Do heavily margined stocks do worse afterwards? Method
Written 2026-10-07, before any result was computed. Page: /leverage/mtf/. Code: pipeline/tipsheet/compute/mtf_insights.py.
The design, thresholds and verdict rule are fixed here and are not to be tuned after seeing results.
The question
When a large share of a stock’s value is bought with borrowed money through the margin trading facility (MTF), is the stock’s return over the following months lower than that of similar stocks with less margin funding?
Sample
- Dates: every month-end from January 2018 (when the size lists below begin and the book was large enough to matter) to the last month-end with a full six-month outcome.
- Stocks: those on AMFI’s large, mid and small cap lists in force at that date (each list used only from 15 days after its half-year ends, as on the rest of the MTF page). Membership is taken as it was then, so stocks later delisted stay in.
Margin intensity
For each stock and month-end: NSE’s funded MTF amount for that stock (Rs crore) divided by its average market cap on the AMFI list in force. A stock with no funding has intensity 0.
Groups
Within each AMFI size group (large, mid, small) separately, the heavily margined stocks are the top tenth by intensity among the stocks of that group. The comparison is every other stock in the same size group. Comparing inside a size group keeps the small-cap tilt of margin funding from passing for an effect of margin.
Timing (no look-ahead)
NSE publishes a day’s MTF figures the next trading day. So the portfolio uses the month-end intensity and is bought at the close of the second trading day after the month-end, and returns run from that close.
Outcome
Price return (adjusted for splits, bonuses and rights; dividends excluded, which affects both groups alike) over the next 1, 3 and 6 months. Each month: the equal-weighted average return of the heavily margined stocks minus that of the comparison, in each size group, then the plain average of the three size groups’ spreads.
Statistics
The mean monthly spread for each horizon, its t-statistic with Newey-West standard errors (lags = horizon in months minus one, for the overlap of the windows), the share of months with a negative spread, and the same means for the first and second halves of the sample.
The verdict rule
“Heavily margined stocks did worse” is said only if, for the 3-month horizon:
- the mean spread is negative with a Newey-West t-statistic of −2 or below, and
- the spread is negative in both halves of the sample.
“Did better” needs the mirror conditions. Otherwise the page says the difference is not clear. The 1- and 6-month results are shown beside it as context, not used for the verdict.
Limits
About 100 month-ends, all from one market cycle with a strong small-cap rally. The funded amount is the loan outstanding, not the market value of the shares, so intensity understates leverage after a price rise. This describes the past and is not a trading rule.