Leap Year Shenanigans: An Extra Day to Die
Why one should look to rates and not to counts
Other people can run with the explainers on why a Leap Day is needed and the Gregorian adjustment such that 2000 was a leap year but 1900 wasn’t…
…let me focus on my area: are leap years more deadly?
OF COURSE, THEY ARE
You’ve got an extra day in which to die….
Funny How Shorter Months Have Fewer Deaths… Kinda
Because I’m an actuary, I’m usually thinking in terms of rates, so I will try to normalize things.
Once upon a time, somebody remarked to me they noticed the drop in deaths from January to February ….and it popped back up in March.
Example from 2019:
Yeah, that’s not exceptional.
January has 31 days, February (usually) 28, and March 31.
There might be a reason that February usually has 10% fewer deaths than the surrounding months.
Calendar Musings
Seasonal Pattern to Deaths
There is a seasonality to deaths in the U.S. (and other non-tropical areas). Even when it’s not a bad flu/pneumonia year, more old people die in the winter and fewer in the summer, and most deaths are of old people.
So February may be shorter - but it’s still in winter.
Anyway, let’s check how things change from January to February to March:
I marked the leap years in green dots compared to the non-leap years. Interesting.
In general, the increase from February to March is less in leap years and more in non-leap years.
Leap years aren’t particularly special
Ain’t it just like an actuary to be a party pooper: leap years are just our mechanism to get the calendar year to get in synch with the earth’s orbit about the sun.
Yes, we stick in an extra day in the year, stick an extra day in a month… it seems the Egyptians may have been the first to get with the idea of making the accounting match up properly with a forced reset every so often by shoving in extra days, but that’s just accounting.
It’s just the label you put on the day.
It’s not that the rates in leap years are particularly higher.
Just that the number of deaths is going to be higher because there’s one more day in which to die.
This is why I’m looking at rates, not the number of deaths.
When we change to rates, we see that generally, as we go from January to February to March, death rates drop.
February to March is always negative in trend. As the weather warms up, the flu, pneumonia, and colds go away. Heart disease deaths also decrease.
January to February is not always negative, and if I look at it more closely, I’m going to guess it will be driven more by flu epidemics and other infectious diseases.
Leap days do not affect rates. Their effect will be on the ultimate count, by multiplying the rate by the number of days.
So does a true year consist of 4 orbits? 16 seasons and 1461 days? We would save on birthday celebrations and wedding anniversaries. Unfortunately living to 25 would be an achievement.