Geostationary satellites are at 35,786 km and the altitude is very precisely known, as is the rotation speed (obviously). This would simply be in contradiction to your assertion.

A relevant point which has to be considered.

So to try and put some values of Ag at this altitude

a rough approximation of the value of the predicted strength of the acceleration applied to a Geo stationary satellite is 225 millimetres a second squared,however if you factor in just earths orbit around its barycentre this causes this value to vary by around 40 millimetres per second squared with out factoring in any other variables.

Then if you consider what i am suggesting you have to allow for the fact half of earths volume is within the last 1315 kilometres of earths radius so the vector angle may not have diminished as much as may be first thought.

But that said it is thought that earths core is much denser than the rest of the planet, which is most likely and will have a direct effect on resulting attraction,Indeed it would change the results of the 13.8 metres per second squared for the 90 degree average interaction on earth to a lesser figure but this does not change the basic principle to which i am alluding.

So these things considered [angle of attraction given the satellites distance and the fact that earth is not a homogeneous sphere]So to just put some rough figures to give a example, that the mass acceleration resulting force at earths surface equal to 12.4 metres per second squared giving a force on the satellite equal to an acceleration of 283 millimetres per second squared.

So only about 20 millimetres PS2 different to the variable that is predicted by the inverse square law as it is presently applied when considering earths orbit around its barycentre, and as there are other variables to allow for it may be within the tolerances of the station keeping that is designed in to these satellites.

There are other points of concern that call in to question using the inverse square from the centre, for example if you take point mass at the centre and two point masses any where else but equal distance from a body [say a satellite] the distance between the masses not at the centre do not increase in the same way as the distance between the point mass at the centre and the satellite.

[remember in reality every particle attracts every other particle from where they actually are not from the centre]

I look forward to comments