Refugee Convention Signatories Never Signed Up for a Mass Influx of Aliens

Summary

• The crisis in America’s asylum system has its roots in the Refugee Convention of 1951’s “non-refoulement” obligation to not return aliens to their persecutors. Paradoxically, the Convention’s drafters emphatically rejected the notion of adding a right to asylum. Nevertheless, the non-refoulement obligation itself is in many cases functionally indistinguishable from a right to asylum (or, at least, to admission) because of the difficulty of finding third countries willing to accept aliens who claim to fear persecution in their homelands.
• Thus, it is crucial to understand what the Convention’s non-refoulement obligation itself entails. In 1994, the U.S. Supreme Court in Sale v. Haitian Centers Council, Inc., concluded that the text of the Convention did not “apply to aliens interdicted on the high seas”, based on the Court’s textual analysis and its taking from the Convention’s negotiating history that “several delegates understood the right of non-refoulement to apply only to aliens physically present in the host country” and “There is no record of any later disagreement with that position.”
• The negotiating history of the Convention makes clear that the negotiators were laser-focused on ensuring that its non-refoulement obligation not require their countries to allow the entry of mass influxes of aliens arriving at their frontiers claiming to be refugees. The negotiating history even suggests that the delegates wanted their countries to retain the option of removing such aliens who succeeded in penetrating their frontiers back to the countries where they claimed to fear persecution.
• The United States now faces a mass influx of purported refugees and a historic and unsustainable backlog of well over a million such aliens in immigration court. The U.S. Congress should feel free to reform our nation’s immigration laws to the extent necessary to end this mass influx in regards to inadmissible aliens attempting to penetrate our borders, and even as to those who have succeeded in doing so, content in the knowledge that the drafters of the Refugee Convention worked to ensure that its non-refoulement obligation would not condemn their countries to having to endure mass influxes of aliens.

What the Refugee Convention Has Wrought

As my colleague Mark Krikorian has written:

[T]he United Nations Convention Relating to the Status of Refugees [Convention]… laid the groundwork for the asylum and refugee system we have now. Its terms originally applied only to people displaced by Nazi and Soviet aggression in Europe, but in 1967 those provisions were extended to the whole world by a sequel called the Protocol Relating to the Status of Refugees [Refugee Protocol].

Harry Truman did not sign the 1951 Convention because he felt it infringed on U.S. sovereignty, but in 1968 Lyndon Johnson reversed course and signed the Protocol, and the Senate ratified it.

Krikorian then described the legacy of these agreements:

• The key threat the Refugee Convention and Protocol pose to U.S. (or any nation’s) sovereignty comes in the form of asylum, rather than refugee resettlement itself.

• [For i]t is the asylum-seeker who instigates the process by getting himself here on his own power, without our consent … who then demands the right to stay based on the same criteria of persecution as a refugee. The number making such claims is unlimited.

• Asylum … represents a profound surrender of sovereignty, a limitation on the American people’s ability to decide which foreigners get to come here from abroad.

• An unprecedented migration wave … has turned the Refugee Convention … [into] a crowbar used by the post-national Left to pry open the borders of democratic societies contrary to the will of their citizens.

• Asylum is driving the current border crisis, with illegal aliens, coached by smugglers and U.S. immigration lawyers, turning themselves in and claiming a fear of return, thus initiating the asylum process.

Krikorian noted that the United States saw “only 2,000 to 3,000 asylum cases a year in the 1970s”. For example, the “1974 Annual Report: Immigration and Naturalization Service” (1974 Annual Report) stated that:

“The Service [INS] has adopted procedures, in collaboration with the Department of State [DOS], to insure that the principle of asylum inherent in the Protocol and Convention, and their explicit prohibition against the forcible return of refugees to conditions of persecution, will be carried out in the case of any such alien who comes within the jurisdiction of the Service. Every precaution is taken to insure that anyone requesting asylum is given an opportunity to fully present his case. This year, 2,716 aliens requested asylum, and 294 requests were granted.

Even with such low numbers, INS’s 1974 Annual Report stated that “Many [asylum] requests were found to be frivolous and were being used as a stratagem to remain longer in the United States or to avoid deportation.”

Krikorian noted that the numbers have since “ballooned, with more than 300,000 claims filed in 2019”. The balloon continues to inflate. According to the Department of Homeland Security’s (DHS) Office of Homeland Security Statistics, in fiscal year 2022, aliens filed 238,841 affirmative asylum applications with U.S. Citizenship and Immigration Services (USCIS). And according to the Department of Justice’s (DOJ) Executive Office for Immigration Review (EOIR, home of the immigration courts and the Board of Immigration Appeals (BIA)) in fiscal year 2023, aliens filed 479,689 defensive asylum applications with the courts (in the context of removal proceedings) and 12,920 affirmative applications, and this fiscal year, aliens are projected to file (annualizing results through the second quarter) over 668,000 defensive applications and over 11,000 affirmative applications.

Fraud has also ballooned. In 2011, Sam Dolnick reported in the New York Times that:

A shadowy industry dedicated to asylum fraud thrives in New York. ... . Immigrants peddle personal accounts ripped from international headlines. ... The embellished stories go in and out of fashion along with the news of the day. ... West Africans claim genital mutilation or harm from the latest political violence. Albanians and immigrants from other Balkan countries claim they fear ethnic cleansing. Chinese invoke the one-child policy or persecution of Christians, Venezuelans cite their opposition to the ruling party, and Russians describe attacks against gay people. ... Of course, thousands of those claims are legitimate. But each cataclysm provides convenient cover stories for immigrants desperate to settle here for other reasons. ... “When there’s a problem anywhere, a horrible slaughter in Somalia, wherever, the first couple of years of those cases are very real,” said ... an immigration lawyer in Manhattan. “Then the next four or five years, they just mimic those stories.”

Three years later, Kirk Semple, Joseph Goldstein, and Jeffrey Singer wrote an article in the New York Times titled “Asylum Fraud in Chinatown: An Industry of Lies”, reporting that:

• A Chinese woman ... had applied for asylum, claiming that she had been forced to get an abortion in China ... and she was anxious about her coming interview. ... She had good reason to be worried: Her claim, invented by her lawyer’s associates, was false. But the lawyer ... told her to relax. ... “You are making yourself nervous. ... All you would be asked is the same few rubbish questions. ... Just make it up[.]” ... [The lawyer] and a paralegal briefed her on the sequence of fictitious events she had to memorize. ... He said asylum was nearly a foregone conclusion: Cases like hers were getting approved without a problem. “It’s too easy,” he said.

• Peter Kwong, a professor at the City University of New York ... said it was an open secret in the Chinese community that most asylum applications were at least partly false, from fabricated narratives of persecution to counterfeit supporting documents and invented witness testimony. ... False asylum petitions are among the most common forms of immigration fraud, in part because they are difficult to detect, experts said. ... [A]sylum fraud cuts across all immigrant groups, officials say. ... And while many such claims are legitimate, officials and industry specialists said, an untold number are not. Mr. Kwong said the cases were easy to fake. “The law itself provides a huge loophole…” he said. ... Narratives and documents are recycled from client to client, with the names and dates changed — though sometimes the lawyers forget to do even that.

Fast forward to the present day, where even President Biden’s impeached DHS Secretary Alenjandro Mayorkas and Attorney General Merrick Garland have acknowledged that:

• [M]any individuals [placed into expedited removal proceedings] who avail themselves of the credible fear [of persecution] process [to allow them to be released from detention and possibly make an asylum claim in immigration court] do not have meritorious claims.

• [M]ost people processed for expedited removal ... will likely establish credible fear and remain in the United States for the foreseeable future despite the fact that many of them will not ultimately be granted asylum.

• [T]he current asylum system — in which most migrants who are initially deemed eligible to pursue their claims ultimately are not granted asylum in the subsequent ... removal proceedings — has contributed to a growing backlog of cases awaiting review by asylum officers and immigration judges.

• [N]oncitizens ultimately found ineligible for asylum or another form of protection are likely to spend many years in the United States prior to being ordered removed.

• [T]he fact that migrants can wait in the United States for years before being issued a final order denying relief, and that many such individuals are never actually removed, likely incentivizes migrants to make the journey north.

In fact, as DHS and DOJ reported in 2018, “significant proportions of aliens who receive a positive credible-fear determination never [even] file an application for asylum or are ordered removed in absentia”. What happens to apprehended aliens who are found to have a credible fear, or to those who are not put into “expedited” proceedings in the first place and are simply given Notices to Appear in immigration court years down the line and released onto our streets? They enter a metaphorical supermassive black hole. Well, the metaphor sort of breaks down here because while not even light itself can escape the gravitational field of a black hole, many such aliens do escape our immigration system. DHS explained in 2018:

[Many] aliens claiming credible fear ... know it will give them an opportunity to stay in our country, even if they do not actually have a valid claim to asylum. As a result, the United States has an overwhelming asylum backlog. ... Most of these claims are not meritorious ... . However, by the time a judge has ordered them removed from the United States, many have vanished.

And as the House Homeland Security Committee’s report on the impeachment proceedings against Secretary Mayorkas explained: “During Secretary Mayorkas’ tenure, the immigration court backlog has more than doubled from about 1,300,000 cases to over 3,000,000 cases. The backlog is destroying the courts’ ability to administer justice and provide appropriate relief in a timeframe that does not run into years or even decades.”

In order to understand how we got here, we need to go back to 1950. As Andrew Schoenholtz, professor from practice at the Georgetown University Law School and director of its Human Rights Institute and its Center for Applied Legal Studies, has explained, “[p]revious international agreements did not address the situation of the numerous refugees produced by the Second World War, so following a study by the Secretary General, the [United Nations’] Economic and Social Council … appointed [an] Ad Hoc Committee” on Statelessness and Related Problems, with 13 countries represented, to “consider the desirability of preparing a new convention regarding the international status of refugees and stateless persons”.

Then, as Schoenholtz explained:

On January 16, 1950, the Ad Hoc Committee convened in Lake Success, New York, and completed its initial work on February 16, 1950, with the adoption of a Draft Convention relating to the Status of Refugees and a Protocol thereto relating to the Status of Stateless Persons. Following discussions of the draft within the Council and among governments, the Ad Hoc Committee met in Geneva from August 14 to August 25, 1950, and prepared a revised draft.

This draft was submitted to the [United Nations’] General Assembly, which convened [a] conference of plenipotentiaries in Geneva from July 2 to July 25, 1951, to complete the drafting. This final drafting provided an opportunity for more governments to participate. Twenty-six states were represented by delegates and two governments by observers.

The Conference of Plenipotentiaries adopted the Convention by a vote of twenty-four to none.

Although, as Krikorian mentioned, the United States was not a party to the Convention, when we acceded to the 1967 Protocol, we committed to observe Articles 2 through 34 of the Convention, which were incorporated by reference in the protocol. Article 33 provides that:

Prohibition of Expulsion or Return (“Refoulement”)

1. No Contracting State shall expel or return (“refouler”) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened on account of his race, religion, nationality, membership of a particular social group or political opinion.
2. The benefits of the present provision may not, however, be claimed by a refugee whom there are reasonable grounds for regarding as a danger to the security of the country in which he is, or who, having been convicted by a final judgement of a particularly serious crime, constitutes a danger to the community of that country.

