences in overall fairness. The existing variation in the values of Gini coefficient for different rate structures can be linked to the difference in the fixed fees that most rate structures have. When fixed fees are relatively high, households that use less water actually bear a greater revenue burden. To illustrate this, I focus on the Spring Hill water utility, which had a
98 uniform rate structure during the duration of the study period. In all the months of 1998, Spring Hill water utility charged a fixed monthly fee of $4.22/month, and a volumetric fee of $0.97 per thousand gallons. Suppose that Household A used 2,000 gallons in a particular month. (In the Whitcomb data se t, some households did actually record a monthly usage of 2,000 gallons). And, suppose Household B used 10,000 gallons in the same month. (10,000 gallons is fairly common quantity for a household to consume). Household A would have a water bill of $6.16. Household B would have a bill of $13.92. In this case, Household B uses five times as much water as household A, but only pays a little more than twice the water bill. If Spring Hill had been using an inclining block structure, then Household B would almost certainly be paying relatively more for its higher consumption. Analysis of the fairness of different rate structures used by the same utility can tell an interesting story as well. Consider Ocoee results. In the early months in 1998, Ocoee had a uni form water rate structure. The Gini coefficient for this period is about 0.133. Later in 1998, Ocoee switched to a three block water rate structure. Surprisingly, this did not seem to affect the overall fairness of the water rate. In 2003, Ocoee adopted a six block structure. For this period, the Gini coefficient was 0.087, significantly less than the observed coefficients in the other two periods, indicating a more fair rate structure. Conclusion The affordability results suggest that unifo rm rates a re not better for low -income households than inclining block rates. In fact, inclining block rates seems to be slightly better for low income households than uniform rates. Also, this analysis indicates that
99 in some cases, low -income households may be s everely limited in the amount of water they can purchase without exceeding their hardship threshold. Therefore, in the case of Florida, lifeline water rates that make non -discretionary use affordable may be justified. This preliminary analysis provides e vidence that uniform rates are not fairer than inclining block rates. In contrast, uniform rate structures can be less fair. Of the five rate structure periods completed so far, Tallahassee (which used a uniform rate structure) had by far the highest Gini coefficient (0.17), indicating a less fair water rate structure. The lowest observed coefficient (0.047) was for Palm Beach, which had a three block structure. The second lowest coefficient (0.087) was observed for the period in 2003 when Ocoee had a six block structure. This evidence suggests that equity and fairness concerns do not constitute a good argument against inclining block structures. In fact, proponents of inclining blocks could use this analysis to argue that inclining block structures are b etter in terms of overall equity and fairness. Limitations of the Study The importance and meaning of the Equity results depend greatly on assumptions made about the affordability of water. More research is needed about when water rates force consumers to exceed hardship thresholds. Also, more research is needed about how much water actually constitutes non-discretionary use. The fairness results are also limited by the underlying assumptions. Many variables (usage characteristics, the relative size o f fixed and volumetric fees, the size and steepness of price blocks, etc) affect the Gini coefficient. Therefore, more research is needed to explore this issue further.
100 Also, it is important to point out that the Gini Coefficient results do not necessarily mean inclining block structures are better for low income users, or for users with low discretionary use. The Gini coefficients and Lorenz curve simply analysis the overall equity of the rate structure periods. Table 6 1. Customer Income Groups. Name Income range Income group 1 <$15,000 per year Income group 2 $15,000 $29,999 per year Income group 3 $30,000 $49,999 per year Income group 4 $50,000 79,000 per year Income group 5 $80,000 $100,000 per year Income group 6 >$100, 000 per year Table 6 2. Estimated Hardship Thresholds for Hypothetical Florida Household Year Monthly Household Income on the Poverty Line Hardship Threshold 1998 $12,138 $40.46 1999 $12,367 $41.22 2000 $12,584 $41.95 2001 $12,999 $43.33 2002 $13,3 57 $44.357 2003 $13,564 $45.21 Table 6 3. Melbourne Rate Structure: 1998. Water Sewer Taxes Per TG Fixed Charge per TG Fixed Charge per TG 2.07 3.28 3.34 4.63 None
101 CHAPTER 7 CONCLUSIONS AND IMPL ICATIONS Competing Objectives of Rate D esign: The Big Picture To a certain extent, each core chapter of this thesis (chapters 4 -6) is a distinct study. Each is largely independent, with different methods and objectives. Despite this, several general conclusions can be made. Inclining block structures can be used to send strong conservation price signals. But, the number of price blocks is not as important as several other factors, including the ratio of fixed to volumetric fees, the nature of the sewer fees, and location of the breakpoints. Neither a large number of price blocks nor a steepness of price blocks guarantee strong price signals. This is because many price blocks have break points that are far above the water use level of an average household. For example, in rate structures w ith three or more price blocks, it is not unusual to see breakpoints at 20 or more thousand gallons per month. Most households are unaffected by these higher rates. Regardless of block steepness, most households with never be affected by the top blocks, and have no additional incentive to conserve. Revenue variability is a potential cause for concern for utilities that use price incentives to promote conservation. There may be a tradeoff between conservation and revenue variability and predictability. M ore empirical research is needed to confirm and define this possible tradeoff. Equity and affordability should not be a major cause for concern. The fairness objective: an equitable rate structure that makes basic, non-discretionary water use affordable need not compete with the other objectives of conservation, cost recovery, and revenue stability.
