Standard for the Flammability (Open Flame) of Mattresses and Mattress/Foundation Sets; Notice of Proposed Rulemaking

CONSUMER PRODUCT SAFETY COMMISSION

16 CFR Part 1633

Standard for the Flammability (Open Flame) of Mattresses and Mattress/Foundation Sets; Notice of Proposed Rulemaking

AGENCY:

Consumer Product Safety Commission.

ACTION:

Notice of proposed rulemaking.

SUMMARY:

The Commission is proposing a flammability standard under the authority of the Flammable Fabrics Act that would address open flame ignition of mattresses and mattress and foundation sets (“mattresses/sets”). The Commission currently has a flammability standard that addresses ignition of mattresses by cigarettes. However, that standard does not address mattress fires ignited by open flames. The proposed standard sets performance requirements based on research conducted by the National Institute of Standards and Technology (“NIST”). Mattresses/sets that comply with the proposed requirements will generate a smaller size fire, thus reducing the possibility of flashover occurring. These improved mattresses should result in significant reductions in deaths and injuries associated with mattress fires. Due to the interaction of mattresses and bedclothes discussed herein, elsewhere in today's

Federal Register

the Commission is publishing an advance notice of proposed rulemaking to begin rulemaking on bedclothes.

DATES:

Written comments in response to this document must be received by the Commission not later than March 29, 2005. Comments on elements of the proposed rule that, if issued in final form would constitute collection of information requirements under the Paperwork Reduction Act, may be filed with the Office of Management and Budget (“OMB”) and with the Commission. Comments will be received by OMB until March 14, 2005.

ADDRESSES:

Comments should be filed by email to

cpsc-os@cpsc.gov

. Comments also may be filed by telefacsimile to (301)504-0127 or mailed, preferably in five copies, to the Office of the Secretary, Consumer Product Safety Commission, Washington, DC 20207-0001, or delivered to the Office of the Secretary, Consumer Product Safety Commission, Room 502, 4330 East-West Highway, Bethesda, Maryland; telephone (301) 504-7530. Comments should be captioned “Mattress NPR.”

Comments to OMB should be directed to the Desk Officer for the Consumer Product Safety Commission, Office of Information and Regulatory Affairs, OMB, Washington, DC 20503. The Commission asks commenters to provide copies of such comments to the Commission's Office of the Secretary, with a caption or cover letter identifying the materials as comments submitted to OMB on the proposed collection of information requirements for the proposed mattress standard.

FOR FURTHER INFORMATION CONTACT:

Margaret Neily, Directorate for Engineering Sciences, Consumer Product Safety Commission, Washington, DC 20207; telephone (301) 504-7530.

SUPPLEMENTARY INFORMATION:

A. Background

On October 11, 2001, the Commission issued an advance notice of proposed rulemaking (“ANPR”) concerning the open flame ignition of mattresses/bedding. 66 FR 51886. The ANPR was the result of the staff's evaluation of fire data over the course of several years and petitions filed by Whitney Davis, director of the Children's Coalition for Fire-Safe Mattresses (“CCFSM”). Although the Commission has an existing mattress flammability standard that addresses ignition by cigarettes, 16 CFR Part 1632, no current Commission standard directly addresses open flame ignition of mattresses. The most common open flame sources are lighters, candles and matches. The Commission is now issuing a notice of proposed rulemaking (“NPR”) proposing a flammability standard to address open flame ignition of mattresses.

1

1

Commissioner Thomas H. Moore issued a statement, a copy of which is available from the Commission's Office of the Secretary or from the Commission's Web site,

http://www.cpsc.gov

.

Characteristics of mattress/bedding fires.

A burning mattress generally provides the biggest fuel load in a typical bedroom fire. Once the mattress ignites, the fire develops rapidly creating dangerous flashover conditions. Flashover is the point at which the entire contents of a room are ignited simultaneously by radiant heat, making conditions in the room untenable and safe exit from the room impossible. At flashover, room temperatures typically exceed 600-800 C (approximately 1100-1470 F). In these conditions, carbon monoxide rapidly increases, and oxygen is rapidly depleted. Mattress fires that lead to flashover are responsible for about two-thirds of all mattress fatalities. This accounts for nearly all of the fatalities that occur outside the room where the fire originated and about half of the fatalities that occur within the room of origin. A mattress that reduces the likelihood of reaching flashover could significantly reduce deaths and injuries associated with bedroom fires. [1&2]

2

2

Numbers in brackets refer to documents listed at the end of this notice. They are available from the Commission's Office of the Secretary (see “Addresses” section above) or from the Commission's Web site (

http://www.cpsc.gov/library/foia/foia.html

).

The size of a fire is measured by its rate of heat release. A heat release rate of approximately 1,000 kilowatts (“kW”) leads to flashover in a typical room. Tests of twin size mattresses of traditional construction (complying with the existing mattress standard in 16 CFR 1632) without bedclothes have measured peak heat release rates that exceeded 2,000 kW in less than 5 minutes. Tests of traditional king size mattresses measured nearly double that peak rate of heat release. [2]

Fire modeling and available test data show that as a room fire grows, a layer of accumulating hot gases and smoke thickens downward from the ceiling. For fires exceeding 600 kW, this layer typically descends to less than three feet from the floor. Heat release rates exceeding 500 kW are generally considered to pose a serious threat of incapacitation and of igniting nearby items. [2]

The objective of the proposed standard is to limit the size of mattress/bedding fires to below 1,000 kW for a period of time by reducing the heat release from the bed, specifically the mattress and foundation, and by reducing the likelihood that other objects in the room will become involved in the fire.

Research has shown that the mattress, foundation and bedclothes operate as a system in bedroom fires. Often the first item ignited is bedclothes, which then ignite the mattress. The gas burners used in the proposed test method are designed to represent burning bedclothes. Research has indicated that bedclothes themselves can contribute significantly to fires, even reaching heat release rates of up to 800 kW. [2&13] Because of the role of bedclothes in mattress fires, the Commission is initiating a rulemaking on bedclothes through an ANPR that is published elsewhere in today's

Federal Register

. The Commission received numerous comments on the mattress ANPR concerning the role of bedclothes and the need for a rule addressing them. These comments are discussed in section J of this document.

NIST research.

The industry's Sleep Products Safety Council (“SPSC”), an affiliate of the International Sleep Products Association (“ISPA”), sponsored a research program at the National Institute of Standards and Technology (“NIST”) to better understand mattress/bedding fires and establish the technological basis for future performance requirements of a standard. NIST has conducted extensive research, which has become the basis for California's open flame mattress standard (Technical Bulletin or “TB” 603) and for the Commission's proposed standard.

The NIST research showed that a full-scale test is the most reliable method for measuring fire performance of mattresses/sets because they contain many materials in a complex construction. Because the order of materials, method of assembly, quantity of materials, and quality of construction, among other factors, can affect fire behavior, the complete product may perform differently in a fire than the individual components would. Based on its research, NIST drafted a full-scale test method for mattresses that uses a pair of gas burners to represent burning bedclothes as the ignition source. Both the Commission's proposed standard and California's TB 603, use this test method. [1&2]

Overview of the proposed standard.

With certain exceptions explained in section G below, the proposed standard requires manufacturers to test specimens of each of their mattress prototypes (designs) before mattresses based on that prototype may be introduced into commerce. If a mattress and foundation are offered for sale as a set, the mattress must be tested with the corresponding foundation. The prototype specimens are tested using a pair of gas burners as the ignition source. The mattress and corresponding foundation, if any, must not exceed a 200 kW peak heat release rate at any time during the 30 minute test, and the total energy released must be less than 15 megajoules (“MJ”) for the first 10 minutes of the test. The proposed standard is discussed in greater detail in section G of this document.

B. Statutory Authority

This proceeding is conducted pursuant to Section 4 of the Flammable Fabrics Act (“FFA”), which authorizes the Commission to initiate proceedings for a flammability standard when it finds that such a standard is “needed to protect the public against unreasonable risk of the occurrence of fire leading to death or personal injury, or significant property damage.” 15 U.S.C. 1193(a).

Section 4 also sets forth the process by which the Commission may issue a flammability standard. As required in section 4(g), the Commission has issued an ANPR. 66 FR 51886. 15 U.S.C. 1193(g). The Commission has reviewed the comments submitted in response to the ANPR and now is issuing a notice of proposed rulemaking (“NPR”) containing the text of the proposed rule along with alternatives the Commission has considered and a preliminary regulatory analysis. 15 U.S.C. 1193(i). The Commission will consider comments provided in response to the NPR and decide whether to issue a final rule along with a final regulatory analysis. 15 U.S.C. 1193(j). The Commission cannot issue a final rule unless it makes certain findings and includes these in the regulation. The Commission must find: (1) If an applicable voluntary standard has been adopted and implemented, that compliance with the voluntary standard is not likely to adequately reduce the risk of injury, or compliance with the voluntary standard is not likely to be substantial; (2) that benefits expected from the regulation bear a reasonable relationship to its costs; and (3) that the regulation imposes the least burdensome alternative that would adequately reduce the risk of injury. 15 U.S.C. 1193(j)(2). In addition, the Commission must find that the standard (1) is needed to adequately protect the public against the risk of the occurrence of fire leading to death, injury or significant property damage, (2) is reasonable, technologically practicable, and appropriate, (3) is limited to fabrics, related materials or products which present unreasonable risks, and (4) is stated in objective terms. 15 U.S.C. 1193(b).

