OPINION OF THE COURT

BECKER, Chief Judge.

Appellant Manny Baker, a Pennsylvania state parolee, was arrested while leaving the parole office for violating the condition of his parole that required him to refrain from driving. Parole officers searched the passenger compartment of the borrowed car that Baker had driven to the parole office. They also searched the trunk of *440 the car and discovered what they believed to be drug paraphernalia. On the basis of what they found in the trunk, the officers searched Baker’s home, where they found weapons and 66 grams of heroin. A federal grand jury indicted Baker for possessing with intent to deliver a controlled substance (heroin) in violation of 21 U.S.C. § 841(a)(1), and for violating the statutes prohibiting felons from possessing firearms, 18 U.S.C. §§ 922(g)(1) & 924(a). Before trial, Baker moved to suppress the evidence seized from his home, but the District Court denied his suppression motion. Baker proceeded to trial, where he was convicted of violating §§ 841(a)(1) and 922(g)(1) and sentenced to almost twenty years imprisonment on weapons and drugs charges.

This appeal which, as we shall explain below, is now in its third discrete phase, raises an important first-impression question: whether the standard Pennsylvania Board of Probation and Parole consent to search form, signed by Baker as a condition of his parole, authorized suspicionless searches of his person, property, and residence. Before reaching this question, we will explain our previous conclusions — memorialized in an unrequited certification to the Pennsylvania Supreme Court — that Baker had standing to object to the search of the vehicle, which he did not own, and that the parole officers lacked reasonable suspicion to search the trunk of that car. These holdings are the predicate for the question whether the consent to search form authorized suspicionless searches because, if Baker lacked standing to object to the search, or if the officers had reasonable suspicion, we could dispose of the case without construing the consent form.

The consent to search form provided:

I expressly consent to the search of my person, property and residence, without a warrant by agents of the Pennsylvania Board of Probation and Parole. Any items, in possession of which constitutes a violation of the parole/reparole shall be subject to seizure, and may be used as evidence in the parole revocation process.

If as a matter of Pennsylvania law the standard consent to search form implies a requirement that parole officers have reasonable suspicion in order to conduct a search of a parolee, the evidence against Baker must be suppressed and his conviction set aside. Because we conclude that the form should be so construed (or more precisely, predict that the Supreme Court of Pennsylvania would so construe it), the order of the District Court denying Baker’s motion to suppress will be reversed.

In August 1996, Baker drove to the state parole office in Philadelphia for a scheduled visit with his parole officer. As a condition of his parole, Baker had agreed not to drive without a license. During the parole visit, a parole agent asked Baker if he had a driver’s license, and Baker responded that he did not. When Baker attempted to drive away after his visit, he was arrested by parole officers for violating this condition of parole.

After Baker was arrested, parole officers searched the passenger compartment and the glove compartment of the ear that Baker had been driving and discovered that the car was registered in someone else’s name. The officers could not figure out how to open the trunk of the car. One of them asked Baker, then in custody at the parole office, how to open the trunk. Baker explained that the engine must be turned on to open the trunk and gave the officer the keys to the trunk. 1 Once the *441 parole officers got into the trunk, they found what they suspected was drug paraphernalia: “several screw top glass vials, oils, empty clear plastic lunch bags and taller incense bags.” The officers then conducted a warrantless search of Baker’s home, which yielded numerous weapons and sixty-six grams of heroin. As we have noted, the District Court found that this search was justified by the reasonable suspicion that arose from the items discovered in the car and denied Baker’s motion to suppress evidence. This appeal followed.

In a prior opinion, United States v. Baker, No. 97-1977 (3d Cir. Jan. 7, 1999), we ordered a limited remand for the District Court to consider whether Baker had standing to challenge the search of the car and whether there was reasonable suspicion to justify the search of the trunk of the car. On remand, United States v. Baker, No. 97-00297, 1999 WL 163631 (E.D.Pa. Mar.17, 1999), the District Court concluded that Baker had standing and that the parole officers had reasonable suspicion to search the trunk of the car. In the wake of the District Court’s opinion, the appeal presented three questions: (1) whether Baker had standing to challenge the search of the car; (2) whether there was reasonable suspicion to search the trunk of the car that Baker drove to the parole office; and (3) if there was not reasonable suspicion, whether the search of the trunk was still valid because of the consent form that Baker signed upon his release on parole. 2

