I’m pretty sure employers are paying a lot more for workers to be in office than for them to work from home. The commute would be a tiny fraction of the electricity office space and other costs they pay for workers to be in office.
Hmm, just to prime the pump with some quick napkin-math... If it's an average of 200 ft²/employee [0] and $45/ft²/month [1] that gives a cost $108k/employee/year.
In comparison, lets imagine someone already making $80k/yr, a 40-mile heavy-traffic roundtrip commute taking 2 hours (1.25x time) and adds $10 in daily gas... That's about +22.5k in costs.
So yes, in direct math it's probably not as big, but having a physical office probably satisfies other hard-to-quantify priorities of corporate executives, such as personal prestige, sense of control, and how things look to investors.
This doesn't add up to me. The priorities of corporate executives are to please the board of directors. BoD priorities are to please shareholders/investors. Shareholders/investors are pleased by growth/profits. That is, the investors care not about "how things look" but instead how they actually are, as in quarterly reports, share price, etc.
If employees are happier and more efficient working from home, it satisfies the priorities of all the above parties. If RTO makes people unhappy and either less productive OR makes them not want to work for that employer, it goes against the incentives and priorities of all the above parties.
CORRECTION: That figure from [1] doesn't actually specify the time period, but other sources [3] imply it is per year rather than per month. This means I overstated the cost of office-space by a factor of 12, and it should probably be more like $9k/employee/year.
That also means the office-rent cost is significantly less than if the employer began paying for employee commutes.