The United States honors Article 33 not through the discretionary statutory relief of asylum (to be eligible for which an alien must be “unable or unwilling to return to [generally, the country of their nationality] … because of persecution or a well-founded fear of persecution on account of race, religion, nationality, membership in a particular social group, or political opinion”. Rather, we honor Article 33 through the mandatory statutory relief of withholding of removal (under which DHS “may not [with certain exceptions] remove an alien to a country if the [secretary] decides that the alien’s life or freedom would be threatened in that country because of the alien’s race, religion, nationality, membership in a particular social group, or political opinion”.

As the Supreme Court explained in 1987 in INS v. Cardoza-Fonseca:

[Asylum] is a discretionary mechanism which gives [DHS] the authority to grant the broader relief of asylum to refugees. As such, it does not correspond to Article 33 of the Convention, but instead corresponds to Article 34 … [which] provides that the contracting States “shall as far as possible facilitate the assimilation and naturalization of refugees ... .” Like [the Immigration and Nationality Act’s (INA) asylum provision], the provision is precatory; it does not require the implementing authority actually to grant asylum to all those who are eligible.

In fact, the representatives to the Ad Hoc Committee and delegates to the Conference of Plenipotentiaries were adamant that the Refugee Convention not establish any right to asylum. Louis Henkin, the U.S.’s representative to the Ad Hoc Committee, stated in an affidavit¹ filed by the respondents in the 1994 Supreme Court case of Sale v. Haitian Centers Council, Inc. that “Those who drafted the Convention recognized that governments were not prepared to commit themselves to grant asylum, even to bona fide refugees” and that the Convention “was not intended to grant or suggest any right to asylum.” And Harold Hongju Koh, professor of international law and director of the Orville H. Schell, Jr., Center for International Human Rights at Yale Law School, has written that “contracting states continued to resist a mandatory obligation of asylum”. But, as Schoenholtz noted, “that decision did not in any way restrict the application of the nonrefoulement obligation.”

In any event, while the decision to grant asylum to an alien who meets the eligibility criteria remains within DHS’s discretion, federal courts in practice have made it very difficult for USCIS and the immigration courts to not grant asylum. In 2004, the Ninth Circuit revealed in Kalubi v. Ashcroft how much like Kabuki theater asylum cases really are:

We have consistently required the BIA to state its reasons and show proper consideration of all factors when weighing equities and denying [asylum] relief.”… This means that the BIA must explain what factors it has considered or relied upon sufficiently that we are able to discern that it “has heard, considered, and decided.”

As the Beatles might have said, you know what the 9th Circuit means.

As mentioned, we honor Article 33 through withholding of removal. Yet immigration judges grant withholding much less frequently than they do asylum — according to EOIR’s “Statistics Yearbook: Fiscal Year 2018”, immigration judges granted asylum in 13,248 cases and withholding in 1,746 cases. The lower numbers are explained by the fact that in immigration court, “[a]n asylum application also generally serves as an application for withholding of removal” and if an alien is granted asylum there is no need to consider eligibility for withholding.

Asylum will always be an alien’s first preference, as it offers many more benefits than does withholding of removal and, additionally, the evidentiary standard is lower. As DHS and DOJ have noted, unlike asylum generally, withholding does not “[p]rohibit the Government from removing the alien to a third country; ... create a path to lawful permanent resident status and citizenship; or … afford the same ancillary benefits (such as derivative protection for family members) and access to Federal means-tested public benefits.” And the evidentiary standard for withholding — “[m]ore likely than not that the alien would be subject to persecution” — is higher than the standard for asylum — generally speaking, that persecution be a “reasonable possibility”.

If the United States were ever to do away with statutory asylum, I would venture a guess that most aliens who would have received asylum would be granted withholding in its stead. Thus, the Convention is still responsible for the bedlam in our asylum system even though it does not actually require the grant of asylum to eligible aliens.

What Were They Thinking? The Negotiating History of the Refugee Convention

Did the representatives participating on the Ad Hoc Committee and the delegates participating on the Conference of Plenipotentiaries really intend to create a Convention that would profoundly surrender their countries’ sovereignty and limit their ability to decide which foreigners would be allowed to enter? That is the primary question addressed by this report. Before diving into the history of the negotiations, I want to highlight three factors that should inject a dose of healthy skepticism regarding claims that the Convention’s aim was in fact such surrender.

First, the delegates on the Conference of Plenipotentiaries rejected non-refoulement absolutism. Louis Henkin was an absolutist, both during his time on the Ad Hoc Committee and later as an academic. He famously said during the Ad Hoc Committee’s 20th meeting, on February 1, 1950, that “[w]hatever the case might be”, a refugee “must not be turned back to a country where his life or freedom could be threatened” and “[n]o consideration of public order should be allowed to overrule that guarantee”.

In 2007, DOS contended in its “U.S. Observations on UNCHR [United Nations High Commissioner for Refugees] Advisory Opinion on Extraterritorial Application of Non-Refoulement Obligations” (U.S. Observations) that Henkin’s “sentiment was not adopted by the Committee (nor was it raised by [George Warren,] the U.S. delegate at the later Conference of Plenipotentiaries)”, and that his “vision thus did not advance beyond his proposal at the initial Committee.” On the other hand, Henkin stated in his Sale affidavit that:

Mr. George Warren attended the Conference of Plenipotentiaries as the U.S. delegate at the July 11 and 25, 1951 sessions, but he and I had no disagreement regarding this point. My statement the previous year … continued as the official position of the United States government with regard to the meaning of Article 33. Mr. Warren took no action to rescind that interpretation or to support any other interpretation.

Koh has taken Henkin’s affidavit to mean that “Henkin's views … represented the official position of the United States government.”

In any event, the Conference of Plenipotentiaries rejected Henkin’s views, as it went on to add paragraph 2 to Article 33, providing an exception to non-refoulement for an alien for whom there are “reasonable grounds for regarding as a danger to the security of the country in which he is, or who, having been convicted by a final judgement of a particularly serious crime, constitutes a danger to the community of that country.”

Second, Henkin himself stated in his affidavit that “[n]o one could anticipate how many refugees might claim the privilege of non-refoulement” and later reflected that:

Th[e non-refoulement] provision entered the Convention as a kind of postscript. It represents a small charitable concession, reflecting … “humanitarian impulses” by the drafters in response to reported incidents of Jews fleeing the Gestapo and being pushed by police of neighboring states back into the hands of their persecutors. States were prepared to make that humanitarian gesture, given its limited nature and limited scope as they saw it. Indeed, I suspect, the non-refoulement provision was seen as not likely to be invoked again. [All emphasis is added throughout this report, unless otherwise specified.]

Third, David Martin, INS’s general counsel during the Clinton administration and DHS’s principal deputy general counsel during the Obama administration, has suggested that:

It is time to put the 1951 Convention in perspective, to remember that it was a product of the dark time in which it was created. That year may have marked the low point of the early Cold War, a time when the world had grown deeply disillusioned about the bright hopes for collective UN action that had sprung up in 1945. Moreover, and this is crucial, it was adopted at a time when national sovereignty loomed large.

To further investigate “what they were thinking”, we need to look to the negotiating history of the Convention, contained in the travaux préparatoires (“official documents recording the negotiations, drafting, and discussions during the process of creating a treaty”) with respect to the Ad Hoc Committee and the Conference of Plenipotentiaries.

The Supreme Court heavily relied on the travaux préparatoires in Sale. In this 8-1 decision, the Court affirmed the federal government’s ability to repatriate aliens interdicted on the high seas without having to screen them for eligibility for withholding of removal, As the Court laid out the question:

The President has directed the Coast Guard to intercept vessels illegally transporting passengers from Haiti to the United States and to return those passengers to Haiti without first determining whether they may qualify as refugees. The question presented is whether such forced repatriation, “authorized to be undertaken only beyond the territorial sea of the United States,” violates [the INA’s withholding provision]

The Court ruled that “neither [the INA’s withholding provision] nor Article 33 ... applies to action taken by the Coast Guard on the high seas” beyond the territorial waters of the United States. The Court based its decision to a significant extent on its reading of the travaux préparatoires, both as to the meaning of Article 33 and the meaning of the INA’s withholding of removal provision (as “the history of the [Refugee Act of] 1980 Act does disclose a general intent to conform our law to Article 33 of the Convention”).

The Court presented what it found to be the key exchanges in the negotiating history. Keep in mind that the travaux préparatoires consist of paraphrases of representatives’ and delegates’ statements (which is exactly what one finds in the “Annals of Congress”, setting forth the floor debates of the U.S. House and Senate from 1789 to 1824):

In early drafts of the Convention, what finally emerged as Article 33 was numbered 28. At a negotiating conference of plenipotentiaries … on July 11, 1951, the Swiss delegate explained his understanding that the words “expel” and “return” covered only refugees who had entered the host country. He stated:

“Mr. ZUTTER (Switzerland) said that the Swiss Federal Government saw no reason why article 28 should not be adopted as it stood; for the article was a necessary one. He thought, however, that its wording left room for various interpretations, particularly as to the meaning to be attached to the words ‘expel’ and ‘return’. In the Swiss Government's view, the term “expulsion” applied to a refugee who had already been admitted to the territory of a country. The term ‘refoulement’, on the other hand, had a vaguer meaning; it could not, however, be applied to a refugee who had not yet entered the territory of a country. The word ‘return’, used in the English text, gave that idea exactly. Yet article 28 implied the existence of two categories of refugee: refugees who were liable to be expelled, and those who were liable to be returned. In any case, the States represented at the Conference should take a definite position with regard to the meaning to be attached to the word ‘return’. The Swiss Government considered that in the present instance the word applied solely to refugees who had already entered a country, but were not yet resident there. According to that interpretation, States were not compelled to allow large groups of persons claiming refugee status to cross its frontiers. He would be glad to know whether the States represented at the Conference accepted his interpretations of the two terms in question. If they did, Switzerland would be willing to accept article 28, which was one of the articles in respect of which States could not, under article 36 of the draft Convention, enter a reservation.” [Emphasis added, except underlined text reflects the Court’s emphasis.]

In a further exchange that day that the Supreme Court did not include, Baron van Boetzelaer, the representative from the Netherlands, stated that:

[He] supported the Swiss representative's observations. He appreciated the importance of the basic principles underlying article 28 but, as a country bordering on others, the Netherlands was somewhat diffident about assuming unconditional obligations so far as mass influxes of refugees were concerned, unless international collaboration was sufficiently organized to deal with such a situation. He recalled that the Swiss representative had already referred to the matter of international collaboration in the general discussion at the third meeting. [Emphasis added.]