Following receipt of supplemental briefing and additional argument, the panel conferred and concluded that: (1) Baker had standing to challenge the search of the car; and (2) that the police lacked reasonable suspicion to search the trunk of the car. On February 18, 2000, we certified to the Pennsylvania Supreme Court (pursuant to Pennsylvania Supreme Court Order No. 197, Judicial Administration, Docket No. 1, filed Jan. 12, 2000) the remaining— and vexing — question whether, under Pennsylvania law, the consent form authorized suspicionless searches.

The Supreme Court did not accept our petition for certification, leaving us to determine whether, under that Court’s opinions in Scott v. Pennsylvania Bd. of Probation & Parole, 548 Pa. 418, 698 A.2d 32 (1997), rev’d on other grounds, 524 U.S. 357, 118 S.Ct. 2014, 141 L.Ed.2d 344 (1998), and Commonwealth v. Williams, 547 Pa. 577, 692 A.2d 1031 (1997), the consent to search form used by the Pennsylvania Board of Probation and Parole, as a matter of Pennsylvania law, authorizes suspicionless searches or implies a condition that any search conducted pursuant to the consent form be founded on reasonable suspicion.

II.

The threshold question presented by these facts is that of standing. Standing to challenge a search requires that the individual challenging the search have a reasonable expectation of privacy in the property searched, see Rakas v. Illinois, 439 U.S. 128, 143, 99 S.Ct. 421, 58 L.Ed.2d 387 (1978), and that he manifest a subjective expectation of privacy in the property searched, see California v. Greenwood, 486 U.S. 35, 39, 108 S.Ct. 1625, 100 L.Ed.2d 30 (1988).

It is clear that a passenger in a car that he neither owns nor leases typical *442 ly has no standing to challenge a search of the car. See Rakas, 439 U.S. at 133-34, 99 S.Ct. 421 (holding that there is no legitimacy to a defendant’s expectations of privacy where the area searched is in the control of a third party). “Fourth Amendment rights are personal rights, which, like some other constitutional rights, may not be vicariously asserted.” Id. at 133-34, 99 S.Ct. 421 (quoting Alderman v. United States, 394 U.S. 165, 174, 89 S.Ct. 961, 22 L.Ed.2d 176 (1969)), “a person who is aggrieved by an illegal search and seizure only through the introduction of damaging evidence secured by a search of a third person’s premises or property has not had any of his Fourth Amendment rights infringed.” Id. at 134, 99 S.Ct. 421.

For these reasons, we have previously suggested that a defendant who had stolen a car and used it in a robbery would not have standing to object to a search of the car. See United States v. Yeager, 448 F.2d 74, 85 (3d Cir.1971) (rejecting challenge to search on basis that “if[the defendant’s] theories were valid, a stolen car used in a robbery could not be searched and objects therein seized by the police without a search warrant”). We have never considered, however, whether an individual who borrows a car and has control over it has a legitimate expectation of privacy in it.

Cases from other circuits suggest that whether the driver of a car has the reasonable expectation of privacy necessary to show Fourth Amendment standing is a fact-bound question dependent on the strength of his interest in the car and the nature of his control over it; ownership is not necessary. Compare United States v. Cooper, 133 F.3d 1394, 1398-99 (11th Cir.1998) (driver of a rental car whose contract to rent the car had expired four days before the search had a reasonable expectation of privacy in the car because he could have extended the contract with a simple phone call); United States v. Angulo-Fernandez, 53 F.3d 1177, 1179 (10th Cir.1995) (driver who was able to produce registration papers in the name of the person from whom he claimed to have borrowed the car had standing); United States v. Rubio-Rivera, 917 F.2d 1271, 1275 (10th Cir.1990) (permission from the owner to use a vehicle supports privacy expectation therein); United States v. Garcia, 897 F.2d 1413, 1417-18 (7th Cir.1990) (driver using vehicle with the permission of an absent owner has a reasonable expectation of privacy therein); with United States v. Padilla, 111 F.3d 685, 687 (9th Cir.1997) (defendants lacked standing to object to the search of car in which they had only a temporary “bailment interest”); United States v. Riazco, 91 F.3d 752, 755 (5th Cir.1996) (defendant lacked standing to object to the search of a rental car in a case in which his name was not on the rental agreement, the rental agreement had expired, and the defendant did not have permission to drive the car from the person who rented the car); United States v. Wellons, 32 F.3d 117, 119 (4th Cir.1994) (driver of a rental car did not have standing to contest the search thereof because he was not listed as an authorized driver on the rental contract); United States v. Ponce, 947 F.2d 646, 649 (2d Cir.1991) (defendant must show legitimate basis for possessing a car, such as permission from the car owner, to have standing); United States v. Sanchez, 943 F.2d 110, 113-14 (1st Cir.1991) (because defendant failed either to show that he had the owner’s permission to use the car or to demonstrate prior use or control of the car, the circumstances tipped in favor of denying the motion to suppress).

