Talked about this earlier. The problem with this estimate is NYC. 8.4 million people. With an IFR of .26%, it would predict 21,840 deaths. There have already been 20,800 deaths, confirmed via positive test data. Tracking "excess deaths" (ie people who died without being tested for COVID-19) this number is over 30,000. So in order for this IFR estimate to be accurate, the outbreak in NYC would have to be essentially over. But the antibody testing shows only about 25% of people having been infected.
Decisions in NYC regarding care facilities, and delayed lockdowns changed the speed / timeline, density affected the infection rate, but even in NYC the hospital system didn't break, so the base "lethalness" that IFR should measure shouldn't be different in NYC than elsewhere. If it doesn't work for a place that experienced a larger outbreak, it isn't accurate. In this case, it's too low.
It's lower elsewhere, because of suppression tactics, delaying the virus penetration(shown in the antibody testing elsewhere being super low). Stop those tactics, and we'll see what happens.