C. The Product

The proposed standard applies to mattresses and mattress and foundation sets (“mattresses/sets”). Mattress is defined as a resilient material, used alone or in combination with other materials, enclosed in a ticking and intended or promoted for sleeping upon. This includes adult mattresses, youth mattresses, crib mattresses (including portable crib mattresses), bunk bed mattresses, futons, flip chairs without a permanent back or arms, sleeper chairs, and water beds and air mattresses if they contain upholstery material between the ticking and the mattress core. Mattresses used in items of upholstered furniture such as convertible sofa bed mattresses are also included. Not included as mattresses are: sleeping bags, mattress pads, or other items used on top of the bed, or upholstered furniture which does not contain a mattress. However, the Commission could decide to address mattress pads or other top of the bed items in its rulemaking on bedclothes.

Under the proposed standard, the mattress must be tested with its corresponding foundation if the mattress and foundation are offered for sale as a set. A foundation is a ticking covered structure used to support a mattress.

According to ISPA, the top four producers of mattresses and foundations account for almost 60 percent of total U.S. production. In 2001, there were 639 establishments producing mattresses in the U.S. [10]

Mattresses and foundations are typically sold as sets. However, more mattresses are sold annually than foundations; some mattresses are sold as replacements for existing mattresses (without a new foundation) or are for use in platform beds or other beds that do not require a foundation. ISPA estimated that the total number of U.S. conventional mattress shipments was 21.5 million in 2002, and is estimated to be 22.1 million in 2003 and 22.8 million in 2004. These estimates do not include futons, crib mattresses, juvenile mattresses, sleep sofa inserts, or hybrid water mattresses. These “non-conventional” sleep surfaces are estimated to comprise about 10 percent of total annual shipments of all sleep products. The value of mattress and foundation shipments in 2002, according to ISPA, was $3.26 and $1.51 billion respectively. [10]

The expected useful life of mattresses can vary substantially, with more expensive models generally experiencing the longest useful lives. Industry sources recommend replacement of mattresses after 10 to 12 years of use, but do not specifically estimate the average life expectancy. In the 2001 mattress ANPR, the Commission estimated the expected useful life of a mattress at about 14 years. To estimate the number of mattresses in use for analysis of the proposed rule, the Commission used both a 10 year and 14 year average product life. Using CPSC's Product Population Model, the Commission estimates the number of mattresses currently in use (

i.e.

, in 2004) to be 233 million mattresses using a ten-year average product life, and 302.6 million mattresses using a fourteen-year average product life. [8&10]

According to industry sources, queen size mattresses are the most commonly used. In 2002, queen size mattresses were used by 34 percent of U.S. consumers. Twin and twin XL are used

by 31.2 percent of U.S. consumers, followed by full and full XL (21 percent), king and California king (11 percent), and all other sizes (2.6 percent). The average manufacturing price in 2002 was $152 for a mattress and $86 for a foundation. Thus, the average manufacturing price of a mattress/foundation set was about $238 in 2002. Although there are no readily available data on average retail prices for mattress/foundation sets by size, ISPA reports that sets selling under $500 represent 40.7 percent of the market. Sets selling for between $500 and $1000 represent 39.2 percent of the market. [10]

The top four manufacturers of mattresses and foundations operate about one-half of the 639 U.S. establishments producing these products. The remainder of the establishments are operated by smaller firms. According to the Statistics of U.S. Businesses Census Bureau data for 2001, there were 557 mattress firms operating these 639 establishments. According to the same data source, all but twelve mattress firms had less than 500 employees. If one considers a firm with fewer than 500 employees to be a small business, then 97.8 percent (557-12/557) of all mattress firms are small businesses. [9&10] The potential impact of the proposed standard on these small businesses is discussed in section M of this document.

D. Risk of Injury

Annual estimates of national fires and fire losses involving ignition of a mattress or bedding are based on data from the U.S. Fire Administration's National Fire Incident Reporting System (“NFIRS”) and the National Fire Protection Administration's (“NFPA”) annual survey of fire departments. The most recent national fire loss estimates indicated that mattresses and bedding were the first items to ignite in 19,400 residential fires attended by the fire service annually during 1995-1999. These fires resulted in 440 deaths, 2,230 injuries and $273.9 million in property loss each year. Of these, the staff considers an estimated 18,500 fires, 440 deaths, 2,160 injuries, and $259.5 million property loss annually to be addressable by the proposed standard. Addressable means the incidents were of a type that would be affected by the proposed standard solely based on the characteristics of the fire cause (

i.e.

, a fire that ignited a mattress or that ignited bedclothes which in turn ignited the mattress). For example, an incident that involved burning bedclothes and occurred in a laundry room would not be considered addressable. [3]

Among the addressable casualties, open flame fires accounted for about 140 deaths (32 percent) and 1,050 injuries (49 percent) annually. Smoking fires accounted for 210 deaths (48 percent) and about 640 injuries (30 percent) annually. Children younger than age 15 accounted for an estimated 120 addressable deaths (27 percent) and 500 addressable injuries (23 percent) annually. Adults age 65 and older accounted for an estimated 120 addressable deaths (27 percent) and 250 addressable injuries (12 percent) annually. [3]

E. NIST Research

Overview.

NIST has conducted extensive research on mattress/bedding fires for SPSC and the Commission. SPSC sponsored several phases of research at NIST to gain an understanding of the complex fire scenario involving mattresses and to develop an effective test method to evaluate a mattress's performance when it is exposed to an open flame ignition source. The first phase of the research program, known as Flammability Assessment Methodology for Mattresses, involved four main objectives: (1) Evaluating the behavior of various combinations of bedclothes, (2) characterizing the heat impact imposed on a mattress by bedclothes, (3) developing burners to simulate burning bedclothes, and (4) testing the burners on different mattress designs to ensure their consistency. NIST's findings, published in NISTIR 6498, established the basis for an appropriate test method and the next phase of the research program. [2]

3

3

NIST publications can be found at NIST's Web site, (

http://fire.nist.gov/bfrlpubs/

).

Phase 2 of the NIST research focused on (1) analyzing the hazard by estimating the peak rate of heat release from a mattress with an improved design, (2) measuring a burning mattress's ability to involve nearby items in the room, and (3) assessing (in a limited way) bedclothes and their contribution to mattress fire hazards. This testing used mattresses with improved flammability properties while the flammability properties of bedclothes remained unchanged. [2] The findings from Phase 2 are detailed in NIST Technical Note 1446, Estimating Reduced Fire Risk Resulting from an Improved Mattress Flammability Standard.

Bedclothes.

During phase 2, NIST conducted tests on twin and king size mattresses with corresponding size bedclothes. In some tests, the bedclothes contributed up to 400 kW to the fire. NIST had previously estimated that a heat release rate that may cause flashover for an ordinary sized room is about 1000 kW. Thus, a mattress that contributes more than 500 kW at the same time as bedclothes are contributing 400 kW could lead to flashover. NIST conducted additional tests concerning bedclothes for CPSC, which are discussed later in this section. [2]

Other objects in the same room.

Part of the NIST study assessed the potential of a bed fire to ignite other objects in the same room. Other objects become involved by either direct flame impingement or by fire generated radiation. Although the location of objects in a bedroom is highly variable, their potential involvement is significantly influenced by their shape and properties relating to ease of ignition. NIST concluded from this research that further reducing the heat release rate from the bed could reduce the potential for ignition of other objects and therefore reduce their contribution to the overall heat release rate. [2]

Modeling.

NIST used fire modeling to explore the effect that heat and toxic gases from bed fires can have throughout a home. Fire modeling is an analytical tool that uses mathematical calculations to predict real-world fire behavior. NIST used this modeling to corroborate test data exploring the predicted levels of heat and toxic gases for the room of origin and outside the room of origin. The modeling suggested that untenable fire conditions would occur within the room, with little difference between a small and large room, at 10 minutes and 25 MJ. [2]

Gas burners' correspondence to bedclothes.

In addition to the research discussed above, NIST conducted separate studies for CPSC. One series of tests evaluated improved mattress designs and further supported the correlation between full scale mattress tests with the NIST gas burners and actual bedclothes. The study, NISTIR 7006-Flammability Test of Full-Scale Mattresses: Gas Burners Versus Burning Bedclothes, found that mattress designs showing good performance when tested with burners also exhibited significantly improved performance when tested with burning bedclothes. [2]

Interaction between mattresses and bedclothes.

