You know how open-plan offices changed behavior? Not because the work changed, but because everyone could see everyone else. The guy who used to nap at his desk stopped napping. The manager who played favorites had to justify it in front of witnesses. The work didn't get harder — the monitoring got cheaper. That's the core mechanism this paper is exploiting, except the open-plan office is a geographic accident where India's electoral constituency boundaries don't line up with its administrative district boundaries, and the people being watched are politicians. The committed claim: increasing the number of elected representatives governing the same administrative area causes new firm creation, lower unemployment, and greater real economic activity. Not correlation — causal, via a geographic discontinuity design. Adjacent regions that are otherwise identical end up with different numbers of politicians purely because of how electoral maps were drawn, creating quasi-random variation. The authors supplement this with a delimitation episode (India's 2008 constituency redrawing) that changes the number of politicians governing the same region over time, giving them within-region variation too. The mechanism isn't that more politicians means more government. It's that non-aligned politicians impose checks and balances on each other. When a single politician controls an area, they can extract rents, steer contracts to cronies, and erect regulatory barriers that protect incumbent firms. Add a second politician from a different party or faction, and suddenly that behavior gets monitored, contested, and exposed. The paper argues this reduces regulatory obstacles, cronyism, and improves public infrastructure provision — all of which lower the cost of firm entry. This sits in a long-running fight in political economics: is political fragmentation good or bad for growth? The Weingast tradition says competition between jurisdictions disciplines government. The common-pool problem literature (Persson and Tabellini, Alesina and others) says more politicians means more spending, more pork, more waste. This paper sides firmly with the competition-and-monitoring camp, but with an important qualifier — the politicians must be non-aligned. Same-party politicians don't check each other. The identification strategy is the paper's strongest asset. Geographic regression discontinuity designs are the gold standard for this kind of question because they compare places that are geographically adjacent — similar in culture, economy, geography — but happen to fall on different sides of a constituency boundary. The delimitation supplement is smart: it addresses the concern that boundaries might correlate with unobserved differences by showing that the same region responds when its politician count changes. Two independent identification strategies pointing the same direction is genuinely reassuring. What's missing is magnitude. The abstract reports direction (more firms, lower unemployment, greater activity) but no effect sizes. We don't know if adding a second politician increases firm entry by 2% or 20%. That's a significant gap for policy relevance — the direction matters less than whether the effect is large enough to justify the coordination costs of overlapping governance. The India-specific context also raises external validity questions: does this mechanism travel to countries with different institutional structures, party systems, or levels of corruption? The obvious next experiment is testing whether the effect reverses when the additional politicians are aligned (same party, same faction). The paper claims non-alignment is key, but without a clean test of aligned vs. non-aligned overlap, the monitoring mechanism is asserted rather than proven. My read: the authors likely explored this but the aligned subsample may be too small for clean identification, or the results were ambiguous. This is probably (a) — a data limitation rather than a hidden negative result.