Then, Theodoli, the representative from Italy, “associated himself with the statements of the Swiss and Netherlands representatives” and went on to say that:

He would like some clarification of the words “expel or return”. Under article 28, no Contracting State was to expel or return a refugee to a territory where his life or freedom would be in danger. On the other hand, he personally felt that a State could not commit itself not to expel or return large groups of refugees who presented themselves on its territory, and who might endanger public security. The Italian delegation would therefore reserve its position on the article, unless some satisfactory explanation was forthcoming. [Emphasis added.]

Von Trützschler, the Federal Republic of Germany’s representative, stated that he “supported the observation of the Netherlands representative concerning countries subject to a large influx of refugees.” (emphasis added).

Then, Herment, the Belgian representative, “drew attention to the fact that in article 28 the prohibition on returning refugees to the frontier could be construed as applying to individuals, but not to large groups. Such was the interpretation placed on it by the Belgian Government.”

Now back to the Supreme Court’s description:

No one expressed disagreement with the position of the Swiss delegate on that day or at the session two weeks later when Article 28 was again discussed. At that session [the 35th meeting on July 25, 1951], the delegate of the Netherlands recalled the Swiss delegate's earlier position:

“Baron van BOETZELAER (Netherlands) recalled that at the first reading the Swiss representative had expressed the opinion that the word ‘expulsion’ related to a refugee already admitted into a country, whereas the word ‘return’ (‘refoulement’) related to a refugee already within the territory but not yet resident there. According to that interpretation, article 28 would not have involved any obligations in the possible case of mass migrations across frontiers or of attempted mass migrations.

“He wished to revert to that point, because the Netherlands Government attached very great importance to the scope of the provision now contained in article 33. The Netherlands could not accept any legal obligations in respect of large groups of refugees seeking access to its territory.

“At the first reading the representatives of Belgium, the Federal Republic of Germany, Italy, the Netherlands and Sweden had supported the Swiss interpretation. From conversations he had since had with other representatives, he had gathered that the general consensus of opinion was in favour of the Swiss interpretation.

“In order to dispel any possible ambiguity and to reassure his Government, he wished to have it placed on record that the Conference was in agreement with the interpretation that the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33.”

“There being no objection, the President ruled that the interpretation given by the Netherlands representative should be placed on record.” [underlined text reflects the Court’s emphasis.]

Did Switzerland and the Netherlands Rule the Day or Rue the Day?

Before we try to ascertain the meaning of these exchanges, we need to determine whether the Swiss and Dutch understanding actually prevailed at the Conference of Plenipotentiaries. There has been a longstanding debate about whether the Conference actually adopted their viewpoint. The Supreme Court in Sale certainly believed that it had:

Although the significance of the President's comment that the remarks should be “placed on record” is not entirely clear, this much cannot be denied: At one time there was a “general consensus,” and in July 1951 several delegates understood the right of non-refoulement to apply only to aliens physically present in the host country. [footnote: The Swiss delegate's statement strongly suggests, moreover, that at least one nation's accession to the Convention was conditioned on this understanding.] There is no record of any later disagreement with that position. Moreover, the term “refouler” was included in the English version of the text to avoid the expressed concern about an inappropriately broad reading of the English word “return.” [Emphasis in original.]

Two years earlier, Judge John Walker, Jr., concluded in his dissenting opinion in the Second Circuit’s 1992 decision in Haitian Ctrs. Council v. McNary that:

The President of the Conference premised his ruling on “there being no objection,” a finding which alone would indicate agreement. Additionally, the Conference recorded the Netherlands delegate’s “interpretation,” not his “views.” Indeed, I cannot fathom why a member would seek to have his “views” formally recorded since that is accomplished by the transcript itself … . [T]he Conference formally recorded the delegate's “interpretation” so as to show its agreement.

Further, Dr. Paul Weis concluded in his commentary on the Convention based on the travaux préparatoires that “It was ruled by the President of the Conference that the Article does not apply to mass migrations.” Dr. Sadako Ogata, United Nations High Commissioner for Refugees, had expressed his “great source of satisfaction” that Weis was able to write the commentary, and noted that Weis “played an active part in the work leading to the preparation of the 1951 Convention”, that he “was a universally recognized expert on International Refugee Law”, and that he “was posthumously awarded the Nansen Medal in recognition of his major contribution to the development of international legal standards for the treatment of refugees”.

Justice Blackman thought otherwise in his dissent in Sale:

• Baron van Boetzelaer's remarks almost certainly represent, in the words of the United Nations High Commissioner for Refugees [UNHCR], a mere “parliamentary gesture by a delegate whose views did not prevail upon the negotiating conference as a whole”. [Emphasis added by UNHCR.]

• That it did not is evidenced by the fact that Baron van Boetzelaer's interpretation was merely “placed on record,” unlike formal amendments to the Convention which were “agreed to” or “adopted.” It should not be assumed that other delegates agreed with the comment simply because they did not object to their colleague's request to memorialize it, and the majority's statement that “this much cannot be denied: At one time there was a ‘general consensus,’” … is wrong. All that can be said is that at one time Baron van Boetzelaer remarked that “he had gathered” that there was a general consensus, and that his interpretation was placed on record.

• In sum, the fragments of negotiating history upon which the majority relies … were never voted on or adopted, [and] probably represent a minority view.

In Haitian Ctrs. Council v. McNary, the Second Circuit’s majority opinion expressed ambivalence:

The Netherlands’ representative asked that it be placed on record “that the Conference was in agreement” with his interpretation, but the President ruled only that “the interpretation given by the Netherlands representative should be placed on record.”… [I]t would … be an equally-fair reading to see this passage as simply recording the views of a dissenting member.

UNHCR’s “Advisory Opinion on the Extraterritorial Application of Non-Refoulement Obligations under the 1951 Convention relating to the Status of Refugees and its 1967 Protocol” (Advisory Opinion) concluded that “these portions of the negotiating history do not warrant the conclusion that the drafters of the 1951 Convention reached consensus about an implicit restriction of the territorial scope of the principle of non-refoulement as provided for in Article 33(1).”

As to commentators, Koh has written that:

[A]part from the Swiss and Dutch speakers, only four others of the twenty-six states represented at the provision’s drafting voiced actual support for the Swiss and Dutch position, not surprisingly, the European countries of Belgium, Federal Republic of Germany, Italy, and Sweden. The President of the Conference then ruled that the Swiss and Dutch interpretation be placed on the record — a record that contains thousands of pages of written material — not that the entire Conference agreed with that statement.

And Andrew Schoenholtz has written that “The fact that no objections were raised does not, in the politeness of international negotiations and in the absence of the interpretation being ‘agreed’ to or ‘adopted,’ indicate that an agreement had been reached.”

I must agree with the Supreme Court, the dissenting Second Circuit judge, and Weis. It is one thing to simply allow a delegate to record his views. But what was recorded was the delegate’s determination that “the Conference was in agreement with the interpretation that the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33”. It would have been legal malpractice for any delegates who disagreed with this interpretation to not object to its placement in the record with a notation that the Conference was “in agreement”.

It’s All About that Mass (Influx)

The exchanges paraphrased above establish that the delegates to the Conference of Plenipotentiaries were laser-focused on ensuring that the Convention not obligate them to accept mass influxes of refugees. The Swiss representative expressed concern that his nation not be “compelled to allow large groups of persons claiming refugee status to cross its frontiers”. The Dutch representative stated that he “supported the Swiss representative's observations” and noted that “as a country bordering on others, the Netherlands was somewhat diffident about assuming unconditional obligations so far as mass influxes of refugees were concerned, unless international collaboration was sufficiently organized to deal with such a situation”. The Italian representative “associated himself with the statements of the Swiss and Netherlands representatives” and went on to state that “he personally felt that a State could not commit itself not to expel or return large groups of refugees who presented themselves on its territory, and who might endanger public security”. The Federal Republic of Germany’s representative stated that he “supported the observation of the Netherlands representative concerning countries subject to a large influx of refugees”.

This is not just my assessment. Alice Edwards, who was chief of UNHCR’s Protection Policy and Legal Advice Section, has concluded that “In the context of mass movements of refugees, the Swiss, Dutch and other representatives at the Conference of Plenipotentiaries wanted it recorded that art 28 (the then equivalent to art 33) did not compel states to allow mass migration across [their] frontiers.” And James Hathaway, professor emeritus at the University of Michigan Law School, concluded in his highly respected book The Rights of Refugees under International Law:

• Beyond the possibility of particularized exclusion under Art. 33(2), the intention to establish a broadly applicable duty of non-refoulement was qualified during the final phase of the drafting process in order to accommodate critical public order and national security concerns which may arise during a “mass influx.”

• [I]t is clear that their intention was not to endorse the routine refoulement of refugees, but rather only to authorize states to defend their frontiers in the event of a threat to their national security engendered by a mass migration of refugees[.]

• There are three significant gaps in the protective ambit of Art. 33 … . [T]he duty of non-refoulement does not bind a state faced with a mass influx of refugees insofar as the arrival of refugees truly threatens its ability to protect its most basic national interests.

  [This is an] implied exception to the duty of non-refoulement for refugees arriving as part of a mass influx … . [S]imply expanding the notion of non-refoulement to mass influx situations would … exact an inappropriately high cost to the collective survival of states of destination that happen to be in the refugees’ path of flight.

Hathaway cited the July 11, 1951, statement of Knud Larsen, the Danish delegate, during the Conference that:

The President [of the Ad Hoc Committee] thought that the … Committee, in drafting article [33], had, perhaps, established a standard that could not be accepted. That Committee … had felt that the principle inherent in article [33] was fundamental, and that it could not consider any exceptions to the article.

Hathaway concluded that “As is clear from this statement … the absolutism of concern to the President was the unwillingness to consider exceptions to the duty of non-refoulement, as for example were argued to be necessary in the event of mass influx.” (Emphasis in original.) Hathaway further wrote that:

• At the Conference of Plenipotentiaries, Switzerland and the Netherlands reasserted the customary understanding that a comprehensive and absolute duty of non-refoulement was untenable in the face of a mass influx. The President agreed, ruling that ‘‘the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33.’’ The French term refoulement was added to the English text of the article following the word ‘‘return’’ to ensure that the duty of non-return was understood to have ‘‘no wider meaning’’ than the French expression, which was agreed not to apply in the event that national security or public order was genuinely threatened by a mass influx.

• At the level of text, th[e] position [that there is no limitation on the scope of the duty of non-refoulement where a state is at grave risk owing to a mass influx] ignores the explicit decision to add the French language word ‘‘refoulement’’ to the English language version of Art. 33 in order to ensure that the traditional civil law understanding of that term (which did not govern in a mass influx) would be formally recognized.

Hathaway did note that “The view that there is an implied limitation on the scope of the duty of non-refoulement where a state is at grave risk owing to a mass influx is, however, often resisted.” Justice Blackman was one of the resisters, asserting in his dissenting opinion that:

[N]o one seriously contends that the treaty's protections depend on the number of refugees who are fleeing persecution. Allowing a state to disavow “any obligations” in the case of mass migrations or attempted mass migrations would eviscerate Article 33, leaving it applicable only to “small” migrations and “small” attempted migrations.

But that is exactly what the delegates whose statements I cited contended, and they were quite serious.

