H & R Block Eastern Tax Services, Inc. v. State, Department of Commerce & Insurance, Division of Insurance

267 S.W.3d 848, 2008 Tenn. App. LEXIS 53, 2008 WL 269514

• Court: Court of Appeals of Tennessee
• Decided: January 30, 2008
• Docket: M2007-00656-COA-R3-CV
• Status: Published

Opinion

OPINION

CHARLES D. SUSANO, JR., J„

delivered the opinion of the court,

in which HERSCHEL P. FRANKS, P.J., and D. MICHAEL SWINEY, J., joined.

This case presents a pure and novel question of law regarding the Tennessee statutory definition of insurance. At issue is the “Peace of Mind” program (“POM program”) offered by H & R Block Eastern Tax Services, Inc. (“Block”) to customers who hire Block to prepare their tax returns. Essentially, Block offers its customers the option of purchasing, for an additional fee, an enhanced version of Block’s basic guarantee of the accuracy of its tax — preparation services. Block promises customers who purchase the POM program that, in the event Block makes an error which results in the customer’s tax liability being initially underestimated, Block will pay up to $5,000 of the customer’s newly revealed tax liability. The Tennessee Department of Commerce and Insurance (“the State”) contends — and the Commissioner of that department declared-that the POM program is a contract of insurance under T.C.A. § 56-7-101 as it was written when this case began, and, therefore, Block is subject to various penalties under that statute because it was not licensed to sell insurance. The trial court upheld the Commissioner’s ruling. We disagree with both the Commissioner and the trial court, and find that the POM program is not properly characterized as insurance under the statute. Accordingly, we reverse.

I.

The facts are not in dispute. We quote, in pertinent part, from the joint stipulation of facts in the administrative record:

Petitioner H & R Block Eastern Tax Services, Inc. (“Block”) and Respondent Tennessee Insurance Division (“Division”) stipulate that the following matters shall be deemed true, without the necessity of further proof thereof:

Block’s Business and Services

1. Block’s primary and core business is that of providing income tax preparation services to taxpayers throughout the United States.

2. Block reported $2.1 billion in fees from tax preparation and related services provided to 18.2 million clients from January 1 through April 15, 2008.

3. Block customers paid an average fee of $119.41 per tax return for the 2003 tax season.

[[Image here]]

10. Block’s services are provided to customers in two different types of offices: standard and premium.

11. Block’s premium offices provide tax consultation and preparation services in private offices rendered by tax preparation professionals with the highest level of tax experience, training and expertise offered by Block.

[[Image here]]

The Basic Guarantee

13. Incidental to and directly associated with its income tax preparation services, Block provides to its customers, at no additional charge, a basic guarantee regarding the accuracy of its tax preparation services (“Basic Guarantee”).

14. The Basic Guarantee provides that, if Block makes an error in the preparation of a tax return, Block will prepare an amended return and pay any interest and penalties which may be due, but will not pay any additional taxes which may be due.

15. Also included in the Basic Guarantee is the promise by Block that it will appear at an audit with the taxpayer to explain how the taxpayer’s return was prepared.

The Peace of Mind Program

16. As a supplement to the Basic Guarantee, Block offers its customers, for an additional fee, the Peace of Mind Program (“POM Program”). [A footnote here states, “The POM Program is provided to all customers, at no additional charge, in Block’s premium offices.”]

17. The POM Program provides that, if Block makes an error in the preparation of a customer’s tax return, in addition to the benefits provided under the Basic Guarantee, Block will pay any additional taxes which may be due to federal, state or local tax authorities, subject to a maximum amount of five thousand dollars ($5,000).

18. However, additional taxes are paid if, and only if, the error is Block’s fault; Block will not pay if the error(s) is due to the fault of the customer (i.e., the customer provides Block with incorrect information).

19. The POM Program is ancillary to and supports Block’s core business of tax preparation.

20. The fee for the POM Program in Block-owned offices is twenty-seven dollars ($27.00) and the fee for tax preparation services is one hundred nineteen dollars ($119.00).

21. During the 2001 tax season, Block offices prepared 14,209,454 tax returns, of which only 2,419,534 customers, or seventeen percent (17%), purchased the POM Program.

22. The amount Block will pay a customer under the POM Program is limited to a cumulative total of five thousand dollars ($5,000).

23. The POM Program is sold only for the current tax year and at the time Block’s tax preparation services are rendered.

24. If the tax return is prepared by someone other than Block, the POM Program is not available.

25. Under the POM Program, taxpayers must notify Block of its error within three years from the filing deadline, not including extensions, for such return.

26. The POM Program addresses the possibility that a Block tax return preparer will make a mistake or its tax preparation software is defective, and the reasonable reliance by the customer that no such mistake will be made and that the tax preparation software is free of defects.

27. Through the POM Program, Block stands behind the accuracy of its tax preparation services. When the tax return is completed, the customer walks away with the assurance that the tax return is accurate and that the refund or liability reflected on the return is correct. The customer may very well alter his financial position based on the information reflected on the return.

28. If Block commits an error and the tax return reflects a refund due, when, in fact, the taxpayer is not entitled to a refund, the taxpayer has a right under the POM Program to be restored to the position that Block said the taxpayer was in at the time that it prepared the return. Such restoration is what is normally contemplated under a warranty.

29. Conversely, if Block commits an error and the tax return reflects that the taxpayer owes money, when in fact, he is entitled to a refund, the taxpayer would be entitled to a refund from the particular tax authority and Block will not pay out any money.

30. Errors in the preparation of a tax return are not uncontrollable events such as an act of God or other external contingency. Whether Block makes an error is completely within the control of the Block tax preparer. Block’s obligations under the POM Program are not triggered by any external contingency.

31. In the event it prepared an inaccurate return, Block is agreeing to compensate the taxpayer for its negligence by paying the taxpayer’s additional taxes (only if additional taxes are due).

32. The taxpayer does not purchase the POM Program solely to cover additional taxes that may be due. Block’s errors do not always result in additional taxes due. Rather, it is just as likely that Block’s error will result in an additional refund for the taxpayer. In this case, Block would only be providing services to the taxpayer and would not be responsible for paying any additional taxes.

