The first day of holidays ... all my children want to do is sleep.
I can understand why. The American Centres for Disease Control says schoolchildren aged up to 13 need nine to 11 hours per night. Teenagers need eight to 10 hours.
What with getting to sleep after 10, and getting up early enough to fly out to school, they don't stand a chance.
Neither do as many as one third of Australian teenagers, according to the most authoritative survey by researchers at the University of South Australia. Many of the teenagers who don't get enough sleep during the week "yo-yo", undersleeping Sunday to Thursday and catching up on Friday and Saturday.
So common is the yo-yo that the researchers reckon the average 17-year-old sleeps just eight hours 46 minutes on Mondays and 10 hours 10 minutes on Saturdays, an extraordinary difference of 84 minutes.
Getting children to sleep earlier is all but impossible. Australian 12-year-olds typically stay up beyond 9pm on school nights, and 15-year-olds beyond 10pm. At 17, boys manage 10.43pm, according to the Australian survey. Girls are little better at 10.35pm. Changing circadian rhythms during puberty means they are simply not tired. But they are forced to get up to leave for school between 7am and 8.30am as if they've had a full night's sleep.
American studies find teenagers who are chronically tired are less able to learn and retain material (especially in the morning), more likely to have car accidents, and more depressed, irritable, defiant and apathetic.
So the Americans are experimenting with asking them to go to school later.
You're probably thinking that makes them go to bed even later, but it doesn't. A review of the six major US studies finds that delaying the start of school by between 25 minutes and one hour boosts weeknight sleep by between 25 and 77 minutes. In other words, far from going to bed later, some of them go bed earlier.
"This provides evidence countering the hypothesis that students will simply stay awake later if school start time is delayed, and verifies the developmental shift in circadian timing that favours phase delay during later childhood," is how the researchers put it. The students surveyed thought better, felt better, and (among those old enough to drive) had fewer car accidents.
Naturally, their parents weren't keen. Getting kids off later is massively inconvenient. There are costs to parents that need to be weighed up against the benefits to children.
Which is where a major new study by the Rand Corporation comes in. It offsets those costs against two of the likely benefits; greater lifetime earnings for students, and fewer car crashes involving students (many of whom can hold a restricted licence aged 16 in the United States).
It examined the effect of delaying start times by just half an hour, from 8am to 8.30am, over a period of 20 years, making the reasonable point that the benefits would grow over time.
The net benefit, offsetting the costs to parents and transport systems against the likely benefits to students across the entire country: $US83 billion over a decade, jumping to $US140 billion over 15 years.
After just five years the benefit cost ratio reached 2 (or 1.7 under different assumptions) meaning the benefits were double the costs. After 20 years the ratio reached 3, a pay-off unheard of for most infrastructure projects, especially in Australia where studies struggle to find anything exceeding 1.
Rand thinks its assessment of benefits is terribly low. It excludes health effects, including reduced obesity, and it excludes the known effects of sleep deprivation on violent crime and substance abuse.
Rand only examined the effect of postponing school start times by half an hour. It's likely that if they were delayed an hour the benefit would be bigger still.
We'd have to adjust, but parents are used to adjusting. It'd be more than worth it.
Peter Martin is economics editor of The Age.