My observations of FM Global

Editor's note: The following article represents the experiences of the author, not necessarily those of NRCA or Professional Roofing magazine.

The old Factory Mutual (FM) system that transformed itself into FM Global has long been proactive in developing criteria for roof system evaluation through its research division, now called FM Research. FM Research started out as a nonprofit arm of the FM system to evaluate the performances of many construction materials and systems, including roof systems and roof components, and assist FM's underwriters in adjusting insurance rates for its insureds.

With the transformation to FM Global came some other changes. FM Property Loss Prevention Data Sheets became FM Global Property Loss Prevention Data Sheets, and the FM Approval Guide now is published by FM Global Technologies LLC. Although the original intent was to provide in-house information to the FM system, FM recommendations have permeated every nook and cranny in the roofing industry and indirectly affect every roofing contractor who deals with FM Global insureds and anyone who subscribes to FM Research recommendations. The yet-to-be-introduced RoofNav computer program being developed by FM Research is supposed to replace its Approval Guide and provide up-to-the-minute information about FM Research-approved roof systems and roof system components. RoofNav promises to be a significant challenge for roofing contractors.

The Approval Guide probably is the least user-friendly, most misunderstood, basically unreadable document ever published. But RoofNav is supposed to resolve these issues. The Property Loss Prevention Data Sheets are much more understandable than the Approval Guide but usually fail to provide the information necessary to install roof systems in some configurations. For example: The only reference in Property Loss Prevention Data Sheet 1-29 to mechanical attachment of a base sheet is over a wood deck 3/4 of an inch (19 mm) thick or thicker using large-head (1-inch- [25.4-mm-] diameter) roofing nails. There is no recognition for fastening of base sheets to cementitious wood fiber decks, poured-in-place gypsum or gypsum planks, or lightweight insulating concrete decks (and no differentiation between perlite or vermiculite aggregate lightweight insulating concrete and lightweight insulating cellular concrete) even though it is certain FM Global covers insureds with these types of roof decks in their inventories. There is no differentiation between base-ply fasteners with 1-inch- (25.4-mm-) diameter heads and base-ply fasteners incorporating FM Research 1-90 stress plates even though common sense would suggest a 3-inch- (76.2-mm-) diameter 1-90 stress plate should increase the uplift properties of base sheets over those of base sheets secured with 1-inch- (25.4-mm-) diameter heads and allow installation of fewer 1-90 fasteners for given uplift specifications.

So a roof system designer must attempt to use the vague information provided in FM Global Property Loss Prevention Data Sheet 1-29 and come up with his best scientific wild-ass guess as to how to define the attachment of roof systems to be installed over these "unmentioned" roof decks so FM Global engineers can approve or accept the proposed roof system specification. It is obvious FM Research and FM Global don't communicate much with regard to the real world of roofing. But because no base-ply fastener supplier will pay FM Research to test fewer fasteners, there probably will be no changes to the data sheet.

In other words, FM Research writes the rules—however incomplete—regarding roof systems for use not only by FM Global but other insurers and designers of roof systems. One would think the rules promulgated by FM Research would, at the very least, apply to all divisions within the FM Global network in determining acceptable roof systems or in dispute resolution in cases of damaged roof systems. But though FM Research writes the rules of the game, when it comes to dispute resolution, the rules often become "irrelevant" to the FM Global engineering group responsible for determining the extent and nature of roof system damage.

A recent experience (and another not-so-recent experience) provides some interesting insight as to how the rules developed by FM Research do not apply to FM Global operations. This, of course, is at FM Global's sole discretion.

The owner of a large office-warehouse complex recently had a relatively large claim against its FM Global insurance policy. Four of the seven roof areas on the facility had blown off in a storm along with several roof systems on adjacent facilities and much accessory metal. There was no question the roofs on four of the seven roof sections were gone, and the FM Global engineer in charge of resolving the claim agreed to pay for the missing roofs. When the owner requested the adjacent appurtenant roof systems be checked for integrity, the FM Global engineer declined and indicated the roof systems were intact on the other areas and no further investigation would be undertaken by FM Global.