33. Block actually provides the tax preparation services which are covered by the POM Program. The POM Program is not being provided by an outside company that has no control over whether the tax returns are prepared correctly-

34. In the case of the POM Program, Block is providing protection against its own errors, not errors committed by the taxpayer or a third party.

35. The additional fee paid by Block’s customers for the POM Program is not identified in the POM sales and marketing documents as a premium; it does not vary from person to person, but instead remains constant regardless of the complexity of the return.

36. Block does not utilize any type of underwriting process to evaluate the likelihood of an error.

(Formatting in original.)

Both the State and Block also mention some additional pertinent facts in their briefs on appeal. We quote from the State’s brief, although the parties’ respective wording is almost identical with regard to these facts:

Block will review the taxpayer’s return within thirty (30) days of its filing to ensure its accuracy. If Block finds an error, it will notify the taxpayer and amend the return at no cost to the taxpayer. The taxpayer is entitled to representation by Block at an audit as long as the taxpayer completes a Power of Attorney form, which is required by the Internal Revenue Service (“IRS”). The Power of Attorney form gives a Block agent legal authority to represent the taxpayer before the IRS ... [including the authority to] represent the taxpayer at meetings with the IRS without the taxpayer’s attendance.

The briefs then proceed to discuss Block’s promised representation of the taxpayer in more detail. Both the Commissioner and the trial court viewed this representation as the “service” element of the POM Program. As will be seen, we disagree with that conclusion, and under our analysis, the details of the representation are not relevant to the resolution of this case.

The State first notified Block of its opinion that the POM program constitutes insurance in a letter from the Assistant Commissioner on May 1, 2002. Although it disagreed with the State’s position, Block agreed to stop offering the POM program in its standard offices until the dispute was resolved. (It continued offering the program in its premium offices, where there is no additional fee for the program.) The State and Block then proceeded to exchange letters throughout the summer and fall of 2002, arguing back and forth about the status of the standard-office POM program. Finally, on December 30, 2002, Block filed a petition with the Division of Insurance asking for either a declaratory order declaring that the POM program is not insurance or a formal hearing on the issue. The Division scheduled a hearing before an Administrative Law Judge (“the ALJ”), which was held on September 24, 2003. The ALJ issued an Initial Order on January 7, 2004, declaring the POM program to be insurance. Block appealed that order, and oral argument was held before Commissioner Paula Flowers on July 21, 2004. Commissioner Flowers (“the Commissioner”) issued her Final Order on September 20, 2004, denying Block’s appeal and again declaring the POM program to be insurance. “Block must, therefore, comply with all applicable legal requirements to offer contracts of insurance in Tennessee should it desire to continue offering its ‘Peace of Mind’ Program,” the Commissioner wrote.

Block filed a petition for judicial review on November 9, 2004, under the Uniform Administrative Procedures Act (“UAPA”), codified at Tenn.Code Ann. § 4-5-322 (2005). The case was not heard, however, until August 2006. In the interim, the legislature passed, and the governor signed, an amendment to T.C.A. § 56-7-101 (Supp.2007) that clarifies the insurance status of tax-preparation guarantees, such as the POM program, as follows:

A contract entered into between a tax preparation service company and a taxpayer, providing for the tax preparation service company to pay the additional tax liability, penalties or interest imposed by a taxing authority on the taxpayer as a result of an error of the tax preparation service, shall not be deemed to constitute a contract of insurance, as long as the tax preparation service has secured, on a form approved by the commissioner, a surety bond from an insurance company licensed in Tennessee for a penal sum in an amount to be determined by the commissioner, which amount shall be not less than one hundred thousand dollars ($100,000), but not more than five hundred thousand dollars ($500,000), with respect to the statewide operations of such tax preparation ser vice and its franchisees engaged in the tax preparation business. In the alternative, the commissioner may accept a deposit of cash or securities in such amount. This bond or deposit shall be subject to suit by the state and by any person who has a cause of action arising from a contract subject to this subdivision [ ].

T.C.A. § 56 — 7—101(b)(3). However, no claim is made that the amendment was retroactive, and therefore this case proceeded under the law as it existed prior to the amendment.

After oral argument on August 11, 2006, the trial court entered an order on February 14, 2007, affirming the Commissioner’s decision that the POM program constitutes insurance under the pre-amendment definition of insurance in TenmCode Ann. § 56-7-101. Block appeals.

II.

Because this case originated with an agency ruling that was appealed under the UAPA to the court below, our usual standard of review as stated in Tenn. R.App. P. 13(d) does not apply. Instead, we look to Tenn.Code Ann. § 4-5-322(h), which sets out the standard by which agency decisions are to be reviewed at both the trial and appellate levels. That section states as follows:

The court may affirm the decision of the agency or remand the case for further proceedings. The court may reverse or modify the decision if the rights of the petitioner have been prejudiced because the administrative findings, inferences, conclusions or decisions are:

(1) In violation of constitutional or statutory provisions;

(2) In excess of the statutory authority of the agency;

(3) Made upon unlawful procedure;

(4) Arbitrary or capricious or characterized by abuse of discretion or clearly unwarranted exercise of discretion; or

(5)(A) Unsupported by evidence that is both substantial and material in the light of the entire record.

(B) In determining the substantiality of evidence, the court shall take into account whatever in the record fairly detracts from its weight, but the court shall not substitute its judgment for that of the agency as to the weight of the evidence on questions of fact.

Our initial task is to “determine whether the trial court properly applied” the above-quoted standard of review. Waste Mgmt., Inc. of Tennessee v. Solid Waste Region Bd. of Metro. Gov’t, No. M2005-01197-COA-R3-CV, 2007 WL 1094131, at *3 (Tenn.Ct.App. M.S., filed April 11, 2007).

The trial court deemed the central question in this case — how Tennessee’s definition of insurance applies to Block’s POM Program — to be a question of law, and thus subject to de novo review even under T.C.A. § 4-5-322(h). As will be seen, we ágree. However, despite prevailing in the trial court, the State now urges us to adopt a more restrictive standard of review than the one used by the trial court. The State argues that “considerable deference” should attach to the Commissioner’s ruling-more deference than the trial court afforded it, even though the trial court upheld the Commissioner’s conclusion. Block responds that no deference is owed to the Commissioner’s ruling. We must settle this threshold issue before we can continue with our analysis of the substance of the case.

