You can't read much at all into a comparison of life expectancies for different countries for one simple (and a whole lot of more complicated) reason - the biggest factor in the differences is not when the old people die, it's how many young people die.

There is this oft-repeated claim that we are living SO much longer than in years past. That in the middle ages the average life expectancy was only 22 and even by the beginning of this century in the United States it was less than 50. Yet history is replete with people living into their 80's and 90's - more than two to four times their life expectancy. That would be like someone living to about 300 years old today. What gives?

The answer is that life expectancy is the average age of a person in a population when they die. If a baby dies at age 1 and another person dies at age 79, the average life expectancy is 40. But does saying that the life expectancy is 40 really tell us anything? If so, what? It's sort of like putting your head in an oven and your feet in a bucket of ice water - on average you should be quite comfortable, right?

Historically, infant mortality was enormous - 30% was not uncommon at all and fewer than half survived to become adults. This gives a life expectancy for this half of the population of something like five years old. In order for the other half of the population to bring that average up to the ripe old age of 22 they have to live, on average, to the age of 40. You then also have to take into account wars and accidents that tend to take a disproportionate share of young adults and hence has a significant impact on the average life expectancy figure. You therefore have to have a lot of people living into their 60's, 70's and 80's to keep the overall average even up to 22.

A much better indicator is to look at the expected years remaining to a person after they have survived these high impact years. Unfortunately, these can be difficult to track down. For instance, in 1900 the average life expectancy for males in the U.S. was only 46.4 years old. But half of the people that were 65 at that time (and hence born well before the Civil War) lived to be 76.2. In 2000, the average life expectancy for males is 73.9 (a pretty staggering 59% increase) but the average expectancy for someone that is 65 today is only 80 - a quite modest increase of less than 5% in the age at which they will most likely die.

So what about Japan's four additional years? Using the data from:

http://www.infoplease.com/ipa/A0004393.html

They have an average life expectancy of 80.9 years and an infant mortality rate of 3.8 per 1000 live births. The U.S. has an average life expectancy of 77.4 years but an infant mortality rate of 6.7 per 1000 live births. So there is part of the difference right there. Then there is the simple fact that the U.S. has higher crime rates than Japan - with many of the resulting deaths being young people - and that the life expectancy figures for the U.S. still reflect losses - again of young people - in Vietnam and even Korea. World War II effects are largely fading out from the statistics though you can certainly see their role if you look at the increases in Japan's life expectancy over the last couple of decades.