The State appeals orders of the Superior Court (Galway, J.) granting the plaintiffs’ declaratory judgment petition, finding the statewide property tax unconstitutional as applied, and requiring the State to repay the $880 million collected under the tax to date. We reverse.
In December 1999, three property owners from three different communities filed a petition for declaratory judgment and injunctive relief challenging the constitutionality of the property tax as applied statewide. The plaintiffs challenged both the assessment method of the tax and the distribution method after collection. After a six-day bench trial, the trial court concluded that the statewide property tax as applied violates Part II, Articles 5 and 6 of the New Hampshire Constitution but that the distribution method does not violate the plaintiffs’ right to equal protection. The trial court ruled that the State had agreed to repay $880 million if the tax was found unconstitutional and ordered that relief.
The issues before us are: (1) whether the trial court erroneously determined that the statewide property tax, as applied, violated Part II, Articles 5 and 6 of the New Hampshire Constitution; and (2) whether the trial court erroneously ruled that the State had an obligation to repay $880 million if the statewide property tax is found to be unconstitutional. The plaintiffs have not appealed the *367trial court’s ruling rejecting their equal protection claim, and therefore we do not address it.
I. Facts
A. The Legislation
Following our decision in Claremont School District v. Governor, 142 N.H. 462 (1997) (Claremont II), the legislature passed House Bill 117, establishing a statewide property tax to help fund the State’s obligation to provide a constitutionally adequate public education. See Laws 1999, ch. 17. House Bill 117, however, contained a provision that allowed certain communities to phase in the full rate of the tax over five years, which this court found unconstitutional. See Claremont School Dist. v. Governor (Statewide Property Tax Phase-In), 144 N.H. 210, 212 (1999) (Claremont III). The legislature subsequently passed House Bill 999, which reenacted the statewide property tax without the phase-in provision and provided that the tax will expire at the beginning of 2003. See Laws 1999, ch. 338.
House Bill 999 imposes a tax at the rate of $6.60 per $1,000 of the value on certain real property across the State. See RSA 76:3 (Supp. 2000). The New Hampshire Department of Revenue Administration (DRA) calculates the total amount of the statewide property tax to be raised by each municipality based upon the equalized value of the property within its borders. See RSA 76:8 (Supp. 2000). The New Hampshire Department of Education calculates the total amount of adequate education funds that each municipality needs to educate its students, based upon the number of students within its borders. See RSA 198:40 (1999). Under the statute’s distribution formula, communities that raise more funds through the tax beyond that necessary to fund an adequate education for their students are required to pay the excess funds to the education trust fund for distribution to communities unable to raise sufficient funds to meet their cost of adequacy. See RSA 198:39, :41, :42, :46 (1999 & Supp. 2000). This has resulted in the characterization of some municipalities as “donor” communities and others as “receiver” communities, although House Bill 999 contains no such designations. The plaintiffs in this case are from “donor” communities.
For the 2000 tax year, $825 million was the amount required to fund an adequate education for each school child in the State. Of this amount, $442 million was raised through the statewide property tax. Of the $442 million, $418 million was distributed directly to the school districts of the municipalities where the money was collected. The remaining $24 million was placed in the education trust fund *368with $383 million from the other revenue sources and distributed to municipalities with insufficient property values to raise an amount necessary to fund an adequate education in their school districts.
B. The Tax Process
Statewide property tax bills reflect the equalized value of property as of April 1 of each year. See RSA 76:2 (1991); RSA 76:8, I (Supp. 2000); RSA 21-J:3, XIII (2000). The selectmen or the assessors in each municipality annually determine the local assessed value of the property within each municipality as of April 1. See RSA 74:1 (1991); RSA 75:8 (1991). Each municipality completes a form that lists the total value of its properties, the changes to its properties, and any exemptions and credits claimed. This information is supplied to the DRA.
RSA 21-J:3, XIII requires the DRA to “[e]qualize annually . . . the valuation of the property as assessed in the [municipalities] in the state ... by adding to or deducting from the aggregate valuation of the property . . . such sums as will bring such valuations to true and market value of the property.” The first step in the equalization process involves a ratio study which compares, within each municipality, the assessed property values submitted by the municipality to the sales prices of all properties sold in the municipality in the prior year. See RSA 21-J:9-a (2000). Sales information is provided to the DRA by a private company with which the State contracts to collect information from the registry of deeds in each county. The DRA then performs a sales screening process so as to include only “arms-length” transfers of property. See id. Assessing officials within each municipality are required to certify the information necessary for the DRA to conduct the annual sales-assessment ratio study. See id.