The State argues in its brief that we should overturn the Commissioner’s decision only if it is “not supported by substantial and material evidence” and is therefore “arbitrary and capricious.” Jackson Mobilphone Co. v. Tennessee Pub. Serv. Comm’n, 876 S.W.2d 106, 110 (Tenn.Ct.App.1993). However, that is a standard for reviewing an administrative agency’s factual judgments. The facts in the instant case are undisputed.

Our review focuses entirely on construing a statute and applying that statute to undisputed facts. Both of these are questions of law. As a general rule, no presumption of correctness attaches when rulings on such questions are appealed. See Davidson v. Lewis Bros. Bakery, 227 S.W.3d 17, 19 (Tenn.2007) (“We review questions of law, including issues of statutory construction, de novo without a presumption of correctness.”); Lawrence County Educ. Ass’n v. Lawrence County Bd. of Educ., 244 S.W.3d 302, 309 (Tenn., 2007) (“If the material facts are undisputed, the interpretation and application of legislative enactments present questions of law unaccompanied by any presumption of correctness.”). However, the State suggests that a different standard should apply where, as here, we are reviewing an agency’s statutory construction and application, rather than a lower court’s.

The Supreme Court has stated that “courts will give full consideration to the ... interpretation of [a legislative] act by administrative boards whose duty it is to enforce it.” Moto-Pep, Inc. v. McGoldrick, 202 Tenn. 119, 303 S.W.2d 326, 330 (1957). More recently, the court opined that “interpretations of statutes by administrative agencies are customarily given respect and accorded deference by courts.” Riggs v. Burson, 941 S.W.2d 44, 51 (Tenn.1997). The Court of Appeals in Profill Dev., Inc. v. Dills, 960 S.W.2d 17, 27 (Tenn.Ct.App.1997), stated that administrative agencies’ decisions “concerning the applicability of technical terms of the statute are entitled to deference” where the agency “has the knowledge, expertise and experience” needed to administer such “technical details.”

However, the Supreme Court has also stated that an Administrative Law Judge’s “construction of [a] statute and application of the law to the facts is a question of law.” Beare Co. v. Tennessee Dept. of Revenue, 858 S.W.2d 906, 907 (Tenn.1993). See also Sanifill of Tennessee, Inc. v. Tennessee Solid Waste Disposal Control Bd., 907 S.W.2d 807, 810 (Tenn.1995) (acknowledging the “substantial and material evidence standard” with regard to factual issues, but stating that statutory construction and application are “question[s] of law that may be addressed by the courts”). This court elaborated on that rule in 2002, declaring that

[j]udicial review of an agency’s construction of a statute and its application of the statute to the facts of the case is a determination involving a question of law. Accordingly, this Court’s review of such matters is de novo without any presumption of correctness.

Jones v. Bureau of TennCare, 94 S.W.3d 495, 501 (Tenn.Ct.App.2002) (citation omitted; emphasis added). Further, Jones specifically distinguished between construction and application of a statute, on the one hand, and “an agency’s interpretation of its own rules and regulations,” on the other. With regard to the latter category, “courts afford deference and controlling weight to such determinations unless plainly erroneous or inconsistent with the regulation.” Id. The instant case, however, involves a statute, not an administrative regulation, and thus neither “deference” nor “controlling weight” should be afforded to an agency interpretation, according to Jones.

A 1976 Supreme Court case suggests the proper resolution of this apparent conflict. In Nashville Mobilphone Co. v. At kins, 536 S.W.2d 385, 340 (Tenn.1976), the court stated:

The Public Service Commission and Mo-bilphone ... insist that the Public Service Commission has interpreted subsection (f) and, citing authorities, they urge upon us the general rule that weight and importance are given by the Tennessee courts to the interpretation of the agency charged with the enforcement or administration of a particular act. We agree that such an interpretation is entitled to consideration and respect and should be awarded appropriate weight, and this is particularly true in the interpretation of doubtful or ambiguous statutes. But administrative interpretation is not controlling and the ultimate determination addresses itself to the courts. Where the court determines that “such construction is erroneous ” it will be “impelled to depart from it.”

(Emphases added; citation omitted.) A recent Court of Appeals case further elucidated the rule: ‘While the courts will give careful consideration to [an agency’s] interpretation of statutes, especially when their interpretation sheds light on legislative intent, statutory construction remains a question of law. Accordingly, the courts review an agency’s interpretation of statutes without a presumption of correctness.” Kidd v. Jarvis Drilling, Inc., No. M2004-00973-COA-R3-CV, 2006 WL 344755, at *4 (Tenn.Ct.App. M.S., filed February 14, 2006) (citation omitted; emphasis added).

Based on our review of these cases, we believe the proper standard of review on these facts 1 is as stated in Kidd and Nashville Mobilphone. The Commissioner’s ruling is “entitled to consideration and respect,” but not necessarily to “deference.” The ruling is neither controlling nor presumed correct. If we find error in either the Commissioner’s interpretation of the statute or her application of the statute to the case’s undisputed facts, we will be “impelled to depart from it.” This rule is consistent with Tenn.Code Ann. § 4-5-322(h), which states that an agency’s decision may be reversed or modified “if the rights of the petitioner have been prejudiced because the administrative findings, inferences, conclusions or decisions are ... [i]n violation of constitutional or statutory provisions” or “characterized by abuse of discretion or clearly unwarranted exercise of discretion.” An error of law is an abuse of discretion “by definition.” Koon v. United States, 518 U.S. 81, 100, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996). See also Weeks v. Corbitt, No. E1999-02698-COA-R3-CV, 2000 WL 782009, at *2 (Tenn.Ct.App. E.S., filed June 20, 2000) (“If the trial court [or in this case, the Commissioner] misconstrues or misapplies the law, its discretion lacks the necessary legal foundation and becomes an abuse of discretion.”).

Accordingly, we decline the State’s invitation to grant the Commissioner’s ruling a greater degree of deference than the trial court did. As noted earlier, the trial court reviewed the Commissioner’s decision de novo, so we take no issue with the court’s ruling on those grounds. However, our inquiry does not end there. As will soon become clear, we believe the trial court— like the Commissioner — committed an error of law in its analysis. Although it has been stated that “the drafters of the Uniform Administrative Procedures Act did not intend that there be a broad, de novo appellate review of a trial court’s disposition of a petition for review,” Waste Mgmt., 2007 WL 1094131, at *3 (summarizing Metro.Gov’t v. Shacklett, 554 S.W.2d 601, 604 (Tenn.1977)), this limitation on appellate review under the UAPA is aimed primarily at the breadth of our factual review, and does not restrict our ability to correct errors of law, whether they occur at the agency level or at the trial court level (or, as in this case, both).