Baker is asserting his own, not a third party’s, expectation of privacy. He came alone in the car to the parole office. Although he did not own the car, he had substantial control over it insofar as he had borrowed it from a friend and had been driving it for four to six weeks. He carried the keys to the car with him into the parole office. Although the defendant and his associates were somewhat vague about who owned the car, there is no evidence in the record that the car was stolen or that Baker knowingly possessed a stolen car. All of these factors lead to *443 the conclusion that Baker had a reasonable expectation of privacy in the car. In Angulo-Fernandez, 53 F.3d at 1179, the Court of Appeals for the Tenth Circuit rejected the argument that the defendant lacked standing because the registered owner, from whom a defendant claimed to have borrowed the car, denied ownership. The court held that “[t]he officer’s testimony established that Mr. Angulo Fernandez had claimed to have borrowed the car from the rightful owner and had produced a registration bearing the owner’s name. Although such evidence may not be determinative of the Defendant’s right to possess the car, absent evidence to the contrary, it is sufficient to meet his burden of demonstrating Fourth Amendment standing.” Id. Similarly, in Garcia, 897 F.2d at 1418, the Court of Appeals for the Seventh Circuit held that a driver who borrowed a car (but was unable to provide the last name of the owner) had standing to object to the search because “[i]f an individual has the owner’s permission to use property, society surely recognizes this as reasonable.” (citation omitted).

We conclude that a discrepancy between an individual’s statement regarding the owner of the car he is driving, and the identity of the owner of the car as reflected by the title and registration, is not enough, by itself, to destroy the driver’s reasonable expectation of privacy when there is clear evidence of continuing possession and control, as well as no evidence that the driver obtained the car illegitimately. Under the circumstances, therefore, Baker had the requisite legitimate expectation of privacy to support standing for Fourth Amendment purposes.

III.

Given that Baker has standing to challenge the search, the next question is whether the District Court erred in concluding that the parole agent had reasonable suspicion to search the trunk of the car. Though officers may lawfully search the passenger compartment of the car incident to arrest, see New York v. Belton, 453 U.S. 454, 460, 101 S.Ct. 2860, 69 L.Ed.2d 768 (1981) (holding that, when a policeman has made a lawful custodial arrest of the occupant of an automobile, he may, as a contemporaneous incident of that arrest, search the passenger compartment of that automobile), such a search incident to arrest does not extend to the trunk of the car, see id. at 460-61 & n. 4, 101 S.Ct. 2860. Thus, in order for the search of the trunk to comport with the Fourth Amendment, it had to have been supported by the level of suspicion required under the Fourth Amendment.

Under normal circumstances, the Fourth Amendment requires government officials to have both probable cause and a warrant to conduct a search. In the case of parolees, however, the requisite level of suspicion is reduced and a warrant is not required. In Griffin v. Wisconsin, 483 U.S. 868, 871-72, 107 S.Ct. 3164, 97 L.Ed.2d 709 (1987), a Wisconsin statute authorized probation officers to conduct warrantless searches of probationers’ homes when there were “reasonable grounds” to believe that contraband would be found there. The Court found that the operation of a state’s probation system presented “special needs,” beyond the need for law enforcement, justifying an exception to the warrant and probable cause requirements of the Fourth Amendment. See id. at 873-74, 107 S.Ct. 3164. The Court noted:

To a greater or lesser degree, it is always true of probationers (as we have said it to be true of parolees) that they do not enjoy “the absolute liberty to which every citizen is entitled, but only ... conditional liberty properly dependent on observance of special [probation] restrictions.”