Following replacement of the four blown-off roof systems, the owner, still concerned about the conditions of the remaining roof systems, requested a private, independent evaluation of the three remaining roof systems. Nondestructive wind-uplift testing was conducted on the three roof areas in question according to FM Global Property Loss Prevention Data Sheet 1-52, which has been the standard for nondestructive wind-damage evaluation of roof systems for many years.

The first paragraph of the data sheet's scope reads as follows: "This data sheet describes two methods for field testing above-deck roofing assemblies on steel and other roof decks to determine if adequate wind resistance is present. Field uplift tests may be needed where inferior construction is suspected or known to be present or where a partial blow-off has occurred. These tests also may be used for confirmation of acceptable wind-uplift resistance on recently completed roof systems."

There certainly was a partial "blow-off" on this facility, and it would appear prudent to determine whether "adequate wind resistance is present" for the roof systems on the balance of the facility.

The company providing the independent testing randomly selected eight locations on the three remaining roof areas for nondestructive wind-uplift testing. The roof membrane in one sample area broke loose at 12 pounds of uplift—short of the first milestone of 15 pounds for one minute per the test protocol. The roof membrane began to lift at 5 pounds negative pressure in all other test areas, or the equivalent weight of the aggregate-surfaced built-up roof (BUR) membrane, and the samples failed by the uplift test criteria almost immediately. The test indicated an uplift "failure" within the roof system by FM Research test criteria in every test area.

Data sheet 1-52 suggests a roof be cut following the uplift test to determine the plane of failure. This step in the test protocol obviously was not necessary because of the overall minimal uplift resistance of the roof system and the fact the plane of failure didn't matter in this instance—the roof membrane was loose in all areas tested. The test information was conveyed to the owner, who then requested a re-evaluation by FM Global.

Following the nondestructive wind-uplift testing, at the owner's request, an FM Global engineer and roof consultant retained by FM Global (who, by the way, declined to be present during the nondestructive testing) visited the site. The consultant made noise about the lack of destructive verification in the test areas but proceeded to remove six samples from the roof systems in areas other than the location of the nondestructive testing even though the test areas were readily identifiable.

The BUR membrane and mopping surface of the fiberglass roof insulation separated cleanly from the fiberglass core board in each destructive test area—in essence verifying the results of the nondestructive uplift testing and resulting in 14 total random sample areas on three roof sections. The consultant retained the samples and returned home to conduct sample analysis of the roof membrane according to ASTM D2829, "Standard Practice for Sampling and Analysis of Built-Up Roofs."

The question in this case was not about the configuration or acceptability of the BUR membrane but about uplift resistance. ASTM D2829 makes no provision for evaluating a BUR membrane sample for separation from the top surface of the roof insulation or separation of the roof insulation from the substrate. The ASTM D2829 test simply is not applicable in this type of situation. The testing of the BUR membrane according to ASTM D2829 was, in this case, pure existence justification by the consultant and provided virtually no information about the plane of separation or adhesion of the roof membrane to the insulation or insulation to the substrate.

The conclusion of the testing included in the consultant's report was that the BUR system specification was an Owens Corning Fiberglas® BUR specification No. 323-TGD (nice information, but it's information that could have been gleaned from surface observation). From the roof membrane sample analysis, it was determined "the adhesion between the steel deck was inadequate and did not comply with the recommendations of the material manufacturer." And, therefore, the problem was not one of wind-uplift damage but rather one of noncompliance with the manufacturer's application recommendations—even though the asphalt quantity and distribution appeared to be in accordance with FM Research recommendations valid at the time of roof system installation.

The final conclusion was there was no evidence of wind damage on the subject roofs. I suppose eight failed nondestructive wind-uplift tests and six separated roof system samples just couldn't be construed to be evidence of wind damage.

Whether the roof system installation complied with the manufacturer's recommendations appears to be moot in this case because FM Global had the opportunity to examine the premises before issuing the insurance policy; apparently took no exception to the existing construction; issued the "replacement value" insurance policy; and, in doing so, essentially accepted the existing conditions. At this point, there was a standoff between FM Global and the owner. So another audit was scheduled to kick more rocks around on the suspect roof areas.

When the FM Global engineer and consultant finished kicking rocks around, they announced, once again, there was no evidence of wind damage and, therefore, no claim on the insurance policy.