We therefore proceed to analyze the legal merits of the decisions below, mindful of our obligation to afford “consideration and respect” to the agency’s decision, and further to avoid “affording] any broader or more comprehensive review [of the factual record made before the administrative agency] ... than [was] afforded to them by the trial court in the first instance.” Metro. Gov’t, 554 S.W.2d at 604.

III.

A.

The Commissioner and the trial court reached the same conclusion about the central issue in this case, as both found that the POM program constitutes insurance, but their rationales are quite different. The Commissioner declared the § 56-7-101 definition unambiguous, but nevertheless employed a common-law standard, the “service-indemnity test,” to clarify the statute’s meaning. The trial court, on the other hand, essentially declared that the statute speaks for itself and disregarded the service-indemnity test, thus potentially creating an even broader definition of insurance than that contemplated by the Commissioner.

The Commissioner’s conclusions of law are as follows, in pertinent part:

Tenn.Code Ann. § 56-7-101 defines a contract of insurance. Pursuant to this definition, “[a] contract of insurance is an agreement by which one party, for a consideration, promises to pay money or its equivalent, or to do some act of value to the assured, upon the destruction or injury, loss or damage of something in which the other party has an insurable interest.Tenn.Code Ann. § 56-7-101(a) (emphasis added). The statute is not ambiguous or somehow contradicted by another statute, and therefore must be applied in accordance with the natural and ordinary meaning of the terms contained therein. Bryant v. Genco Stamping & Mfg. Co., Inc., 33 S.W.3d 761, 765 (Tenn.2000).

In exchange for the payment of an additional fee by the taxpayer beyond the standard cost for tax return preparation services, Block promises to pay up to five thousand dollars ($5,000) of the taxpayer’s tax liability, and Block also promises to — or rather, assumes total discretion to — appear before, and challenge the findings of, the Internal Revenue Service if the taxpayer’s legal liability is found to exist due to Block’s error. The taxpayer has an insurable interest at stake to protect against his/her legal liability for payment in additional taxes to the Internal Revenue Service if Block erred in preparation of the tax return. The possibility of economic loss to the taxpayer for payment of additional taxes is indemnified by Block up to $5,000 in exchange for the additional fee paid by the taxpayer to participate in the POM program. Thus, in light of the above, the POM program would appear to be a contract of insurance under Tenn.Code Ann. § 56-7-101 (a).

Tennessee’s case law is consistent with this conclusion, as the Tennessee Supreme Court has held that a contract of insurance is created if the contract (here, payment of an additional fee by the taxpayer to obtain the protections of the POM program) is contingent upon the interest (here, the taxpayer’s legal liability for payment of additional taxes); or the principle [sic] object and purpose of the contract is indemnification of risk. Citizens Ins. Co. v. Ayers, 88 Tenn. 728, 13 S.W. 1090, 1091 (1890).

Petitioner’s reliance upon Tennessee Attorney General Opinion No. 85-038, is misplaced. Whereas an automobile parts warranty is a promise of service (i.e. — to replace any automobile parts which do not perform satisfactorily), the sole assurance provided through purchase of the POM program by the taxpayer is indemnification for up to five thousand dollars ($5,000) of additional taxes. The supposed “service” feature of the POM program is minimal at best, and perhaps wholly illusory. Providing representation to affirm the validity of the tax return before the Internal Revenue Service cannot be considered the primary, or even a substantially valuable, feature of the POM program for the taxpayer; indeed, control of this defense of the tax return must be surrendered to Block by the terms of the agreement, the purpose of such almost certainly being to allow Block to minimize it’s [sic] own potential liability to the taxpayer.

It is therefore CONCLUDED that Block’s Peace of Mind program is a contract of insurance in the State of Tennessee. Block must, therefore, comply with all applicable legal requirements to offer contracts of insurance in Tennessee should it desire to continue offering its “Peace of Mind” program.

(Formatting in original.) As can be seen, the Commissioner fundamentally bases her ruling on what is known as the “service-indemnity test,” which balances the respective importance of a contract’s service and indemnity components to determine whether the contract constitutes insurance. The Commissioner held that the POM program is insurance because its service element, in the judgment of the Commissioner, is not its “primary, or even a substantially valuable, feature” — in other words, because the indemnity portion of the contract predominates over the service portion. Intriguingly, the Commissioner based her decision on the service-indemnity test despite declaring TenmCode Ann. § 56-7-101, which contains no mention of any such test, to be unambiguous.

The trial court went further than the Commissioner, implying that a contract may be insurance even where the indemnity element does not predominate over service. The opinion is a clear rejection of the service-indemnity test that the Commissioner had relied upon. The court declared that it “does not find it necessary to search for tools such as criteria from other states, to review or balance the service versus indemnity components in this contract.” According to the court, the indemnity element only needs to be “significant to the contract” and “not trifling or de minimis.” If “the indemnity provision would affect or influence the customer’s decision to purchase the POM Program,” that is enough to make the program insurance, according to the court.

The court acknowledged that its ruling goes further even than the State asked it to go. Both parties, it said, are “assuming that not all contracts for services will be regulated as insurance in this state,” but the court apparently begs to differ. It writes that “a straightforward application of the statutory definition” leads to the conclusion that it “arguably encompasses all contracts for future services after a contingent loss or damage occurs.” The court says this sweepingly broad interpretation — which could vastly expand the common understanding of what constitutes insurance — is simply the consequence of faithfulness to the literal text of the statute in question: “The Court is not empowered to balance one significant term against another given the straightforward definition in Tennessee’s law.”

B.

Before we can assess the merits of either analysis, we must first look to the statute in question. It states as follows, in pertinent part:

A contract of insurance is an agreement by which one party, for a consideration, promises to pay money or its equivalent, or to do some act of value to the assured, upon the destruction or injury, loss or damage of something in which the other party has an insurable interest; and it is unlawful for any company to make any contract of insurance upon or concerning any property or interests or lives in this state, or with any resident thereof, or for any person, as insurance agent or insurance broker, to make, negotiate, solicit, or in any manner aid in the transaction of such insurance, unless and except as authorized under the provisions of this title[.]

