Justice ENOCH

delivered the opinion of the Court.

The question is whether the Federal Insecticide, Fungicide, and Rodenticide *23 Act 1 (“FIFRA”) and its implementing regulations preempt state common-law damage claims for crop damage — specifically, damage to peanut plants. Ordinarily in federal preemption cases, we are asked to decide the scope of preemption, whether congressionally expressed or implied. But in this case, it is not the breadth of preemption, but rather the breadth of the congressionally created exception to express preemption that determines the outcome.

As we explain, Congress has dictated that state actions regarding product labeling are preempted to the extent that the content of the product label is regulated. But Congress has permitted its regulatory body, the Environmental Protection Agency, to choose NOT to regulate product labeling with respect to how well a product works, that is, the product’s “efficacy.” Central to this case, the EPA has also chosen to define product efficacy to include “target area phytotoxicity,” that is, the effect of a particular product or combination of products on the crops that are deliberately sprayed. Simply put, the EPA does not regulate herbicide labels regarding how well a product works, and this includes if the product actually injures the crops it was intended to assist. Because of the EPA’s choice not to regulate, and therefore because there are no labeling or packaging requirements regarding crop damage imposed under FIFRA, we conclude that state common-law claims about target area crop damage are not preempted. Thus, the Geyes’ claims are not preempted.

Terry Geye and his son Brandon are peanut farmers. The summary judgment evidence shows that in 1993 they treated part of their peanut crop with a mixture of the herbicides Pursuit and Prowl, both of which American Cyanamid manufactures. In selecting the Pursuit-Prowl combination, the Geyes claim they relied on various labels and advertisements that specifically stated that Pursuit could be “tank mixed” with Prowl. The advertisements also stated that Pursuit was a sound choice for crop safety and that it does not cause injury to peanut plants. But the Geyes allege that applying the Pursuit-Prowl mix to their fields actually injured their peanut plants. The Geyes assert that the Pur-suiL-Prowl mix stunted root growth and inhibited foliage development which resulted in a 3,000-pound per acre reduction in crop yield.

The Geyes sued American Cyanamid. They alleged breach of express and implied warranties, strict liability, and violation of the Texas Deceptive Trade Practices Act. American Cyanamid filed a motion for summary judgment, asserting that FIFRA preempted the claims. The trial court agreed and dismissed the claims. The court of appeals reversed, holding that FIFRA did not preempt the Geyes’ crop-damage claim. 2 We affirm the judgment of the court of appeals.

Under the Supremacy Clause of the United States Constitution, the laws of the United States are “the supreme Law of the Land; ... any Thing in the Constitution or Laws of any State to the Contrary notwithstanding.” 3 “A state law is preempted and ‘without effect’ if it conflicts with federal law.” 4 Congressional *24 intent determines whether a federal statute preempts state law. 5 Preemption maybe determined by the express provisions provided by Congress. 6 It may also be implied if the statute’s scope indicates that Congress intended federal law to occupy the field, or when state law actually conflicts with federal law. 7 Finally, preemption based on an actual conflict may still exist even if the claim is not expressly preempted under the relevant statute. 8

This is an express preemption case. FIFRA contains an express preemption clause that provides that a “State shall not impose or continue in effect any requirements for labeling or packaging in addition to or different from those required under [FIFRA].” 9 Preemption of state law claims under this provision is linked to the labeling and packaging requirements imposed by FIFRA. Under this Act, Congress has given the EPA the role of evaluating and determining the content of pesticide labels. 10

The Administrator is authorized ... to prescribe regulations to carry out the provisions of this Act. Such regulations shall take into account the difference in concept and usage between various classes of pesticides, including public health pesticides, and differences in environmental risk and the appropriate data for evaluating such risk between agricultural, nonagricultural, and public health pesticides. 11

Consequently, the EPA regulations define “the domain expressly pre-empted.” 12 Because the EPA’s labeling requirements determine what state actions are preempted, we cannot know whether the Geyes’ crop-damage claim is preempted until we determine what the EPA requires for product labels concerning crop safety.