Once sales are screened, the DRA establishes an individual sales/assessment ratio based upon the sales and assessment information from each municipality for each property used in the study. The DRA then calculates three ratios based upon the assessment and sales information for each municipality. First, the median ratio is calculated by arraying the individual ratios from the highest to the lowest. The median ratio is the middle. Second, the mean ratio is calculated by averaging the individual ratios. Third, the weighted mean or aggregate ratio is calculated by dividing the sum of the assessed values for all properties in a community for which there have been sales by the sum of their sale prices. The DRA then determines which ratio best equalizes all of the property in each municipality. Most often, the DRA chooses the median ratio.
*369The DRA next uses the selected ratio to adjust the municipality’s assessed values, either upward or downward, in order to approximate full value. As the DRA explains in its brochure:
Adjustments are not made to any individual properties. Rather, the total value of all property in town is adjusted based upon the comparison of recent property sales with local property assessments. For example, if the comparison of recent sales indicates that on the average, the town is assessing property at 90% of market value, then the total local assessed value of the town would be increased by 10% in order to approximate the town’s full value. If the comparison indicates that, on the average, the town is assessing at 105% of market value, then the total local assessed value would be decreased by 5%.
Once the equalized property valuations are determined, each municipality’s equalized valuation is multiplied by $6.60 per $1,000 to determine the portion of the statewide property tax each municipality must raise. See RSA 76:8. To determine the rate paid by individual taxpayers, each municipality divides the uniform rate by the percentage that the DRA used to equalize its property.
The final step in the DRA’s annual ratio study is to calculate the coefficient of dispersion (COD) for each municipality. The COD measures the percentage that individual municipality ratios, on average, deviate from the median ratio. A COD of ten, for example, indicates that the sales/assessment ratios in a municipality, on average, deviate from the median sales/assessment ratio by ten percent.
II. Standard of Review
Establishing “the rules by which each individual’s just and equal proportion of a tax shall be determined is a task of much difficulty, and a very considerable latitude must be left to the legislature on the subject.” Opinion of the Justices, 77 N.H. 611, 615 (1915) (quotation omitted). In reviewing the statewide property tax, we determine only whether there is a “clear conflict with the Constitution” in the tax as applied, and do not concern ourselves with whether the tax is “wise, reasonable, or expedient.” Petition of Boston & Maine Corp., 109 N.H. 324, 325-26 (1969) (quotation omitted). As we have often stated, our task is neither to establish educational policy nor to determine the appropriate mechanism for its funding. See Claremont II, 142 N.H. at 475. The statewide property tax law, like any legislative act, is presumed constitutional *370and will not be declared invalid except upon “unescapable grounds.” Niemiec v. King, 109 N.H. 586, 587 (1969) (quotation omitted).
We review the trial court’s legal conclusions and application of law to fact de novo. See Bursey v. CFX Bank, 145 N.H. 126, 129 (2000). The trial court’s “[findings of fact . . . are binding on us unless they are not supported by the evidence or are erroneous as a matter of law.” Blagbrough v. Town of Wilton, 145 N.H. 118, 124 (2000) (quotation omitted).
III. Part II, Article 5
A. In General
Pursuant to Part II, Article 5 of the New Hampshire Constitution, the legislature is granted “full power and authority ... to impose and levy proportional and reasonable assessments, rates, and taxes upon all inhabitants of, and residents within, the said state; and upon all estates within the same.” In order for a tax to be proportional, all property in the taxing district must be valued alike and taxed at the same rate. See Opinion of the Justices, 99 N.H. 525, 527 (1955). Each taxpayer’s property must be valued at the same percentage of its true value as all the taxable property in the taxing district and “shall be valued within a reasonable time before the tax is assessed.” Bow v. Farrand, 77 N.H. 451, 451-52 (1915). A change in either the rate or the valuation affects the tax. Opinion of the Justices, 99 N.H. at 527; see Opinion of the Justices, 76 N.H. 609, 611 (1913).
Taxes must not merely be “proportional, but in due proportion, so that each individual’s just share, and no more, shall fall upon him.” Rollins v. Dover, 93 N.H. 448, 449 (1945) (quotation omitted). “[A]s any one’s payment of less than his share leaves more than their shares to be paid by his neighbors, his non-payment of his full share is a violation of their constitutional right.” Id. Nonetheless, “[t]he demand of constitutional equality in taxation anticipates some practical inequalities.” City of Berlin v. County of Coos, 146 N.H. 90, 94 (2001). “Absolute mathematical equality is not obtainable in all respects if taxation is to be administered in a practical way.” Id. (quotation omitted).