TenmCode Ann. § 56-7-101(a). In 2005, as noted earlier, a subsection was added to § 56-7-101 specifically addressing tax preparation guarantees such as this one. However, it took effect after the events of this case, and retroactivity is not alleged. Therefore our statutory analysis is limited to the above-quoted language broadly defining an insurance contract.

“[T]he intention of the legislature is to be learned from the words it has used; ... and, if that intention is expressed in a manner devoid of contradiction and ambiguity, there is no room for interpretation or construction, and the judges are not at liberty, on consideration of policy or hardship, to depart from the words of the statute.” Austin v. Memphis Pub. Co., 655 S.W.2d 146, 148 (Tenn.1983) (quoting Heiskell v. Lowe, 126 Tenn. 475, 153 S.W. 284, 290 (1912)). The threshold question for us to answer, therefore, is whether the statute in question is “devoid of ... ambiguity.” We disagree with both the trial court and the Commissioner on this point. In our view, the statute’s definition is not only exceedingly broad — arguably encompassing many types of future-services contracts that are not generally considered insurance, such as extended product warranties offered by retailers — but it is also inherently circular. Included in its definition of a “contract of insurance” is a requirement that the contract cover “something in which the other party has an insurable interest.” TenmCode Ann. § 56-7-101(a) (emphasis added). The statute does not specify what types of interest are “insurable,” which is surely a crucial question in determining what constitutes “insurance.”

Tennessee courts have previously found statutes to be ambiguous because of circular language. For example, in 1993 the Supreme Court declared unconstitutionally vague a statutory definition of “excess violence” as “the depiction of acts of violence in such a graphic and/or bloody manner as to exceed common limits of custom and candor, or in such a manner that it is apparent that the predominant appeal of the material is portrayal of violence for violence’s sake.” Davis-Kidd Booksellers, Inc. v. McWherter, 866 S.W.2d 520, 532 (Tenn.1993) (quoting Tenn.Code Ann. § 39-17-901(4) (1991)). The court specifically noted that this definition “ends with the circular phrase of ‘violence for violence’s sake.’ ” Id. at 531. Similarly, in a 2001 case concerning the statutory definition of “equine activity,” this court wrote:

Equine activities include “[rjides, trips, hunts, or other equine activities of any type, however informal or impromptu, that are sponsored by an equine activity sponsor.” Tenn.Code Ann. § 44-20-102(3)(E). This rather circular definition conveys more than one meaning. Arguably, it could include any activity involving an equine. It could also be construed less broadly because the General Assembly may not have intended to grant qualified immunity to a tortfeasor whenever the tortious activity somehow involves an equine.... Because of these two possible constructions of Tenn.Code Ann. § 44-20-102(3)(E), we turn to other familiar canons of statutory construction.

Friedli v. Kerr, No. M1999-02810-COA-R9-CV, 2001 WL 177184, at *4 (Tenn.Ct.App. M.S., filed February 23, 2001) (emphasis added; footnotes omitted). In a footnote accompanying the word “circular,” the court reiterated that “[t]he term ‘equine activities’ forms part of the definition of ‘equine activity.’ ” Id., n. 10.

In the instant case, “insurable interest” is a key component of the definition of “insurance.” Because the statute does not clarify what it means by “insurable interest,” we believe this circular language renders the definition of insurance sufficiently ambiguous that we, like the court in Fried-li, must “turn to other familiar canons of statutory construction” to ascertain the proper interpretation of the definition.

C.

“Our role in construing statutes is to ascertain and carry out the legislative intent without unduly restricting or expanding the statute’s coverage beyond its intended scope.” State v. Sliger, 846 S.W.2d 262, 263 (Tenn.1993). “Statutory terms draw their meaning from the context of the entire statute, and from the statute’s general purpose. We give these words their natural and ordinary meaning!.]” BellSouth Telecomm., Inc. v. Greer, 972 S.W.2d 663, 673 (Tenn.Ct.App.1997) (citations omitted). A statute should be read “without any forced or subtle construction which would extend or limit its meaning.” Nat’l Gas Distribs., Inc. v. State, 804 S.W.2d 66, 67 (Tenn.1991). “Moreover, we will not apply a particular interpretation to a statute if that interpretation would yield an absurd result.” State v. Flemming, 19 S.W.3d 195, 197 (Tenn.2000).

“The legislative process does not always produce precisely drawn laws. When the words of a statute are ambiguous or when it is just not clear what the legislature had in mind, courts may look beyond a statute’s text for rehable guides to the statute’s meaning.” BellSouth, 972 S.W.2d at 673. Such “reliable guides” can include the statute’s historical background and legislative history. In the instant case, however, we are presented with no such historical information. 2 We assume this is a consequence of the statute’s ancient origins. The definition in question was apparently drafted sometime in the 19th century. A 1911 case, American Sur. Co. of New York v. Folk, 124 Tenn. 139, 135 S.W. 778, 779 (1911), quotes language identical to the current statute, with the exception of an extra comma after the word “damage,” as coming from “Chapter 160, § 2, Acts of 1895.” Id.

Given the statute’s long history, one might presume that prior cases would provide helpful guidance in interpreting the definition and applying it to the facts of this case. Yet we have been unable to find any such guidance in the case law. Although prior cases interpreting the definition of insurance do exist, their contexts are so different as to be generally unhelpful. This case tests the outer limits of TenmCode Ann. § 56-7-101 in a way that no prior Tennessee case has done. That in itself is arguably significant; it appears the courts of this state have never before been presented, in the 113 or more years of this statute’s existence, with the claim that a program such as this one should be regulated as insurance.

On the other hand, the Tennessee Attorney General was presented with a similar claim — and his opinion is now cited by Block as being averse to the State’s current position. In 1986, the Department of Commerce and Insurance — the same state department involved in the instant case— inquired as to whether the Attorney General’s office was of the opinion that “extended warranties and service contracts ... fall within the definition of ‘insurance’ in Tenn.Code Ann. § 56-7-101.” Tenn. Op. Atty. Gen. No. 85-038, 1986 WL 222674, at *1 (February 19,1986). Specifically, the Attorney General was asked about automobile warranties offered by new car dealers, and also about automobile warranties offered by “an independent company apparently not affiliated with the dealer or manufacturer.” The Attorney General opined that neither contract would constitute insurance under TenmCode Ann. § 56-7-101 because both would fail the service-indemnity test. We will consider this Attorney General opinion in more detail later.

