Выдержки о наиболее распространенных рыночных мифах от JP Morgan из обзора Quant Forensics
January effect
In the case of the MSCI Europe index, we find a 73.3% probability for the month
of January to give the right indication. Even more impressive, the chance for the
index to end the year up when January was up is close to 87.5%.Very strong results
for such a simple test.
The Month of the Year Effect
The coefficients are never statistically significant, despite what the chart
above seems to suggest. Only the months of April, September and
December are close to being significantly different from the rest of the year.
October: worst month of the year?
We have tested several indices from various countries and regions for the month of
October and found similar results with gains being more or less important but still
relatively significant.
The strategy above clearly shows that the worst month of the year is September and
not October.
The Day of the Week Effect
In the UK, these week-day effects (Friday and Monday) seem to be highly significant
and highly statistically different from returns on the rest of the week. Fridays
particularly exhibit the highest returns and Mondays the lowest.
On a more global basis, the MSCI Europe index shows that when incorporating a
larger number of countries, these effects tend to cancel each other out and globally
the estimates of our excess returns are never statistically significant.
The Turn of the Month Effect
For the MSCI Europe as well as for the FTSE index, conclusion is the same: returns
and statistical numbers generated by the “Turn of the Month” effect are highly
significant9.
“Sell in May and Go Away”
For the past 50 years at least, the stock
market has performed immensely better between early November and early May of
the following year, than it has during the rest of the year.Above all, it is interesting to see the hit rate (percentage of positive returns) of the ”Sell in May” strategy reaching more than 72%, against only 52% for the buy-andhold
portfolio.
In the case of the MSCI Europe index, we find a 73.3% probability for the month
of January to give the right indication. Even more impressive, the chance for the
index to end the year up when January was up is close to 87.5%.Very strong results
for such a simple test.
The Month of the Year Effect
The coefficients are never statistically significant, despite what the chart
above seems to suggest. Only the months of April, September and
December are close to being significantly different from the rest of the year.
October: worst month of the year?
We have tested several indices from various countries and regions for the month of
October and found similar results with gains being more or less important but still
relatively significant.
The strategy above clearly shows that the worst month of the year is September and
not October.
The Day of the Week Effect
In the UK, these week-day effects (Friday and Monday) seem to be highly significant
and highly statistically different from returns on the rest of the week. Fridays
particularly exhibit the highest returns and Mondays the lowest.
On a more global basis, the MSCI Europe index shows that when incorporating a
larger number of countries, these effects tend to cancel each other out and globally
the estimates of our excess returns are never statistically significant.
The Turn of the Month Effect
For the MSCI Europe as well as for the FTSE index, conclusion is the same: returns
and statistical numbers generated by the “Turn of the Month” effect are highly
significant9.
“Sell in May and Go Away”
For the past 50 years at least, the stock
market has performed immensely better between early November and early May of
the following year, than it has during the rest of the year.Above all, it is interesting to see the hit rate (percentage of positive returns) of the ”Sell in May” strategy reaching more than 72%, against only 52% for the buy-andhold
portfolio.