Motors are sized based on the application requirements.
High starting torque (required for a high inertia load) usually translates into light loading at rated speed and voltage. The driving factor is NOT the efficiency under the loaded condition - it is the ability to start the process in the first place.
Environmental conditions also affect design choices - high ambient temperatures, high altitudes, and presence of hazardous (e.g. combustible or flammable) materials will necessitate design changes that ultimately may have the machine operating somewhat less efficiently than might otherwise occur.
Oversizing also leads to worse power factor, which again reduces the overall efficiency of the system. High currents cause power loses in cables and transformers. IE3 motors have a better efficiency than IE2 motors and IE2 motors have a better efficiency than IE1 motors. But the inverse is true for power factor and worse under light loads. So without power factor correction, the efficiency of the system may be worse.
The power quality of the system upstream of the motor terminals can also affect design decisions. If there is noticeable harmonic content, the machine may be "overdesigned" to accommodate the added thermal and voltage stresses applied to the winding - which may well lead to a reduction in observed "efficiency".
As one more consideration - the cost of operating a motor with a slight efficiency penalty may well be significantly lower than the capital cost of having to purchase a new machine rating (and all the other regulatory and infrastructure costs that may also result) when the process changes somewhere down the road in the next decade.
Check ALL the reasons for the motor sizing as "steady state" operation may not be the driving factor in the choice.
* Starting a high-inertia load requires a very high starting torque, which may only be obtainable by using a larger-than-steady-state rating.
* Perhaps the environment is hazardous, and exceptionally low surface temperatures are necessary to avoid accidental ignition of the surrounding medium.
* The application has a requirement for rapid reversal, which would indicate a very high peak torque capability ... and a low magnetic saturation.
* Regulatory requirements for increased efficiency (measured as electrical efficiency, not including power factor) tend to result in operating at lower steady-state values in an attempt to limit temperature rise - and therefore copper losses. This is because the highest efficiency point is almost always at something below the full rated nameplate power output of the machine (although to be fair, it certainly does not occur down in the 25-40 percent rated load region!).
The cost of "future process changes" is indeed a factor in the decision to "oversize". This is especially true for larger equipment, where the initial purchase price plus the cost of "lost" energy during operation of the machine can be far lower than the expense associated with loss of production plus facility infrastructure changes plus a new machine down the road.