Theory teaches us that liquidity varies with volatility and asymmetric information. Using daily returns for all points seen in 2010, the three vols (all annualised) for these three underlyings work out to:
- INR/USD: 6.6%
- Nifty: 15.2%
- Reliance: 24.5%
The data that I got was for the top 20 prices in the order book. This shows the following picture of how impact cost when buying (in basis points) varies with transaction size (measured in million rupees):
The blue line is the impact cost of Reliance. The quantity available in the top 20 prices is low, and the impact cost therein goes up rapidly from around 1 basis point `at the touch' to around 9 basis points at the end. This is the combination of high asymmetric information and high volatility.
The black line is the impact cost for the April Nifty futures. Within the top 20 prices, roughly Rs.100 million or $2 million was available at this time. A single market order for $1 million would get done at a one-way impact cost of roughly 2 basis points and a single market order for $2 million would get done in a bit less than 3 basis points.
The INR/USD futures yields a lot of quantity within the top 20 prices: all the way out to Rs.250 million or roughly $5 million or 5000 contracts. The impact cost for $1 million is below 1 basis point and that at $2 million is below 2 basis points - so this is more liquid than the currency forward. I found it intruiging that at $2 million (i.e. Rs.100 million) the impact cost for Nifty and for the INR/USD were not that different.
To look at these orderbooks in realtime, use the links in the previous blog post.
Is it possible to factor out the effects of vol so as to be left with effects only due to differences in information asymmetry? Thanks.
ReplyDeletei think you have diktat in a good way.thanks
ReplyDeletei like that you discussed the liquidity in really well manner.
ReplyDelete