102 More specific conclusions are discussed below. Characteristics of Water Rates used by Florida Utilities The analysis of current rate structure used by the sixteen sample utilities indicates that in Florida, there is a trend toward rate structures that are more oriented toward conservation. Specifically, the preferred type of rate structure in Florida now appears to be inclining block with at least three price blocks. It appears that water conservation is an objective that utility managers in Florida take seriously, and it appears that many of them have confidence in block pricing as a tool that be used to achieve this objective. But, at the same time, many utilities in Florida still rely heavily on fixed fees, probably in an effort to maintain stability and predictability in the revenue stream. Because of these fixed fees, inclining block structures do not always translate to stronger conservation pr ice signals. In the future, utility managers need to consider reducing the fixed fees if they really want to encourage conservation. Another broad conclusion is that sewer fees often represent a large portion of residential water bills and residential r evenue for utilities. These fees may obscure or complicate the price signals received by consumers. Any long term plan that involves conservation water rates needs to incorporate sewer rates. Most consumers see both water and sewer charges on their mont hly bill. In addition, and sewer fees are often linked to water use. Finally, the results suggest that changes in the type of rate structure are relatively uncommon for Florida utilities. While most utilities will change rates over time, these cha nges tend to be minor. Usually, the fixed or the volumetric charge will increase or decrease slightly. Utilities seem more reluctant to change the type of rate structure-
103 from uniform to inclining block or vice versa. In fact, during the entire duration of the study, only two utilities (Ocoee and Toho) actually changed the type of rate structure they used. This may be due to the fact that large changes in rate design can be hard for consumers to understand and/or accept. Effectiveness of Water Rates in Stimulating Water Conservation The analysis suggests that uniform or inclining block type of rate structure alone does not guarantee stronger conservation price signals. First, uniform rates can have average and marginal prices that are just as high as prices associated with inclining blocks rates. Second, the incentive scatterplot (Chart A 19) shows that the conservation incentives (measured as reduction in water bill associated with 40% reduction in water use) tend to be comparable for uniform rat es and for block rates with three or less price blocks. However, there is some evidence that price signals are significantly stronger for block rates with more than three price blocks. Third, the revenue distribution analysis shows that the level of utili ties fixed fees can be more important for stimulating water conservation than the number and steepness of price blocks. For example, Tampa has stronger price signals than St. Petersburg or Miami, even though St. Petersburg and Miami had more price blocks. The marginal and average price results suggest that some (but not all) utilities tend to increase rates to approximate inflation, so that real rates stay fairly constant. A few utilities in the sample increased prices at a rate that exceeded inflation. This may be due to debt service fees, water supply increases, infrastructure expansion needs, or other costs that result in higher revenue needs.
104 Revenue The results show that average monthly revenue was, in many cases, higher in periods with block pricing However, if minimizing short term revenue variability is a major priority, then inclining block rates should be implemented carefully. There is some evidence that block prices are correlated with increased revenue variability, both in terms of individual household bills and overall utility revenue. This increase in variability appears to be relatively small, however. There is no evidence that block pricing leads to a vastly more volatile revenue stream than uniform pricing. Equity and Affordability Co nsumer fairness, and affordability should also not be used to justify the use of water rates that lack conservation price signals. There is little or no evidence that inclining block rate structures lead to a more inequitable billing distribution or to water that is less affordable for low income households. In fact, there is some limited evidence that block structures may be slightly more equitable, at least when equity is defined by income groups. There also appears to be no clear correlation between ty pe of rate structure and water affordability for low income households. Some block structures make water relatively affordable to low income households, and other do not. But there is no consistent pattern in these results. The fixed fees and the price of water for low levels of usage are usually the only important factors for determining the affordability of a rate structure. Study Limitations Limitations of the research methods are discussed in each of the core chapters (4 -6). This se ction briefly supplements those discussions.
