@DocMcQuinn First, 20% by antibodies means 60%-100% actual infected & recovered, based on recent studies that looked at what fraction of recovered have detectable T-cell & mucosal Ab but not serum Ab responses -- a finding that's also consistent with several other observations and studies.
@DocMcQuinn NYC showed 20%-25% IgG+ in two randomized surveys conducted by the NYS Dept of Health as far back as late April, so presumably the great majority of NYC residents have been infected by now -- not surprising, given their high death toll & the total collapse of their death curve.
@DocMcQuinn Stockholm blood donors were 30% positive for COVID-19-specific T cells in May (about 3X the percentage of seropositives) and the Swedish FHM estimated a few days ago that the number of infected is now about 40%, which seems reasonable given another couple of months have passed.
@DocMcQuinn The population mortality in NYC is about 2.5X the population mortality in Stockholm County (~0.023-0.028% vs 0.010%), which is consistent with the difference in the estimated percentage of the population that has had Covid -- ~100% vs ~40%.
@DocMcQuinn The Stockholm mortality curve isn't quite as close to the zero bound as NYC's, but in both cities it's clear that vast bulk of the mortality which will occur (at least unless and until there's a "second wave") has already occurred and the epidemic is essentially over. 
@DocMcQuinn Thus, both cities have attained herd immunity locally, in the sense that idealized, pedagogical concept is applicable in the setting of a global pandemic. Importantly, Stockholm has reached HI at a much lower percentage infected & with correspondingly lower population mortality.
@DocMcQuinn Also importantly, as HI thresholds depend on context such as public health measures in place, Stockholm has attained HI in the context of a sustainable public health response (no lockdown) & NYC has attained it in the context of an unsustainable public health response (lockdown.)
@DocMcQuinn That last point is a bit academic for NYC as HI was only achieved there with the infection of essentially their entire population. As their immunity is now ~100%, it won't matter what they do or don't do there in terms of public health measures for the foreseeable future.
@DocMcQuinn I fully expect the HI threshold to be lower than in Stockholm, let alone in NYC, in most places. In many places it's probably zero, even in the absence of lockdown.
@DocMcQuinn However, looking at the mortality trends in Sweden, I would guess the national average won't be *dramatically* lower than for Stockholm -- it looks like at the end of the day population mortality outside Stockholm County will reach at least half of the asymptote in the capital.
@DocMcQuinn Also, some of that advantage (Non-Stockholm vs Stockholm) will not be due to a lower HI threshold but will instead reflect learning curve improvements in treatment & public health responses (e.g., better protection of care home populations.)
@DocMcQuinn My guess is the variation in population mortality across the nation here will probably be higher than in Sweden, partly due to the greater heterogeneity of the US but mainly since several North East states, notably NY & NJ, colossally screwed up, which probably won't be repeated.
@DocMcQuinn In that respect, we seem to be recapitulating the experience in Italy, where the first-hit region of the country did much, much worse than the rest of the country. The situation in Sweden looks like it will end up more homogenous.
@DocMcQuinn Despite these differences, things tend to even out at the national level and all three of the countries I mentioned -- Sweden, the US and Italy, look set to end up at roughly the same terminal population mortality, i.e., roughly 0.06%.