B. Burden of Proof
The plaintiffs sought a declaration that the tax “as applied” violated Part II, Article 5. To prevail upon their declaratory judgment petition, the plaintiffs were required to show a “present legal or equitable right and an adverse claim.” Delude v. Town of *371Amherst, 137 N.H. 361, 363 (1993) (citation omitted). Such a petition cannot be based upon a set of hypothetical facts. See id. “There must be some action taken and harm done before [a statute] will be invalidated.” Id. In the context of a challenge to a tax statute, the plaintiff is required to prove the “practical operation and effect” of the tax was unconstitutional. Austin v. State Tax Comm’n, 114 N.H. 137, 139 (1974), reversed on other grounds, 420 U.S. 656 (1975) (quotation omitted).
As the trial court recognized, the plaintiffs had the burden of demonstrating the statewide property tax was unconstitutionally disproportionate, as applied. See Chasan v. Village District of Eastman, 128 N.H. 807, 818-19 (1986) (plaintiffs have burden to prove disproportionality under Part I, Article 12 of New Hampshire Constitution). An “as applied” constitutional claim, in particular, requires proof of harm to the plaintiffs. See Delude, 137 N.H. at 364; Austin, 114 N.H. at 141. That the tax assessment procedures “may have allowed for illegal results does not end the matter. Rather, we must turn to the validity of the . . . procedures, as applied, to determine whether the plaintiffs in this case were injured.” Macioci v. Commissioner of Revenue, 438 N.E.2d 786, 793 (Mass. 1982).
The plaintiffs did not assert that their own property appraisals were invalid or that they paid the statewide property tax based upon something more than the market value of their properties. Rather, they asserted that because properties in other municipalities were appraised at less than market value and because the DRA’s equalization process is fundamentally flawed, other taxpayers paid the statewide tax based upon something less than market value. Accordingly, the plaintiffs claimed they bore more than their fair share of the statewide property tax burden.
We have not previously had the opportunity to articulate the burden of proof where the taxpayer’s challenge is based upon the under-assessment of other taxpayers. In the Claremont cases, we were faced with either facial challenges to the education tax or challenges to the rate of taxation. In Claremont II, for instance, the issue was whether municipalities in the State had the same rate of taxation. See Claremont II, 142 N.H. at 470-71. At the time, the DRA computed a different property tax rate for each school district, based upon the revenue needed to meet that school district’s budget. Claremont II, 142 N.H. at 467. This system resulted in property tax rates that differed among municipalities by as much as 400 percent. Id. at 470. Because the taxes were raised to fulfill a State purpose, we determined the State was the relevant taxing district, and we held that, in the context of a statewide tax, these varying property *372tax rates violated Part II, Article 5. Id. at 470-71. To comply with Part II, Article 5, the statewide property tax “must be administered in a manner that is equal in valuation and uniform in rate throughout the State.” Id. at 471.
In Claremont III, the issue was whether the tax, on its face, treated taxpayers in property rich communities differently from taxpayers in other communities. See Claremont III, 144 N.H. at 213 (quotation omitted). The legislation at issue in Claremont III established a statewide property tax “at the uniform rate of $6.60 on each $1000 of the value of taxable property.” Id. (quotation omitted). The act, however, phased in the uniform rate of taxation over five years for fifty property rich towns across the State, while imposing it immediately upon the remaining towns. Id. We determined that the differential treatment afforded the property rich communities was not justified and was “so arbitrary as to serve no useful purpose of a public nature.” Id. at 216 (quotation omitted). Accordingly, we held the phase-in provision violated Part II, Article 5. Id. at 216-17.
Courts interpreting uniformity clauses in other state constitutions have required taxpayers challenging the under-assessment of other taxpayers to establish a constitutional violation by proving more than merely that the others are under-assessed. For instance, taxpayers in New Mexico must show “substantial” inequality such that the inequality “amounts to an intentional violation of the essential principle of practical uniformity.” Hannahs v. Anderson, 966 P.2d 168, 176 (N.M. Ct. App. 1998), cert. denied, 972 P.2d 351 (N.M. 1998) (quotation omitted). Taxpayers in Kansas are held to a similar burden.
