The Supreme Court of Korea’s Oct. 16 decision in the long-awaited NAVER Shopping case delivered a resounding defeat for the Korea Fair Trade Commission (KFTC). In finding that NAVER—a local search-based platform that competes with Google in Korea—had not violated Korean competition law, the court overturned both the KFTC’s January 2021 decision and the December 2022 decision of the Seoul High Court.
The court’s judgment is particularly notable for the broader antitrust community, because it represents—along with the social backlash against the Korean government’s platform-regulation initiatives of the early 2020s—one of the most significant pushbacks against the government’s recent pursuit of the EU’s regulatory approach to the digital economy.
Just as the previous attempts to transplant Digital Markets Act (DMA)-style regulation failed amid social resistance, the enforcer’s hasty effort to make a “self-preferencing case” (misaligned with Korea’s market and institutional context) has now been halted by this ruling. It is a rightful correction—one that offers valuable lessons for non-EU jurisdictions seeking to design their own digital competition governance frameworks.
Background
The facts and procedural history of the case largely resemble those in Google Shopping. NAVER, a search platform that was then dominant in the markets for general search and comparison-shopping services in Korea, was accused of abusing that dominance in the e-commerce marketplace (called “open-market service”) by favoring sellers that used its own e-commerce service (“Smart Store”) over others in the search results of its comparison-shopping service—e.g., by manipulating its search algorithm to lower the rankings of rivals and boost the visibility of Smart Store listings.
In January 2021, the KFTC sanctioned the conduct, imposing a corrective order and a fine of approximately KRW 26.7 billion, invoking the prohibitions against:
- discrimination as an abuse of dominance (for restricting competition) (then Art. 3-2(1)(3); now, Art. 5, Monopoly Regulation and Fair Trade Act (MRFTA));
- discrimination as an unfair trading practice (for lessening competition) (then, Art. 23(1)(1); now, Art. 45(1)(2), MRFTA); and
- deceptive inducement (another form of unfair trading practice) (then, Art. 23(1)(3); now, Art. 45(1)(4), MRFTA).
Roughly speaking, Korea’s unfair trading practice (UTP) provisions correspond to Section 5 of the U.S. Federal Trade Commission Act. They allow the KFTC to address, within the framework of competition-law enforcement, not only incipient anticompetitive conduct that can lessen competition, but also certain acts related to consumer protection or unfair competition, such as deceptive practices.
The Seoul High Court upheld the KFTC’s decision on appeal in December 2022. And as noted above, the Supreme Court recently overturned the lower court’s ruling and found against the KFTC.
Further Context
Beyond the factual background, however, there are several additional points that may help to understand the context of the Supreme Court’s judgment, which I’ve already discussed in my Kluwer Competition Law Blog post and on SSRN. I will offer a brief outline here, as well.
First, regarding the search bias issue, it is notable that the KFTC has a demonstrated pattern of following the EU’s actions against Google. In 2013, the enforcer opened investigations into local platforms NAVER and Daum, which were resolved in 2014 by consent decisions. These followed the European Commission’s commitment discussions with Google in 2013 and 2014.
The KFTC subsequently sanctioned NAVER in October 2020 for self-preferencing (the decision was formally dated Jan. 27, 2021), which followed the Commission’s June 2017 Google Shopping decision. That the KFTC’s action was influenced by the EU’s case is rather evident from both its press conference transcript and the enforcer’s decision itself.
But while the KFTC sought to replicate the European Commission’s approach, the underlying market realities were not themselves comparable. For example, the Commission found (see, e.g., T-612/17, paras 182-183) that Google held a deeply entrenched—almost “superdominant”—position in the general search market in the EU. By contrast, the KFTC’s decision made clear (No. 2021-027, paras 251 and 267) that NAVER—with around 70% market share—still faced competitive pressure in both Korea’s general search market (from Daum and Google) and the nation’s comparison-shopping market (from Daum, Danawa, and Enuri).
Moreover, from a dynamic perspective, the current market situation casts doubt even on the KFTC assessment that NAVER was dominant, although not superdominant. In the Korean general search market, NAVER has lost a substantial portion of its share (around 30%) to Google since the mid-2010s. Also, unlike the KFTC’s market definition, which treated comparison shopping as an independent market, it is now more widely recognized that NAVER competes with Coupang, often dubbed the “Amazon of Korea.”
