Jul 14, 2003

A lot of talk in financial circles since yesterday has centered around the large decline in the long-term bond market over the past month. It usually pays for stock investors to be aware of bond market movements, as bonds and stocks have been strongly (inversely) correlated for the past few years, as the chart below illustrates.

However, we can see clearly that the stock rally beginning in March of this year was NOT accompanied by a similar drop in bonds, as virtually every other rally during the bear market was. The table below shows the relationship between 30-year bonds and the S&P 500 for several different time periods:

In the table, “correlation” shows the statistical relationship between movements in the two markets – a correlation of 0 would mean they do not move together at all, while -1 would mean they are perfectly inversely correlated (one goes up while the other goes down), and +1 would mean they move in lockstep (they both go up or both go down). The chances that these correlations exist due to chance are virtually zero, so it appears as though something about this intermarket relationship has broken down, particularly since the equity market low in March. Since we can somewhat safely assume now that what is good for bonds is good for stocks going forward (though the relationship is quite a bit weaker than the inverse movements seen previously), and considering the huge drop in bonds yesterday, it’s time to take a look at how the sentiment picture in bonds is shaping up.

Let’s first take a look at the Commitments of Traders positions for 30-year bond futures. The chart below is just a simple net difference between large commercial traders (the “smart” money) and small speculators (the “dumb” money). As the indicator (the tan area) moves higher, that shows that commercial traders are becoming more net long and/or small specs are becoming more net short. Both of those would tend to be positive for appreciation in bond price, as large commercial traders tend to be right on major directions in interest rates while small traders tend to be wrong.

We can see from the chart how watching this information tipped us off to the last big rally beginning this Spring (as the indicator went positive), then became more negative as bonds rallied. While commercials never became as short as they had at previous bond market peaks (and small specs didn’t get as long), we still had a decent heads-up in mid-June that perhaps it was time to expect an increase in rates, or at least a slowing of the decline as bonds began to look toppy. Currently, this “variance” between the two groups is neutral, as it has oscillated back and forth around the zero line for the past couple of weeks.

Next, let’s take a look at an indicator more commonly applied to the equity markets – the put/call ratio. Here we are looking at the volume of puts traded on 30-year bond futures compared to the volume of calls traded. As the ratio heads higher, more and more volume is flowing into puts relative to calls – this is normally interpreted as a sign of pessimism or fear among bond traders, as they scramble to own protection for current bond positions, or speculate on a decline in price (and rise in yields). For a variety of reasons, mostly that bond investors tend to be more sophisticated than their equity-trading cousins, put/call ratios on bond futures tend to be one of the least consistent sentiment indicators available. Still, they can be worth watching.

Here, we can see that in mid-May, the put/call ratio became very low as traders rushed to speculate on (or hedge against) rising bond prices by picking up calls at a much greater pace than they did puts. Not surprisingly, this coincided with a short-term peak in the bond market, which was then exceeded a couple of weeks later before this most recent failure. Currently, the put/call ratio is neutral as it has oscillated between very low and very high on a daily basis during the past two weeks.

Next is investor sentiment as gauged by Market Vane, Inc.. Market Vane polls a group of CTA’s (Commodity Trading Advisors) and creates an index based upon their recommended outlook for a variety of markets. Similar to other sentiment measures, this is best interpreted in a contrarian manner – when the advisors are too optimistic, it often leads to a decline in price; when they are pessimistic to an extreme, then it is usually reasonable to expect a trough in bond price sometime soon.

As of the most recent data available, these advisors were still relatively constructive on bonds. At just over 70% bulls, the advisors were about as pessimistic as they were at the other short-term bond lows over the past six months or so. However, they are nowhere near the extreme levels of 40% seen at the major lows in 2001 and 2002. While they are off the peak reading of 87% in mid-May, they still appear to be quite bullish.

Lastly, I want to show two charts derived from the Rydex funds. These charts show the relative amount of asset flow going into the bullish bond fund, then the asset flow into the bearish bond fund.

Here, we can see that the asset flow into the fund that bets on rising bond prices is quite depressed, and is on a par with what was seen around other lows in bond price over the past year. Now let’s look at the bearish bond fund:

Contrary to what we saw in the long bond fund, the short bond fund asset flow is very high and is approaching what was seen at the major low in October 2002. This tells us that Rydex traders have had a very clear preference for betting on declining bond prices over the past couple of weeks. This is fine, as they have been right on the mark, but now that they are entering an extreme, it appears as though they are pressing their bets. This type of aggression is usually not rewarded.