NIST's work for CPSC also reinforced observations from previous NIST research on the interaction between the mattress and bedclothes. NISTIR 7006. Tests on improved mattress designs with burning bedclothes as the ignition source tend to have two distinct heat release rate peaks. The first peak is predominantly

from the burning bedclothes, while the second is predominantly from the mattress and foundation. In tests of good performing mattress designs, NIST found the second peak (

i.e.

, from the mattress/foundation) to be comparable or lower than the first peak (

i.e.

, from the bedclothes) and to occur appreciably later. [2]

Mattress size.

In another study conducted by NIST in 2004 for CPSC, NIST found that a twin size mattress made in a design that yields a very low peak heat release rate (less than 50 kW) tested with gas burners behaves essentially the same as a queen or king size mattress of the same design. Mattress designs that yield a moderate heat release rate peak (greater than 100 kW, but less than 200 kW) tend to behave the same for the first 30 minutes in twin size and king size. After ignition with the burners, the fire is localized (

i.e.

, its spread is limited) and is not affected by the mattress size. [2]

NIST evaluated the same mattress designs and sizes with burning bedclothes. NIST found the mattress size to have an apparent effect during these tests due to the differences in the size and fuel load of bedclothes. In tests of “well performing” mattress designs with burning bedclothes, the early heat release rate peak when testing a king size mattress was triple that when testing a twin size mattress. This was driven by the burning bedclothes. Mattress designs that showed a moderate heat release rate peak when tested with gas burners resulted in more serious fires when tested with burning bedclothes, especially in king size mattresses. [2]

F. Existing Open Flame Standards

In the mattress ANPR the staff reviewed 13 existing tests or standards relevant to open flame hazards associated with mattresses/bedding. These included Technical Bulletin (“TB”) 129, TB 121, and TB 117 from California, the Michigan Roll-up Test, and Boston Fire Department (“BFD”) 1X-11, as well as standards from ASTM International (formerly the American Society for Testing and Materials) (ASTM E-1474 and ASTM E-1590), Underwriters Laboratories (UL 1895 and UL 2060), the National Fire Protection Association (NFPA 264A and NFPA 267) and the United Kingdom (British Standard (“BS”) 6807 and BS 5852). 66 FR 51886.

As directed by California Assembly Bill 603, California's Bureau of Home Furnishings and Thermal Insulation (“CBHF”) adopted Technical Bulletin 603 (“TB 603”), an open flame fire standard for mattresses and mattress/box spring sets and futons. TB 603 is expected to go into effect January 1, 2005 and applies to items manufactured for sale in California. The California standard incorporates the same test methodology as the Commission's proposed standard. Both are based on testing and research conducted at NIST. Both TB 603 and the Commission's proposed standard require that mattresses not exceed a 200 kW peak heat release rate during the 30 minute test. However, the Commission's proposed standard requires that mattresses not exceed a total heat release of 15 MJ in the first ten minutes of the test, while TB 603 sets test criteria of 25 MJ in the first 10 minutes of the test.

The Commission believes that the research NIST has conducted, discussed above, establishes the most appropriate basis for an open flame mattress standard. Several comments on the ANPR also expressed this view (see section J of this document).

G. The Proposed Standard

1. General

The proposed standard sets forth performance requirements that each mattress/set must meet before being introduced into commerce. The test method is a full scale test based on the NIST research discussed above. The mattress specimen (a mattress or mattress and foundation set, usually in a twin size) is exposed to a pair of T shaped propane burners and allowed to burn freely for a period of 30 minutes. The burners were designed to represent burning bedclothes. Measurements are taken of the heat release rate from the specimen and energy generated from the fire. The proposed standard establishes two test criteria, both of which the mattress/set must meet in order to comply with the standard: (1) The peak rate of heat release for the mattress/foundation set must not exceed 200 kW at any time during the 30 minute test; and (2) The total heat release must not exceed 15 MJ for the first 10 minutes of the test.

2. Scope

The proposed standard applies to mattresses and mattress and foundation combinations sold as sets. Mattress is defined, as it is in the existing mattress standard at 16 CFR 1632, as “a resilient material or combination of materials enclosed by a ticking (used alone or in combination with other products) intended or promoted for sleeping upon.” The proposed standard lists several types of mattresses that are included in this definition (

e.g.

, futons, crib mattresses, youth mattresses). It also refers to a glossary of terms where these items are further defined.

Specifically excluded from the definition of mattress are mattress pads, pillows and other top of the mattress items, upholstered furniture which does not contain a mattress, and juvenile or other product pads. Mattress pads and other top of the bed items may be addressed in the Commission's rulemaking on bedclothes.

Like the Commission's existing mattress standard, the proposed standard allows an exemption for one-of-a-kind mattresses and foundations if they are manufactured to fulfill a physician's written prescription or manufactured in accordance with comparable medical therapeutic specifications.

3. Test Method

The proposed standard uses the full scale test method developed by NIST in the course of its research. Based on the NIST work, the Commission believes that a full scale test is necessary because of the complexities of mattress construction. Testing individual components will not necessarily reveal the likely fire performance of the complete mattress.

Under the proposed standard, the specimen (a mattress and corresponding foundation if they are to be offered for sale together as a set) is exposed to a pair of T-shaped gas burners. The specimen is to be no smaller than twin size, unless the largest size mattress or set produced of that type is smaller than twin size, in which case the largest size must be tested.

The burners impose a specified local heat flux simultaneously to the top and side of the mattress/set for a specified period of time (70 seconds for the top burner and 50 seconds for the side burner). The burners were designed to represent the local heat flux imposed on a mattress by burning bedclothes. The heat flux and burner duration were derived from data obtained from burning a wide range of bedding items. As discussed above, NIST test results using the burners have been shown to correlate with results obtained with bedclothes.

The proposed standard allows the test to be conducted either in an open calorimeter or test room configuration. Tests have shown that either configuration is acceptable. Although room effects (

i.e.

, the size and characteristics of the room) can be a factor in mattress flammability performance, test data show that room effects do not become an issue until a fire reaches about 300 to 400 kW.

Because the proposed standard limits the peak rate of heat release to 200 kW, room effects should not be an issue in the test. Preliminary analysis of data from the inter-laboratory study (discussed in section I) does not suggest any significant differences between tests based on either test configuration. The NIST test method allowed a third test configuration, essentially a smaller test room than described in the proposed standard. However, in addition to safety concerns, using the burners in the smaller size room is awkward. Only one laboratory in the country uses this configuration. Therefore, the Commission decided to propose only the two configurations.

4. Test Criteria

The proposed standard establishes two test criteria that the specimen must meet to pass the test. The peak rate of heat release must not exceed 200 kW at any time during the 30 minute test, and the total heat release must not exceed 15 MJ during the first 10 minutes of the test. [2&8]

Setting the peak rate of heat release limit at 200 kW (during the 30 minute test) ensures a less flammable mattress, reducing the contribution from the mattress, while taking into account that bedclothes and other room contents are likely to contribute to the fire. Numerous technologically feasible mattress designs are available that can meet the 200 kW criterion. Limiting the peak rate of heat release represents a significant improvement in performance compared to the 16 CFR part 1632 cigarette ignition standard for mattresses and will have the most impact on available escape time. A peak rate of heat release lower than 200 kW could limit the mattress design approaches that would meet the standard, thus increasing costs. [2&8] We note that California's TB 603 also prescribes a 200 kW peak rate of heat release.

The proposed standard requires that the total heat release in the first 10 minutes of the test must not exceed 15 MJ. This early limit ensures that the mattress will have little involvement in the fire initially and provides a substantial increase in escape time by slowing the rate of fire growth and severity. The mattress's initial performance is important because if the mattress becomes significantly involved in the early stages of the fire, this will greatly limit the time a person has to escape. [2]

The proposed 15 MJ limit in the first 10 minutes takes into account that bedclothes, and possibly other items, will be burning during this initial period and will contribute significantly to the fire. The Commission believes that the types of ticking (i.e, the outermost fabric or material that covers the mattress) currently used on mattresses can continue to be used with the 15 MJ/10 minute criteria. [2] This will allow manufacturers considerable flexibility in their mattress designs because they should be able to change tickings without affecting the mattresses performance under the test method, except in the unusual case where the ticking itself is part of the fire resistance design.

