If you have parents and grandparents who lived (or are living) well into old age you can usually expect a long life yourself. However, the common assumption that this is a consequence of good genes is flawed, an enormous new study has found.

When the discovery of genetics explained why some diseases run in families it was an easy leap to attribute differences in lifespans to the passing on of important genes. However, a different sort of inheritance can be even more important. If your family members lived so long because they're rich enough to afford the best healthcare (or live somewhere that doesn't ration care by wealth) there's a high chance you share their good fortune.

Scientific efforts to quantify the genetic contribution to lifespan variations have generally come out with estimates of 15-30 percent. That's probably a lot lower than many non-scientists' guesses, but a new study in Genetics indicates it's still too high.

Ancestry.com provided anonymized data on the years and places of birth, years of death and familial relationships of more than 400 million people, to Dr Graham Ruby of Calico Life Sciences. Ruby looked at those born in the 19th and early 20th centuries to see how much lifespans correlated by different sorts of connections.

A tendency for siblings or parents and children to have similar longevity could be attributed to shared genetics, but (Targaryens aside) that shouldn't apply to spouses. Yet Ruby found married couple's lifespans were more closely linked than oppositely gendered siblings.

This might be accounted for by living in the same conditions. However, that does not explain Ruby's observation there is a correlation between people's lifespans and those of their spouses' siblings. Indeed, even cousins-in-law and more distant relatives by marriage proved somewhat predictive for an individual's lifespan, despite sharing neither homes nor genetics.

Ruby attributes this to assortative mating, where people marry those who resemble them in background and other characteristics. "What assortative mating means here is that the factors that are important for lifespan tend to be very similar between mates," Ruby said in a statement.

As Ruby noted drily; "Generally, people get married before either one of them has died," so people are usually unaware of their partner's life-expectancy, instead choosing for other reasons that happen to provide a longevity match.

Once Ruby controlled for factors that drive assortative mating, such as wealth, he found genetics accounts for a maximum of 7 percent of lifespan variation. He argues previous studies were ethnically or geographically limited, and these contributed to overestimates.

Lifespan is a good proxy for health in general, giving the work broader implications. The findings also tend to undermine narratives that attribute success or failure in life to genetic factors, suggesting environments are more important.