We are trying to optimize Lighting in fillrate. Integrated graphics chips are very slow in memory access.
You might want to look at The Compact YCoCg Framebuffer. It uses a 2-channel buffer to store luminance for every pixel and the two chroma components in half the pixels each, forming a checkerboard. It also uses an edge-aware upsampling filter at the end of the frame to reconstruct the missing chroma components and convert back to RGB.
You could extend this to a single 3-channel buffer to store the diffuse luminance, specular luminance, and checkerboarded chroma (assuming based on the diagram in your question that it suffices to store just one pair of chroma values).
You might also consider using R10G10B10A2 format instead, unless you really need the range of the floating-point format. It's the same amount of bandwidth but will give you a good deal more precision (R11G11B10F only gives you 5–6 bits of mantissa per component). Plus, you can use the 2 bits of alpha to store a couple of flags, such as your diffuse/specular flag.
If you are ok with specular being only white, you could put diffuse and specular into RGBA.
It looks like that is what unity does for it's deferred rendering.
Someone discovering this issue: http://forum.unity3d.com/threads/specular-color-based-on-light-color.134412/
The docs mentioning it: https://docs.unity3d.com/Manual/RenderTech-DeferredLighting.html