Other considerations...
A smaller inverter may avoid electrical upgrade costs that will never pay themselves back. For example, suppose a residential 100A service could only take a 3800W inverter, but the sold system is 5kW based on the customer energy needs. Upgrading the service will cost $3000 but will likely only result in $1000 more savings from increased energy production over 25 years. The extra kW of modules may only deliver 75% of what the rest do, but still is cost effective at delivering energy at lower than utility cost.
On utility scale systems, the contract arrangements may greatly benefit having the inverter deliver a more or less constant full output for as much of the day as possible. And since adding more modules has gotten cheaper and cheaper, it often makes sense to just add modules to make that happen.