105 Perhaps the most significant limitation to this study is that it does not analyze the differences in the household water use induced by changes in water rate structures. Instead, the study focuses on characteri stics of the water rates that can influence household water use decisions (such as marginal and average prices and total household bill). To make strong policy recommendations about conservation water rates, the study results should be used in combination with water demand elasticity analysis such as Whitcomb (2005) and Unel (2010). Chapter 5 examines monthly utility revenue level and its variability, as well as the distribution of household water bills. Frequency distributions for inclining block rate st ructures show significant difference in household water bills, but it is not clear if it is caused by variations in water rates between price blocks or by changes in household water usage And, while very basic econometric models are used to examine fact ors influencing variability in household water bills, no advanced econometric techniques were utilized to analyze the data in more depth. Finally, there are two broad limitations to this study related to the characteristics of the empirical data used. F irst, data sample include only few utilities that changed the type of rate structure (i.e., shifted from inclining block to uniform rates or from uniform to inclining block). Changes in unit water rates implemented by most utilities during the 6 year period were not significant enough to cause noticeable changes in utility revenue, customer fairness or affordability, or water conservation incentives. While differences in rate structures among utilities were significant, the comparison of results is complic ated by the variation in the number of households sampled from each utility service area.
106 Future studies should take advantage of cases where significant changes in the type of rate structure have occurred. The other related limitation is that the indivi dual rate structure periods were uneven in terms of length, number of households, and time of year. Therefore, comparing them between utilities on the basis of any criteria is very difficult. Ideally, future studies will focus on rate structure periods t hat change with more consistency. Study Contribution Despite these significant limitations, this research does make a number of important contributions to the literature. Most importantly, it provides a wide variety of empirical evidence for Florida, which is unique in terms of its policy and demographic characteristics. This study contributes to the current policy discussions in Florida about what constitute a conservation oriented rate, and identifies Florida utilities that send the strongest water conservation price signal. Using Florida-specific data, the study also demonstrates that conservation water rates allow utilities to achieve revenue stability and customer equity/water affordability objectives. This study helps fill a gap in empiric al research on the topic of conservation water rates in the state. Another important contribution of this study is the analytical procedure developed to evaluate water rate structures. This procedure integrates various methods used in the literature f or quick andeasy evaluation of water rates given limited information available about customer and utility characteristics. Specifically, effectiveness / conservation price signals were evaluated based on a) average and marginal prices faces by an average customer, b) reduction in average customer water bill resulting from 40% decrease in water usage, and c) relative shares of utility revenues from fixed and volumetric charges (estimated based on a sample of households). Further, simple
107 averages and standard deviations were used to examine utility revenue level and variability. Gini coefficients and Lorenz curves were used to examine customer equity. Finally, water affordability was examined based on the percent of income spent on water by a customer on the poverty line. Although all these methods have been used in other studies, no study was found that would apply all of them to conduct quick but comprehensive evaluation of water rate structures. Future studies could easily take advantage of these methods to examine the rate design implications in other geographic areas. Recommendations for Future Research Probably the most important recommendation involves conservation effectiveness. Even though extensive empirical work has been done about price and income elasticity of residential water demand, more is still needed. It is not entirely clear how effective conservation rates are, even when price signals are strong. Time is another relevant factor that should be studied in the context of wat er rate design. Previous research suggests a possible time lag effect in residential response to conservation rates (which can be especially important for seasonal and drought water rates). This study does not examine this possibility, but future studies should. Utilities and regulatory agencies sometimes rely on nonprice mechanisms (education programs, rebates, irrigation restrictions, etc.) But, there is relatively little empirical evidence about the effectiveness of these programs or about their inter action with conservation rates. Future research should attempt to address this gap in knowledge. Identifying if and when price and nonprice based conservation polices are effective should be a top priority of research in the near future.
