Thursday, May 28, 2015

New Highs In Cumulative A/D Is Not Confirmation Of A Healthy Market

Summary: Cumulative advance-decline (A/D) is one of the most popular ways to measure market health using breadth data. But cumulative A/D has given no warning before any drop in equities of 5% or more since the 2009 low, including the 20% drop in 2011. That cumulative A/D made a new high last week is not confirmation that the underlying market is healthy.

* * *

In September 2014, the cumulative advance-decline (A/D) line for the S&P hit a new high. A month later, the stock index was down 10%.

Twice in December 2014 and again in February 2015,  the cumulative A/D line hit new highs and within two weeks the stock indices lost 5% (twice) and 4%.

Among the ways of measuring market health, this is one of the consistently least useful. There are two big problems with it.

The first is conceptual. Cumulative A/D takes the number of stocks moving up on a day and subtracts the number of stocks moving down. That sum is then added to yesterday's total.

If a lot of small stocks move up by one penny on low volume but slightly fewer large companies move down a dollar on high volume, the cumulative A/D line moves higher.  Or, if a large number of defensive stocks move up while "risk-on" cyclical stocks move down, the cumulative A/D line moves higher. Neither one of these is a healthy sign.

The second problem is empirical. Every fall of 5% or more since the 2009 low has started from a new high in cumulative A/D. This includes several 10% declines and the 20% decline in 2011. It's hard to be enthusiastic about a measure of stock market health that gives no warning before a 20% fall in equities. These instances are shown below below (cumulative A/D in top panel and the S&P in the lower panel; the zig zag is set for declines greater than 5%).


Wednesday, May 27, 2015

Tuesday's Sell Off Low Likely To Be Retested

Summary: Yesterday, stocks fell more than 1% on intense selling pressure. Trin spiked over 2 for the first time since early March. Declining volume was more than 8 times more than advancing volume. It would be very unusual for that level of selling pressure to dissipate immediately. More likely, Tuesday's low will be retested in the days ahead.

* * *

On Tuesday, US equities sold off hard. The S&P dropped more than 1%. Of significance, selling pressure was intense: declining volume on the NYSE was more than 8 times greater than advancing volume, the highest since late January.

Moreover, Trin (also called the Arms Index) closed above 2.0 yesterday. Trin is a breadth indicator. It is derived by dividing the advance-decline ratio for issues by that for volume. A close over 2 means that down-volume was twice down-issues; in other words, stocks fell on relatively high volume.

A spike higher in Trin like yesterday's can often be near a low in the equities market. That is especially true if the market has been selling off for a week or more. In this case, a high in Trin marks capitulation. A relevant post on this indicator is here.

This makes today's action noteworthy. The S&P gapped up overnight and then rose nearly 1%; the Nasdaq rose 1.6%. Yesterday's low was not retested.

It would be unusual if Tuesday's low remains uncontested in the days ahead. Down momentum like yesterday's almost always takes more than one day to dissipate.

Let's look at recent examples.

The first chart looks at the past year: the S&P is in the top panel, Trin is in the middle panel and declining volume relative to advancing volume is in the lowest panel. Every instance of a Trin spike higher involves at least one lower low in the S&P, even if it came after a strong bounce like today's (note September 2014 and January 2015).


Saturday, May 23, 2015

A Set Up For Volatility To Soon Rise

Summary: The S&P index is making new highs. The trend is higher. But in the process, the trading range has become very tight. In the past, this has been followed by higher volatility and limited near term upside for equities.

* * *

Over the past five weeks, the trading range for the SPX has contracted dramatically. Just 18 points separates the last 5 weeks’ closes (2118, 2108, 2116, 2123, 2126). 4 of these 5 closes are separated by less than 0.5%.

Moreover, the range between the weekly open and close for SPX the last 4 weeks has been 11 points,  6 points, 7 points and 5 points. That averages to a mere 0.3% open/close range each week.

Periods of contraction in the market are typically followed by expansion during which volatility increases. Presented below are several studies that suggest this is likely over the next week or two.

First, Salil Mehta of Georgetown notes that SPX has gone 8 days without a 1% intraday decline from a high. Streaks this long are rare and typically end after 9 days, meaning that a 1% decline is due next week. This implies an intraday visit to roughly the 2100 level is ahead (his post is here).


Thursday, May 21, 2015

GDP Seasonal Adjustment: There's No Smoking Gun

Summary: First quarter GDP growth was stronger than originally estimated. When adjusted for seasonality, the economy grew 1.8%, which is consistent with the average growth rate over the past three years. It is also consistent with a wide variety of economic data (employment, wages, housing, consumption) released in the past several months. And, finally, it is consistent with the message being sent by the treasury market.

* * *

This week, the Bureau of Economic Analysis (BEA) announced that it would be revising the seasonal adjustment methodology it uses in estimating quarterly GDP.  For anyone following US macro data, this announcement was of no surprise (read further here and here).

For at least 15 years, the BEA's methodology has consistently underestimated first quarter GDP. If BEA's methodology worked perfectly, first quarter seasonally adjusted real GDP growth should not be consistently higher or lower than growth in any other quarter.

That the methodology needed revising was obvious. First quarter GDP was 1% lower than the remainder of the year from 2000-10 and 2.3% lower from 2010-14 (data from the SF Fed; full post here).



Wednesday, May 20, 2015

Investor Opinions Have Become Extremely Uniform, And That's Not Good

Summary: Investor opinions have become extremely uniform. By some measures, they are the most uniform in 25 years. In the past, this has corresponded to a period where equities have lacked significant upside momentum. That appears to be a quite likely outcome until investors become more varied in their market outlook than they are today.

* * *

Last week, the number of investment advisors who are bullish US equities dropped below 50%. For many, this was seen as sign of bearishness.

Indeed, relative to the past 16 months, investment advisors were relatively bearish, to the point where the S&P has had a strong tendency to move higher in the ensuing weeks (yellow shading; chart from Investors Intelligence).