Even Louis Henkin stated in his affidavit that “Some delegates did express the concern that the right of non-refoulement should not become a vehicle for requiring the admission of massive numbers of migrants.” But he then curiously stated that:

But none of our discussions suggested any intent to condone a state's denying the right of non-refoulement to members of groups, without individually determining — through screening or other means — whether each individual qualified for refugee status … . By expressing a caveat about mass migrations, the delegates were confirming that the right of non-refoulement attached to individual refugees and not to groups.

I am not sure exactly what Henkin meant to say here, but as Hathaway has pointed out:

In the context of individuated applications for protection, it is feasible for states scrupulously to avoid peremptory acts of refoulement. The applicant can be admitted to the state’s territory and removed if ultimately adjudged to constitute a serious risk to either national security or the safety of the community. In contrast, it is not usually practical for a country overwhelmed by a mass influx of refugees to engage in … detailed, case-by-case analysis of risks to its own well-being. Governments therefore wanted the assurance that in truly exceptional circumstances, they could engage in peremptory refoulement to the extent truly necessary to protect their most critical national interests.

What Is a Mass Influx and How Does the Convention Allow a Signatory to Respond?

The question arises of exactly what constitutes a mass influx. In 2004, UNHCR concluded in its “Conclusion on International Cooperation and Burden and Responsibility Sharing in Mass Influx Situations” (Conclusion) that:

[M]ass influx is a phenomenon that has not been defined, but … may … have some or all of the following characteristics: (i) considerable numbers of people arriving over an international border; (ii) a rapid rate of arrival; (iii) inadequate absorption or response capacity in host States, particularly during the emergency; (iv) individual asylum procedures, where they exist, which are unable to deal with the assessment of such large numbers.

Alice Edwards contended that “There is neither a minimum number, nor speed of arrival, for a ‘mass influx.’” She stated that the absorption capacity criteria “has been said to add ‘an important qualitative dimension, making it impossible — and indeed undesirable — to establish in precise quantitative terms how large, or sudden, a refugee flow must be in order to be [so] characterized[.’]” She further contended that:

[T]he size and scale of a movement of refugees is not the decisive criterion as to whether a state is empowered to delay or suspend the application of the 1951 Refugee Convention’s provisions … . [T]he decisive criterion is rather whether the mass influx presents a threat to national security within the specific context of war or other grave and exceptional circumstance; it is thus not enough for a state to declare a particular refugee movement as a ‘mass influx’ to circumvent Convention obligations.

Hathaway concluded that under “the limited scope of the mass influx exception as conceived by the drafters of the Convention: states are allowed to deny entry to refugees only in truly exceptional circumstances, and only to the extent truly necessary to protect their most critical national interests”. The denial of entry is “[j]ustified in the case of mass influx only where it is the sole realistic option for a state that might otherwise be overwhelmed and unable to protect its most basic national interests”. He went on to argue that:

• Most fundamentally, there can be no question of avoiding the duty of nonrefoulement under this implied exception … where the numbers arriving and the resources of the receiving state are such that security concerns can be addressed under the individuated exceptions set by Art. 33(2).

• Because it is such an exceptional measure, suspension of protection from refoulement must be carried out in a way that is minimally invasive of the human dignity of refugees. While the acute risk to states inherent in particular circumstances will sometimes justify blunt refusals of protection, the limited right of states to engage in refoulement should not be interpreted as a form of carte blanche to practice unnecessary harshness.

Hathaway gave an example of a situation he did not believe to meet this bar:

[T]he American interdiction of the boats of fleeing Haitians was an infringement of the rule against refoulement: it simply could not reasonably be said that the circumstances in the country of destination were so fragile, or the number of asylum-seekers so massive in relation to adjudicative and reception resources, that the orderly assessment of claims would have exposed the receiving state to an unacceptable risk.

He then gave an example of a case that likely did:

A more compelling case can … be made for the legality of the border closings by Zaire and Tanzania in the face of refugee flows from Rwanda and Burundi. Both states had been overwhelmed by hundreds of thousands of refugees, and were faced by the imminent prospect of additional flows at the time of the border closures. At least in the case of Zaire, there was also good reason to believe that internal security could be threatened by the entry of refugees, many of whom were suspected of having committed serious criminal offenses. The decisions to suspend border crossings were moreover of limited duration, while efforts to secure international resources to protect refugees were being pursued.

UNCHR concluded in its “The Principle of Non-Refoulement as a Norm of Customary International Law. Response to the Questions Posed to UNHCR by the Federal Constitutional Court of the Federal Republic of Germany in Cases 2 BvR 1938/93, 2 BvR 1953/93, 2 BvR 1954/93” (Response to the Questions Posed) that:

[As to] statements by various representatives regarding the non-applicability of Article 33 paragraph 1 to situations of “mass migrations” it is clear that what was intended was not simply the arrival of a large number of asylum-seekers, but the arrival of asylum-seekers in such large numbers as to go beyond the capacity of the asylum-state to handle without the ¬co-operation of the international community.

While conceding that these fears were “understandable”, UNHCR contended that:

Later events have shown that with further development and strengthening of international co-operation in dealing with refugee problems, these … fears have been largely overcome and the principle of non-refoulement is now regarded as fully applicable in large-scale influx situations. In this connection, mention should be made of Conclusion No. 22 … adopted by the UNHCR Executive Committee in 1981 on the Protection of Asylum-Seekers in Situations of Large-Scale Influx … [including the] state[ment] that in “all cases of large-scale influx the principle of non-¬refoulement — including non-rejection at the frontier — must be scrupulously observed”.

However, Conclusion No. 22 also stated that:

A mass influx may place unduly heavy burdens on certain countries; a satisfactory solution of a problem, international in scope and nature, cannot be achieved without international cooperation. States shall, within the framework of international solidarity and burden-sharing, take all necessary measures to assist, at their request, States which have admitted asylum-seekers in large-scale influx situations.

Hathaway argued that:

Conclusion No. 22 actually suggests an understanding of the duty of non-refoulement which disallows state parties any prerogative to deny entry to refugees in a mass influx situation so long as there is reason to believe that the risk to their critical national interests occasioned by the mass influx will be countered by timely assistance from other states. [Emphasis added.]

But, as Hathaway acknowledged, UNHCR’s Conclusion 85 states that:

[I]nternational solidarity and burdensharing are of direct importance to the satisfactory implementation of refugee protection principles; [we] stress[] however … that ... the meeting by States of their protection obligations should not be dependent on burden-sharing arrangements first being in place, particularly because respect for fundamental human rights and humanitarian principles is an obligation for all members of the international community.

Hathaway countered that:

Not only is this conclusion not clearly oriented to the mass influx situation, but it also begs the question since — if the duty of non-refoulement does not extend to circumstances where fundamental interests are threatened by a genuine mass influx — a call to respect protection obligations even without burden-sharing is not infringed by recognition of the implied exception to the duty of non-refoulement.

Of course, it is important to note that there is no indication that the representatives and delegates who negotiated the Convention were aware of any of these hypothesized limitations, or that they agreed that they should be implicitly read into Article 33. Other than the Italian delegate, who expressed concern regarding a mass influx that might endanger public security, the other delegates expressed the need to prevent mass influxes, without any qualifiers.

When Hathaway discussed an “implied exception”, he was raising the concept of “derogation”. As to derogation, Alice Edwards has explained that:

Under international law, derogation is a method of putting other states “on notice” that the notifying state intends to suspend or curtail specific norms or obligations that would apply in ordinary circumstances, owing to a war or other comparable emergency making it impossible for the state in question to comply fully. Derogation is thus “the legally mandated authority of states to allow suspension of certain individual rights in exceptional circumstances of emergency or war”. It sends a message that the suspension of rights is “necessary, temporary, and lawful”. Rooted in the doctrine of state necessity (or self-preservation), derogation is reserved for exceptional situations that are temporary in nature … . “This compromise … enables continued adherence to the principle of the rule of law and faithfulness to fundamental democratic values, while providing the state with adequate measures to withstand the storm wrought by the crisis”.

The Convention has two articles relevant to the power of derogation. Article 8 provides:

Exemption from Exceptional Measures

With regard to exceptional measures which may be taken against the person, property or interests of nationals of a foreign State, the Contracting States shall not apply such measures to a refugee who is formally a national of the said State solely on account of such nationality. Contracting States which, under their legislation, are prevented from applying the general principle expressed in this article, shall, in appropriate cases, grant exemptions in favour of such refugees.

Edwards explained that Article 8 “was designed to ensure that there could be no derogation on discriminatory grounds”.

Article 9 provides:

Provisional Measures

Nothing in this Convention shall prevent a Contracting State, in time of war or other grave and exceptional circumstances, from taking provisionally measures which it considers to be essential to the national security in the case of a particular person, pending a determination by the Contracting State that that person is in fact a refugee and that the continuance of such measures is necessary in his case in the interests of national security.

Edwards explained that:

• In relation to what constitutes a threat to national security for the purposes of art 9, it is generally understood to be broader than armed conflict, and could include threats to overthrow the state by illegal means, terrorism or other serious disturbances to public order. Further, as the United Nations Security Council has noted on a number of occasions that mass displacement can itself constitute a threat to international peace and security, it is arguable that such situations could, where evidence is substantiated, qualify as an ‘emergency’ or ‘national security threat’ justifying derogation under art 9. In fact, mass influxes were also in the minds of the drafters.

• For derogation from Convention obligations in situations of mass influx to be permitted, the mass influx must be within the context of war or otherwise create a situation of grave and exceptional circumstance that threatens national security; or alternatively, the influx itself must threaten the life of the nation.

Does derogation apply to all the commitments made by signatory nations? Edwards has concluded that:

• Very few rights were in fact mentioned explicitly in the derogation debates and thus it begs the question whether the language of “[n]othing in this Convention” … can be taken to mean that there are no limits on rights fallings within the scope of art 9. Rather, it is argued that the text reflects merely that nothing in the Convention prohibits derogation. It does not follow as a consequence that derogation is permissible against all of the rights in the Convention. [Emphasis added throughout.]

• [D]erogation under the 1951 Refugee Convention is subject to the general rules of international law, namely: it applies only in “war or other grave and exceptional circumstance”; the measures taken need to be proportionate to the emergency in question and last only as long as the threat exists (and hence they are temporary); it does not apply to the most fundamental of rights, including the principle of non-refoulement.

Yet, Edwards admitted that:

[Paragraph 2 of Article 33] permits lawful refoulement in respect of refugees who pose a threat to national security or, having been convicted of a particularly serious crime, a danger to the community. Such an exception speaks against art 33 being able to achieve the status of a jus cogens norm [of international law], against which no derogation is permitted.

Thus, independent of the question of whether Article 33’s non-refoulement obligation applies outside the territory of a nation is the question of whether the Convention allows for derogation of non-refoulement. The answer comes down to the implied exception postulated by Hathaway.

