Granularity of the Bid/Ask spread relative to STD/ATR

It's a measure every trader should apply whether to even make a decision in a market.

Only derivatives with the smallest ratio, should be traded.

The higher the 'granularity', lower the probability any decision made will yield a profit.

 

(bid-ask + commission) / vol = standardised measure of costs. Low costs == good.

GAT

 

It all comes out in the wash. The most liquid markets are the most thoroughly arb'ed. E.g. factor trading, in the academic literature, generally 'works' best on smallcaps.

 

A little food for thought:

(Fresh-cooked from a noon-time scrape.)

 

It does, but the question is whether the return is still there after costs, so you need to know the costs. Also for many phenomenon (eg trend following) the pre-cost returns are the same across different instruments, regardless of costs.

Anyway I think it's important to know costs. At some point standardised costs are too just high for the thing to be tradeable, when vol gets way too low (Japanese CHF STIR a few years ago, Bobl more recently).

GAT

PS I have to say I hate the word "arb'd" being used to mean something other than arbitrage....

 

Maybe. And maybe not.

Sure, it makes us all feel warm and fuzzy with regard to the Efficient Market Hypothesis that lets us make sense of the world. But in practice? In harsh, repeatable, ugly reality?

Go inspect the SPX options bid-ask spreads.
Check out the weekly(s); check out the sub-weekly(s). Check out their respective volume. (Check out their respective OI! Doesn't matter.) As a conclusion, it would be safe to say that, regardless of volume, there is a consistent bid-ask spread constructed by market-makers of about ~50% at a roughly ten-delta, 1-strike vertical spread. A factor that consistently shrinks that bid-ask spread is time.

Now, though, check out the monthlies. You'll find the bid-ask spread is now ~400%. And the volume is upwards of 5x or 6x or more. Yowie!

There is a simple explanation for that dissimilarity, and totally in keeping with the EMH, but it's not immediately apparent. At any event, *volume* and *liquidity* are very much not the same thing. Yowie.

*[BTW -- a quick cheat: check out the .pdf I posted in this thread earlier....]

 

There's no "maybe." Factor trading, in the academic literature, generally 'works' best on smallcaps is unambiguously true. If you don't believe me, ask GAT.

 

No one challenges your bibliographic skills.

 

http://schwert.ssb.rochester.edu/f532/jfe12_ff.pdf
Table 2

GAT

 

And the more interesting question, does it still work after costs?
http://faculty.som.yale.edu/Tobiasmoskowitz/documents/TradingCostofAssetPricingAnomalies.pdf

table IV

GAT