Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.08405 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914086551289856 |
|---|---|
| author | Xu, Yang |
| author_facet | Xu, Yang |
| contents | This paper studies a linear production model in team networks with missing links. In the model, heterogeneous workers, represented as nodes, produce jointly and repeatedly within teams, represented as links. Links are omitted when their associated outcome variables fall below a threshold, resulting in partial observability of the network. To address this, I propose a Generalized Method of Moments estimator under normally distributed errors and develop a distribution-free test for detecting link truncation. Applied to academic publication data, the estimator reveals and corrects a substantial downward bias in the estimated scaling factor that aggregates individual fixed effects into team-specific fixed effects. This finding suggests that the collaboration premium may be systematically underestimated when missing links are not properly accounted for. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_08405 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Team Networks with Partially Observed Links Xu, Yang Econometrics This paper studies a linear production model in team networks with missing links. In the model, heterogeneous workers, represented as nodes, produce jointly and repeatedly within teams, represented as links. Links are omitted when their associated outcome variables fall below a threshold, resulting in partial observability of the network. To address this, I propose a Generalized Method of Moments estimator under normally distributed errors and develop a distribution-free test for detecting link truncation. Applied to academic publication data, the estimator reveals and corrects a substantial downward bias in the estimated scaling factor that aggregates individual fixed effects into team-specific fixed effects. This finding suggests that the collaboration premium may be systematically underestimated when missing links are not properly accounted for. |
| title | Team Networks with Partially Observed Links |
| topic | Econometrics |
| url | https://arxiv.org/abs/2505.08405 |