This paper investigates the conditions of platform workers, with a focus on the monopsony power wielded by platforms. Broadly, labor-supplying platforms can be divided into two categories: web-based and location-based (Jang et al., 2020). For illustrative purposes, this paper examines ICT software developers as a primary example of web-based freelance platforms and delivery app workers for location-based labor platforms. While no singular case can encapsulate the entirety, each instance provides valuable insights for its category and contributes to a comprehensive understanding.

Here, distinct methodologies are employed to assess the monopsony power in each context. For the ICT software developer platform, the elasticity of labor supply was directly estimated using data from actual job postings and applications to ascertain the extent of labor monopsony. Conversely, for delivery apps, surveys were conducted among delivery app workers. Their feedback, combined with a consumer-focused analysis that accounts for dynamics, offers insights into the multi-homing status and the ease of switching between platforms from both worker and consumer viewpoints. Based on these findings, the paper draws conclusions about the monopsony power inherent to each platform.

Labor monopsony power is inversely related to labor supply elasticity. This elasticity can be empirically quantified by estimating the degree to which labor suppliers adjust to a decrease in labor prices. If labor supply cannot be reduced despite a decrease in its price, this signals strong monopsony power on the side of labor demand. However, the discourse on labor monopsony within the platform economy possesses a distinct uniqueness, setting it apart from broader dialogues. Specifically, when considering competition among platforms, it is vital to account for both (or multifaceted) sides of the consumer-supplier relationship beyond merely assessing the platform’s monopoly status or market concentration.

Ⅲ.  Status of Platform Workers: (1) Web-based Platforms - ICT Software Developers

This section examines web-based freelance labor markets, with a specific emphasis on the ICT software developer labor market. This particular market epitomizes a shift toward freelancer-centric operations. As companies grapple with the complexities of managing software development staff and swift technological evolution, they increasingly turn to specialized external firms or individual experts (Lee et al., 2019). In the system integration (SI) domain, software customization tailored to individual companies results in a marked reliance on freelance professionals.

This section estimates the monopsony power of platforms, particularly focusing on in-house contractors potentially more at risk. Data was collected from one of the leading IT platforms through daily web scraping of online project postings from April to September 2022. This dataset is used to measure the correlation between changes in per capita monthly income and the duration of job vacancies. Fundamentally, the analysis investigated how vacancy duration lengthens (or shortens) as the offered income rises (or falls). This relationship between job vacancy duration and income fluctuations represents labor supply elasticity.

The results in Table 1 reveal that an increase in per capita monthly income corresponds to shorter vacancy durations. Without controlling the specificities of contracts or the characteristics of individuals or firms, a 1% rise in per capita monthly income led to a 0.2% reduction in job vacancy duration (measured in days). This observation could be attributed to the blending of different contract types in the dataset. After accounting for various variables, including job sector, project scale, geographical location, and capability for remote work, the estimates become markedly more pronounced. They approach a value of 1, especially when controlling for firm-fixed effects.

Given the average labor supply elasticity is estimated to be near 1 in the preceding study (e.g., Webber, 2015), the labor monopsony of domestic ICT freelance platforms is on par with that of an average user in a conventional labor market. This value is also five to ten times greater than the existing labor supply elasticity estimates for online platforms (e.g., Dube et al., 2020). Such findings indicate that the monopsony power wielded by these platforms in the local ICT freelance market may not be as strong as predicted.

These findings indicate that, even for the potentially more vulnerable in-house contractors in the ICT software developer market, there seems to be little compelling need for active governmental intervention beyond standard contractual safeguards. Historically, ICT software developer platforms have predominantly functioned as information providers and transaction facilitators, with a number of platforms engaging in meaningful competition.

Ⅳ. Status of Platform Workers: (2) Location-based Platforms - Delivery App Workers

Next, this study turns its attention to the delivery app market as a prime example of location-based platforms. According to the Online Shopping Survey by Statistics Korea (2023), online food product transactions saw a marked increase, rising from 9.7 trillion won in 2019 to 26.6 trillion won in 2022.

The discussion surrounding the employment status of delivery app workers has been longstanding. From the perspective of this paper, this highlights potential concerns about substantial monopsony power wielded by delivery app platforms. Unlike the ICT software developer platforms discussed earlier, delivery apps are not simply viewed as intermediary platforms. These platforms exert significant influence over riders' compensation and working conditions, suggesting their dominant position vis-à-vis the riders.