If we take a step back and look at the various market participants, the following themes emerge:

Overall, I don’t see much from a sentiment perspective that would have me excited about trying to pick an intermediate-term bottom in bonds right now. Sentiment is only one piece of a very big, very complex puzzle (especially in the bond market, and ESPECIALLY after a move like yesterday), but it can give us an excellent heads-up when a confluence of extremes across different investor groups is reached. As of today, we’re beginning to see some of the signs that are often seen near longer-term low points, but not many.

So what does this mean for stocks? If we assume that the recent positive relationship between stocks and bonds will continue for the foreseeable future, then the outlook is not very good for equities. The reason is because there is not much evidence that bonds are about to form a bottom here, which obviously means that their most likely direction is a continued decline. Since bonds and stocks have shown a relatively strong positive correlation, a continued decline in bonds would suggest a decline in equities by association. As I said above, this relationship is weaker than the negative relationship that had been established, and it has only been in place for a short time, so stocks don’t HAVE to decline if bonds do. But that’s the way the signs are pointing.

Federal Reserve Chairman Alan Greenspan recently testified before Congress in his semi-annual Humphrey-Hawkins address. Love him or hate him, it is impossible to argue that his words do not affect the financial markets. A subscriber asked me to look at how the stock market performed around these speeches, and the results are quite negative. I could only find these speeches going back to 1996, so we have a small sample size of 14 occasions. On the day of the testimony, the S&P 500 closed in positive territory only 5 times (not including yesterday). The day after the speech was a bit worse, as the S&P closed higher only 4 times, with an average loss of 0.50%. Over the short-term, the S&P often had great difficulty making much headway. Three days after the speech, the S&P showed an average loss of 1.0%, with 9 of the 14 occurrences being negative. Within those three days, the high of the day of the speech was exceeded 43% of the time (by an average of 0.1% and a maximum of 2.7%), but the low of the day of the speech was exceeded 79% of the time (by an average of 1.6% and a maximum of 6.1%). After five days, the average return was negative 1.8%, with 10 of the occurrences being negative. Within those 5 days, the high of the day of the speech was exceeded 50% of the time (by an average of 0.3% and a maximum of 3.8%) while the low of the day of the speech was again exceeded 79% of the time (by an average of 2.7% and a maximum of 11.3%). These results are quite skewed to the negative side.

Once we get past the first five days after the speech, the results become more in line with any other random period. After 30 and 60 days, overall the market was slightly more negative than it was during other random periods, but after 90 days there was no difference between post-speech periods and any other period. Part of reason for these longer-term negative results may simply be the time of year these speeches are given. As I said, they are semi-annual, with most of them occurring in mid-February and again in mid-July. As we all know, the middle summer months are particularly weak for the equities markets, so some of the negative bias could be attributed to that. Below, I’ve posted charts going back to 1996 showing the dates these speeches were given (the blue vertical lines).

This week ends with option expiration, and there have been some distinct patterns over the past year surrounding this event. The following table outlines how the S&P 500 reacted around each of them:

Month	Day Before	Day Of	Day After	Day +3	Day + 5
Jul ‘02	(2.7%)	(4.1%)	(3.6%)	(0.7%)	0.9%
Aug ‘02	1.1%	(0.3%)	2.6%	2.5%	1.3%
Sep ‘02	(3.0%)	0.2%	(1.5%)	(0.5%)	(2.5%)
Oct ‘02	2.3%	0.6%	1.6%	1.3%	1.5%
Nov ‘02	2.3%	0.6%	(1.0%)	0.7%	2.4%
Dec ‘02	(0.8%)	1.4%	0.1%	(1.1%)	(2.4%)
Jan ‘03	(0.5%)	(1.6%)	(1.5%)	(1.5%)	(6.1%)
Feb ‘03	(1.1%)	1.3%	(2.0%)	(2.6%)	(1.0%)
Mar ‘03	0.2%	2.4%	(3.7%)	(3.1%)	(4.0%)
Apr ‘03	(1.2%)	1.6%	(0.1%)	3.0%	0.5%
May ‘03	0.8%	(0.4%)	(2.5%)	(2.2%)	(1.3%)
Jun ‘03	(1.5%)	0.1%	(1.3%)	(2.1%)	(2.1%)

Avg Return	(0.3%)	0.1%	(1.1%)	(0.5%)	(1.1%)
% Positive	42%	67%	25%	33%	42%

NOTE: Day + 3 shows the percentage change between expiration and the day three trading days after. Day + 5 shows the percentage change between expiration and the day five trading days after.

