📈 economics

Career networks, bureaucratic authority, and political selection

This paper analyzes career histories of Chinese officials from 1979 to 2020 to demonstrate that exposure to high-authority counterparts through co-service networks serves as a key observable indicator of favorable organizational screening and selection into administrative elite positions.

Original authors: Doojin Ryu, Han Sun

Published 2026-08-13

📖 7 min read🧠 Deep dive

CC BY 4.0

Original authors: Doojin Ryu, Han Sun

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a giant, invisible ladder stretching up into the clouds. This isn't a ladder of wood or metal, but a ladder of careers, where people climb from one rung to the next to reach the very top. In the world of political science, researchers have long been obsessed with figuring out who gets to climb this ladder and why. For decades, the big question was: Is it about who you know? Is it about being from the same hometown as the boss? Or is it about having a secret "sponsor" who whispers your name to the right people?

Think of it like a high-stakes video game. In many games, your character's success depends on their "network"—the other players they've teamed up with. But in real-world politics, especially in China, the rules are tricky. We can't see the secret handshakes or the whispered promises. We can only see the public record: the list of jobs a person has held and the organizations they worked in. The big mystery has been whether the size of your network matters, or if it's actually about who is in your network. Does it matter if you've worked with a regular clerk, or if you've shared an office with someone who holds the keys to the kingdom? This paper dives into that exact puzzle, using a massive amount of historical data to see if the "who" matters more than the "how many."

The Great Career Map: Who You Worked With Matters More Than How Many

Imagine you are a detective trying to solve the mystery of how people get promoted to the very top of the Chinese government. You have a huge stack of old notebooks (public career records) spanning from 1979 to 2020. These notebooks list every job a mid-level official held. The researchers, Doojin Ryu and Han Sun, decided to use these notebooks to draw a massive map of connections.

They didn't try to guess who was friends with whom or who owed a favor to whom. Instead, they looked at something much more concrete: co-service. If two people worked in the same organization at the same time, they were connected on the map. But here is the twist: not all connections are created equal.

Think of your career network like a group of friends you've made at school. If you've hung out with a thousand kids who are all in the same grade as you, that's a big network. But if you've hung out with just a few kids who are already the student council presidents or the principals, that's a different kind of network. The researchers wanted to know: Does having a big network of regular friends help you become the principal? Or does having a few connections to the already powerful people help you more?

The Big Discovery: It's About the "High-Reach"

The paper found a clear answer. When they looked at the data, they discovered that the most important factor wasn't the number of people you worked with, but the power of the people you had worked with in the past.

They created a special score called "High-Authority Reach." This score counts how many times an official had worked alongside someone who, by the previous year, had already reached the very top tier of government (the provincial-ministerial level).

Here is the magic number: For every standard jump up in this "High-Authority Reach" score, an official's chance of getting promoted to the elite tier went up by 1.2 percentage points. While that might sound small, remember that the average chance of getting promoted in any given year was only 4.5%. So, this single factor boosted their chances by about 27%. That is a huge deal in the world of career climbing!

What It's NOT About

The researchers were very careful to rule out some popular ideas.

It's not just about having a big crowd: When they looked at the sheer number of people an official had worked with (network volume), it helped a little bit on its own. But once they added the "High-Authority Reach" into the mix, the importance of just having many connections dropped significantly.
It's not just about hometown ties: Some people thought being from the same hometown as the leaders was the secret sauce. The data showed that while hometown ties had a tiny effect, they weren't the main driver.
It's not about secret sponsors: The researchers tried to guess who might have been a "sponsor" (a powerful mentor) by looking at who an official worked near. But these guesses didn't show a clear positive link to promotion. In fact, some of these "sponsor" guesses even showed a negative link, suggesting that maybe being too close to a stalled career track can actually hurt you.

The "Why" Behind the Magic

So, why does working with powerful people help? The paper suggests it's not necessarily because those powerful people are pulling strings behind the scenes (though they might be). Instead, it's like a quality stamp.

Imagine a school where the best students are always grouped together for special projects. If you see a student who has worked on projects with the top students, you assume that student must be pretty good too. The researchers argue that in the Chinese government, the people who get assigned to work with high-ranking officials are likely those who have already passed a strict "screening" process. They were chosen for a reason. So, having a history of working with the elite is a public signal that says, "This person has already been vetted and found worthy by the system."

The Long Game

One of the coolest findings is that this isn't just about who you are working with right now. The researchers looked at connections made four to six years before a promotion. Even those old connections still mattered! This suggests that your career path is like a long trail of breadcrumbs. If you walked on a path that led to powerful people years ago, that path is still glowing when you try to climb the final ladder.

They also found that the age of the connection mattered. Connections made a long time ago (more than eight years back) were actually just as strong, if not stronger, than recent ones. This reinforces the idea that it's about the long-term "track record" of who you've been associated with, not just the friends you made last week.

The Bottom Line

In simple terms, this paper tells us that in the Chinese bureaucracy, your career history is a public resume that says more than just your job titles. It reveals the power level of the people you've stood next to.