When confronted with the results of the nondestructive testing conducted according to FM Research protocol, the FM Global consultant opined the separation of the roof membrane from the fiberglass core board occurred because the roof was "old." He never explained the reason that physical degradation sufficient to result in clean separation of the roof membrane from the insulation core board had occurred.

The only reason separation of a roof membrane mopping surface from a fiberglass core board occurs is because the roof has been subjected to relatively continual heavy foot traffic or wind uplift. Because there was no mechanical equipment on any of the three roofs and traffic was limited to an occasional maintenance walkover, it is unlikely heavy traffic accounted for the separation.

The FM Global engineer's response was most interesting. When the logic of the separation was explained and results of the nondestructive wind-uplift testing and destructive testing were once again called to his attention, he announced FM Global Property Loss Prevention Data Sheet 1-52 was "irrelevant" and FM Global engineering did not use FM Research testing criteria in adjusting claims—that was "FM Research stuff."

When queried why the data sheet was not rescinded if it was, in fact, "irrelevant," his response was, "That's their stuff, not ours." He went on to indicate the owner was not obligated to ensure his roof system was installed in accordance with FM Research recommendations any more that he was obligated to use the FM Research criteria in his adjusting decisions.

What he failed to say to the owner was that the owner certainly has the option not to use FM Research recommendations in the construction of his roof systems but that if the roof systems were not installed according to FM Research recommendations, his roof system classification would be downgraded and his insurance rates adjusted upward accordingly.

Funny. I was led to believe that in the transformation of the FM system to FM Global that all the original FM carriers and the research group were brought together as one big, happy family. It became clear there was no common ground within the FM Global organization.

In a similar but slightly different scenario, a fire in an FM Global insured's warehouse resulted in burning out the pre-engineered building insulation and burning through several metal roof panels just under the ridge of the roof on one side. The metal building supplier was long gone from the commercial market, and the acquisition of matching metal roof panels no longer was possible as an option. The FM Global consultant's solution to the problem was to partially replace the metal roof panels (he didn't mention where to locate the replacement panels) and install new pre-engineered building insulation in the burned-out area held in place by chicken wire secured to the purlins.

FM Global Property Loss Prevention Data Sheet 1-31 is specific on this matter and states, "Any existing batt-type insulation (standing- and lap-seam roofs) should be removed or held in place with steel liner panels (expanded metal or wire mesh is not acceptable) to avoid shielding of sprinkler water."

So why would the FM Global consultant make an "unacceptable" recommendation? That answer should be obvious: to minimize the amount of the claim that I'm not sure ever was resolved. Tarps remained as the permanent roof cover on the facility at the time of my last visit—more than a year later.

FM Global certainly has the right to research systems and materials for its underwriting information, and many of the recommendations FM Research promulgates have become "standards" in the roofing industry as a result of its efforts. FM Research's recommendations have been beneficial in reducing not only FM Global's losses but the losses of other carriers and many building owners.

But it seems FM Global wants to write the rules, enforce the rules when they are in its best interest and ignore the rules when they are not in its best economic interest. It seems FM Global has developed a double standard. Somehow, that doesn't make good business sense and certainly doesn't appear to be fair to its insureds.

So what does this mean for roofing contractors?

Consider an FM Global insured who receives a denial of a claim with a roof consultant's report indicating the roof in question never was constructed according to manufacturer recommendations. You can be sure there will be no reference to the fact that testing conducted on behalf of FM Global was inappropriate for the condition. The fact is a typical building owner has no idea what an ASTM test method is or what it includes; he only thinks a reference to ASTM should be valid. Who will the owner be looking to for restitution?

And if FM Global engineers take the position that any given FM Research test protocol or recommendation is "irrelevant" for claim settlement, how can they take the position that all other data sheets are relevant with respect to roof system installation? Given this logic, it would appear all FM Research recommendations should be rescinded, leaving the roofing industry absent of much valid, useful information.

I overheard an FM Global insured say he bought into FM Global for the engineering assistance, not the insurance. I hope he never has a major claim. I guess at least a portion of any wind-damaged roof system should end up in an adjacent parking lot before an FM Global adjuster appears to assess the damage.

Dick Baxter is president of CRS Inc., Monore, N.C.