When Do Bull Markets Begin?

zeelotes

Newbie
Messages
2
Likes
0
Introduction

This study seeks to find when major bull markets have begun over the past forty years -- I went as far back as 1966 since that is when I have data for the NYSE and I wanted all the data to match. The thought behind this study is that if we know when major bull markets begin, we can invest a bit more aggressively from those dates.

Bull Markets Defined

For the purposes of this test, I look for periods when the underlying index has risen by greater than 20% without a drop of more than 20%. By using this approach I find the major peaks and troughs in the market. I also require that the rise had to endure for more than 40 consecutive weeks to be considered a bull market. Believe it or not, there are times that the market rises by greater than 20% just to go on to fall again almost immediately. A good example of this is the rise from 9/21/2001 on the S&P 500 -- it rose 21.40% in fifteen weeks only to go on to fall again in one of the biggest bear markets we've seen in the past century.

So, we have two requirements:

1.) The market must rise by more than 20%
2.) The rise must endure longer than 40 weeks (about 75% of a year)

Major Bull Markets in the NYSE

This first bit of data starts with the NYSE. There have only been eight times when the market has risen by more than 20% in this index. Of these eight, all but one fits within our definition of a bull market. The only exception is the rise from 9/21/2001 where it rose 20.89% over 26 weeks. The question now becomes: When did these bull markets begin?

NYSE_20.png


Six of the seven took place in year two of the Presidential Cycle -- four of the seven began from a low set in October. As we can see here, 86% of the major bull markets of the last forty years began in year two of the Presidential Cycle.

Major Bull Markets in the S&P 500

Since the S&P 500 is one of the most tracked indexes in the world, the next study is based on this index. Once again, in order to match the NYSE, I only go back to 1966. In this case there have been ten times when the market rose more than 20%. Two of those ten took place in 2001 -- one from 3/22/2001 and the other from 9/21/2001. In both cases, the market when on to fall within a very short period of time since the market was still fully within the clutches of the bear. In the first case it only took nine weeks before the peak was reached, but in the second, it lasted a bit longer at 26 weeks.

Once again, of the six times that fall within our definition of a bull market, there is only one occurrence that fell outside of year two of the presidential cycle -- and that was in 1987.

SP_20.png


Let me throw in the Dow for good measure:

Dow_20.png


Bull markets have historically begun in year two of the presidential year, and often, within October of that year.

But what happens if we lower the bar a bit from 20% to 15% -- in other words, we keep the second part of our definition the same requiring a 40 week rise, but we change the amount of rise required to 15%. Let's see what happens in this case.

Major Bull Markets in the NYSE Using a 15% Rise

Let me give you all the data first, and then I'll make some observations.

NYSE_15.png


Of the thirteen times when the market rose more than 15% on the NYSE, four are disqualified due to the brevity of their time in bull mode. Those are the ones I've listed at the bottom. The ones in the top part of the table fall within our definition of a bull market -- after revising the rise from 20% to 15%. Of the nine that qualify, only one falls outside of year two of the cycle -- 1987 once again. You'll note the dates -- 1966, 1970, 1974, 1978, 1982, ##, 1990, ##, 1998, and 2002. Of the ten in the cycle -- only two did not register as the beginning of a bull market. From 1966 to 2005 bull markets have begun in year two all but two of the ten times. Statistically significant? You decide.

Major Bull Markets in the S&P 500 Using a 15% Rise

This is the same as above but applied to the S&P 500. As you can imagine the results are similar, if not the same.

SP_15.png


Of the thirteen times when the S&P 500 has risen by more than 15%, four are disqualified due to the length of time that it rose. Once again, the other nine all fall within year two except for 1987. Half of the eight land in October of year two -- in fact, the one in 1998 that landed in August I'd consider an error anyway, since if I change to using the low price, rather than the close price, the trough lands on 10/8/1998.

When Do Bull Markets Begin?

At least for me -- the question is answered. How about you?

Note: This is my first post on this board. I'm not sure if I'm posting in the correct place, so feel free to correct me if I'm not.
 
Last edited:
Statistical Validity ?

zeelotes said:
Introduction

When Do Bull Markets Begin?

At least for me -- the question is answered. How about you?

Note: This is my first post on this board. I'm not sure if I'm posting in the correct place, so feel free to correct me if I'm not.

Hi zeelotes

Welcome to T2W. This is an interesting post and you have clearly spent some time on the statistical analysis.

I'm not a statistician but know a little about it. To me it seems that looking at just 8 events under your 20% rule and 13 events under your 15% rule does not constitute a high enough population to be statistically valid. The problem here, of course, is that (depending upon your definition of a bull market) that is all that you have available.

The sort of questons I would ask would be:

  • How many presidential cycles were there during the 40 years ?
    What other "causes" have you tried to take into account ?

It is a common error to link an "effect" with a "cause" just because the latter coexists or slightly pre-dates the former. However this relationship may be random or both the "effect" and "cause" may be "effects".

Having said that, though, one might consider reasons for such a link if it does indeed exist. The second year of a presidency should be a period of stability. The uncertainty of the election has gone and the market has by now evaluated the impact of the president and his/her policies and there is some time before the instability of the next election arises.

Charlton
 
Charlton:

Thanks for the welcome and your response to my research. Let me begin by answering your first question concerning how many Presidential Cycles have taken place over the past four years.

The Presidential Cycle follows the period of time that a President is in office -- four years -- so over the past forty years we have 40 / 4 = 10.

As to the cause / effect relationship: many have proposed answers to this, but first let me state unequivocal agreement with your two main thoughts -- the lack of data points, and also the importance of finding a cause / effect relationship before jumping to a conclusion about somethings validity.

Since so much has been written about this subject let me refer you to a few sites that have offered an answer:

Sy Harding
Gold Eagle

A quick search in Google will find you many, many more, including research from the Fed, academia and Wall Street. Those a whole lot smarter than I have found the theory useful and worthy of being exploited. My post seeks to validate the truth of their combined wisdom.
 
May we thus conclude that a cyclical bull market will probably commence in the U.S. equities market in the second half of 2006?
 
Top