We can see from the table that the day before expiration (tomorrow in our current situation) has not had much of a bias, although it has made some wide swings. The day of expiration, Friday, has been positive 8 of the 12 months, while the following Monday has been positive only 3 times. The last 6 have been negative, and the market has quickly recovered only once, in April of this year. One other tidbit is that expiration Friday has often had a quite narrow range, while the following Monday’s range has averaged significantly greater. Though expiration is often labeled as “wild”, it is in fact usually quite tame. The days that follow, however, are indeed quite volatile.

For the sixth time since the low in March, our shortest-term sentiment model, STEM.MR, is in oversold territory. Each of the other occurrences resulted in some sort of a bounce – at least 10 S&P points, but usually much greater. As I’ve said the last few times, if the market cannot rally off of this reading, then that is telling us that something fundamental may have changed about how the market is reacting. The market passed each of the other tests, but now we again need to see if instead of traders buying dips, the mentality may be switching to selling rallies. If that is the case, then these short-term oversold readings won’t have much of an impact and we will continue lower. That will be an early sign that the larger trend may be changing. My stance remains the same – if we are under 1015 on the S&P, then I prefer to be on the short side of the market for anything longer than day-trades.

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.

Over the past week, assets in the two Rydex funds that are pegged to the Nasdaq 100 on the long side (i.e. the funds rise as the NDX does) have hit new 52-week highs. While this seems like it should be a bearish sign of trader over-confidence, it’s not terribly surprising considering the performance the Nasdaq 100 has enjoyed. Meanwhile, in the S&P 500, even though the index itself is very near a new yearly high, assets in the funds that are supposed to do well when the index does are not even close to setting new yearly records. Also, assets in the funds that bet on the short side are not at 52-week lows. If the fund flows followed their underlying indexes directly, then we should expect to see the long-side fund assets hit new highs as the S&P 500 does, and the short-side fund assets hit new lows. The fact that we do not shows us that there is an underlying psychology at work here (besides just the usual ebb and flow of fund assets).

Sometimes we will see the long-side funds hit new record highs even though the S&P rallied a small amount – this shows us that the Rydex traders are quite confident of further upside, and not surprisingly it often marks at least a short-term high in the market since these traders are wrong more than they are right at the extremes. I’ve been experimenting with different ways to express this relationship between fund flows and the underlying indexes, and the result is the Rydex Enthusiasm Index. This is a fairly straightforward calculation that takes into account how much assets in a particular fund rise or fall compared to the NAV of the fund itself. So, for example, if the NAV of a long-side fund rises 2% on a given day, but the assets in that fund rise 10%, then we can assume that traders are more confident of the upside than they “should” be (I know, this is a pretty big assumption). We can be a bit more confident of that assumption if assets in the mirror-image short-side fund, which should have dropped somewhere around 2%, actually drop 15%. When we spread this across all 8 index funds, and look over a period of 5 days, the underlying psychology of the Rydex traders should become clear, and is less distorted by idiosyncrasies in any one fund on any one day. Are they bubbling over with excitement, and pouring more assets into certain funds than they rightfully should be? Or are they fighting the market, and pulling assets out of funds that are actually rising? This index gives a glimpse into the mindset of the traders by comparing asset flows to index performance, which our other Rydex indicators do not do.

I have placed red arrows at those times when the Rydex timers got ahead of themselves by throwing more money at long-side funds (and withdrawing money from short-side funds) than they should have based on the performance of the market itself. Green arrows show the opposite – when traders became pessimistic about the market’s prospects and decided to bet heavily against it, even though it may not have been justified based on how the market itself was performing. The arrows are not based on any mechanical methodology – I put them there based on when the indicator reached an extreme and then reversed.

Like any indicator, this is not perfect. As you can plainly see, several false signals were given over the past year. Fund flows can be influenced by a single large trader, or seasonal tendencies that are not yet apparent, or a hundred other reasons – it is not an idyllic substitute for trader opinion. Also, these figures are averaged over a period of 5 days. Perhaps 3 days or 10 days would give a better picture, but I am NOT a fan of optimizing indicators for the sake of making them look pretty. There is no guarantee that what worked well in the past will work well in the future.

