Jul 21, 2003

http://www.sentimentrader.com

Thursday, July 24, 2003 8:20 PM EST

I’ve seen some suggestions recently that the low VIX readings we have been seeing (we dropped below 20 briefly today) were normal considering we are entering the Summer months. This is partly true, as the chart below illustrates.

We can see from the graph that the average daily range (i.e. the high minus the low, or how much the VIX travels intraday) subsides considerably during the late Spring / early Summer months of May, June and July. Not surprisingly, those three months also show the lowest average VIX reading of right around 20. We can also see that entering the Fall, volatility tends to pick up dramatically, beginning in August and peaking in October, when the average range rises to well over 3.0 per day (meaning the VIX travels an average of 3 points from high to low intraday) and the average VIX reading of 26 is 6 full points higher than it is in June. No doubt this has a lot to do with the major sell-offs seen during September and October.

Certainly some part of the recent low VIX readings can be attributed to seasonality. Let’s take a look at a “corrected” VIX, which removes the seasonal bias from the indicator, and compare it to the traditional indicator.

There is not a whole lot of difference between a seasonally-corrected VIX (the red line) and the traditional VIX (the tan line) as quoted by the Chicago Board Options Exchange (CBOE). Interestingly, for the past couple of months the VIX has actually been running a bit higher than its long-term mean value for these months, which you can see above as the red line has been above the tan line for quite some time.

Let’s look at this one other way. The chart below shows a 10-day moving average of the deviation of the VIX from its long-term mean values for each month of the year. This type of chart would assign a much greater significance to a VIX reading of 19 if it occurred in October than it would if it occurred in June, since the VIX tends to be significantly higher in October than it does in June. A reading of 19 in the fall would likely signal much more investor complacency than if it had occurred during the early summer.

We can see that over the past five years, when the VIX stretched more than about 50% from its mean value for a given month, it preceded a short- to intermediate-term market rally without exception. Conversely, when the VIX began to fall back to (or drop below) it historical monthly averages, the S&P 100 (OEX) often had difficultly making much headway, with a couple of notable exceptions in 1998. Currently, the VIX is hovering just above its monthly average, as I noted above. This puts it close to other periods seen in the past few years when the VIX dropped below its monthly average, each of which of course lead to a market decline soon afterward. We’re not yet quite as extreme as those other instances, but getting there.

I’ve been asked several times over the past few days why I haven’t mentioned the bullish configuration of OEX options, namely the put/call ratio. The reason is because I don’t think the traders there are being particularly bullish, despite the low level of the traditional put/call ratio. As you may know, I view the OEX put/call ratio in a non-contrarian manner, meaning I think low OEX put/call ratios are bullish for the market, while high OEX put/call ratios are bearish. This is exactly the opposite of the equity-only put/call ratio. OEX traders simply tend to be better market timers than those who trade equity options. Lest you not believe me, here is a chart of the 10-day moving average of the OEX put/call ratio:

We can clearly see that generally, high OEX put/call ratios coincided with imminent market declines while low ratios tended to precede market rallies. Obviously, the indicator is not perfect, and gave some major false signals such as in October 2002. However, in general, the indicator is fairly good when it reaches an extreme. Currently, the ratio is the lowest it has been since the low in July 2002 after peaking at a very high level in early June. This appears to be quite bullish, so what’s the problem?

The issue I have is that open interest is not confirming. Besides volume, the CBOE also reports how many option contracts are “open”. A contract is opened when a trader buys or sells a new option position. Open interest in the series will increase if a buyer enters a new long position while the counter-party (the seller) to the trade enters a new short position. Conversely, open interest will decline if the buyer is closing an existing short position while the counter-party (the seller) is closing an existing long position. It can all get very confusing, but basically in the OEX you want to see high call open interest when approaching a market low, and high put open interest when approaching a market top. That tells us that OEX traders are building new positions, and not just manicuring existing ones.

