A California surprise, Part I
Something unexpected happened when California ordered its utilities to save water: the state’s investor-owned private utilities out-conserved local governments.
California’s long-term drought began as early as 2007, but intensified to crisis conditions by 2012. Conditions worsened, and in response 2015 Governor Jerry Brown and the California State Water Resources Control Board imposed restrictions on 408 drinking water utilities designed to reduce urban water usage by 25% statewide. The order required utilities to cut water use, but left individual utilities to choose the means by which to achieve conservation. The mandate assigned each utility its own conservation target, with standards ranging from 4-36% reductions relative to 2013 levels. These standards were formulaic, and varied based on utilities’ historical water consumption.
These conservation rules were in place for twelve months—June 2015 through May 2016—and applied to both local government utilities and private, investor-owned utilities. Conservation rules were assigned based on historical demand patterns and supply considerations only, not on ownership or governance.
Happily, the State of California has shared utility-level conservation data lavishly—a boon to water policy researchers! Over the past year, I’ve been sifting through that mountain of data with Youlang Zhang and David Switzer to see how California’s conservation efforts have fared. We’re discovering some fascinating things. The first of our studies is now forthcoming in Policy Studies Journal.
Restricting the flow
Faced with water scarcity, communities frequently restrict residential outdoor water use, such as car washing and especially lawn/garden irrigation. These water restrictions are effective in driving immediate reductions in water consumption. In California those restrictions typically take the form of limiting the number of days when outdoor irrigation is allowed each week. The graph below shows how public and private utilities regulated outdoor irrigation during the drought.
Eyeballing that graph, there doesn’t appear to be much difference between public and private utilities. But after adjusting statistically for a host of factors like utility size, demographic composition, and hydrological conditions, it turns out that private, profit-seeking, investor-owned utilities restricted irrigation about 4% more than public, local government utilities. That may not seem like much, as we’ll see it’s actually huge.
Meeting the mandate
We were also interested in what made utilities more or less likely to comply with the state’s conservation rules. Overall compliance was about 53%–that is, on average 53% of utilities reached their conservation targets each month. We modeled compliance statistically, and found a number of interesting correlates of success and failure. But most notable was a yawning gap between public and private sector: after adjusting for other factors, private utilities were nearly twice as likely as similar public utilities to meet the state’s conservation standards.
Finally, we analyzed overall conservation during the mandatory conservation period. And again, we found that, after accounting for other factors, private utilities conserved an average of 3% more water each month than their public counterparts during the mandatory restriction period. Although this difference is small in percentage terms, it reflects an enormous difference in absolute volume of water. This plot presents the distributions of conservation results from June 2015-May 2016 for local government utilities (green), and what it would have been if each utility had saved 3% more:
The areas within the white bars on the right side of the distribution represent the conservation that didn’t happen due to differences in ownership. Three percent greater conservation would have boosted public utilities’ restriction compliance rate from 51 to 62 percent.
In substantive terms, three percent greater conservation by California’s local government utilities during the mandate period would have reduced the state’s water consumption by 54.6 billion gallons—enough to supply the City of San Francisco for more than two years.
So what happened?
California is once again in the midst of a hot, dry summer; other parts of the world are, too. So it’s worth trying to figure out what’s behind the public-private disparity in drought response. Although it’s surprising at first blush, it’s actually a logical result of the institutions that govern water in America generally and California specifically. My next post will explain why.*
*Spoiler: as usual, it’s about money and politics. If you can’t wait for the next post, you can read the forthcoming article.
Terrible, horrible, no good, very bad measurement, part 4
My criticism of average bill ÷ Median Household Income (MHI) as a measure of household-level water affordability isn’t especially new. Lots of other people have pointed out the problems with this conventional methodology, and I’ve been presenting and publishing these arguments for more than twelve(!) years. But golden numbers are stubborn, and bad habits are hard to break—even when people know better.