The point of this exercise was to see whether traders have been buying the rally over the past few days with the same gusto they bought the previous ones, based on how the market itself has performed. The answer, it appears, is no – at least so far (today’s numbers have not yet been released). Currently, the Enthusiasm Index level is at 0.10, less than half what was seen at previous momentum peaks over the past year. If this rally peters out here, it would be the least-hyped peak since November 2002. While the rally could stop here, and indeed there are a multitude of indicators suggesting that it should, it will be unusual in that Rydex traders will be rewarded for their caution.

Back in early June, I presented the results of a study I did regarding low levels of bearishness among the respondents to Chartcraft’s Investor’s Intelligence survey (check the June 12^th daily comment in the archives on the site for reference), and I’ve been asked to give an update. At the time, I showed that the average number of trading days between the initial low bearish reading and a top before at least a 5% correction set in was 49 days, with a rise in the S&P of an average of 9% in that time. Technically, we’ve already satisfied that 5% drop with the decline into late June, but just barely (it declined 5.2% high to low). Since the S&P set a new intraday high today by a tiny margin, let’s pretend for a moment that the S&P didn’t quite satisfy the 5% drop criteria. In that case, we’ve now gone 25 trading days since the initial low bearish reading was given, almost exactly ½ of the average of the other readings throughout the history of the survey, and longer than only 3 out of the other 13 occurrences. We’ve rallied about 3% since the first low reading was given, which is one of the smallest rises seen, and well below the average of 9%. It was quite rare to see the market decline immediately after such a display of lopsided bullishness, and we’re seeing that pattern continue this time around.

In the weekly reports that came out this weekend, there was not much of a change. The sentiment surveys stayed relatively stable (the AIM model rose a small 0.8%), and the Commitments of Traders data showed both large commercial traders and small speculators not changing their positions by any large degree. The one report worth noting is the NYSE Members report, which showed another decline in specialist shorting activity and a rise in public shorting (as a percentage of total shorting activity). Currently, the specialist short ratio, which is simply total specialist shorts as a percentage of the total, is at 30%. This is one of the lowest ratios since 1996, being eclipsed only by the 20% seen in the aftermath of 9/11 and the odd 29% anomaly seen in early May of this year. Public shorting as a percentage of the total is currently registering 50%, which is the highest since mid-April of this year.

The 20% spread between the two groups of traders (i.e. 50% public shorting minus 30% specialist shorting) is quite rare. After the low in April 1997, when this spread reached 23%, it did not approach 20% again until 9/11, and it has only occurred a few times since then. Other than a very premature signal in June of last year, each of the others has been a relatively good (though often fairly early) buy signal. The odd thing about the current reading is that it comes after a large rally, whereas the others occurred after a sustained – and often precipitous – decline.

I try to not guess the reasons behind these types of readings, but if I had to venture an opinion, I think that the large amount of convertible bond issuance seen last quarter may have something to do with the strange readings we’ve seen here. Many hedge funds use a “convertible arbitrage” strategy whereby they buy convertible bonds and sell the underlying stock short. As convertible bond issuance rises, there is more need to sell short stock to hedge their position. I may be mistaken on this, but I believe the very large GM convertible bond issuance took place during this reporting week, so that may have had something of an effect. The reason I mention this at all is that if the increased convertible issuance has had an impact on these readings, then this may not be as bullish as I have been assuming. If the market rises, these funds will probably not remove their hedge by buying back their shares, as that would leave them with unintended risk. If these were just regular speculative funds betting on a market decline, then that would give us a more predictable short-covering base. In any event, my thoughts here are little more than an educated guess, so unless I learn something more substantial regarding this, I don’t want to write off these readings. Historically, including recently, a low specialist short ratio – and a high public short ratio – have been bullish for the market. We have that situation now (or at least as of two weeks ago, the most recent data available), so the default assumption should be that it is a bullish development going forward.

Late last week, I said I thought it was likely we would see a bounce, but accompanied by lessening momentum and therefore unlikely to carry far. It has already carried all the way to the mid-June highs in the S&P, but evidence of lessening momentum is debatable. I’ve been giving short positions the benefit of the doubt unless we exceed those mid-June highs and held there, which we have not yet done. Unless that happens, I continue to prefer to concentrate on the short side for longer-term trades (out several weeks at least).

Time Frame	Correlation	Relationship
01/01/01 – 12/31/02	-0.76	Strongly Negative
01/01/03 – 03/11/03	-0.86	Strongly Negative
03/12/03 – 07/15/03	+0.66	Positive