delivered the opinion of the Court,

Once again this Court is called upon to determine whether the funding of Texas public schools violates the Texas Constitution.1 Three groups of school districts raise three separate challenges.

The plaintiffs, 47 districts led by West Orange-Cove Consolidated Independent School District,2 which educate over a fourth of the State’s more than 4.3 million school children, contend that property taxes, though imposed locally, have become in effect a state property tax prohibited by article VIII, section 1-e of the Texas Constitution, because the State leaves districts no meaningful discretion to tax below maximum rates. Article VIII, section 1-e states simply: “No State ad valorem taxes shall be levied upon any property within this State.”3 We held in Edgewood III that “[a]n ad valorem tax is a state tax when it is imposed directly by the State or when the State so completely controls the levy, assessment and disbursement of revenue, either directly or indirectly, that the authority employed is without meaningful discretion.”4

The other two groups, intervenors, totaling an additional 282 districts, also educate about a fourth of the State’s school children. One group is led by Edgewood Independent School District,5 the other by Alvarado Independent School District.6 *752Intervenors contend that funding for school operations and facilities is inefficient in violation of article VII, section 1 of the Texas Constitution, because children in property-poor districts do not have substantially equal access to education revenue.

All three groups also contend that the public school system cannot achieve “[a] general diffusion of knowledge” as required by article VII, section 1 of the Texas Constitution, because the system is underfunded.

Article VII, section 1 states:

A general diffusion Of knowledge being essential to the preservation of the liberties and rights of the people, it shall be the duty of the Legislature of the State to establish and make suitable provision for the support and maintenance of an efficient system of public free schools.7

This provision sets three standards central to this case. One is that the public school system be efficient. In Edgewood I, we held:

There is no reason to think that “efficient” meant anything different in 1875 [when article VII, section 1 was written] from what it now means. “Efficient” conveys the meaning of effective or pro*753ductive of results and connotes the use of resources so as to produce results with little waste; this meaning does not appear to have changed over time.8

As applied to public school finance, we added, constitutional efficiency requires that “[c]hildren who live in poor districts and children who live in rich districts must be afforded a substantially equal opportunity to have access to educational funds.”9 We have referred to efficiency in the broader sense as “qualitative”, and to efficiency in the context of funding as “financial”.10 The parties have also referred to financial efficiency as “quantitative”.

Another standard set by the constitutional provision is that public education achieve “[a] general diffusion of knowledge ... essential to the preservation of the liberties and rights of the people”.11 We have labeled this standard “adequacy”,12 and the parties have adopted the same convention. The label is simply shorthand for the requirement that public education accomplish a general diffusion of knowledge. In this context, the word “adequate” does not carry its broader dictionary meaning: “[cjommensurate in fitness; equal or amounting to what is required; fully sufficient, suitable, or fitting.”13 Our responsibility in this case is limited to determining whether the public education system is “adequate” in the constitutional sense, not in the dictionary sense. That is, we must decide only whether public education is achieving the general diffusion of knowledge the Constitution requires. Whether public education is achieving all it should — that is, whether public education is a sufficient and fitting preparation of Texas children for the future — involves political and policy considerations properly directed to the Legislature. Deficiencies and disparities in public education that fall short of a constitutional violation find remedy not through the judicial process, but through the political processes of legislation and elections.

A third constitutional standard is that the provision made for public education be “suitable”. We have mentioned this requirement only once, in Edgewood IV:

Certainly, if the Legislature substantially defaulted on its responsibility such that Texas school children were denied access to that education needed to participate fully in the social, economic, and educational opportunities available in Texas, the “suitable provision” clause would be violated.14

In essence, “suitable provision” requires that the public school system be structured, operated, and funded so that it can accomplish its purpose for all Texas children.

Article VII, section 1, makes it “the duty of the Legislature” to provide for public education.15 The judiciary’s role, though important, is limited to ensuring that the constitutional standards are met. We do not prescribe how the standards should be met.

In this case, the district court, after a five-week bench trial, found in favor of *754the school districts on all their claims except for inefficient operations funding and enjoined the defendants16 (collectively “the State defendants”) from continuing to fund the public schools.17 The court issued its judgment on November 30, 2004, but stayed the effect of its injunction for ten months, until October 1, 2005, “to give the Legislature a reasonable opportunity to cure the constitutional deficiencies in the finance system”.18 The Legislature convened in regular session in January 2005, and while it gave much attention to public education issues, it did not reach consensus. After adjournment, the Governor called the Legislature into special session on June 21, 2005, and that session was in progress when we heard oral argument in this case on July 6. That session also ended without enactment of public education legislation, and the Governor immediately called a second special session to convene July 21. Thirty days later, the Legislature again adjourned without enacting public education legislation. The district court’s injunction has been stayed by the State defendants’ appeal.19

We now hold, as did the district court, that local ad valorem taxes have become a state property tax in violation of article VIII, section 1-e, as we warned ten years ago they inevitably would, absent a change in course, which has not happened.20 Although the districts have offered evidence of deficiencies in the public school finance system, we conclude that those deficiencies do not amount to a violation of article VII, section 1. We remain convinced, however, as we were sixteen years ago, that defects in the structure of the public school finance system expose the system to constitutional challenge.21 Pouring more money into the system may forestall those challenges, but only for a time. They will repeat until the system is overhauled.

The judgment of the district court is modified and affirmed in part, reversed in part, and remanded for reconsideration of the award of attorney fees.

I

We begin by summarizing first the structure of the public school finance sys*755tem in Texas as relevant to the issues in this case, then the evidence regarding the adequacy of public education thus financed, and finally the procedural background of the case. The record contains evidence through the end of the 2003-2004 school year, and our discussion of the present status of the system generally refers to that time frame unless otherwise noted.

A

The basic structure of Texas’ present public school finance system derives from Senate Bill 7 enacted by the Legislature in 1993.22 We have twice described the system thoroughly,23 including its historical evolution,24 and will not repeat here all that we have said before. In 1995, we held in Edgewood IV (among other things) that the system under Senate Bill 7 did not violate article VII, section 1, or article VIII, section 1-e of the Texas Constitution but noted that the system was “minimally acceptable only when viewed through the prism of history.”25 The parties in this case contend that the operation of the system has changed since Edgewood IV, and so in the discussion that follows we include several comparisons between then and now.

Texas has a little over 4.3 million children in public schools, and the number is growing by more than 72,500 per year. More than half qualify for federally subsidized, free or reduced-price lunches and are therefore categorized by the State as economically disadvantaged.26 About 15% have limited proficiency in English. According to the State defendants’ expert, the annual cost of public education is $30-35 billion, or about $7,000-8,000 per student, depending on what expenses are counted. More than half of the cost is funded by ad valorem taxes imposed by independent school districts on local property. The State funds only about 38% of the cost, down from about 43% in Edge-wood IV,27 the lowest level in more than 50 years.28 The balance, usually around 8-9%, comes from the United States government.

There are 1,031 independent school districts 29—more than four times the number of counties. A fourth of public school students are educated in 12 districts in seven counties;30 half are educated in 45 dist*756ricts.31 The largest district, Houston ISD, has 211,499 students, more than the combined student population in half of all the other districts put together.32 Two-thirds of the districts have fewer than 1,200 students each; half have fewer than 700 each; almost a fourth have fewer than 350 each; 11 districts have fewer than 60 each.33 Divide Independent School District in Kerr County, the smallest, has 10 students.34

The Legislature’s decision to rely so heavily on local property taxes to fund public education does not in itself violate any provision of the Texas Constitution,35 but in the context of a proliferation of local districts enormously different in size and wealth, it is difficult to make the result efficient — meaning “effective or productive of results and connot[ing] the use of resources so as to produce results with little waste”36 — as required by article VII, section 1 of the Constitution.37 Compensation must be made for disparities in the amount of property value per student so that property owners in property-poor districts are not burdened with much heavier tax rates than property owners in property-rich districts to generate substantially the same revenue per student for public education. According to the evidence, in 2001, Dew ISD in Freestone County had an adjusted taxable value of $300,384,388, a “weighted average daily attendance” (“WADA”) of 147.43 students,38 and thus $2,037,488/ WADA, while Boles ISD in Hunt County had an adjusted taxable value of $8,831,414, a WADA of 876.95 students, and thus $10,071/WADA. This 200-to-l disparity was 700-to-l in Edgewood I.39 Also, many districts have been created as tax havens40 — lots of property and few students — allowing property owners to escape paying their fair share of the cost of public education in Texas and making it more difficult to achieve efficiency.41 A *757system that operates with an excess of resources in some locales and a dearth in others is inefficient, as we held in Edgewood I42 and Edgewood II.43 Summing up in Edgewood III, we said:

The inefficiency was this gross disparity both in tax burden and in tax spending. To put it graphically, in some areas of the state, education resembled a motorcycle with a 1000-gallon fuel tank, and in other areas it resembled a tractor-trailer rig fueled out of a gallon bucket. Some vehicles were flooded, some purred along nicely, and some were always out of gas. A fleet of such vehicles is not efficient, even though a few of them may reach their destination. We did not hold that efficiency requires absolute equality in spending; rather, we said that citizens who were willing to shoulder similar tax burdens, should have similar access to revenues for education.44

The large number of districts, with their redundant staffing, facilities, and administration, make it impossible to reduce costs through economies of scale.45 Bigger is not always better, but a multitude of small districts is undeniably inefficient. The justification offered for this situation is that as a matter of public policy, public schools should be locally controlled, although it has never been clear why the legitimate benefits of local control are so entirely inconsistent with efficiency in funding.46 Districts are firmly entrenched and powerfully resistant to meaningful change, and while matters have improved somewhat over the past century, the number of school dis*758tricts has not declined significantly in the past two decades.47

The purpose of Senate Bill 7 was to try to make funding public education with local property taxes efficient by reducing the effects of the vast disparities among the more than 1,000 independent school districts. School maintenance and operations (“M & 0”) are funded separately from facilities. Tax rates set yearly are capped at $1.50/$100 valuation for M & O48 (except for seven districts in Harris County49 ), as they have been for sixty years,50 and $0.50/$100 valuation for debt service on facilities (referred to as “I & S”, for “inter-est and sinking fund”).51 For M & 0, disparities in available revenue among the school districts are reduced in two ways: by supplementing property-poor district tax revenues with state funds through the Foundation School Program (“FSP”) under chapter 42 of the Education Code,52 and by “recapture” — a scheme under chap-fer 41 of the Education Code by which property, tax revenue is taken from property-rich (“chapter 41”) districts and given to property-poor (“chapter 42”) districts— referred to by some as “Robin Hood”. Chapter 41 districts educate 12.3% of Texas students.

The FSP has two tiers for M & 0. Tier 1 guarantees to all districts that tax at or above the rate of $0.86 per $100 valuation (and all districts but one do) a basic allotment of $2,537 per student in “average daily attendance” (“ADA”), subject to various special allotments and adjustments for district and student characteristics.53 Thus, any district with less than $295,000 value/ADA ($2,537 = .0086 x $295,000) receives FSP funds to supplement local revenue as if it had that much property value per student, up to an $0.86 tax rate. The basic allotment includes a per capita distribution (usually $250-30054) for each stu*759dent from the Available School Fund (“ASF”),55 which consists of certain appreciation from the Permanent School Fund, as required by the Constitution.56 Districts with $295,000 value/ADA or more receive no Tier 1 state funds, although they do receive the ASF distribution. Tier 2 guarantees that for each $0.01 of tax rate above $0.86, the yield will be $27.14/ WADA57 — the yield a district would have if it had $271,400 value/WADA ($27.14 = .0001 x $271,400). Thus, a district with only $100,000 value/WADA could generate only $10/WADA in local tax revenue for each $0.01 of tax rate, and the FSP would add $17.14/WADA to make up the difference. A district taxing at the maximum $1.50 rate is thus guaranteed $1,736.96/ WADA (($1.50 — $0.86)58 x $27.14). Districts with at least $271,400 value/WADA receive no Tier 2 funds. In Edgewood IV, the Tier 1 basic allotment was $2,300/AD A, and the Tier 2 guaranteed yield was $20.55/WADA, or $1,315.20/WADA at a tax rate of $1.50.59

Recapture helps fund the FSP60 and further equalizes access to revenue among districts. Most districts with more than $305,000 value/WADA, and which therefore receive no funds under FSP Tier 1 or Tier 2 for M & O, must transfer that excess value- — in practical reality, the tax revenue derived from it — to the State or other districts for distribution under the FSP to chapter 42 districts that have less.61 In Edgewood IV, the statutory retained value/WADA cap was $280,000.62 A chapter 41 district may choose to effectuate the transfer in one of five ways:63 (1) consolidate with a chapter 42 district, to reduce the value/WADA to $305,000 or less;64 (2) detach territory to a chapter 42 district, to achieve the same effect;65 (3) purchase “average daily attendance credits” from the State;66 (4) agree to pay to *760educate students in a chapter 42 district;67 or (5) consolidate tax bases with a chapter 42 district — combining the finance mechanism while leaving district administration independent.68 The third and fourth options call for a district to simply write a check directly to the State or other districts.69 Options (1), (2) and (5) are rarely used.

The net effect of recapture, generally speaking, is that a district with more than $305,000 value/WADA must pay, either to the State or to another district or districts directly, its local tax revenue that exceeds what the retained value generates. Thus, for example: a district with a 10,000 WADA and $366,000/WADA in property value, taxing at a $1.50 rate, for a revenue of $5,490/WADA, would be required to purchase 2,000 credits from the State at $5,490 each, totaling $10,980,000, to increase its deemed WADA to 12,000, reducing its deemed value/WADA to $305,000, leaving it $4,575/WADA for its own use. For the 1993-1994 school year, the first under Senate Bill 7, 99 chapter 41 districts transferred $433 million. For 2003-2004, 134 chapter 41 districts transferred over $1 billion. For 2004-2005, the amount of recapture is estimated to be over $1.2 billion. Thus, recapture has doubled in less than a decade, and in 12 years it may have almost tripled.

Several other statutory provisions reduce recapture payments and thus in effect raise the chapter 41 districts’ average actual retained value/WADA above the statutory limit of $305,000. A chapter 41 district receives an early-agreement discount of the lesser of 4% of its total recapture payment or $80/student for agreeing to the payment by September l,70 and an efficiency discount of the lesser of 5% or $100/student for agreeing to pay a chapter 42 district or districts directly, rather than *761sending the payment to the State.71 These discounts saved chapter 41 districts $43.4 million in recapture payments in the 2003-2004 school year, but the efficiency discount also benefitted chapter 42 districts, who received $81.4 million more than they would have had the recapture payments been made through the State. Neither of these discounts existed when Edgewood IV was decided.