The remarkable persistence of a bad idea
Over the years I’ve presented to hundreds of utility professionals and spoken personally with scores of managers, analysts, and rate consultants about the pathologies of %MHI and the virtues of alternative approaches. The reception is universally warm and agreeable, as most water professionals genuinely care about affordability and immediately recognize the fundamental flaws of the conventional approach.
Alas, there’s an and yet.
Even well-informed specialists continue to use and promote the tried-and-false conventional methodology. Researchers who recognize that average-bill-as-%MHI is deeply flawed employ it anyway because it’s easy and widely recognized (for example). Managers who know that %MHI is a misleading statistic continue to put it in front of their elected officials because it’s familiar and they feel that they have to use this metric because everyone else does, and because they believe it’s an EPA standard (it isn’t). Advocates, analysts, and rate consultants who I like and respect persist with the conventional approach in their studies, even when they know these metrics are fundamentally flawed (many have told me as much!).
Examples abound. The Alliance for Water Efficiency has a nice tool that’s designed to help water utilities model the financial impacts of various rate structures. Sensibly enough, their model includes an assessment of affordability. Unfortunately, it uses the familiar flawed metric:
UNC’s Environmental Finance Center continues to feature average-bill-as-%MHI as the sole affordability indicator on its rates dashboards. Folks at EFC know about the problems with this metric (they blogged about it here), but continue to display it prominently nonetheless.
Easy metrics die hard, it seems.
Water and sewer ratemaking is a niche specialty (to put it mildly). That’s good news, because if the community of specialists who analyze and design rates for a living get affordability metrics right, there’s a good chance that the utilities they serve will get affordability right, too.
I’ve developed better ways to measure affordability; others are working on this issue, too. At this stage there’s no consensus over the best metrics (naturally, I think mine are great). But abandoning the flawed measurement convention is an important first step.
Terrible, horrible, no good, very bad measurement, part 3
As my last couple of posts explain, the conventional method of measuring household-level water affordability is to divide a utility’s average residential bill by its community’s Median Household Income (MHI). If the resulting percentage is less than 2.0 or 2.5 (4.0 or 4.5 for water and sewer combined), then water is deemed “affordable;” if it’s greater, then water is “unaffordable.”
This post explains the fourth, and in my view most serious, problem with the conventional approach: the arbitrary threshold that it uses to define affordability.
An arbitrary standard
I’ve never—and I mean never—seen a theoretical or empirical rationale for a water/sewer affordability standard as a function of %MHI. Apparently the 2%MHI threshold emerged from the mists of federal regulatory history in a 1970s-era USDA rural grant program. At some point EPA began using the metric as part of its regulatory enforcement program. Despite those strange and ill-fitting origins, the 2.0 (or 2.5) %MHI affordability threshold is now held up as a definitive measure of household-level affordability, apparently for no other reason than convention.
The affordability of anything is rarely a strictly yes/no phenomenon—things are more or less affordable relative to the costs of other things and the resources available to pay for them. But rather than deal with those nuances, analysts simply cite precedent and invoke average-bill-as-%MHI as an EPA “standard” (though as I’ve noted before, it’s not an actual EPA standard, and the agency never intended %MHI to be a measure of household-level affordability).
The 2% or 4%MHI threshold has become what organizational theorists call a golden number. Golden numbers are standards that have no basis in theory or evidence, but are so widespread that they take on independent importance simply by virtue of their familiarity. The trouble with golden numbers like 2%MHI is that they preempt or short-circuit serious deliberation over values. Managers of a utility that satisfies the %MHI threshold can use the standard as an excuse dismiss affordability concern, even if many of its customers struggle to pay their bills.
Worse yet, a golden number like 2%MHI that has a tangential relationship to an EPA guideline puts a veneer of legitimacy on what is ultimately an arbitrary norm. Performance metrics reflect an organization’s values and help guide management toward decisions consistent with those values. In that regard, the golden number that predominates water rate analysis fails fundamentally.
Communities and their elected leaders should set affordability standards consistent with their values. Average bill <2%MHI distorts the affordability picture and distracts from meaningful consideration of the issue.