THE NEW CSE REPORT ON THE UPPER SAN PEDRO BASIN
Another report has just come in about the water situation in the Upper
San Pedro Basin. In July, New Mexico's "Center for Sustainable
Economy" (the "CSE") issued "Fort Huachuca and the San Pedro
River: Improving Water Deficit Liability Calculations Through
Economic Modeling," whose full text is online at
1. Report page 2: "Currently, groundwater is
being pumped well in excess of aquifer recharge rates, leading to a
water deficit in the upper basin of between 5,144 and 7,000 acre feet
(af) per year, although more recent estimates place that deficit as
high as 10,800 ...."
-- However, the size of the deficit only has meaning if compared to the
size of the supply; drawing 1 gallon of water from a 5 gallon can is
noticeable, but taking 1 gallon from a 5000 gallon pool is
imperceptible. At the present rate of deficit, the Upper San
Pedro Basin has a 2000-year supply of water. If we must reduce
the problem to one number, the number shouldn't be "10,800 acre feet,"
it should be "1/20 of 1%".
-- It has been commented that the problem is not that the San Pedro
Basin will run out of water, but that enough water will be used so that
the San Pedro River stops flowing, and creates a disaster for the
ecology along the river. I'd say that the people who think the
San Pedro is in crisis must decide which they value more: a water
supply, or the ecology along the river. Preserving the ecology
requires wasting water. When flowing, the San Pedro is a shallow
sheet of water in a very hot place -- maximizing water loss by
2. Report pages 2-4: With respect to
groundwater, Fort Huachuca considers groundwater uses "attributable" to
the Fort to be uses "associated with water used on-post in base
operations, water used by 'people who live in the Sierra Vista
sub-watershed due to the presence of Fort Huachuca,' and water use
associated with 'off-post induced economic development' that would not
occur but for the presence of the Fort." In other words:
people "attributable" to the Fort include people assigned to the post,
people living off them, and people living off the people living off the
people assigned to the post.
-- CSE is increasing water use "attributable to" the Fort, by
increasing the number of people "attributable" to the Fort. This
results in a higher percentage of water use "attributable" to the
Fort. However, the same procedure could be applied to water use
"not attributable" to the fort, and that percentage would go up
too. But of course, percentages can't add up to over 100%, so, in
a process called "renormalization," all percentages must be reduced
until they add up only to 100 -- and you end up where you
started. CSE, an advocacy group, is so fixed on its goals that it
has forgotten basic mathematics.
3. Report page 5: the procedure used to calculate the Fort's water usage in 2002.
In 2002, the "attributable" population was calculated at 34,993, which in turn was the sum of two smaller groups.
-- The first group was 26,531 military personnel, contractors,
retirees, survivors, and family members (all of whom may live on or off
the Fort). This number was not an actual count, but was reduced
"for double counting and for those non [sic] living in the Sierra Vista
subwatershed (3%)." Note that reducing by 3% is the same as
multiplying by .970; such a "multiplier" pops up again, two paragraphs
-- The second group was 7093 people in "an interrelated and
interdependent population of induced employees and family
members." This number came from a calculation, not a count.
-- -- The number of enlisted personnel assigned to the Fort was
4199. That was multiplied by .332288, "the assumed employment
multiplier for enlisted military personnel".
-- -- -- It has been commented that using enlisted personnel amounts to
excluding officers and warrant officers. However, "enlisted
military personnel" is the term used in CSE's report. Clearly
there is an error here; whose, is not clear.
-- -- --- Using an "assumed" multiplier is bad enough, but using a
6-digit multiplier is ridiculous; a 6-digit multiplier is accurate to
one part in a million, but Cochise County only has roughly 100,000
people, using a 6-digit multiplier in Cochise County implies accuracy
down to 1/10 of a person -- common sense is left behind. In any
event, the result of this calculation was multiplied by .968, "a
discount factor to account for those not living in the Sierra Vista
subwatershed." Again, this number gives the appearance of more
precision than really exists, and there is no possible good reason for
using .968 here, versus .970 for the same factor two paragraphs up.
