To the OP: the answer is, it depends. Australia's national statistical office, for example, does not use one.
If you're talking about some poverty statistic, you'll have to look at the methodology.
The most accurate picture would be reflected by looking at the private (market) income of a household, and adding any social transfers in cash (cash government benefits), as well as adding the value of any social transfers in kind (e.g rent-assisted housing, Medicare).
On the flipside, you would subtract direct taxes (e.g. income taxes) and the income paid by consumers indirectly (e.g. VATs or GST, taxes on production). The latter part is very involved but nevertheless an important aspect of understanding the totality of the tax and transfer system.
You should also add the value of housing services that flow to owner-occupiers from owning their own house instead of renting. $400 a week in income for someone who has to rent is a lot different than $400 a week income for someone who owns their own home without a mortgage.