California's TB 603 prescribes a 25 MJ limit in the first 10 minutes of the test. However, NIST research, supported by fire modeling, has shown that untenable fire conditions can occur in a room from a fire producing 25 MJ in the first 10 minutes of a test. This represents the total contribution from all possibly involved items. That is, a fire that reaches a size of 25 MJ within 10 minutes could limit a person's ability to escape the room. According to the mattress industry and available test data, there are numerous technologically feasible approaches to mattress designs for meeting the proposed 15 MJ /first 10 minute limit. [2]

The 30 minute test duration is related to, but not equivalent to, the estimated time required to permit discovery of the fire and allow escape under typical fire scenarios. A mattress complying with the proposed criteria under the 30 minute test is estimated to provide an adequate time for discovery of and escape from the fire under certain conditions or assuming the bedclothes do not contribute to the extent of posing a hazardous condition early in the fire. Compared to current scenarios, this is a substantial increase in estimated escape time. The effectiveness of the estimated escape time is based on timely escape from the potentially hazardous conditions. [2&3]

Multiple test results indicate that a large number of mattress designs (using a range of fire retardant barrier technologies) can perform well in tests with gas burners for 30 minutes. Many of the tested designs are able to meet the proposed test criteria for 30 minutes, but perform erratically after 30 minutes. The number of failures, test variability, and performance unreliability increases after 30 minutes. A substantial range of technologically feasible and viable solutions and design choices exist that meet the proposed test criteria for 30 minutes. [2] We note that California's TB 603 also includes a 30 minute test duration.

The Commission considered proposing a 60 minute test duration. However, as discussed above, after 30 minutes, test variability increases, costs increase, and substantially fewer technologically feasible design approaches are available to meet the test. Most importantly, it is unclear from available data that much additional benefit would accrue with a 60 minute test.

5. Prototype Testing

The proposed standard requires, with certain exceptions, that mattress manufacturers must test specimens representative of their mattress/set prototype (design) before introducing a mattress/set into commerce. Mattresses then produced based on the prototype mattress must be identical in all material aspects of their components, materials, and method of construction to the prototype. The term “manufacturer” is defined as “an individual plant or factory at which mattresses and/or mattress and foundation sets are manufactured or assembled.” The definition includes importers. As in the existing mattress standard (16 CFR part 1632), this definition refers to the establishment where the mattress is produced or assembled, not the company. Thus, the plant or factory producing or assembling the mattress/set is required to conduct prototype testing. This is also true for importers. However, there are three exceptions to this requirement.

A manufacturer is allowed to sell a mattress/set based on a prototype that has not been tested if the prototype differs from a qualified prototype (one that has been tested and meets the criteria) only with respect to: (1) The mattress/foundation size (e.g twin, queen, king etc.); (2) the ticking, unless the qualified ticking has characteristics that are designed to improve the mattress's test performance; and/or (3) any other component, material or method of construction, provided that the manufacturer can show, on an objectively reasonable basis, that such change will not cause the prototype to exceed the specified test criteria. The third numbered option allows a manufacturer to construct and test a “worst case” prototype and rely on it to cover a range of related designs without having to perform additional testing. If a manufacturer chooses to take this approach, he/she must maintain records documenting that the change(s) will not cause the prototype to exceed the test criteria (

see

§ 1633.11(b)(4) of the proposed rule).

When conducting prototype testing, the manufacturer must test a minimum of three specimens of the prototype in accordance with the test method

described, and all of the mattresses/sets must meet both of the test criteria discussed above. If any one prototype specimen that the manufacturer tests fails the specified criteria, the prototype is not qualified (even if the manufacturer chooses to test more than three specimens).

The Commission believes that three specimens is the appropriate minimum number for testing. Numerous research studies have typically used replicates of three for tests using the developed gas burners. This is also the number industry has generally used as it has researched and developed options for meeting the requirements of California's TB 603. Preliminary analysis of the inter-laboratory study also indicates that three replicates are appropriate to accurately characterize mattress performance. [2] Moreover, because small changes in mattresses' construction or components can affect their flammability, testing more than one mattress will provide a better indication of their performance. [1]

6. Pooling

The proposed standard allows for one or more manufacturers to rely on a given prototype. Under this approach, one manufacturer would conduct (or cause to be conducted) the full prototype testing required (testing three prototype specimens), obtaining passing results, and the other manufacturer(s) may then produce mattresses/sets represented by that prototype so long as they conduct one confirming test on a specimen they produce. If the mattress/set fails the confirming test, the manufacturer must take corrective measures, and then perform a new confirmation test that must meet the test criteria. If a confirmation test specimen fails to meet the test criteria, the manufacturer of that specimen must also notify the manufacturer of the pooled prototype about the test failure. Pooling may be used by two or more plants within the same firm or by two or more independent firms. As discussed in the initial regulatory flexibility analysis, pooling should reduce testing costs for smaller companies. Once they have conducted a successful confirmation test, pooling firms can produce mattresses based on a pooled prototype and may continue to do as long as any changes to the mattresses based on the pooled prototype are limited to the three discussed above: (1) Size of the mattress/foundation; (2) the ticking, unless the qualified ticking has characteristics that are designed to improve the mattress's test performance, and/or (3) any component, material or method of construction that the manufacturer can show (on an objectively reasonable basis) will not cause the prototype to exceed the specified test criteria.

7. Quality Assurance Requirements

Research and testing indicates that small variations in construction of a mattress/set (

e.g.

missed stitching around the side of the mattress) can affect the fire performance of a mattress. Therefore, the proposed standard contains strict requirements for quality assurance. Each manufacturer must implement a quality assurance program to ensure that the mattresses/sets it produces are identical in all material respects to the prototype on which they are based. This means that at a minimum, manufacturers must: (1) Have controls in place on components and materials to ensure that they are identical to those used in the prototype; (2) designate a production lot that is represented by the prototype; and (3) inspect mattresses/sets produced for sale. The Commission is not requiring manufacturers to conduct testing of production mattresses. However, the Commission recognizes the value of such testing as part of a quality assurance program. Therefore, the Commission encourages manufacturers to conduct random testing of mattresses/sets that are produced for sale.

8. Recordkeeping

The proposed standard requires manufacturers to maintain certain records to document compliance with the standard. This includes records concerning prototype testing, pooling and confirmation testing, and quality assurance procedures and any associated testing. The required records must be maintained for as long as mattresses/sets based on the prototype are in production and must be retained for three years thereafter.

The purpose of these recordkeeping requirements is to enable manufacturers to keep track of materials, construction methods and testing. Thus, if a manufacturer produced a mattress/set that failed to meet the test criteria, he/she should be able to use the records to determine the prototype on which the failing mattress was based, as well as the components and method of construction that were used. This information would help the manufacturer correct the problem that caused the mattress to fail the test criteria.

9. Other Requirements: Labeling, One of a Kind Exemption, and Policy on Renovation of Mattresses

Under the proposed standard, each mattress/set must bear a permanent label stating the name and location of the manufacturer, the month and year of manufacture, the model identification, prototype identification number, and a certification that the mattress complies with the standard. By placing the certification on the mattress, the manufacturer is attesting that the specific mattress would comply with the test criteria if tested.

The proposed standard allows an exemption for a one-of-a-kind mattress/set if it is manufactured in response to a physician's written prescription or manufactured in accordance with comparable medical therapeutic specifications.

Subpart C of the proposed standard restates the policy clarification on renovation of mattresses that is in Subpart C of the existing mattress standard (16 CFR Part 1632). The policy statement informs the public that mattresses renovated for sale are considered by the Commission to be newly manufactured for purposes of the requirements of the proposed standard.

H. Effectiveness Evaluation

To determine the potential effectiveness of the proposed standard, CPSC staff conducted an effectiveness evaluation, focusing primarily on reduction of deaths and injuries. The staff's analysis is explained in detail in the memorandum “Residential Fires Involving Mattresses and Bedding.” [3] The evaluation was based primarily on review of CPSC investigation reports that provided details of the occupants' situations and actions during the fire. Staff reviewers identified criteria that affected the occupants' ability to escape the fires they had experienced. The staff used these criteria to estimate percentage reductions in deaths and injuries expected to occur under the much less severe fire conditions anticipated with improved designs of mattresses that would comply with the proposed standard. The staff then applied these estimated reductions to national estimates of mattress/bedding fire deaths and injuries to estimate numbers of deaths and injuries that could be prevented with the proposed standard. [3]

As stated in section D of this document, the most recent national fire loss estimates indicated that mattresses and bedding were the first items to ignite in 19,400 residential fires attended by the fire service annually

during 1995—1999. These fires resulted in 440 deaths, 2,230 injuries and $273.9 million in property loss each year. Of these, the staff considers an estimated 18,500 fires, 440 deaths, 2,160 injuries, and $259.5 million property loss annually to be addressable by the proposed standard (

i.e.

, of the type that the proposed standard could affect based on the characteristics of the fire). [3]

Overall, CPSC staff estimates that the proposed standard may be expected to prevent 80 to 86 percent of the deaths and 86 to 92 percent of the injuries presently occurring in addressable mattress/bedding fires attended by the fire service. Applying these percentage reductions to 1998-2002 estimates of addressable mattress/bedding fire losses, staff estimates potential reductions of 310 to 330 deaths and 1,660 to 1,780 injuries annually in fires attended by the fire service when all existing mattresses have been replaced with mattresses meeting the new standard. There may also be reductions in property damage resulting from the proposed standard, but data are not sufficient for the staff to quantify this impact. [3]

I. Inter-Laboratory Study

An inter-laboratory study was conducted with the support of the SPSC, NIST, and participating laboratories to explore the sensitivity, repeatability, and reproducibility of the NIST test method. All of the participating labs conducted multiple tests of eight different mattress designs. The mattress designs varied critical elements (

e.g.