The delivery app market, characterized by the dominance of three primary apps, might experience constraints in individual platform market power due to competition among them. In simpler terms, an individual delivery app platform could struggle to enforce pay and working conditions distinct from its competitors. In the platform economy, especially in the absence of a clear dominant platform, the ability of consumers and delivery workers—on both ends of the platform economy—to easily switch between or use multiple delivery app platforms concurrently becomes especially pertinent. Therefore, it is crucial to investigate the reasons behind platform choices from the viewpoints of both consumers and delivery app workers.

This paper leverages data on the app preferences of Korean consumers to examine the standing of delivery apps and the intensity of competition in the consumer market. Panel data from a survey documenting the behavioral tendencies of Korean delivery app users sheds light on the evolving competitive dynamics over time.

The data reveals a significant surge in both the overall user count and the proportion of multi-homing users, especially in the latter half of 2020 (Figure 1). Accounting for potential duplicate counts of multihoming users, the adjusted user count has shown a steady increase since 2019, with a sharp uptick in the latter half of 2020. This rise coincides with the rapid emergence of Coupang Eats, where the portion of consumers using multiple apps soared from roughly 40% to over 60% during this timeframe. Initially, this trend was prominent in the capital region, but similar patterns emerged slightly later in noncapital regions.

In addition, the market concentration index (Herfindahl-Hirschman Index: HHI) for (major) delivery apps was already notably high before 2020, surpassing 4,000. However, there was a sharp decline in the latter half of 2020 (Figure 2). While it remains at a level that’s hard to deem low, it’s nearing the 2,500 benchmark, which is typically viewed as high. Similarly, this trend initially emerged in the capital area and subsequently manifested in non-capital regions after a brief delay.

Another perspective to consider involves delivery app workers. The KDI survey carried out in September 2022 incorporated detailed queries regarding the use of multiple platforms and the costs tied to transitioning between them, thus providing a richer dataset than earlier surveys.

Usage patterns among platform workers can be generally classified into shifters and non-shifters. Shifters are those who interact with multiple platforms and move between them. The KDI survey highlighted that approximately 47% of delivery (agency) app riders used two or more delivery apps simultaneously (multi-homing). On the other hand, 14% relied on a single delivery app but might switch to another over time (single-homing/shifting). The remaining 40% consistently used only one delivery app (Table 2).

Responses on the reasons behind platform usage patterns suggest that, while there is fierce competition among delivery app platforms to attract more consumers, the rivalry for securing riders might be less intense (Lee and Han, 2022). The survey data reveals that riders predominantly engage with multiple platforms “due to insufficient income or to enhance income,” suggesting that their engagement across several platforms might not be entirely by choice. Many riders committed to a single platform cited “the challenges of juggling multiple apps” and the perceived “benefits of remaining loyal to one platform.” Such observations hint at potential intrinsic or externally imposed costs in transitioning between platforms. As a result, a surge in consumer demand may not necessarily translate to better conditions for those working on these platforms.

Upon closer examination of the working conditions of delivery app workers, many of these concerns become more apparent. About 71% of these workers expressed that their jobs resemble traditional paid work despite the limited control the apps have over tasks, working hours, and idle intervals (Table 3). While these sentiments are subjective and may not accurately reflect reality, a notable portion of respondents mentioned that their pay is determined by the apps and that there is a system in place to evaluate their performance through mechanisms like ratings and reviews (Table 3). Detailed feedback suggests a high likelihood of work limitations or even account suspension based on such performance evaluations (Lee and Han, 2022).

In sum, these findings suggest that while there are reasons for those concerns about labor monopsony in relation to delivery apps, there is also a growing chance that their influence may be diminishing over time. From a consumer perspective, competition among platforms is intensifying, with newer entrants diluting market concentration and more users opting for multi-homing. However, for delivery app workers or riders, this heightened competition does not necessarily lead to better working conditions because elements like involuntary participation due to insufficient earnings and high switching costs complicate shifts between platforms. Moreover, even with the flexibility to determine their own working schedule, there is still a considerable level of oversight exerted by platforms through post-task evaluations and task allocation algorithms.

Ⅴ. Conclusion and Policy Recommendations: Designing Social Protection Based on Platforms’ Monopsony Power

Discussions regarding the protection of domestic platform workers have primarily revolved around the issue of recognizing their status as employees. While such distinctions are undeniably vital, the rigid dichotomy of employee versus business owner impedes a deeper understanding of the intricate nature of their real-world roles. Consequently, this often results in inadequate measures to address the protection needs of a vast majority of platform workers.