108 Also, there i s no consensus methodology that utilities use to design rates. More research is needed to determine why utilities select specific conservation mechanisms. The rate design objectives discussed in this study: conservation, revenue, revenue stability, equit y, affordability, are undoubtedly important to most utilities. However, there are probably other objectives that are worthy of consideration (such as meeting regulatory requirements). More information is needed about what these objectives are, and how ut ilities balance and prioritize the objectives they have. Two related issues are management and ownership. The management and governance structure of a utility is likely to have a major impact on how rate decisions are made. More research is needed to explore this issue. The same can be said for ownership. Most utilities in this sample are publicly owned. But, within Florida and throughout the United States, some utilities are privately owned. These two types of ownership each have advantages and disadvantages, and more research could help define them. More research could be used to explore how ownership affects conservation rate design decisions. Finally, the results of this study suggest a both challenge and an opportunity. Whitcomb (2005) pr oduced a very rich data set, which has been used in the original study, in materials designed for public consumption, in the Waterate simulation model, and in this study. Future studies of similar depth and magnitude should be undertaken. The combination of use data, billing data, and tax assessor data makes examination possible of a wide variety of issues. But it is strongly recommended that future studies of this kind take advantage of more utilities that show variation in the type of rate structure us ed. Rather than compare rate structure periods with very different
109 characteristics, an ideal future study would focus on a small number of utilities that changed the type of rate structure used throughout the study. This could serve as a kind of natural experiment that would make results and conclusions more meaningful.
110 REFERENCES Agarwal, Anil, Marian S. delos Angeles, Ramesh Bhatia, Ivan Chret, Sonia Davila Poblete, Malin Falkenmark, Fernando Gonzalez Villarreal Torkil Jnch -Clausen, Mohammed At Kadi, Janusz Kindler, Judith Rees, Paul Roberts, Peter Rogers,Miguel Solanes, and Albert Wright. (2000). Integrated Water Resource Management. Global Water Partnership Technical Advisory Committee. Copyright Global Wate r Partnership 2000. Accessed at: www.cepis.ops oms.org/bvsarg/i/fulltext/tac4/tac4.pdf on June 1, 2009. Agthe, Donald and Billings, Bruce. (1987). Equity, Price Elasticity, and Hous ehold Income Under Increasing Block Rates for Water American Journal of Economics and Sociology vol. 36, number 3, 273-283. Alliance for Water Efficiency (AWE). (2008). Conservation Oriented Rate Structures. Alliance for Water Efficiency (AWE) Resourc e Library accessed at: http://www.allianceforwaterefficiency.org/1Column.aspx?id=712 on March 1, 2009. American Water Works Association (AWWA). (2000). Principles of Water Rates, Fees, and Charges American Water Works Association Manual of Water Supply Rates, M1. (Fifth Edition). published by American Water Works Association. American Water Works Association (AWWA) and Raftelis Financial Consultants. (2006). Water and Wastewater Rate Survey published by American Water Works Association. Arnold, Lee (chairman). (2003). Improving Floridas Water Supply Management Structure Report for Florida Council of 100 prepared by the Council Water Management Task Force. Accessed at: www.floridacouncilof100.com/documents/waterreportfinal.pdf on February 10, 2009. Bakker, Karen. (2001). Paying for Water: Water Pricing in England and Wales Transactions of the Inst itue of British Geographers, Vol. 26, No. 2. Baumann, Duane D., and Boland, John. (1997). Urban Water Demand Management and Planning New York, NY: McGraw -Hill. Beecher, Janice A., Landers, James R., and Mann, Patrick C. (1991). Integrated Resources Planni ng for Water Utilities Silver Spring, MD: The National Regulatory Research Institute. Beecher, Janice A., and Laubach, Anne. (1989). Compendium on Water Supply, Drought, and Conservation Silver Spring, MD: The National Regulatory Research Institute. Beec her J.A., P.C. Mann, Y. Hegazy, and J.D. Stanford. (1994). Revenue Effects of Water Conservation and Conservation Pricing: Issues and Practices National Regulatory Research Institute Report 94-18 Silver Spring, MD: The National Regulatory Research Instit ute.