, the barrier—sheet or high-loft, the type of mattress—single or double-sided) and the style of mattress (

e.g.

, tight or pillow top). [2]

Preliminary analysis of the data does not suggest either unreasonable sensitivities (

i.e.

, significantly different test results when minor variations in test procedure are made) or practical limitations in the test protocol. The preliminary analysis suggests that some mattress designs perform more consistently than others. The type of barrier appears to have a significant impact on the performance and repeatability of performance of all mattress designs tested. However, the uniformity of other components and the manufacturing process can also affect the variability in fire performance. [2]

The inter-lab tests also appear to confirm earlier observations that mattresses constructed with currently available barrier technologies are able to limit the fire severity for a substantial but not indefinite time. Most of the tested mattress designs could meet the proposed requirements if the test ended at 30 minutes, but appeared to perform erratically after 30 minutes. [2]

The preliminary analysis, supported by earlier data, suggests that significant variability exists among currently available mattress designs. Although products appear to be moving toward consistency, manufacturers clearly need to control components, materials, and methods of construction. Thus, quality assurance measures, as required in the proposed rule, are important. [2]

The inter-lab study was only recently completed, and the discussion above is based on the staff's preliminary analysis of the results. A final report on the inter-lab study is expected by the end of 2004 and will be available to the public.

J. Response to Comments On the ANPR

On October 11, 2001, the Commission published an ANPR in the

Federal Register

. 66 FR 51886. During the comment period, the Commission received sixteen written comments from businesses, associations and interested parties representing various segments of the mattress and bedding industries. After the close of the comment period, the Commission received a number of additional comments, including one from the California Bureau of Home Furnishings and Thermal Insulation urging the Commission to adopt California's TB 603 as a federal standard. Significant issues raised by all of these comments are discussed below. [14&15]

Mattress Comments

1.

Comment.

Commenters agree that the hazards associated with mattress fires appear to be clearly identified. All of the commenters support the need for an open flame standard for mattresses and initiation of federal rulemaking.

Response.

CPSC agrees that mattress and bedding fires continue to be one of the major contributors to residential fire deaths and civilian injuries among products within CPSC's jurisdiction. The most recent national fire loss estimates indicate that mattresses and bedding were the first items to ignite in 19,400 residential fires attended by the fire service annually during 1995—1999. These fires resulted in an estimated 440 deaths, 2,230 injuries, and $273.9 million property loss annually. In these fires, the bedclothes are most frequently ignited by a small open flame source. The burning bedding then creates a large open-flame source igniting the mattress and creating dangerous flashover conditions, the point when the entire room and its contents are ignited simultaneously by radiant heat.

The proposed standard is designed to address the identified hazard of flashover resulting from open flame ignition of mattresses, usually from burning bedclothes. Under the proposed standard, mattresses and mattress/sets are exposed to gas burners, simulating burning bedclothes. Mattresses are required to meet two performance criteria that minimize the possibility of or delay flashover for a period of time. Mattresses must not exceed 200 kW peak heat release rate during the 30 minute test, and the total heat released must be less than 15 MJ for the first 10 minutes of the test.

2.

Comment.

Most commenters endorsed the direction of the mattress flammability test development research underway at NIST and encouraged the CPSC to issue a technologically practicable, reasonable standard. More recent commenters suggest California TB 603 be adopted as the federal standard.

Response.

CPSC agrees with the technical approach suggested by the NIST research. A majority of the commenters agreed that preventing flashover from mattress fires would appropriately address the risk and that a full scale test with an ignition source comparable to burning bedclothes could achieve that objective. They strongly supported the NIST approach and discouraged the adoption of any existing standards.

Before California's adoption of TB 603, one commenter suggested using a modification of the small-scale British test, BS 5852, for smoldering and flaming ignition of upholstered furniture seating composites. However, a full-scale rather than small-scale test is generally considered the most reliable method for measuring performance of a product that contains many materials in a complex construction, such as a mattress. NIST research confirmed that a full-scale test of the mattress was needed to measure its performance when exposed to burning bedclothes or the representative set of gas burners. NIST's comprehensive, scientifically based research program was designed to address the open-flame ignition of mattresses and bedclothes under controlled conditions closely resembling those of real-life fire scenarios. The program focused on understanding the dynamics of fires involving mattress and bedclothing assemblies and on developing an appropriate and technologically practicable methodology to effectively measure the hazard.

NIST subsequently prepared a test method which the state of California incorporated into their TB 603,

“Requirements and Test Procedure for Resistance of a Mattress/Box Spring Set to a Large Open-Flame” in 2004. The proposed standard is also based on the test method developed by NIST. Research on mattress and bedclothes fires conducted by NIST for CPSC and the industry provides the basis for the test criteria specified in the proposed standard. Manufacturers and suppliers have demonstrated that mattress designs complying with these performance criteria and suitable for the residential market can be produced.

3.

Comment.

One commenter requested the exclusion of certain product categories, such as mattresses used for therapeutic reasons and in healthcare environments, from an open flame standard.

Response.

The proposed standard includes all mattresses, including those used in or as part of upholstered furniture items. “One-of-a-kind” mattresses and foundations are defined as physician prescribed mattresses to be used in connection with the treatment or management of a named individual's physical illness or injury. These products may be exempted from testing under the proposed standard in accordance with the rules established by the Commission. The proposed standard requires them to be permanently labeled with a warning statement indicating that the mattress and foundation have not been tested under the standard and may be subject to a large fire if exposed to an open flame.

4.

Comment.

In October 2003 the California Bureau of Home Furnishings (CBHF) urged the Commission to adopt their new standard, TB 603. Subsequently, a number of commenters expressed written support for adopting the TB 603 test methodology and performance criteria.

CBHF claimed that harmonization of California and federal standards would avoid a number of potential problems. They noted potential problems such as possible federal preemption and negative impacts on interstate commerce. Since TB 603 is a newly developed methodology, CBHF suggested that an inter-laboratory study be conducted before a potential adoption of TB 603 by CPSC. They noted that data obtained from an inter-laboratory study would verify the credibility of the test method.

Response.

An inter-laboratory study was conducted with the support of SPSC, NIST, CBHF, and other participating laboratories to collect additional data and confirm the test protocol developed by NIST. A number of laboratories participated in the study to evaluate sensitivity, repeatability, and reproducibility of the test protocol. While the final report is not yet available, preliminary analysis of the data does not suggest either unreasonable sensitivities or practical limitations in the test protocol.

The Commission's proposed standard is similar to California's TB 603. The proposed standard and TB 603 use the same test method and limit the peak rate of heat release of a mattress or mattress/foundation to 200 kW. TB 603 also limits the size of the fire produced in the first 10 minutes of the test to 25 MJ. According to NIST research, untenable fire conditions could occur in a room from a fire of this size. Unlike TB 603, the staff's draft proposed standard requires that the mattress contribute no more than 15 MJ to the early fire scenario. This ensures that the mattress will have little involvement in the fire for the specified period of time. This lower limit partially compensates for the contribution of an uncertain combination of burning bedclothes on the bed, helping to preserve tenable conditions for egress.

5.

Comment.

Two commenters recognize the sophistication and complexity of the test method used in California TB 603 and potentially in a federal standard. They suggest that CPSC explore laboratory accreditation programs to insure test labs are properly qualified to conduct this complex test.

Response.

The interlaboratory study may identify laboratory practices, equipment, and other related factors that must be controlled to ensure consistent and accurate test results. The report and findings of the study will be available to the public; and appropriate guidance can be provided to interested laboratories. While accrediting test laboratories is not a CPSC function, the Commission supports industry and commercial laboratory development of such a program.

6.

Comment.

A commenter expressed concerns about environmental impact and consumer sensitivity to flame retardants that may be used in mattresses, whether topically applied or integrated into fibers. The commenter recommends requiring a label that discloses the use of flame retardants in the mattress and provides a source of more specific information.

Response.

Mattress fire performance can be improved by incorporating fire retardant chemicals into component materials or by using materials that are inherently fire resistant. Flame retardant chemicals are already widely used in other applications. More than one billion pounds of different flame retardant chemicals are currently used annually in the United States, including applications in many consumer products. There are also flame resistant (FR) materials that may be used for mattress barriers that have other consumer product applications. For example, melamine resins, which can be used in FR barriers, are also used in many laminated counter tops.

Based on available data, the Commission believes that there are available options for meeting the standard without posing an unacceptable health risk to consumers or significantly affecting the environment. Moreover, as described in section N of this preamble, even if a method used by some manufacturers to meet the standard were suspected of posing an unacceptable risk, there would be regulatory and other mechanisms that can be used to control that particular method. The staff is planning to conduct migration and exposure studies on various FR chemicals that could be used to meet the standard.

