6

Curious if there's a formula for this. There are some notorious cases of software (accidentally) and people (purposefully) quickly unloading a given security, causing the price to (temporarily, generally) plummet.

Is there a general formula for determining how many issues can be transacted in a given time period so that the market price is minimally impacted?

Eric
  • 163
  • 5
  • By "formula", I think that what you really want is some empirically tested theory on how trade volume causally effects price fluctuations. I'm curious to see if this has been addressed in academic literature. – jmbejara Dec 09 '14 at 22:40
  • @jmbejara I would think with every other formulaic indicator that works with ema/sma/volume/liquidity/spread someone would have done this math somewhere. I'm using "formula" in the sense that since algorithms no doubt use this math, it's been written as a formula somewhere. – Eric Dec 09 '14 at 22:46
  • This is a case where practitioners and academics are going to differ. To refer to this as "doing the math" implies that there is some sort of theory out there that describes this phenomenon. Practitioners might have some empirically driven theory that they have tested whereas academics (in some cases) might argue that they don't even know why people trade assets in the first place! So, I don't know what a good answer to this is. – jmbejara Dec 14 '14 at 22:53

1 Answers1

3

There are countless models, from different perspectives (theory/econometric/etc.) on price impact of trade volume. The seminal model in market microstructure---the Kyle model (85 Econometrica) addresses this very issue. There are many descendents of this model.

To demand a "formula" is a little simplistic. The usual framework for these models, if micro-founded, is strategic. The price impact is a property of the resulting game-theoretic equilibrium. In the Kyle model, one has

$$ \Delta p = \lambda\cdot Q $$

where $\Delta p$ is the price adjustment of market maker when the incoming total order flow is (only market orders are considered) $Q$. The coefficient $\lambda$ (Kyle's lambda) is a measure of price impact, and is determined by the severity of the adverse selection problem faced by the market maker.

Ubiquitous
  • 16,925
  • 4
  • 37
  • 84
Michael
  • 2,619
  • 13
  • 22