Tuesday, April 15, 2008

Overnight and Daytime Market Regimes



The charts above decompose the S&P 500 Index (SPY) and 10-year Treasury rates ($TNX) into two components: changes that occur from close to open (overnight) and those that occur from open to close (day). For purposes of comparison, the charts are set to an arbitrary index value of 100 on 12/31/04. It doesn't take much analysis to see that the overnight and day markets behave quite differently. In fact, the correlation between overnight price changes and subsequent day changes is -.05 for SPY and .03 for $TNX. These, in essence, are separate markets.

The top chart illustrates how much of the stock market's bullish trend since 2005 has been a function of overnight price change. Indeed, a pure daytrader experienced none of the benefits of this bull run. To be sure, overnight exposure brings its risks, but closing positions at day's end also has greatly dampened reward.

Notice how, for the most part, interest rate changes have been much more pronounced during the day session compared with overnight: swings up and down tend to be larger. A good part of the trending behavior in rates has occurred during the day--at least until recently.

Which gets us to one of the most interesting aspects of this exercise. Until January of this year, much of the drop in stock prices since mid-2007 occurred during the day session. Similarly, much of the fall in interest rates (flight to quality) also occurred during the day. Since January, however, the day behavior of stocks has been relatively muted--as has been the day behavior of rates. Instead, we've seen pronounced overnight weakness in both stocks and rates since January.

This shift in regimes may be quite meaningful, reflecting a thematic shift in the markets. Much of the drop in shares and rates from mid-2007 through the January lows was a function of credit fears, whose epicenter has been in the U.S. Since January, however, U.S. stocks and rates have been responding increasingly to global recession fears, weakness in global share prices (across Asia most notably), and preopening economic and earnings reports related to recession.

It's interesting that the number of common stocks on the NYSE making fresh 52-week lows hit their highest level in January at exactly the time this overnight/day regime shifted. I believe that January low represents a pivotal point at which markets shifted their focus, such that overnight events--and the global economic picture--are now weighing more heavily on stocks.
.

RELEVANT POSTS:

Are Technical Indicators Relevant for Daytraders?

How to Lose Money Buying in an Uptrend

The Multiple Personality of the Stock Market
.

7 comments:

Trader Steve said...

As you said before, market changes its character without our knowings. Your analysis always gives me very timely and early warning to the emerging changes. Thanks.

Ryan said...

Brett,

Another great post. Thanks for all the hard work. One of these days, I'm going to sit down and read all of the posts you've ever posted. It'll be like a twilight zone marathon, except not as scary. I have a question for you. Can you point me to a post where you listed the 17 stocks in your institutional basket? Thanks again,

dgov

Ziad said...

Regarding your earlier twitter comment about how yields have been rising in the last couple of days but stocks not following suite in step with the normal regime that has been around for quite a while, I think it might have to do with the fact that with inflation running so high and the fed having less and less of a margin to reduce rates, the bond market is looking out to rising, or at least not declining rates in the future (The $TNX is at multi-year support too). When you have these types of expectations taking stage, bonds and stocks tend to move together instead of in opposite directions.

As for money flowing out of bonds and into stocks as a sort of sentiment indicator, it could just be that money flowing out of bonds now is going into commodities and not the stock market. This dynamic, coupled with the fact that yields are at multi-year supports, could mean a regime change is about to come regarding the relationship between stocks and bonds, at least temporarily.

Brett Steenbarger, Ph.D. said...

Hi dgov,

I'll list the stocks in my basket in the next post covering technical strength--

Brett

Brett Steenbarger, Ph.D. said...

Hi Ziad,

Yes, I think it *is* possible that we'll see rising rates (especially at the short end as the Fed stops easing) and an end to the flight to quality trade vis a vis stocks. Thanks for the observations--

Brett

MarketSci.com said...

Brett - I know that this is a bit of an old post, but it still has me intrigued. I’ve tried running some of my strategies through either just the daytime or overnight regimes to see if it improves the signals. Nothing fruitful as of yet, but will continue to throw it against the wall. Just a thought I’m having as I write this (but I haven’t tried yet): I’m wondering if the much less noisy overnight data could be used in a trend following strategy – perhaps less prone to getting shook out in those volatile (but trend-less) daytime moves. Downside is I don’t know how much stock I’d put in a trend-follower with only 15 years of testing. Just some thoughts…

I did a write up on my blog duplicating your numbers but extended back to ’93: http://marketsci.wordpress.com/2008/07/21/the-markets-are-nocturnal-daytime-vs-overnight-performance/

Hope all is well,
ms

MarketSci.com said...

Brett - interestingly, gold shows the same stark contrast between daytime and overnight regimes (albeit not as straightforwardly as the S&P 500). Nearly the entire bull run up since 2000 looks to be the result of the overnight markets. Data here:
http://marketsci.wordpress.com/2008/07/22/gold-is-nocturnal-too-sometimes/

Enjoy,
michael