When a journal loses its place in major citation databases, why do fraud networks move their papers elsewhere?

This explores why organized research-fraud operations (paper mills, brokers and cooperating editors) move their output to other journals after a journal is dropped from citation indexes, and what that says about how fraud is organized. The short answer from the corpus is that the move is rational for a business. Paper mills are not lone cheaters. They are networks that sell authorship and publications, and the product only has value if it lands in an indexed venue. Richardson and colleagues document mills sharing image banks (2,213 articles with duplicated images), groups of editors in several countries passing submissions to each other with retraction rates above 50%, and a steady shift of output to new journals whenever a journal loses indexing Does scientific fraud operate through organized networks or individual actors?. Deindexing removes the value of that journal, so the network sends its pipeline somewhere else. The brokers, image banks and editor contacts all survive the move.

This matters because enforcement that targets one journal at a time is aimed at the wrong level. The fraud lives in the network, so punishing one venue shifts the problem instead of removing it. A security paper outside publishing shows the same pattern: attackers evaded six malicious-skill scanners, with 96% success, because each scanner judged skills one at a time. Feedback from the scanners let attackers make each piece look less suspicious while the harmful chain stayed intact Can attackers evade skill scanners by refining individual skills?. Deindexing can work like that scanner feedback. It tells a mill which venue has been flagged, but it says nothing about the coordination across venues. The comparison is an analogy, not a study of paper mills, but it suggests where detection would need to look: at shared images, shared editors and shared submission patterns across journals.

There is also pressure on the supply side. One demonstration had LLMs write 288 complete finance papers from 96 statistically significant signals, each with an invented theoretical rationale Can AI generate hundreds of fake academic papers automatically?. When manuscripts are this cheap to make, the scarce resource for a mill is no longer the paper. It is a venue that will accept and index it. That makes access to cooperative editors and still-indexed journals more valuable and makes moving between journals more likely. Fraud is also finding new targets in the review process itself. Eighteen arXiv manuscripts were found with hidden instructions telling AI reviewers to rate them favorably Are hidden AI prompts in preprints a deceptive research practice?.

The collection only goes so far here. Just one note studies paper-mill journal-hopping directly, and it describes the behavior more than it explains the economics. The corpus has no data on how quickly networks relocate, which kinds of journals they choose next, or whether deindexing slows them down at all. The idea worth taking away is that a fraud network treats journals as replaceable storefronts. Effective countermeasures probably have to follow the people and shared assets across venues rather than shut down one storefront at a time.