If you have read my past posts, you would have seen that the technical analysis I use in my investment process is based not on patterns but on the analysis of price/volume relationships. I do this in order to gauge the supply and demand of a particular investment at varying price points. Over the last few weeks, I have been working on putting an objective structure around that analysis (albeit imperfect) to use as a tool to better time entry and exit points, first on the market as a whole and then for particular investments.
What I developed is a diffusion index that measures the overall, near-term price/volume relationship in the stock market. I will not go into the details of how a diffusion index is calculated, but if you would like explore, more details can be found at the Conference Board website here. That said and for those who do not know, a diffusion index is a measure that calculates the proportion of components contributing positively to some measure or index. For instance, the Institute of Supply Management's Purchasing Managers' Index is a diffusion index. Diffusion indexes typically indicate neutral indication at the 50 level, are more positive above 50, and are more negative below 50. The same is true for the index I developed.
My diffusion index looks at price/volume characteristics of the overall market, here defined as the S&P 500. Stronger measures are generated when the the price and volume increase in unison, with a higher value attributed to greater volume increases. In contrast, weaker or negative measures are input into the calculation when price and volume move in opposite directions, again with greater absolute values attributed to greater volume levels. The later, in my estimation, indications an increases in demand while the former indicates an increase in supply relative to demand.
I provide three charts (since 1997) detailing the results followed by performance characteristics and general conclusions.
Diffusion Index
Rolling Sum of Indicator
Slope of Indicator
What I see in the above charts is the rolling sum of the indicator (which is the raw numerical value of the data prior to the diffusion index calculation) will tend to follow the market while both the slope of the indicator and the diffusion index tend to fall off and diverge from the price path of the S&P 500 before market tops. Other visual aspects of the charts also look interesting, but as you know looks can be deceiving. So I turned to the objective data, and the data results are just as interesting.
Since the early 1950's, I calculated the average 6-month return on any trading day to be about 4% and the batting average (or the percentage of instances the market's performance was positive) to be 67.5%. I also calculated the average 6-month return and batting average of the market following varying scenarios using the above three measures. Strong positive performance was logged when the diffusion index was greater than or equal to 60 and the indicator had increased over the last month. Stronger than normal performance was also logged following a diffusion index that fell below 30, and when the slope dropped below -0.10. I think these last two scenarios provided strong performance because they indicated a wash-out affect. Taking these three scenarios and averaging the return yields an average 6-month performance of 6.9% and a batting average of 76.7%.
As for the negative scenarios, I also looked at three different scenarios and averaged the stock market performance hence. The three scenarios in question are when the diffusion index is above 80 (indicating very strong demand that is unlikely to be sustained), a diffusion index below 40 that is declining in conjunction with a falling indicator measure, and when the slope of the indicator has been in decline. Following these three instance, the average 6-month performance was logged in at 1.4% and the batting average was just 55.6%.
I think the diffusion index and other measures are likely more useful as tools to gauge long-term turning points in the market and I am shy of using them as daily or weekly trading tools. That said, I do not think this diminishes their potential usefulness. It is my opinion these measures could be powerful tools, especially in conjunction with other measures and analyses. For instance, the weighted average standardized VIX measure or the long-term technical indicator. Other avenues for use could include timing long-term entry or exit points for individual stocks or funds. I plan on updating you on the diffusion index weekly or when interesting opportunities arise.
What I developed is a diffusion index that measures the overall, near-term price/volume relationship in the stock market. I will not go into the details of how a diffusion index is calculated, but if you would like explore, more details can be found at the Conference Board website here. That said and for those who do not know, a diffusion index is a measure that calculates the proportion of components contributing positively to some measure or index. For instance, the Institute of Supply Management's Purchasing Managers' Index is a diffusion index. Diffusion indexes typically indicate neutral indication at the 50 level, are more positive above 50, and are more negative below 50. The same is true for the index I developed.
My diffusion index looks at price/volume characteristics of the overall market, here defined as the S&P 500. Stronger measures are generated when the the price and volume increase in unison, with a higher value attributed to greater volume increases. In contrast, weaker or negative measures are input into the calculation when price and volume move in opposite directions, again with greater absolute values attributed to greater volume levels. The later, in my estimation, indications an increases in demand while the former indicates an increase in supply relative to demand.
I provide three charts (since 1997) detailing the results followed by performance characteristics and general conclusions.
Diffusion Index
Rolling Sum of Indicator
Slope of Indicator
What I see in the above charts is the rolling sum of the indicator (which is the raw numerical value of the data prior to the diffusion index calculation) will tend to follow the market while both the slope of the indicator and the diffusion index tend to fall off and diverge from the price path of the S&P 500 before market tops. Other visual aspects of the charts also look interesting, but as you know looks can be deceiving. So I turned to the objective data, and the data results are just as interesting.
Since the early 1950's, I calculated the average 6-month return on any trading day to be about 4% and the batting average (or the percentage of instances the market's performance was positive) to be 67.5%. I also calculated the average 6-month return and batting average of the market following varying scenarios using the above three measures. Strong positive performance was logged when the diffusion index was greater than or equal to 60 and the indicator had increased over the last month. Stronger than normal performance was also logged following a diffusion index that fell below 30, and when the slope dropped below -0.10. I think these last two scenarios provided strong performance because they indicated a wash-out affect. Taking these three scenarios and averaging the return yields an average 6-month performance of 6.9% and a batting average of 76.7%.
As for the negative scenarios, I also looked at three different scenarios and averaged the stock market performance hence. The three scenarios in question are when the diffusion index is above 80 (indicating very strong demand that is unlikely to be sustained), a diffusion index below 40 that is declining in conjunction with a falling indicator measure, and when the slope of the indicator has been in decline. Following these three instance, the average 6-month performance was logged in at 1.4% and the batting average was just 55.6%.
I think the diffusion index and other measures are likely more useful as tools to gauge long-term turning points in the market and I am shy of using them as daily or weekly trading tools. That said, I do not think this diminishes their potential usefulness. It is my opinion these measures could be powerful tools, especially in conjunction with other measures and analyses. For instance, the weighted average standardized VIX measure or the long-term technical indicator. Other avenues for use could include timing long-term entry or exit points for individual stocks or funds. I plan on updating you on the diffusion index weekly or when interesting opportunities arise.
Thanks, I will check it out.
ReplyDelete