Water meters still don't reduce consumption

In 2009, Joe Buchanan wrote the following report that was published by Right to Water, a New Zealand group working to support the human right to water. I'm publishing it again because it's pretty much fallen off the web, and because Kapiti mayor Ross Church is claiming that introducing volumetric charging will reduce consumption by 25%. I suspect he's rehashing these statistics which, as Joe's paper shows, don't stand up to even a cursory examination.

The impact of metering on water use: a brief look at the statistics
Joe Buchanan
Wellington: Right to Water, November 2009.

Context
Throughout New Zealand, meters are being mooted as a tool for conserving water or reducing water use. Nationally, the Turnbull Group, Water NZ and the Institute of Professional Engineers (IPENZ) are promoting meters and volumetric charging on these grounds. Many local councils are also considering or implementing metering. Advocates of meters often cite impressive sounding figures about how much water meters can save. And many people, while concerned with the social impact of metering, especially on large families and those on low incomes, now believe the conservation benefits are worth it. Right to Water looked closely at the statistics to see if the evidence stacks up.  

Summary
 We found statistics were being cherry-picked, inflated and massaged to support pro-metering arguments. The most common errors in statistics purporting to show that water meters result in reduced demand were:
 

• Comparing actual residential use in metered areas with gross water use (including commercial and leakage) in non-metered systems, divided by the permanent population of the region. This will always show lower usage in metered systems, even if actual residential use is the same.
• Attributing all reductions in water use to the introduction of meters. Per capita water use has been decreasing in many parts of the country over the last decade or so, in both metered and non-metered systems.
• Cherry-picking statistics that are favourable to the pro-meter argument. This is especially common in references to international studies, which show a wide range of impacts from meters, including zero reduction in demand. These studies often acknowledge low confidence in the figures and that findings are not transferable.
  

Some examples:

 ONE
“Water meters can result in a 15% reduction in water use” e.g. TV3 news report October 25 2009  

This figure appeared in the book “Water: Use Less -Save More: 100 Water-Saving Tips for the Home” (The Chelsea Green Guides) by Jon Clift and Amanda Cuthbert (2007).

When contacted the author acknowledged that the figure was “picked up... from the various UK water companies who cite this figure.”

This figure appears to have no real basis. It appears to be an amalgam from various studies or PR but is neither indicative or predictive for the New Zealand situation.

TWO
“Studies have shown that the introduction of water meters results in a reduction of water use.” Beacon Pathways report (1)

The report cites two international studies as evidence for water savings from meters (2). Both studies however, are equivocal. They each cite mainly US and UK studies that show anything from 0–50% reduction in consumption following water meters.

Both studies acknowledge: 

• a drop off in savings after an initial period that savings might reflect other conservation  programs (such as public education)
• low confidence in the figures
• most savings are from reductions in outdoor use and are most effective where there is a large disparity in summer and winter use

The highest saving cited in these studies is from the Solomon Islands in the 1960s.

In general these studies show savings from water meters are made where summer use greatly exceeds winter use (for example parts of the USA where summer demand can be five to six times winter demand) or where pricing makes water use prohibitive (e.g. third / majority world countries). Neither situation applies to Wellington. 

THREE

“In Nelson… [when water meters were] introduced several years ago, peak summer demand reduced by 37 per cent” Murray Gibb (Water New Zealand) Dominion Post October 19 2009 (3)
 
The 37% was not a reduction in total peak period water use, as suggested. This dramatic figure was attained by first subtracting “typical” winter use from total use, then calculating the reduction in the remainder which was termed
“discretionary” use.
 
The actual reduction in water use between the peak periods used was 17%, and this includes any reductions in commercial use and reductions in leakage. 
 
It also includes reductions arising from education campaigns and simple year-to-year variation (over the same period water use decreased in areas that did not install meters). 
 
Using the same methodology Wellington’s peak water use decreased by 23% between the 2007/8 and 2008/9 summers.   There is no basis for claiming that reduced water demand in Nelson resulted from the introduction of water meters.

FOUR

“Meters can result in a 40% reduction in consumption when compared to non-metered systems” IPENZ report (4)  from Beacon Pathway paper. (5)

This claim is accompanied by an impressive graph that claims to show several unmetered Council areas that use far more water than unmetered regions. 
 
The figures are cherry picked and misleading. In this graph, and the table in the Beacon Pathways report on which it is based, the per capita usage shown for non-metered areas is total use, including system leakage and commercial use, apportioned across the number of residents. Conversely, in metered areas, the Right to Water figures shown are average amounts used by households alone. This comparison is either deliberately misleading or disingenuous. 
 
The figure shown for Manukau is around 190 litres per person per day. Gross use in Manukau is above 300 litres per person per day when commercial use and leakage is included (6), putting use into the range shown for unmetered areas. Similarly the figure shown for Tauranga is just over 200 litres per person per day, whereas gross use since metering is around 335 litres per person per day (6).
 
The lowest figure shown is for Nelson, about 160 litres per person per day. When the figure is recalculated using gross water use/permanent population to make it comparable to the figures shown for non-metered areas, Nelson usage rises to around 500 litres per person per day (7). This is higher than figures shown for some non-metered areas.
 
Some areas use more water for various reasons: horticultural use, climate, soil types, types of housing etc. Queenstown showed the highest per capita use  in these statistics, but the Queenstown Lakes Council estimate nearly 50% of the volume supplied is lost by system leakage, and that high use arises from commercial use (commercial use in Queenstown is not metered, aside from three hotels) and a large visitor population which is difficult to account for (8).
 
The figure for Kapiti (650 litres per person/day) is not the actual water use, it is a Council target. Water use across the Kapiti region varies, with very high use in Greater Otaki, which the Kapiti Coast District Council acknowledges includes large volumes used by commercial horticulture and lost due to leakage from ageing pipes (9).

It is particularly surprising that IPENZ, a professional organisation who one would expect to understand statistics, should promote these figures. Calculating per capita usage by including commercial and leakage does not reflect residential usage.   
                                               

References
1
 Best practice water efficiency policy and regulations (WA7060/3); A report prepared for Beacon Pathway Limited, May 2008
2
 A & N Technical Services, Inc. 2000. BMP Costs & Savings Study: A guide to the data and methods for cost-effectiveness analysis of urban water conservation Best Management Practices (BMPs) and Inman, D. and Jeffrey, P., 2006.
Urban Water Journal
, Vol. 3, No. 4, p. 127
–
143
3
 The 37% figure was from the Nelson City Council Water Supply Asset Management Plan 2009
–
2012).
4
IPENZ: Water: New Zealand’s valuable resource (undated).
 
5
 The figures appear to be from Lawton & Birchfield, Beacon Pathway Ltd 2008. Making New Zealand policy water conservation friendly. SB08 conference paper (Table 1).
 
6
 Blakemore, R. & Burton, S. 2007. An assessment of peak daily demand in Tauranga city and implications for water supply management.
7
 Statistics provided by Nelson City Council.
 
8
 Queenstown Water Demand Management Plan, Queenstown Lakes District Council, December 2008
 
9
 Kapiti Coast District Council 2002. Water Matters: Kapiti Coast District Sustainable Water Management Strategy.