How market makers price earnings

[srinir]

looks like ZOES (zoeys kitchen) one of those growthy stocks has negative surprise today

 

[Sig]

Hey sig thanks for the reply. So I have purchased data that gives me the last 50 quarters of company's price move, estimate and surprise. I also get the Monthly return leading into earnings, weekly returns, and daily return right before earnings. After running regressions on MANY companies here is what I found. Earnings surprise has very little correlation to stock move after earnings, I can put up multiple graphs of you do not believe me. Also the volatility before earnings has no correlation to the earnings move. In fact I have applied multiple linear regression models, autocorrelation models, garch models to earnings moves and NOTHING correlates. HOWEVER 70% of the companies returns after earnings are normally distributed with a kurtosis of 3!!! When I said average i meant to say first standard deviation. Have you found any correlations between earnings returns an a variable or variables? That would help alot thanks sig!!!

So first I apologize for my simplistic reply, it sounds like you've looked at this more and with more rigor than I have. Your findings are very interesting, I've put the subject on my list of things to research some more. A couple of random thoughts that you may or may not have already considered:
-Did you look at the distribution of the difference in implied volatility before earnings release minus the implied after versus the actual change in stock? I think the key question here is if the market is systematically over or under estimating the volatility associated with earnings release. If the volatility collapse immediately after earnings release is significantly greater than the stock move would warrant, for example, it would show an over pricing of volatility around earnings, which is what your findings hint at.
-It's actually exceptional that you'd ever see a univariate normal distribution of returns with any large sample of stock returns, this is actually very uncommon. That you should be seeing it at a time when it's even less likely than usual is also something worth looking into.

 

[TheBigShort]

In image 1 those are the distributions of the last 48 earnings for multiple companies that recently had their earnings. (histogram with 20 buckets)
Image 2: same companies. Y axis = price change close to close after event. x axis = surprise

@srinir @truetype @Sig

 

[TheBigShort]

So first I apologize for my simplistic reply, it sounds like you've looked at this more and with more rigor than I have. Your findings are very interesting, I've put the subject on my list of things to research some more. A couple of random thoughts that you may or may not have already considered:
-Did you look at the distribution of the difference in implied volatility before earnings release minus the implied after versus the actual change in stock? I think the key question here is if the market is systematically over or under estimating the volatility associated with earnings release. If the volatility collapse immediately after earnings release is significantly greater than the stock move would warrant, for example, it would show an over pricing of volatility around earnings, which is what your findings hint at.
-It's actually exceptional that you'd ever see a univariate normal distribution of returns with any large sample of stock returns, this is actually very uncommon. That you should be seeing it at a time when it's even less likely than usual is also something worth looking into.

I have looked at the difference in jump vol and diffusion vol. (actually started a thread on it, but didnt get much attention). Something I want to look into but don't have data for it is, close to open returns (gap). because if the gaps are larger/smaller than what the jump vol predicts, there for sure would be money to be made. Especially in the lower market cap stocks. BTW the DXC trade worked out nice. Bot back for $5 ($3.30 gain). Obviously a trade that made money is not necessarily a good trade... but I think it was a good trade.

 

[Sig]

I have looked at the difference in jump vol and diffusion vol. (actually started a thread on it, but didnt get much attention). Something I want to look into but don't have data for it is, close to open returns (gap). because if the gaps are larger/smaller than what the jump vol predicts, there for sure would be money to be made. Especially in the lower market cap stocks. BTW the DXC trade worked out nice. Bot back for $5 ($3.30 gain). Obviously a trade that made money is not necessarily a good trade... but I think it was a good trade.

Better lucky and right than right and unlucky sometimes right! Agree that this is an interesting subject, will look up your other post as well.

 

[murray t turtle]

In liquid underlying's the price of the options before earnings are driven by supply and demand. Let's say a non liquid underlying before earnings. Where only 100 contracts are sold. How do market makers price the earnings event vol?

%% Books have been written on that; but same way we would= to make a profit. BUT like an insurance co, they are good @ risk control[usually,LOL]. I ignore most of IBD options comments on earnings/options.100 contracts doe$ not sound liquid @ all, unles$ its QQQ or SPY options.............................................................I returned my daytradeing option book

 

[TheBigShort]

%% Books have been written on that; but same way we would= to make a profit. BUT like an insurance co, they are good @ risk control[usually,LOL]. I ignore most of IBD options comments on earnings/options.100 contracts doe$ not sound liquid @ all, unles$ its QQQ or SPY options.............................................................I returned my daytradeing option book

I wouldnt consider trading earnings day trading. Its a very statistical process, where if you have a good model and good execution you should be able to make money in the long run

 

[truetype]

In image 1

How do you like RStudio on Mac?

 

[newwurldmn]

I wouldnt consider trading earnings day trading. Its a very statistical process, where if you have a good model and good execution you should be able to make money in the long run

I think it will be very capital intensive and returns will be mediocre (by retail standards).

 

[TheBigShort]

How do you like RStudio on Mac?

I have a mac and pc, I really find no difference if anything the mac runs a bit faster when doing large computations. Do you use R studio or something else?