Predicting what the stock market will do in the next 12 months is tantamount to predicting coin flips.
Even after what might feel like an already long bull market, it's a mistake to think that a lengthy duration is a sure-fire sign that stocks will soon fall.
"One common refrain we hear from skeptical investors is that bull markets typically last no more than five years, and thus the market is poised for a correction," Bank of America Merrill Lynch's Savita Subrmanian. "Our work shows that the length of bull markets has varied over time-from two years to nine years-and the dispersion of duration is quite high (a 2-year standard deviation)."
In other words, some bull markets will last a few years, and some will last many years.
What's particularly interesting is what Subramanian noticed about what five-year returns tell you about what'll happen next.
"We find no relationship between historical 5-year returns and subsequent 12-month returns," she wrote. Using some simple regression analysis, she found an R-square of 0.0002. That's about as low and R-square gets. (R-square is a statistical measure which reveals how well a regression line - the line of best fit you see - explains the relationship between two variables. The higher the R-square, the better that relationship is explained.)
Even after what might feel like an already long bull market, it's a mistake to think that a lengthy duration is a sure-fire sign that stocks will soon fall.
"One common refrain we hear from skeptical investors is that bull markets typically last no more than five years, and thus the market is poised for a correction," Bank of America Merrill Lynch's Savita Subrmanian. "Our work shows that the length of bull markets has varied over time-from two years to nine years-and the dispersion of duration is quite high (a 2-year standard deviation)."
In other words, some bull markets will last a few years, and some will last many years.
What's particularly interesting is what Subramanian noticed about what five-year returns tell you about what'll happen next.
"We find no relationship between historical 5-year returns and subsequent 12-month returns," she wrote. Using some simple regression analysis, she found an R-square of 0.0002. That's about as low and R-square gets. (R-square is a statistical measure which reveals how well a regression line - the line of best fit you see - explains the relationship between two variables. The higher the R-square, the better that relationship is explained.)
No comments:
Post a Comment