Id. at 874, 107 S.Ct. 3164 (quoting Morrissey v. Brewer, 408 U.S. 471, 480, 92 S.Ct. 2593, 33 L.Ed.2d 484 (1972)) (omission and alterations in original). The Court went on to hold that the special needs of the probation system, including the need to supervise probationers, see id. at 875, 107 S.Ct. 3164, justified a lower standard for searches of a probationer’s property, see *444 id. at 878, 107 S.Ct. 3164. Specifically, the Court held that a search could be conducted on the basis of such “reasonable grounds” as information indicating that there might be weapons in a probationer’s home. See id. at 880, 107 S.Ct. 3164.

In United States v. Hill, 967 F.2d 902, 909 (3d Cir.1992), we extended the holding in Griffin to parolees and concluded that a parolee’s car or home can be searched on the basis of reasonable suspicion. alone, even in the absence of an authorizing state statute such as that in Griffin. In Hill, the defendant was arrested for violating parole. His apartment was then searched and two guns were seized. See 967 F.2d at 904-05. We affirmed the district court’s denial of Hill’s motion to suppress the guns, concluding that “parole may be an even more severe restriction on liberty because the parolee has already been adjudged in need of incarceration.” Id. at 909. Quoting Latta v. Fitzharris, 521 F.2d 246, 249 (9th Cir.1975), we noted that “parole authorities have a ‘special and unique interest in invading the privacy of parolees under their supervision.’ ” Hill, 967 F.2d at 910. “[I]t is reasonable to allow a parole officer to search whenever he reasonably believes that it is necessary to perform his duties. The decision to search must be based on ‘specific facts,’ but the officer need not possess probable cause.” See id.; see also Terry v. Ohio, 392 U.S. 1, 21, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968) (holding that reasonable suspicion requires “specific and articulable facts which, taken together with rational inferences from those facts, reasonably warrant the intrusion” of a warrantless search).

In Hill, the officers acted on a report from the parolee’s estranged wife that he had committed several parole violations, including keeping drugs and guns in the home that they jointly owned. See 967 F.2d at 904, 911. We concluded that these facts were specific enough to give rise to reasonable suspicion. We also rejected Hill’s argument that, once he was arrested, any special need justifying a lower standard vanished, concluding that “the parole agents’ ‘interest in inspecting [Hill’s] place of residence did not terminate upon his arrest; if anything, it intensified.’ ” Id. at 911 (quoting Latta, 521 F.2d at 252); see also United States v. Jones, 152 F.3d 680, 686 (7th Cir.1998) (rejecting the argument that “the State’s special supervisory need is diminished when a parolee is in custody”).

Following remand in this case, the District Court concluded that the search of the trunk was lawful because (1) Baker had no driver’s license (which was a violation of his parole); (2) Baker could not produce documentation demonstrating that he owned the car; and (3) the parole officer might reasonably have concluded that Baker “might have been in further violation of his parole (the car might have been stolen) or that the trunk of the ear might contain evidence of a further violation of his parole such as drug paraphernalia.”

We disagree that there was justification to search the trunk. The parole officers’ actions were not based on “specific facts” giving rise to suspicion that there would be some evidence of a further violation of parole in the trunk. In Hill, the parole officers acted pursuant to information regarding specific allegations of parole violations. See Hill, 967 F.2d at 911. Here, by contrast, there were no similar allegations. There is no evidence in the record that the parole agents who searched Baker’s trunk had yet discovered the inconsistencies between the title/registration and Baker’s statements regarding the ownership of the car when they searched the trunk. 3 Even if they had, we do not think that a mere suspicion that a car might be stolen justifies a search of the trunk of that car: *445 There is little reason to think evidence of the car’s rightful owner would be found in the trunk. 4

Similarly, neither Baker’s violation of his parole by driving a vehicle nor his failure to document that he owned the vehicle can give rise to a reasonable suspicion that he was committing other, unspecified, unrelated parole violations — the evidence of which might be found in the trunk. Cf. Knowles v. Iowa, 525 U.S. 113, 118, 119 S.Ct. 484, 142 L.Ed.2d 492 (1998) (“Once [the defendant] was stopped for speeding and issued a citation, all the evidence necessary to prosecute that offense had been obtained. No further evidence of excessive speed was going to be found either on the person of the offender or in the passenger compartment of the car.”).

