Posted by Ray Rui Gao (University of Southern California) and Yifei Lu (University of Illinois Urbana-Champaign), on

Monday, November 3, 2025 *

Ray Rui Gao is an Assistant Professor of Accounting at the Marshall School of Business, University of Southern California, and Yifei Lu is an Assistant Professor of Accountancy at the Gies College of Business, University of Illinois Urbana-Champaign. This post is based on their article forthcoming in the Journal of Accounting Research.

Every year, public companies list a set of “peer” firms in their proxy statements to benchmark top executive pay. These compensation peer group disclosures might seem routine, but they contain rich information about who competes for top executive talent. In our forthcoming study (Journal of Accounting Research, in press), we tap into this data in a novel way by using network analysis of peer group disclosures to map the managerial labor market. Our study generates novel managerial labor market classifications and competition measures, offering novel insights beyond traditional product industry-based measures. In a sense, it is the managerial labor market analog to the network-based product market classifications by Hoberg and Phillips (2010, 2016).

From Product Industry Peers to Compensation Peer Network

Traditionally, researchers assumed that a company’s competitors for executive talent were mostly those in the same product industry or of similar size. However, this can be misleading. Many firms recruit executives from outside their product industry: think of a retail company hiring a tech executive to develop the e-commerce business (e.g., Nike hired John Donahoe from ServiceNow as CEO in 2019). We argue that we can better capture these complex relationships by looking at the network of disclosed peers across all firms. In practice, most large companies disclose a peer group of around 15–20 firms for pay benchmarking (over 97% of S&P 500 firms do so). Because boards consider multiple factors, such as industry, size, and specific talent needs, when selecting these peers, each disclosure embeds rich information about a firm’s talent competitors. For example, Delta Air Lines in its proxy statements in 2019 mentioned that “our peer group is composed of three major U.S. airlines and eighteen other companies in the hotel/leisure, transportation/distribution/machinery/aerospace/ defense, and retail industries,” and that “in order to retain and attract the talent we need, Delta must compete with these types of companies”. Still, any one company’s list is limited in scope. By aggregating all peer relationships across firms, we let the market itself reveal a broader set of potential talent competitors, including those a firm might not have identified on its own. We then construct a network “map” of companies connected by peer relationships. If Company A lists Company B as a peer, we draw a link between A and B. Collectively, these links form a web that shows which firms see each other as comparable in the market for managerial talent.

This network-based approach reflects competition in multiple dimensions, not just industry, but also size, geography, specialized skill sets, and so on. It’s a dynamic view that can evolve as companies update their peer groups annually. Each firm’s position in the network indicates its labor market neighborhood: companies closely linked in the network likely compete for the same pool of executives. This is a big shift from static product industry classifications. For example, an innovative media company might appear in the peer networks of tech firms if both are hunting for executives with digital content expertise, even if traditional product industry classifications wouldn’t pair them. By analyzing the overall peer network, our study essentially lets “the crowd” of boards reveal how they perceive the landscape of executive talent competition.

A Better Predictor of Executive Moves

One of the main findings from our research is that our network-derived classification of the managerial labor market better predicts executive movements between companies than conventional measures. In tests looking at executive career moves, the companies that end up hiring each other’s executives tend to be those closely connected in the peer network. In fact, we show that this network classification outperforms a comprehensive list of traditional predictors (e.g., within the same product industry, size similarity, geographic proximity) in explaining where top executives move when they change jobs. This makes intuitive sense: firms that are closely linked in the peer network compete for similar managerial talent. By capturing these less obvious connections, our peer network approach provides a more complete view of the executive talent marketplace. The results from the executive move prediction model using our network-based labor classifications also have practical implications for boards: by examining their network neighbors rather than relying solely on self-selected peers, boards can better identify potential executive candidates and where their own talent might go.