Ill

FIFRA is a comprehensive federal statute regulating pesticide use, sales, and labeling, and granting enforcement authority to the EPA. 13 The Act provides a detailed scheme for regulating the content and format for labeling herbicides. Under FI-FRA, all herbicides sold in the United States must be registered with the EPA. 14 Each manufacturer must submit to the EPA a statement that includes a “complete copy of the labeling of the pesticide, a statement of all claims to be made for it, and any directions for its use.” 15 Each manufacturer must also submit “the complete formula of the pesticide” 16 and “a full description of the tests made and the re- *25 suits thereof upon which the claims are based.” 17 After evaluating this information, the EPA then registers a product that “performfs] its intended function without unreasonable adverse effects on the environment.” 18

But despite FIFRA’s comprehensive nature, Congress authorized the EPA in 1978 to choose not to require the submission of data relating to the “efficacy” of products. 19 “Efficacy” refers to how well a product works or, as defined by the EPA, to the “product’s ability to control the specific target pest or produce the specified plant or animal response when the product is applied in accordance with the label directions ..., precautions, and limitations of use.” 20 Congress said: “In considering an application for the registration of a pesticide, the [EPA] Administrator may waive data requirements pertaining to efficacy, in which event the Administrator may register the pesticide without determining that the pesticide’s composition is such as to warrant proposed claims of efficacy.” 21 The EPA, acting under this authorization, has chosen not to collect efficacy data for any products except in a specific set of circumstances that are not relevant here. 22 Essential to this case, the EPA has also chosen to include “target area phytotoxicity” within the concept of efficacy, and therefore the EPA has chosen not to collect data concerning “target area phytotoxicity.”

“Target area” refers to the “area intentionally treated with a pesticide when label use directions are followed [i.e., the peanut field].” 23 The EPA further defines “target area plants” as “all plants located within the target area, and includes both desirable and undesirable species [i.e., the peanuts and the weeds].” 24 “Target area phytotoxicity” then describes the toxic effect to both desirable and undesirable plants within the target area. The Geyes’ peanuts are desirable plants that are “target area plants” and are within the “target area.”

The fact that the EPA has chosen not to collect data concerning target area phyto-toxicity is found principally in two places. First, the data table found in 40 C.F.R. § 158.540 identifies the data that manufacturers must submit regarding plant protection. The table limits the submission of target area phytotoxicity to “Special Review” circumstances under 40 C.F.R. § 154.1 and in “certain public health situations.” 25 American Cyanamid does not as *26 sert that either of these reviews occurred in this case.

Second, the EPA has authored a set of Pesticide Assessment Guidelines that “contain the standards for conducting acceptable tests, guidance on evaluation and reporting of data, definition of terms, further guidance on when data are required, and examples of acceptable protocols.” 26 Specifically, Pesticide Assessment Guidelines, Subdivision J: Hazard Evaluation Non-target Plants provides:

(d) Target area phytotoxicity testing waiver of requirements.

(1) The Administrator has determined that efficacy test data include target area phytotoxicity testing data, and that data submittal for such testing may be waived, by his authority under FIFRA Sec. 3(e)(5)[136a(c)(5) ] for most kinds of pesticide products.... Such products generally include all pesticides whose uses result in direct or indirect application to plants in the target area such as agricultural, lawn, and garden use.

(2) Even though the Administrator will ordinarily waive the requirement for submittal of target area phytotoxicity test data as indicated in paragraph (b)(1) of this section, he reserves the authority to require such data on a case-by-case basis whenever the Administrator deems that such data are necessary to evaluate the acceptability of a product for registration. If it is determined that data phytotoxicity for a pesticide are necessary, the Agency will promulgate the specific target area phytotoxicity data requirements by letter to a specific registrant or by general notice. 27

As well, the EPA has limited the submission of data for the specific problem that the Geyes complain about — tank mixtures:

(5) Tank Mixtures. When tank mixtures are recommended on product labeling, a study may be required on a case-by-case basis to demonstrate the extent of antagonism and synergism with respect to detrimental effects on nontarget plants by the products of tank mixtures. 28

The EPA did not evaluate whether the tank-mixed PursuiG-Prowl combination had a toxic effect on peanut plants. It therefore could not evaluate the label claims made by American Cyanamid which stated that the combination was safe to use on peanut plants. It thus exercised no regulatory authority over American Cyan-amid’s label that specifically allowed the tank mixing of Pursuit and Prowl.