The 1986 opinion was not the first time the Attorney General had been asked to weigh in on the statutory definition of insurance. In 1979, the Attorney General opined, “The definition of insurance [in what is now Tenn.Code Ann. § 56-7-101] clearly contemplates that provision of future services constitutes insurance. Still, it cannot be the case that every contract for future services is one of insurance.” Tenn. Op. Atty. Gen. No. 79-254, 1979 WL 33863, *3 (May 23, 1979) (emphasis added). Instead of deeming all future-services contracts as insurance, the Attorney General wrote, it is “appropriate to consider the elements of a contract which mark it as one of insurance.” The opinion further stated that indemnity and contingency are “essential elements of insurance.” Id. The Attorney General reaffirmed this stance in 1981, writing:

As pointed out in a previous opinion issued from this office, [the statute’s] exceedingly broad definition might be read to include all contracts for future services. As also pointed out, however, all arrangements which would fall literally within this definition should not perforce be treated as insurance in legal contexts. It is necessary to determine whether the basic elements of insurance — contingency and indemnity — are dominant in the particular contract.

Tenn. Op. Atty. Gen. No. 81-068, 1981 WL 142731, at *1 (January 30, 1981) (citations omitted; emphasis added).

The State’s position in the instant case appears to be a retreat from those opinions about the definition of insurance under Tenn.Code Ann. § 56-7-101. In its brief, the State argues that “the plain meaning of the statute clearly permits the Commissioner to identify the POM Pro gram as a contract of insurance.” Yet the aforementioned line of Attorney General opinions makes clear that the State historically has not regarded any and all future-services contracts falling under the literal ambit of § 56-7-101’s “plain meaning” as therefore necessarily constituting insurance.

“Although opinions of the Attorney General are not binding on courts, government officials rely upon them for guidance; therefore, [such] opinion[s are] entitled to considerable deference.” State v. Black, 897 S.W.2d 680, 688 (Tenn.1995). Thus, for instance, the Court of Criminal Appeals has held that a statute challenged as facially overbroad may survive where the Attorney General has adopted a limiting construction. State v. Burkhart, No. 01C01-9804-CC-00174, 1999 WL 1096051, at *6 (Tenn.Crim.App., filed December 6, 1999). See also Scott v. Ashland Healthcare Ctr., Inc., 49 S.W.3d 281, 287 (Tenn.2001) (agreeing with Attorney General’s interpretation of what legislature intended in enacting nursing-home licensing law); Ferrell v. Cigna Prop. & Cas. Ins. Co., 33 S.W.3d 731, 737-38 (Tenn.2000) (agreeing with Attorney General’s interpretation of the definition of “necessary” in statute relating to judicial absence and appointment of substitute judge).

Attorney General opinions are particularly persuasive when they have been consistently repeated. Thus, in Tennessee Bd. of Dispensing Opticians v. Eyear Corp., 218 Tenn. 60, 400 S.W.2d 734 (1966), the Supreme Court declined to follow a single Attorney General opinion that had not been “the consistent interpretation of the statute by the State,” but opined that “[w]here a consistent interpretation has been given such should be given persuasive weight by this Court.” Id. at 739. See also Murfreesboro Bank & Trust Co. v. Evans, 193 Tenn. 34, 241 S.W.2d 862, 864 (1951) (stating that, where an Attorney General opinion had been “consistently followed by the Department [for nearly a decade], [it] should be given persuasive weight by the Court”).

We also believe that Attorney General opinions are especially persuasive where, as here, the State is a litigant arguing for a more expansive statutory interpretation that it had previously disavowed — particularly where that interpretation is one that has apparently never been asserted in any recorded case since the disputed language was written, more than a century ago.

In any event, we agree with the Attorney General that the statute’s “exceedingly broad definition might be read to include all contracts for future services,” yet such a reading would be absurd; “it cannot be the case that every contract for future services is one of insurance.” In light of the circular reference to “insurance” and “insurable interest,” it seems clear that the statute implicitly incorporates some common-sense notion of what sort of contract can constitute insurance, and therefore it cannot have been the legislature’s intent to include such contracts as simple product warranties offered by a retailer, which do not pass such a common-sense test.

D.

The problem, of course, is defining that “common-sense test” more precisely, and then applying it to the facts of this case. A good place to start is the very test relied upon — albeit applied incorrectly, in our view — by the Commissioner: the service-indemnity test.

The first appearance of the service-indemnity test in Tennessee apparently occurred in the aforementioned 1986 Attorney General opinion regarding whether automobile warranties can constitute in surance. 3 Tenn. Op. Atty. Gen. No. 85-038, 1986 WL 222674. In it, the Attorney General quotes a California case for the proposition that “[t]he critical question is whether indemnity is a ‘significant financial proportion of the business,’ as opposed to service.” Id. at ⅜2 (quoting People ex rel. Roddis v. California Mut. Ass’n, 68 Cal.2d 677, 68 Cal.Rptr. 585, 441 P.2d 97, 101 (1968)). We agree with the Attorney General that this test is an excellent common-sense gauge of whether or not a contract is insurance. As will be seen, however, we disagree with the Attorney General’s application of the test.

The Attorney General in 1986 analyzed the service-indemnity question in terms of whether the benefit that the company provides to the customer is a “service” or pure “indemnity” — i.e., money. “[IJndemnity is not a significant object” of the warranties in question, according to the Attorney General, because the warranties are

designed to provide servvce to the purchaser. Either plan provides that when a covered automobile part malfunctions, the purchaser can get it replaced or fixed at no charge except the deductible. If he does not get the part replaced or fixed, he does not receive a payment. This is the essence of service as opposed to indemnity.

Id. at *3 (emphases added).

The Attorney General’s narrow focus on what the contract “provide[s] ... to the purchaser” — “service” or “payment” — is mirrored by the Commissioner’s judgment in the instant case. Directly contrasting this case to the 1986 opinion, she wrote that

[wjhereas an automobile parts warranty is a promise of service (i.e.-to replace any automobile parts which do not perform satisfactorily), the sole assurance provided through purchase of the POM program by the taxpayer is indemnification for up to five thousand dollars ($5,000) of additional taxes.