111 Bhatt, Nisith R. and Cole, Charles A. (1985). Impact of Conservation on Rates and Operating Costs Journal of Water Resource Planning and Management vol. 111, number 2, 192206. Billings, R. Bruce, and C. Vaughn Jones. (1996). Forecasting Urban W ater Demand. Denver, CO: American Water Works Association (AWWA). Bjornlund, J. (2002). The Socio -economic Structure of Irrigation Communities: Water Markets and the Structural Adjustment Process The Journal of Rural Society, vol. 12, number 2, 123145. Booker, J. F., Michelsen, A. M. & Ward, F. A. (2005). Economic impact of alternative policy responses to prolonged and severe drought in the Rio Grande Basin Water Resources Research, 41(2), W02026. Borisova, Tatiana and Rawls, Colin. (2009). Conservation Water Rates for Residential Customers: A Practical Overview Article submitted to the Florida Water Resources Journal. Brown, R. R., L. Sharp, and R.M. Ashley. (2005). Implementation impediments to institutionalising the practice of sustainable urban water management Paper Presented at the 10th International Conference on Urban Drainage, Copenhagen/Denmark, 21-26 August 2005. Accessed at: http://arts.monash.edu/ges/r esearch/nuwgp/pdf/conference brownet al -05.pdf on September 3, 2009 California Department of Water Resources. (2009). Urban Water Needs Analysis Integrated Water Management. Accessed at: www.water.ca.gov/urbanwatermanagement/wateranalysis/index.cfm on December 7th, 2008. Carriker, Roy and Boriosova, Tatiana. (2009). Public Policy and Water in Florida. Electronic Data Information Source. Institute of Food and Agricultural Sciences. Carter, David (1997). Water Supplies and Environmental Externalities: Prospects for Efficient Water Pricing in Florida. Masters thesis. Gainesville, FL: University of Florida, Department of Food and Resource Economics. Carter, D.W. and J.W. Milon. (2005). Price Knowledge in Household Demand for Utility Services Land Economics, vol. 81, number 2, 256283. Cavanaugh, S.M., W.M. Hanemann, and R.N. Stavings. (2002). Muffled Price Signals: Household Water Demand Under Increasing-Bl ock Prices FEEM Working Paper No. 40. Milan, Italy: Fondazione Eni Enrico Mattei. Clunie, J. (2004). Can a Water Rate Structure Really Promote Conservation? Presentation at U.S. Conference of Mayors, Urban Water Council. October 2122, 2004. Accessed at : www.usmayors.org/urbanwater/foxwoods04/Clunie.ppt on January 14, 2009.
112 Chestnutt, Thomas W., Anil Bamezai, and Michael Hanemann. (1993). Revenue Instability Induced by Conservation Rate Structures Working Paper number 695. Berkley, CA: University of California at Berkley. Chestnutt, Thomas W., and Beecher, Janice. (1998). Conservation Rates in the Real World. Journal of the American Water Works Association, vol. 90, issue 2, 60 -70. Cummings, Ronald G. and Nercissiantz, Vahram. (1992). Use of Water Pricing as a Means for Enhancing Water Use Efficiency in Irrigation: Case Studies in Mexico and the United States. World Bank Technical Paper Series Washington, DC: The World Bank. Cuthbe rt R.W. and P.R. Lemoine. (1996). Conservation-Oriented Water Rates. Journal of the American Water Works Association, vol. 88, number 11, 6878. Dalhuisen, J.M., R.J.G.M. Florat, H.L.F. de Groot, and P. Nijcamp. (2003). Price and Income Elasticities of Res idential Water Demand: A Meta Analysis. Land Economics vol. 79, 292308. Daniel, Patty and Lingo, Jennie. (1997). Conservation-Oriented Rate Structures for Water Utilities Tallahassee, Fl: Florida Public Service Commission. Deslatte, Aaron. (1998). Crist calls new Federal Water Ruling Disappointing The Orlando Sentinel Orlando, FL: The Orlando Sentinel. Accessed at: http://blogs.orlandosentinel.com/news_politics/water_wars/ page/2/ on August 2, 2008. Daigger, Glen T. Evolving Urban Water and Residuals Management Paradigms: Water Reclamation and Reuse, Decentralization, and Resource Recovery. Water Environment Research, Volume 81, Number 8, August 2009 pp. 809823(15) Dina r, Ariel. (2008). The Political Economy of Water Pricing Reforms Oxford, United Kingdom: Oxford University Press. Environmental Finance Center (EFC), (2009). Rate Dashboards University of North Carolina, Environmental Finance Center. Accessed at: www.efc.unc.edu/RatesDashboards/ on July 12, 2009. Environmental Finance Center. (2007). Water Price Signals in Georgia University of North Carolina, Environmental Finance Center. Accessed at: http://www.efc.unc.edu/publications/pdfs/Water_Price_Signals_in_Georgia.pdf on April 4, 2009. Environmental Protection Agency. (2002). Rate Options to Address Affordability Co ncerns for Consideration by District of Columbia Water and Sewer Authority Copyright EPA. Washington, DC. Accessed at: www.epa.gov/waterinfrastructure/pricing/pdfs/AffordO ptions.pdf on March 1, 2009 Espey, M., J. Espey, and W. D. Shaw. (1997). Price Elasticity of Residential Demand for Water: A Meta Analysis. Water Resources Research, vol. 33, number 6.