Also, as we said above, most districts may retain only $305,000 value/WADA, but there is an exception: for a district taxing at the maximum $1.50 rate, recapture cannot reduce its revenue/WADA, excluding the ASF distribution, below the level for the 1992-1993 school year.72 This exception was designed to mitigate the impact of Senate Bill 7 on the wealthiest districts and was initially intended to last only three years,73 but it has become permanent and has even been increased.74 There are 34 of these so-called “hold-harmless” districts, educating less than 1% of Texas students.75 On average they retain $421, 373/WADA instead of $305,000/WADA, thereby saving about $38 million in revenue that would otherwise have been transferred to the FSP. This raises the average retained value/WADA of all chapter 41 districts to $341,457. In Edgewood IV we held that the effect of the hold-harmless districts was not so great as to render the entire system inefficient, especially since they were to be phased out in three years.76

Senate Bill 7 thus retains in its design a gap in available per-student M & 0 revenue attributable to property-wealth disparities among school districts. We discussed this gap in Edgewood IV. To compare its size then and now, we must exclude hold-harmless districts, discounts, and other factors that effectively raise the statutory cap on a district’s retained wealth/WADA, disregard for purposes of a benchmark comparison the differences between ADA and WADA and other Tier 1 and Tier 2 formula differences, and assume a maximum tax rate of $1.50. With these assumptions, the FSP guarantees $4,273.96/student,77 while a district with a tax base of $305,000/student has $4,575/student — a difference of $301.04/student, or 7%. Under the statutory parameters that existed in Edgewood IV, this gap was $584.80/student, or 16%.78 If a $300 ASF distribution is added to the *762non-FSP revenue, the gap is enlarged to 14% at present and 24% in Edgewood TV.

In actual operation, however, this gap is wider. According to the intervenors’ expert, on average, at a tax rate of $1.48, chapter 41 districts’ revenue is $5,457/ WADA while chapter 42 districts’ revenue is $4,330/WADA, a difference of $1,127/ WADA or 26%. By comparison, at the time of Edgewood TV (as reflected in the record but not our opinion), the average tax rate was only $1.17, chapter 41 districts’ average revenue was $3,510/WADA, and chapter 42 districts’ average revenue was $3,005/WADA, a difference of $505 or 17%. The proportional size of the gap in actual operation has thus increased by about half, from 17% to 26%. But as we have noted, we did not consider in Edge-wood IV the effect of hold-harmless districts that would have made the gap much larger, and other discounts and factors that would affect these figures did not exist. According to the intervenors’ expert, these elements together contribute at least $599 to the present difference. Assuming they would have impacted the calculations at the time of Edgewood IV similarly, the increase in the gap since then would be much smaller.

Looking to the extremes rather than at averages, with similar tax rates near the maximum, districts at or above the 95 percentile level of property value per student have $5,895/WADA, while districts at or below the 5 percentile level have only $4,217/WADA, a difference of $1,678, or 40%. In Edgewood IV, this gap was projected to be about $600 — actually, according to the evidence, $4,440 vs. $3,868, or 16% — with hold-harmless districts phased out and all districts taxing at a $1.50 rate.79

To generate the same revenue per student that the FSP guarantees to an average chapter 42 district that taxes at the maximum $1.50 rate, taking into account differences between Tier 1 and Tier 2 formulas, the average chapter 41 district need only tax at the rate of $1.33. A different comparison was made in Edgewood IV. There we calculated that to generate $3,500/WADA, which the trial court had found to be the cost of an adequate education — or in the words of' article VII, section 1 of the Texas Constitution, “[a] general diffusion of knowledge”80 — districts at or below the 15 percentile level of property value per student, averaging a $26.74 yield per $0.01 of tax, were required to tax at a $1.31 rate while districts at or above the 85 percentile level, averaging a $28.74 yield per $0.01 of tax, needed only a $1.22 rate.81 Thé parties in this case have not attempted to replicate this calculation for current data.

Since the 1993-1994 school year, which we reviewed in Edgewood IV, M & O tax rates have migrated to the $1.50 maximum. That year, most districts’ tax rates were below $1.20; now, only about 2% of the districts, with less than one-fourth of 1% of the students, tax below $1.20. The concentration of districts at the higher tax rates is shown in the following table:

*763[[Image here]]

In the 1998-1994 school year, school districts spent only 83.3% of the revenue that could have been generated at maximum tax levels for public education; now they spend over 97%. The trial court found that—

any remaining capacity is not realistically available because accessing this capacity would require (1) a virtually 100% tax collection rate (practically impossible); (2) the repeal of any property tax exemptions (politically improbable); and (3) a district to have stable or increasing property values. In other words, these percentages represent virtual full funding for most of the larger districts in the system.

Up to this point we have been describing the financing of school maintenance and operations. For instructional facilities (as opposed to facilities used for administration and extracurricular purposes),82 the FSP includes what may be considered a third tier83 that partially equalizes access to funding up to the maximum $0.50 tax rate to support bonds.84 Through the Instructional Facilities Allotment (“IFA”), the State guarantees districts a yield of $35/ADA for each $0.01 of I & S tax rate for bonds for new facilities, with certain exceptions,85 up to a maximum of the lesser of $250/ADA or $100,000,86 for the life of the bonds. To retire preexisting debt, the Existing Debt Allotment (“EDA”) guarantees districts $35/ADA for each $0.01 of I & S tax rate up to $0.29, with certain exceptions.87 However, unlike the FSP Tier 1 and Tier 2, both the IFA and the EDA are subject to funding being appropriated by the Legislature. A district that receives an EDA grant for debt service in one biennium has no guarantee that the grant will be renewed at all or at the same level for the life of the debt, and must assume the risk that the assistance provided will be limited. Neither the IFA nor the EDA assists districts too poor to levy taxes in the first place.

Property-poor districts are given priority for IFA funding88 but not for EDA. New and existing IFA awards now total about $270 million, but substantial requests have gone unfunded. In the 2002-*7642003 school year, 520 school districts got $457.5 million in EDA allotments. Together, the two allotments equalize districts’ access to revenue for 90% of eligible debt service. I & S rates are excluded from Tier 2 allotments89 and are not used in recapture calculations,90 so that districts’ retained wealth for I & S taxes is not capped as it is for M & 0 taxes. Thus, property-rich districts have more than 20 times as much value/WADA to tax for facilities as property-poor districts. By contrast, in Edgewood IV, I & S tax rates for debt service were included within Tier 2 allotments91 and used for recapture calculations to cap retained value.

The district court found:

Lacking sufficient funding, property-poor districts such as the Edgewood In-tervenors have been unable to provide adequate facilities for all the children in their districts. Substandard conditions include: overcrowded schools and classrooms; out-of-date buildings, equipment and fixtures; inadequate libraries, science labs, cafeterias, gymnasiums, and other school facilities.

The court identified health and safety concerns raised by some conditions, like inadequate heating, air conditioning, and ventilation, and science laboratories without emergency eye washes, fume hoods, exhaust fans, and other safety features. The court found that inadequate facilities negatively impacted student scores on standardized tests, and that “property-poor districts like the Edgewood Intervenors lack all the facilities essential to providing students a learning environment in which to attain a general diffusion of knowledge.”

The State makes a few other contributions to public education finance besides the programs and allotments we have described. It paid districts $110 per student for the 2003-2004 school year, and it has funded other projects, like Head Start and the High School Completion Initiative. But 95% of all funds for public education flow through the Foundation School Program, including the IFA and the EDA, and are thereby equalized among the districts. The other 5% includes tax revenue that is not recaptured, taxes above the $1.50 M & O level in seven districts, and I & S tax revenue that exceeds the IFA and EDA yields or is not included under these allotments. On the whole, about 85% of the student population resides in districts with revenue equivalent to a district with $271,400/student.

B

The finance system we have described funds an education system with four integrated components: a state curriculum, a standardized test to measure how well the curriculum is being taught, accreditation standards to hold schools accountable for their performance, and sanctions and remedial measures for students, schools, and districts to ensure that accreditation standards are met.

The Legislature has prescribed the following basic public school curriculum:

Each school district that offers kindergarten through grade 12 shall offer, as a required curriculum:

(1) a foundation curriculum that includes:

(A) English language arts;

(B) mathematics;

(C) science; and

*765(D)social studies, consisting of Texas, United States, and world history, government, and geography; and

(2) an enrichment curriculum that includes:

(A) to the extent possible, languages other than English;

(B) health;

(C) physical education;

(D) fine arts;

(E) economics, with emphasis on the free enterprise system and its benefits;

(F) career and technology education; and

(G) technology applications.92

The Legislature has also required that “[t]he State Board of Education [‘SBOE’], with the direct participation of educators, parents, business and industry representatives, and employers shall by rule identify the essential knowledge and skills of each subject of the required curriculum that all students should be able to demonstrate”.93 After years of consultation and study, over a thousand public meetings, and thousands of public comments, the SBOE adopted the Texas Essential Knowledge and Skills (“TEKS”) curriculum for use beginning in the 1998-1999 school year. The Legislature has also required the SBOE to “determine curriculum requirements for the minimum, recommended, and advanced high school programs that are consistent with the required curriculum”.94 In 2000, after study and input, the SBOE revised these programs, making them more difficult and restricting future participation in the minimum program. Beginning in the 2004-2005 school year, no high school student may be enrolled in the minimum program unless the student, the student’s parent or guardian, and a school administrator agree.95

To correspond to the curriculum changes, the Legislature required the development of a new state standardized test — the Texas Assessment of Knowledge and Skills (“TAKS”) test — to replace the Texas Assessment of Academic Skills (“TAAS”) test.96 The TAKS test, developed after consultation with educators and testing experts and first given in the spring of 2008, has harder questions, covers more subjects — five (reading/English Language Arts, writing, math, science, and social studies) instead of three (for most of TAAS’s duration) — and is given at more grade levels.97 A student must pass portions of the test for promotion to the fourth and sixth grades (and in school year 2007-2008 to the ninth grade),98 and cannot graduate high school without passing an exit-level test first administered in the eleventh grade.99 A student may take the test as many as three times in order to pass it for promotion,100 and for any student who fails any part, the district must provide accelerated instruction,101 an individualized graduation plan,102 and study guides to the student’s parents.103 A stu*766dent may retake a necessary exit-level test any time it is administered.104 There are special tests for Spanish-speaking students, students with limited English proficiency, and disabled students.105

At the Legislature’s direction,106 the SBOE determined after public input what scores would constitute passing — “cut scores” — and decided that they should be lower at first, increasing over three years, to give teachers and students time to adjust to the new and more difficult test. To pass the 2004 TAKS test, an 11th grader was required to answer correctly 37 out of 73 questions (50.7%) on the reading test, 24 out of 55 questions on the science test (43.6%), and 25 out of 60 questions (41.7%) on the math test. The passing rates for the 2004 TAKS test, calculated statewide and for five different student populations — African-American, Hispanic, white, economically disadvantaged (“ED”),107 and limited-English-proficiency (“LEP”)108— are shown in this chart:109

[[Image here]]

[[Image here]]

These passing rates were somewhat lower than those for 2002, the last year the TAAS test was given. TAAS passing *767scores had increased significantly leading up to 2002, as shown in the following chart for reading, math, and writing tests in grades three through eight and ten (1994-2002), and social studies and science tests in the eighth grade (1995-2002):

[[Image here]]

In 1994 the minimum passing rate for an “academically acceptable” rating was 25%, and by 2002 it had climbed to 55%.

For accountability, schools and districts are rated “exemplary”, “recognized”, “academically acceptable”, or “academically unacceptable”110 based on “academic excellence indicators” chosen by the Commissioner of Education.111 Those indicators are standardized test scores, high-school completion rates, and seventh- and eighth-grade dropout rates. For each accountability rating, the required test passing rate must be met in each of five student groups — all students, African-American, Hispanic, white, and economically disadvantaged. The completion rate is the percentage of students entering the ninth grade who have either completed or are continuing their high school education four years later. If any school in a district is rated “academically unacceptable”, the district cannot be rated “exemplary” or “recognized”. With certain exceptions, the minimum requirements for each rating are as follows:

[[Image here]]

These test passing rate requirements remain the same for the three years 2004-2006 that the test cut scores are phased in, then they increase incrementally for three years to the point that a district must have a test passing rate of at least 70% for all student groups in all subjects to be rated “academically acceptable”. After 2006, GED recipients will no longer be counted as completers. In 2005, the maximum dropout rate for an “academically acceptable” rating falls to 1%, and in 2007 a broader definition of dropout will be used.

Before the change to the TAKS test in 2003, many districts improved their ac*768creditation rating, but at the same time the number of “academically unacceptable” districts also grew. After the change to the harder test, ratings predictably slid, although the number of “academically unacceptable” districts also declined. The following table summarizes these trends:

[[Image here]]

As sanctions for an “academically unacceptable” rating, the Commissioner of Education may, among other things, order a school board to hold a public hearing on the deficiency,112 order the school board president and superintendent to appear before the Commissioner,113 or order an on-site evaluation and recommendations for reform.114 After a year, the Commissioner may appoint a board of managers in place of the school board.115 After two years, the Commissioner may annex the district to an adjoining district.116 (For example, the Commissioner has recently announced her intention to annex the Wilmer-Hutchins Independent School District to the Dallas Independent School District, pending pre-clearance by the United State Department of Justice under the federal Voting Rights Act.117)

Academic success is also measured by the National Assessment of Educational Progress (“NAEP”) achievement test, as witnesses for all parties at trial acknowledged. In 2000, controlling for socioeconomic and family characteristics, Texas was first out of 47 states overall, first for white students, fifth for African-American students, ninth for Hispanic students, first for fourth- and eighth-graders in math, and second in rate of improvement. In 2003, Texas rankéd first in the nation in closing the gap between African-American and white fourth-graders in math, and second in the nation in closing the gap between Hispanic and white fourth-graders in math and reading. But unadjusted NAEP data, which may more accurately reflect college preparation, showed Texas sinking to 37th among the states in fourth-*769grade and eighth-grade reading, although it had risen to 22nd in fourth-grade math and remained 34th in eighth-grade math.

Because more students are failing the TAKS test than were failing the TAAS test, and because passing the TAKS test is now required for promotion to the fourth and sixth grades, the districts must spend more for remediation through summer school, remedial classes, curriculum specialists, reduced class-size, and more math and science teachers. There is a worsening undersupply of teachers, aggravated by high attrition and turnover. Additionally, the percentage of LEP and ED students, who generally cost more to educate, has increased. The FSP provides an extra bilingual education allotment for LEP students118 and an extra compensatory education allotment for ED students,119 but the attendance weights used to determine those allotments have not increased since 1985.

Based on the eleventh-grade exit-level TAKS test, the percentages of student groups meeting the college-readiness standards of the Texas Higher Education Coordinating Board for English and math are shown in this table:

[[Image here]]

In 2003, Texas ranked last among the states in the percentage of high school graduates at least 25 years old in the population. Texas also has a severe dropout problem: more than half of the Hispanic ninth-graders and approximately 46% of the African-American ninth-graders leave the system before they reach the twelfth grade. The gaps between white students on the one hand and African-American and Hispanic students on the other are especially troublesome since the African-Americans and Hispanics are projected to be about two-thirds of Texas’ population in 2040. According to the plaintiffs’ expert, if these gaps are not reduced, Texas will “have a population that not only will be poorer, less well-educated, and more in need of numerous forms of state services than its present population, but also less able to support such services ... [and] less competitive in the inereas-ingly international labor and other markets.”