It's All About that Frontier

Delegates to the Conference of Plenipotentiaries were also laser-focused on ensuring that the Convention not obligate their countries to permit the entry of aliens who show up at the frontier of their countries claiming to be refugees. However, Louis Henkin stated in his Sale affidavit that:

I know no basis for th[e] view [that the terms of Article 33 do not extend to persons located outside the territory of the relevant nation]. There is nothing to support it in the language of Article 33 or in the record of the proceedings of the Committee that drafted it. I do not recall any expression to that effect by any member of the Committee.

Well, Henkin obviously missed the goings-on at the Conference of Plenipotentiaries, where the Swiss delegate stated that “States were not compelled to allow large groups of persons claiming refugee status to cross its frontiers” and the Dutch delegate had stated that “The Netherlands could not accept any legal obligations in respect of large groups of refugees seeking access to its territory.”

In 1987, Judge Harry Thomas Edwards wrote in an opinion concurring in part and dissenting in part in the D.C. Circuit’s opinion in Haitian Refugee Ctr. v. Gracey that:

The negotiating history of this convention reveals that delegates to the Conference of Plenipotentiaries at which it was adopted considered and rejected the argument … that Article 33 [always] prohibits a contracting state from returning a refugee to a country in which he or she would be endangered. They did so by agreeing that “expulsion” would refer to a “refugee already admitted into a country” and that “return” would refer to a “refugee already within the territory but not yet resident there.” Thus, the Protocol was not intended to govern parties’ conduct outside of their national borders.

The delegates further agreed that the “possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33.

Then, in Sale, the Supreme Court concluded that:

The text of Article 33 … fits with Judge Edwards’ understanding that “‘expulsion’ would refer to a ‘refugee already admitted into a country’ and that ‘return’ would refer to a ‘refugee already within the territory but not yet resident there.’ Thus, the Protocol [through its incorporation of the Convention] was not intended to govern parties’ conduct outside of their national borders.”… From the time of the Convention, commentators have consistently agreed with this view.

Whether or not it was fair to say that commentators “consistently” agreed with this view, the Court cited Nehemiah Robinson — “Art. 33 concerns refugees who have gained entry into the territory of a Contracting State, legally or illegally, but not to refugees who seek entrance into [the] territory”— and Atle Grahl-Madsen — “[Non-refoulement] may only be invoked in respect of persons who are already present — lawfully or unlawfully — in the territory of a Contracting State” and Article 33 “does “not obligate the Contracting State to admit any person who has not already set foot on their respective territories”.

The Court also noted that “[t]he Convention's failure to prevent the extraterritorial reconduction of aliens has been generally acknowledged (and regretted)[]”, citing commentators including Prince Sadruddin Aga Khan:

Does the non-refoulement rule ... apply to refugees who present themselves at the frontier or only to those already within the territory of the Contracting State? One would like … to be able to opt immediately in favor of the broader interpretation … . [But] States do not see it that way ... . There is thus a serious gap in refugee law as established by the 1951 Convention … and it is high time that this gap should be filled.

And, again, Nehemiah Robinson said: “If a refugee has succeeded in eluding the frontier guards, he is safe; if he has not, it is his hard luck. It cannot be said that this is a satisfactory solution of the problem of asylum.”

The Court also cited a 1949 UN “Study on the Position of Stateless Persons presented by the Secretary-General”. As to that study, DOS stated in its U.S. Observations that:

The study had used the term “expulsion” to mean “the juridical decision taken by the judicial or administrative authorities whereby an individual is ordered to leave the territory of the country.”… The study used the term “reconduction,” which it regarded as the equivalent of “refoulement,” to mean “the mere physical act of ejecting from the national territory a person who has gained entry or is residing therein irregularly.” The study explicitly opined that “reconduction” did not “signify the act of preventing a foreigner who has presented himself at the frontier from entering the national territory.”… The [Ad Hoc] Committee on Statelessness later replaced the term “reconduction” with “return,” which expresses the same sentiment as reconduction -- the mere act of ejection without the accompanying judicial process that is implicit in the term “expulsion.” The interpretation of these terms by the UN study thus confirms that Article 33 originated in an intention only to bar removal of individuals from within the Contracting State’s territory.

To be fair, what the study said was that “[a]lthough terminology varies to some extent on this point, the term “refoulement” is not used in this study to signify the act of preventing a foreigner who has presented himself at the frontier from entering the national territory.” That statement is considerably less emphatic than “explicitly opin[ing]”.

It Depends On What the Meaning of “Expel”/“Return” Is

The Supreme Court noted in Sale that “At a negotiating conference of plenipotentiaries … the Swiss delegate explained his understanding that the words ‘expel’ and ‘return’ covered only refugees who had entered the host country,” concluding that:

[E]ven if we believed … the intent of some signatory states [was] to protect all aliens, wherever they might be found, from being transported to potential oppressors, we must acknowledge that other signatory states carefully — and successfully — sought to avoid just that implication. The negotiating history, which suggests that the Convention's limited reach resulted from a deliberate bargain, is not dispositive, but it solidly supports our reluctance to interpret Article 33 to impose obligations on the contracting parties that are broader than the text commands.

Justice Blackman emphatically disagreed:

The ordinary meaning of “return” is “to bring, send, or put (a person or thing) back to or in a former position.”... [B]y taking instead as its starting point the assumption that “return,” as used in the treaty, “has a legal meaning narrower than its common meaning,” … the majority leads itself astray.

The straightforward interpretation of the duty of non-return is strongly reinforced by the Convention's use of the French term “refouler.” The ordinary meaning of “refouler,” as the majority concedes … is “to repulse, ...; to drive back, to repel.” Larousse Modern French-English Dictionary … . Thus construed, Article 33.1 … reads: “No contracting state shall expel or [repulse, drive back, or repel] a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened[.]”

But as DOS argued in U.S. Observations (noting that the document’s “arguments [were] largely drawn from the submissions of the United States to the United States Supreme Court in Sale….”):

• The most natural reading of this language is that it expresses a prohibition against removal of a refugee from the Contracting State to a foreign territory in the specified circumstances, irrespective of the manner in which the removal might be accomplished. “Expel” means “to force or drive out.”… Similarly, one meaning for the French word “refouler” is “expel (aliens).”… Under this meaning, “return (‘refouler’),” like “expel,” connotes not merely transfer, but instead ejection of an alien from within the territory of the Contracting State.

  As for any suggestion that “return (‘refouler’)” must have some other meaning … in order to avoid redundancy in the terms “expel” and “return (‘refouler’),” an examination of the two terms indicates that they are not in fact redundant. “Expulsion” in this context typically connotes “the formal process whereby a lawfully resident alien may be required to leave a state, or be forcibly ejected therefrom,” as opposed to removal of an individual who is not lawfully resident … . Read in this way, “expel” does not reach all classes of aliens present in a country — it covers only those cases in which the alien’s presence is lawful. Accordingly, the term following “expel” in Article 33, “return (‘refouler’),” rounds out the prohibition by ensuring that aliens whose presence in the country is unlawful are also covered. There is nothing in the text or context to suggest that the use of both “expel” and “return (‘refouler’)” was meant to encompass individuals within a State’s territory as well as individuals outside of a State’s territory.

• UNHCR’s analysis is … incorrect in its suggestion that, by adopting the particular language that it did, the Conference of Plenipotentiaries must have meant to approve only the Dutch delegate’s understanding that a Contracting State would have no obligation to accept a mass migration of refugees across its borders, but not the specific meaning he attached to the terms in the text … dictating that Article 33.1 related only to refugees who had already entered the territory of the Contracting State…. This argument ignores the rationale of the Dutch delegate’s conclusion — that there would be no obligation to accept a mass migration because he and the other delegates agreed with the Swiss delegate’s underlying interpretation of both “expel” and “return (‘refouler’)” as applying only to aliens who had already entered the territory of the Contracting State.

And DOS concluded upon reviewing the negotiating history that “Read together, the words ‘expel’ and ‘return (refouler)’ in Article 33.1 can thus only be understood to embody a deliberate decision by the Contracting States to incorporate a territorial limitation into the Convention’s provision on non-refoulement.”

Judge John Walker, Jr., the dissenting judge in McNary, added another consideration:

• The discussion which immediately followed the President's ruling [regarding agreement with the Dutch delegate’s interpretation] clears up any residual ambiguity:

  “Mr. Hoare (United Kingdom) remarked that the Style Committee had considered that the word 'return' was the nearest equivalent in English to the French term 'refoulement'. He assumed that the word 'return' as used in the English text had no wider meaning.

  “The President suggested that in accordance with the practice followed in previous Conventions, the French word ‘refoulement’ (‘refouler’ in verbal uses) should be included in brackets and between inverted commas after the English word ‘return’ wherever the latter occurred in the text.”

• The … suggestion[] made by the President w[as] adopted unanimously… .

  These statements evidence the delegates' unanimous agreement that “return (‘refouler’)” be limited to non-resident aliens within a contracting state's territory. Only by reading Mr. Hoare's “refouler” as different from Baron van Boetzelaer's can one escape this conclusion. Both, however, were attending the same conference, and Mr. Hoare, who spoke immediately following the Baron, expressed no disagreement with his interpretation. Thus, there is no basis to suggest that they were interpreting the word differently.

However, Timothy Lynch, professor at the University of Missouri-Kansas City School of Law, wasn’t sure what to make of the inclusion of “refouler” in the English text:

• [Regarding] the inclusion of the parenthetical “refouler” after the word “return” in the English text[, n]o doubt it is a curious inclusion, and, in fact, it was not a word that was used in the English drafts until it was inserted only moments before the final text of the provision was adopted and only an hour or two before the final text of the entire Convention was adopted.

• At one point, midway through the [C]onference [of Plenipotentiaries], while discussing the text of the non-refoulement provision, the Swiss representative referred to the term “refoulement” and asserted that the term “could not be applied to a refugee who had not yet entered the territory of the country.” Later, in the afternoon of the last day of negotiations after three straight weeks of negotiating the final text of the Convention, the representative from the United Kingdom asserted that he considered the word “return” in Article 33.1 to have no wider meaning than the French word “refoulement,” as that term had been explained by the Swiss representative. Then immediately thereafter, it was decided at the suggestion of the conference president that the French verb “refouler” would be inserted in parentheses in the English version of Article 33.1 after the word “return” “in accordance with the practice of previous Conventions.”

Lynch concluded that “It is challenging to understand the minds of the various plenipotentiaries and the conference president regarding this last-minute insertion by reading the summary of their negotiations.” He then even surmised that “one gets the sense reading the summary proceedings of the three-week conference that by the end the plenipotentiaries were tired and eager to finish their task, adopt a final text, and conclude their weeks of negotiation and get out of Geneva.”

Finally, Lynch has argued that the representatives and delegates negotiating the Convention were concerned about any obligation to permit the entry of aliens arriving at the frontier — mass influx or not.