113 Florida Department of Environmental Protection. 2002. Florida Water C onservation Initiative Final Report. Accessed at: www.dep.state.fl.us/water/waterpolicy/docs/WCI_2002_Final_Report.pdf on November 13, 2008. Florida Department of E nvironmental Protection. (2004). Joint Statement of Commitment for the Development and Implementation of a Statewide Comprehensive Water Conservation Program for Public Water Supply. FEDP 2004. Accessed at: www.dep.state.fl.us/water/waterpolicy/docs/JSOC --new_small.pdf on November 13, 2008. Florida Department of Environmental Protection (FDEP). (2008). Water Conservation Tallahassee, FL: Florida Department of Environmental Protection. Accessed at: www.dep.state.fl.us/swapp/Conservation.asp on November 13, 2008. Florida Department of Environmental Protection (FDEP). (2006). Water Resource Implementation Rule Tallahassee, FL: Florida Department of Environmental Protection Accessed at: www.dep.state.fl.us/legal/Rules/shared/62-40/6240.pdf on November 13, 2008. Florida Section of Amer ican Water Works Association (FSAWWA). (2008). Water Conservation: Draft Issue Papers. Florida 2030: A vision for Sustainable Water Infrastructure Published by the American Water Works Association. Accessed at: ht tp://www.florida2030.com/ on July 19th, 2009. Gainesville Regional Utilities (GRU). (2008). Annual Report: 20072008. Accessed at www.gru.com/Pdf/AnnualReport2008/2007-2 008_GRU_Annual_Report.pdf on September 10th, 2008. Gaudin S. 2006. Effect of Price Information on Residential Water Demand. Applied Economics 2006, vol. 38, issue 4, 383 -393 Gleick, Peter H. (1996). Basic Water Requirements for Human Activities: Meeting Basic Needs. Pacific Institute for Studies in Development, Environment, and Security Accessed at: https://waterportal.sandia.gov/literature/waterecondev/basicwater.pdf on May 19th, 2009. Gonzales -Gomez, F. and Garcia Rubio, Miguel A. Efficiency in the management of urban water services What have we learned after four decades of research? Hacienda Pblica Espaola / Revista de Economa Pblica, 185-(2/2008): 39-67 2008 Instituto de Estudios Fiscales. Griffin, Ronald C. (2001). Effective Water Pricing Journal of the American Water Resource Association vol. 37, number 45.
114 Hanemann, William Michael, Thomas Chestnutt and Anil Bamezai. (1993). Revenue instability induced by conservation rate structures: an empirical investigation of coping strategies CUDARE Working Paper 695 Berkeley, CA: University of California at Berkeley Hotelling, Harold. (1931). The Economics of Exhaustible Resources Journal of Political Econom y, Vol. 39:137175. Howe, Charles W. (2002). Policy Issues and Institutional Impediments in the Management of Groundwater: Lessons from Case Studies. Environment and Development Economics vol. 7, issue 4. Howe, Charles W. and Goemans, Chris. (2002) Eff ectiveness of Water Rate Increases following Watering Restrictions. Journal of the American Water Works Association vol. 94, issue 10. Huffaker, R., Whittlesen, N., Michelsen, A., Taylor, R. and McGuckin, T. (1998) Evaluating the effectiveness of conse rvation water pricing programs. Journal of Agricultural and Resource Economics 23, 12 19. Johns, Grace M. project manager. Economic Incentive Approaches to Water Resource Management Hazen and Sawyer: Environmental Engineers and Scientists. Copyright 1997. Jones, D.M. (1974). Conservation and Utility Earnings: A policy predicament Public Utility Fortnightly vol. 93, number 9, 25 30. Kenney, D. (2003) Water allocation and management in the western United States: An Overview Paper presented to the Inter national Working Conference on Water Rights: Institutional options for improving water allocation. Hanoi, Vietnam, 12 15 February 2003. Kenny, Douglas and Christopher Goemans, Roberta Klien, Jessica Lowrey, and Kevin Reidy. (2007). Residential Water Dem and Management: Lessons from Aurora, Colorado. Journal of the American Water Works Association vol. 44, issue 1. Mankiw, Gregory N. (2004). Principles of Economics South-Western Publishing. Fourth Edition. McLarty, Rebecca and Heaney, James. (2008). Wa ter Savings from Conservation Rate Systems. Working Paper, Deptartment of Environmental Engineering Sciences, University of Florida Michelsen A.M., McGuckira J.T., Stumpf D.M. (1998). Effectiveness of Residential Water Conservation Price and Nonprice Programs America Water Works Association Research Foundation. Published by the American Water Works Association.