[UJniformity and equality in a constitutional and statutory sense does not require mathematical exactitude in the assessment valuation of property for taxation. . . . Mere excessiveness of an assessment or errors in judgment or mistakes in making unequal assessments will not invalidate an assessment, but the inequality or lack of uniformity, if knowingly high or intentionally or fraudulently made, will entitle the taxpayer to relief.
Appeal of Andrews, 851 P.2d 1027, 1030-31 (Kan. Ct. App. 1993).
As the Massachusetts Supreme Judicial Court has explained, “practically it is impossible to secure exact equality or proportion in the imposition of taxes but. . . the statutory aim should be towards that result by approximation at least.” Bettigole v. Assessors of Springfield, 178 N.E.2d 10, 15 (Mass. 1961) (quotation and brackets *373omitted). This equality of approximation “requires attempting equality of assessment in absolute good faith and the best abilities of the public officers charged with making valuations.” Id. When a taxpayer alleges that he or she pays more than the taxpayer’s just proportion, not because his or her land is assessed in excess of its fair cash value, but because the land of others is intentionally assessed at even lower percentages of full, fair value, the Massachusetts Supreme Judicial Court requires proof of specific facts showing a “widespread scheme of intentional discrimination,” rather than merely isolated lack of uniformity, to prove the tax is unconstitutionally excessive. Macioci, 438 N.E.2d at 797. In Montana, as well, temporary inequalities will not establish a uniformity clause violation. Rather taxpayers must show intentional and systematic discrimination, constructive fraud or arbitrary action. Patterson v. State Dep’t of Revenue, 557 P.2d 798, 803 (Mont. 1976).
We now hold that to establish a violation of the uniformity clause based upon the under-assessment of other taxpayers, a taxpayer must prove a systematic pattern of taxation that is not proportional and reasonable. To prevail, the taxpayer must prove specific facts showing a “widespread scheme of intentional discrimination.” Macioci, 438 N.E.2d at 797. Isolated or inadvertent lack of uniformity will not suffice. See id. Intentional discrimination may be proved with direct evidence, or indirectly with proof of inequality or lack of uniformity substantial enough that intent may be inferred. See Hannahs, 966 P.2d at 176.
Thus, to prevail upon their declaratory judgment petition, the plaintiffs were required to establish they were harmed by the practical operation of the statewide property tax. Because the plaintiffs did not claim their own assessments were based upon something other than market value, but instead claimed that other taxpayers were under-assessed, the harm the plaintiffs were required to show was widespread intentional discrimination.
C. Evidence of Flaws in System
Evidence at trial established that the statewide property tax system as currently applied has significant shortcomings. In particular, there are five areas which raise serious concerns regarding the system’s ability to achieve taxpayer proportionality and equality. As we explain in section D, these concerns, although serious, are insufficient to establish a constitutional violation.
*374 1. Lack of Standards for Local Assessments
The DRA relies upon officials within each municipality to report the value of the taxable property therein. The DRA’s Deputy Director for the Property Appraisal Division, Linda Kennedy, testified at trial that while local officials in each municipality are responsible for obtaining assessment values and reporting them to the State, different methods may be used to update property values within each municipality. Kennedy also testified that while the DRA audited assessments in municipalities with poor CODs in 1999, no comprehensive auditing procedures are yet in place to verify that assessments are being performed correctly and accurately at the local level. Kennedy acknowledged that the municipality’s assessments, along with the sales price data, form the “keystone” of equalization.
Kennedy testified that under the DRA’s administrative rules, a municipality may update property values in one of three ways. It may do a full “revaluation,” which entails a full measure and list of all property within the municipality. A full revaluation involves physically inspecting property exteriors and may include inspecting property interiors. A municipality may also do a partial revaluation, which is a revaluation of a neighborhood or a class of property within a municipality. The third revaluation option is to use a sales analysis to update property assessments. The DRA’s manual for assessors advises that revaluations and updates are “periodical necessities” to ensure equity within a municipality.
Some municipalities have undertaken a full revaluation of property recently, while others have not done so in decades. Standards promulgated by the International Association of Assessing Officers (IAAO) provide that
property should be physically reviewed and reappraised at least every six years. This can be accomplished in at least three ways: (1) reappraising all properties at periodic intervals (e.g., every four to six years); (2) reappraising properties on a cyclical basis (e.g., one-fourth or one-sixth each year); and (3) reappraising on a priority basis as indicated by assessment-ratio studies or other considerations, while still ensuring that all properties are physically reviewed at least every sixth year.
