This question falls within the 20 year rule, because by 1996, then US Ambassador to the UN Madeleine Albright was questioned on 60 Minutes regarding the veracity of the 500,000 figure and whether it was "worth it".
Albright, and the US government at the time, claimed the figure was inaccurate and based on flawed methodology. I am curious if any of our illustrious contributors could elaborate on this claim, and whether any new information has come out since then as to the truth of that figure.
I believe that the charge was half a million Iraqi children, not just Iraqis. And while I have no access to any figures that might justify the claim, it is certainly the case that three people charged by the UN with administering the sanctions regime - Dennis Halliday, and then his successor Hans von Sponeck along with another UN official, Jutta Burghardt - resigned, stating that in their opinion the sanctions regime was being manipulated by the UK/US governments in a manner that they believed was "genocidal" in its effect.
Source: Hans von Sponeck "A Different Kind of War: The UN Sanctions Regime in Iraq"
The estimate came from an August 1995 FAO survey: the figure of 567,000 deaths came from two of the researchers (S Zaidi & MC Fawzi, Health of Baghdad's children, The Lancet 346, 2 Dec 1995, p 1485). The FAO findings are here: while the organisation didn't put a number on total excess mortality, it found a doubling of infant deaths and a 388% increase in child deaths (ie at ages 1-4) in Baghdad after 1990, as well as evidence of severe ongoing malnutrition.
Without the Lancet letter I haven't managed to reconstruct how its authors arrived at their figure: assuming 750,000 annual births and pre-war infant mortality of 5% and child mortality of 1.3%, and speculating that these might have fallen to 4.1% and 0.9% without the war and sanctions, I get about 400,000 excess deaths using the FAO multipliers. It could be that the authors assumed higher fertility and/or pre-1990 mortality, or a faster "normal" downward trend.
One of the authors of the December piece later (link below) reported finding far lower mortality in a follow-up survey in April 1996. But others also found high mortality: Ali et al suggest around 50,000 annual excess deaths among under-5's in 1991-95, half the earlier estimate.
If there was an error in the 1995 findings, it might be partly down to their basis in results for Baghdad, which may well have suffered a sharper proportional fall in well-being and rise in mortality than other areas less favoured in the 1980s. Later studies were also suspected of suffering from an unintended urban bias that may have overstated deaths.
The 500,000 figure was first mentioned in a 1995 Lancet paper by Zaidi and Fawdi, and was later cemented following a 1999 UNICEF child and maternal mortality survey. Both were based on surveys (conducted in cooperation with local authorities) and tried to estimate the increase in child mortality. The deaths were typically attributed to malnutrition, lack of medicine, and lack of clean water: all these were certainly among the effects of the sanctions, with the first one partly ameliorated by the Oil For Food program.
There are however two key questions about the number. First the mortality rate estimates are heavily disputed. An additional survey by Zaidi in 1996 got far lower results. The UNICEF survey showed a reversal in mortality rates between the Kurdish north and Saddam controlled south that may have indicated data manipulation. And a number of postwar UN-funded surveys have continued to get lower results (except in the Kurdish zone where they often match up).
The second question is the extent to which the excess deaths can all be blamed on sanctions. This is a complex political question, and there are various claims and counterclaims that I'm not able to do justice to.