Put more simply, when the customer makes a successful claim under his automobile warranty, he gets service; when the customer makes a successful claim under the POM program, he gets money, or the benefit of a monetary payment to the tax authorities.

We believe this analysis misses the point entirely. As stated in the California case quoted by the Attorney General in 1986, the focus should be on what “proportion of the business ” indemnity occupies, “in the context of the plan as a whole.” California Mut. Assoc., 68 Cal.Rptr. 585, 441 P.2d at 101 (emphasis added). That is a far broader question than the one asked and answered by the Commissioner, namely what proportion of the direct benefit to the customer is indemnity as opposed to service.

The nature of the benefit to the customer — service or money — certainly should not be the dispositive issue in this case. In fact, under the terms of the statute, it is arguably not relevant at all. Tenn.Code Ann. § 56-7-101 states that insurance involves the payment of “money or its equivalent, or ... some act of value.” In other words, it is possible to have an insurance contract where the benefit to the customer is purely service, not money. But is the converse also true? Can a future-services contract not be insurance even if the benefit to the customer is purely (or predominantly) money, not service? It seems to us that the answer must be yes; otherwise, it would be impossible for a company without an insurance license to warranty the product it sells in any case where that product inherently involves money. Why, for instance, should it be possible for a company to provide a simple warranty for automotive services, but not for financial services? 4

In any event, the Commissioner clearly committed an error of law in her analysis on this point. Where she writes that “[t]he supposed ‘service’ feature of the POM program is minimal at best, and perhaps wholly illusory,” she is referring to the fact that, in the event of a claim under the POM program, Block provides customers with legal representation in any audit proceedings. The Commissioner may be right that this aspect of Block’s guarantee is so self-serving as to be minimal or even illusory. But, again, this conclusion misses the point. The question is not whether the “payout” is service-based or indemnity-based. The question is whether the contract, as a whole, is primarily a service guarantee or a promise of indemnity.

E.

The service-indemnity test, properly applied, looks not just at the nature of the benefit, but at the core essence of the contract itself: is it a contract of service, or a contract of indemnity? Moreover, in applying the test, it is important to “look[ ] behind the technical aspects of the contracts involved to the realities of the business.” Garrett v. Forest Lawn Mem’l Gardens, Inc., 505 S.W.2d 705, 710 (Tenn.1974). Both the Commissioner and the trial court failed to follow that rule when they chose to analyze the POM program as if it were an independent entity, unrelated to Block’s broader tax-preparation services. Such an approach makes no sense. The POM program, which guarantees the accuracy of Block’s tax-preparation services, is inextricably linked to those services. The fact that a customer desiring the POM guarantee must pay a separate fee to Block, above and beyond the normal tax-preparation fee that it pays to Block, does not divorce the POM program from its broader context — namely, Block’s preparation of the customer’s tax returns. If Block were not providing tax services, there would be nothing for its POM program to guarantee. Indeed, the stipulated facts specify that a customer cannot purchase the POM guarantee for taxes prepared by someone else; the program is available only to customers who have their taxes prepared by Block. Furthermore, the customer’s “separate” payments to Block — the tax-preparation fee and the POM fee — must in fact be simultaneous: customers must purchase the POM program at the same time that they purchase the tax-preparation services. These characteristics make the POM program similar to a typical extended product warranty at an electronics retailer: both are optional add-ons, both cost extra money, yet neither can sensibly be viewed as independent of the underlying purchase that they guarantee, because they can only be obtained in connection with that larger purchase. 5

The argument that Block’s POM program cannot sensibly be viewed as independent from Block’s tax-preparation services is buttressed by the fact that the POM program only sometimes costs extra, and the State appears to implicitly concede that it is not insurance in those cases where there is no extra cost. The POM fee, which the State contends is an insurance premium, is waived for customers at Block’s “premium” offices; those customers pay more for their tax preparation in exchange for a variety of perks — including getting the POM program for no additional charge. The State does not attempt to argue that the POM program constitutes insurance when it is offered at the “premium” offices; only at the “standard” offices, where it costs extra, is it insurance, according to the State. But why should there be any difference between paying more for tax preparation and getting a “free” guarantee, versus paying two separate fees to the same company, one for tax preparation and one for the guarantee of that preparation? This is a distinction without a difference, and again, it fails Garrett’s directive to “look[] behind the technical aspects of the contracts involved to the realities of the business.” Id. The notion that this program is insurance, but would cease to be insurance if Block were to simply eliminate the POM fee and raise the cost of tax preparation for all customers, is logically indefensible. The difference between insurance and non-insurance cannot possibly hinge on such a flimsy distinction.

Block’s POM program falls on the “service” side of the service-indemnity test not because Block provides its customers with the “service” of legal representation in the event they are audited, but because the core essence of the program is that of a tax-preparation service with an added guarantee. The stipulated facts concede this point: “The POM Program is ancillary to and supports Block’s core business of tax preparation.” Moreover, it seems to us intuitively obvious that the POM program exists not just to increase Block’s profits from existing customers, but to entice new customers to hire Block for their tax-preparation needs. As the program’s name suggests, Block is selling “peace of mind,” not just to existing customers but to prospective ones as well. The message is clear: hire us to do your taxes, and for a small additional fee, we’ll give you a guarantee that nobody else can. We’ll promise that you can rely on our work — no unhappy surprises. As the stipulated facts put it:

Through the POM Program, Block stands behind the accuracy of its tax preparation services. When the tax return is completed, the customer walks away with the assurance that her tax return is accurate and that the refund or liability reflected on the return is correct. The customer may very well alter her financial position based on the information reflected on the return.

Thus, the POM program protects “the reasonable reliance by the customer that no ... mistake will be made [by the tax preparer].”

The fee schedule bears out that the POM program is indeed “ancillary” to the tax-preparation service — in other words, that service, not indemnity, is the predominant feature of the package Block sells to customers who purchase its tax-preparation services and the POM guarantee. The average tax preparation fee is $119, while the POM program costs $27, or less than 23 percent of the cost of the service. Clearly, most of Block’s revenue is coming from the tax-preparation service. The POM program produces profits in its own right — as do product warranties — but this does not make it a separate entity from the underlying service.

