Elliot Wave Update
By Larry Katz

In our early July report we had argued that, although higher highs were likely, the market should experience some downside pressure into late August, as a number of short term cycle lows were due. This did come to pass, although the weakness was a lot more pronounced in the DJIA than the S&P 500. Our preferred count for the past several months has the late January low (January 23 for the S&P 500, and January 30 for the DJIA) as the orthodox low of a large fourth wave triangle from the early 1994 high (see chart #3). At this point we see no reason to alter that count. In early July we had given two possible alternatives in regards to the post May 19 rally. Our preferred count at that time labeled the advance as wave 5 of iii from the January low, while the alternate count argued that this rally was only wave .3 of 3 from the mid march low (see chart #1). At this time both counts are still quite valid. However, the continued strength into mid July has changed that outlook in favor of third of a third count for a number of reasons. First off, the rally from the May 19 low into the peak of July 17 traveled an almost perfect 1.618 times the length of the April 19 to May 15 advance on the S&P 500, while the DJIA was less than thirty points away from a perfect hit, both on a print and hourly basis. This is a normal relationship for third waves, and alone is a strong argument in favor of this count. We had, in early July, argued that the post may 19 rally did have a number of the characteristics associated with fifth wave advances, and up until that point that was correct. However, the continued strong showing by both breadth and volume, as well as the fact that the daily new highs reached peak readings during this rally, are also strong arguments in favor of the third of a third count. The post May 19 rally can still be counted as wave 5 of iii, however, it has to be counted as an extended fifth wave. One of ElliottŐs rules is that fifth wave extensions are always doubly retraced. The rally from the July 19 low to the August 2 peak satisfied part of that double retracement, and can be counted as either a "b" wave or an X wave. However, the decline from the July 28 ( orthodox) peak did not violate the low of July 17, the presumed "a" wave of the pattern. In the case of the S&P 500, it did not even come close. This decline can be counted as a large "c" wave failure, but the depth of the failure is a bit unrealistic. The best count allows that the decline from July 28 into the late August low was a small fourth wave. This count also argues that the rally from July 19 to July 28 was wave .5 of 3, completing a five wave sequence from the April 19th low and also the middle three of the entire advance from the late January low (see chart #1). Alternate counts allows that the post May 19 rally was wave 5 of iii (see alternate count chart #1), and that the averages are either still in wave iv or are in the early stages of wave v. However, the weight of the evidence argues in favor of our preferred count and, for now, that is how we are approaching the wave pattern.

Our longer term count has the orthodox low of Super Cycle wave (IV) completing a near twenty one year triangle pattern from the 1928 (29) high into the lows of 1949. Nothing has changed in regard to that count (see chart # 2). Earlier in this report we pointed out that a nearly one year triangle pattern was complete at the late January low. This triangle is best counted as wave 4 from the 1987 low, putting the averages currently in wave 5 from the 1987 low (see chart #3). Although the post 1987 advance can be counted as an extended fifth wave from the 1974 low (see alternate count chart #2), Fibonacci time and price relationships argue that it is a third wave up from the 1982 lows. First off, the extent of the gains from 1987 has exceeded even the most optimistic relationships for extended fifth waves, as the post 1987 rally has traveled more than the net travel from the 1974 low to the 1987 peak. Normally, fifth wave extensions travel about equal to the net travel of waves one through three. Time relationships are OK, as the 1987 rally has traveled about .618 the time of waves one through three. However, the post 1987 advance has traveled 1.618 wave (1) (1982-1987), while also generating a number of characteristics of third waves, such as strong breadth and volume. The weekly A/D line just hit fresh all time highs, this is not what one would expect to see near the peak of a large fifth wave. Yes, the daily A/D line is still well below its 1994 peak, however, at the end of the last large fifth wave (1966) the A/D line had been diverging for what turned out to be a multiple of years. We would expect to see the same sort of pattern at the peak of the fifth wave from 1974 as well. Fibonacci relationships point to potential targets for wave 5 in the 4895-5105 level. This is the area where wave 5 (1995) would be equal in length to wave 1. This is also the area where wave (3) ( 1987-1995?) will have traveled near 1.618 the length of wave (1) (1982-1987).

