If we take the ML estimate, I would say that the 1/9th chance distributes evenly. So it is 1/9th loss for both cases.
I have found that large corporations are able to capitalize better on incremental improvements and responsibilities are more narrow: hence they can tolerate less overall employee productivity.
Also, they might be more experienced at managing people and understand that advantages that working from home bring. For example, taking care of your children so that you live in constant fear of loosing your job :-)
I'd disagree, based on my observations in both environments. Law of large numbers shows that it is easier to hide unusual behavior when number of samples is large and the observer is focused on aggregate-based performance measurement.
I have found that large corporations are able to capitalize better on incremental improvements and responsibilities are more narrow: hence they can tolerate less overall employee productivity.
Also, they might be more experienced at managing people and understand that advantages that working from home bring. For example, taking care of your children so that you live in constant fear of loosing your job :-)