What impact cost measures
Start with the price everyone quotes. Between the best buying price and the best selling price sits a midpoint — the ideal price, the number you would trade at if the market would let you deal any size at one price. It will not.
Impact cost is how far your actual average fill lands from that ideal price, expressed as a percentage, for a stated order size.
That last clause is the whole idea. Impact cost is meaningless without a size attached, because a share that can absorb ten of your shares invisibly may move a rupee for two thousand. The exchange therefore publishes it against a specified order value, so that two shares can be compared on the same footing.
And it is not a fee. It is 01.09.02's slippage, defined precisely enough that a machine can compute it from the order book, continuously, for every share listed.
- 200 shares available at ₹100.50
- ₹20,100
- 300 more at ₹101.20
- ₹30,360
- 500 more at ₹102.00
- ₹51,000
- 1,000 more at ₹103.50
- ₹1,03,500
- Total paid for 2,000 shares
- ₹2,04,960
Now double it, and then add everything else
You will want to sell one day, into the same thin book. Assume it is no worse then, which is optimistic, and the round trip costs roughly 5% in impact cost alone.
Put that beside chapter 01.09 so far. The statutory charges on a trade came to a fraction of a percent. The spread on a liquid share cost 0.1%. This one line is twenty to fifty times either of them.
So the share has to rise about 5% before you are level — and that is before brokerage, before STT, and before the spread you also paid. A thin share is not simply riskier. It is more expensive to own, measurably, in a number published before you buy it.
Impact cost is a property of the share, not a verdict on the person trading it.
Where the exchange uses it against a company
This is not an academic measure. NSE uses impact cost as an eligibility test for index inclusion: a company is only considered for the main indices if its impact cost, for a specified order size, stays low enough over a sustained period of observations.
The reasoning is practical rather than moral. An index is meant to be trackable — a fund must be able to buy and sell the constituents in size without the act of trading moving the price against every unit holder. A share nobody can deal in fails that test whatever its business looks like.
Which gives you a free piece of information. If a share is not in an index, one of the reasons may be that dealing in it costs too much, and the exchange has already done that arithmetic. 01.05.03 described how an index is built; this is one of the gates.
The asymmetry that catches people
Impact cost is usually quoted as one number, which hides something important: it is rarely the same going in as coming out.
You choose when to buy. You can wait for a day with more activity, split the order, or simply not buy. The buying side of impact cost is largely voluntary.
You do not always choose when to sell. If you need the money, or the company announces something bad, you are selling into a book that has thinned at precisely the moment everybody else wants out too — which is 01.09.02's correlation, in its most expensive form.
So read the published figure as a floor rather than an estimate. It is what dealing costs on an ordinary day. The day you most want to trade will not be an ordinary day.