The commenter suggested labeling of chemically treated components as a possible requirement of the standard, to inform consumers of the materials used. The Commission questions whether such information would be of practical value to consumers. Simply stating that a mattress component has been chemically treated does not indicate to the consumer whether the mattress poses any health risk or not. The proposed standard requires manufacturers to maintain records specifying details of all materials, including flame retardant treatments applied and inherently flame resistant materials, used in each mattress design (prototype). This will allow identification of relevant mattresses and mattress/sets if an unacceptable risk is identified.

7.

Comment.

Another commenter recommended test provisions in the standard that address the long term durability of the flame retardant chemicals used in mattresses to ensure they continue to meet the fire performance requirements.

Response.

It is expected that most manufacturers will use some kind of flame resistant barrier material to protect the mattress components with the greatest combustible fuel load from exposure to an open flame. Flame resistant barriers for mattresses may take several forms, including ticking fabrics, woven and non-woven interlinings, and battings. It is likely that these barriers will be made with an inherently flame resistant fiber (

e.g.

, para-aramid or fiberglass) or by treatment with flame retardant chemicals, many of which are incorporated within the fiber, foam, or

other material. At this point in the development of technologies that may be used to meet TB 603 or the proposed standard, the staff has seen no evidence that suggests that changes in these materials over time will occur or affect fire performance.

8.

Comment.

One commenter expressed concerns about the potentially severe economic impact of a federal regulation, similar to TB 603, on small businesses.

Response.

The Commission acknowledges that the cost of testing, record keeping, and quality control/quality assurance programs could be disproportionately higher for small businesses. While these costs are estimated to be a little over one dollar per mattress per year for average-sized establishments, they could be substantially higher for some small mattress producers. The proposed standard, however, allows manufacturers to pool their prototype qualification and testing, and thus these costs can be mitigated. Moreover, if manufacturers produce mattress/set constructions for longer than a year or use a worst-case prototype to represent other mattress constructions, these costs will be lower. It is also expected that some barrier suppliers or independent laboratories would be willing to do the testing and quality control/assurance programs for small producers in exchange for a small charge. Therefore, the proposed standard is expected to minimize the impact on small businesses, while maintaining the benefits resulting from the standard.

The Commission is requesting comments from small businesses on the expected economic impact of the requirements of the proposed standard and the proposed effective date of 12 months after publication of the final rule in the

Federal Register

.

9.

Comment.

One commenter reported that some juvenile or crib mattresses, while meeting the 200 kW peak rate of heat release requirement, produce large amounts of flaming droplets that have the potential for spreading flames beyond the mattress. TB 603 does not address these flaming droplets.

Response.

The objective of the proposed standard is to reduce the size of mattress/bedding fires and, thereby reduce the likelihood of or delay the development of flashover conditions in the room. Based on research conducted by NIST, performance criteria were developed to limit the size of the mattress fire and reduce the likelihood of it involving other objects in the room. The Commission believes that, while the proposed standard may be less effective in isolated circumstances, the objective of the standard can be met with the performance criteria specified: maximum 200 kW peak heat release rate during the 30-minute test and maximum 15 MJ total heat release in the first 10 minutes of the test. Laboratory tests of currently marketed crib mattresses of which the Commission is aware show unacceptable performance in one or both of these fire performance measures. Like full-size mattresses, these crib mattresses would also need to be improved to meet the requirements of the proposed standard.

10.

Comment.

One commenter suggested that a 60-minute test duration is needed in the standard to allow for fire and rescue workers to respond and help occupants escape.

The commenter notes that the longer test time will allow emergency responders to assist vulnerable citizens to escape fires involving mattresses and bedding. They report that response times can vary widely among local circumstances, from approximately 16 minutes to an hour or more.

Response.

To estimate the proposed standard's potential effectiveness, the staff reviewed in-depth investigations that provided detailed information about fires that ignited mattresses and bedding, details of the occupants' situation, and occupants' actions during the fire. Most investigations also included documentation from the fire department that attended the fire. The in-depth investigations involved fires occurring during 1999-2004, and included a total of 195 deaths and 205 injuries. In some of these cases, even with traditional mattresses and bedding, other members of the household present at the time of the fire and emergency responders arriving within as little as 5 minutes were able to rescue victims.

With improved mattresses, those complying with the 30-minute test specified in the proposed standard, the fire growth is slowed considerably and flashover conditions are delayed, making successful rescue efforts of family members and emergency responders more likely. The Commission estimates that 310 to 330 deaths and 1,660 to 1,780 injuries resulting from mattress and bedding fires could be prevented annually by the proposed standard. A maximum additional 80 deaths and 280 injuries, considered addressable by the draft standard, might be further reduced with a 60-minute test. However, actual reductions would likely be much lower. This is because those considered likely to die or be injured in conditions associated with a proposed 30 minute test are those incapable of acting on their own and with no potential rescuer in the occupancy. Even with more time, in such circumstances, the fire continues to progress, and the chances of rescue are unpredictable.

Based on the preliminary regulatory analysis, the expected benefits of the proposed standard, incorporating a 30-minute test, are greater than the costs. The regulatory analysis also considered alternatives to the proposed standard, including a 60-minute test; neither this nor the other alternatives was shown to increase expected net benefits.

11.

Comment.

A few commenters expressed the need to maintain protection from the threat from cigarette ignitions while considering an open flame standard.

Response.

The standard that addresses cigarette ignition resistance, the Standard for the Flammability of Mattresses and Mattress Pads, codified as 16 CFR part 1632, remains in effect unless it is modified or revoked by the Commission in a separate rulemaking proceeding. If such a rulemaking occurred, the Commission would thoroughly evaluate the need for maintaining both an open flame standard and the standard for cigarette ignition resistance.

Bedclothes Comments

Comment.

Most of the commenters refer to the impact of burning bedclothes on mattress/bedding fires and express opinions on the potential scope of an open flame mattress standard. Some commenters urge the Commission to limit the scope of a standard to mattresses while opposing commenters recommend that either the scope be expanded to incorporate bedclothes or bedclothes should have ignition standards of their own.

Commenters in support of regulating bedclothes believe that studying the impact of burning bedclothes is appropriate and would assist in the development of better performing, safer products. They note that bedclothes contribute to the intensity and spread of the original ignition source often involved in mattress fires. Therefore, burning bedclothes become a significant ignition source to the mattress and impact the burning characteristics of the mattress and foundation. They further note that bedclothes alone have been shown to generate a fire large enough to pose a hazard and can alone be the cause of ignition to nearby items. According to these commenters, improving the flammability of certain bedding items, such as filled items, is economically feasible. One commenter claims that mattress fires cannot be adequately addressed without also

considering the flammability of bedclothes.

In support of limiting the scope to mattresses and not regulating bedclothes, some commenters identify bedding items as an uncontrolled variable. They claim that there is no way to predict the type of bedclothes that may be involved in an incident at any given time; the number and type of items used by consumers is indefinable and consumers select items based on season, fashion, and climate. In addition, according to these commenters, there is no objective method to determine if consumers would use regulated bedclothes; there is little data to suggest that regulating some selected items will have an impact on the hazard; and flammability performance should not be based on what consumers may or may not use as bedclothes. These commenters also state that most U.S. textile manufacturers already voluntarily test for small open flame ignition of bedclothes using ASTM voluntary test methods. They assert that the additional burden and expense of any regulation on bedclothes would be substantial and could not be justified.

Response.

The Commission notes that bedclothes substantially contribute to the complexity and magnitude of the mattress fire hazard. In laboratory tests peak heat release rates as high as 800 kW were observed from some larger bedclothes items. This presents a clear risk of flashover; and this heat release rate is much higher than that allowed for a mattress/set in the proposed standard. The extent to which bedclothes can be modified in a manner that is technologically practicable and economically feasible is unclear at this time. However, reducing the contribution of certain high fuel load bedding items to a mattress/bedding fire is desirable. The Commission is issuing an ANPR for a bedclothes flammability standard. The Commission believes that such a standard could increase the likelihood that mattress/bedding fire losses are effectively reduced.

K. Preliminary Regulatory Analysis

The Commission has preliminarily determined to issue a rule establishing a flammability standard addressing the open flame ignition of mattresses. Section 4(i) of the FFA requires that the Commission prepare a preliminary regulatory analysis for this action and that it be published with the proposed rule. 15 U.S.C. 1193(i). The following discussion, extracted from the staff's memorandum titled “Preliminary Regulatory Analysis of a Draft Proposed Standard to Address Open-Flame Ignitions of Mattresses,” addresses this requirement. [8]

1. Introduction

There were an estimated 18,900 fires where the first item ignited was mattress/bedding in 1998 (the last year for which detailed data comparable to previous years are available). These fires caused an estimated 2,260 civilian injuries, 410 deaths, and $255.4 million in property losses. As discussed elsewhere in this document, NIST has conducted extensive research and developed a test methodology to test open flame ignition of mattresses.