Additionally, the markets in which the alleged abuses occurred differed. In contrast to the EU case, where Google’s rivals were significantly disadvantaged by Google’s conduct in the comparison-shopping market (Case AT.39740, e.g., section 7.2.3), the Korean case concerned NAVER leveraging its comparison-shopping dominance in the e-commerce marketplace. But the actual impact of the conduct in the latter market was unclear, and the causal link between the conduct and the alleged effects in the e-commerce marketplace also remained highly contestable. (See, e.g., Kim & Song 2021, pp. 166-167, 172-173; Kang & Lim 2021, p. 187, 214-215; and Kim 2022, pp. 58-61). Moreover, it was questionable whether traffic from NAVER’s comparison-shopping service played as significant a role in Korea’s e-commerce marketplace as traffic from Google’s general search service did for comparison-shopping providers in the EU case.
Third, it is also noteworthy that, in Korea, there was always doubt as to the soundness of “self-preferencing” as a theory of harm. Prior to the Commission’s 2017 Google Shopping decision, there had been broad consensus among Korean competition-law scholars that “self-preferencing” was typically a manifestation of competition on the merits, rather than an unlawful restriction of competition, except where the self-preferencing amounted to tying or an abusive refusal to deal.
Indeed, at that time, the U.S. Federal Trade Commission’s (FTC) 2013 closure of its investigation of Google’s self-preferencing (under Section 5 of the FTC Act) was cited favorably in Korean scholarship. As I understand it, this broad consensus also reflected the local market conditions described above. While the post–Google Shopping debate appears to have taken on a more nuanced tone amid the government’s push for platform regulation (e.g., Kim 2024), it is noteworthy that the self-preferencing doctrine itself has still not been widely endorsed as an independent theory of harm in Korea.
These contextual points help us to understand the Supreme Court’s reasoning and ruling more clearly.
Key Findings
What, then, did the Supreme Court of Korea decide? The key findings can be identified as follows.
Abuse of Dominance
The first point to note is the Court’s strong, explicit, and repeated rejection of a general obligation of “equal treatment” under Korea’s competition law. In this case, the Court made clear in three separate contexts that even a dominant platform operator is not required to treat its own products or services equally with those offered by other businesses, unless other laws provide a legal basis for such treatment.
Of course, technically speaking, this was partly because three different provisions were applied in the case. Nevertheless, it seems telling that the Court, in each of its analyses, repeatedly emphasized that no obligation of equal treatment exists under Korean competition law (pp. 7-8, 13, 15).
Second, it appears this decision substantially curtails the role of “intent or object” in establishing abuse of dominance. Ever since the Supreme Court’s 2007 POSCO decision, which fully incorporated the effects-based (or more economics-oriented) approach into Korean competition law, it has been required that an enforcer must prove exclusionary effects or concerns thereof, and intent or object, in order to establish abuse, saving an exception that—once actual effects were proved—intent or object could be presumed.
In practice, however, the KFTC has increasingly faced difficulties in proving effects and has shown a tendency to rely on internal communications or similar evidence to establish subjective intent or object—particularly in digital-platform cases. A similar pattern appears in this case as well. But the court drew a clear line by stating that “the concern about competitive harm cannot be established solely on the basis of the intent or object to restrict competition.” (p. 5)
In my view, the Court’s curbing of the tendency to rely on intent or object, in contrast to the lower court’s stance that gave weight to them (Jeong 2023, p. 111), makes this ruling noteworthy.
Specifically, in this case, the KFTC selected five instances (out of dozens) of NAVER’s search-algorithm adjustments that led to outcomes favorable to itself (more precisely, to the sellers using NAVER’s Smart Store), and based its finding on anticompetitive intent or object. In the KFTC’s view, those five instances explicitly revealed that NAVER, as a latecomer to the e-commerce marketplace segment, had intentionally manipulated its comparison-shopping algorithm to support its own service, while analyzing the potential favorable or unfavorable effects of each algorithmic adjustment (see, e.g., press conference transcript, p. 22).
NAVER, in contrast, argued that these adjustments were made to enhance the diversity and accuracy of search results, and that it was unjustified to single out only those few instances—out of more than 50 adjustments in total—as having been carried out with any special intent or object (No. 2021-027, paras 325–328).
The Supreme Court took a broader view, acknowledging the business reality that analyzing the potential effects of algorithmic adjustments is part of normal business activity. It also held that enhancing the diversity of search results may be an important factor in improving the quality of the comparison-shopping service; it therefore viewed that such intent can be indicative of competition on the merits (p. 9). The court further observed that some degree of bias is inevitable in the process of improving quality, and rejected the enforcer’s approach of inferring anticompetitive intent or object from a few selectively chosen instances (p. 10).
