Isn't barrier penetration more a function of bullet design than caliber?
Yes. But there are a variety of factors that go into it.
In this case early bullets for the .357 Sig tended to break up prematurely as they were essentially bullets designed for the lower velocities of the 9mm, meaning that they were literally bullets built for the 9mm. To take full advantage of the increased velocity and energy of the .357 Sig required that bullets specifically developed for it be used. In some cases specifically built for barrier penetration.
The idea that the .357 Sig was developed specifically for superior barrier penetration is inaccurate and largely driven by gun writers reading backward and internet drivel. The 9mm round was chosen for it because that made sense from an economic and design viewpoint as has been explained earlier.
I like the following tests done by a fella on another forum because they illustrate the variety of factors involved in penetration.
The tests aren't exhaustive, and had their limitations, but they do help to see the factors involved. Bullet construction, bullet diameter and sectional density, velocity, weight of the bullet, etc. There is a dialectic in their relationship.
Note here that most handgun rounds will easily penetrate auto glass and auto and truck bodies, etc. Numerous tests have shown this as well as daily experience over decades. But if you need to go through a windshield, through a person and into the rear seat or trunk a well designed .357 Sig round may do it better than a 40 or 45. But then, with the right bullet so could the .357 Magnum, 38 Super, 9x23 Winchester, etc. but they don't fit in the same sized gun.