dissenting in part.
I agree with my colleagues that the record supports the BIA’s finding that the events occurring before García-Cruz left *10Guatemala to enter the United States in May of 2012 did not constitute the type of persecution that creates a presumption of a well-founded fear of future persecution. I also agree that the record supports the BIA’s finding that García-Cruz could have, in any event, safely moved elsewhere in Guatemala.
I cannot agree, however, that we have jurisdiction to adjudicate Gareia-Cruz’s newly minted argument that the agency failed to consider adequately the relevant factors in finding that it would be “reasonable” to expect him to relocate within Guatemala. Congress has stated that we “may review a final order of removal only if ... the alien has exhausted all administrative remedies available to the alien as of right.” 8 U.S.C. § 1252(d)(1). This means “that theories not advanced before the BIA may not be surfaced for the first time in a petition for judicial review of the BIA’s final order.” Makhoul v. Ashcroft, 387 F.3d 75, 80 (1st Cir. 2004). We have further concluded that “[t]his exhaustion requirement is jurisdictional; that is, it constitutes a limitation on our power of review.” Mazariegos-Paiz v. Holder, 734 F.3d 57, 62 (1st Cir. 2013); see also Sousa v. INS, 226 F.3d 28, 31 (1st Cir. 2000) (“Whatever our own views, we are bound by precedent to apply the INA exhaustion requirement in a more draconian fashion.”).
As my colleagues concede, García-Cruz never raised before the BIA his challenge to the thoroughness of the IJ’s consideration of the factors set forth in 8 C.F.R. § 1208.13(b)(3) in finding that relocation within Guatemala would be reasonable. In order to leap over this significant jurisdictional hurdle, my colleagues declare that the BIA sua sponte raised and exhausted the issue of whether relocation would be reasonable. They then assert that we can find, essentially, that the BIA’s sua sponte exhaustion of that issue was not exhaustive enough, because the BIA did not also sua sponte evaluate factors that García-Cruz never faulted the IJ for allegedly failing to consider.
To explain my disagreement, I first recount exactly what transpired before the agency. On the subject of relocation, the IJ stated the following:
Given that the Patriota party has gained ... ground [in Guatemala] and the president of the country is part of that party, the Court finds that the respondent could internally relocate and that, although it would be economically difficult to do that, it is reasonable to expect internal relocation rather than come to the United States.
... Finally, the respondent’s wife and children remain in Guatemala in a town nine to 10 hours away from Chixocol and the Court finds that it is reasonable for the respondent to relocate there and that he could do so safely.
It is plain from the above that the IJ made two relevant findings, consistent with the two-step analysis described in Matter of M-Z-M-R-, 26 I. & N. Dec. 28, 31-32 (B.I.A. 2012): (1) García-Cruz could safely relocate within Guatemala, and (2) it would be reasonable for him to do so.
As my colleagues acknowledge, in his appeal to the BIA, García-Cruz raised a question concerning only the first of those two findings. In Garcia-Cruz’s own words, the question was “[wjhether respondent can safely relocate within Guatemala.” With respect to that question, he advanced only a single, specific argument (again in his words): “Respondent cannot safely relocate within Guatemala.” And in support of that argument, he stated only:
Respondent has demonstrated that he was targeted by political enemies from UNE y GANA, a national party. As a highly visible public supporter of this party, he runs the risk of being identi*11fied and targeted throughout the country.
The fact that respondent’s wife and children have not been harmed after fleeing their hometown should be given very little weight. Respondent’s wife and children were not public supporters of the party and were not themselves politically active. Accordingly, it is unlikely that respondent’s persecutors have been able to determine the identity and location of his family. In contrast, respondent’s persecutors can easily recognize him on sight. Given these facts, the fact that respondent’s wife and children have so far escaped harm does not indicate that he can also safely relocate within Guatemala.
(citations omitted). The BIA directly responded to Garcia-Cruz’s argument with the following:
[T]he Immigration Judge found that the respondent would be able to relocate to another area in Guatemala. In this regard, the Immigration Judge noted that the respondent’s wife and children remain in Guatemala in a town 9 or 10 hours away, strong evidence that the respondent could do so as well.
We see no clear error of fact or mistake of law in the Immigration Judge’s assessment.
(citations omitted).
