From Rates

Making it Work

A Kansas water utility gets affordability measurement right

And lo, there arose from the Kansas City suburbs a mighty measurement

Recently we’ve seen progress in affordability measurement, as more water utilities are using better metrics when evaluating affordability.* Last year I published a new methodology for measuring water and sewer utility affordability (AR20and HM), and followed that up with a national assessment using those metrics. AR20  is the Affordability Ratio of basic water and sewer service price divided by disposable income at the 20th percentile household income. HM  is basic water and sewer service expressed in Hours of labor at Minimum wage.These metrics seek to capture the trade-offs that low-income households must make in paying for water and sewer services. Utilities have begun to use these and other improved metrics, which is encouraging!

Too hard?

The main objection I’ve seen to the real-world use of AR20 is that it can seem too complicated. You need to know the community’s 20th percentile income and essential non-water/sewer costs of living in order to calculate AR20. But there’s no convenient, off-the-shelf source for those numbers. You have to think about economic conditions in your community.

I usually include housing, health care, taxes, food,and home energy as essential non-water/sewer costs, and I use statistical models to estimate those expenses. Statistical models are important in my research because I’m analyzing affordability across hundreds of communities.  Apparently that’s led some to think that regression models of consumer data are the only way to estimate AR20,  which can seem impossibly difficult.

Fortunately, it’s not really that hard. Since publishing my 2018 article, I’ve heard from folks in utilities large and small about efforts to use these metrics exactly as they were intended: adapting AR20  to fit local needs, calculating it with local data, and using it to shape local decisions.

WaterOne’s excellent measurement adventure

An especially encouraging case is WaterOne, a special district that provides drinking water to a population of about 425,000 in the Johnson County suburbs southwest of Kansas City. As in plenty of other utilities, WaterOne’s leaders have long been interested in the affordability of their service, but had also long used the conventional 2%MHI to gauge affordability. Dissatisfied with that nonsensical number, WaterOne’s financial planning team decided to use AR20  to assess affordability and help guide policy for their own system.

Local calibration

From WaterOne’s Affordability Ratio paper

Making AR20 work for WaterOne required adapting it to local preferences and conditions in a few ways. First, the original AR20 was calculated for water and sewer combined; since WaterOne provides only drinking water, its AR20 calculation included only water rates, not sewer rates. Second, WaterOne analyzed its own customer data and decided that 45 gallons per capita per day and a 2.6-person household were the appropriate basic water consumption level for its customers (my published studies assume a 4-person household at 50 gpcd). Third, WaterOne chose to exclude home energy from their calculation of essential non-water/sewer costs. Rather than constructing an econometric model to estimate essential non-water costs, WaterOne’s finance team used available data and guidelines from the Census, IRS, USDA, and Bureau of Labor Statistics to estimate appropriate costs for its service area.

The result was a WaterOne-specific AR20 that showed the remarkable difference between the conventional %MHI method and the more meaningful AR20. After they’d done all that work, WaterOne staff contacted me to ask for feedback. We had a terrific phone call with WaterOne managers where I offered some comments on their execution, but I didn’t have much of a critique to give—they pretty much got it right.

From analysis to decision

The results were reported with the district’s 2019 budget and written up in a white paper for WaterOne’s governing board. Crucially, the paper uses the affordability metrics to frame a discussion of goals and guidelines, not to declare WaterOne’s rates “affordable” or “unaffordable” according to some arbitrary threshold. They also warned against comparing AR20 values to my published works and to other systems’ AR20 values, since WaterOne’s AR20 is based on different assumptions and WaterOne’s values may not align with others.’

Measurement principles in practice

Complexity isn’t an excuse for crummy measurement; it’s a reason to be careful with measurement. A modicum of creativity can get you there. Want to know what low-income families pay for health insurance locally? Go to Need an estimate for local low-income housing costs? Check Ask local charitable organizations what low-income families pay for food or home energy. You don’t need a PhD or advanced econometric skill to do sound affordability analysis.

WaterOne answered the affordability measurement challenge with a thoughtful, nuanced analysis that applied community values to the best available data. Adapting AR20 for  WaterOne—WOAR20?—is a fine example of how utilities can put measurement principles into practice.

*The conventional approach to water affordability measurement (average bill as a percentage of Median Household Income (%MHI) is deeply flawed, as I’ve blogged previously. Despite its well-document problems,use of %MHI remains widespread, mainly because it’s easy and familiar.

My estimates use publicly-available Consumer Expenditure Survey data and Ordinary Least Squares regression. They’re not especially sophisticated.

Right on the money

A California surprise: update

Post-drought porn

California has been enjoying a great deal of rain and snow over the past several months—a pleasant rebound in precipitation after the brutal drought that plagued the state from 2011-2017. It’s now early 2019, reservoirs are full, the mountain snowpack is deep, and water managers in the Golden State are breathing easier than they have in a long time. Though water use has crept up since the end of the drought, overall water consumption remains lower than its pre-drought levels.

A California surprise

A surprising finding emerged from my analysis of California’s drought data with Youlang Zhang and David Switzer: the state’s private, investor-owned utilities conserved significantly more water than did local government utilities during the crisis. We linked the difference in drought response to the institutions that govern water finance. Nerds interested readers can read the full study in Policy Studies Journal for the details.

In a blog post last summer, I observed that a public-private conservation gap of 2-3% persisted in 2017 even after the drought ended, and wrote that financial imperatives would likely cause the trend to continue:

This consistent public-private difference lends greater weight to the idea that rate decoupling facilitates water conservation for private utilities, and that political constraints hamper public sector conservation. If 2018 holds to form, public and private conservation will converge in the spring and diverge again in the summer and autumn.

If financial and political considerations are really behind the public-private differences in conservation, then it stands to reason that the greatest differences would come during summer months, when water demand–and therefore rate revenue–fluctuations are greatest.