For the past couple of weeks, what we have seen is a persistently high open interest in OEX puts versus OEX calls. Expirations always play havoc with open interest ratios, since open contracts are effectively closed out by expiring, but if you look at the OEX open interest chart on the site, you can see how high peaks in the open interest ratio (put open interest divided by call open interest) often corresponded to market highs, while low troughs in the open interest ratio corresponded to market lows. This reinforces the fact that OEX traders often build up substantial put positions versus call positions near market highs, and substantial call positions versus put positions near market lows.

Unlike the low in July 2002, when both the put/call ratio and the open interest ratio were very low (suggesting OEX traders were concentrating on call options, and building up new positions), currently the put/call ratio is low while the open interest ratio is rather high, and is closer to that seen near market peaks, not market lows. To try to simplify this a bit, I created a single indicator which combines the activity between the two ratios. The chart below shows the percentage of put volume going into new positions versus the amount of call volume going into new positions, on a 10-day moving average basis. If the indicator (the red line) is high, then that means that a significant amount of the put volume has been going into new positions, while call volume has been going to close out existing positions, or opening a smaller amount of new positions. This is bearish. Conversely, if the ratio is low, then they have been building up call positions and closing out (or concentrating less on) put positions, which is bullish.

Let’s take one day as an example. On July 15^th, 2003, there were 39,943 calls traded on the OEX index, but call open interest only increased by 1,243 contracts. On that same day, there were 26,619 puts traded and put open interest increased by 12,590 contracts. So, only 3% of call volume went into new positions, while 47% of put volume did. This seems a lot more bearish to me than the put/call ratio 0.66 would have indicated.

The chart below shows this type of activity over the past year:

We can see from the chart that generally, when OEX traders are moving into new put positions to a greater degree than new call positions (pushing the red line above 0%), it has preceded at least short-term market weakness a good amount of the time after becoming extreme. On the other hand, when the preference is for calls and the indicator drops below 0%, then it corresponded quite well with tradable market lows after an extreme was seen. Currently, this indicator is well above 0%, and is at a point more closely associated with market peaks than market lows. This tells us that there has been a clear preference among OEX traders for new put positions as opposed to new call positions, and from their history, that is a bearish indication for the market going forward.

The broader market has not been able to rally (and hold) from the oversold readings I mentioned on Tuesday. While there is still a chance for it to do so, the odds don’t look particularly inviting. If we don’t see a more substantial bounce soon, then this would be yet another confirming indication that the character of the market has changed and we are heading lower. At the time, I said that my preference remained with shorting strength. I don’t see any reason from a sentiment perspective to change that outlook now.

- Jason Goepfert

Disclosure: long OEX puts, long SPX puts

This disclosure is not intended as trading advice in any form. It is meant as a note to subscribers that the author may have a position directly affected by the market outlook reflected in the commentary. Although the author takes great pains to remain objective in any commentaries, it is only fair that readers should know that the author may have taken positions in accordance with his market outlook. Positions can and do change at any time, without notice to the reader.

Tuesday, July 22, 2003 8:33 PM EST

“Selling short” involves selling a security first, then buying it back later - hopefully at a lower price. It is essentially the same as a regular stock transaction, only in reverse, and allows one to profit from a decline in price. “Short interest” is defined as the total number of shares of a stock that have been sold short, and not yet covered. The New York Stock Exchange keeps these short interest figures for the stocks traded on that exchange and reports the data with a delay once per month. Watching these short interest figures ebb and flow gives us a clue as to how much potential “fuel” there is from those holding short to possibly cover those shorts (by buying shares) and thus provide some support for the market.

For the month ended July 15^th, short interest on the NYSE declined 3.2% from what was seen in June. Such a drop is quite rare, as short interest has shown a secular rise since the 1940’s, and went parabolic along with the market beginning in the early 1980’s. In fact, since 1980, only 16% of the months have shown a month-to-month drop in short interest of 3% or greater. The closest recent comparison is January 2001, which showed exactly a 3% drop from December 2000.

