India's regulatory architecture is staffed by officers who rotate every two to three years; the formal system has no mechanism to transfer the tacit knowledge that accumulates with a posting. Yet matters proceed across transfers, complex regulatory questions resolve over multi-year horizons exceeding any single tenure, and trajectories survive succession. An ecosystem of external actors fills the memory discontinuity the architecture creates. What does this layer do that the system itself does not?
The transfer cycle's anti-capture logic is a settled feature of Indian administration; what follows from the choice is less examined. The architecture creates a memory discontinuity by design; the consequences of that discontinuity shape the entire ecology around the file.
The first consequence is procedural. Every transfer creates a specific institutional event: the successor has the right to fresh review of any matter the predecessor was considering. An approval-in-principle from the previous officer carries no automatic obligation forward. A matter at an advanced stage of consideration may be sent back for fresh review, not because of any defect but because the new officer wants to satisfy themselves independently before concurring. Each successor has the right of fresh review; an approval-in-principle from the predecessor carries no procedural obligation forward.
The second consequence is informational. The incoming officer inherits the file but not the tacit knowledge that surrounds it: which counterpart in another ministry actually decides, which industry association's representations carry analytical weight, which sector is being quietly watched politically, which previous representation was rejected and on what grounds. None of this appears in any briefing note. The briefing captures formal positions; it does not capture institutional fluency. The new officer reaches fluency the same way the predecessor did, through months of lived experience, by which time the cycle is preparing to move them.
The third consequence is behavioural. When institutional memory resets every two to three years, the system cannot distinguish a good-faith actor from a bad-faith actor, and applies equal institutional suspicion to all. A company that has spent five years building trust with the previous officer arrives at the new officer's desk identical to a company that has spent five years failing to comply. Both must re-establish trust from a baseline of institutional unfamiliarity. Trust earned with one incumbent does not transfer to the next; it is held by the officer, not by the institution.
The fourth consequence is ecological. The memory discontinuity rotation creates is not allowed to remain empty. It is filled by actors outside the architecture: industry associations with sustained relationships across ministries, regulatory consultancies that maintain continuity through officer transitions, law firms that hold the institutional history of their clients' matters, and government-affairs advisers whose value lies precisely in remembering what officers cannot. These actors are not workarounds the system tolerates; they are products the system produces. The architecture's choice of rotation as an anti-capture instrument generated a structural demand for memory that the architecture itself does not supply. The intermediary class fills that demand. The external memory layer is not a workaround the system tolerates but an architectural product the system produces; the demand for it is created by the choice of rotation the architecture itself made.
This layer occupies an institutionally peculiar position. The system formally does not acknowledge it. There is no statutory provision for consultants in policy discourse, no procedural recognition of advisers in regulatory submissions, no formal role for industry associations beyond the limited consultation structures that some ministries operate. Yet the layer exists, operates continuously, and is structurally necessary for the architecture to function. The intermediary remembers what the file does not capture: the conversations across transfers, the implicit understandings, the sequence in which arguments were made and rejected, the officers who took an interest and the officers who did not.
For organisations engaging with the regulatory system, this changes what good engagement looks like. The official channel is the file. The institutional memory channel is not the file. A representation that lands on a new officer's desk without an account of where the matter previously stood is asking the officer to begin from zero on a question that has institutional history. Companies that engage well work through the external memory layer, not around it; they recognise that the system itself has no instrument to retain the history their matter accumulates over multiple officer tenures.