Friday, October 16, 2009

Galleon Group Raj Rajaratnam's Insider Trades


There was a pretty big insider trading bust today, involving high level executives investor relations firms, and Raj Rajaratnam of the Galleon Group hedge fund. From what I've been able to gather from the charges, I compiled a set of charts for various of the insider trades. It excludes some in stocks that no longer trade publicly (such as Hilton). Judge the charts for yourself. It would be interesting to develop a more systematic study of trading patterns prior to big jumps, and see how insiders show their hands in the form of the stock's price.


The charges appear here.

Tuesday, October 13, 2009

List of publicly traded companies with industry and sector classification from from Yahoo! Finance

Quantitative traders and investors doing research often need a comprehensive list of publicly trader US equities along with their sector and industry classification. The Yahoo! Finance industry browser is great for that, but getting the data into a neat form requires some programming ability. Since I often see people requesting a list of companies with industry and sector classification, I have decided to post the one I use. It contains the vast majority of US-listed stocks (excluding some very small OTC-BB names). The rows are tab-delimited, and the columns are ticker,sector,industry,market cap (in billions as of 10/1/2009', company name. I hope this is useful for you.

The list is available here .

Wednesday, July 1, 2009

Short-term VIX


The heatmap below shows 10 minute SPY returns depending on the 20-minute fast stochastic oscillator (called kfast here) applied to 5-second bar closes of SPY(vertical axis) and VIX(horizontal axis). The oscillator measures the current price relative to the price range in the last 20 minutes. So, when it takes on a value of 0 it means a new 20-minute low, whereas 1 would be a new 20 minute high. The color of each bin is the mean divided standard deviation (i.e. sharpe ratio, not annualized) for all the 10 minute returns that fall in it.

We can see that rise in SPY with drop in VIX is generally bullish (bottom left orange corner). Rising VIX along with SPY is bearish (bottom right), and so is falling SPY and dropping VIX (top left).

The data used here is 5-second bars for SPY and VIX for 370 trading days starting on 2008/01/10.

Friday, April 24, 2009

Intraday market action leading up to FOMC (Fed) Meetings


You know the drill. We will look at the last several FOMC meetings, and the days that preceded it to gain a sense of what market behavior to expect for the week of April 27. There is a Fed meeting on Wednesday, April 29.

The attached graph consists of a series of subplot. In each subplot we show the price of SPY for all the days up to and including FED day. So, the very last color (yellow) is the day of the Fed meeting. The only exception is the bottom right subplot in which I use the current day (Friday, April 25th) in order to give a sense of what the market has been like in the period leading up to now.

One notable thing is how many fakeouts (mostly down, then up) there are on the day of the announcement, and the overall high volatility of the days preceding it. This is always talked about by traders, and it is nice to observe directly in the data.

Friday, March 20, 2009

Options Expiration Intraday Action


Friday March 20 2009 is quadruple witching expiration. Traders believe that expiration days are difficult to trade and market makers and others play a variety of tricks in order to settle their positions.

I am not sure what the proper way to trade options expiration is, but thought I should assemble intraday prices of SPY (ETF tracking S&P 500) from options expiration Fridays (and some Thursdays) of the last year. Each square in the grid is the intraday SPY price for the options expiration Friday in the title, as well as the trading days preceding it. There are a few bad prints in the data, which I did not have time to correct, but I think it is still quite useful.

These are interesting to look at and maybe guide your decision making somehow, though admittedly the sample is small.