Basically no, the US uses the narrow definition of refugee in its domestic law.
Also, in the US case, complementary protection does not have a (single) name, although a catchall "withholding of removal" covers some of it, but also covers categories not reflected in international classifications. The temporary protected status (TPS) probably comes closest to meeting the extended refugee definition, but TPS has a distinctly group-based nature.
A 2014 comparative study concluded that
In Canada, complementary protection is well integrated into the RSD [refugee status determination] regime; in
the United States, it is unpredictable and highly fragmented; and in Australia,
complementary protection is a highly politicized discretionary power of the
Immigration Minister. [.....]
Like its RSD regime, the American complementary protection regime is incredibly
complex and multifaceted. There are almost a dozen different forms of relief
from deportation that have been layered onto existing legislative frameworks
over time, leading to an almost incomprehensible web of legal categories. Some
forms of relief lead to permanent residency status; many do not. Some forms are
connected to international legal obligations and others stem from domestic politics
or foreign policy objectives. As with the American RSD regime, the broader
world of humanitarian protection in the United States involves a lot of different
players, is highly legalistic, and has varied over time and presidential administrations.
As Kanstroom (2010:118) recently put it, “this patchwork quilt of specific
protective measures evidences certain deep underlying humanitarian urges that
have garnered specific political support.” In other words, the complexity of the
system is not the product of a holistic or coordinated vision for protection, but
of the ad hoc nature of migration policymaking in a fragmented institutional
landscape.
In fact, the United States has a major complementary protection avenue
that predates its RSD system. Before the 1980 Refugee Act established an
RSD process based on the Convention definition, American immigration law
included a provision that would prevent deportation of anyone who could demonstrate
a “clear probability of persecution” if returned to their home country
(Immigration and Nationality Act, Section 243(h)). This provision is called
withholding of removal, and prior to 1980, it was purely discretionary and was
not linked to a path to permanent residency or eventual citizenship. When the
Refugee Act inserted an RSD system based on the 1951 Convention definition
into American law, it layered this new RSD process onto the existing framework,
leaving the withholding of removal component intact. The only change the
Refugee Act made to the withholding of removal category was to shift it from
a discretionary form of relief to a mandatory form of protection. This change
was made in order to comply with Article 33 of the UN Convention—the nonrefoulement pledge.
The relationship between the new asylum provision and the old withholding
provision in the law was the subject of confusion for several years after the 1980
Act was passed. There is no evidence in the historical record surrounding the
passage of the Act to suggest that Congress deliberately intended to create two distinct forms of protection. Rather, this outcome seems to be an oversight; the
asylum provision was barely discussed in the hearings on the bill because it was
eclipsed by a focus on overseas resettlement issues—in particular, the immediate
crisis in Southeast Asia (Martin 1990, Hamlin and Wolgin 2012). It took two
separate Supreme Court decisions to resolve the question of whether the “clear
probability of persecution” standard of proof for withholding of removal was different
from the new “well-founded fear of persecution” standard the Act had created
for RSD (INS vs. Stevic, [467 U.S. 407, 1984] and INS vs. Cardoza-Fonseca,
[480 U.S. 421, 1987]) Initially, the BIA [Board of Immigration Appeals] found that the two standards were
“not meaningfully different and, in practical application, converge” (Matter of
Acosta [Interim Decision #2986, 1985]). By this logic, withholding of removal
had been subsumed into the new category of refugee protection. However, in
Cardoza-Fonseca, the Supreme Court concluded (without reference to any supporting
evidence) that because Congress had left the two forms of relief separate
from one another when they passed the 1980 Act, they intended for them to be
two distinct forms of relief.
Today, withholding of removal requests are considered at the same time as
an asylum claim in the same hearing, but they have split standards. Withholding
of removal has a greater burden of proof, and still does not result in a full legal
status. Thus, it becomes relevant when it is used as a safety net for asylum seekers
who do not meet the refugee definition but can show that they will be persecuted
if deported.
It goes on discussing withholdings under
- CAT [convention against torture]
- IIRIRA (seldom successful) if causing “exceptional and extremely unusual hardship” to an American citizen or permanent resident family member
- temporary protected status (TPS); this is probably the closest thing to a complementary protection as envisaged by UNHCR, and the most flexible, but TPS is also a fairly political decision as to which groups (rather than individuals) get this designation
- T and U visas: for victims and witnesses of human trafficking who are willing to testify
These withholding paths however are not all integrated in the RSD process. The Liberians TPS case is perhaps illustrative as to how complex and ad-hoc this can be:
Finally, in the cases of several national groups, TPS has been anything but
temporary. For example, a large group of Liberians had their TPS renewed
for decades with no pathway to permanent residency on the horizon, leaving
them in a legal limbo (Wasem and Ester 2008). When their TPS finally expired
in 2007, President Bush invoked his power of Deferred Enforced Departure
(DED), which had been created to replace EVD as a discretionary presidential
power to protect those who did not qualify for TPS (U.S. Department of
Homeland Security 2007). The Liberians have continued to have their DED
status renewed every 18 months since 2007, and although they can work legally
in the United States and often have citizen children, they are not on the path to
citizenship.
And concludes
Overall, the complementary protection system in the United States has been
cast as a wide net, with many crisscrossing threads, but also some large holes.
The RSD regime includes a channel for making CAT claims and complying
with nonrefoulement obligations, but the rest of the protection avenues have not
been integrated with that system. In addition, the long-term trend in the United
States is unclear. Some beneficiaries of complementary protection, such as the
Liberians with (not-so) temporary status, and the mostly female victims of
human trafficking, have received protection that would not have been available
to them 20 years ago. Others, such as Haitians, Central Americans, and undocumented
immigrants who cannot prove extreme hardship to a U.S. citizen, are
seeing their protection options contract. Thus, protection is not holistic and can
be very inconsistent over time, depending on whether the individual or national
origin group in question is politically sympathetic. [... Thus,] in the U.S. case, complementary protection
has diversified (albeit inconsistently) the circumstances that are officially recognized
as valid reasons for seeking protection.
A 2014 US study basically disagrees that a EU-style subsidiary protection is superior to TPS:
by requiring applicants to make an
individualized showing of harm, the subsidiary protection regime requires individualized
hearings and a fairly lengthy adjudication process in which an applicant’s credibility and
supporting documentation is assessed. (Wettergren and Wikstrom 2013, 569). A process
like TPS, in contrast, generally does not require a lengthy hearing, as individuals may be
approved or denied on the basis of more “objective” criteria. Given that the goal of TPS
is to provide short-term humanitarian relief, there is merit to retaining a process that is
relatively simple, straightforward, and can be implemented fairly quickly.
Of course, the problem with TPS is that in practice it seems to have proved the old adage that temporary solutions become permanent (a fact acknowledged in the same US study.) A 2018 update on TPS indicates that Trump has massively terminated the program on several countries in Latin America... but these decision were met with legal challenges. The main argument in these appears to be the fact that the administration seemingly lacks objective standards for making TPS determinations. (Ironically, that was considered a strength in the 2014 paper quoted above.) According to Vox, the Trump administration is expected to ultimately prevail on these matters at the Supreme Court, given the latter's deference toward executive privileges on matters of immigration.
And yes, there's an opposite scholarly opinion that TPS is worse than subsidiary protection:
While it extends protection beyond those who face persecution, the availability of TPS is severely limited; it applied only when US officials exercise their discretion to designate a country's nationals for TPS protection and it protects only those nationals already in the US. It falls far of subsidiary protection in the EU.
Interestingly, this latter article claims that in the Matter of LS (2012) the BIA essentially created a de facto US equivalent of subsidiary protection.