Not sure why the formatting for the stats in my earlier reaction was all on line (it was not in the edit window and I can not longer edit my earlier reaction), but these are the stats in better readable format:

• motorway :: missing maxspeed: 62 , total: 13163 , pct with maxspeed: 99.5
• trunk :: missing maxspeed: 50 , total: 8224 , pct with maxspeed: 99.4
• primary :: missing maxspeed: 446 , total: 45602 , pct with maxspeed: 99.0
• secondary :: missing maxspeed: 2055 , total: 67339 , pct with maxspeed: 96.9
• tertiary :: missing maxspeed: 7817 , total: 108843 , pct with maxspeed: 92.8
• unclassified :: missing maxspeed: 62620 , total: 219803 , pct with maxspeed: 71.5
• residential :: missing maxspeed: 78225 , total: 361891 , pct with maxspeed: 78.4

I did an analysis for the Netherlands using osmium and some python code:

motorway :: missing maxspeed: 62 , total: 13163 , pct with maxspeed: 99.5 trunk :: missing maxspeed: 50 , total: 8224 , pct with maxspeed: 99.4 primary :: missing maxspeed: 446 , total: 45602 , pct with maxspeed: 99.0 secondary :: missing maxspeed: 2055 , total: 67339 , pct with maxspeed: 96.9 tertiary :: missing maxspeed: 7817 , total: 108843 , pct with maxspeed: 92.8 unclassified :: missing maxspeed: 62620 , total: 219803 , pct with maxspeed: 71.5 residential :: missing maxspeed: 78225 , total: 361891 , pct with maxspeed: 78.4

In the Netherlands all traffic signs are available as open government data, so that might explain the pretty good numbers.