We also demonstrate that the advantage of the network approach lies in its ability to capture the multidimensional and dynamic nature of competition. Traditional categories like product industry codes are one-dimensional and static; they don’t update as business models evolve or new competitors emerge. In contrast, companies adjust their peer groups as their demand for managerial talent changes, effectively updating the network. Our methodology leverages this dynamic feature to reflect real-time shifts in the labor market. For example, if firms in the renewable energy space start looking for executives with expertise in AI or cybersecurity, these executives will soon show up in changed peer lists and hence in new network connections. Our findings demonstrate that the multidimensional and dynamic properties of the network-based groupings contribute to their ability to predict executive turnover.

Managerial Labor Market Competition Influences Pay Design and Performance

Beyond mapping who competes with whom, the peer network also allows us to develop network-based competition measures to measure how intense the competition is for each firm’s executives. For instance, how many other companies name your firm as a peer (an indicator that many are eyeing your talent), and how tightly clustered your peers are (indicating a group of firms all seeking very similar executives). Validating these measures, we find that firms facing more fierce managerial competition use more retention strategies through pay design, including offering higher compensation, allocating a greater share of compensation to equity, and using longer vesting periods. These tools make it harder for a rival to poach their executives by ensuring the manager has substantial financial incentives to stay over the long term. This evidence suggests that boards are responsive to the external market for talent: when outside opportunities abound, they design compensation to retain talent.

Our study also sheds light on some controversial pay practices using the competition measures we developed. Executive pay at certain firms often draws criticism. For example, rewarding CEOs for industry-wide or market-driven gains (“pay for luck”) or failing to adjust compensation based on peer performance (a lack of “relative performance evaluation”). We argue that in many cases, intense labor market competition can help explain these practices. When a company anticipates that other firms may make attractive offers to its CEO or key executives in a strong market, it may preemptively match those potential offers. What might appear as overpaying or leniency in isolation could actually be a defensive move to retain valued talent in a highly competitive environment. This perspective doesn’t excuse all instances of high pay but provides important context: boards may be weighing the cost of higher compensation against the even greater cost of losing leadership to a rival.

Interestingly, we also find that a competitive executive labor market can motivate better performance, essentially creating a tournament-like incentive. When executives know that the potential gain from moving up to a better-paying firm is large, they have stronger reasons to excel, as outstanding performance can attract attention from other firms or position them for a step up the ladder. Consistent with this idea, we find that CEOs facing larger “tournament prizes”, the pay gap between their current job and the next best opportunity identified using our labor classifications, deliver stronger firm performance. In a separate analysis, we also show that the portion of pay driven by labor market competition predicts superior future performance, suggesting that firms operating in more competitive talent markets tend to retain higher-productivity managers. In other words, an active market for managerial talent doesn’t just drive up pay; it can also push managers to perform better, ultimately benefiting shareholders. This perspective adds nuance to the governance debate: while some worry that a fluid CEO job market fuels escalating pay, it might also spur those executives to achieve stronger results to win the next promotion.

Implications for Boards, Researchers, and Regulators

• Identifying a Broader Pool of Managerial Talent Competitors

Our network approach aggregates the peer choices disclosed by thousands of firms, revealing how individual decisions collectively map the executive labor market. This broader view identifies not only a company’s direct competitors for talent, but also potential competitors that emerge through indirect links, firms connected through shared peers. Understanding this expanded network helps boards anticipate where poaching threats or recruitment opportunities may arise and helps researchers characterize the true boundaries of managerial labor markets.

• Linking Pay Design to Executive Labor Market Competition

Our network measures quantify how intensely firms compete for executive talent. Firms located in dense parts of the network face more rivals and greater retention needs. We find that these firms rely more on equity-based pay and use longer vesting periods to retain executives. Knowing where a firm sits in this network helps boards design and calibrate compensation packages and helps researchers better evaluate pay practices.

• New Empirical Toolkit for Managerial Labor Market Competition

The SEC’s 2006 enhanced compensation disclosure rules require firms to list compensation benchmarking peers and describe peer selection criteria in proxy statements. We turn these disclosures into a new empirical toolkit: (i) labor market classifications that identify a firm’s competitors for executive talent and (ii) competition measures that quantify how intense that competition is. These measures provide a foundation for future research on managerial labor markets and executive pay.

The full paper is available here for those interested in the detailed analysis and evidence.