Reasonable suspicion requires more specificity than these parole officers had in this case. Thus, we are satisfied that the search of the trunk was not supported by reasonable suspicion. The evidence procured in the search of the trunk and the house must therefore be suppressed as the fruits of an illegal search unless the consent form provided a basis for the search.

IV.

The parties agree that the proper construction of the consent form that Baker signed turns on two cases decided by the Pennsylvania Supreme Court, Commonwealth v. Williams, 547 Pa. 577, 692 A.2d 1031 (1997), and Scott v. Pennsylvania Bd. of Probation & Parole, 548 Pa. 418, 698 A.2d 32 (1997), rev’d on other grounds, 524 U.S. 357, 118 S.Ct. 2014, 141 L.Ed.2d 344 (1998). Baker contends that under these precedents the consent form that he signed must be construed as a matter of Pennsylvania law to imply a condition that the parole officers have reasonable suspicion to justify any search of his person, property, or residence. If Baker is correct, the evidence procured in the search of the house must be suppressed because the consent form he signed did not authorize searches in the absence of reasonable suspicion.

The United States disagrees, contending that the Pennsylvania Supreme Court’s resolution of Williams and Scott rested on federal constitutional grounds rather than the proper construction of the Pennsylvania consent form. More specifically, the government submits that these cases hold that an unqualified consent to search executed as a condition of parole can justify a search consistent with the Fourth Amendment only where reasonable suspicion is present, but do not decide whether — under Pennsylvania law — the form should be construed to provide unqualified consent to search. Since we are not bound by Pennsylvania’s interpretation of the Fourth Amendment, this reading of Williams and Scott would leave us free to construe the forms anew.

We agree with the government that the Pennsylvania Supreme Court has never directly construed the form as a matter of Pennsylvania law. Nevertheless, as we shall explain, although we believe that it is arguable that the form waives the reasonable suspicion requirement, Williams and Scott strongly suggest that Pennsylvania would construe the form to preserve the reasonable suspicion requirement.

In Williams, the Pennsylvania Supreme Court reviewed a parolee’s claim that a search conducted pursuant to a consent form identical to the one at issue here violated his Fourth Amendment rights. 5 The Court analyzed the conflicting inter *446 ests involved in the case, the relevant precedents (state and federal, including Griffin), and the approaches taken by other states, and adopted the approach that:

the parolee’s signing of a parole agreement giving his parole officer permission to conduct a warrantless search does not mean either that the parole officer can conduct a search at any time and for any reason or that the parolee relinquishes his Fourth Amendment right to be free from unreasonable searches. Rather, the parolee’s signature acts as acknowledgment that the parole officer has a right to conduct reasonable searches of his residence listed on the parole agreement without a warrant.

Id. at 1036.

Significantly, the approach adopted requires a court to conduct what amounts to a Fourth Amendment analysis when a parole officer conducts a search pursuant to the consent form:

A search [conducted pursuant to the form] will be deemed reasonable if the totality of the evidence demonstrates: (1) that the parole officer had a reasonable suspicion that the parolee had committed a parole violation, and (2) that the search was reasonably related to the parole officer’s duty.

Id. (citation omitted). As such, the Williams case seems to be based on the Pennsylvania Supreme Court’s interpretation of the Fourth Amendment (especially insofar as it purports to adopt the “middle ground” approach adopted by other states, which approach seems simply to require that parolees acknowledge that parole officers can conduct reasonable searches without a warrant, see id. at 1037). Indeed, the Court concluded that this approach “comports with the Fourth Amendment protection afforded to parolees by the United States Supreme Court in Griffin, supra, and in our sister states.” Id. Nevertheless, the Court in Williams concluded that the parole agents had reasonable suspicion to search the defendant, so it did not have to decide whether the form authorized suspicionless searches. See id. at 1037-38.