In my view, this is a rightful rejection of the unrealistic perception that only perfectly pure and impeccably neutral conduct can be legitimate. Given the realities of business practice, and of self-interested human behavior more broadly, expecting overly moralistic standards is hardly reasonable.
The third noteworthy point is that the court explicitly ruled that, when a platform’s self-preferencing practices are accused of abusively leveraging the platform’s dominance in another market, the enforcer still must prove anti-competitive effects, and a causal link between the conduct and the alleged effects must also be examined.
According to the decision, the KFTC and the Seoul High Court’s findings of actual or potential anticompetitive effects (i.e., “effects or concerns thereof”) were made merely on the basis that, following the conduct in question, NAVER’s market share, transaction volume, and the number of sellers using its Smart Store platform increased more than those of its rivals. The apparent assumption is that there is a greater need to intervene when dominance in one market is leveraged into another (p. 7). But on this reasoning, any conduct that involves deliberate leverage of dominance could amount to an abuse, regardless of whether there were any actual effects, or even whether such effects were likely.
In reality, however, at the time of the conduct, NAVER had a major competitor (Gmarket/Auction) with nearly a 50% share in the e-commerce marketplace (Kim 2022, pp. 43-45). And the transaction volumes of other competitors did not fall; in fact, they increased, although to a lesser extent than NAVER’s own transaction volumes (Kim 2022, p. 49). Moreover, the growth of NAVER’s Smart Store could be more plausibly attributed to the introduction of its payment service, rather than to the algorithmic adjustments at issue (Kim & Song 2021, p. 166). While it is true that a finding of anti-competitive effects does not necessarily require proof of actual outcomes, disregarding such market realities gives the impression that the KFTC was determined to create a Korean self-preferencing case and then reverse-engineered the legal finding for that pre-set conclusion.
Fortunately, this flaw was corrected by the Supreme Court’s judgment. The court rejected the lower court’s assumption that conduct affecting another market—namely, the leveraging of dominance—is inherently more restrictive of competition. It held that, even in leveraging cases, the enforcer must still demonstrate a “concrete concern” of anticompetitive effects and establish a causal link—that is, “a reasonable likelihood,” of connection between the conduct and the resulting outcome (p. 8).
Accordingly, after considering the market realities described above, the court emphasized that “effective competition” had continued and found that the enforcer and the lower court should have further examined whether the subsequent expansion of NAVER’s business resulted from the alleged anticompetitive conduct, from legitimate competition on the merits, or simply from the overall growth of the industry (pp. 8-9).
Such a jurisprudentially sound ruling leads me to believe it is the Supreme Court, rather than the KFTC, that better reflects the spirit of the EU Google Shopping case (e.g., C-48/22 P, paras 186, 224).
Unfair Trading Practices
With that said, there are other areas where the Supreme Court’s judgment may have gone too far. In particular, on those counts where the KFTC invoked unfair trade practice (UTP) provisions—specifically, discriminatory treatment as lessening competition and deceptive inducement—the court’s reasoning may be overly stringent.
First, apart from the specific judgment on the legality of NAVER’s conduct, some may find the threshold to recognize a protectable consumer perception was set too high, particularly for digital cases.
Regarding whether the algorithmic adjustments at issue amounted to deceptive inducement, the Supreme Court applied the well-established standard of “ordinary consumers with normal trading experience and attention” (e.g., 2014Du15047) (pp. 11, 13). Under this standard, the court held that such consumers would have understood that the rankings were not determined solely by the characteristics of the goods themselves, and found that the algorithmic changes were not sufficient to reach the level of illegitimate deception in the sense of Korean competition law, capable of impairing consumers’ rational choice (pp. 13-14).
One might wonder, however, whether a somewhat more relaxed standard might have been permissible under the interpretation of the UTP provision. Much like Section 5 of the U.S. FTC Act, UTPs partly serve a consumer-protection function within the Korean competition-law enforcement framework. If consumers may be more easily misled in digital environments than in offline settings, a somewhat more nuanced and consumer-friendly threshold may be appropriate, broadening the scope of what counts as deception (see here for an opposing view). This threshold may, in turn, have implications for the ongoing Coupang case, which concerns allegedly deceptive forms of self-preferencing (see my Kluwer Competition Law Blog post or SSRN).