As a federal regulation, Section 158.540 is entitled to deference. 29 This regulation establishes that manufacturers submit target area phytotoxicity data only when the pesticide is subject to “Special Review” 30 or in “certain public health situations.” 31 Nothing in the record suggests that the EPA evaluated the specific claims made on the label stating that Pursuit could be tank mixed with Prowl.

While Subdivision J is not entitled to the same deference as an agency regulation, it may have the “power to per *27 suade.” 32 The EPA’s choice not to evaluate efficacy claims has been consistent. In a proposed rule, first announced in 1979, the Agency explained that the choice “stemmed from a need to reduce the amount of resources devoted to reviewing product performance so that additional effort could be devoted to the evaluation of health and safety data.” 33 In choosing not to evaluate the data regarding efficacy, the EPA relied on the marketplace to ensure that products were effective and did not “impart any detrimental effects (particularly on crops, ornamentals and other desirable plants) for which they could be liable.” 34 The EPA’s reliance on the marketplace found its way into Subdivision J: “The Agency has determined that target area phytotoxicity data does not need to be submitted because the registrants are generally willing to accept the overall responsibility of the product [with] respect to efficacy and phytotoxicity.” 35

The information contained in the As sessment Guideline was originally published by the EPA as a proposed rule on November 3, 1980. 36 The proposed rule was to be designated as a regulation to be published in Section 40 C.F.R. §§ 163.120-1 to 163.120-4. The specific choice not to collect data on target area phytotoxicity now found in Subdivision J was in this proposed rule. As a proposed rule, it was subject to public comment and consideration. 37 In 1981, it appears that the EPA decided not to codify Section 162.120 as a regulation and instead proposed that the information contained in proposed Section 162.120 become Subdivision. 38 According to the EPA, this was to “limit the regulation to a concise presentation of the data requirements and when they are required.” 39 The EPA adopted this approach, and the current regulation specifically references the Assessment Guidelines. 40 It appears that the Assessment Guidelines are essentially the EPA’s directions to pesticide manufacturers on what tests to conduct, how to conduct those tests, and what data is required for registration.

In sum, the EPA has had a relatively consistent approach to not collecting data on target area phytotoxicity since at least 1980. Including target area phytotoxicity within the concept of efficacy reflects the conclusion that a product will not be considered effective by consumers if it harms the plants that it was intended to assist. It also reflects the EPA’s assumption that if a manufacturer states that a product is safe to use on peanut plants, that manufacturer has made sure that the product is in fact “safe for use on peanut plants.” Finally, the fact that the information in Subdivision J was subject to public disclosure *28 and comment is an important factor in determining the level of deference it should receive. 41 Subdivision J persuades us that the EPA does not regulate label claims about target area phytotoxicity.

In concluding that the Geyes’ claims were not preempted, the court of appeals relied on an opinion letter issued by the EPA, Pesticide Regulation (PR) Notice 96-4. 42 American Cyanamid takes issue with the court of appeals’ reliance on the opinion letter. PR Notice 96-4 describes the product labeling process and details the EPA’s choice not to collect efficacy data. PR Notice 96-4 essentially concludes that state law actions should not be preempted because the EPA does not determine if a product will be “efficacious or will not damage crops or cause other property damage.” 43 The court of appeals did substantially rely on PR Notice 96-4: “Our decision is based upon the information contained in [Notice 96-4 ]> and we conclude that we should quote extensively from the EPA regulation.” 44 American Cyanamid argues that PR Notice 96-4 is a “narrowly focused legal brief thinly disguised as an EPA guidance document; it is by no means a ‘regulation.’ ” We agree with American Cyanamid. PR Notice 96-4 is not a regulation and the court of appeals erred in its substantial reliance on the Notice.

But the EPA’s choice not to collect target area phytotoxicity data, as found in 40 C.F.R. § 158.540 and the EPA’s Assessment Guideline, Subdivision J, Hazard Evaluation Nontarget Plants convinces us that the conclusion reached by the court of appeals was correct. The combination of these two sources demonstrates that the EPA does not regulate product labeling with respect to a product’s target area phytotoxicity. Thus the Geyes’ crop-damage claim is not preempted.

