First and Last Hour Correlation

We are currently running some analysis on intraday market action and its potential impact on the longer-term performance of the market.  One of the things we did was calculate the 100-day rolling correlation between the percentage change of the S&P 500 in the first hour of the trading day and the last hour of the trading day.  We also calculated the 100-day rolling correlation between the full day's change and the first hour and last hour. 

As shown in the first chart below, the 100-day correlation between the S&P 500's change in the first hour of the day and the last hour of the day had a 20-year peak of 0.54 on September 18th, 2008.  Currently (over the last 100 days), the correlation is back down to 0.08, so the change in the first and last hour haven't been correlated at all recently.  (A correlation of 1 means they would have moved the exact same amounts on a daily basis over the last 100 days.) 

 

Interestingly, the correlation over the last 100 days between the change in the first hour of the day and the change over the entire day is at its highest level in 20 years.  Conversely, the 100-day correlation between the change in the last hour of the day and the change over the entire day has dropped dramatically from a 20-year high last September to the low end of its historical range.

Based on these correlation numbers, the bulk of the trading day's move has been coming in the early morning over the last few months, while the last hour hasn't had much of an impact at all.  We still have a lot of analysis to do before we can see what impact the correlations might have on the future performance of the market, but we're wondering if anyone wants to theorize on what these recent trends are telling us or forecasting.

 

 

Comments

[Random?]

Does it necessarily have to mean anything? It could just be random data. You can plot the correlation between % change in the number of pigeons on wall st. and the % change for the day and probably end up with a similar looking chart.

[dave]

I think that it's (unfortunately) the impact & prevalence of daytrading over longer timeframe trading. Too often the trend is intraday instead of interday.

[Joe Calhoun]

When I started in this business way back when, one of the grizzled veterans told me that the dumb money trades in the morning and the smart money trades in the afternoon. After 20+ years of dealing with all kinds of clients from retail investors to institutional traders, I still find that to be generally true. So factor that into your analysis....

[Norman]

It might have something to do with 'flash trading'. The big boys drive up prices in the morning to suck in the amateur momentum traders to whom they then sell the morning's purchases. Someone is being zoomed.

[Paul Teetor]

International markets are connected, often more than we recognize. International investors are still active in the morning, whereas they are largely gone home by the time New York closes in the afternoon. I wonder if the international trading is giving more weight to morning activity?

[Prepared Investor]

So the trend for the day has been set in the first hour, lately .. Reminds me of those trader rules where one waits until the first hour to see which direction the market heads and then setups the trades ..

May be there are more day traders of this sort these days ..;) just kidding ..

[miles]

when dumb money moves the needle, it's a sucker's rally

[Tamas]

It is not clear whether you include any opening gaps in the first hour of trading.

Index data is notoriously inaccurate for opening prices so I would appreciate if you would clarify how you have treated opening gaps.