If you want to know who will become a top leader, don't just count how many people they know. Look at the rank of the people they used to work with. If an official has a history of working alongside the "big shots," they are much more likely to become a "big shot" themselves. It's not about who you know today; it's about who you've been seen with yesterday, and the day before that. The system seems to reward those who have been part of the "high-authority" circle, treating that history as a badge of honor and a sign of trust.

Technical Summary: Career Networks, Bureaucratic Authority, and Political Selection

Problem and Motivation
Existing literature on political selection within the Chinese party-state widely acknowledges that personal ties shape career advancement. However, the operationalization of "connections" remains ambiguous. Prior studies rely on inferred constructs—such as factional alignment, patron-client relationships, shared native places, or co-service networks—that often require expert judgment to assign labels or assume specific relational dynamics (e.g., loyalty or sponsorship) not directly observable in public records. These measures frequently yield conflicting results; for instance, hometown ties to incumbents have been associated with both promotion premiums and penalties depending on the context. Furthermore, many relational measures are difficult to reproduce or compare across studies.

This paper addresses the measurement problem by restricting attention strictly to observable features contained within public career records. The authors ask which record-based features of career networks are systematically associated with selection into administrative elite positions when loyalty, factional support, and sponsorship are not directly observable.

Data and Methodology
The study utilizes position-level career histories for Chinese officials at the bureau level (the rank directly below the provincial-ministerial threshold) from 1979 to 2020. The main sample comprises 3,408 officials across 38,472 official-years, of whom 1,714 make their first entry into the provincial-ministerial elite.

The authors construct record-based measures of career networks based on documented co-service: two officials are connected if their careers overlap in the same organization at the same time. The core innovation lies in how these connections are weighted:

Authority Scoring: Each office is assigned an authority score based on observable attributes: formal administrative rank, party-leadership responsibility, personnel control, central placement, and resource-allocation responsibility.
Network Measures:
- Embeddedness (Embed): The volume of distinct co-service counterparts.
- Authority-Weighted Exposure (AuthExp): Connections weighted by the authority score of the counterpart's office.
- High-Authority Reach (HighAuth): A threshold measure counting documented co-service records where the counterpart holds a provincial-ministerial or higher office as of the preceding year ( $t-1$ ).
- AuthRes: The residual of authority exposure after regressing on raw network volume, isolating the authority component independent of network size.

The dependent variable is a binary indicator for an official's first entry into the provincial-ministerial tier in the following year ( $t+1$ ). The analysis employs linear probability models (with logit robustness checks) including year, province, sector, and official fixed effects, clustering standard errors at the official level.

Key Results

Dominance of High-Authority Reach: In joint specifications, HighAuth exhibits the largest positive standardized coefficient among all network measures. A one-standard-deviation increase in HighAuth is associated with a 0.012 (1.2 percentage point) increase in the probability of elite entry, representing approximately 27% of the sample's 4.5% entry rate.
Incremental Information: While network volume (Embed) is positively associated with entry in univariate models, its coefficient diminishes and becomes statistically insignificant in joint specifications when HighAuth is included. Conversely, HighAuth remains robustly positive.
Robustness to Alternatives: The positive association of HighAuth persists after controlling for static native-province background (NativeTie), inferred sponsor proxies (Sponsor, SponsorB), and network position metrics (bridge position, eigenvector centrality). Notably, inferred sponsor proxies do not show positive coefficients in joint specifications.
Temporal and Structural Stability:
- The association remains positive in official fixed-effects models, suggesting it is not driven solely by time-invariant individual characteristics.
- The effect holds when HighAuth is measured 2, 3, and 4 years prior to the outcome, indicating the signal is not merely a reflection of immediate pre-promotion placement.
- Decomposition by tie age reveals that long-standing ties (formed >8 years prior) yield the largest point estimates, suggesting the association reflects long-run career sorting rather than recent networking.
- When decomposing authority into components, formal-rank exposure retains a positive association in joint specifications, while functional components (party leadership, personnel authority, etc.) do not independently retain significance.

Contributions
The paper makes three primary contributions to the political economy of bureaucracy:

Measurement: It offers a measure of political embeddedness constructed exclusively from public career records using a transparent coding framework, avoiding the need to infer unobservable loyalty or factional affiliation.
Empirical Hierarchy of Predictors: It demonstrates that HighAuth (threshold reach to high-authority offices) carries a larger standardized coefficient than raw network size or graded authority exposure in joint specifications, distinguishing it from general network volume.
Theoretical Distinction: It distinguishes between graded authority-weighted exposure and threshold exposure to high-authority offices, showing that the latter serves as a distinct marker of career networks that have previously received favorable organizational screening.

Significance and Claims
The authors claim that public career trajectories contain observable information consistent with organizational screening within a nonmarket hierarchy. The findings suggest that HighAuth acts as a marker for career tracks that have already been vetted by the organization, rather than simply reflecting personal patronage or factional alignment. By showing that record-based authority reach predicts elite selection even when controlling for network volume and other relational proxies, the study argues that the structure of an official's documented career path—specifically their reach to high-authority counterparts—provides incremental information about their likelihood of advancement. This challenges the reliance on inferred relational constructs and highlights the informational value of observable bureaucratic hierarchies in personnel selection.