Both the plaintiffs’ and the State’s experts endorsed this- IAAO standard at trial. The plaintiffs’ expert testified that there can be no consistency or reliability in a system that does not perform regular *375revaluations of its taxable property and that without current, consistent data there can be no proportionality across taxpayers. The State’s expert likewise agreed that while trending can be effective to update property data for short periods, full revaluations are required at regular intervals.
2. Lack of Standards for Exclusion of Sales
Upon receiving information from each municipality of the total assessed value of its properties, the DRA conducts a ratio study comparing the assessed values of each municipality’s properties to the sales prices of all properties sold in the prior year in each municipality. Evidence at trial underscored the degree of discretion afforded the DRA in choosing the sales information included in the State’s ratio study.
Kennedy testified that only “arms-length” sales are included in the ratio study and that she excludes all sales that she determines are not “arms-length.” She also testified, however, that the DRA does not have any set procedures for including or excluding sales data; rather, those decisions are based solely upon her judgment.
In addition, the trial court found that as a result of the DRA’s sales screening
the sample sizes used in the sales analysis are very limited, with many exclusions, and with no set procedures or verification process for exclusion of sales. So many sales are excluded from the study, that these samples give very little indication of what percentage of town property is actually being used in the equalization process, or how indicative the samples are of the fair market value of property in any town across the State. In fact, the evidence presented demonstrates that some town[s’] ratios were calculated from studies based on as few as two property values.
The plaintiffs’ expert also expressed concern about the extremely small samples used in the ratio study. He testified that the sample size in the DRA ratio study significantly affects the reliability of the statistics, especially when there are no confidence intervals calculated to monitor their reliability, and that change is necessary to achieve fairness in the study. Likewise, the State’s expert testified that: (1) many of the sample sizes used in the ratio study were too small to generate confidence in the ratios produced; and (2) no confidence intervals were calculated by the DRA to determine whether its statistics were reliable.
*376 3. Lack of Guidelines for Choice of Ratio
Once sales are screened, Kennedy establishes an individual ratio for each property used in the study by dividing the assessed value of the property by its reported sales price. She also determines the ratio to be used to equalize all of the property in each municipality. While in most instances the median ratio is chosen, other ratios may be used. Kennedy acknowledged that her choice of ratio affects the amount of tax a municipality must pay and that the DRA has no written policy governing her choice of ratio.
A High CODs
As part of the DRA’s annual ratio study, Kennedy calculates the COD for each municipality. The COD is defined by the DRA as the average percentage that ratios deviate from the median ratio for the community. The COD is a measure of equity within a municipality. The lower the COD, the greater the uniformity of assessments in a municipality and therefore the closer the municipality is to achieving equity of property values.
The DRA classifies CODs between zero and five percent as indicating excellent equity, between six and ten percent as good equity, between eleven and fifteen percent as average equity, between sixteen and twenty percent as borderline equity, and anything over twenty percent as unacceptable equity. The DRA’s “Ratio and Coefficient of Dispersion Guidelines” recommend COD standards of ten percent or less for single-family residential property in new, more homogeneous areas, and fifteen percent or less for single-family residential property in older, heterogeneous areas and income producing properties in large, urban jurisdictions. The outside standard of twenty percent is recommended only for rural and seasonal property and vacant land.
Evidence at trial showed that under the DRA’s own definitions, during the tax years in question at least half of the municipalities in the State had borderline to unacceptable CODs overall and unacceptable CODs in at least one class of property. In the DRA’s 1998 Equalization Summary, 53.2% (124/233) of municipalities with a calculable COD had a COD of over fifteen percent. When groupings by different classes of property were analyzed, 51.7% (122/236) of municipalities had at least one class of property with a COD over twenty percent. In the 1999 Equalization Summary, it was reported that 45.5% (107/235) of municipalities had a COD of over fifteen percent. The stratified analysis revealed that 65.7% (148/225) of municipalities had at least one class of property with a COD of over twenty percent.
*377There is no statutory authority for the DRA to set rules on CODs. While the DRA recommends that municipalities stay within the COD guidelines, it did not have the authority to require municipalities to take corrective action. Under House Bill 999, however, the DRA now may petition the New Hampshire Board of Tax and Land Appeals to issue an order requiring a municipality to reassess property. See RSA 21-J:3, XXV (2000). Moreover, the DRA is authorized to monitor local assessments and supervise local assessors to ensure that appraisals are accurate. See RSA 21-J:11 (2000).