Not actually the author. I might have doctored the image a bit, too.

Now that full 2018 data are posted, it’s time to revisit conservation performance for the Golden State’s water utilities. Was the forecast valid? Did private systems conserve more than public systems again last year?

Are my water conservation predictions any better than my NCAA Tournament picks?*


Overall urban water use remained significantly lower in 2018, with average monthly conservation of about 14% compared with 2013. The public-private disparity in overall conservation also persisted. This graph plots average conservation (relative to the same month in 2013) for public and private utilities from January-December 2018:

Data: California State Water Resources Control Board

As you can see, public and private conservation moved in pretty close parallel through 2018, but private utility conservation was consistently higher than public. The difference was negligible during winter months, but during the May-September peak demand season, California’s investor-owned utilities saved an average of 2.3% more than their local government counterparts.

As always when discussing water, it’s important to give percentages some context. Had public utilities saved at the same rate as private utilities in 2018, the difference would have been about 27 billion gallons—more water than San Francisco uses in a year.

Decoupling, man. Decoupling.

The persistent seasonal swing in public-private water conservation suggests that the difference is due to differences in outdoor irrigation behavior. That the pattern is now consistent over three years adds to the mounting evidence that rate decoupling encourages conservation for investor-owned water systems.

This isn’t a story of environmental angels or devils, it’s about governance institutions and the incentives that they create. In light of the political challenges of managing local government water finance, it’s impressive that public utilities have continued to conserve as much as they have—a testament to local water managers’ commitment to efficiency in the face of political headwinds.


*They could hardly be worse. I didn’t get a single Final Four team right.

Taxing questions

Devils (and angels) in the details, Part 5

The ironic regressivity of a luxury tax

In early January the California Water Board (SWRCB) published its long-anticipated draft proposal for a statewide low-income water bill assistance program. I’ve blogged about it over the past few weeks*; in this final (I think) post on the proposal, I’ll look at how the SWRCB proposes to pay for the estimated $606 million annual program.

How to raise 600 million dollars

The SWRCB recommends paying for the new water assistance program through “progressive revenue sources… in order not to burden some of the residents that the program seeks to serve.” To that laudable end, the proposal calls for two new taxes:

  • A .25% tax on personal incomes over $1 million; and
  • A sales tax on bottled water.

The income tax would generate an estimated $466 million annually, while the bottled water tax would generate $154 million. Under California’s Byzantine pubic finance laws both taxes would require supermajorities in the state legislature. The bottled water tax would also have to pass a ballot referendum.

Progressive taxation is crucial to any low-income assistance program, since the whole point is to transfer resources to people who are short on them. A millionaire’s tax makes sense from that perspective; a person whose income tops $1 million annually likely has little difficulty paying his/her water bill, and it’s doubtful that an additional 0.25% tax will much constrain productivity or lifestyle for people with seven-figure incomes.

Robbing Peter

The bottled water tax is thornier.

It’s common for those of us who work on American drinking water issues to think of bottled water as a luxury good. Bottled water is orders of magnitude more expensive than tap water, after all. It also carries some severe negative externalities: it’s lightly regulated, uses lots of energy to produce and transport, and empty bottles create a huge solid waste problem. For all those reasons, taxing bottled water would be progressive, in theory.

Luxury good? 

In practice, the progressivity of a bottled water tax isn’t so clear. Counterintuitively, in America bottled water consumption is negatively correlated with income—that is, poor and working-class Americans drink much more bottled water than do middle-class and wealthier Americans. Study after study after study show that low-income people and members of racial/ethnic minorities are much more likely to drink bottled water.

Why do lower-income people pay orders of magnitude more for bottled water, when affordable tap water is available? From a health and efficiency perspective, that isn’t rational behavior. It might be cultural, it might be taste/odor preferences, it might be about distrust in government, or it might be something else entirely. In some cases, people served by the smallest, poorest communities that suffer from poor water quality might need bottled drinking water. Raising the cost of bottled water might have the perverse effect of pushing low-income households to drink more soda and sugary beverages.

Whatever the reason for the income-bottled water relationship, the distributional effect of the SWRCB’s proposed tax is clear: the poor will bear a disproportionate burden of any bottled water tax. When you consider that a significant proportion of eligible households will never actually participate in the assistance program, a bottled water tax becomes doubly regressive.

So where do we get the other $150 million?

There’s an intuitive political appeal to using a water-related tax to raise money for water bill assistance. What kinds of water taxes could be progressive? Here are a couple of half-baked ideas (I don’t have the time or data to bake ’em).

You see “Fiji Girls.” I see tax collectors in cocktail dresses.

  1. A tax on “luxury” bottled water. Poor folks aren’t buying $5.00 bottles of FijiWater; they’re buying $9.99 cases of Ozarka at Walmart. A sales tax on water that retails for more than $1.00 per liter would spare the lowest-income households, and probably wouldn’t push them into drinking sugary beverages.
  2. A tax on residential tap water consumption over 12,000 gallons per month. At 50 gallons per capita per day (reasonably efficient indoor use), a family of four uses about 6,000 gallons per month. In the vast majority of situations, residential water consumption beyond 12,000 gallons a month is for discretionary outdoor use. A main drawback to this kind of tax is that it would irritate water utility managers, who don’t want to act as the state’s tax collectors.

A combination of these two taxes could generate significant revenue without putting the revenue burden of low-income assistance onto the people that it’s intended to help.


*In the past few posts, I’ve summarized the proposal, discussed its potentially perverse incentives for ratemaking, pondered its implications for struggling small systems, and options for administering assistance.


The SWRCB’s carefully avoids alluding to political appeal, but does note that “fees on bottled water or alcohol would have a nexus to water use.”