I make the assumption that we are discussing spacers as used in Overlanding types of use, and not on some sort of trailered buggy.
I see this subject not as any empirical evidence proves it right, but more like scientific theory's where one piece of contrary evidence proves it invalid. From:
http://www.expeditionportal.com/forum/showpost.php?p=286259&postcount=20
Grim Reaper said:
Wanna hear my loosing a 35inch bfg off a really lanky 75 Jimmy at 65mph with my kids in the truck becasue of a spacer story? You should see what a 90lb wheel/tire rolling 40+mph does after hitting a curb bouncing 20ft in the air and then slamming into the passenger door of a Ford Ranger (it totaled the truck with $1800+ in damage). Never again will I run a spacer.
You can buy Eagle 589's in a 17.
If they were 15x7 IFS versions I'd be all over them. I run them on my 4Runner. I run Eagle 589's on my Suburban.
For me the cost of a single failure is too high. What if there had been an occupied child's seat in that Ranger's passenger side?
The issue of an aluminum spacer is not it's strength. It is that aluminum, unlike low carbon steel, has a finite fatigue life. Those that mention how many thousands of miles or years that they have on their spacers are closer to a failure now than when brand new. More time in service means that they are closer to the coming failure.
What will determine when the spacer fails (it
will eventually fail) is how highly loaded it is and how long it is in service. Unfortunately there is no simple way to predict when the failure will happen because it is too dependent on how good a job the designer did in reducing the internal stress' and how good a job the mfg shop did in following the design, and also how perfect the blank they started with was (Nothing is ever perfect!).
If the Mfg has done all of their homework they know what the internal stress' are and, based on Fatigue Life calculations for the alloy used, also know what the maximum expected service life should be. None of this will account for an unknown imperfection in the metal, but it is a reasonable indicator of when the spacers should be replaced. Not quite as "plug and play" as saying "a loading of X lbs. has a service life of Y thousand miles" since the loading usually varies over time. Obviously, the lighter the loading, the longer the service life will be. In a very lightly loaded case the service life may effectively be infinite, but absent mfg's guidelines or published stress data how do you know?
The actual position of the WMS is insignificant to the loading of the wheel bearings (& other joints if so equipped). Where the contact patch centerline is relative to the WMS is the important item. Witness a U.S. 2WD dually pick-up. The front axle WMS is
way outside of the contact patch center, but the bearings are not.
In the case of dual rear tires the need for both tires to be operational becomes paramount less the now dislocated contact patch center be moved an exceptional distance away from being centered on the bearings.
As to the choice of AR 589's, I know of one set of 6 that saw 10 plus years of regular pre-run and chase service in Baja and in the US. Once that truck was raced because the desert race truck was not ready. Those wheels were retired when one did finally develop a crack. The one mentioned with the problem is a classic example of an unknown imperfection causing a catastrophic result. Empirical evidence is handy, but in most cases one can find counter evidence to any single claim. The percentage of failures is what matters in this sort of analysis.