California Technical Bulletin (TB) 603, which is based on the use of NIST test burners designed to mimic the local thermal insult (heat flux levels and duration) imposed by burning bedclothes, is scheduled to become effective in California January 1, 2005. TB 603 requires all mattress/foundation sets, mattresses intended to be used without a foundation, and futons to meet the following pass/fail criteria: (1) The peak heat release rate (“PHRR”) does not exceed 200 kW during the 30 minute test, and (2) the total heat release does not exceed 25 mega joules (MJ) in the first 10 minutes of the test.

Large mattress manufacturers may eventually produce TB 603-compliant mattresses for sale nationwide, because of legal liability and production logistics. In the short-run, however, some manufacturers may limit their sale of TB 603-complying mattresses to California. Sealy's president and CEO said that “[they] plan to be ready by the end of this year [2004] if a national retailer wants the same product” with fire resistant technology, but will not convert all production by January 2005 (Furniture Today, March, 10, 2004). Smaller producers are more likely to wait until they have a better idea of enforcement efforts in California, or until a federal standard is adopted. The mattress industry, represented by ISPA, supports the development of a mandatory federal standard (Furniture Today, May, 10, 2004). A Federal standard would eliminate the uncertainty that may result from having different flammability standards for different states.

2. The Proposed Standard: Scope and Testing Provisions

The proposed standard will apply to all mattresses, where the term mattress means a ticking (

i.e.

, an outer layer of fabric) filled with a resilient material used alone or in combination with other products intended or promoted for sleeping upon. This definition is discussed further in section G.2. above.

A typical innerspring mattress construction might include ticking; binding tape fabric; quilt cushioning with one or more separate layers; quilt backing fabric; thread; cushioning with one or more separate layers; flanging; spring insulator pad; spring unit; and side (border) panels. Options for meeting the standard include the use of one or a combination of the following: fire resistant ticking; chemically treated or otherwise fire resistant filling products; or a fire blocking barrier (either a sheet style barrier, sometimes called a fabric barrier, or a high-loft barrier, sometimes called a fiber barrier). The fire blocking barrier is placed either directly between the exterior cover fabric of the product and the first layer of cushioning materials, or beneath one or more “sacrificial” layers that can burn without reaching the proposed heat release constraints.

There are already over twenty different vendors of fire resistant materials associated with the production of mattresses, including barriers, ticking, foam, tape, and thread. These materials include chemically treated cotton, rayon, and/or polyester, melamine, modacrylic, fiberglass, aramid (Kevlar”), or some combination of them. The cost of using sheet barriers is higher than using high-loft barriers, since sheet barriers are thin and therefore could not be substituted for an existing foam or cushioning layer. There is also concern that some sheet barriers, unlike high-loft barriers, may reduce the comfort of the sleeping surface.

To qualify a prototype, three mattresses/sets must be tested and must pass the test requirements. To obtain a passing result, each mattress/set must pass a 30 minute test, where the PHRR does not exceed 200 kW and the total heat release does not exceed 15 MJ in the first 10 minutes of the test. If any of the sets fail, the problem must be corrected, the prototype must be retested and pass the test (in triplicate). Manufacturers may sell any mattress/set based on a qualified prototype. Manufacturers may also sell a mattress/set based on a prototype that has not been tested if that prototype differs from a qualified prototype only with respect to (1) mattress/foundation size; (2) ticking, unless the ticking of the qualified prototype has characteristics designed to improve performance on the burn test; and/or (3) any component, material, or method of construction that the manufacturer can demonstrate, on an objectively reasonable basis, will not cause the prototype to exceed the test criteria specified above.

If one or more establishments (plants within the same firm) or independent

firms choose to “pool” prototypes, then each pooling plant or firm is required to test one mattress/set for confirmation testing. If that set fails, then the plant or firm will need to test another mattress/set after correcting its production to make sure that it is identical to the original prototype.

A pooling firm may sell other mattresses that have not been tested by the pooling firm if they differ from the pooled prototype only with respect to (1) mattress/set size; (2) ticking, unless the ticking of the qualified prototype has characteristics designed to improve performance on the burn test; and/or (3) any component, material, or method of construction that the manufacturer can demonstrate, on an objectively reasonable basis, will not cause the prototype to exceed the test criteria specified above.

3. Products and Industries Potentially Affected

According to ISPA, the mattress producers' trade organization, the top four producers of mattresses account for almost sixty percent of total U.S. production. In total, there are 639 establishments (as of 2001) that produce mattresses in the U.S., using the U.S. Department of Commerce NAICS (North American Industry Classification System) Code 33791 for mattresses. The top four producers account for about half of the number of all these establishments. The number of establishments has been declining over time due to mergers and buy-outs. Total employment in the industry, using the NAICS Code 33791, was 25,500 workers in 2001.

The mattress manufacturing industry has three key supplying industries: spring and wire product manufacturing, broad-woven fabric mills, and foam products manufacturing. Depending on the type of fire resistant barrier chosen by different manufacturers, the demand for foam padding for mattresses might decline if it were replaced by the high-loft barrier in the construction of the mattress and foundation. This would be offset by an increase in the demand for the high-loft barrier. If sheet barriers were chosen by some mattress producers, then sales of, and employment by, the sheet barrier suppliers would increase. Since the sheet barriers would not replace other inputs, there would most likely be no offsetting effect on other industries. Fiberglass, melamine, and aramid producers may also be affected to the extent that they are used to produce fire resistant materials used in mattress production.

Manufacturers of bedclothes may also be affected by the proposed standard. Sales of bedclothes may increase or decrease based on whether consumers view bedclothes as complements or substitutes for a new mattress/set (complements are goods generally consumed together, substitutes generally substitute for each other). For example, if people tend to buy all parts of a new bed (mattress, foundation, and bedclothes consisting of a comforter, pillows, and sheets) at the same time, then an increase in the quantity of mattresses sold would cause an increase in sales of bedclothes. If, alternatively, people tend to have a fixed budget from which to buy all mattresses and bedding items, then an increase in the quantity of mattresses sold would lead to a decrease in sales of bedclothes. Also, if the decision to buy a new mattress (or mattress/set) involves buying a mattress that is much thicker than the one currently in use, then consumers will most likely buy new sheets (and possibly matching pillowcases and other bedclothes items) to fit the new thicker mattress.

If the cost increase is relatively small or there is no resulting increase in the price of a mattress/set, then the demand for bedclothes will only be affected if consumers place a higher value on the safer mattress and replace their current mattress sooner than they would have with no standard in place. An increased demand for the safer (and thicker, if the current mattress is relatively old) mattress will likely result in an increased demand for sheets that fit the newer mattresses. This effect, however, is not directly resulting from the adoption of the proposed standard since the thickness of the mattress need not be increased by the presence of either type of barrier. It is the result of the increased utility some consumers may derive from the safer mattress and the consequent increase in demand for bedclothes. The increased demand for safer mattresses would most probably lead to an increase in sales and employment in the spring and wire products, broad-woven fabric, and foam products industries, as well as in the mattress and bedclothes industries.

Other producers that could potentially be affected, if the price change associated with producing compliant mattresses is significant, are those of other substitute products, like airbeds, waterbeds, * * * etc. that contain no upholstered material and would, therefore, not be covered by the proposed standard. Their sales may increase as a proportion of total bedding products.

4. Characteristics of Mattresses Used in U.S. Households

The total number of U.S. conventional mattress shipments was 21.5 million in 2002 and is estimated to be 22.1 in 2003 and 22.8 in 2004. Mattress shipments have grown at an average rate of three percent over the period 1981 to 2004. Unconventional mattresses (including futons; crib mattresses; juvenile mattresses; sleep sofa inserts; and hybrid water mattresses) are estimated to be about ten percent of the total market. This yields an estimated total number of mattresses produced domestically of 25.3 million in 2004. The value of mattress and foundation shipments in 2002, according to ISPA, was $3.26 and $1.51 billion respectively.

The CPSC Product Population Model (PPM) estimate of the number of mattresses in use in different years is based on available annual sales data and an estimate of the average product life of a mattress. Industry representatives assert that the average consumer replaces a mattress/set after ten years. A 1996 CPSC market study estimated the average expected life of a mattress to be 14 years. The PPM estimates the number of (conventional and non-conventional) mattresses in use in 2004 to be 233 million mattresses, using a 10-year average product life, and 302.6 million mattresses, using a 14-year average product life. These two numbers are later used to estimate the pre-standard baseline risk and the expected benefits of the proposed standard.

