A comparison between per capita 'confirmed' infection rates and CFR for several countries.  Data as at 23 March.  NZ separated at the top as generally not comparable to the other countries.  First case here only 25 days ago whereas the other countries listed have had confirmed infection for more than double that time. NZ probable cases are included as is the case (in part?) for China. 

In the table, CFR (Case Fatality Rate) is calculated "No. Deaths divided by No. Confirmed Cases 7 days earlier" as per https://www.worldometers.info/coronavirus/coronavirus-death-rate/#correct.  Numbers of cases & deaths generally sourced from Wikipedia (plus alternate US data from https://www.worldometers.info/coronavirus/).  Numbers of confirmed cases are subject to the extents and criteria of testing which vary between jurisdictions (highest in South Korea). 

A few observations:

• Taiwan & Japan have the lowest per capita cases by a large margin, with China the third lowest (lower even than Australia).  Of course the corresponding figure for Hubei province would be a lot higher (but not as high as for Italy)
• The other 'gold standard' state listed, Singapore, has 10 times the per capita rate of cases in Taiwan and Japan, though it is lower than most parts of Europe and the US. 
• Singapore is though another jurisdiction where imported cases are a significant factor (only locked down the border a day or two ago)  
• Singapore has the lowest CFR (0.8%), though with only two deaths that is very sensitive to any further deaths.  Other countries with low CFR rates are South Korea, Germany, and Australia (1.9%). Germany stands out in that regard from the other European countries listed, which have CFR values of 13% to 23%.  Again for Taiwan, the CFR is based off only 2 deaths.  Also, the current number of cases in Taiwan is three times what it was 7 days ago (hence the 3% figure). 
• Notably, CFR is not well correlated with confirmed cases per capita.

The table data only tells part of the story.  The rates of increase in cases are also telling.  In almost all cases there is initially exponential increases in cases above 150.  The exponential phase lasts for a further 10 - 20 days before starting to roll off, depended on when restrictions and lockdowns are imposed.

Though still initially exponential the rates of growth in numbers in Japan and Singapore were notably lower than in China, the US, Europe (including Germany), and Australia (data not plotted for all countries in the table).  Singapore, Australia, the US, and possibly UK are still in the exponential phase, with the US still heading for 100,000 cases in the next few days, and potentially 300,000 in about 6 days. 

Japan is a particularly interesting case.  Not just a low number of cases per capital, and low initial growth in cases, but appears to have now 'rolled off' the exponential increase in cases.  Interesting because from a discussion with a colleague in Tokyo this morning:

"On COVID-19, here in Tokyo still relatively ok.  The office is split in two teams that don’t overlap. Abe has special powers for now, but nothing has been really enforced. ... We can go outside and all that stuff"

No lock down, offices still working near normally, and presumably still crowded subway trains.

It does make me wonder if wearing masks is possibly a significant factor, despite what we have been told.  Also likely applicable to Singapore, etc.