5. Lack of Verification
As noted above, there are few written guidelines and few quality control mechanisms. The DRA does not have comprehensive procedures to verify that assessments are being performed correctly and accurately at the local level. Expert testimony indicated that all local data should be validated through audit procedures. The trial court found that the need for local audits is heightened when the localities perform their own assessments and sales screening, because each locality may have a vested interest in achieving lower CODs and lower ratios, which directly affect the assessment of taxes. In addition, testimony highlighted the need to review the method by which sales are screened and to establish confidence intervals to monitor the reliability of the statistics used in the ratio study. Testimony provided by both sides’ experts supports the trial court’s finding that “New Hampshire and the DRA need to . . . perform audits of towns within the State to validate data, and calculate confidence and reliability indicators for all of their statistics used.”
D. Conclusion
We now decide whether, based upon the evidence presented to the superior court, the plaintiffs carried their burden of proving the statewide property tax was unconstitutional as applied under Part II, Article 5. We conclude that they did not.
The current system of administering the statewide property tax raises serious concerns as to whether it is proportional and reasonable, as required by Part II, Article 5. In section C we have described the evidence that supports the factual basis of our concerns. Much of that evidence came from expert witnesses for the plaintiffs and the State. They analyzed data and offered opinions upon various facets of the statewide property tax system, as well as *378upon the system as a whole. After reviewing this testimony, the trial judge observed:
The parties extensively debated at trial the statistical calculations and charts created by both [the State’s expert] and [the plaintiffs’ expert]. Errors were noted and recalculations performed, and the only point ultimately agreed upon was that none of the charts reflected the actual taxing situation occurring in New Hampshire.
The trial judge noted that the experts for both sides relied upon data from the DRA and “[i]f the underlying DRA data is flawed or unreliable, all calculations based on that data are also flawed and unreliable.” The trial judge found that “none of the charts presented reflect the actual values and ratios used by the DRA” and concluded that “the numbers presented in all of the charts lack s.upport to make them meaningful to the Court.”
Notwithstanding the absence of conclusions based upon reliable statistical data, the trial judge ruled:
What is clear from the expert testimony presented by both sides, is that without uniformity in assessments, sales, and ratios, without adequate sample sizes, without confidence intervals, and without full physical revaluations of property every five years to establish a base for a successful taxing system, the tax as applied is hot equal in valuation [ ]or proportional across the State.
We agree that the plaintiffs failed to present statistical data that provided a comprehensive and reliable picture of the actual operation of the statewide property tax system. In our view, however, the absence of such evidence is fatal to their claim. What the plaintiffs proved was that the taxing system is flawed. What they did not prove is that the flaws resulted in a systematic pattern of disproportionate taxation. Nor did the plaintiffs prove that the flaws produced such substantial inequality that the inequality must be deemed intentional.
For instance, although there was evidence that the DRA has discretion to choose different ratios to equalize the value of property in different municipalities, the plaintiffs’ experts did not testify that the use of different ratios resulted in the widespread failure to tax municipalities, and, derivatively, municipal taxpayers, at full market value. Based upon a study of thirty-three municipalities, the plaintiffs’ experts testified that if the median ratio for each municipality had been used, the difference between the statewide *379property tax the State actually collected in 1998 and the statewide property tax the State would have collected had they used the plaintiffs’ experts’ equalized values was .6 percent. As the State’s expert explained, “it really doesn’t make much difference” whether the median, mean or weighted mean ratio is used. “They’re all pretty close together and they’re usually not significantly different.”
The plaintiffs’ primary evidence of the flaws in the DRA’s equalization process was a study of thirty-three municipalities. At trial, one of the plaintiffs’ experts admitted that because he used the DRA’s data, his study had the same flaws as the DRA’s equalization study. The plaintiffs’ expert testified that “to make a statistically valid analysis of assessments to other indications of market value, we would have either had to use more years of sales or commissioned appraisals to be done, and we would have then produced for one or more municipalities an equalization study that fully complied with the IAAO standards on ratio studies.” Such a study would have given consideration “to the representativeness of the sales in the community to make sure that all types of property are adequately reflected [and] all value ranges are adequately reflected.” The expert admitted that no such study had been conducted.
The plaintiffs also relied upon evidence that several municipalities have “unacceptable” CODs. A COD measures only the average variation of assessment/sale ratios from the median ratio. A high COD indicates that, on average, the assessment/sale ratios within a municipality vary from the median ratio to a large degree. It is a measure of uniformity within a municipality only. The State’s experts testified that a COD of twenty percent or less complies with IAAO standards. In other words, the State’s experts testified that it is permissible for assessment/sale ratios within a municipality to vary from the median ratio by at least twenty percent. The plaintiffs’ expert admitted that, even if municipalities have CODs as high as twenty, forty or fifty percent, equalization can cure the disparities between municipalities.