[L]ater commentators have often made the argument that these exchanges among the plenipotentiaries leads to the conclusion that the Convention’s non-refoulement provision contains a non-rejection-at-the-frontier element for individual refugees but not for large numbers of refugees. The records of these negotiations do not support such a conclusion. Indeed, it is impossible to read into the plain language of Article 33.1 any allowance for affording different treatment to individuals within a mass influx compared to individuals without a mass influx. The text does not support such a conclusion. There is simply no reference at all to mass influxes or other large groups of refugees in Article 33.1. Instead, these negotiations strongly suggest that there is no non-rejection element at all and that the negotiating States were particularly concerned with ensuring they avoided incurring any obligations regarding mass migrations. [Emphasis in original.]

Of course, the negotiating countries were particularly concerned with ensuring they avoided incurring any obligations regarding mass migrations. One could argue that while Article 33 itself doesn’t differentiate between mass and minor influxes, only mass influxes could justify derogation of non-refoulement obligations.

Did the Supreme Court Really Mean That?

The Supreme Court in Sale ruled that “We do not read [Article 33] to apply to aliens interdicted on the high seas.” But the majority opinion muddied the water when it stated in dicta (as Justice Blackman reminded the other justices) that “these translations imply that ‘return’ means a defensive act of resistance or exclusion at a border rather than an act of transporting someone to a particular destination”. Justice Blackman jumped on this inconsistency, stating that “the majority agrees that the Convention does apply to refugees who have reached the border … . The majority thus cannot maintain that Van Boetzelaer's interpretation prevailed.”

Did the Court actually mean that Article 33 doesn’t apply to aliens interdicted on the high seas but does apply to aliens “repulse[d]”, “repel[led]” or “drive[n] back” at the border — such as by means of a border wall? This seemingly cannot be, because the Court had just concluded that “the Protocol was not intended to govern parties’ conduct outside of their national borders”. As Lynch has noted:

On the one hand, the Court makes it clear that the Convention has no extraterritorial application, that it does not govern a State party’s behavior outside its borders, and suggests the Convention would apply only to refugees “already admitted into a country” and those who are “already within the territory but not yet resident there.” On the other hand, [it] suggests that … Article 33 … might include a prohibition against ‘defensive acts of resistance or exclusion at a border[]’ [if it would result in refoulement].

Giving the Sale majority the benefit of the doubt, I would suggest that what it might have meant by “defensive act of resistance or exclusion at a border” was acts countering aliens who had already penetrated the border, who had already crossed the frontier. Then, there would be no inconsistency in the Court’s statements.

Humanitarian Concerns

Various objections have been raised to the conclusion that the Convention and the Protocol were not intended to govern signatory countries’ conduct outside of their national borders. One has been that such a conclusion clashes with the humanitarian goal of the Convention.

Along with his contention that an alien “must [never] be turned back to a country where his life or freedom could be threatened”, Louis Henkin also famously stated at the Ad Hoc Committee’s 20th meeting that:

It did not … follow that the convention would not apply to persons fleeing from persecution who asked to enter the territory of the contracting parties. Whether it was a question of closing the frontier to a refugee who asked admittance, or of turning him back after he had crossed the frontier, or even of expelling him after he had been admitted to residence in the territory, the problem was more or less the same.

In its Response to the Questions Posed, UNHCR argued that:

[I]t would be difficult and even illogical to conclude that Paragraph 1 … only applies to refugees present in the territory of a Contracting State and not to refugees who present themselves at the frontier, even if their rejection would oblige them to return to a territory where their life or freedom would be threatened. Stated in other terms, a refugee in the territory would enjoy all the guarantees of the Article … whereas if he presents himself at the frontier his plight and the dangers facing him in the event of return could simply be disregarded. Such a conclusion would be wholly artificial and inconsistent with the article's humanitarian purpose.

But as the Supreme Court in Sale simply stated:

[A] treaty cannot impose uncontemplated extraterritorial obligations on those who ratify it through no more than its general humanitarian intent. Because the text of Article 33 cannot reasonably be read to say anything at all about a nation's actions toward aliens outside its own territory, it does not prohibit such actions.

And as DOS argued:

UNHCR additionally argues that any interpretation that construes Article 33.1 as not extending to actions taken with respect to aliens outside of a Contracting State’s territory “would be fundamentally inconsistent with the humanitarian object and purpose of the 1951 Convention and its 1967 Protocol.”… [T]he fact that a treaty may be characterized as generally serving a humanitarian purpose cannot support an effort many years after its entry into force to rewrite the treaty by stretching its terms without limit so long as the final result can be described as serving a humanitarian purpose.

The Convention Failed to Spell Out that Non-Refoulement Applies at the Frontier

The negotiators could have — but didn’t — specify that non-refoulement applied to aliens at the frontier — despite the facts that 1) they were aware of a prior treaty that had done so; 2) an early draft of the Convention proposed by the UN’s Secretary General had done so; and 3) the French representative to the Ad Hoc Committee had requested that they do so.

At the Ad Hoc Committee’s 20th meeting, Pierre Ordonneau, the French representative, stated:

Mr. Henkin had been correct in emphasizing that the expulsion of refugees could not be regulated without dealing at the same time with the question of refusal of admittance, which raised a serious legal and humanitarian problem … .

[T]o be complete and clear, the article should include a first paragraph dealing with non-admittance in general, a second on expulsion in general, a third on non-admittance at the frontier and expulsion to territories where the life and liberty of the refugee might be threatened, and a last paragraph on the special measures to be applied in connexion with refugees unable to obey the expulsion order. 

And at the Committee’s 21st meeting the next day (February 2, 1950), Ordonneau “pointed out that article 24 as originally drafted by the Secretariat had covered both expulsion and non-admittance” and that “The reference to non-admittance at the frontier (refoulement) had been omitted from the new draft, which was thus incomplete.”

The Secretary-General had submitted a memorandum dated 3 January, 1950, to the Ad Hoc Committee on statelessness and related problems containing the text of a draft Convention to advance the Committee’s discussion. The first paragraph of the secretary general’s draft article proposed that:

Each of the High Contracting Parties undertakes not to remove or keep from its territory, by application of police measures, such as expulsions or non-admittance at the frontier (refoulement) refugees (and stateless persons) who have been authorized to reside there regularly, unless the said measures are dictated by reasons of national security or public order. 

The Secretary General’s memo commented that:

The sovereign right of a State to remove or keep from its territory foreigners regarded as undesirable cannot be challenged.

Nevertheless, expulsion or non-admittance at the frontier are serious measures in any event; they are especially serious in the case of a refugee who cannot be sent back to his country of origin and whom no other State can be compelled to accept. There is little likelihood that a foreign country will consent to receive a refugee whose expulsion has been ordered and who is thereby stamped as an undesirable. As every frontier is barred to a refugee whose expulsion has been ordered, only two possibilities are open to him, either not to obey the order and to go into hiding to avoid being caught or to cross a frontier illegally and clandestinely enter the territory of a neighbouring country. In that country too he must go into hiding to avoid being caught. In either case, after a certain time he is discovered, arrested, prosecuted, sentenced and escorted to the frontier after serving his sentence. Caught between two sovereign orders, one ordering him to leave the country and the other forbidding his entry into the neighbouring country, he leads the life of an outlaw and may in the end become a public danger.

I[n] this way measures of expulsion or non-admittance at the frontier, intended to protect law and order, achieve opposite results when an attempt is made to apply them to refugees without taking into account their peculiar position. 

And the article’s third paragraph provided that:

Each of the High Contracting Parties undertaken in any case not to turn back refugees to the frontiers of their country of origin, or to territories where their life or freedom would be threatened on account of their race, religion, nationality or political opinions.

The memo commented on this text:

The text of paragraph 3 reproduces that of the 1933 Convention (Article 3, paragraph 2 [“It undertakes in any case not to refuse entry to refugees at the frontiers of their countries of origin.”] but with an addition which takes into account not only the country of origin, but also the other countries where the life or freedom of the refugee would be threatened for the same reasons.

Of course, there is a crucial difference between “not to refuse entry to refugees at the frontiers of their countries of origin” (the 1933 Convention) and “not to turn back refugees to the frontiers of their country of origin” (the Secretary General’s draft Convention). The former only relates to aliens at the frontier of the country into which they seek entry, while the latter does not.

In reply to a question by Ordonneu, Giraud from the UN Secretariat’s Human Rights Division “explained that the original draft of article 24 had been based on article 3 of the 1933 Convention”. In fact, the title of the draft Article 24 in the Secretary’s memo noted “(See 1933 Convention, Article 3)”. What was the 1933 Convention? Guy Goodwin-Gill and Jane McAdam explained in their book The Refugee in International Law,² that “[n]ot until after the First World War … did international practice begin to accept the notion of non-return, and only in 1933 does the first reference to the principle that refugees should not be returned to their country of origin occur in an international instrument”, referring to the League of Nations’ 1933 Convention Relating to the International Status of Refugees. Article 3 of the 1933 Convention provided in part that:

Each of the Contracting Parties undertakes not to remove or keep from its territory by application of police measures, such as expulsions or non-admittance at the frontier (refoulement), refugees who have been authorized to reside there regularly, unless the said measures are dictated by reasons of national security or public order. 

UNHCR had concluded in its Response to the Questions Posed that “The wording of Article 3 of the 1933 Convention made it clear that the principle of non-refoulement as therein defined also applied to rejection at the frontier.” Lynch noted that:

This treaty [no longer in force] was only ratified by nine countries, one of which, the United Kingdom, made an explicit reservation and refused to consent to the obligation “not to refuse entry to refugees at the frontier of their countries of origin.”… The lack of widespread participation and the explicit refusal by the United Kingdom to consent to non-rejection at the frontier are telling.

Lynch said that even “more importantly this 1933 treaty demonstrates that the issue of non-rejection at the frontier was a known issue among States, diplomats, and international humanitarian and refugee circles even in 1933”. The Ad Hoc Committee’s report of February 17, 1950, stated that:

In drafting this convention the Committee gave careful consideration to the provisions of previous international agreements. It sought to retain as many of them as possible in order to assure that the new consolidated convention should afford at least as much protection to refugees as had been provided by previous agreements.

The Ad Hoc Committee made two telling comments on its draft Convention:

• It was the opinion of the Committee that the obligation not to return a refugee to the country where he was persecuted did not imply an obligation to admit him to the country where he seeks entry. The return of a refugee-ship, for example, to the high seas could not be construed as a violation of this clause. 

• The Convention of 1933 contains a provision of this kind.

UNHCR concluded in its Response to the Questions Posed that “The reference to the 1933 Convention in the Secretary-General’s Memorandum … and in the Report of the Ad Hoc Committee … would seem to indicate that there was to be no deviation from the standard laid down in the pre-war instruments.” But UNHCR’s conclusion ignores the inconvenient truths that the Committee displayed conscious knowledge of the 1933 Convention’s language but still failed to include the text “non-admittance at the frontier (refoulement)” and “not to refuse entry to refugees at the frontiers of their countries of origin”. The Committee’s draft hardly contained a provision of the “kind” contained in the 1933 Convention.

DOS concluded in its U.S. Observations that Article 33 “contains no express statement or other affirmative indication that it was intended to impose obligations on a contracting state outside its own territory.”

