This Doesn’t Always End Badly But If It Does…

On Tuesday afternoon, just before the close, we published a quick post on market breadth called “It’s Hard To Call This Healthy.”

Essentially, that was just a short reminder that despite Tuesday’s rebound, this market is looking thin on all kinds of measures.

The overarching point there, and in multiple other quick posts we’ve blasted out recently, is that it’s important to look under the hood. And because it doesn’t take much in the way of homework to do so, you won’t be afforded much in the way of sympathy if bad breath finally catches up to benchmarks that, while failing to floss, have managed to obscure the halitosis with Listerine’s latest best-selling flavor: “minty fresh mega cap.”

Then again, when everyone is all aboard the passive SPY/QQQ bandwagon, all that matters are the headline numbers which, in case you haven’t noticed, have demonstrated a rather remarkable tendency to rise come hell or high water for the better part of a decade.

So what to make of Tuesday’s action in the U.S. which saw the S&P, the Nasdaq, and the Russell all up 1% or more…

….the Dow post its best day since April…

… but on SPX volume that was 20% below the 30-day average?

Well, Bloomberg’s Cameron Crise decided to take a look, and as usual, he’s done some interesting math and been kind enough to share it with the world (lucky you).

Read more below…

Via Bloomberg

My colleagues have mentioned how dire U.S. equity volume was today despite the rip in stocks. Conventional wisdom suggests that one shouldn’t believe melt-ups that don’t have big buying behind them, but is that actually right? I decided to take a look.

Since the start of 2010 the S&P 500 has had an average daily gain of 3.6 bps and average one week gain of 17.6 bps (the numbers don’t quite match up due to rounding.)

I filtered all trading days that have shown at least a 0.5% gain with volume 15% below the one month average, then compared subsequent performance on a one- and five-day horizon.

On balance, it looks like there is something to the old saw.

Average one-day performance after a big low-volume rally was -8 bps over a sample of 110 observations. One week performance was even worse, with an average loss of 18 bps, quite a contrast to the aggregate performance of stocks over the past seven years or so. A word of caution, however: returns are not evenly distributed. While the median one-day return was -5 bps, the median one week return was actually positive 20 bps. It turns out that the averages are driven down by a tendency toward large losses after low-volume rallies.

Over our sample period, if the market does go down the week after a low-volume rally, it falls by an average of 215 bps. That’s quite a bit worse than the average losing week for the full sample of 159 bps. In sum, there’s no guarantee that the market will drop over the next five days, but if it does: look out below!