Undergrounding isnt a slam dunk winner either.  While it certainly is more attractive from a visual standpoint and more secure from wind/storm damage it does not mean there are drawbacks.  

From a servicing standpoint, overhead lines are far easier to observe, inspect and repair than underground lines.  I discussed this and other pros and cons with a line engineer with SCE.  According to the engineer, his crews prefer servicing and troubleshooting overhead lines because the entire portion is visible.  If there are issues with an undergrounded circuit between transformers or other equipment, they pretty much have to replace the entire segment.  

Longer term, underground equipment and lines do have usable lifespans.  Which means there will come a point in the future where these lines WILL have to be replaced.  How long that lifespan is is dependent on the quality of the equipment and wires as well as how much load those circuits handle.  According to the engineer, early undergrounding equipment lifespans were approximately 30 years, give or take.  Most of what goes underground today will last 50 years or so, provided the electrical load remains at the rated capacity. 50 years is a pretty good length of time but as we know simply putting off a problem (the eventual need to replace the lines) into the future does not mean it will be addressed (or could be addressed) when that time comes. Repair and or replacement of undergrounded lines will almost certainly be more difficult to utility companies than the original undergrounding was (mostly performed by the developers).  

And that assumes that the raw material/specialized equipment is readily available and affordable.  

On a personal note, I lived in a subdivision (in Virginia) that was undergrounded in the early 60's with lines with a 30 year lifespan.  As expected by the late 80's the system was showing its age.  We would lose power every time it would rain more than one inch.  The utility could never locate the problematic section so after a year or so of undependable service they replaced the entire distribution section.

Makes you wonder which method is better in the long term.

    This issue has been discussed extensively in the DC area, where much of the system relies on overhead cables and power has been lost over large areas in recent storms. However, there are some underground cables and there have been problems in actually finding where these cables are - in some cases the records just aren't very good.
Stssgabs100
Technologies have been developed that solve some of these issues, with some of the same basic ideas (though obviously modified) as have been developed for replacing sewer lines, without requiring surface disruption.
Then you hook up a radio signal generator at one end of the cable and track it with a hand carried reciever and mark the cable run on the ground with spray paint.
This is good in theory, but it is sometimes lacking in practice.  A couple examples:

We were placing a traffic signal pole at an intersection for a 'hurry up' job.  The DOT rep brought the utility companies together on site and said, "We are drilling here (pointing to the ground) Monday with a 48" auger down to 22'.  Is anyone in the way? Yes or No."  An old worker from the local water company said there might be a problem because he didn't remember the water line being where it was marked (in the nearby ditch).  So they dug it up.  The 8" water line made a jog in the middle of the sidestreet but the locate guys didn't pick up on it because of the high water table.  Our pole foundation was actually closer to the locate markers, by 10', than the water line was.

Next: Fiber Optic contractors for the telephone companies are notorious for putting a couple big loops (10+') of fiber at the end of a run.  They just throw it out to the side and bury it.  The locates come along, and miss the big loop to the side and go to the junction box.  I know power would be in conduit, but 'stuff happens'.

The best underground utility locator in the world: a Backhoe.