-- -- The number of contractors and government civilians supported by
the Fort was 5879. That was multiplied by .383932, "the assumed
multiplier for contractors and civilians," again giving a false
appearance of precision. The result was multiplied by the ".968"
multiplier -- as if .970 weren't already more precise than this
-- -- The sum of the above two paragraphs was multiplied by .7325, "the
assumed proportion of family members [not] holding induced jobs."
-- -- The result of the above multiplication was multiplied by 2.74,
"the ratio of jobs to population in the Sierra Vista
subwatershed." This multiplication gave the final result of 7093.
That's how the total figure of 34,993 was calculated for 2002.
4. Report pages 5-6: the heart of CSE's position
In December 2006, Robert Carreira, at Cochise College, came up with not
34,993 people, but only 18,543. Most of the reduction came from
simply counting heads, and eliminating any use of "economic
multipliers." CSE disagrees strongly with Carreira, and would not
only restore economic multipliers, but make them more important.
As noted above, CSE applies its "multiplier" approach only to
Fort-related water use, not to "non-Fort" use. The result is an
artificial inflation of the Fort's effect on water use. If the
"multiplier" approach is applied to both sets of users, the net result
is: no change in their respective shares of water use. CSE
would impose great complications that achieve nothing.
5. Report pages 6-10: CSE's criticisms of mere "head counting"
CSE groups multiplier effects into three main categories
-- "the money Fort Huachuca spends on salaries .... In 2005, the
Fort estimated this amount to be $491 million, supporting 13,379
jobs.... Off duty, such personnel ... induce water use in the
local economy by spending money on goods and services from local
businesses that use water .... when off base military personnel
play golf, buy food, landscape their homes, wash their cars, or eat out
they induce water use at golf courses, supermarkets, nurseries,
carwashes, and restaurants. These businesses, in turn, purchase at
least some of their goods and services locally, inducing further use of
water by suppliers."
-- -- It has been noted that the "$491 million, 13,379 jobs" may be double-counting some money and jobs.
-- "direct procurement of goods and services by Fort Huachuca,
estimated [in 2005 at] 339.4 million in Sierra Vista alone....
businesses that provide ... laundry, hospitality, or maintenance
services or food, medicines, equipment, technology, or other goods all
use water themselves ... and also induce further water use by
purchasing at least some of their goods and services locally.
Another pathway is water use by households that benefit from all this
-- "the economic activity generated by the Fort just because of its
presence.... visits by out of area family members, by those on
business with the Fort, or by those touring the Fort's museums or
historical sites.... Another example is the retiree population
located in Sierra Vista due to the presence of Fort Huachuca."
-- -- It has been noted that CSE includes tourists to the Fort, but
omits all other tourists. This, of course, is another example of
CSE's unfair overreaching to increase the share of water use
attributable to the Fort.
-- -- A reader has noted an interesting point: "when they figure jobs/businesses attributable to the Fort, do
they take into account that Sierra Vista is the only large
shopping center for everybody else in Cochise County? For instance, if
WalMart sells $xmillion per year in Sierra Vista, they aren't only
selling to Army personnel. Even without the Fort, the rest of the
county would need a main shopping and entertainment center closer than
Tucson." People will want to chew over the implications of this for themselves.
And so on. In short: money is spent not just once but over
and over, and its effect is multiplied every time it is spent.
6. Report pages 10-13: CSE's criticism of "arbitrary assumptions"
None of these criticisms relates to CSE's main point about using
economic multipliers, not "head counts," so these criticisms are
discussed as briefly as possible.
"Assumption 1: Population growth necessary to account for groundwater deficit"
-- CSE asserts that it has caught Carreira in an arithmetic
error: Carreira either tripled a population increase, or tripled
the actual per capita water usage.
-- The error does not exist. Carreira stated that a certain
figure was "per year" over several years, but CSE incorrectly read that
as a total figure for the entire year.
-- Even if the arithmetic error existed, it could not validate the "paradigm shift" that CSE wishes to impose.
"Assumption 2: Water use intensity is constant across all population groups"
-- CSE notes, correctly, that both earlier models "contain an implicit
assumption that the water use profile of the population attributable to
Fort Huachuca is no different than that of the general population in
the Sierra Vista subwatershed." However, because CSE does not
state any way in which Fort personnel might use water differently than
others, the safest course is use the assumption in the earlier models.