It is difficult to quantify the cost of an adequate education — one that achieves a general diffusion of knowledge. The parties offered competing cost function studies prepared by economists, examining statistical relationships between spending and student performance, taking into account student and school characteristics. We do not attempt to describe here the detailed procedures used in the studies but focus only on the conclusions. The study offered by plaintiffs and intervenors, done by Dr. Jennifer Imazeki and Dr. Andrew Res-chovsky (the “I/R study”), concluded that to achieve a 55% statewide pass rate on the 2005 TAKS test would require additional spending of from $1,653 billion to $6,171 billion — between $401 and $1,511 more per student. The study offered by the State defendants, done by Dr. Lori Taylor (the “Taylor study”), concluded that *770some 17% of school districts — 117 out of 695 studied120 — could not achieve a 55% pass rate in 2003 at a $1.50 tax rate without additional revenue of $563-$731 million. The district court accepted the I/R study and found the Taylor study flawed in several respects. Despite those flaws, the court continued, the Taylor study showed that school funding was insufficient to provide an adequate education in many districts. The court also found that both studies underestimated the costs of meeting accreditation standards. As the State defendants noted, however, the cost studies and court findings overlook the reality that almost all schools are meeting accreditation standards with current funding.

C

Four of the plaintiff school districts initiated this action in April 2001, alleging that the $1.50 maximum M & O tax rate had become in effect a state property tax prohibited by article VIII, section 1-e of the Texas Constitution,121 because they and other districts had lost all meaningful discretion to tax at any lower rate.122 Forty school districts intervened in two groups, six with Edgewood ISD and 34 with Alvarado ISD, opposing the plaintiffs’ position but asserting that the public school finance system was inefficient, inadequate, and unsuitable in violation of article VII, section 1 of the Texas Constitution,123 because the State does not provide sufficient funding.124 About three months after suit was filed, the trial court dismissed the plaintiffs’ claims on the pleadings, concluding that the plaintiffs could not prove an unconstitutional state ad valorem tax unless about half of the school districts, more or less, were taxing at maximum rates, and by the plaintiffs’ own admission, less than a fifth of the districts were at the cap at the time.125 The court of appeals affirmed, not because so many school districts were still taxing below maximum rates, but because it concluded that the plaintiffs had not alleged that taxing at maximum rates was necessary just to provide an accredited education, rather than being used for additional programs.126

We reversed. The Legislature, we said, is constitutionally obligated “to make suitable provision for a general diffusion of knowledge through, free public schools”,127 and because it “has chosen to rely heavily on school districts to discharge its duty”,128 school districts must tax at levels necessary to achieve the constitutional mandate 129 as well as to meet statutory accreditation standards that the Legislature has imposed to achieve a general diffusion of knowledge.130 If school districts are forced to tax at or near maximum rates to meet constitutional and statutory requirements, then control over local ad valorem tax rates and spending effectively shifts to the State, depriving school districts of any meaningful discretion to tax below the rate *771cap set by the State or to spend on programs other than those required by the State and the Constitution.131 The result, we again warned as we had in Edgewood IV,132 would be a state ad valorem tax in violation of article VIII, section 1-e.133 We concluded that the plaintiffs’ pleadings had fairly alleged that such a violation was occurring.134

After we remanded the case to the trial court, 285 other school districts were added as plaintiffs or intervenors. The plaintiffs joined the intervenors in their article VII, section 1 claims that the public school finance system is inadequate and unsuitable, but not in their claims that the system is inefficient. To facilitate trial of the case, the plaintiffs, the Edgewood intervenors, and the State defendants agreed that each group would designate a few districts, which they called “focus districts”, on which the evidence would center.135

On November 80, 2004, the district court rendered judgment for the plaintiffs on all their claims and for the intervenors on all but one of their claims. In extensive findings and conclusions, the court held that local ad valorem taxes had become a state ad valorem tax in violation of article VIII, section 1-e, that the public school finance system is inadequate and unsuitable in violation of article VII, section 1, and that the funding of school facilities is inefficient in violation of article VII, section 1. The court refused to find that the funding of school maintenance and operations is also inefficient. The court enjoined the State defendants “from giving any force and effect to the sections of the Education Code relating to the financing of public school education (Chapters 41 and 42 of the Education Code) and from distributing any money under the current Texas school financing system until the constitutional violations are remedied.” The court stayed this injunction until October 1, 2005, “to give the Legislature a reasonable opportunity to cure the constitutional deficiencies in the finance system”. Finally, the court awarded the plaintiffs and intervenors $4,273,120.50 in attorney fees through proceedings in this Court.136

The State defendants and each of the two intervenor groups filed separate, direct appeals to this Court.137 We noted probable jurisdiction in all three cases,138 consolidated them,139 and expedited briefing and oral argument.

*772II

At the outset, the State defendants challenge the district court’s subject matter jurisdiction on three grounds: that the plaintiff and intervenor school districts lack standing to assert any of their constitutional claims, that their claims under article VII, section 1 are nonjusticiable political questions, and that article VII, section 1 is not self-executing and thus cannot be enforced by court action. With one exception, we have previously rejected all of these contentions, either expressly or implicitly, in this case when it was last before us or in the other cases in which the constitutionality of the public school finance system has been at issue. In none of our prior cases has a school district’s standing to challenge the public school finance system under article VII, section 1 been challenged, and we have not specifically addressed that issue.

To the extent we have already spoken to these issues, the State defendants urge us to reconsider. Our prior decisions have not ended litigation over school finance once and for all, and the State defendants argue that this is because the courts cannot give sufficiently certain meaning to the constitutional standards. Each new case, they argue, threatens to drag the courts inescapably into a morass of policy-making where they do not belong and from which they will not be able to extricate themselves, endlessly second-guessing the detailed structures of public education. We think our prior opinions on these matters are clear enough and remain correct, but because the issues are important, we address each of the State defendants’ arguments in turn.

A

In the plaintiffs’ earlier appeal, we held “that the plaintiff school districts in this case have standing to assert their claims.”140 At that time, the plaintiffs’ only claims were under article VIII, section 1-e. The State defendants argue that to allow the plaintiffs and intervenors standing to raise any of their constitutional claims is inconsistent with the general rule that governmental entities do not possess constitutional rights, citing our 1966 decision in Deacon v. City of Euless,141 the court of appeals’ decision in Nueces County Appraisal District v. Corpus Christi People’s Baptist Church, Inc.,142 and the dissent in the prior appeal.143 In Deacon, we held that a home-rule city could not complain that legislation affecting its power to annex territory was unconstitutionally retroactive,144 but we did not establish a broad rule that a governmental entity cannot sue to declare a statute unconstitutional. The court of appeals noted this in Nueces County, stating that Deacon did not “establish an mondad rule that a county may never attack the constitutionality of a state statute”,145 and held that a county was the proper party to challenge a statute that allowed churches to regain tax exemptions they had lost by failing to timely file claims.146 Thus, of the three authorities on which the State defendants rely, only the prior dissent in this case lends support to their position.

In answer to the dissent, we said:

*773In Nootsie, Ltd. v. Williamson County Appraisal District, we held that a county appraisal district had standing to seek a declaratory judgment that the Legislature had unconstitutionally defined open-space land for tax purposes to include ecological laboratories. We see no difference in the standing of an appraisal district to assert its claims in Nootsie and the standing of the school districts here.147

Following Nootsie, we held in Proctor v. Andrews that the City of Lubbock had standing to challenge the constitutionality of a statute requiting arbitration of disciplinary disputes with its police officers “ ‘because it is charged with implementing a statute it believes violates the Texas Constitution.” ’148 That interest, we said, gave the City “ ‘a sufficient stake in [the] controversy to assure the presence of an actual controversy that the declaration sought will resolve.” ’149 The City’s claims in Proctor were that the statute was an unconstitutional delegation of legislative power to a private authority and an impermissible infringement on the constitutional powers of home-rule cities. A year later, in Wilson v. Andrews, we held that the City also had standing to challenge the statute on due process and equal protection grounds.150 “[T]he constitutional demands of standing,” we explained, “are that there is (a) a real controversy between the parties, which (b) will be actually determined by the judicial declaration sought. Under this standard, Lubbock indeed has standing.”151

The State defendants argue that Nootsie is distinguishable because there, the appraisal district had an “interest, as a tax-collecting entity, in ensuring the collection of those tax obligations legally due”, whereas here, the school districts “are merely representing the potential interests of students and parents in their districts.” But this argument with respect to article VII, section 1 ignores what we said in the plaintiffs’ earlier appeal, that the Legislature has required school districts to achieve the goal of a general diffusion of knowledge.152 This requirement is expressly imposed by section 11.002 of the Education Code, which states:

The school districts and charter schools created in accordance with the laws of this state have the primary responsibility for implementing the state’s system of public education and ensuring student performance in accordance with this code.153

School districts’ interest in discharging this duty is not merely representative of constituent students and taxpayers. And with respect to article VIII, section 1-e, the State defendants’ argument overlooks the fact that school districts’ interests in not collecting an illegal tax may conflict with taxpayers’ interest. For one thing, *774school districts’ concerns could be met simply by raising the cap on ad valorem taxes, something that could well be expected to aggravate taxpayers’ concerns.

The dissent agrees that school districts have standing to challenge public school finance under article VIII, section 1-e154 but argues that they have no standing to assert challenges under article VII, section 1 because that provision confers rights only on school children, not districts.155 We think the guarantee of public free schools assured by article VII, section 1, extends not only to school children but to the public at large, which is vitally concerned that there be a general diffusion of knowledge. We agree that the provision creates no rights in school districts,156 but such rights are not a prerequisite for standing to assert that the provision has been violated. Standing to assert a constitutional violation depends on whether the claimant asserts a particularized, concrete injury. As we recently explained in Brown v. Todd:

under Texas law, standing limits subject matter jurisdiction to cases involving a distinct injury to the plaintiff and “a real controversy between the parties, which ... will be actually determined by the judicial declaration sought.” Texas Workers’ Compensation Comm’n v. Garcia, 893 S.W.2d 504, 517-18 (Tex. 1995); see also State Bar v. Gomez, 891 S.W.2d 243, 245 (Tex. 1994).

... [W]e may look to the similar federal standing requirements for guidance. [Texas Ass’n of Bus. v. Texas Air Control Bd., 852 S.W.2d 440, 444 (Tex. 1993).] “To meet the standing requirements of Article III [of the United States Constitution], ‘[a] plaintiff must allege personal injury fairly traceable to the defendant’s allegedly unlawful conduct and likely to be redressed by the requested relief.’” [Raines v. Byrd, 521 U.S. 811, 818-819, 117 S.Ct. 2312, 138 L.Ed.2d 849 (1997)] (quoting Allen v. Wright, 468 U.S. 737, 751, 104 S.Ct. 3315, 82 L.Ed.2d 556 ... (1984)) .... The United States Supreme Court has “consistently stressed that a plaintiffs complaint must establish that he has a ‘personal stake’ in the alleged dispute” and that the injury suffered is “concrete and particularized.” Id. at 819, 117 S.Ct. 2312 ... (citing Lujan v. Defenders of Wildlife, 504 U.S. 555, 560-61, 112 S.Ct. 2130, 119 L.Ed.2d 351 ... (1992)).157

The dissent argues that school districts have suffered no injury because they have lost no funds that belong to them, and that “[t]he injury [the districts have] alleged in this case was suffered only by school students”.158 But the school districts have alleged the very same injury that the appraisal district alleged in Nootsie and the city in Proctor, which is that they are being required to implement unconstitutional statutes. And this is also the same *775injury that gives the districts standing to complain that local ad valorem taxes have become a state property tax in violation of article VIII, section 1-e.

The dissent argues that “[t]he districts do not complain that they are affirmatively compelled to perform unconstitutional teaching, testing, or any other services; they complain only that they are underfunded.”159 The dissent disregards the plaintiffs’ statement in their brief:

the school districts stand in precisely the same position as the county appraisal district in Nootsie: all are required to implement statutes that they regard as unconstitutional.

Likewise, the Edgewood intervenors asserted standing in the district court because “they have been charged with implementing statutes which they believe violate the Texas Constitution.”

The dissent repeatedly states that government agencies do not have standing to sue for increased funding,160 tacitly assuming that funding of governmental functions is always a matter of policy and allocation of resources. The dissent’s statements are not true when funding is required by the Constitution, as the districts claim here. In Vondy v. Commissioners Court, we held that by providing that justices of the peace be compensated by salary, article XVI, § 61161 requires commissioners courts to set reasonable salaries.162 Similarly, we held in Mays v. Fifth Court of Appeals that a commissioners court must pay a district court’s court reporter the salary determined by the district court as authorized by statute.163 The dissent attempts to distinguish these cases as “involv[ing] nondiscretionary ministerial acts”,164 but the commissioners court in Vondy had discretion to determine what salary was reasonable, and making that determination was not simply a ministerial act. Similarly, the Legislature has discretion under article VII, section 1 to determine how to structure and fund the public education system to achieve a general diffusion of knowledge. However, in Vondy, as in this case, governmental discretion is circumscribed by the Constitution. Article VII, section 1 requires that public school finance be efficient and adequate to provide a general diffusion of knowledge. The school districts have standing to insist that this provision be obeyed.

Finally, the dissent argues that to allow school districts standing to challenge public school finance under article VII, section 1 distorts the constitutional issues because school districts will advance only their own interests and not those of students or others, and “fundamental reforms may be overlooked if school districts may *776assert Article VII claims by themselves”.165 Of course, a party’s standing to assert a claim does not depend on its ability or willingness to look out for interests other than its own, and the dissent has no authority to the contrary. Not all districts share the same view of the public school system. The plaintiffs and intervenors do not, and more than two-thirds of the districts have not joined this action. The fact that districts disagree among themselves and may also disagree with some students, parents, teachers, and taxpayers does not deprive them of standing to assert the claims they have. The dissent argues that the record in this case would look far different if it could be brought only by individuals and not by districts, but this ignores two facts: individuals who the dissent thinks would take different positions were free to intervene, and essentially all of the arguments made by the school districts in this case have been made by individuals in prior cases. The suggestion that the plaintiff and intervenor districts could not find one student out of 4.3 million to join in asserting their positions, when dozens have joined prior actions, cannot be taken seriously.

Like the appraisal district in Nootsie and the city in Proctor and Wilson, the school districts here have a real controversy to be resolved in this case, -and thus they have standing. The dissent’s charge that we have abandoned judicial restraints like standing because this is a noteworthy case166 is simply not true. We do not suggest, of course, that only school districts have standing to raise the issues that have been raised here. Prior cases challenging public school finance have involved individual claimants as well as school districts. The State defendants do not contest that individuals would have standing to raise the claims in this case. The interests of individual taxpayers in suitable, adequate, efficient public education and in avoiding a state property tax might well diverge from those of their school districts. But individuals’ standing to assert these constitutional claims does not deprive school districts of standing to assert the same claims.

Accordingly, we conclude that the plaintiff and intervenor school districts have standing to assert the claims made in this case.

B

Preceding our decision in Edgewood I, a divided court of appeals held that whether the public school finance system is efficient within the meaning of article VII, section 1 “is essentially a political question not suitable for judicial review.” 167 We firmly rejected that view:

This is not an area in which the Constitution vests exclusive discretion in the legislature; rather the language of article VII, section 1 imposes on the legislature an affirmative duty to establish and provide for the public free schools. This duty is not committed unconditionally to the legislature’s discretion, but instead is accompanied by standards. By express constitutional mandate, the legislature must make “suitable” provision for an “efficient” system for the “essential” purpose of a “general diffusion of knowledge.” While these are admittedly not precise terms, they do provide a standard by which this court must, when called upon to do so, measure the constitutionality of the legislature’s actions. *777We do not undertake this responsibility lightly and we begin with a presumption of constitutionality. Nevertheless, what this court said in only its second term, when first summoned to strike down an act of the Republic of Texas Congress, is still true:

[W]e have not been unmindful of the magnitude of the principles involved, and the respect due to the popular branch of the government.... Fortunately, however, for the people, the function of the judiciary in deciding constitutional questions is not one which it is at liberty to decline.... [We] cannot, as the legislature may, avoid a measure because it approaches the confines of the constitution; [we] cannot pass it by because it is doubtful; with whatever doubt, with whatever difficulties a case may be attended, [we] must decide it, when it arises in judgment.