To find from the foregoing that the BIA raised—or even acknowledged—a challenge to the IJ’s step-two, reasonableness finding, one logically must point to language that one would not expect to find were the BIA simply discussing and rejecting Garcia-Cruz’s argument that it was not safe for him to relocate. My colleagues point to no such language. Rather, and without explanation, they point to the fact that the BIA said García-Cruz “would be able to relocate.” Yet this is what one would well expect the BIA to say in rejecting Garcia-Cruz’s argument that he could not safely relocate. My colleagues otherwise point (again without explanation) to the BIA’s observation that the fact that Garcia-Cruz’s wife and Children remained in Guatemala provided strong evidence that García-Cruz could do so as well. But this statement, too, was directly responsive to Garcia-Cruz’s argument that the experience of his wife and children did not mean that he could safely relocate as well. In short, there is no basis for reading into the BIA’s opinion any indication that it was addressing or even aware of any challenge to the IJ’s step-two, reasonableness finding. Indeed, the v,ery fact that my colleagues fault the BIA for not expressly weighing any of the step-two, reasonableness factors listed in 8 C.F.R. § 1208.13(b)(3) underscores my point: the BIA did not weigh those factors because it did not need to do so in order to respond to Garcia-Cruz’s challenge to the IJ’s step-one, safety finding. To conclude otherwise is to conclude that the BIA decided on its own to raise a question about whether the IJ’s consideration of the § 1208.13(b)(3) factors was thorough enough, and then decided not to say anything at all about the regulation, the factors, or the thoroughness of the IJ’s review. If this is exhaustion, then fatigue must be pandemic at the BIA.
The foregoing explains why the BIA did not hint at or acknowledge the reasonableness issue. But even such a hint or ac-knowledgement would not have been enough to justify my colleagues’ finding of sua sponte exhaustion. Our precedent requires the BIA to “squarely address[ ]” an issue in order for us to find that it sua sponte exhausted the issue. Mazariegos-Paiz, 734 F.3d at 63; see also Velerio-Ramirez v. Lynch, 808 F.3d 111, 117 (1st Cir. 2015) (finding that the BIA sua sponte exhausted the issue of which law applied *12because the BIA “addressed applicable law directly” by “remarking on the IJ’s erroneous use of removal law, stating that [petitioner’s] application is governed by deportation law, and making [an] additional unbriefed determination” regarding the issue). Applied with straight-faced rigor, this “squarely address[ ]” test ensures that we recognize sua sponte exhaustion only when “the agency makes clear its wish to entertain the argument.” Garcia-Carbajal v. Holder, 625 F.3d 1233, 1239 (10th Cir. 2010) (Gorsuch, J.).
I am at a loss to explain how the BIA’s opinion “squarely addressed” a challenge to the IJ’s finding not merely that it would be safe for García-Cruz to relocate, but also that it would be reasonable for him to do so. Perhaps what my colleagues mean to say is that if one discusses the first step of the two-step 8 C.F.R. § 1208.13(b)(2)(ii) inquiry, then one necessarily “exhausts” all challenges to the second step as well. But if this were so, then we would hold that García-Cruz himself, merely by challenging the IJ’s first-step finding as to safety, also exhausted his claim that the IJ committed procedural error in making the reasonableness determination required by the second step of the analysis. And since my colleagues correctly do not so hold, they must think that the BIA raised something that García-Cruz did not, and then squarely addressed it. Yet I can find no explanation in my colleagues’ opinion of how this is so.
The exhaustion requirement established by 8 U.S.C. § 1252(d)(1), like most issue-preservation rules, ensures a modicum of repose and orderliness in the narrowing of issues that occurs as a dispute progresses through several layers of review. This requirement preserves the agency’s statutory prerogative to go first, while also avoiding the delay and uncertainty that would result from multiple rounds of agency review. See Mazariegos-Paiz, 734 F.3d at 62-63 (citing, inter alia, SEC v. Chenery, 332 U.S. 194, 200-01, 67 S.Ct. 1760, 91 L.Ed. 1995 (1947)). Sua sponte exhaustion does not hinder the achievement of these goals, so long as we are confident that the agency really did do what it would have done had the petitioner properly raised an argument. The “squarely address[]” rule provides that assurance only because it demands unambiguous evidence that the agency exhausted the issue on its own. By contrast, my colleagues’ dowsing for sua sponte, collateral exhaustion buried in the disposition of properly raised issues provides no such assurance. The notion that the agency itself, rather than the petitioner, may satisfy § 1252(d)(l)’s exhaustion requirement already stretches the boundaries of our jurisdictional grant. At least, though, we can say that the purposes of the exhaustion requirement are served when the BIA, for whatever reason, has squarely addressed the petitioner’s otherwise unpreserved basis for challenging the IJ’s decision. See id. When we take the yet further step of gleaning exhaustion from a record as bare as this one, on issues that, as here, raise no constitutional considerations, we abandon both statutory text and congressional purpose, and place ourselves in direct conflict with not only our own precedent, see id., but also the precedent of at least one other circuit, see Garcia-Carbajal, 625 F.3d at 1238-39. Such, I fear, may tempt Congress to conclude that the ' circuit courts of appeals cannot be trusted with our already limited power of review in this domain. I therefore respectfully dissent from the order remanding this case to the BIA.