Second, a similar concern arises with the Court’s finding on discriminatory treatment as “lessening competition.” In its assessment of the UTP discrimination, the court repeated its reasoning under the abuse-of-dominance analysis, concluding that neither restricting nor lessening competition had been sufficiently proven (also finding that the requirement of “significance” in discrimination was not established) (pp. 15-16). Considering the legislative intent and utility of the UTP provisions, particularly in cases of incipient violations, one might question whether the court demanded an overly high evidentiary standard in this example, as well.
Of course, the court’s decision to dismiss the UTP claims is not entirely without justification. And the primary issue in the case lies in the abuse-of-dominance violation rather than the UTPs. It is likely that, even if the court had applied a more relaxed standard to the UTP provisions, the outcome would have differed only slightly, apart from a somewhat mitigated degree of sanction (as UTP violations entail a lighter level of sanction).
Moreover, the KFTC has shown a recent tendency to invoke UTP provisions almost habitually as a kind of fail-safe to increase the likelihood of success in abuse-of-dominance cases. In that context, it may be understandable that the Supreme Court wanted to place a check on this scattergun or shotgun-like enforcement practice.
Reflections and Broader Implications
Around the turn of the 2020s, the specter of platform regulation was haunting Korea, as in many other jurisdictions. While the KFTC’s sanction against NAVER was an enforcement action, rather than regulation in a formal sense, it was nonetheless part of a broader wave of regulatory tightening on digital platform companies.
Fundamentally, similar to the platform-regulation initiatives pursued around the same time by the same authority, this enforcement action can be seen driven not necessarily by market or institutional efficiencies or by local needs, but by policymakers’ pursuit of the EU model under conditions of policy uncertainty—one that was perceived as a socially legitimate solution that could be pursued with little cost (see, e.g., DiMaggio and Powell 1983, p.151).
Arguably, such top-down isomorphic actions—which may be rational at the individual level yet collectively irrational—can produce higher inefficiencies and stronger social or political resistance as the contextual misfits widen. In that sense, this judgment represents—much like the social backlash against the government’s platform regulation initiatives—an important pushback against the government’s pursuit of the EU’s regulatory approach to the digital economy, despite the obvious contextual differences.
As discussed, the Korean NAVER Shopping case differed from Google Shopping in its market realities and the academic landscape. NAVER’s market position, the effects of its conduct, the causal link between the two, and the academic support for treating self-preferencing as an independent theory of harm were all far less robust. The KFTC’s decision to nonetheless press ahead with the case appears, if anything, to have been driven by a prior determination to create a Korean self-preferencing case—and by an effort to fit the ill-suited facts to that predetermined conclusion.
It is fortunate that the Supreme Court’s decision ultimately corrected, albeit belatedly, the misguided isomorphic push for self-preferencing (for more on Korea’s isomorphic moves in digital-competition policy, see my 2025 ASCOLA Asia Conference presentation).
Of course, drawing on established competition governance models elsewhere is by no means inherently undesirable. Generally speaking, it may enable faster policy design by reducing learning costs and minimizing trial and error. It also offers advantages in promoting the international interoperability of institutions—both regulatory measures and enforcement practices—by aligning domestic systems with global norms, thereby reducing transaction costs, and facilitating cross-border coordination.
As noted earlier, however, a closer look reveals that it was actually the Supreme Court of Korea, rather than the KFTC, that effectively and successfully internalized the virtues of the European Court of Justice’s Google Shopping judgment in Korea—embracing the conclusion that self-preferencing is not inherently anticompetitive, and reaffirming that it is the enforcer’s obligation to prove a causal link and all constituent elements of an infringement to the requisite legal standard.
I believe the Korean experience offers broader insights for non-EU jurisdictions seeking to navigate the tension between policy convergence and contextual adaptation in designing their own digital-competition governance frameworks.
Lastly, from a domestic perspective, while there remain some regretful aspects in the court’s interpretation of the UTP provisions, these seem rather technical, as discussed above. I am inclined to think that the outcome might have been different, in favor of the KFTC, had the enforcer instead relied solely on the UTP provisions.
In that regard, I look forward to the Supreme Court’s forthcoming judgment in the Coupang case, where deceptive self-preferencing is being challenged purely under the UTP framework (see my Kluwer Competition Law Blog post or SSRN).
———————-
(I am grateful to Professor Jinha Yoon (Korea University) for generously sharing her insights in response to my questions about this case. Before joining academia, she served as the judicial researcher in charge of the case at the Supreme Court of Korea and was substantially involved in drafting the NAVER Shopping judgment. I am also deeply thankful to Professor Lazar Radic (IE University) for his constructive feedback on an earlier draft of this post. All remaining errors are, of course, my own.)