115 Miniaci, Raffaele, Carlo Scarpa, and Paola Valbonesi. (2008). Distributional Effects of Price Reforms in the Italian Utility Markets Fondazione Eni Enrico Matte. Copyright Milan, Italy. Mizgalewicz, Pawel Jerzy. (1991). An analysis of Monthly Water Demand in Phoniex Arizonia. Masters Thesis. The University of Teaxs at Austin. Accessed at: www.ce.utexas.edu/prof/maidment/GISHYDRO/pawel/phoenix/x2v6a.pdf on February 4th, 2009. Morgan, Douglas W. (1987). How Fair Are Water Tarrifs? Some Evidence from California Evaluation Review, vol. 11, p. 33. Sage Publications. Minnesota Department of Natural Resources (Minnesota DNR). 2008. Conservation Rates published by the Minnesota Department of Natural Resources. National Water Commission (NWC). (2009). Water Pricing and Water Reform Government of Australia, Nati onal Water Commission. Accessed at: http://www.nwc.gov.au/www/html/601 water -pricing andwater -reform.asp on January 14th, 2009. Nauges C. and A. Thomas. 2000. Privately operated Water Utilities, Municipal Price Negotiation, and Estimation of Residential Water Demand: The Case of France. Land Economics vol. 76 Nida C. S. Eskaf. 2009. Water and Wastewater Rates and Rate Structures in North Carolina NC League of Municipalit ies and Environmental Finance Center, School of Government, UNC. Accessed at: http://www.efc.unc.edu/projects/NCWaterRates.htm on November 21st, 2009. Nieswiadomy, Michael and Molina, David. (1 989). Comparing Residential Water Demand Estimates Under Decreasing and Increasing Block Rates Using Household Data. Land Economics vol. 65, number 3. ODea, Gerald and Cooper, Jennie. (2008). Water Scarcity: Does it exist and can price help solve the pro blem? Independent Pricing and Regulatory Tribunal (IPART). Accessed at: www.myoung.net.au/water/policies/Water_Scarcity.pdf on January 23rd, 2009. Policy Research Initiative (PRI). (2 005). Economic Instruments for Water Demand Management in an Integrated Water Resources Management Framework. Synthesis Report PRI Project, Sustainable Development. PRI. Accessed at: http://www.policyresearch.gc.ca/doclib/SR_SD_EconomicInstruments_200502_e.pdf on March 22, 2009. Purdam, Elizabeth D. (2002). Florida Waters: A Water Resources Manual from Floridas Water Management Districts Published by the Florid a Water Management Districts.
116 Raftelis G.A. (2005). Water and Wastewater Finance and Pricing: A Comprehensive Guide Third Edition. Boca Raton, FL: Taylor & Francis Group. Reepen, Deena and Merriam, Chip. (2009). A Comprehensive Water Conservation Progra m for South Florida South Florida Water Management District. Presentation at the SFWMD Governing Board Workshop April 9th, 2009. Rogers, Peter, Radhika Silva, and Ramesh Batia. 2002. Water is an Economic Good: How to use Prices to Promote Equity, Efficiency, and Sustainability Water Policy 2002 vol. 4, 1-17. Rubin, Scott, J. (1998). A Nationwide Look at the Affordability of Water Service. Public Utility Consulting. Accessed at: http://publicutilityhome.com/speeches/AWWA1998.pdf Saarinen, Phyllis Park (1993). A political economic analysis of the Florida Water Allocation Process under scarcity conditions Masters thesis. Gainesville, FL: University of Florida, Department of Food and Resource Economics. Saunders, Margot Freeman, Phyllis Kimmel, Maggie Spade, and Nancy Brockway. (1998). Water Affordability Programs The American Water Works Research Foundation. Accessed at: http://books.google.com/books?hl=en&lr=&id=8eDRms8y9FYC&oi=fnd&pg=PR17&dq= Saunders+M+1998+water+affordability&ots=ORP1tY4p7j&sig=CO9eSJJd2Q5XXET5rl1 YulJuMFM#PPR5,M1 on January 27th, 2009. Schleich J. and T. Hillenbrand. 2009. Determinants of residential water demand in Germany. Ecological Economics vol. 68, issue 6, 1756 -1769 Smith, William James and Wang, Young-Doo. (2008). Conservation Rates: The best new Source of Urban Water during Drought Wate r and Environment Journal Vo.22, issue 2. South Florida Business Journal (SFBJ). (2008). Sugar crop hurt by water shortage American City Business Journals. Accessed at: http://southflorida.bizjournals.com/southflorida/stories/2008/03/03/daily37.html on February 1st, 2009. South Florida Water Management District. (2008). Working Draft Water Conservation Program Plan. South Florida Water Management District. Acc essed at: www.sfwmd.gov/pls/portal/docs/PAGE/PG_GRP_SFWMD_NEWS/PORTLET_CONSE RVESUMMIT/TAB19809840/FIN AL_WATERCONSERVATIONPROGRAM_DRAFT032 108.PDF on December 10th, 2008. Stallworth, Holly. (2000). Conservation Pricing of Water and Wastewater United States Environmental Protection Agency. Accessed at: www.p2pays.org/ref/21/20599.pdf on February 3rd, 2009.