Four months later, in Scott v. Pennsylvania Bd. of Probation & Parole, 548 Pa. 418, 698 A.2d 32 (1997), rev’d on other grounds, 524 U.S. 357, 118 S.Ct. 2014, 141 L.Ed.2d 344 (1998), the same Court squarely faced the question whether a parolee is entitled to the protections of the Fourth Amendment in his parole revocation hearing when he signed a consent to search form as a condition of his parole. The Scott Court construed its decision in Williams as a decision balancing the interests of the government and parolees. The Court noted that, “[i]n deciding what effect the signing of the consent provision had on Williams’s limited rights under the Fourth Amendment, we ‘balanced] the governmental interests involved in granting parole and supervising parolees with that interest of the private individual, i.e., the parolee, which has been affected by the governmental action.’ ” Scott, 698 A.2d at 35 (1997) (quoting Williams, 692 A.2d at 1035). This passage suggests that the Pennsylvania Supreme Court read its earlier decision as having been based on weighing of the Fourth Amendment interests involved and not the meaning of the form as a matter of Pennsylvania law. 6

*447 At the same time, Scott notes that “applying Williams to the instant case, we hold that Appellee has a Fourth Amendment right against unreasonable searches and seizures that is unaffected by his signing of the consent to search provision.” 698 A.2d at 36. Although the Court does not explain why (ie., whether it is interpreting Pennsylvania law), in our view this statement suggests that signing the form does not effect a waiver of any of the signatory’s rights under the Fourth Amendment.

The problem we have in construing the Williams and Scott decisions stems from the fact that, instead of engaging in a two-stage analysis, asking first whether the form as a matter of Pennsylvania law provided for suspicionless searches or required reasonable suspicion, and second whether the form as construed under Pennsylvania law comported with the requirements of the Fourth Amendment, the Pennsylvania Supreme Court did not go beyond reasoning that under the Fourth Amendment a search of a parolee required reasonable suspicion, and it construed the form in light of that requirement. The Court did not independently analyze the meaning of the form as a matter of Pennsylvania law. 7 Accordingly, we need to decide whether, as a matter of Pennsylvania law, Baker consented to a suspicionless search of his person and property by signing the form.

By its terms, there are (at least) two possible readings of the Pennsylvania standard form. First, it can be understood to authorize suspicionless searches, because there is no caveat modifying the phrase “I expressly consent to the search of my person, property and residence” that would suggest that the background rule of reasonable suspicion applies. This interpretation does make sense of the plain language of the consent form and the common understanding of what it means to consent to a search in the Fourth Amendment context. Cf. Florida v. Jimeno, 500 U.S. 248, 251, 111 S.Ct. 1801, 114 L.Ed.2d 297 (1991) (“The standard for measuring the scope of a suspect’s consent under the Fourth Amendment is that of ‘objective’ reasonableness — what would the typical reasonable person have understood by the exchange between the officer and the suspect?”) (citations omitted); *448 United States v. Kim, 27 F.3d 947, 957 (3d Cir.1994) (“Of course [the defendant] could have limited his consent to certain items, but he had the burden to express that limitation.”) (citing Jimeno, 500 U.S. at 252, 111 S.Ct. 1801). If Pennsylvania had adopted this interpretation, however, the reason for the extended discussion of reasonable suspicion in Williams and Scott would be unclear.

The second construction of the form is that it waives only the warrant requirement. On this reading, the phrase “I expressly consent to the search of my person, property and residence, without a warrant” is limited in the sense that it waives only the warrant requirement and leaves the reasonable suspicion requirement in place. The difficulty with this position is that Griffin (which we have applied to searches of parolees) holds not only that reasonable suspicion (rather than probable cause) is the requisite level of suspicion for searches of probationers, but also that no warrant is required when conducting such searches. See Griffin, 483 U.S. at 876-77, 107 S.Ct. 3164 (holding that the special needs of the probation system made the warrant requirement impracticable); see also Hill, 967 F.2d at 910 (holding that in light of the special needs of a parole system, the warrant requirement does not apply to searches of parolees).

