Generally, no, it doesn't make sense to shape a port at its physical port bandwidth.
If you don't shape an interface, it will congest when and if the transmission rate cannot keep up with the offered rate.
The default, for not shaped interfaces, there's usually a default software FIFO queue. Often only when that overflows, are there drops.
Again, no, there's little reason to shape an interface at it physical bandwidth.
Where shapers can be useful, is shaping for "downstream" bandwidth that's less the the interface bandwidth, and when that's done, managing the congestion as you desire.
For example, assume you have a site in the US and Britain. They have local connections to some "cloud" technology. The Britain side has an E1, 2 Mbps, while the US side has a T1, 1.5 Mbps. This means, the Britain site can send 500 Kbps faster than the US side can accept.
If this example, if you do nothing, congestion may form at the cloud's egress to the US site, where you have may have no say in how congestion is managed. But if you shape the Britain traffic, to the US side, at 1.5 Mbps, it will congest first at your Britain shaper providing you the option to manage that congestion.