We acknowledge that many jurisdictions that have considered this issue have reached the opposite result. 45 But all these cases are based on the premise that the EPA regulates whether products are toxic to target area crops. For example, the California Supreme Court considered the efficacy waiver in a case factually similar to ours — Etcheverry v. Tri-Ag Service, Inc. 46

The plaintiffs in Etcheverry claimed that applying two products, Guthion and Mor-estan, reduced its walnut production, resulting in about $150,000 in damages. 47 The California Supreme Court held that the plaintiffs’ crop-damage claims were preempted by FIFRA, concluding that the EPA’s waiver of efficacy data was “beside the point” and “irrelevant” 48 because the plaintiffs were complaining about phyto- *29 toxicity and not efficacy. But in so holding, the California Supreme Court did not draw the essential distinction between phytotoxicity and target area phytotoxicity that the EPA historically has.

Likewise in Taylor AG Industries v. Pure-Gro, 49 the Ninth Circuit did not consider the EPA’s decision to not collect target area phytotoxicity data. The court described the EPA review as “rigorous” and that “the EPA approves each label only after a careful review of the product data and the draft label.” 50 Of course, that may be true for phytotoxicity generally, but it is not true for target area phyto-toxicity. With respect to target area phy-totoxicity, the EPA makes no review. And without that review and approval process, there certainly can be no federal regulation. And without EPA regulation, there can be no preemption.

In Quest Chemical Corp. v. Elam, 51 we considered FIFRA’s preemptive effect in the context of a personal injury suit. We reversed the court of appeals’ judgment that allowed the plaintiffs strict liability and implied warranty claims to escape federal preemption. We stated, “FIFRA preempts all common law tort suits against manufacturers of EPA-registered pesticides which are based solely upon claims relating directly or indirectly to labeling.” 52 Quest does not control this case. The fundamental difference is that Quest involved a claim for personal injury while this case involves a claim for target crop damage. The EPA continues to collect data concerning safety to humans. Regulating labeling concerning these health issues remains one of the Agency’s primary mandates. After all, the EPA used concern for health and safety as one of the justifications for not evaluating efficacy claims:

The decision to [not collect and review] efficacy [data] as an Agency policy stemmed from a need to reduce the amount of resources devoted to reviewing product performance so that additional effort could be devoted to the evaluation of health and safety data, and from a desire to reduce regulatory burdens in pesticide registration. 53

This case is unlike many cases in the preemption area where we struggle to determine the breadth of federal preemption expressed in a federal statute. For here we must determine the breadth, not of the preemption itself, but of an exception to that preemption founded on Congressional authorization for the EPA to specifically choose to not collect efficacy data. Acting on that authorization, the EPA has chosen not to evaluate whether a product will be toxic to the crops it was intended to assist. Because the EPA does not evaluate whether a product will be toxic to the crops that it was intended to assist, the EPA does not regulate a product’s labeling claims on this subject. Because the scope of FIFRA’s preemption is dependent on what the EPA regulates, FIFRA does not preempt the Geyes’ common law crop-damage claim. For these reasons we affirm the court of appeals’ judgment.

. 7 U.S.C. §§ 136-136y.

. 32 S.W.3d 916, 921.

. U.S. Const, art. VI, cl. 2.

. Hyundai Motor Co. v. Alvarado, 974 S.W.2d 1, 4 (Tex.1998) (quoting Maryland v. Louisiana, 451 U.S. 725, 746, 101 S.Ct. 2114, 68 L.Ed.2d 576 (1981)).

. Worthy v. Collagen Corp., 967 S.W.2d 360, 366-67 (Tex. 1998) ("We are ... bound to give effect to the will of Congress.”).

. See, e.g., Cipollone v. Liggett Group, Inc., 505 U.S. 504, 517, 112 S.Ct. 2608, 120 L.Ed.2d 407 (1992).

. See, e.g., Freightliner Corp. v. Myrick, 514 U.S. 280, 287, 115 S.Ct. 1483, 131 L.Ed.2d 385 (1995).