The State points out that the POM program is supplemental to Block’s basic guarantee, but that fact is not dispositive. Extended product warranties often supplement manufacturer’s warranties — and for an additional and separate fee, just like POM. The mere fact that customers can opt for a lesser degree of protection, at a lesser cost, does not make the POM program independent of the service it guarantees. If the guarantee was offered by an entity independent of H & R Block — ie., if a third party offered to indemnify Block customers in the event of a mistake by Block-this would be an entirely different question, because then the guarantee would truly be independent of the tax-preparation service. Similarly, if the guarantee covered more than just Block’s own errors in preparing the customer’s taxes, it might potentially make more sense to treat it as insurance. But as the stipulated facts point out:

The POM Program is sold only for the current tax year and at the time Block’s tax preparation services are rendered .... If the tax return is prepared by someone other than Block, the POM Program is not available.

[[Image here]]

Errors in the preparation of a tax return are not uncontrollable events such as an act of God or other external contingency. Whether Block makes an error is completely within the control of the Block tax preparer. Block’s obligations under the POM Program are not triggered by any external contingency.

[[Image here]]

Block actually provides the tax preparation services which are covered by the POM Program. The POM Program is not being provided by an outside company that has no control over whether the tax returns are prepared correctly.... In the case of the POM Program, Block is providing protection against its own errors, not errors committed by the taxpayer or a third party.

The element of contingency, central to any common-sense definition of insurance, is missing where the customer is “insured” only against losses caused directly by the “insurer.” This would be like a car-insurance company only covering accidents that are caused by its own employees. Less absurdly, it is similar to a manufacturer’s warranty, which protects the buyer against defects, but not against damage caused by the buyer or by third parties.

As an additional ground for finding the element of contingency absent, it is worth noting that the “loss” suffered by the customer is not really a “loss” at all, at least not as that term is normally defined for insurance purposes. The purported “loss” is simply the sum of additional tax liability above and beyond what the customer expected to pay — but the amount of one’s tax liability is not a contingency. Generally speaking, an individual’s tax liability is fixed from the moment the clock strikes midnight on January 1. H & R Block does not have the power to reduce or expand a customer’s tax liability, only to calculate it accurately or inaccurately. An inaccurate calculation may lead to newly discovered tax liability, but not to new tax liability. Thus, an error by Block does not create an insurable “loss,” but simply an unfulfilled expectation.

In any event, we believe the most sensible analytical framework is the holistic service-indemnity test described herein. Of course, even under this analysis, the POM program could still be designated as insurance if we were to accept the trial court’s holding that the indemnity element need only be “significant to the contract” and “not trifling or de minimis.” After all, paraphrasing from the court’s ruling, it is certainly true that “the [POM Program cjould affect or influence the customer’s decision to purchase the [tax-preparation service].” However, the same could be said of any extended warranty — or, for that matter, a basic manufacturer’s warranty- — influencing a customer’s decision to buy, for example, a camcorder or a plasma TV. Likewise, a customer might decide not to take his car to a particular repair shop if the shop does not offer a warranty for its work. In this sense, any guarantee associated with a contract for the sale of goods or services is likely to be “significant to the contract.” Are all such guarantees now to be considered insurance? Under the trial court’s “significant to the contract” test, as modified by our more expansive understanding of “service” under the service-indemnity test, the answer would seem to be yes. Yet we do not think it could have been the legislature’s intent to sweep so broadly when it defined insurance. After 113 years in which the courts have never been asked to treat simple warranties as insurance, we can find no legal, logical or public-policy justification for adopting a definition so expansive that it would do precisely that.

IV.

For all of the reasons stated above, we reverse the judgment below and find in favor of Block. Under Tennessee law as it existed when this controversy began, the POM program was not subject to insurance regulation under T.C.A. § 56-7-101. Costs on appeal and at the trial court level are taxed to the State of Tennessee, Department of Commerce and Insurance, Division of Insurance. This case is remanded to the trial court for collection of costs at that level, pursuant to applicable law.

1 . We do not here address whether a more deferential standard applies to an agency’s interpretation of "technical details” in statutes where the agency is better suited than the courts to understand the technicalities in question. We do not regard the definition of insurance under Tenn.Code Ann. § 56-7-101 as a "technical detail.”

2 . The closest thing to a "legislative history” argument in either party’s brief is Block’s contention that the legislature’s subsequent amendment of Tenn.Code Ann. § 56-7-101, which was apparently passed in direct response to the Commissioner’s ruling in this case, indicates that the legislature intended all along to exclude tax contracts such as this one from the ambit of the statute. According to Block, the amendment was a mere clarification of that original intent. We reject this argument. A statutory amendment may be a clarification of original intent, or it may be a modification indicating that the original intent was quite the opposite. We do not have adequate information to make a judgment as to which category this amendment falls into, and we therefore do not consider the amendment at all in our analysis of this issue.

3 . As noted earlier, the Attorney General was asked about two types of automobile warranties: those offered by new car dealers, and those offered by "an independent company apparently not affiliated with the dealer or manufacturer.” The Attorney General concluded that neither type of warranty would constitute insurance under the facts presented. Whether the warranty was offered by the seller or by a third party was not seen as dispositive. “We are of the opinion that this distinction should not control whether a given plan falls within the definition of insurance,” the Attorney General wrote. That statement does not necessarily contradict our analysis in the instant case, because the factual situations are quite different. However, to the extent that the Attorney General may have been saying that the seller/third-party distinction can never be a crucial or controlling factor in determining whether a future-services contract constitutes insurance, we disagree. In the context of the facts presented by the instant case, we believe the seller/third-party distinction is extremely important.

4 . Indeed, if any future-services contract whose payout is primarily monetary must therefore be regarded as insurance, it seems that even Block’s "Basic Guarantee” — which promises to pay any interest and penalties that arise because of an error-might well be considered insurance too.

5 . The main difference is that, at the electronics store, the purchase being guaranteed is a good, whereas at H & R Block, the purchase being guaranteed is a service. Yet we see no principled difference between warranting goods and warranting services. Moreover, there are plenty of examples of non-insurance warranties where the warranted product is a service, automotive repair warranties being perhaps the most obvious.