This paper began by analyzing the status of platform workers through the perspective of platforms’ monopsony power over labor demand. It conducted an empirical analysis to estimate monopsony power in the prime examples of both web-based freelance and location-based labor supply platforms. For the ICT software developer platforms (web-based freelance platforms), the estimated monopsony power aligns with general labor market levels and does not raise immediate concerns. On the other hand, delivery apps (location-based platforms) presented riders with hurdles such as algorithm-driven oversight and barriers to platform switching. These findings suggest potential concerns about pronounced monopsony power. Furthermore, the recent rise in competition among platforms and increased consumer multi-homing call for careful monitoring of these evolving dynamics.

As platform sectors exhibit varied competitive conditions and the protective requirements of platform workers differ, adopting universal criteria might risk either under-regulation or over-regulation. This complexity arises because platforms’ monopsony power is not uniform across sectors. Furthermore, its magnitude can vary, reflecting shifts in the competitive landscape.

To safeguard platform workers effectively while still fostering platformdriven innovation, policies must be flexible yet unified, with the primary aim of curbing platforms’ monopsony over labor demand and averting potential abuses. Moreover, given that intensifying competition among platforms might not directly alleviate issues tied to their labor monopsony power, there’s a pressing need to bridge the divide between antitrust policies, which aim to enhance competition, and labor policies tailored to protect workers with limited bargaining capabilities.

In situations where there is ambiguity in classifying platform workers as either paid or self-employed, the prudent course of action starts from categorizing them as self-employed given the current regulatory landscape. However, their social protection should be proportionally determined by assessing the platform’s monopsony power. When a significant degree of this monopsony power is either directly measured or reasonably expected, the extent of regulatory intervention can be proportionately heightened. When the monopsony power decreases, these interventions may need to be adjusted accordingly.

For instance, when a platform’s monopsony power surpasses a designated threshold, it might be interpreted as the platform having a “superior bargaining position.” Platforms with this distinction could then be subject to regulations aimed at preventing the misuse of such power. Although this criterion could be relevant for a broader group of workers, it holds special significance for platform workers, given that a platform’s monopsony power can be readily quantified through data analysis.

Furthermore, regulations on algorithmic transparency can be incrementally fortified based on the extent of monopsony power. Previous discussions have shed light on self-preferencing issues. Specifically, concerns have been raised by mobile app developers regarding the ranking systems in app stores like Google and Apple. Debates have also emerged around perceived self-favoritism in certain domestic mobility apps. As labor monopsony power becomes more pronounced, the push for greater algorithmic transparency intensifies in parallel. It is posited that platforms should be obligated to establish transparent criteria, incorporate post-implementation human reviews, and offer mechanisms for users to challenge automated task allocations. Such interventions should be viewed as crucial “guardrails” in the rapidly evolving landscape of AI.

Competition among platforms can lead to improved working conditions, including better compensation, as they vie for platform workers, emphasizing the important role of traditional competition policies. Yet, even when multiple platforms are in competition, restricted worker mobility might allow a single platform to maintain monopsony control. This underscores the importance of actively reducing barriers for workers transitioning between platforms or into wage-based roles. In situations where a platform demonstrates significant monopsony power over labor demand, considering measures such as limiting multi-homing restrictions, curbing preferential treatment clauses, or enhancing data portability becomes essential. Broadly speaking, including the standpoint of labor policy, there’s a potential need to discourage non-compete clauses, commonly included by default, or even to review their removal. Additionally, the standardization of work-related data for each platform worker, facilitating its transferability or validation as evidence of work experience, is also worth consideration.

Also, there is a compelling need to establish basic protections for all labor providers since promoting competition among platforms or reinforcing fair trade policies among businesses might be insufficient to protect individual workers. Such protections encompass provisions like safety guidelines, health regulations, and workers’ compensation. These fundamental safeguards serve not only as a basic human rights guarantee, but also as a countermeasure to platforms’ monopsony power. Limiting these protections solely to “platform workers” could result in unintended outcomes. Consequently, extending these safeguards universally to all labor providers is recommended. However, setting this protection bar too high might inadvertently hamper the growth of the platform economy or lead to unintended advantages in regulatory grey areas. In this regard, it is crucial to determine these standards carefully with a comprehensive review of the current economic situation.