Intraday Models
Drill Down: SPY Arbitrage Model Context Model Context Model Correlation VIX Indicator
Drill Down: SPY Arbitrage Model Context Model Context Model Correlation VIX Indicator
| © Capital Context CDS data courtesy of CMA |
| Suffusion theme by Sayontan Sinha |
Hi Dave, brilliant site – thanks. Quick question, I get the context/ES model from having read about it on ZH but do you have an explanation of the divergences in your other models? Do you collect stats on the frequency/timescales of divergences correcting or any other analysis to demonstrate the predictive capability of your models? Many thanks.
The stock indicator and VIX indicator are daily models. We provide the intraday feed mainly to keep an eye out for when the market and model cross (which has trading implications for our TAA strategy), not as the basis for an intraday trading strategy. On average, stocks go from expensive to cheap or cheap to expensive every 2.5 weeks, but this year has been a bit anomalous with the indicator staying expensive almost every day since mid-March. The VIX indicator crosses every 8.5 days on average and we’re still exploring the best way to use it in a trading strategy.
For the stock indicator, check out Introducing the Capital Context Corporate Index and Credit-informed Tactical Asset Allocation for more information on how it works and how we use it.
Dave. Ive been following your commentary and this trade for sometime.. . I was wondering if you can actually say what is in the context model. Proportions, actual contracts.
I know that you look at a currency pair,(EUR/USD or AUD/JPY) CL, as well as the 2-10-30 curve (or the 2-5 -10)
Is it possible to give exact contracts.. Do you use cash treasuries or Futures.. and contracts?
If you can it would be much appreciated.
Thanks
Matt
Dear MadridMatt,
For the CONTEXT model, factors and weights change regularly as our model senses changes in the relationships – and we recalibrate. This is why we created the site. The factors include, as you said, FX carry pairs (e.g. AUDJPY, EURJPY), TSYs (outrights and curves – we like 2s10s30s as a carry driver), Commodities (e.g. Oil and Gold often factor), credit, and swaps (e.g. EUR-USD swap spreads).
We tend to use the cash markets as opposed to futures in order to avoid technical roll issues but futures could be used to implement the convergence trades.
For other models, such as SPY Arb, the components do not change but again their weights do change (are recalibrated). Clients are given the weights daily for execution purposes.
With regard to the Stock and VIX Indicators, the frameworks are relative-value models (which we disclose) we have used for many years in capital structure arbitrage research but are based on our proprietary credit indices.
I hope this helps.
Cheers
Tim
Hi,I would like to become a client. I figured a way to implement the vix model by gamma scalping but i would like to work with you guys in other things. How fo I contact you?
Thank you very much.
Hi Pepiolas,
Thanks for your interest! We’re putting the finishing touches on a pay version of the site that will, among other things, make the hedge ratios for some of the models available. Keep an eye on the site for the announcement.
Dave
I am curious as to the last time your Stock Indicator Model had a divergence that was negligible and not this differential of 40-50 SPY points?
At the beginning of November the difference was down to about 10 SPY pts. However, the stock indicator has shown stocks as expensive for almost three months now and for almost the entire year of 2011.
That is a little different than the past few years but is an indication that, even when stocks improve, corporate credit spreads remain wide.
Dave
Is there a service available to view the intraday models without a delay? Thank you.
Dear Finn,
Yes. We will be starting a subscription service very soon that offers all the frameworks in real-time with no delay. Thanks for your patience as we finalize infrastructure testing.
Regards
Tim
CC,
I’d like to use the Stock Indicator model and I’m interested in the C-Ranks and Factor Scores. If possible, I’d like to get at least an idea of the general form of each model so I have at least some intuition when trying to apply them.
Here’s my impression thus far: 1) Based on the paper posted on SSRN it appears that the Stock Indicator model is an intensity (hazard-based) model.
2) For the Factor Scores, are the factor scores linear; i.e, is a score of 55 10% better than a score of 50? Or, is there some sort of distribution or other type of scale used? For instance, is a GLM used to fit into the factors into a distribution? Do the factors include any intensity- or structural-type credit factors?
3) Do the C-ranks incorporate ideas from 1) and 2) above? In other words, do the C-ranks include some combo of credit-implied equity values, equity-implied credit values, as well as relative rankings on accounting-type measures?
P.S. I apologize in advance if this is the wrong forum for these questions.
P.P.S. The site seems like a great idea. I just want to make sure that I know what I don’t know.
Thx,
Lou
BTW, you’re welcome to rewrite or condense my original question from 12/5 as necessary. I’m still hoping for an answer, though.
thx,
Lou
Hi Lou,
Sorry for the delay in getting back to you. We can’t go into too much detail about the models however let me give you some quick answers.
1) Yes, the Stock Indicator uses default probabilities which are derived from a hazard-based models. For the stock indicator, we convert the full term structure of corporate credit spreads into hazard rates then on to default probabilities. Some color is provided at the end of the SSRN paper.
2) The factor scores are percentile ranks. So a 55 is in the 55th percentile, etc. This is a common approach when ranking credits to avoid wide spreads dominating a particular model.
3) For each credit covered, we create a credit-equity relative value model (which helped inspire the stock indicator approach). So, for a given credit term structure, we have an expected equity level and vice versa. Then, based on our rankings of risk factors, we can extract an equity C-Rank and a credit C-Ranks.
Hope this helps. Feel free to drop us a line at services@capitalcontext.com if you’d like to follow up.
Cheers,
Dave
Dave,
I noticed that the Stock Indicator model is no longer on the Intraday Models page. What happened? Is it coming back? FWIW, I thought that idea was the most interesting given the paper on SSRN.
~Lou
Hi Lou,
The Stock Indicator is really intended more as an daily model. We may put a daily history up on the stock page in the future. For now, you can see it as the gauge on the left-hand side of the page.
Dave
Dave,
A daily history for the Stock Indicator would be really helpful. The Stock Indicator has a body of research behind it and I’ve found it to be a great shortcut to get a credit-implied valuation on the equity markets (albeit, at a daily interval). Also, the historical data could only help for users to test the model results.
~Lou
Hi,
As per the context model correlation to ES. If we are closer to one then is the context model correlating exactly with the ES. Could you tell me the standard deviation data for this,
Regards,
Scott
So are these the raw materials? for intraday
http://finance.yahoo.com/echarts?s=spy#chart1:symbol=spy;range=1d;compare=tbt+xlf+smh+hyg+xiv;indicator=volume;charttype=line;crosshair=on;ohlcvalues=0;logscale=on;source=undefined
Some of the ETFs you mention are indeed components of SPY Arb.
Dear Dave,
When comparing the equities market with the credit market, which credit market do you use for comparison? HYG? JNK? IEF? TLT?
We use a proprietary credit index which better correlates to the equity market. More info here: http://capitalcontext.com/2011/09/21/introducing-the-capital-context-corporate-index/
Dear Dave,
Where can I check out the CDS market quotes?