. See Geier v. American Honda Motor Co. Inc., 529 U.S. 861, 884, 120 S.Ct. 1913, 146 L.Ed.2d 914 (2000); see also Great Dane Trailers, Inc. v. Estate of Wells, 52 S.W.3d 737, 741 (Tex.2001).

. 7 U.S.C. § 136v(b).

. See id. § 136w(a)(l).

. Id.

. Cipollone, 505 U.S. at 517, 112 S.Ct. 2608.

. Wisconsin Pub. Intervenor v. Mortier, 501 U.S. 597, 601, 111 S.Ct. 2476, 115 L.Ed.2d 532 (1991).

. 7 U.S.C. § 136a(a).

. Id. § 136a(c)(l)(C).

. Id. § 136a(c)(l)(D).

. Id. § 136a(c)(l)(F).

. Id. § 136a(c)(5)(C).

. See id. § 136a(c)(5).

. Bernard A. Schneider, EPA, Pesticide Assessment Guidelines, Subdivision G: Product Performance 36 (1982) (available from the National Technical Information Service pursuant to 40 C.F.R. § 158.108).

. 7U.S.C. § 136a(c)(5).

. See 40 C.F.R. § 158.640(b)(1) (waiving efficacy data "unless the pesticide product bears a claim to control pest microorganisms that póse a threat to human health and whose presence cannot readily be observed by the user”); 40 C.F.R. § 158.540(b)(1) (waiving target area phytotoxicity data unless required for "Special Review and certain public health situations”).

. Robert W. Holst & Thomas C. Ellwanger, EPA, Pesticide Assessment Guidelines, Subdivision J: Hazard Evaluation Nontarget Plants 18 (1982) (available from the National Technical Information Service pursuant to 40 C.F.R. § 158.108).

. Id. (emphasis added).

. 40 C.F.R. § 158.540(b)(1).

. 40 C.F.R. § 158.108.

. Subdivision J at 15.

. Subdivision J at 29 (emphasis added).

. See Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-44, 104 S.Ct 2778, 81 L.Ed.2d 694 (1984).

. 40 C.F.R. § 154.1.

. 40 C.F.R. § 158.540(b)(1).

. Skidmore v. Swift & Co., 323 U.S. 134, 140, 65 S.Ct. 161, 89 L.Ed. 124 (1944).

. Interim Final Regulation Relating to Conditional Registration, 44 Fed.Reg. 27932, 27938 (May 11, 1979) (to be codified at 40 C.F.R. pt. 162.).

. Proposed Data Requirements, 47 Fed.Reg. 53192, 53196 (Nov. 24, 1982) (to be codified at 40 C.F.R. pt. 158).

. Subdivision J at 7.

. Proposed Guidelines for Registering Pesticides in the United States: Subpart J: Hazard Evaluation: Nontarget Plants and Microorganisms, 45 Fed.Reg. 72948, 72961 (Nov. 3, 1980) (to be codified at 40 C.F.R. pts. 162, 172).

. See Subdivision J at 1-13 (discussing public comment received on proposed regulation).

. Proposed Data Requirements, 47 Fed.Reg. at 53192.

. Id.

. 40 C.F.R. § 158.108.

. See Christensen v. Harris County, 529 U.S. 576, 586-87, 120 S.Ct. 1655, 146 L.Ed.2d 621 (2000).

. Pesticide Regulation (PR) Notice 96-4 (June 3, 1996), available at http:// www.epa.gov/opppmsdl/PR Notices/pr96-4.html.

. Id.

. 32 S.W.3d at 918.

. See, e.g., Andrus v. AgrEvo USA Co., 178 F.3d 395 (5th Cir.1999); Kuiper v. American Cyanamid Co., 131 F.3d 656 (7th Cir.1997); Taylor AG Indus, v. Pure-Gro, 54 F.3d 555 (9th Cir.1995); Worm v. American Cyanamid Co., 5 F.3d 744 (4th Cir.1993).

. 22 Cal.4th 316, 93 Cal.Rptr.2d 36, 993 P.2d 366 (2000).

. Id. at 368.

. Id. at 375.

. 54 F.3d at 560.

. Id.

. 898 S.W.2d 819, 820 (Tex.1995).

. Id.

. Interim Final Regulation Relating to Conditional Registration, 44 Fed.Reg. at 27938.