"Assumption 3: Validity of DES [Arizona Department of Economic Security] population data"
-- CSE notes that Carreira "questions the methodology of the DES [and]
suggests the U.S. Census data has greater accuracy and should be used"
but nonetheless relies "heavily in [sic] the DES data for period
analysis not covered by the Census." Here, the CSE is merely
carping, not actually criticizing. Even if Census data is more
reliable than DES data, DES data may still be acceptable when Census
data does not exist -- and Carreira acknowledges the changes in
validity between the two types of data.
"Assumption 4: All same-sized Arizona communities are identical"
-- Here as above, CSE is carping, not criticizing. CSE is
complaining about Carreira's argument that because many Arizona cities
are growing, Fort Huachuca isn't the only reason that Sierra Vista is
growing. On the face of things, Carreira's argument is correct;
there is natural population growth, by an excess of births over deaths,
and also, as Cochise County residents know, there is great pressure
from realty interests to build large housing tracts in Cochise County,
especially as "bedroom suburbs" for Tucson. However, CSE does not
acknowledge natural population growth, nor show any awareness of the
pressure from builders to proceed with "bedroom suburbs" -- which, as
residents of Cochise County know, is immense.
"Assumption 5: All new jobs are filled locally"
-- CSE notes that Carreira "assumes that an increase in jobs
automatically correlates to a reduction in unemployment.... This
is an oversimplification" and that Carreira also "assumes that if new
jobs are created then family members of active military fill
[them]. This assumption ... is unfounded in economic
literature." Actually, Carreira's point is merely that if
military family members are counted as "attributable" to the Fort, and
they fill jobs that are "attributable" to the Fort, they must not be
double counted. Recognizing that a correction must be made is not
the same as stating that the correction must be set at 100%, as CSE
"Assumption 6: Economic multipliers are not appropriate to estimate population growth"
-- CSE argues that Carreira's choice, of using head-counting instead of economic multipliers, has two major problems:
-- -- that using multipliers must be right, because economists have
been using multipliers for 50 years, especially since computers came
-- -- -- That is, "all the kids do it" so it must be right.
-- -- that Carreira's estimate is "of people who are directly connected
to the Fort. This implied assumption is that secondary effects and
tertiary effects (indirect and induced) are inconsequential...."
-- -- -- That is, Carreira is wrong because CSE says so, a classic example of arguing "from authority."
-- CSE's arguments are typically used between demanding children and
resistant parents; they have no place in a serious discussion between
7. Report pages 14-19: CSE's proposed theories of calulation
CSE states that "Given the inherent problems with population-based
models for modeling Fort Huachuca's water deficit liability an economic
based model may be a far better approach" and to "suggest three
alternative approaches to modeling Fort Huachuca's water deficit
liability based on its overall economic impact."
-- the "Gross regional product approach," "to assign water deficit
liability based on the share of regional economic activity attributable
to Fort Huachuca in the Sierra Vista subwatershed."
-- the "Water intensity - labor market approach," "to consider the
labor associated with the Fort's impact via the number of employees
attributable to the Fort in primary, indirect and induced activities
and the average water use per employee." This result would be
measured against "the total water consumed in the sub-watershed."
-- -- this section has some standard figures for water usage per
employee per day in different kinds of work: a government office,
136; a "primary metal industry," 1318; and "for food and kindred
products," 1967. So to use less water, shrink farming, and expand
government office work.]
-- the "Econometric approach," "to use econometrics," meaning to find a
set of variables that can be manipulated in a mathematical formula. For
instance, the county's total water deficit could be the sum of one
formula for "economic activity associated with the Fort" plus another
formula for "economic activity ... not associated with the Fort."
The goal is to determine "the estimated share of the deficit
attributable to the Fort".
CSE does not show any calculations, but merely concludes that "All of
these models are likely to find a water deficit liability greater than
the 54.3% found by Fort Huachuca's 2002 biological assessment. In fact,
preliminary data suggest that a model based on share of economic
activity may find the Fort's water deficit liability to be as high as
-- If 80% is the highest number that CSE's equations, using
"preliminary data," have generated, then 80% cannot be a representative
result. A report that gives one extreme result, but does not show
the entire range of results, with the data and equations that led to
them, is not ready for publication.