As Lynch put it:

• [T]he notion of non-rejection at the border was understood as an issue well before the adoption of the Refugee Convention, yet the explicit recognition of such a notion is conspicuously absent from the Convention.

• The drafting committee chose to omit the principle of non-rejection at the frontier in its non-refoulement provision … . [T]he adopted text did not contain any explicit language addressing admission or non-rejection at the frontier.

Thus, the representatives and the delegates negotiating the Convention failed to address the issue of non-rejection at the frontier, even though the 1933 Convention and the Secretary’s draft had addressed the issue (if only as to refugees authorized to reside regularly in a country or at the frontiers of their countries of origin) and the French representative had urged them to do so. The Ad Hoc Committee had even commented that “return of a refugee-ship … to the high seas could not be construed as a violation of th[e non-refoulement] clause.”

Why Did the Refugee Convention Specify with Regard to Other Articles that They Applied Only to Aliens Within a Nation’s Territory?

The obverse of the question as to why Article 33 did not specify its extraterritorial applicability is why other articles did in fact specify that they applied to only to aliens within a country’s territory. In Haitian Ctrs. Council v. McNary, the Second Circuit observed that:

[O]ther articles of the Refugee Convention that have a more limiting effect on the term “refugee”. See, e.g., Article 4 (“refugees within their territories”); Article 15 (“refugees lawfully staying in their territory”); Article 17.1 (same); Article 18 (“refugee lawfully in their territory”); Article 19.1 (“refugees lawfully staying in their territory”); Article 21 (same); Article 23 (same); Article 24.1 (same); Article 26 (“refugees lawfully in its territory”); Article 27 (“refugee in their territory”); Article 28 (“refugees lawfully staying in their territory”); Article 31.1 (refugees who “enter or are present in their territory without authorization”); Article 32.1 (“refugee lawfully in their territory”).

The court concluded that:

• Article 33.1's prohibition against “return” plainly applies to all refugees, regardless of location. This reading is borne out by the language used in other articles of the Refugee Convention that have a more limiting effect on the term “refugee”.

• Had the parties to the Refugee Convention meant to limit [the] application [of the non-refoulement obligation] in that way, we would expect a wording of that section in line with, for instance, Article 4 (“refugees within their territories”). But the contracting states did not so limit Article 33.1.

However, Lynch explains away this apparent anomaly by noting that the articles that do specify their applicability only to aliens within a country’s territory “only serve to distinguish the rights of refugees who are lawfully within the territory of a State from the rights of those who are unlawfully within the territory of the State”. (Emphasis in original.)

DOS went further and argued that the language of these other articles actually supports the contention that Article 33 does not apply outside a nation’s territory:

These other references to the territorial scope are consistent with an overall reading of the instrument that, absent some express provision to the contrary, the Convention applies within the territory of Contracting State. … [The reading of the Convention that because] certain provisions include a requirement that the refugee must be within the territory of the Contracting State … [while] the territorial scope of Article 33.1 is not made explicit … there is no territorial limitation [in Article 33.1] … is not a natural reading of a treaty text, nor would it be reasonable to impose on treaty drafters a reading that every provision of a treaty would apply extraterritorially absent an express limitation in its text. Such a reading would be particularly illogical where the only express indicators of the drafters’ intentions all show the intent that the instrument would not apply extraterritorially… [and] is flawed because it fails to acknowledge that the provisions that include a requirement that the refugee must be within the territory of the Contracting State demonstrate that the territorial limit of the Convention is evident throughout the Convention. The logical conclusion … is not that Article 33.1 does not carry any territorial limits; it is, to the contrary, that in context, Article 33.1, like these other provisions in the Convention, is limited to the territory of the Contracting State. [Emphasis in original.]

I assess DOS’s argument as quite weak. If the drafters’ intent was that the Convention as a whole would not apply extraterritorially, they would have most likely included a provision up front stating this fact. That, at least, is what I would have done in drafting legislation. I believe Lynch’s response to be much more powerful.

What to Do About Paragraph 2?

The Supreme Court in Sale contended that the language of Article 33’s paragraph 2 bolstered the Court’s conclusion that the non-refoulement obligation did not apply outside of a nation’s territory:

Under the second paragraph of Article 33 an alien may not claim the benefit of the first paragraph if he poses a danger to the country in which he is located. If the first paragraph did apply on the high seas, no nation could invoke the second paragraph's exception with respect to an alien there: An alien intercepted on the high seas is in no country at all. If Article 33.1 applied extraterritorially, therefore, Article 33.2 would create an absurd anomaly: Dangerous aliens on the high seas would be entitled to the benefits of 33.1 while those residing in the country that sought to expel them would not. It is more reasonable to assume that the coverage of 33.2 was limited to those already in the country because it was understood that 33.1 obligated the signatory state only with respect to aliens within its territory.

And DOS argued that:

[P]aragraph 2 of Article 33 confirms that paragraph 1 … applies only to aliens inside the territory of a Contracting State. Paragraph 2 states that the benefit of Article 33 may not be claimed by a refugee who is a danger to the security of “the country in which he is.” This paragraph, the only reference to territory in the Article, contemplates that a refugee is covered only if he is “in” a “country” of refuge. Accordingly, read as a whole, Article 33 applies only to removal by a Contracting State of a refugee who is within the territory of that Contracting State.

Justice Blackman disagreed with these contentions:

The signatories’ understandable decision to allow nations to deport criminal aliens who have entered their territory hardly suggests an intent to permit the apprehension and return of noncriminal aliens who have not entered their territory … . One wonders what the majority would make of an exception that removed from the Article's protection all refugees who “constitute a danger to their families.” By the majority's logic, the inclusion of such an exception presumably would render Article 33.1 applicable only to refugees with families.

Far from constituting “an absurd anomaly,”… the fact that a state is permitted to “expel or return” a small class of refugees found within its territory but may not seize and return refugees who remain outside its frontiers expresses precisely the objectives and concerns of the Convention. Nonreturn is the rule; the sole exception … is that a nation endangered by a refugee's very presence may “expel or return” him to an unsafe country if it chooses. The tautological observation that only a refugee already in a country can pose a danger to the country “in which he is” proves nothing.

However, while that tautological observation may not prove anything, it is still the case that an alien outside a country’s territory may still pose a sufficient danger to that country that the country would not want to be forced to permit the alien to enter. Thus, if the Convention’s drafters’ Article 33’s first paragraph were to apply outside of a country’s territory, it would not have limited the applicability of paragraph 2 to aliens within its territory. That, at least, is what I would have done in drafting legislation.

UNHCR argued in Response to the Questions Posed that:

[Paragraph 2] cannot in any way be taken as an indication that the Article has no application to rejection at the frontier. It can rather be taken to support the opposite conclusion. It would indeed be wholly inconsistent to provide a refugee in the territory of a Contracting State with all the safeguards contained in paragraph 2 while at the same time giving Contracting States an unqualified right to reject them at the frontier and to return them to a country of persecution.

While that may be wholly inconsistent, it would be a perfectly reasonable inconsistency if the drafters of the Convention intended to treat refugees within a country’s territory inconsistently with refugees outside a country’s territory.

Did the Concern About a Mass Influx End After the Masses Crossed the Frontier?

It is certainly controversial to maintain that the drafters of the Refugee Convention intended signatory nations to have no Article 33 non-refoulement obligations regarding masses of aliens at their frontiers claiming to refugees. It would be even more controversial to maintain that that the drafters of the Refugee Convention intended signatory nations to have no Article 33 non-refoulement obligations regarding masses of aliens who have already entered their territory claiming to be refugees. But there is strong evidence for the latter contention.

If a nation’s non-refoulement obligations apply only to masses of aliens who have not yet entered the nation’s territory, the drafters’ distinction between mass influxes and “attempted” mass influxes, and their use of the word “across” in the negotiations, would make no sense. Remember that Baron van Boetzelaer, the Dutch representative, stated during the deliberations that:

• According to th[e Swiss] interpretation, article 28 would not have involved any obligations in the possible case of mass migrations across frontiers or of attempted mass migrations.” 

• [H]e wished to have it placed on record that the Conference was in agreement with the interpretation that the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33.

And, remember that the president ruled that “the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33.” 

Merriam-Webster defines “across” (when used as a preposition) first as “from one side to the opposite side of : OVER, THROUGH”, giving as an example “swam across the river”, and as second “on the opposite side of”, giving as an example “lives across the street from us”. Thus, the plain meaning of the phrase “across frontiers” is traversing from one side of a country’s frontier to the other, i.e. actually entering the country’s territory. And this, in turn, provides a meaningful distinction between mass influxes and “attempted” mass influxes. A mass influx is a successful penetration of the frontier, the border, while an attempted mass influx describes an unsuccessful attempt to penetrate the frontier, the border. Consequently, the non-refoulement obligation would not apply to a group of aliens who have succeeded in a mass migration across a country’s frontier and are now present in the country’s territory. And the obligation would not apply to a group of aliens “attempting” a mass migration — who have reached a country’s frontier but have not yet succeeded in penetrating the frontier.

Andrew Schoenholtz’s use of the phrases “may have crossed or are attempting to cross the border” and “let in (across frontiers)” in his commentary provide support for my reading of “across”. So do Goodwin-Gill and McAdam in their use of the phrase “mass movement to or across an international frontier”, as does Professor Koh in his use of the phrase “let all of the refugees in (‘across frontiers’)”, as do the respondents in Sale who used the phrase in their Brief for Respondents “let all of the refugees in (‘across frontiers’).”

Treating for non-refoulement purposes groups of aliens who have succeeded in crossing a nation’s frontier and those at the frontier attempting to cross makes sense on a policy level. Louis Henkin stated during the deliberations of the Ad Hoc Committee that “Whether it was a question of closing the frontier to a refugee who asked admittance, or of turning him back after he had crossed the frontier, or even of expelling him after he had been admitted to residence in the territory, the problem was more or less the same.” And he stated in his Sale affidavit that “There was no thought of distinguishing … between those who had already entered a state's territory and those who were merely knocking on its doors.” Of course, Henkin did not want there to be any exemption whatsoever from the non-refoulement obligation, but he didn’t believe there was any fundamental difference between removing refugees who had already entered a nation’s territory and those still seeking entry.

Did the drafters of the Convention intend for signatory nations to not have a non-refoulement obligation to permit refugees or purported refugees to remain who had already entered their territory en masse? Andrew Schoenholtz has written that the “mass migration concern” exists regardless of whether a group of aliens has entered a country’s territory or is at its frontier:

[M]ass migration has no bearing on the issue of where a refugee must be in order to trigger the nonrefoulement obligation of Article 33. As the Dutch delegate indicated, those involved in a mass migration may cross or attempt to cross borders, so that they may be inside the territory of a country of refuge or outside those borders. The mass migration concern is simply not one about a refugee's location in relation to a country of refuge, but rather focuses on the number of refugees that may have crossed or are attempting to cross the border. [Underlined text was emphasized in the original.]

And Theodoli, the Italian representative on the Ad Hoc Committee, “personally felt that a State could not commit itself not to expel or return large groups of refugees who presented themselves on its territory, and who might endanger public security”.

