As far as I know, usually an Ethernet connection either negotiates the maximum rate advertised by both ends (regardless of actual cable quality), or fails negotiation and falls back to 10 Mbps half-duplex. If two 100 Mbps devices are connected with a really bad cable, I would expect them to still negotiate 100 Mbps and just work poorly.
But I’ve seen various cases where devices actually switch to a lower rate:
A computer should have a 1 Gbps connection (and does for some time after connecting), but several minutes later switches to 100 Mbps full-duplex on its own. Reconnecting the cable brings 1 Gbps back, but a few hours later it’s at 100 Mbps again.
(The cable is 4–5 meters Cat5e. Computer’s owner insists it’s because the Ethernet cable is running right next to AC mains cable, though posts on SU usually say that shouldn’t be a problem.)
An Ubiquiti device should have a 100 Mbps connection (and does for some time), but several days later I find it using 10 Mbps full-duplex. Merely reconnecting the cable brings 100 Mbps back, but several days later the problem repeats again.
(The cable is Cat5e, somewhere around 20–40 m. Two similar devices exhibit this problem. They’re both on the roof, and I’m not sure if the cable is an “outdoors” one, although it is at least FTP from what I remember. It wasn’t professionally installed, that’s for sure.)
In either case, the speed negotiation seems to succeed (it’s not a half-duplex mode, and both the device and the switch agree on the selected mode), I just have no idea what the mechanism is? Port statistics on all devices always show 0 errors received, for what it’s worth.
(I guess at least the 1st case could be explained by the “Ethernet@Wirespeed” feature, but do they apply even after all 4 pairs were determined to work? And in the second case, it’s only a 10/100 device with a two-pair connection anyway.)
There are network cable testers available (e.g. from Fluke) which measure the physical properties of ethernet cables if you’re unsure about the cabling.
You also might want to check the power supplies of Ethernet switches in between. It has happened several times to me that dying power supplies (usually seen after 5 years of use) create the strangest connection problems.
About the fact, that some cables are installed outdoors:
I had cheap SFTP cables for indoor use routed on the outside of a house for 10 years in direct sunlight without any visible cracks to the outer PVC layer after that period. YMMV if the cables are exhibited to very hot or cold conditions.