Morton v. Gordon, Dallam 396, 397-398 (Tex. 1841). If the system is not “efficient” or not “suitable,” the legislature has not discharged its constitutional duty and it is our duty to say so.168

We reaffirmed this position in the plaintiffs’ earlier appeal in this case, extending it to include not only efficiency but the other standards in article VII, section 1:

The final authority to determine adherence to the Constitution resides with the Judiciary. Thus, the Legislature has the sole right to decide how to meet the standards set by the people in article VII, section 1, and the Judiciary has the final authority to determine whether they have been met.169

Nevertheless, the State defendants argue that the history of school finance litigation over the past two decades requires us to reconsider whether the court of appeals in Edgewood I was right after all, and that the issues of adequacy, suitability, and efficiency under article VII, section 1 are all nonjusticiable political questions because they involve either “a textually demonstrable constitutional commitment of the issue[s] to a coordinate political de*778partment; or a lack of judicially discoverable and manageable standards for resolving [them]”—the two principal tests set out in the United States Supreme Court’s landmark decision in Baker v. Carr.170 The first test is satisfied, the State defendants continue, by the language of article VII, section 1, which expressly makes the establishment, support, and maintenance of free public schools “the duty of the Legislature”.171 The second test has been demonstrated, according to the State defendants, because “[sjixteen years of school finance litigation ha[ve] disproved the Court’s expressed hope in Edgewood I that it could pass upon the legitimacy of the system without intruding upon the Legislature’s province.”

The tests of Baker v. Carr define non-justiciable political questions for purposes of demarcating the separation of powers in the federal government under the United States Constitution.172 Assuming that the same tests would serve equally well in defining the separation of powers in the state government under the Texas Constitution, we disagree with the State defendants that either of the tests categorically precludes the judiciary from deciding the issues raised in this case under article VII, section 1. As we explained in the first appeal in this case, “[b]y assigning to the Legislature a duty, [article VII, section 1] both empowers and obligates.”173 The Constitution commits to the Legislature, the most democratic branch of the government, the authority to determine the broad range of policy issues involved in providing for public education. But the Constitution nowhere suggests that the Legislature is to be the final authority on whether it has discharged its constitutional obligation. If the framers had intended the Legislature’s discretion to be absolute, they need not have mandated that the public education system be efficient and suitable; they could instead, have provided only that the Legislature provide whatever public education it deemed appropriate. The constitutional commitment of public education issues to the Legislature is primary but not absolute.

Nor do we agree with the State defendants that the constitutional standards of adequacy, efficiency, and suitability are judicially unmanageable. These standards import a wide spectrum of considerations and are admittedly imprecise, , but they are not without content. At one extreme, no one would dispute that a public education system limited to teaching first-grade reading would be inadequate, or that a system without resources to accomplish its purposes would be inefficient and unsuitable. At the other, few would insist that merely to be adequate, public education must teach all students multiple languages or nuclear biophysics, or that to be efficient, available resources must be unlimited. In between, there is much else on which reasonable minds should come together, and much over which they may differ. The judiciary is well-accustomed to applying substantive standards the crux of which is reasonableness. This is not to say that the standards in article VII, section 1 involve no political considerations beyond the judiciary’s power to determine. *779We have acknowledged that much of the design of an adequate public education system cannot be judicially prescribed. Litigation over the adequacy of public education may well invite judicial policy-making, but the invitation need not be accepted. The judiciary’s choice is not between complete abstinence from article VII, section 1 issues, and being, in the State defendants’ words, “the arbiter of education and policy, overseeing such issues as curriculum and testing development, textbook approval, and teacher certification”. Rather, the judiciary’s duty is to decide the legal issues properly before it without dictating policy matters. The constitutional standards provide an appropriate basis for judicial review and determination.

The State defendants argue that if the standards of article VII, section 1 had judicially manageable content, litigation over the constitutionality of the public education system would not have lasted as long as it has. It is true, of course, as this case illustrates, that disagreements over the construction and application of article VII, section 1 persist. But such disagreements are not unique to article VII, section 1; they persist as to the meanings and applications of due course of law, equal protection, and many other constitutional provisions. Indeed, those provisions have inspired far more litigation than article VII, section 1, which has been at the heart of only a few lawsuits in two decades. Moreover, the continued litigation over public school finance cannot fairly be blamed on constitutional standards that are not judicially manageable; the principal cause of continued litigation, as we see it, is the difficulty the Legislature has in designing and funding public education in the face of strong and divergent political pressures.

To this point, we have assessed the State defendants’ arguments as if the tests of Baker v. Carr would apply under the Texas Constitution. If they do—a question we need not reach—their application is limited. In the federal system, political questions are a rarity. The United States Supreme Court has held only two issues to be nonjusticiable political questions: whether the military was properly trained,174 and whether the impeachment trial of a federal judge may be conducted before a Senate committee instead of the entire Senate.175 The Court did not hold the one-man-one-vote congressional apportionment issue in Baker v. Carr to be a political question, and it has refused to hold issues to be political questions in at least seven other cases.176 The Court did not even discuss the doctrine in Bush v. *780Palm Beach County Canvassing Board177 and Bush v. Gore,178 eases in which the winner of the 2000 national presidential election was at stake, certainly a “political issue” as conventionally understood, if not within the meaning of Baker v. Carr. Some have questioned whether the political question doctrine has any real vitality at all.179 This Court has never held an issue to be a nonjusticiable political question, and we have referred to the doctrine only in passing.180 The courts of appeals have applied the doctrine only rarely.181

A few state supreme courts have refused to adjudicate constitutional challenges to public school finance on the ground that the issues were nonjusticiable political questions,182 but many others have rejected the argument.183 Like the majority of these states, we conclude that the separa*781tion of powers does not preclude the judiciary from determining whether the Legislature has met its constitutional obligation to the people to provide for public education.

c

The State defendants’ third challenge to the district court’s jurisdiction is that article VII, section 1 is not self-executing and thus does not allow for court *782action to enforce its provisions. The concept of a constitutional provision as self-executing, long-recognized in the law, was restated by this Court in 1898 in Mitchell County v. City National Bank of Paducah, Ky.:

“A constitutional provision may be said to be self-executing if it supplies a sufficient rule by means of which the right given may be enjoyed and protected, or the duty imposed may be enforced; and it is not self-executing when it merely indicates principles, without laying down rules by means of which these principles may be given the force of law.”184

In that case, Mitchell County issued bonds to obtain funds for building bridges and a courthouse and jail.185 When the county refused to pay on the bonds, the holder sued.186 Article XI, section 2 of the Texas Constitution states.that “[t]he construction of jails, court-houses and bridges ... shall be provided for by general laws”,187 and section 7 of the same article prohibited certain cities and counties from incurring debt for certain projects without imposing taxes for a sinking fund to repay the debt.188 The Legislature had enacted statutes authorizing the kind of bonds Mitchell County had issued, but the county argued in part that the statutes improperly added to the constitutional requirements.189 We held that the constitutional provisions were self-executing only in prohibiting conflicting laws and did not preclude statutes passed to effectuate the provisions.190

We have used the doctrine of .self-executing constitutional provisions to preclude judicial action only once. Article XVI, section 59 of the Texas Constitution makes “[t]he conservation and development of all of the natural resources of this State ... public rights and duties” and requires the Legislature to “pass all such laws as may be appropriate thereto”.191 In City of Corpus Christi v. City of Pleasanton, we held that this provision did not provide a basis to declare “what types of conduits and reservoirs may be used for the transportation and storage of water, lawfully obtained and lawfully used”, and especially, to “declare the use of certain types of conduits and reservoirs to be. unlawful when the Legislature itself, by necessary implication, has declared them to be lawful.” 192 The Legislature had acted to conserve river water, and the constitutional provision did not authorize us to impose a *783different, specific requirement.193

The standards of article VII, section 1— adequacy, efficiency, and suitability — do not dictate a particular structure that a system of free public schools must have. We have stressed this repeatedly. In Edgewood I, we wrote: “Although we have ruled the school financing system to be unconstitutional, we do not now instruct the legislature as to the specifics of the legislation it should enact; nor do we order it to raise taxes.”194 In Edgewood II, we said: “We do not prescribe the means which the Legislature must employ in fulfilling its duty.”195 In Edgewood III, we reiterated: “As before, we do not prescribe the structure for ‘an efficient system of public free schools.’ ... We have not, and we do not now, suggest that one way of school funding is better than another, or that any way is past challenge, or that any member of this Court prefers a particular course of action ..., or that one measure or another is clearly constitutional.”196 But like the constitutional provisions in Mitchell County, article VII, section 1 dictates what the public education system cannot be: it cannot be so inadequate that it does not provide for a general diffusion of knowledge, or so inefficient that districts which must achieve this general diffusion of knowledge do not have substantially equal access to available revenues to perform their mission, or so unsuitable that it cannot because of its structure achieve its purpose.

Thus, we agree with the State defendants that article VII, section 1 does not provide the courts a basis for declaring what education or finance systems will alone satisfy its standards. But the provision is self-executing insofar as it prohibits any system that fails to meet those standards.

Ill

We come, then, to the question whether the public education system violates the requirements of article VII, section 1.

A

The State defendants argue that in determining whether the public school finance system is adequate, efficient, and suitable, the test should be whether there is any rational basis for the Legislature to have fashioned the system the way it has. The plaintiffs and intervenors disagree, arguing that a rational-basis test is too deferential to the Legislature, but they have not suggested a clear alternative. In Edgewood I and II, we did not find it necessary to articulate a standard of review; the public school finance system was simply not “efficient” by any stretch of the word. Nor did we state a standard of review expressly in Edgewood IV, though it was a much closer case.

But the standard of review the Court applied in Edgewood IV is apparent from the opinion. For example, the property-poor school districts complained that the public school finance system was inefficient because as Senate Bill 7 was structured, the richest districts could always raise on average $600/WADA more than the poorest districts.197 We concluded that this gap was not unreasonable, given that the poorest districts could still provide an adequate education at a slightly higher tax rate.198 The property-poor districts also *784complained that the system was inefficient because Senate Bill 7 excepted a few “hold harmless” districts from full recapture for a three-year phase-in period.199 Again, we held that this accommodation was not so unreasonable as to render the entire system inefficient.200 In essence, we refused to find a constitutional violation when the challenged aspect of the system was not arbitrary. This comports with what we said in 1931 in Mumme v. Marrs:

The purpose of [article VII, section 1] as written was not only to recognize the inherent power in the Legislature to establish an educational system for the state, but also to make it the mandatory duty of that department to do so.... The Legislature alone is to judge what means are necessary and appropriate for a purpose which the Constitution makes legitimate. The legislative determination of the methods, restrictions, and regulations is final, except when so arbitrary as to be violative of the constitutional rights of the citizen.201

The State defendants argue that there is no difference between a rational-basis test and a test based on arbitrariness. An action is arbitrary when it is taken without reference to guiding rules or principles 202—in other words, the State defendants argue, when it has no rational basis. But the phrase, “rational basis”, is more often associated with the minimal requirement a classification must meet to be consistent with the constitutional guarantee of equal protection when no suspect class or fundamental right is involved.203 In that context, the idea is that the government is permitted to give classes disparate treatment, notwithstanding the constitutional guarantee, as long as it has a rational basis for doing so. The same idea does not fit in the context of article VII, section 1. That provision does not allow the Legislature to structure a public school system that is inadequate, inefficient, or unsuitable, regardless of whether it has a rational basis or even a compelling reason for doing so. But the provision does allow the Legislature, of necessity, much latitude in choosing among any number of alternatives that can reasonably be considered adequate, efficient, and suitable. These standards do not require perfection, but neither are they lax. They may be satisfied in many different ways, but they must be satisfied.

Article VII, section 1 allows the Legislature a large measure of discretion on two levels. The Legislature is entitled to determine what public education is necessary for the constitutionally required “general diffusion of knowledge”, and then to determine the means for providing that education. But the Legislature does not have free rein at either level. “[T]he Legislature may [not] define what constitutes a general diffusion of knowledge so low as to avoid its obligation to make suitable provision imposed by article VII, section 1.” 204 And “[w]hile the Legislature certainly has broad discretion to make the myriad policy decisions concerning education, that *785discretion is not without bounds.” 205 It would be arbitrary, for example, for the Legislature to define the goals for accomplishing the constitutionally required general diffusion of knowledge, and then to provide insufficient means for achieving those goals. If the Legislature’s choices are informed by guiding rules and principles properly related to public education— that is, if the choices are not arbitrary— then the system does not violate the constitutional provision.

For article VII, section 1, as for other provisions, “[t]he final authority to determine adherence to the Constitution resides with the Judiciary.” 206 As we have said, “ ‘a mere difference of opinion [between judges and legislators], where reasonable minds could differ, is not a sufficient basis for striking down legislation as arbitrary or unreasonable.” ’ 207 At the same time, “legislative action ... is not without bounds.” 208 In assessing challenges to the public education system under article VII, section 1, courts must not on the one hand substitute their policy choices for the Legislature’s, however undesirable the latter may appear, but must on the other hand examine the Legislature’s choices carefully to determine whether those choices meet the requirements of the Constitution. By steering this course, the Judiciary can assure that the people’s guarantees under the Constitution are protected without straying into the prerogatives of the Legislature.