On this interpretation, the consent to search form has no effect — it simply states the legal standard for searching a parolee that would apply even in the absence of the form. It is hard to understand why Pennsylvania would feel the need to require parolees to consent affirmatively to the background rule of warrantless searches, but not to consent affirmatively to the background rule of reasonable suspicion, if the form was intended simply to set forth the governing standards to search parolees. Nevertheless, Williams characterizes the parole agreement as an “acknowledg[ment] that agents of the Parole Board could conduct a warrantless search,” id. at 1037, which supports the view that the consent form effects no waiver and serves only as an acknowledgment of the background rule.

In these circumstances, it is our job to predict what Pennsylvania would do if it were to construe the form solely as a matter of Pennsylvania law. Based on our reading of Pickron, 535 Pa. 241, 634 A.2d 1093 (1993), and Williams, coupled with the tenor of the discussion in Scott, we conclude that Pennsylvania would construe the consent form to include an implicit requirement that any search be based on reasonable suspicion. We begin with Pick-ron, in which the Court decided that, in the absence of consent or a statutory or regulatory scheme authorizing suspicion-less searches of parolees, the Fourth Amendment required reasonable suspicion to search parolees. That decision is consistent with Griffin and Hill. Then in Williams, the Court noted that, following Pickron, the Pennsylvania General Assembly enacted two statutory provisions authorizing searches of a parolee by state and county parole officers based on reasonable suspicion. See Williams, 692 A.2d at 1035 n. 9 (referring to 61 PA. Cons.Stat. §§ 331.27a, 331.27b). As such, Pennsylvania’s statutory scheme adheres to the reasonable suspicion rule.

More significantly, nothing in Scott suggests that the consent form authorizes the suspicionless searches not provided for by the statutory scheme, and at least one statement suggests that the consent form is similarly limited. As we have noted, Scott states that “applying Williams to the instant case, we hold that Appellee has a Fourth Amendment right against unreasonable searches and seizures that is unaffected by his signing of the consent to search provision.” 698 A.2d at 36. But if the form waived the reasonable suspicion requirement, it would affect a parolee’s substantive rights.

When engaging in a predictive exercise, as we do here, we sometimes must not only read the lines but also between the lines, as it were. The Pennsylvania Supreme Court’s consistent application of the *449 Fourth Amendment reasonableness inquiry to the consent form in Williams and Scott, and its repeatedly expressed concern with balancing the interests of the state with the privacy interests of the parolee, also suggest to us that that Court does not believe that the form authorizes suspicionless searches. Rather, the analysis engaged in, which focuses on the reasonableness of the search, would be unnecessary if the form authorized a search not otherwise permitted under the Fourth Amendment. See, e.g., Williams, 692 A.2d at 1036 (“A search will be deemed reasonable if the totality of the evidence demonstrates: (1) that the parole officer has a reasonable suspicion that the parolee had committed a parole violation, and (2) that the search was reasonably related to the parole officer’s duty.”) (citations omitted). Accordingly, we conclude that Pennsylvania would construe the standard form as requiring reasonable suspicion to conduct a search, and thus that the consent form that Baker signed did not authorize suspi-cionless searches of his person, property, or residence.

For the foregoing reasons, we conclude that the search of the trunk was not founded on reasonable suspicion, and that the consent form did not authorize a suspicion-less search thereof. The fruits of the search, including the evidence found in the search of Baker’s house, must therefore be suppressed. See Wong Sun v. United States, 371 U.S. 471, 83 S.Ct. 407, 9 L.Ed.2d 441 (1963). The order of the District Court denying Baker’s motion to suppress will be reversed and the case remanded to the District Court for further action consistent with this opinion.

. In a letter brief filed after the first stage of this appeal, the government advanced the following theoiy of consent for the first time:

the defendant consented to the search and seizure of his bag from the trunk of the car when he identified the bag as his and instructed the parole agent in how to open the trunk. Schneckloth v. Bustamonte, 412 U.S. 218, 93 S.Ct. 2041, 36 L.Ed.2d 854 (1973) (consent to search must be voluntary, but the government is not obliged to prove that the suspect was warned of any right to refuse consent).

*441 Appellee's Letter Br., Nov. 27, 1998. As the government did not make this argument in the District Court and did not renew it in its brief following remand, and as Baker timely objected that the argument had been waived in his response to the government’s letter brief, see Appellant's Response to Appellee’s Letter Br., Dec. 1, 1998, we will not consider this argument on appeal.