Further, evidence that several municipalities have high CODs does not establish substantially disproportionate taxation. As the commissioner of the DRA explained at trial, proof that a municipality has an unacceptable COD for one type of property does not mean that all of the assessment/sale ratios in the municipality vary from the median ratio to a significant degree. The particular type of property with the unacceptable COD may represent only a small percentage of the total value of property in that municipality. Similarly, the DRA’s 1999 equalization survey showed that munici*380palities with CODs of over twenty percent represent only seven percent of the total equalized value of property in the State.
The plaintiffs argued that because municipalities, such as the City of Keene, have not revalued property through physical inspections in many years, the taxing system necessarily results in disproportionate taxation. The evidence showed, however, that even though Keene has not revalued its property through a city-wide inspection in thirty years, the process it uses to assess property results in assessments that are within an acceptable range of market value.
By contrast, in Claremont II, the plaintiffs submitted clear and unambiguous evidence that the property tax as applied required taxpayers in different communities to pay different property tax rates. The evidence demonstrated that one municipality imposed a property tax rate that was over 400 percent higher than that in another municipality. Claremont II, 142 N.H. at 470. In this case, there was no evidence that any municipality was assessed the tax at anything other than the uniform rate. More importantly, on this record, we simply cannot determine the impact upon the plaintiffs or upon any other taxpayers of the system’s administrative shortcomings.
It was the plaintiffs’ burden to prove that the statewide property tax was unconstitutional as applied. Not only did they fail to produce evidence that their own property taxes were disproportionate, but they also failed to show that the taxes of other property owners were disproportionate. They did not demonstrate that the property of any taxpayer was not “valued alike,” meaning at fair market value. Opinion of the Justices, 99 N.H. at 527; see State Ex. Rel. Stephan v. Martin, 608 P.2d 880, 886 (Kan. 1980) (uniformity clause requires that property taxation be assessed upon equal basis; equal basis provided by legislature is “fair market value in money”). Thus, in spite of evidence that the current system has deficiencies, the evidence was insufficient to show a systematic pattern of disproportionate taxation. See, e.g., Gitney v. Berks County, 635 A.2d 737, 743 (Pa. Commw. Ct. 1993).
IV. Part II, Article 6
A. Violation
Part II, Article 6 of the New Hampshire Constitution states that “there shall be a valuation of the estates within the state taken anew once in every five years, at least, and as much oftener as the general court shall order.” The plaintiffs and the trial court interpret Part *381II, Article 6 to require full physical inspections of all property within the State every five years. We disagree.
“In interpreting an article in our constitution, we will give the words the same meaning that they must have had to the electorate on the date the vote was cast.” Claremont School Dist. v. Governor, 138 N.H. 183, 186 (1993) (quotation omitted). “In so doing, we must place ourselves as nearly as possible in the situation of the parties at the time the instrument was made, that we may gather their intention from the language used, viewed in the light of the surrounding circumstances.” Id. (quotation and brackets omitted).
In New Hampshire, from the period of the establishment of the colony to the Revolution, a system of taxation was built upon the theory of taxing every person according to the person’s income. See Opinion of the Justices, 76 N.H. 588, 593 (1911). This system continued until 1789, “when the income theory was permanently abandoned and the taxation of every person in proportion to his estate was substituted.” Id. at 593-94 (quotation and ellipses omitted).
From the adoption of the constitution in 1784 until 1833, “all property taxed was not annually appraised, but much of the property taxed was rated by legislative act,” the value of which was set by the legislature at least every five years. Id. at 593. Inventories based upon these values were prepared by the selectmen in each municipality and submitted to the legislature. Taxes were apportioned among municipalities based upon the values contained in the inventories. See M. ROBINSON, A HISTORY OF TAXATION IN New Hampshire 38-39 (1902).
When Part II, Article 6 was adopted in 1784, “[i]t was known how estates had been valued since 1770, and in ordering a new valuation every five years like legislative action was expected.” Opinion of the Justices, 76 N.H. at 595. The framers thus did not intend for there to be physical inspections of property every five years. Rather, they intended that the legislature value large classes of property every five years by legislative act.
“After the change in the theory and practice of taxation in 1789, more attention was given to the valuation of property included in the inventory and to the taxation of new forms of personal and real property.” Id. at 594.