Whether the statutory provisions creating the public school system are arbitrary and therefore unconstitutional is a question of law.209 To the extent that this determination rests on factual matters that are in dispute, we must, of course, rely entirely on the district court’s findings. But in deciding ultimately the constitutional issues, those findings have a limited role.210

B

The framers of the Texas Constitution of 1876 premised their mandate of free public education on the axiom that “[a] general diffusion of knowledge [is] essential to the preservation of the liberties and rights of the people”.211 The delegates to the Constitutional Convention of 1875 were deeply divided over how best to provide for a general diffusion of knowledge, finally adopting article VII, section 1 by a vote of 55 to 25.212 No subject was more controversial213 or more extensively debated.214 *786Some believed that the job of educating the public should be done through private institutions rather than by public ones.215 Nearly all were for local control, having chafed under the centrally controlled schools of the Reconstruction Era. Many insisted on guaranteeing adequate funding sources, while others were concerned that the expense would overburden taxpayers.216 Still, none questioned that a general diffusion of knowledge was essential to self-governance.217

The truth of the axiom had long been, and remains, beyond doubt. The framers may have borrowed the phrase from Thomas Jefferson’s 1778 draft of “A Bill for the More General Diffusion of Knowledge”, which provided for free public education in Virginia, observing that “the most effectual means of preventing [tyranny] would be, to illuminate, as far as practicable, the minds of the people at large”.218 George Washington used the same phrase a few years later in his 1796 farewell address, encouraging his countrymen:

Promote, then, as an object of primary importance, institutions for the general diffusion of knowledge. In proportion as the structure of a government gives force to public opinion, it is essential that public opinion should be enlightened.219-

The importance of public education was also recognized on the frontier. In 1836, Texas declared its independence from Mexico, citing as one of the principal reasons that the Mexican government had—

failed to establish any public system of education, although possessed of almost boundless resources, (the public domain,) and although it is an axiom in political science that, unless a People are educated and enlightened, it is idle to expect the continuance of civil liberty, or the capacity for self-government.220

Adopted at the same time, Texas’ first Constitution provided: “It shall be the duty of Congress, as soon as circumstances will permit, to provide, by law, a general system of education.”221

*787Under article VII, section 1 of the Constitution of 1876, the accomplishment of “a general diffusion of knowledge” is the standard by which the adequacy of the public education system is to be judged.222 To achieve such a system, the Legislature has chosen to use local school districts.223 Borrowing from two statutory pronouncements, the district court concluded:

To fulfill the constitutional obligation to provide a general diffusion of knowledge, districts must provide “all Texas children ... access to a quality education that enables them to achieve their potential and fully participate now and in the future in the social, economic, and educational opportunities of our state and nation.” TEX. EDUC. CODE § 4.001(a) (emphasis added). Districts satisfy this constitutional obligation when they provide all of their students with a meaningful opportunity to acquire the essential knowledge and skills reflected in ... curriculum requirements ... such that upon graduation, students are prepared to “continue to learn in postsecondary educational, training, or employment settings.” TEX. EDUC. CODE § 28.001 (emphasis added) ....

We agree, with one caveat. The public education system need not operate perfectly; it is adequate if districts are reasonably able to provide their students the access and opportunity the district court described.

The system the Legislature has devised prescribes an education curriculum, and by means of accreditation standards, holds schools and districts accountable for teaching it. Schools and districts rated “academically acceptable” provide what we have referred to as an accredited education, and we have presumed, simply in deference to the Legislature, that such an education achieves a general diffusion of knowledge.224 The district court found that the plaintiffs and intervenors have rebutted this presumption. The court’s principal reasons, set out in detailed findings and conclusions, may be summarized as follows:

• TAKS tests (and other such tests) cover only a small part of the prescribed curriculum;

• the cut scores and passing rates for TAKS tests (or other such tests) are too low and are set, not to reliably measure achievement, but to ensure a low rate of failure;

• completion and dropout rates are understated and unreliable, in fact fewer than 75% of all students and 70% of minority students complete high school, and this high attrition, worse in larger districts, is unacceptable;

• other important factors in determining whether a general diffusion of knowledge has been achieved, like college *788preparedness of graduates, for example, are not considered in rating schools and districts “academically acceptable” and reflect unfavorably on the system;

• the requirements for an “academically acceptable” rating are set to assure, not that there will be a general diffusion of knowledge, but that almost every district will meet them;

• the prescribed curriculum and TAKS testing have been made more demanding while funding to satisfy statutory requirements has not kept pace, producing budget pressures that have resulted in—

• a shortage of qualified teachers, an increase in teachers having to teach outside their fields, and high attrition and turnover rates;

• difficulty in providing special programs and remediation for students at risk of not completing their education;

• there has also been a lack of funding to meet increased federal requirements, like the No Child Left Behind Act;225

• the changing demographics of the student population — with a majority being economically disadvantaged, 15% having limited proficiency in English, and both groups continuing to grow— have increased education costs while funding has lagged;

• the I/R econometric study correctly shows that the cost of an accredited education exceeds available per-student revenue.

The State defendants contend that the district court focused too much on “inputs” to the public education system — that is, available resources. They argue that whether a general diffusion of knowledge has been accomplished depends entirely on “outputs” — the results of the educational process measured in student achievement. We agree that the constitutional standard is plainly result-oriented. It creates no duty to fund public education at any level other than what is required to achieve a general diffusion of knowledge. While the end-product of public education is related to the resources available for its use, the relationship is neither simple nor direct; public education can and often does improve with greater resources, just as it struggles when resources are withheld, but more money does not guarantee better schools or more educated students. To determine whether the system as a whole is providing for a general diffusion of knowledge, it is useful to consider how funding levels and mechanisms relate to better-educated students. This, we think, is all the district court did.

The State defendants also contend that the district court equated, erroneously, statutory expressions of the Legislature’s aspirational goals and “mission statements” with the constitutional standard. As we read the district court’s findings and conclusions, however, we think the court did no more than try to draw from statutory language the Legislature’s understanding of a general diffusion of knowledge. In section 4.001(a) of the Education Code, for example, the Legislature has expressly linked the stated mission of public education to the constitutional standard:

The mission of the public education system of this state is to ensure that all Texas children have access to a quality education that enables them to achieve their potential and fully participate now and in the future in the social, economic, and educational opportunities of our state and nation. That mission is grounded on the conviction that a gener*789al diffusion of knowledge is essential for the welfare of this state and for the preservation of the liberties and rights of citizens.226

In section 28.001, the Legislature has labeled specific knowledge and skills “essential”, just as a general diffusion of knowledge is:

It is the intent of the legislature that the essential knowledge and skills developed by the State Board of Education under this subchapter shall require all students to demonstrate the knowledge and skills necessary to read, write, compute, problem solve, think critically, apply technology, and communicate across all subject areas. The essential knowledge and skills shall also prepare and enable all students to continue to learn in postsecondary educational, training, or employment settings.227

These clear, affirmative statements cannot be dismissed as merely hopeful rhetoric; rather, the Legislature must be presumed to have chosen its words deliberately. Nor can these words be read to describe a public education system that the Legislature believes would not only meet but exceed constitutional requirements. The specific reference to the constitutional standard in section 4.001(a) and the repeated use of the word “essential” in section 28.001 does not allow it. To avoid improper policy-making of its own, the district court properly looked to legislative policy statements.

But while we think these statutory provisions properly inform the construction and application of the constitutional standard of a general diffusion of knowledge as understood by the Legislature, they cannot be used to fault a public education system that is working to meet their stated goals merely because it has not yet succeeded in doing so. The district court did not find that the system is so designed that it cannot accomplish a general diffusion of knowledge as defined by the statutory provisions just quoted. Rather, the district court found that the system is not producing a general diffusion of knowledge because the State has not provided sufficient funding.

In the extensive record before us, there is much evidence, which the district court credited, that many schools and districts are struggling to teach an increasingly demanding curriculum to a population with a growing number of disadvantaged students, yet without additional funding needed to meet these challenges. There are wide gaps in performance among student groups differentiated by race, proficiency in English, and economic advantage. Non-completion and dropout rates are high, and the loss of students who are struggling may make performance measures applied to those who continue appear better than they should. The rate of students meeting college preparedness standards is very low. There is also evidence of high attrition and turnover among teachers statewide, due to increasing demands and stagnant compensation. But the undisputed evidence is that standardized test scores have steadily improved over time, even while tests and curriculum have been made more difficult. By all admission, NAEP scores, which the district court did not mention, show that public education in Texas has improved relative to the other states. Having carefully reviewed the evidence and the district court’s findings, we cannot conclude that the Legislature has acted arbitrarily in structuring and funding the *790public education system so that school districts are not reasonably able to afford all students the access to education and the educational opportunity to accomplish a general diffusion of knowledge.

We recognize that the standard of arbitrariness we have applied is very deferential to the Legislature, but as we have explained, we believe that standard is what the Constitution requires. Nevertheless, the standard can be violated. There is substantial evidence, which again the district court credited, that the public education system has reached the point where continued improvement will not be possible absent significant change, whether that change take the form of increased funding, improved efficiencies, or better methods of education. Former Lieutenant Governor Ratliff, the author and principal sponsor of Senate Bill 7 in 1993, echoed the considered judgments of other witnesses at trial when he testified:

I am convinced that, just by my knowledge of the overall situation in Texas, school districts are virtually at the end of their resources, and to continue to raise the standards ... is reaching a situation where we’re asking people to make bricks without straw.228

But an impending constitutional violation is not an existing one, and it remains to be seen whether the system’s predicted drift toward constitutional inadequacy will be avoided by legislative reaction to widespread calls for changes.

C

The district court concluded that the public school funding system is inefficient in violation of article VII, section 1, but only in its provision of facilities for districts, not as the intervenors also claim, in its provision for the maintenance and operation of the schools. In Edgewood IV, we stated:

An efficient system of public education requires not only classroom instruction, but also the classrooms where that instruction is to take place. These components of an efficient system — instruction and facilities — are inseparable.229

By this we meant not only that the constitutional requirement of efficiency applies to both instruction and facilities, but also that the requirement must be applied to the two components together. Article VII, section 1 requires “an efficient system of free public schools”,230 considering the system as a whole, not a system with efficient components. Accordingly, we consider (i) whether the inefficient provision of facilities found by the district court makes the entire system inefficient, or (ii) whether, as the intervenors contend, the system is inefficient for reasons apart from its provision of facilities.

For the system to be efficient, “ ‘districts must have substantially equal access to. similar revenues per pupil at similar levels of tax effort.” ’231 In the earlier appeal in this case, we explained:

Because constitutional efficiency does not require absolute equality of spending, we expressly acknowledged [in Edgewood I] that “local communities would [not] be precluded from supplementing an efficient system established by the legislature”, but we added that “any local enrichment must derive solely *791from local tax effort.” In other words, the constitutional standard of efficiency requires substantially equivalent access to revenue only up to a point, after which a local community can elect higher taxes to “supplement” and “enrich” its own schools. That point, of course, although we did not expressly say so in Edgewood I, is the achievement of an adequate school system as required by the Constitution. Once the Legislature has discharged its duty to provide an adequate school system for the State, a local district is free to provide enhanced public education opportunities if its residents vote to tax themselves at higher levels. The requirement of efficiency does not preclude local supplementation of schools.232

The State defendants note that only one of the Alvarado ISD intervenors and none of the Edgewood ISD intervenors is rated “academically unacceptable”, and a third of the former and two of the latter have the higher “recognized” rating. Therefore, the State defendants argue, by providing an accredited education, the intervenors are providing for a general diffusion of knowledge, and “[a]bsent gross funding disparities akin to those invalidated by Edgewood I, a system that provides a general diffusion of knowledge is also presumptively efficient as a matter of law .... Because such enormous disparities do not exist today, the Court need not examine the system’s funding in detail .... ” The intervenors argue that significant funding disparities do exist and have worsened in the decade since we decided Edgewood IV, when we held that the system was “minimally acceptable only when viewed through the prism of history.”233 The intervenors point to the following:

• although 95% of public school funds are equalized through the FSP, 5% are not;

• the State tries to equalize revenue in districts with at least 85% of the student population, but it barely achieves this goal, while the Court in Edgewood II rejected a system that equalized revenue in districts with 95% of the student population;

• gaps in the available M & O revenue per student are worse than they were in Edgewood IV;

• the additional tax rate that poor districts must have to generate the same per-student revenue as wealthy districts has grown from $0.09 in Edgewood IV to $0.17.

The State defendants counter, in reverse order: that the tax-rate difference calculations of $0.09 in Edgewood IV versus $0.17 in this case are not comparable; that the per-student revenue gap in Edgewood IV would have been comparable to the gap today if “hold harmless” districts had not been excluded from the calculations; that the districts with 5% of the student population that were excluded from the funding system in Edgewood II resulted in far worse disparities; 234 and that non-equalization of only 5% of funding is not signifi*792cant. In all, the State defendants argue, the funding gaps cited by the intervenors are no worse than they were in Edgewood IV and do not render the entire system inefficient.

Given the closeness of the decision in Edgewood IV, the Court might well have reached a different conclusion had the “hold harmless” districts been presented as a permanent part of the system architecture. Now, however, it appears that the provisions favoring those districts reduce recaptured funds only about 4%. Although neither the State defendants nor the plaintiffs has made any effort to justify continuation of these districts, we cannot say that they render the entire system inefficient.

There is much evidence that many districts’ facilities are inadequate, but it is undisputed that some 25% of the districts levy no I & S taxes. The State defendants argue that disparities among districts in available facilities are not proof of inefficiency absent evidence that the districts’ needs are similar. They contend that facilities needs vary widely depending on the size and location of schools, construction expenses, and other variables. We agree that such evidence is necessary and lacking. The State defendants also argue that to prove constitutional inefficiency the in-tervenors must offer evidence of an inability to provide for a general diffusion of knowledge without additional facilities, and that they have failed to do so. Again, we agree. Efficiency requires only substantially equal access to revenue for facilities necessary for an adequate system.

The intervenors argue that constitutional efficiency does not permit substantially unequal access to funds to supplement an adequate education, but we have previously rejected this argument. In the earlier appeal in this case, we explained:

As long as efficiency is maintained, it is not unconstitutional for districts to supplement their programs with local funds, even if such funds are unmatched by state dollars and even if such funds are not subject to statewide recapture. We caution, however, that the amount of “supplementation” in the system cannot become so great that it, in effect, destroys the efficiency of the entire system. The danger is that what the Legislature today considers to be “supplementation” may tomorrow become necessary to satisfy the constitutional mandate for a general diffusion of knowledge.235 Supplementation must be just that: additional revenue not required for an education that is constitutionally adequate. For such supplementation we have never held that districts must have substantially equal access to funds.

Accordingly, we conclude that the public school finance system is not inefficient in violation of article VII, section 1.

The dissent concludes that the public school system is inefficient because it is not competitive.236 The dissent complains that this case “once again focuses on short-term funding rather than long-term solutions”237 and asserts that the Court has applied constitutional efficiency to mean only “ ‘equal ability to raise taxes.” ’238 With respect to the meaning of “efficiency”, the dissent is simply wrong. In Edgewood I, we said that “ ‘[efficient’ conveys the meaning of effective or productive of *793results and connotes the use of resources so as to produce results with little waste”.239 Efficiency implicates funding access issues, but it is certainly not limited to those issues. It is true that the plaintiffs and intervenors here have focused on funding, but parties to a lawsuit are entitled to choose the issues to be raised. We cannot dictate how the parties present their case or reject their contentions simply because we would prefer to address others. Perhaps, as the dissent contends, public education could benefit from more competition, but the parties have not raised this argument, and therefore we do not address it. The dissent calls a court that limits itself to the issues raised by the parties “a pushover”.240 We disagree. Here, even the State defendants do not criticize the plaintiffs and intervenors for failing to broaden their arguments to include other aspects of the public education system besides funding.

In Edgewood III, we explained that—

although the issues brought before us in Edgewood I, Edgewood II, and now Edgewood III, have all been limited to the financing of the public schools, as opposed to other aspects of their operation, money is not the only issue, nor is more money the only solution.... In Edgewood I we stated: “More money allocated under the present system would reduce some of the existing disparities between districts but would at best only postpone the reform that is necessary to make the system efficient.” 777 S.W.2d at 397. We are constrained by the arguments raised by the parties to address only issues of school finance. We have not been called upon to consider, for example, the improvements in education which could be realized by eliminating gross wastes in the bureaucratic administration of the system. The Legislature is not so restricted.241

The Legislature may well find many ways of improving the efficiency and adequacy of public education — ways not urged by the parties to this case — that do not involve increased funding.