. In the first stage of the appeal, Baker also claimed that the District Court erred in "failing to recognize its authority to depart from the career offender status on the grounds that the appellant’s criminal history status seriously overrepresented his past conduct and significantly differed from the heartland.” On remand, the District Court clarified that it was fully aware of its authority to depart on this ground, and Baker makes no argument on this issue at this time.

. The District Court found that, before the parole agent searched the trunk, he knew that "the defendant could not produce documents demonstrating that he owned the car.” However, Agent Rnorr testified that the relevant portions of the conversation regarding the ownership of the car occurred after the agents searched the trunk. See App. at 28a ("Q: After you searched the trunk you returned to your office, is that right? A: Yes. Q: And when you went back inside did you have a *445 conversation with anyone? A: I asked Mr. Baker who owned the vehicle ... ”).

. The officer testified that, when he conducted a title search, he learned that the car was not listed as a stolen vehicle. See App. at 37a (“Q: There is no evidence at all that the vehicle was stolen, correct? A: When I ran it it was not on the computer listed as stolen but the name did not match the person who was in possession of it.”).

. In Commonwealth v. Pickron, 535 Pa. 241, 634 A.2d 1093, 1098 (1993), the Pennsylvania Supreme Court recognized that a parolee has limited Fourth Amendment rights, but held that the Fourth Amendment prohibited "the warrantless search of a probationer or a pa *446 rolee’s residence based upon reasonable suspicion without the consent of the owner or a statutory or regulatory framework governing this kind of search.” Pickron specifically reserved the question whether an agreement signed by a parolee giving consent to a war-rantless search could survive constitutional scrutiny. See id. at 1098 n. 6; Williams, 692 A.2d at 1035.

. When Scott was reviewed by the United States Supreme Court, that Court noted the contention "that the Pennsylvania Supreme Court's opinion was at least ambiguous as to whether it relied on state or federal law to determine the extent of respondent's consent .... ” Pennsylvania Bd. of Probation & Parole v. Scott, 524 U.S. 357, 362 n. 3, 118 S.Ct. 2014, 141 L.Ed.2d 344 (1998).

The basis for the Pennsylvania Supreme Court decision had been debated during oral argument before the Court:

QUESTION: I’m not sure you’re right in saying that the Pennsylvania supreme court based its interpretation of the consent form on its reading of the Fourth *447 Amendment. I thought they might have just based it on an interpretation of Pennsylvania law.

GENERAL FISHER: Mr. Chief Justice, we argue very strongly that they did, in fact, interpret it based on their reading of the Fourth Amendment, and that is where we disagree. We believe that a search of a parolee's residence may be without suspicion, or a suspicionless search.

QUESTION: Well, you — suppose this consent form were redrafted to make clear that the person paroled is consenting to a search without a warrant and to a search without any suspicion whatever. Suppose it were drafted that way. Now, what do you think the Pennsylvania courts would say to that?

GENERAL FISHER: It’s — we believe that the Pennsylvania courts would say, based on what they said in this case, that we could not have that kind of consent form, because we believe they have—

QUESTION: — the Fourth Amendment?

GENERAL FISHER: Because the Fourth Amendment requires reasonable suspicion.

QUESTION: It just wasn't clear to me whether that’s what they said, because they could have meant the consent form just didn’t cover it.

GENERAL FISHER: No, Justice O'Con-nor, we think they said very clearly that the Fourth Amendment requires reasonable suspicion.

Pennsylvania Bd. of Probation & Parole v. Scott, Transcript of Oral Argument, 1998 WL 154625 at *5-6 (Mar. 30, 1998).

While this colloquy hardly resolves the issue, we set it forth for background and for ease of reference when this issue comes before the Pennsylvania Supreme Court, as it doubtless some day will.

. Pennsylvania follows the judicial policy that courts should not reach constitutional questions if they can decide a case upon non-constitutional grounds. See Gartner v. Commonweath, 79 Pa.Cmwlth. 141, 469 A.2d 697, 700 (1983). However, neither the Williams nor the Scott Court explicitly invoked this principle as a basis for its decision.