On the other hand, the Swiss delegate had stated that “[T]he word [‘return’] applied solely to refugees who had already entered a country, but were not yet resident there. According to that interpretation, States were not compelled to allow large groups of persons claiming refugee status to cross [their] frontiers.” And the Dutch delegate had stated that “The Netherlands could not accept any legal obligations in respect of large groups of refugees seeking access to its territory.” Were they in fact concerned only about groups of aliens at but not yet across their countries’ frontiers? Yet, what was placed in the record was that “the Conference was in agreement with the [Dutch representatives’] interpretation that the possibility of mass migrations across frontiers or of attempted mass migrations was not covered by article 33”, which lends itself to being read as encompassing both groups of aliens at but not yet across a nation’s frontier and groups who have already successfully crossed.

I must conclude that the evidence is inconclusive.

Does Non-Refoulement Require a Country to Admit or Grant Asylum to an Alien?

UNHCR concluded in its Advisory Opinion that the representatives’ and delegates’ statements of concern about mass influxes were simply “expressions of concern related to a possible obligation to grant asylum to large numbers of arrivals in mass influx situations”. Additionally, Goodwin-Gill and McAdam have written that “States were not prepared to include in the Convention any article on admission of refugees, non-refoulement in the sense of even a limited obligation to allow entry may well have been seen as coming too close to the unwished-for duty to grant asylum.” Henkin stated in his Sale affidavit that at the Ad Hoc Committee’s meeting on February 10, 1950, “I expressed agreement that Article 33 imposed no requirement of admission [probably in the sense of granting lawful status] to refugees” and “as I understood Article 33, and as I believe all members of the Committee understood Article 33, it amounted to this: a state would not be obliged to admit any refugee”.

However, the foregoing does not absolve countries of their non-refoulement obligation. Koh concluded that:

The position of the drafters … [was that] even if a country has no obligation under Article 33 to accept mass migrations for admission or asylum, “whatever the case might be ... [the refugee] must not be turned back to a country where his life or freedom could be threatened.”

Andrew Shoenholtz similarly explained that:

[T]he power to refuse admission is not to be confused with the power to return a refugee to a country of persecution. While the Convention does not limit a contracting state's power to refuse admission, it does generally remove that State's power to return refugees to countries of persecution.

Weis even suggested in his commentary that “The State admitting the refugee … may even expel him to another country willing to admit him.” And Louis Henkin similarly stated at the Ad Hoc Committee’s 20th meeting that “if the State concerned wished to get rid of the refugee at all costs, it could send him to another country or place him in an internment camp”.

In the real world, however it is often impossible to find a third country willing to accept such an alien. Sir Leslie Brass, the United Kingdom’s representative, explained to the Ad Hoc Committee on February 1, 1950, that “If no other country was willing to receive … a delinquent [refugee who “obstinately refused to abide by the laws of the country which had granted him hospitality”], he would perforce have to be sent to the only territory which would admit him, and which was, in fact, the country where his life or liberty might be in danger.” Leslie Chance of Canada, the Committee’s chairman, stated in regard to such refugees: “Practically speaking. There was nowhere where a refugee could be sent. Even without paragraph 1 the reception country would perforce have to keep such a refugee, as it could not possibly return him to a country in which his life would be in danger.” Even the UN secretary stated in his memo to the Ad Hoc Committee that “There is little likelihood that a foreign country will consent to receive a refugee whose expulsion has been ordered and who is thereby stamped as an undesirable.” Thus, regarding such aliens, there really is no difference between assuming a non-refoulement obligation and being obligated to admit a purported refugee at the frontier.

Does It Matter What the Drafters Were Thinking?

Does the intent of Article 33’s drafters matter if there is a plain meaning to the text of the Article?

Justice Blackman wrote in his dissent that “It is axiomatic that a treaty's plain language must control absent ‘extraordinarily strong contrary evidence,’” quoting the Supreme Court’s 1982 decision in Sumitomo Shoji America, Inc. v. Avagliano. Actually, a more complete quote from Avagliano provides a somewhat more nuanced meaning: “Our role is limited to giving effect to the intent of the Treaty parties. When the parties to a treaty both agree as to the meaning of a treaty provision, and that interpretation follows from the clear treaty language, we must, absent extraordinarily strong contrary evidence, defer to that interpretation.”

Blackman went on to state that “Reliance on a treaty's negotiating history (travaux preparatoires) is a disfavored alternative of last resort, appropriate only where the terms of the document are obscure or lead to ‘manifestly absurd or unreasonable’ results,” citing Article 32 of the Vienna Convention on the Law of Treaties. Actually, a fuller presentation of the Vienna Convention provides a quite different meaning. The first paragraph of the Vienna Convention’s Article 31 provides that “A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.” (Emphasis added.) And Article 32 states that:

Recourse may be had to supplementary means of interpretation, including the preparatory work of the treaty and the circumstances of its conclusion, in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31:

(a) leaves the meaning ambiguous or obscure; or

(b) leads to a result which is manifestly absurd or unreasonable. 

Justice Blackman went on to state that:

[E]ven the general rule of treaty construction allowing limited resort to travaux preparatoires “has no application to oral statements made by those engaged in negotiating the treaty which were not embodied in any writing and were not communicated to the government of the negotiator or to its ratifying body.” [citing the Supreme Court’s 1934 decision in Arizona v. California]. There is no evidence that the comment on which the majority relies was ever communicated to the United States Government or to the Senate in connection with the ratification of the Protocol.

But, of course, the president of the Ad Hoc Committee had ruled that “the interpretation given by the Netherlands representative should be placed on record”, and I am not aware of any evidence that the comment on which the majority relied was not communicated to the U.S government or to the Senate.

Even assuming that we should start and stop with the “plain meaning” of Article 33, we need to ask whether Article 33 has a plain meaning. Justice Blackman certainly thought that it does: “The terms are unambiguous. Vulnerable refugees shall not be returned. The language is clear, and the command is straightforward; that should be the end of the inquiry.” UNHCR agreed, stating in its Response to the Questions Posed that:

As regards [paragraph 1 of Article 33], the Office of UNHCR considers that its application to rejection at the frontier follows from its clear wording. It is difficult to conceive that the words “return” and “refouler” are not sufficiently broad to cover a measure of this kind and are limited to refugees who have already entered the territory of a Contracting State.

And Alice Edwards concluded that “The final version of art 33, on its plain reading, does not include [a mass influx] exception. In fact, the language … indicates that refoulement ‘in any manner whatsoever’ is not allowed.”

If the majority in Sale thought that Article 33 has a “plain meaning”, that meaning would be quite at odds with the meaning assumed by Justice Blackman and UNHCR. But it is not clear that the Court thought there to be such a plain meaning. The Court, after all, found that “the text … of Article 33 … [is] completely silent with respect to the Article’s possible application to actions taken by a country outside its own borders”.

Thus, it is entirely appropriate, and even necessary, to look to the intent of the drafters of the Refugee Convention, at the very least to analyze its terms “in their context and in the light of [the Convention’s] object and purpose” and to “confirm the meaning” of the text, if not in order to determine its meaning in the first place.

We might also want to look to the text of the Convention, only not the text that Justice Blackman had in mind. Article 1 of the Convention provides that an alien otherwise meeting the terms of the definition of a refugee must be “outside the country of his nationality” (unless having no nationality). Thus, as Lynch has pointed out:

[S]ince the Convention’s non-refoulement provision is only enjoyed by “refugees” as defined in the Convention, that is to say, by people who are at least outside their home countries, the Convention’s non-refoulement provision would expressly not apply to nationals of a bordering State who appear at that border asking for asylum but have not actually left their country.

Hathaway concurs: “Until and unless refugees actually leave their own state, they are not legally entitled to protection against refoulement, or to any other refugee rights.”

Additionally, Lynch has pointed out that:

[T]he ordinary meaning of the Convention’s non-refoulement provision does not easily accommodate an interpretation that would include a non-rejection-at-the-frontier obligation … . The text speaks of the “expulsion or return” of persons “to” the frontiers of threatening territories. If someone is at such a frontier and has not yet entered or been admitted to the destination country, they cannot be “expelled or returned” “to” that frontier, for they are already there.

Conclusion

The crisis in America’s asylum system has its roots in the Refugee Convention’s non-refoulement obligation. Paradoxically, the Convention’s drafters emphatically rejected the notion of adding a right to asylum. Nevertheless, the non-refoulement obligation itself is in many cases functionally indistinguishable from a right to asylum (or, at least, admission) because of the difficulty of finding third countries willing to accept aliens who claim to fear persecution in their homelands.

Thus, it is crucial to understand what the Convention’s non-refoulement obligation itself entails. In 1994, the U.S. Supreme Court in Sale v. Haitian Centers Council, Inc., concluded that the text of the Convention did not “apply to aliens interdicted on the high seas”, based on the Court’s textual analysis and its taking from the Convention’s negotiating history that “several delegates understood the right of non-refoulement to apply only to aliens physically present in the host country” and “[t]here is no record of any later disagreement with that position”.

The negotiating history in the Conference of Plenipotentiaries makes clear that delegates were laser-focused on ensuring that the Convention’s non-refoulement obligation not obligate their countries to allow the entry of mass influxes of aliens arriving at their frontiers claiming to be refugees. The negotiating history even suggests that the delegates wanted their countries to retain the option of removing such aliens who succeeded in penetrating their frontiers back to the countries where they claimed to fear persecution.

The United States now faces a mass influx of purported refugees and a historic and unsustainable backlog of well over a million such aliens in immigration court. The U.S. Congress should feel free to reform our nation’s immigration laws to the extent necessary to end this mass influx in regards to inadmissible aliens attempting to penetrate our borders, and even as to those who have succeeded in doing so, content in the knowledge that the drafters of the Refugee Convention worked to ensure that its non-refoulement obligation would not condemn their countries to having to endure mass influxes of aliens.

End Notes

¹ The affidavit can be found in the version of Harold Hongju Koh’s THE HAITIAN CENTERS COUNCIL CASE: Reflections on Refoulement and Haitian Centers Council (35 Harv. Int'l L.J. 1 (1994)) available on Lexis/Nexis.

² Guy Goodwin-Gill and Jane McAdam noted in The Refugee in International Law (3rd ed.) (2007), that “[a]s a matter of fact, anyone presenting themselves at a frontier post, port, or airport will already be within State territory and jurisdiction”. Lynch agreed:

[C]ountries’ external border security operations are normally either just within their borders or in an airport transit zone. That is to say, people who arrive at a country’s external borders and apply to enter a country are normally already within the territory of the destination country when they actually face border security personnel, often just slightly so. This on-the-ground reality is a function of the logistical and operational requirements of establishing and managing border control systems. Borders have no thickness … and countries normally situate such operations within their own territory, not within their neighbors’ territories.

However, Lynch was quick to note that “[I]t may be the case that internationally legally mandated non-refoulement rights are not triggered at ports of entry until someone passes through this security and inspection (or “transit” or “international”) zone, this Article … is … agnostic to this assertion.”