D

The district court concluded that the public education system is not “suitable” as required by article VII, section 1 242 for the same reason it concluded that the system is inadequate and inefficient, that is, because the funding is insufficient. Neither the court nor the parties have differentiated suitability from the constitutional standards of adequacy and efficiency, but the requirement of suitability is not merely redundant of the other two. Rather, it refers specifically to the means chosen to achieve an adequate education through an efficient system. For example, we indicated in our prior opinion in this case that if the funding system were efficient so that districts had substantially equal access to it, and the education system was adequate to provide for a general diffusion of knowledge, but districts were not actually required to provide an adequate education, “the Legislature’s use of districts to discharge its constitutional duty would not be suitable, since the Legislature would have employed a means that need not achieve its end.” 243 In Edgewood IV, the property-rich districts argued that the State’s heavy reliance on local tax rev*794enue was unsuitable.244 We rejected the argument, not because it misinterpreted the standard, but because the reliance on local revenue does not prevent the system from providing a general diffusion of knowledge:

Certainly, if the Legislature substantially defaulted on its responsibility such that Texas school children were denied access to that education needed to participate fully in the social, economic, and educational opportunities available in Texas, the “suitable provision” clause would be violated. The present record, however, does not reflect any such abdication.245

Neither the structure nor the operation of the funding system prevents it from efficiently accomplishing a general diffusion of knowledge. The State may discharge its duty to make suitable provision for free public schools through school districts by relying on local tax revenues, even as heavily as it now does. Such reliance, especially given the multitude and diversity of school districts, inevitably makes it difficult to achieve efficiency because of the vast disparities in local property wealth, but efficiency is not impossible. We have suggested that these difficulties might be avoided by fundamental changes in the structure of the system, but the possibility of improvement does not render the present system unsuitable for adequately and efficiently providing a public education. Accordingly, we conclude that the system does not violate the constitutional requirement of suitability.

IV

The final constitutional question is whether the State’s control of local taxation for education amounts to a state property tax in violation of article VIII, section 1-e. We agree with the district court that it does.

As we have set out above, local tax rates have increased markedly since 1993-1994. Then, only 2% of the districts, with 1% of the students, were taxing at the $1.50 maximum M & 0 rate; now, 48% of the districts, with 59% of the students, are taxing at the cap, and 67% of the districts, with 81% of the students, are taxing at or above $1.45. In 1993-1994, 90% of the districts, with 85% of the students, had tax rates below $1.40; that group has now shrunk to 20% of the districts, with 10% of the students. The State defendants acknowledge this shift but argue that school districts tax at or near maximum rates in order to generate revenue for local' supplementation and discretionary purposes, not because State requirements for an accredited education force them to do so. The State defendants point to instances in which school districts:

• have made budget cuts without losing accreditation, demonstrating that not all the revenue generated at maximum tax rates is necessary to provide an accredited education;

• have provided educational programs not required for an accredited education;

• maintain an optimum fund balance — a reserve of funds — for contingencies;

• have chosen to raise teacher salaries above the state-mandated minimum; and

• have voluntarily increased homestead exemptions.246

The State defendants argue that because school districts exercise some discretion in *795taxing and spending for education, the plaintiffs’ claim that local taxes have become a state property tax is disproved as a matter of law.

We held in Edgewood III that “[a]n ad valorem tax is a state tax ... when the State so completely controls the levy, assessment and disbursement of revenue, either directly or indirectly, that the authority employed is without meaningful discretion.” 247 In that case, the State’s control of county education district taxation was direct and absolute, but we did not limit our holding to that situation. We stated that there was “a spectrum of other possibilities” along which varying degrees of state control might fall, and we rejected the argument that the State could circumvent the constitutional prohibition of a state property tax merely by slight variations in the degree of that control:

Each case must necessarily turn on its own particulars. Although parsing the differences may be likened to dancing on the head of a pin, it is the Legislature which has created the pin, summoned the dancers, and called the tune. The Legislature can avoid these constitutional conundra by choosing another path altogether.248

While the Legislature did abandon the system of county educational districts after Edgewood III, the system adopted by Senate Bill 7 allows the possibility of the same constitutional violation by setting maximum tax rates. We explained in Edgewood IV:

However, if the cost of providing for a general diffusion of knowledge continues to rise, as it surely will, the minimum rate at which a district must tax will also rise. Eventually, some districts may be forced to tax at the maximum allowable rate just to provide a general diffusion of knowledge. If a cap on tax rates were to become in effect a floor as well as a ceiling, the conclusion that the Legislature had set a statewide ad valorem tax would appear to be unavoidable because the districts would then have lost all meaningful discretion in setting the tax rate.249

The State defendants in this case have taken a rigid view of this admonition, arguing that there is no constitutional violation unless all school districts, or at least most of them, are required to tax at the absolute maximum rate for no other reason than to provide an accredited education. In our earlier opinion in this case, we rejected the argument that impermissible state control depended on the number of districts affected. “The concern,” we said, “is not the pervasiveness of the tax but the State’s control of it.” 250 We also rejected the argument that districts must be forced absolutely to the limit of the cap.251 The issue, we emphasized, is the lack of meaningful discretion.252

The dissent takes a position similar to the State defendants’, arguing that there can be no state ad valorem tax unless districts are absolutely forced to tax at maximum rates, and that “each and every district must prove it had no other choice.” 253 This misstates the test. If a district were absolutely forced to tax at the maximum rate, it would lack not just *796meaningful discretion, it would lack any discretion whatsoever. Moreover, the dissent’s requirement that a district be left with no choices ignores the realities of the education process. The State does not mandate the expenditures necessary for accreditation; rather, it leaves largely to school districts the decisions on how best to expend education funds to achieve accreditation. Those decisions involve professional judgment and experience, and the methods of meeting accreditation standards vary depending on student demographics and school location. It is simply impossible to trace the impact on accreditation of each dollar spent for programs and teacher salaries. Recognizing these realities, we observed in Edgewood III that State influence on district taxing and spending cannot be measured exactly but must be gauged along a spectrum of possibilities.254

Meaningful discretion cannot be quantified; it is an admittedly imprecise standard. But we think its application in this case is not a close question. The district court found that the plaintiffs’ “focus districts” for which evidence was offered “lack ‘meaningful discretion’ in setting their local property tax rates.” Contrary to the dissent’s assertion, this finding was supported by evidence other than conclu-sory opinions of district superintendents. The district court detailed evidence showing how the districts are struggling to maintain accreditation with increasing standards, a demographically diverse and changing student population, and fewer qualified teachers, while cutting budgets even further. The district court found that due to inadequate funding: 52.8% of the newly hired teachers in 2002 were not certified, up from 14.1% in 1996; more teachers were being required to teach outside their areas of expertise; and attrition and turnover were growing. The court cited the higher costs of educating economically disadvantaged students and students with limited English proficiency, noting that 90% of the growth in the student population has come from low-income families. And as set out in more detail above, the district court noted the increased curriculum, testing, and accreditation standards, and the increased costs of meeting them. These are facts, not opinions. The State defendants point to evidence of some discretionary spending on programs not essential to accreditation, but there is also evidence that such programs are important to keeping students in school.

The State defendants point out that though facing increased challenges, the focus districts have met or exceeded all accreditation requirements, but importantly, one cannot infer from that fact that the districts could lower taxes and still meet those requirements. The district court credited evidence that districts statewide are spending over 97% of the revenue that would be available if every district taxed at maximum rates, up from 83% in 1993-1994. Only about a third of the districts with about a fifth of the student population exceed minimum accreditation standards.255 This is a marked decline from 2001, when over 60% of the districts with well over half of the student population exceeded minimum accreditation standards.256 The current situation has be*797come virtually indistinguishable from one in which the State simply set an ad valo-rem tax rate of $1.50 and redistributed the revenue to the districts.

The State also controls the expenditure of more than $1 billion in local tax revenues recaptured from 134 districts, which educate 12.3% of the students, requiring that they be effectively redistributed to the other districts. The number of districts and amount of revenue subject to recap-tee have almost tripled since 1994. The State’s control of this local revenue is a significant factor in considering whether local taxes have become a state property tax.

The dissent argues that the plaintiffs cannot prove that local ad valorem taxes have become a state property tax with evidence that most districts now tax at maximum rates when few did ten years ago, or that virtually all of the revenue available through local taxes is now being spent, or that among school districts at maximum tax rates accreditation rates have declined, or that the State controls the redistribution of more than $1 billion in local taxes.257 Even if each category of evidence would not, by itself, prove a constitutional violation, all of this evidence taken together, along with the extensive record before us, clearly shows that school districts have lost meaningful discretion to tax below maximum rates and still provide an accredited education. In reaching this conclusion, we do not alter any standard we have previously announced, as the dissent charges, or adopt positions the Court has previously rejected, as the dissent suggests. The question, as we stated in Edge-wood III, is whether school districts have meaningful discretion to tax below maximum rates, and the answer is that they do not.

The district court also determined that the maximum tax rate for purposes of this analysis should be $1.35 rather than $1.50 because school districts must have $0.15 of tax rate — 10% of the maximum — available for local supplementation. Thus, in the district court’s view, almost all school districts are taxing at maximum rates. The State defendants object that districts have no constitutional right to local supplementation, and therefore such expenditures should not be considered in determining whether school districts have meaningful discretion to tax below maximum rates. We agree that local supplementation is not a constitutional right, but it is part of the purpose of FSP funding. Section 42.301 of the Education Code states in part: “The purpose of the guaranteed yield component [Tier II] of the Foundation School Program is to provide each school district with the opportunity to provide the basic program and to supplement that program at a level of its own choice.” 258 Although the statute does not promise any particular level of supplemental funding, local supplementation is made a core component of the system structure, necessitated by the basic philosophy of the virtue of local control. The State cannot provide for local supplementation, pressure most of the districts by increasing accreditation standards in an environment of increasing costs to tax at maximum rates in order to afford any supplementation at all, and then argue that it is not controlling local tax rates.

Accordingly, we conclude that the public school finance system violates article VIII, section 1-e of the Texas Constitution. Various legislative proposals during the past year to remedy perceived problems *798with the public education system and its funding would reduce the maximum ad valorem tax rate and allow it to be exceeded for certain purposes. While we express no view on the appropriateness of any of these proposals, we are constrained to caution, as we have before, that a cap to which districts are inexorably forced by educational requirements and economic necessities, as they have been under Senate Bill 7, will in short order violate the prohibition of a state property tax.

V

We come at last to the issue of the relief to be granted. The dissent argues, although the State defendants do not, that the district court’s injunction was over-broad and is not warranted by our holding. The dissent argues that only the statute which caps ad valorem tax rates at $1.50, section 45.003(d) of the Education Code, should be enjoined, and only in those districts that are forced by accreditation requirements to tax at maximum rates.259 The dissent contends that the district court’s injunction impacts districts where no constitutional violation has been shown and “looks too much like ‘enjoin now and worry later”’.260

It is worth repeating that the dissent again raises an argument the State defendants have not made themselves. The dissent ignores the central role of the $1.50 cap in the public education finance system. The FSP guarantees state funding to property-poor districts only up to a maximum tax rate of $1.50.261 Thus, a property-poor district allowed to tax above the maximum rate would not be entitled to FSP Tier 2 guaranteed yield funding at a higher rate. Property-rich districts allowed to tax above the maximum rate would be entitled to keep all of the revenue generated. If property-poor districts needed no additional revenue to provide an adequate education, there would be no constitutional violation.262 But for districts that need additional revenue, the funding system would be inefficient. The equalization necessary for efficiency that the combination of the FSP, the tax rate cap, and recapture is intended to effectuate would be destroyed if the cap were removed. This would create a structural flaw in the system itself, just as if the basic allotment and guaranteed yield of the FSP were significantly reduced.

The tax rate cap that makes the public education funding system a state property tax is also intended to keep the system efficient. The two roles of the cap are inseparable. To remove the cap so as to allow districts meaningful discretion in setting tax rates at higher levels would be to increase the revenue disparity among the property-rich and the property-poor districts, creating the financial inefficiency that the cap is intended to prevent. Local ad valorem taxes, which we have determined to be a prohibited state property tax, provide more than half the revenue for the public school system. The constitutional violation cannot be corrected without *799raising the cap on local tax rates or changing the system.

The Constitution does not require a particular solution. We leave such matters to the discretion of the Legislature. To end the constitutional violation, we agree with the district court that the use of the current system must be enjoined. The district court delayed the effect of its injunction until October 1, 2005, to allow the Legislature time to respond. Since the injunction issued, the Legislature has undertaken to respond in a regular session and two special sessions. Its inability to do so appears to be due not to any lack of expertise in the issues but to the absence of agreement. At this point in time, it is unlikely that material changes could be made in the public education system that would affect the current school year. School districts will next begin to prepare budgets and set tax rates in the summer of 2006.263 To allow the State ample time to fully consider structural changes in the public education system, and to allow the system time to adjust to those changes, we postpone the effective date of the district court’s injunction to June 1, 2006.

VI

The district court awarded the plaintiffs and intervenors attorney fees under the Declaratory Judgments Act.264 As we have said:

[T]he Declaratory Judgments Act entrusts attorney fee awards to the trial court’s sound discretion, subject to the requirements that any fees awarded be reasonable and necessary, which are matters of fact, and to the additional requirements that fees be equitable and just, which are matters of law.265

Because we have concluded that the plaintiffs are entitled to only a part of the relief granted by the district court, and that the intervenors are entitled to no relief, we reverse the award of attorney fees and remand the ease to the district court to reconsider what award of attorney fees, if any, is appropriate. We express no opinion on this issue.

[[Image here]]

More than half a century ago, in Brown v. Board of Education, the United States Supreme Court wrote:

Today, education is perhaps the most important function of state and local governments. Compulsory school attendance laws and the great expenditures for education both demonstrate our recognition of the importance of education to our democratic society. It is required in the performance of our most basic public responsibilities, even service in the armed forces. It is the very foundation of good citizenship. Today it is a principal instrument in awakening the child to cultural values, in preparing him for later professional training, and in helping him to adjust normally to his environment. In these days, it is doubtful that any child may reasonably be expected to succeed in life if he is denied the opportunity of an education. Such an opportunity, where the state has undertaken to provide it, is a right which *800must be made available to all on equal terms.266

Since then, especially in this Information Age, education as a fundamental basis for our future has grown by orders of magnitude.

More than thirty years ago, in San Antonio Independent School District v. Rodriguez, the first case to challenge the constitutionality of the public school finance system in Texas, the United States Supreme Court stated:

The need is apparent for reform in tax systems which may well have relied too long and too heavily on the local property tax. And certainly innovative thinking as to public education, its methods, and its funding is necessary to assure both a higher level of quality and greater uniformity of opportunity. These matters merit the continued attention of the scholars who already have contributed much by their challenges. But the ultimate solutions must come from the lawmakers and from the democratic pressures of those who elect them.267

As we have said since Edgewood I, structural changes, and not merely increased funding, are needed in the public education system to meet the constitutional challenges that have been raised.

The judgment of the district court is reversed insofar as it declares a violation of article VII, section 1, and awards attorney fees and costs, and the issue of attorney fees and costs is remanded to the trial court. The effective date of the injunction is modified to June 1, 2006. In all other respects, the judgment is affirmed in accordance with this opinion.

So ordered.

dissenting.