Changes in the valuation of real estate were especially marked. In 1803, the valuation of land was advanced, especially when compared with live stock, one of the chief items of the inventory. Again in 1830, a general reduction of *382nearly twenty-five per cent was made in the valuation of all classes of property that were arbitrarily rated by legislative act. In a few cases, however, the valuation was increased. Such a general readjustment indicates clearly that values were undergoing radical changes, and that the old method of rating large classes of property by the legislature for a term of years had become, under the influence of a new industrial era, unsuited to the conditions. A fixed list was extremely convenient while values remained stable, and proved equitable and hence satisfactory in its workings. When values began to change rapidly and new forms of property were appearing, the market price was nearer the just price than any that could be fixed months in advance by the legislature.
Id. at 594-95 (quotation omitted). Accordingly, in 1833, the legislature abandoned the rating system. See id. at 594. In its place, the legislature adopted “a less arbitrary method . . . providing that all taxable property should be appraised at its true value in money by the local assessors.” Id. at 595 (quotation omitted). “The purpose of the act of 1833, from its terms, evidently was to change the mode in which property was to be valued, so as to provide for an appraisal, as a more equal mode than an arbitrary valuation at certain sums for different kinds of property.” Smith v. Burley, 9 N.H. 423, 427 (1838). Individual property appraisals were thus not required until 1833.
Part II, Article 6 governs the valuation of property for taxation purposes. While it does not require physical inspections, it does require that property be assessed at market value at least every five years. See Opinion of the Justices, 76 N.H. at 596. Property assessment entails both a physical description of the property’s features and size and a determination of the unit values for different classes of property.
The legislature has enacted a property assessment process that we believe, if complied with, would satisfy and, even exceed Part II, Article 6’s requirements. The legislature continues to require that taxable property be assessed “at its true and full value in money.” RSA 75:1 (1991). Since at least 1876, the General Court has required local assessors and selectmen annually to examine all the real estate in their respective cities and towns and to reappraise any real estate that has changed in value since the last appraisal. See RSA 75:8 (1991); Hill v. Marvin, 98 N.H. 519, 522-23 (1954). The legislature has also granted the DRA authority to “monitor apprais*383ais of property and supervise appraisers” to ensure that appraisals “comply with all applicable statutes and rules.” RSA 21-J:11, II. This system provides a reasonable process for determining the market value of property within each municipality every year.
The evidence at trial, however, established that some municipalities are not complying with the statutory directive to appraise property annually, or even the constitutional mandate to appraise it every five years. For instance, at trial, Kennedy testified that as of 2000, approximately seventy municipalities had failed to perform a full revaluation, partial revaluation, or update of property values since 1994 “that [the DRA] had knowledge of.” This failure is a violation of the State’s obligation to enforce Part II, Article 6. See Opinion of the Justices, 76 N.H. at 595.
B. Remedy
Based upon this violation of Part II, Article 6, the plaintiffs seek to declare the entire statewide property tax scheme unconstitutional as applied. We cannot declare a taxing scheme invalid as applied, however, absent proof of harm. See Delude, 137 N.H. at 364; Austin, 114 N.H. at 141. The plaintiffs did not prove harm. While the evidence showed the State has failed to ensure municipal assessments every five years, it did not show that this failure resulted in disproportionate taxation. See Opinion of the Justices, 76 N.H. at 596 (Part II, Article 6 requires “appraisals of the property to be taxed at its fair value, so that proportional taxes might be laid”); Thompson v. Kidder, 74 N.H. 89, 95 (1906) (“Article 6 . . . was intended to secure proportional assessments by requiring frequent revaluations”). We thus reverse the trial court’s declaration that the statewide property tax is unconstitutional as applied, and, accordingly, reverse its order requiring the State to return all of the statewide property taxes already collected.
We conclude, however, that the State must implement appropriate enforcement measures to ensure that each municipality assesses property within its borders every five years, as required by Part II, Article 6. The record in this case shows that steps are already being taken to enforce the current statutory scheme. Because the present statewide property tax is scheduled to expire in 2003, the State should have at least until then to develop effective enforcement measures. The people of this State should be able to rely upon the good faith and common sense of the executive and legislative branches to take the necessary action, not just because by doing so the State may avoid future successful legal challenges, *384but because that is the essence of our constitutional form of government.
Reversed.
NADEAU, DALIANIS and DUGGAN, JJ., concurred ; BROCK, C.J., and BRODERICK, J., dissented.