117 Tsur, Yacov and Ariel Dinar. (1997). The Relative Efficiency and Implementation Costs of Alternative Methods Pricing Irrigation Water World Bank Economic Review Washington, DC: The World Bank. Turner, V.K. (1999). Public Good or Commodity? Institutional Differences in Water Management Strategies and Conservation Outcomes in Tucson and Phoenix Global Institute of Sustainability. Accessed at: http://dcdc.asu.edu/dcdcmain/pdf/turner.pdf on July 14th, 2009. Unel, Burcin, Chunrong Ali, and Sanford Berg. (2010). An Analysis of Residential Water Demand: Does Consumer Attitude Matter ? Working Paper. University of Florida. United States Census Bureua. (2000). Small Area Income and Poverty Estimates Accessed at: www.census.gov/cgi bin/saipe/saipe.cgi on August 10th, 2009. United States Department of Health and Human Services. (2009). Pri or HHS Poverty Guidelines Published by the Federal Register, 2009. Accessed at: http://aspe.hhs.gov/POVERTY/figures -fed-reg.shtml on July 25th, 2009. US Water News. (2002). Florida drought over, but water shortage still critical US Water News 2002. Accessed at: www.uswaternews.com/archives/arcconserv/2flodro1.html on December 1st, 2008. Wang Y. -D., Smith W.J. and J. Byrne. (2005). Water Conservation-Oriented Rates: Strategies to Extend Supply, Promote equity, and Meet Minimum Flow Levels American Water Works Association, Denver, Co. Water and Wastes Digest. (2007). Increasing Freshwater Demand Increases Growth in Wa ter Recycling and Reuse Systems Market. Water and Wastes Digest Water Infastructure Network. (2000). Clean and Safe Water for the 21st Century Copyright Water Infastructure Network (WIN). Accessed at: www.wi n -water.org/reports/ on Septermber 1st, 2009. Western Resource Advocates, Colorado Environmental Coalition, and Western Colorado Congress. (2004). Water Rate Structures in Colorado: How Colorado Cities Compare in Using this Important Water Use Efficiency Tool Water Resource Advocates. Accessed at: http://www.westernresourceadvocates.org/media/pdf/Colorado%20Water%20Rate%20 Structures.pdf on December 15th, 2008. Whitcomb, John B. (2005). Florida Water Rates Evaluation of Single Family Homes Funded and prepared by the Florida Water Management Districts. Accessed at: www.swfwmd.state.fl.us/documents/reports/water_rate_report.pdf on June 4th, 2008. White, S.B. and Fane, S.A. (2007). Designing Cost Effective Demand Management Programs in Australia Institute for Sustainable Futures
118 BIOGRAPHICAL SKETCH Colin Rawls was born in Raleigh, North Carolina. He grew up in Vero Beach Florida, in Indian Rive r County. Living near the coast of Florida and the Indian River Lagoon, Colin developed an interested Floridas water resources. He graduated from Vero Beach High School in 2003. He then attended the University of Florida. In 2007, he graduated with a bachelors degree in Economics and a minor in history. As a graduate student, he has worked as intern with the University of Florida Water Institute and an assistant coach with the University Speech and Debate Team.