Lets weigh up the possibilities:
1.Every monitor is different and how they handle scaling will differ.
2.Each monitor can have different refresh rates which may also impact latency.
3.The scaling information needed by the display is traveling through display cables and may slighty limit the combining Refreshrate and resolution bandwidth causing latency.
4.GPU scaling may increase GPU workload which adds to a GPU bound scenario that adds latency or a reduction in frame rate that also adds latency.
Since I have CapFrameX Im able to research the soft statistics, my system latency is pretty low and stable:
For Apex Legends. (DX11)
Variances show that GPU, No scaling, Override scaling set by games is the most stable.
While enabling ingame scaling there are significant stutters so its better to leave this for the GPU.
Using the GPU, full screen aspect, Override it adds a bit of 8ms variance.
Display scaling does the same but with more 4ms variance.
The spikes in frametime are more frequent and higher with ingame scaling:
Framerate averages also show you have higher framerates when using GPU scaling.
Given these results GPU, no scaling, Override is the most stable.
Its possible that information sent to the display that "no scaling is needed" requires some GPU usage, removing that requirement by stating the GPU should do the scaling and that no scaling is required reduces GPU impact.