Because total volume has also shown a secular rise, it is somewhat misleading to look at short interest all by itself. The most common way to correct for volume patterns is to create a “short interest ratio”, which is the short interest figure divided by some type of volume reading. Usually, the ratio is constructed by using an average of the past month’s volume, but volume is very seasonal, so I prefer to use an average of the past year. The theory behind the ratio is that periods of high short interest show extreme investor pessimism, and should be bullish for the market going forward because if the market begins to show some strength, those holding short will likely begin to buy back their shares, thus providing an additional source of demand. Conversely, periods of low short interest mean that traders have covered some degree of their short positions, often a sign of confidence about rising prices.

The chart below shows this short interest ratio:

The ratio as shown above has done a reasonably good job at timing the market long-term. We can see that short interest spiked higher near the bottom in late 1990, and kept rising as the market did. This is one of the few sentiment indicators which correctly stayed relatively bullish throughout the 1990’s. Beginning in 1999, short interest tailed off dramatically, finally becoming extremely low (compared to recent history) and bottoming out in August 2000 as the market formed its long-term peak. From that point, the short interest ratio rose steadily, finally peaking again in August 2002.

The fall recovery, and the rally through this Spring, has been accompanied by flat or declining short interest, at the same time average volume has continued its steady rise. At this point, the ratio between short interest and volume is at its lowest point since 1993. This is especially surprising considering the large number of convertible bond deals that have been brought to market over the past few months, as both May and June set new deal records, with a total of 101 deals closed according to ConvertBond.com. One of the largest customers for convertible bonds is hedge funds, who buy the bonds and then sell the underlying shares short as a hedge. The fact that short interest has declined while convertible issuance has increased so dramatically raises my antennae a bit, since it may indicate more public short covering than previously thought. This would be a bearish indication for the market going forward.

It’s important to keep in mind that there are still 7.8 billion shares short and outstanding on NYSE issues. There have been about 425 million shares covered (bought back) since the peak short interest in the summer of 2002. However, considering that total volume on the NYSE has been hovering around 1.7 billion shares daily, we could see an entire week of volume consisting solely of short-covering before it would be depleted. For those who are curious, these 7.8 billion shares amount to around 2.2% of total shares outstanding for NYSE issues.

Last week, I said that it would be telling if the market could not rally off the short-term oversold conditions that were present, as it had every other time since March. In fact, we didn’t recover right away, and that was the first sign that the mentality was switching to selling rallies instead of buying dips. Now, the market is facing the same test on a larger time frame. The 10-day average of the up issues ratio (i.e. advancing issues as a percentage of both advancing and declining issues, approximately equal to the advance/decline line) is now oversold, as is the 10-day average of the up volume ratio. These ratios have been this oversold three other times this year (see the charts on the site for details). In mid-March, it lead to the initial stages of this multi-month rally. In late June, it lead to a bounce over the next couple of weeks. However, in late January, it didn’t lead to much at all. The market was able to muster only a few days of bouncing around before resuming its slide. That type of activity is what is seen in strongly (or newly) trending markets, as oversold readings get pushed by the wayside and price continues lower. We saw the same thing in reverse as severely overbought readings were ignored as the market rallied throughout April and May. If the broader market cannot maintain current levels and/or begin another stage of the rally soon, then this will be yet another sign that the larger trend has changed and lower prices are much more likely. If the rally is going to resume, it should do so this week. If it continues to fail, I would look to become much more aggressively short on any subsequent bounces.

Our shortest-term indicators relieved their moderate oversold condition with today’s rally and most are now in some form of neutral. I don’t see much of an edge here for either direction in the short-term, but my preference remains with shorting strength for longer-term trades as long as we remain under 1015 on the S&P 500.

- Jason Goepfert

Disclosure: long OEX puts, long SPX puts