It has got to do with how the TT's are designed.
Normally, when a schedule is being designed, first they assume that the train will continuously run at the MPS for the entire stretch between 2 stops, and come up with an initial schedule. Then they will add buffer times everywhere to accommodate the speed restrictions and other constraints.
Despite the added buffer, the assumption is that the LP runs at MPS everywhere possible. Now, in practice, depending on the route, the LP's may...
more... not wish to run at MPS sometimes. For example, in a route with very frequent speed restrictions, it would be really straining for the LPs to continuously increase and decrease the speed for short distances, and they will prefer to just cover the section at a moderate speed. This, in turn, will cause a loss in punctuality. And overspeeding is actually something IR takes very, very seriously, so LPs tend to play it safe and not hit MPS.
So, by reducing the MPS in the design of the TTs, they are providing some room for the LPs. Now, since the TT is designed for a slightly lower booked speed, they can maintain the TT even if they keep a little margin from the MPS while running.