In the name of “efficiency,” several school districts again ask the Texas courts to close the Texas public schools unless the Texas Legislature increases funding. Over the last two decades, we have been asked to do this every two or three years, and have generally complied.

The Court goes too far by doing so again today. First, the Court finds school districts are forced to tax at the highest possible rate only because some of them do. Second, though only five percent of the State’s school districts claim a single statute is unconstitutional, the Court enjoins the State from distributing any money under the current Texas school financing system, an order that applies to every school district in Texas. Thus, because some districts get too little state money, "all districts may get none. It is hard to see how this will help Texas school children.

Yet the Court also does not go far enough. By failing to demand an “efficient system” as the Texas Constitution requires, or to demand standing and proof as Texas law requires, this case once again focuses on short-term funding rather than long-term solutions.

Of course, the true goal of this litigation is to put pressure on the Texas Legislature. We demanded legislative changes by holding the Texas school-finance system unconstitutional in Edgewood I,1 Edgewood *801II,2 and Edgewood III;3 we warned that we might do so again soon in Edgewood IV4 and West Orange-Cove I.5 The Court fulfills that threat today. But there is no end in sight; if the past is any indication, the new funding will not last long, and public education will not change much.

Before we bequeath Edgewood VIII, IX, and X to our grandchildren, we should consider whether we might do more by doing less. As the Court fails to do so today, I respectfully dissent.

I. The Constitution & Efficiency

Since statehood in 1845, every Texas Constitution has required the Legislature to “make suitable provision for the support and maintenance of public schools.”6 But when Texans adopted the current Constitution in 1876, they added a new word— the Constitution now requires “suitable provision for the support and maintenance of an efficient system of public free schools.”7

Were we drafting a constitution today, we might choose a different standard— perhaps an “exemplary” or “comprehensive” or “progressive” or “safe” system of public schools. But in 1876, the people of Texas adopted “efficient” as the constitutional standard, and until that Constitution is amended no court can adopt any other.

When this Court issued Edgewood I in October 1989, we recognized that an “efficient” system would “produce results with little waste.”8 Nevertheless, we have applied the term in every case since then to require only one thing — “substantially equal access to similar revenues per pupil at similar levels of tax effort.”9 In other words, “efficient” has meant only “equal ability to raise taxes.”

Perhaps this made sense in 1989 — before the Berlin Wall fell, before the Soviet Union collapsed, and before state-run businesses everywhere proved uncompetitive. Perhaps back then a government system was “efficient” if it could get sufficient public funding.

*802But surely not now. Today, we know that one thing above all else makes service providers efficient: competition. Even formerly communist countries recognize how efficiency is produced — not by protectionism, not by higher taxes, and not by state control, but by freedom for competition.

Yet the school districts that brought this case never once suggested in six-weeks’ evidence that competition might make the Texas school system more efficient. No one considered fundamental reforms that efficiency might demand. No school expert considered whether it might be efficient to consolidate tiny school districts or redundant school administrations. No one asked whether it might be efficient to transfer students across district lines, or transfer funds to private providers that could meet their needs better. Instead, this trial focused entirely on getting more state funding through more taxes — all else in the system to remain exactly the same.

This, of course, is perfectly natural. Few of us welcome competition, not even judges.10 Competition is often painful, and requires us to make hard choices we would rather avoid.

But long-standing rules of Texas law do not allow us to wink at these omissions here. First, because Article YII’s education guarantee is a right that belongs to school children rather than school districts, the latter have no standing to assert this claim. Every party in this case was a school district, and every witness in the six-week trial was a school employee or school expert. Not a single attorney represented solely the interests of school students and their families — who might actually favor the broader educational options or lower taxes competition might bring. By overlooking standing, this trial focused too much on the priorities of school districts, and not enough on the priorities of school families.

Second, because Article YIII’s constitutional prohibition of state property taxes is violated only if a school district must tax at the statutory maximum, each district had to prove it was forced to do so. The 47 plaintiff districts alone asserted this, but none proved it. No school district addressed, no expert studied, and none of the trial judge’s 679 findings mentioned why districts were “forced” to make expenditures that other public and private schools often forego, or that other government entities often provide. Nor did anyone consider whether competition or other fundamental reforms might make the system more efficient so that less money was necessary. By lowering the burden of proof, this trial focused on whether school expenditures were reasonable rather than required.

My colleagues say our review of “efficiency” must be limited to funding because “[w]e cannot dictate how the parties present their case.”11 This Court is not usually such a pushover. When we interpret contracts, statutes, and (above all) constitutions, we are constrained by what they say, not the parties’ briefs. The constitutional guarantee invoked here requires an efficient system of public schools; it cannot be used to demand more funding for an inefficient system.

Nor can we avoid our duty by suggesting that the Legislature demand efficiency *803when we will not.12 If efficiency is a justi-ciable question (as the Court holds), then we cannot simply suggest that someone else look into it.

The author of the current school-finance system testified at trial that school districts “were no more wasteful or inefficient than any other State agency or State institution.” But that is not the constitutional standard. For whatever reason, the Texas Constitution mandates efficiency primarily in the State’s courts13 and schools;14 they must meet a higher standard because that is what the Constitution requires. If “efficiency” truly means “producing results with little waste,” then someday we ought to apply it to that purpose.

II. Article VII & Standing

A general diffusion of knowledge being essential to the preservation of the liberties and rights of the people, it shall be the duty of the Legislature of the State to establish and make suitable provision for the support and maintenance of an efficient system of public free schools.

Texas Constitution, Article VII, § 1

While acknowledging evidence that the public school finance system is inadequate, unsuitable, and inefficient, the Court nevertheless finds no violation of Article VII because “an impending constitutional violation is not an existing one.”15 We have tried this before, accepting the current system while lamenting it, and warning that the result might be different next time.16

But this is the first time we have entertained such complaints in a courtroom with no students. While standing normally requires only an allegation of injury, a two-part test governs standing to challenge the constitutionality of a statute: (1) an allegation of actual or threatened injury under the statute, and (2) an allegation that the statute unconstitutionally restricts the plaintiffs own rights.17 As all concede, the public-education guarantee in Article VII of the Texas Constitution is a right that belongs to school students, not school districts. Yet only the latter were represented at trial, and as the trial made clear, the interests of the two are not necessarily the same.

Standing is required by two guarantees in the Texas Constitution — separation of powers18 and open courts.19 We should not violate these two constitutional provisions in order to decide whether the State violated two others.

*804A. A Question We Have Never Addressed

This is the first Article VII school-finance case brought solely by school districts, without a single family or school student as plaintiff.

In Edgewood I, 68 school districts and “numerous individual school children and parents” filed suit.20 Edgewood II involved subsequent proceedings in the same suit with the same parties.21 Edgewood III was brought by “numerous school districts and individual citizens.”22 Edgewood IV was filed by “hundreds of school districts ... as well as many parents and local officials.”23

None of these cases approved school-district standing under Article VII. Nor did they approve such standing implicitly, as standing cannot be waived and may be raised during any later appeal.24

To the contrary, in Edgewood IV, we held that section 3 of Article VII granted no constitutional rights to school districts:

Article VII, section 3 does not create any “rights.” It only authorizes the Legislature to establish school districts and to empower the districts to levy taxes for specific purposes. The school districts’ rights, to the extent they exist, are derived solely from the statutes that the Legislature may enact under the authority granted in section 3.25

Similarly, section 1 of Article VII does not create any rights for school districts; in fact, it does 'not even mention them. To the extent school districts assert injury here, they cannot do so for any violation of this constitutional right.

While school districts participated in all our prior Article VII cases, their standing was immaterial because school families participated too. When several parties make the same claim for declaratory or injunctive relief, standing for some renders standing for the remainder immaterial.26 Federal law is to the same effect.27 As all our prior cases included parties whose sole interest was the education of their children, the State had nothing to gain by objecting to school-district standing, and the judgments would have been no different if it had.

There is certainly no “broad rule that a governmental entity cannot sue to declare a statute unconstitutional.”28 But there is no broad rule that they always have such standing either. Just because school districts have standing to bring some claims *805does not mean they have standing to bring all claims.

Instead, standing depends on the nature and source of the claim being made.29 While school districts have standing to pursue an Article VIII claim,30 that does not mean they have standing to pursue an Article VII claim. We have never suggested otherwise, until today.

B. Standing We Have Never Recognized

Before today, we have never held that government agencies have standing to sue the State for a bigger budget.

The school districts allege they have insufficient money to carry out their duties, but it is not money for their own account. As we held long ago, school districts hold money only as trustees for school students:

School funds are held to be trust funds for educational purposes. Such funds do not belong to the district or to the officers of the district, but are merely held by them in trust for the public.31

The injury alleged in this case was suffered only by school students: to the extent school districts must cut courses, or eliminate extracurriculars, or hire less-qualified teachers, it is the students who suffer the concrete, personal harm rather than the districts themselves.

The school districts alleged only that inadequate state funding limited their ability to perform their official duties. Both state and federal courts have rejected standing by government officials to bring such claims.32 Thus, we held in Brown v. Todd that a city councilman lacked standing to challenge a mayor’s personnel policy that did not apply to him, but merely infringed his ability to set such policies.33 Similarly, the United States Supreme Court recently held that grant recipients but not members of Congress had standing to challenge the Line Item Veto Act(though the Act granted standing to both), as the former actually lost money while the latter lost only their discretionary power to dispense it.34

This is not a case like Nootsie, Ltd. v. Williamson County Appraisal District, in which a public entity was compelled to affirmatively grant a tax exemption it believed unconstitutional.35 The districts do *806not complain that they are affirmatively compelled to perform unconstitutional teaching, testing, or any other services;they complain only that they are underfunded.

The Court’s suggestion that we have recognized standing before in these circumstances is indefensible. In Vondy v. Commissioners Court, we ordered commissioners to pay a constitutionally required salary when they had refused to pay any.36 In Mays v. Fifth Court of Appeals, we ordered commissioners to pay a statutorily allowed raise which they had ignored.37 Both cases involved nondiscre-tionary ministerial acts;38 neither involved a dispute between an agency and the State about whether the former’s budget was big enough.

The Court justifies standing here because “the Legislature has required school districts to achieve the goal of a general diffusion of knowledge.”39 But that gives them no rights against the State. As we noted in Edgewood IV, the State can abolish school districts completely, or enlarge or diminish their powers.40 Further, the Texas Constitution requires the Legislature to provide for many things—roads and bridges,41 the Legislative Redistricting Board,42 the Judicial Conduct Commission,43 and the salaries of thousands of public employees.44 These are all important items, and some may be underfunded; but surely all do not have standing to sue the State for more.

In every analysis of standing, “the plaintiff must contend that the statute unconstitutionally restricts the plaintiffs rights, not somebody else’s.”45 This the school districts cannot do.

C. Priorities We Have Never Approved

One reason courts require standing is amply demonstrated by the evidence in this trial, which tended toward a wish-list for school district employees.

Eight superintendents testified for the school districts at trial, each listing what they needed or what they would do if they had more money. Their priorities were almost identical: more bilingual teachers, more certified teachers, more certified librarians, more teacher training, higher salaries, better benefits, smaller classes, and longer school years.

Each of these may be important. But if eight families from the same districts had testified at trial, is this what they would have listed? Assuming all could not be fully funded, would they have listed them in the same order? We simply do not know.

We do know that, for most of us, our priority as employees is higher salaries, *807while our priority as customers is lower prices. Both may be possible when competition increases efficiency, innovation, and productivity. But at some point the two inevitably conflict, and some compromise is necessary. Because the trial here included only education providers and no education customers, the evidence may not accurately reflect where that line should be drawn.

Moreover, fundamental reforms may be overlooked if school districts may assert Article VII claims by themselves. Here, for example, not a single expert witness studied the possible savings that might accrue from consolidating some of the State’s 1,031 school districts. This Court has repeatedly lamented the “crazy-quilt pattern of small school districts,”46 as a result of which “duplicative administrative costs are unavoidable.”47 The plaintiffs’ experts confirmed that smaller districts have “the highest level of expenditures per student, as one would expect,” because of “diseconomies of scale.” Yet not a single school district or expert witness suggested any consolidations.48

It is unrealistic to ask school boards and administrators to recommend their own abolition, or lower salaries for themselves or any employees. Such potential conflicts between the interests of school districts and school families prevent the former from claiming standing to represent the latter. We have recognized representative standing in some circumstances,49 and sometimes state agencies may assert standing on behalf of their constituents.50 But we have done so only when the goals of a group and its members are so closely aligned that there is no reason to require participation by one in a suit by the other.51 That is not the case here.

In its final analysis, the Court dispenses with standing generally, because (1) students and families were free to intervene, and (2) the districts could find students and families to back their claims. Even if we assume that poor families can hire lawyers, or school districts can recruit sham plaintiffs to bolster their claims, it is hard to see what that has to do with the standing of the parties actually before us. More important, such arguments could be made by every party who lacks standing, including millions of taxpayers,52 or the father *808whose challenge to the Pledge of Allegiance was recently rejected for lack of standing.53 Normally, this Court strictly enforces standing so that we retain our proper role;54 hopefully today’s exception is good for this case only.

Standing is not a technicality; it is essential to any court’s authority to decide a case.55 We cannot abandon it in noteworthy cases; indeed, that is when adherence to legal standards is most important. As the United States Supreme Court recently noted, courts must be “especially rigorous” in requiring proper standing when asked to declare the actions of the other two branches of government unconstitutional.56 The school districts alone cannot meet such standards here.

III. Article VIII & Discretion

No State ad valorem taxes shall be levied upon any property within this State.

Texas Constitution, Article VIII, § 1-e

The 47 plaintiffs, mostly property-rich school districts, bring a claim that Article VIII, section 1-e of the Texas Constitution is violated by a tax-rate ceiling in a single subpart of a single statute.57 Unlike Article VII, Article VIII was intended to benefit school districts, and thus they have standing to assert this claim.58

In Edgewood III, we declined to adopt a precise test for violations of Article VIII because state control over property taxes presents “a spectrum of possibilities.”59 Instead, we held that a tax violates Article VIII if the State so completely controls the levy, assessment, and disbursement of revenue that school districts are “without meaningful discretion.”60 In Edgewood IV, we explained that districts lose such discretion when they are “forced to tax at the maximum allowable rate just to provide a general diffusion of knowledge.”61

This appeal turns on whether the plaintiffs proved they were “forced” to tax at the maximum rate. In reviewing the evidence, the Court contradicts everything we have said about such evidence before, and adds new “factors” we apparently overlooked before. This is too imprecise; a *809legal standard cannot turn on entirely different evidence from one case to the next.

A. The Wrong Standard: Everybody Else Does It

The Court points to several statewide trends as evidence of an Article VIII violation. But in our previous cases, we held that evidence just like this could not show an Article VIII violation.

First, my colleagues suggest that school districts are forced to tax at maximum rates because about half of them do. While we have never stated in detail what the Article VIII standard means, we have stated one thing it does not mean — “the number of districts taxing at maximum rates is not determinative.”62 In West Orange-Cove I, we expressly rejected arguments that an unconstitutional state property tax must control the rates in every district (the State’s position) or most districts (the trial court’s conclusion); instead, we held that an ad valorem tax is unconstitutional if it is imposed by the State, no matter how many districts it covers.63 If the State could not use prevailing tax rates to prove the school districts should lose, why can the school districts now use them to prove they should win?