Short term it appears that the averages are in the early stages of wave 5 of iii from the late January low. The august 29, low in both the DJIA and the S&P 500 appears to have completed wave 4 of that pattern. In the case of the S&P 500, the August decline took the form of a small triangle pattern (see chart #4), while the DJIA is best counted as tracing out a double or triple three. However, the pattern is a lot clearer on the S&P 500, as it has moved above its late July peak and is approaching initial targets for the post triangular thrust. On the other hand, the DJIA continues to lag, and although it is most likely in wave 5 of iii, it has not yet moved above initial resistance, which is the .618 retracement of the July 18 to August 29 decline. That level is 4685+/-9, and until it is broken decisively on the upside, there is still a chance that the DJIA is still in wave 4. A move above that level should allow for a move to 4751+/-18, where wave 5 will be .618 wave 1, or 4862 +/-29, where 5 will equal 1. Initial targets for the S&P 500 are at 568.29+/-1.26, that is the area where wave 5 will travel approximately the length of the widest part of the triangle, a normal relationship. However, additional targets for the S&P 500 are 575.69+/-1.98, where wave 5 will be .618 wave 1, and 585.37+/-3.23, where wave 5 will equal wave 1. Both of these targets are measured from the orthodox low of wave 4.

The weakness that was expected into the late July mid-August time frame did come to pass, as a number of our trading cycles were scheduled to bottom at that time. With the exception of the twenty week cycle, which has peaked and is not due to bottom until the first week of October, all of our short term cycles are now moving higher. This should have some positive influence on prices until at least mid September. Some weakness into the first week of October is expected as both the ten and twenty week cycles bottom, but following the resolution of those lows the market should again move higher.

In our early July report we were not sure whether the bonds had peaked in early June, or if one more push above the June 2 high was needed to complete the pattern. It turns out that the bonds did make one more attempt at the highs, which ended in a small failure as they equaled, but did not surpass, the June 2 peak. The decline from those highs into the lows of mid August can be counted as a five wave sequence, although the fourth wave of that pattern needs to be counted as an irregular correction. However, the July/August decline stopped right at an important support level given in early July. This level corresponded to a .618 retracement of the advance from late April to early July. The fact that the decline stopped right at some important Fibonacci support, as well as the fact that the July/August decline may be counted as either a five or seven wave pattern, leaves open the possibility for a move above the July 2 peak. This could set up for a test of the October 1993 peak in the bond market, and challenge the five wave count from October 1993 to November 1994. However, the October 1993 to November 1994 decline is still best counted as a five wave pattern (see chart #5), arguing that this rally, although it may extend a bit more, should not move above the October 1993 peak. Initial resistance in the 113 2/32 +/- 15/32, the .618 retracement of the July/August decline has been slightly exceeded. However, a move above the 116 level, the peak of July 7, will be needed to confirm that the rally from last November is still intact. The pattern from the November 1994 lows in the bonds is best counted as corrective, as each leg up has unfolded in what is best counted as a series of three wave patterns. This lends further support to the fact that 1), the 1993-1994 decline was a five, and 2) that the rally from last yearŐs low should not exceed the 1993 peak. Sentiment numbers have also deteriorated markedly, as both Market Vane and Consensus Inc. have shown a large jump in the bullish percentage. This has occurred in the shorter end of the maturity spectrum, as the numbers on T-Bill bulls has also jumped sharply. However, this jump to the bullish camp is a bit more pronounced in the Market Vane numbers, although the Consensus Inc. survey is showing its fair share as well. Put to call ratio have also dropped sharply, indicating a willingness to bet on higher prices. In fact, these ratios have reached their lowest, and most bearish, levels in several months. This is just one more notch in the bearish case. One note of importance: All our numbers on the bonds are based on a constant bond chart, and not on any one single contract. Therefore, these calculations should be used as guidelines only, and not applied to any one futures contract.

In sum, our preferred count suggests that the averages are close to completing wave iii from the late January low. Targets for wave 5 of iii are 4751+/-18 and 4862+/-29 for the DJIA, and 575.69+/- 1.98 and 585.27+/-3.23 for the S&P 500. Cycles, with the exception of the ten and twenty week cycles, remain favorable and support further near term progress. Some weakness from mid September into the early part of October is possible, as both the ten and twenty week cycles are due to bottom. Following those lows the rally should continue. However, a number of other important cycles should be near peak levels at that time indicating a slowing in the advance. Alternate counts allow that the peak in mid July completed wave iii, and that the averages are now in the early stages of wave v to the upside. However, Fibonacci relationships support the preferred count, as does the wave structure. Longer term Fibonacci relationships reveal potential targets for the completion of the post triangular thrust, as well as the entire five wave affair from the 1987 low, in the 4895-5105 level. Important support can be found at the 4552-4570 level for the DJIA, and 553-555 for the S&P 500. These levels are the fourth wave of previous degree, and are now considered important support zones intermediate term. We will keep you abreast of any changes in the daily reports.