Does the Nikkei Predict Moves in the S&P500?

It's Sunday night as I type this, but the Monday trading day is already half over in Japan. I can see the Nikkei is down about 3% for the day, and I'm wondering if this will have any impact in the US, once the US markets open tomorrow morning. Like all good analytical finance people, I decided to run some tests using regression models.

Question One: Does the daily point move in the Nikkei225 predict the daily point move in the SP500?

Answer to Question One: Yes, but not much. I ran my regression based on daily prices from Thursday, January 4th, 2007 through Friday, April 11, 2008. It turns out the Nikkei does significantly predict a little over 1% of the total price move in the SP500. The problem is it doesn't predict the other 98%+, so this doesn't really present a trade-able strategy (click here for regression results).

Question Two: Does the Monday point move in the Nikkei225 predict the Monday point move in the SP500? Since a lot of newsworthy things can happen over the weekend, perhaps Monday's Nikkei price move has more predictive power.

Answer to Question Two: Yes, it has more predictive power, but still not enough. Based on my regression results, the Monday price move in the Nikkei predicts about 11% of the total Monday variation in the SP500. This is certainly more significant that my previous test, but it's still not good enough since it doesn't predict the other 89% of the variation (click here for regression results).

Question Three: Does the Monday point move in the Nikkei225 predict the change in price for the SP500 from Friday's close to Monday's open? A lot can happen from the SP500 open on Monday to the SP500 close on Monday, so perhaps a better test is simply Friday's SP500 close to Monday's SP500 open.

Answer to Question Three: Yes, it has some significant predictive power, but still not enough. This regression predicts about 9% of total variation which is better than question one (above) and worse than question two (also above). Unfortunately, this test does not explain the other 91% of the total variation (click here for regression results).

Bottom Line: I cannot think of any trade-able strategies based on these regression results (not yet, anyway). At least this puts my curiosity to rest (for now). Let's hope the other 89% of tomorrow's variation in the SP500 (the 89% not explained by the Nikkei) is great news!

Comments

[Jan Sopoci]

What about adding in some of the other global indexes (Hang Seng, ASX, Europe, etc.) that are open for business before the US markets open? Reason I mention it, is the "meltdown" that happened the day after MLK day. Was holding a position in SDS, and logged onto Bloomberg just before going to bed, about 2AM, CST, and saw EVERY major index was down by 3, 4, 5%. Exited the position the next day at a tidy profit.

[mark.d.hines]

Hi Jan,
That's not a bad idea. I'll try to run some regressions using more global indexes to see if I can increase the explanatory power of the model. Thx, Mark.