Second, the Court reverses field by concluding that elose-to-maximum rates show that many districts lack meaningful discretion. Only two years ago, we said close counts neither way: “It may be that a school district taxing at $1.47 instead of $1.50 has exercised meaningful discretion, but that is not necessarily the case.”64 The number of districts taxing in this range simply cannot tell us whether “a single district ... is constrained by the State to tax at [this] particular rate.”65

Third, the Court finds it important that districts are taxing and spending 97 percent of the revenue that would be available if every district taxed at maximum rates.66 But in Edgewood IV we noted, and school district witnesses conceded at trial, that financial incentives in the current school-finance system encourage school districts to tax at maximum rates even if they don’t have to.67 The current system does not force districts to tax at maximum rates merely by providing incentives for them to do so.

Fourth, the Court announces today that substantial transfers of tax revenues from rich districts to poor districts are “a significant factor” in rendering the current system unconstitutional.68 Of course, we demanded something along these very lines when we required equalized funding in *810Edgewood I. Further, we held such transfers constitutional in Edgewood IV;69 today’s opinion appears to adopt the dissent in the latter case.70

Finally, the Court supports its constitutional conclusion by noting a “marked decline” since 2001 in the number of districts that “exceed minimum accreditation standards.” 71 We have never before tied constitutional analysis to testing or accreditation scores, and today’s reference shows why we should be reluctant to enter that hotly debated area. For example, if the base year in this trend were 1994 rather than 2001, then there has been a marked increase in the number of districts exceeding minimum standards. Further, as the standards themselves are rising, declining scores may or may not reflect actual declines. And the “minimum” standard referenced here is “academically acceptable”; nothing in this rating system proves the State is “forcing” every school district to rate above average.

Surely we were not mistaken in all our previous cases. If revenue transfers and accreditation scores were relevant to Article VUI’s standard, it is curious that we have never mentioned them before. And merely looking at average tax rates cannot tell us whether any district was “forced” to that level or arrived there via “meaningful discretion.”

Whether any school district in Texas has lost “meaningful discretion” is not a standard that can be proved by statewide trends. School districts are not forced to tax or spend money just because everyone else does it.72 The standards this Court has established require more specific evidence of a violation of Article VIII.

B. The Right Standard: What Must This District Do?

The school districts cannot establish a violation of Article VIII by proving that their current budgets are customary, or even reasonable; the tax cap they challenge is unconstitutional only if they proved they were forced to tax at that rate.

By definition, districts are not “forced” to make discretionary or voluntary expenditures. Of course, some expenditures may be mandatory de facto, even though not mandatory de jure.73 For example, Texas school boards or administrators who cut football programs or drill teams (as the State’s attorneys bravely suggest) may soon find themselves looking for other occupations.

*811But the Court adopts a standard far too low by holding that districts axe “forced” to tax at maximum rates whenever their “professional judgment and experience” suggests they should.74 Undoubtedly, school districts want to give their students the best education possible, and an educator’s professional judgment would deem anything less to be undesirable. But in Edgewood IV, we rejected a claim that districts were “forced” to transfer revenues “because the various alternatives are all undesirable.”75 By equating professional preferences with coercion, my colleagues again follow the dissent rather than the majority in Edgewood IV.76

The districts did offer examples of expenditures that were mandatory, and programs that were cut. But as proof that districts are forced to tax at maximum rates, both are non séquitos. Proving that some programs are mandatory does not prove that all others are too. Nor does it follow from cuts in one program that no further cuts can be made. To the contrary, the reluctance the superintendents expressed at trial about such cuts served to prove, if anything, their reluctance to cut any programs at all.

Moreover, the State’s trial evidence of discretionary spending did not focus on remedial-reading or bilingual-education programs. Instead, the State pointed to undisputed expenditures for swimming pools, nature trails, athletic stadiums, tennis courts, and unconventional classes such as broadcast journalism, ceramics, power lifting, ballet, film critique, lego robotics, advanced mariachi, and culinary arts.

It is true that several superintendents testified that all these programs were needed to keep students in school. But if we take these claims at face value then nothing schools spend is discretionary. “[A] claim will not stand or fall on the mere ipse dixit of a credentialed witness.” 77 These opinions alone cannot support the trial court’s judgment, both because they are conclusory,78 and because the question is a legal one.79 This Court is not usually so generous in treating such testimony as “facts, not opinions.”80

*812Further, none of the school districts explained why they were “forced” to maintain athletic facilities or library services that local governments often provide, or unconventional classes that might be available through local community colleges or the internet. No one would suggest that communities can run their fire, police, or utility departments through a school district’s budget, thus shifting those costs to the State or richer districts. The trial court could not simply assume there were no alternative providers; the school districts had to prove it.

Similarly, several superintendents conceded paying the highest starting salaries in their region, or special stipends to attract particular types of teachers. Considering the importance of what they do, no one can begrudge teachers higher salaries; but these contribute to a violation of Article VIII only if school districts had no choice. If surrounding public or private schools pay less, it was the districts’ burden to prove why they could not.

When pressed to explain such expenses, district witnesses repeatedly pointed to the demands of their local communities. But again, local demand must be proved, not merely asserted. As no students or families testified at trial, the only proof was the conclusory assurances of school administrators.

In a democracy, community demand is proved by elections, not anecdotal hearsay. In many instances, schools can buy property using school bonds (which require electoral approval) or the general operations budget (which does not). We cannot tell from this record which programs had been approved at an election, or what percentage of the community actually participated. Surely a district cannot avoid elections on expensive programs, or schedule them to ensure low voter turnout,81 and then claim they were forced to adopt those programs by their community.82 Without such proof here, we simply cannot tell.

Finally, because fundamental reforms were never considered, we do not know whether they might allow districts to drop rates below the tax ceiling. School districts cannot spend money inefficiently (subverting Article VII) to “force” themselves to the tax ceiling (subverting Article VIII), as these articles must be construed consistently to give effect to both.83 School districts may have good reasons to avoid consolidating, or starting school later in the year, or increasing class size so that teachers’ salaries could be increased too. But they are forced to make current expenses only if saving money through such alternatives was impossible, not just unpopular.

Of course, had the trial judge required specific evidence that the districts were forced to incur substantially all their current expenses, it would have been much more difficult for the districts to prove an Article VIII violation. But proving a stat*813ute unconstitutional is not supposed to be easy. We must presume the current system is constitutional, and interpret it whenever possible in a manner that renders it so.84 This presumption is “especially strong” when statutes relate to taxation,85 and “especially important” when we deal with politically charged subjects like the schools.86

There was plenty of evidence at trial that public schools are being asked to carry increasingly heavy burdens, burdens that private schools often do not bear. For example, as one superintendent noted, “it is not easy to remove employees in the public sector.” Accountability and testing systems have raised expectations that somehow all schools and school children can be at or above average. Teachers and administrators face the risk that the failure of their students will cause their own professional efforts to be labeled “academically unacceptable.” And as all the witnesses agreed, a growing stream of immigrants with little formal schooling or English proficiency requires that public schools not only leave no child behind, but go back at great expense and pick up more as soon as they arrive.87

Nevertheless, the Article VIII standard is not whether educational expenditures are reasonable, or important, or far-sighted, or what a community would prefer, but whether a district is forced to make them. Before the courts can declare the State’s school-finance system unconstitutional, each and every district must prove it had no other choice. Here, none did.

IV. Equity & Overbroad Relief

Permanent injunctions “must be narrowly drawn,”88 and “the record must contain evidence supporting each injunctive provision.” 89 This one meets neither standard.

It is neither true nor “worth repeating” that these standards can be ignored because the State asks for no injunction rather than a narrower one. A court must craft an equitable injunction even if it is *814not precisely what either party wants.90 If the rule were otherwise, the Court should not postpone the injunction here until June 2006 — as neither party asked for that. Hopefully, today’s rule is once again good for today’s case only.

A. Too Many Districts

First, there is no evidence to support a constitutional violation in every school district in Texas.

Out of 1,031 school districts in Texas, only 329 filed suit, only 47 asserted the single constitutional claim the Court affirms, only 9 presented proof on that claim in any detail, and only 3 called a witness to prove it at trial. On this narrow basis, the Court declares the school-finance system in every district unconstitutional, and enjoins state funding for them all. This is too broad.

As we recently noted, it has always been the law of equity that a permanent injunction “must not grant relief which is ... more comprehensive or restrictive than justified by the pleadings, the evidence, and the usages of equity.”91 Thus, for example, a permanent injunction against protests at five physicians’ homes is too broad if the evidence shows protests occurred at only four.92 Similarly, evidence of flies and foul odors from a 10-acre feedlot does not justify a permanent injunction extending to an entire 450-acre ranch.93 An injunction may extend as far as the evidence, but no further.

In their Article' VIII claim, the plaintiffs did not challenge the tax-rate cap facially,94 but only as it applied to them. “In an as-applied constitutional challenge, we must evaluate the statute as it operates in practice against the particular plaintiff.”95 Yet the trial court did not even try to evaluate how the property-tax cap operates in practice against most of the 47 plaintiffs, much less the other 984 districts covered by the statewide permanent injunction. As the question is one of constitutionality, we cannot simply presume that all districts are alike.

The trial judge pointed to evidence from nine “focus districts” and the testimony of a dozen superintendents as proof that loss of meaningful discretion was “systemic/statewide.” But there was no evidence these districts were, statistically representative of all others. To the. contrary, the handful of successful focus districts were un representative — 78 percent of the plaintiffs’ focus districts were poor districts, while 72 percent of the actual plaintiffs were rich ones.

Nor did the parties agree that proof about the focus districts proved anything *815about the rest. Even if they had, such an agreement would be unenforceable. In Terrazas v. Ramirez, we reversed a permanent injunction that ordered election redistricting based on an agreement by all the parties (including the Governor and Attorney General),96 noting that such agreements are generally unenforceable in cases affecting the public:

Apportionment affects every person in the State, yet only a very few parties can be involved in any lawsuit challenging redistricting. The trial court must attempt to consider the interests, not only of the parties in the case, but of others who are not present. For this reason, the agreement of the parties in a reapportionment lawsuit cannot alone be conclusive of either the validity of the statute or, if it is found to be invalid, the relief to be granted.97

Similarly, as schools and property taxes affect far more Texans than the parties at this trial (none of whom, again, were simply taxpayers or families of school children), the trial court could not grant relief covering districts as to which there was no proof.98

In a state as diverse as Texas, some programs and expenses may be mandatory in one district, but supplemental in another. Even if a dozen districts proved that they were forced to incur all their expenditures (which none did), that would not justify an injunction extending beyond them.99

This is not a class action. No class has been certified, and given the individual ways in which each school district spends money, it is unlikely any could be. But even if one was, we could not grant relief extending to nonparty school districts without a “rigorous analysis.”100 Yet the Court today grants a statewide injunction affecting hundreds of nonparty school districts without class certification, evidence, analysis, or even an explanation. This looks too much like “enjoin now and worry later.”101

B. Too Many Statutes

Second, there is no evidence to support an injunction against every statutory aspect of the Texas school-finance system.

The Court finds only one constitutional violation — that the tax-rate ceiling in subsection 45.003(d) of the Education Code violates Article VIII. As already noted, there is no evidence showing this is the case in every school district in Texas. But even if there were, that would justify nothing beyond declaring this one subsection unconstitutional.

When we declared a single provision of the Water Code an unconstitutional delegation to landowners, we did not enjoin all water quality regulations in Texas.102 *816When we found a single provision of the Tax Code unconstitutional, we did not enjoin all taxes; to the contrary, we reformed the lower court’s injunction to make it narrower.103 When we found an absolute two-year statute of limitations for medical malpractice claims unconstitutional as applied to minors, we did not enjoin the entire statute but merely tolled limitations for minors.104

In each of these cases, we narrowly limited our orders to the legislation we found unconstitutional. By the same standard, if the Legislature imposed a property tax on the nine Texas counties whose names begin with “J”, surely we would declare only that statute unconstitutional; we would not stop all state funding in those counties, much less in the other 245.

But today the Court does precisely that, finding one subsection unconstitutional as applied to nine focus districts, and then affirming an injunction against the entire Texas school-finance system. This injunction includes most of Chapters 41 and 42 of the Texas Education Code — a collection of almost 100 different statutes. This is far too broad.

The Court acknowledges that the single violation here could be corrected by limiting relief to that single statute.105 But it imposes far more sweeping relief, on the ground that we must “leave such matters to the discretion of the Legislature.”106 In other words, rather than enjoining a single statute in a handful of districts, the Court enjoins scores of statutes across the entire State — in deference to the Legislature. Reasonable people may question whether this is very much deference.

It is true that we have enjoined the entire school-finance system before, but never for grounds as limited as those here.107 In Edgewood I and II, there was a “fundamental flaw” in the system, “not in any particular provisions but in its overall failure to restructure the system.”108 By holding that Article VII required the entire system to “draw revenue from all property at a substantially similar rate,”109 our ruling could not be narrowly limited to a small part.

Similarly, because the statute we held unconstitutional in Edgewood III mandated a state property tax in every Texas county, the injunction we issued had to *817cover every county too.110 Nor could we limit relief to the portion of the system held unconstitutional, as there would have been little financing left over for schools without it.111

By comparison, nothing about the Article VIII claim here inevitably extends to the whole school-finance system. Surely a single violation of Article VIII anywhere cannot justify an injunction shutting down school finances everywhere.

The Court says the current system cannot survive without the tax-rate cap, because “for districts that need additional revenue, the funding system would be inefficient.” 112 But the Court cannot have it both ways — if school districts “need” more funding, then current funding cannot be adequate for a general diffusion of knowledge; conversely, if the current funding is adequate (as the Court explicitly holds), then the cap only affects supplemental spending. As the Texas Constitution does not guarantee equal supplemental spending,113 the cap is hardly “central” to a constitutional system.114

Of course, it is no mystery why the plaintiff school districts never asked for narrower relief. If only section 45.003(d) were declared unconstitutional, they would once again have meaningful discretion to set tax rates as they wish, and could raise them to pay for all the programs they say their communities demand. But they also might find out at the next election that their beliefs about community demand were somewhat exaggerated.

Instead, by enjoining school-finance across the state, the school districts here hope to obtain funding from sources other than those within their own borders. Raising revenues from outside sources is unlikely to make school districts more accountable or more efficient. Neither equity nor the Texas Constitution allows school districts to demand supplemental programs on condition that someone else pay for them.

* * * * * *

The Court closes by reminding the Legislature how important education is to the future of this State and its people. This seems an odd way to conclude an opinion that rejects every claim except that the Legislature has imposed a statewide ad valorem tax. If our goal is to improve education, we should not enjoin the entire school-finance system on collateral grounds to pressure the Legislature to change it.

But we should demand efficiency, as that is what the Texas Constitution requires. Recognizing the common meaning of “efficient” would not require us to abandon our previous school-finance cases, or the equity for Texas schools they require. But we cannot keep overlooking the one standard the Texas Constitution explicitly demands. Nor do we help Texas school children by insisting “efficient” means nothing beyond equal access to taxes.

Someday, the Texas school system must become “efficient” by 21st century stan*818dards. As that is what the Texas Constitution requires, we should start that process today.