Why Does BSE Close Higher? Stale Prices, Thin Trading and a Closing-Auction Experiment

In earlier work I found that thousands of Indian companies end each day with two slightly different official closing prices, one on NSE and one on BSE, and that for the least-traded companies BSE’s close tends to sit above NSE’s. This paper asks why. Using 491,644 pairs of closing prices for 2,212 companies (2025-10-20 to 2026-10-26), I test four explanations. First, BSE’s close is often stale for thin stocks: for the least-traded fifth, each 1% rise in a stock’s NSE price goes with a gap 6.6 basis points lower, as if BSE had not caught up. Staleness does not explain the bias itself, however: on days when NSE barely moved, BSE still closed 12.0 basis points higher on average. The bias is concentrated on days with almost no trading on BSE. After a gap, BSE does about 84% of the correction the next day, so BSE’s close is the noisy price. Thin stocks also trade on BSE mostly on days when they fall on NSE, and a BSE close that does not fully follow the fall sits above NSE’s. I date a rise in the typical gap to 2026-01-19 that is not explained by changes in the data, the mix of companies or market volatility. Finally, I use the introduction of a closing auction for stocks with derivatives on 2026-08-03 as a natural experiment. For the most-traded fifth of companies, the typical gap between the two exchanges’ closes rose from 3.7 to 8.0 basis points, while other companies barely changed. Separate auctions on each exchange appear to make the two official closes agree less, not more.