This analysis focuses principally on queen-size mattresses because they are the most commonly used. In 2002 queen-size mattresses were used by 34 percent of U.S. consumers. Following the queen-size are the sizes: Twin and Twin XL (31.2 percent), Full and Full XL (21 percent), King and California King (11 percent), and all other (2.6 percent). ISPA data reflect that the average size of a mattress is increasing. The average manufacturing price in 2002 was $152 for a mattress of average size and $86 for a foundation of average size. Hence the average manufacturing price of a mattress/set was about $238 in 2002.

There are no readily available data on average retail prices for mattress/foundation sets by size. ISPA, however, reports that mattress/foundation sets selling for under $500 represent 40.7 percent of the market. Mattress/foundation sets selling for between $500 and $1000 represent 39.2 percent of the market.

5. Trends in Mattress/Bedding Residential Fires, Deaths, Injuries, and Property Losses

Open-Flame Ignition

. The staff estimates average annual mattress/bedding fires from open-flame ignitions (including candles, matches and lighters) to have been 8,367 and 6,367 over the 1993-95 and 1996-98 periods respectively. This represents a reduction of 23.9 percent. The resulting average mattress/bedding deaths, injuries, and property losses from open-flame ignitions have decreased by 28.2 percent, 22.1 percent, and 5.6 percent respectively, over the 1993 to 1998 period. When adjusted for inflation, the decrease in the value of property losses becomes 37.7 percent.

Smoking Material Ignition

. The staff estimates average annual mattress/bedding fires from smoking material ignition (including cigarettes, cigars, and pipes) to have been 7,733 and 6,067 over the 1993-95 and 1996-98 periods respectively. This represents a reduction of 21.6 percent over the 1993 to 1998 period. Average annual deaths, injuries, and property losses due to mattress/bedding smoking material ignitions have decreased by 4.7 percent, 19.7 percent, and 9.7 percent, respectively, over the same period. When adjusted for inflation, the decrease in the value of property losses becomes 40.4 percent.

Other Ignition Sources

. The staff estimates average annual mattress/bedding fires from other ignition sources (including sparks, embers, or flames escaping from fueled equipment, arcs or sparks from electric equipment, small torches, hot embers, and fireworks, heat escaping from fueled equipment, molten material, short circuit arc, and heat overloaded equipment) to have been 8,633 and 7,767 over the 1993-95 and 1996-98 periods respectively. This represents a reduction of 10 percent over the 1993 to 1998 period. Average annual injuries and inflation-adjusted property losses have decreased by 13.8 percent and 38.7 percent respectively. Average annual deaths increased by 51.7 percent (from 97 to 147). This increase offsets the decrease in deaths resulting from open-flame and smoking material ignition fires. The annual average number of deaths from all ignition sources remained unchanged over the period, equal to 510.

6. Expected Benefits of the Proposed Standard

The expected benefits of the proposed standard are estimated as reductions in the baseline risk of death and injury from all mattress fires, based on a CPSC staff study of fire investigations from 1999-2004. Risk reductions are then calculated on a per-mattress-in-use basis based on estimates of the number of mattresses in use. The monetary value of expected benefits per mattress is derived using current (

i.e.

, 2004) estimates for the value of a statistical life and the average cost of a mattress fire injury. To derive the monetary value of expected benefits over the life of a mattress, the expected annual benefits are discounted (using a three percent discount rate), and then summed over the expected life of the mattress. The analysis considers mattress lives of 10 and 14 years.

The potential benefits of the proposed standard consist of the reduction in deaths, injuries, and property damage that would result. Since the objective of the proposed standard is to reduce the likelihood of flashover or increase the time before flashover occurs, and not to reduce fires, changes in property losses associated with the proposed standard are hard to quantify. Property losses are expected to decline but the extent of the decline cannot be quantified. Consequently, for purposes of this analysis, no reduction in property losses is assumed. That is, all expected benefits from the proposed standard are in the form of prevented deaths and injuries. This underestimates net benefits, since there will likely be some benefits from reduced property losses.

The proposed standard is expected to reduce the likelihood of flashover resulting from fires started by smoking materials or other ignition sources, as well as those started by open-flame ignition. Fires, injuries, and property losses resulting from smoking material ignition and other ignitions, and deaths from smoking material ignition are lower for the 1996-98 period than the 1993-95 period. (Deaths from other ignition sources are more than 50 percent higher). Any additional reduction in these figures due to the proposed open-flame ignition standard will translate into societal benefits, as will be discussed in the benefit-cost analysis (Section 8).

Estimates of the effectiveness of the proposed standard are based on a CPSC staff evaluation of in-depth investigation reports of fires (including details of the occupants' situations and actions during the fire) occurring in 1999-2004 in which a mattress or bedding was the first item to ignite, the fire was of the type considered addressable by the proposed standard, and a civilian death or injury resulted. Most of the investigations also included documentation from the fire department that attended the fire. Some incident reports were initiated from death certificates with follow-up documentation from the fire department. This resulted in a total of 195 deaths and 205 injuries in the investigations to be evaluated. The distribution of mattress ignition sources was not representative of all fires involving mattresses and thus the data were weighted to match the NFIRS-based national fire data distributions.

Evaluations of the fire incidents by CPSC staff reviewers used the results of NIST testing (Ohlemiller, 2004; Ohlemiller and Gann, 2003; Ohlemiller and Gann, 2002) conducted to assess the hazard produced from burning mattresses and bedclothes. Specifically, the evaluations were based on the expectation that occupants in bed when the fire ignited but able to escape the burning bedclothes in the first three to five minutes faced a minimal hazard. Occupants in direct contact with burning bedclothes for a longer period (5 to 10 minutes) would be subject to potentially hazardous levels of heat release. If the burning bedclothes did not ignite other non-bedding items or produce flashover at this time, heat release would subside temporarily and then begin to increase as the involvement of the mattress increased.

These conditions would allow occupants 10 to 15 minutes to escape the room of origin before the situation in the room would become untenable. Since the proposed standard is expected to slow the rate of fire spread and hence increase escape time, assuming that bedclothes do not contribute enough heat to pose a hazardous condition, it was assumed that no deaths would occur among people who were outside the room of origin at the time of ignition, unless they entered the room later or were incapable of exiting on their own. The analysis focused on reduction of deaths and injuries because the proposed standard is designed to limit fire intensity and spread rather than prevent ignition.

Each investigation was evaluated by CPSC staff reviewers to identify the features related to the occurrence of a death or injury once the fire was ignited. These included casualty age, casualty location when the fire started (at the point of ignition, in the room of origin but not at the point of ignition, or outside the room of origin), whether the casualty was asleep, or suffered from additional conditions likely to increase the time needed to escape, whether the casualty engaged in fighting the fire, and whether a rescuer was present. All of these conditions were used to determine a range for the likelihood that each individual death or injury would have

been prevented had the draft proposed standard been in effect. Percentage reductions of deaths (injuries) within subcategories of heat source and age group were applied to equivalent subcategories of the national estimates based on the NFIRS and NFPA data for 1995-1999. The estimated reductions per category were summed and the overall percentage reductions were calculated as the percent of addressable deaths (or injuries) that would have been prevented if the likelihood of flashover were reduced in the first 30 minutes and victims had 10 to 15 minutes of escape time.

The staff indicates that the proposed standard is expected to reduce all addressable deaths from mattress/bedding fires by 80 to 86 percent and reduce all addressable injuries from mattress/bedding fires by 86 to 92 percent. The results vary only slightly by source of ignition. These estimated effectiveness percentages result in the prevention of an estimated 310 to 330 deaths and 1660 to 1780 injuries annually, for the 1998-2002 period.

The staff's analysis presents the estimated annual deaths and injuries that are expected to be prevented by the proposed standard, based on average figures for 1998-2002. For purposes of this analysis, it is assumed that the annual deaths and injuries prevented by the proposed standard equal the average annual deaths and injuries prevented for the 1998-2002 period. The analysis is conducted as if the standard went into effect in 2004. All dollar estimates are based on constant 2004 dollars. A discount rate of 3 percent and average expected lives of a mattress of 10 and 14 years are also assumed.

The estimated ranges of deaths and injuries prevented are calculated by applying the range of percent reductions to average addressable deaths and injuries for the period 1998-2002. Staff's analysis also presents the risk reduction in deaths and injuries that would result from the proposed standard (per million mattresses). Based on the estimated number of mattresses in use (described in Section 4) and an average expected life of 10 years, the annual reduction in the risk of death equals 1.33 deaths per million mattresses (310 deaths divided by the estimated 233 million mattresses in use in 2004) to 1.42 per million mattresses (330 deaths / 233 million mattresses). The estimated reduction in the risk of injury, similarly calculated, equals 7.12 to 7.64 injuries per million mattresses for an estimated 10-year life of a mattress. The estimated risk reductions for an estimated 14-year life of a mattress are 1.02 to 1.09 deaths and 5.49 to 5.88 injuries per million mattresses.

Annual risk reductions resulting from the proposed standard are used to derive the monetary benefits from reduced deaths and injuries. The estimated reduction in the risk of death is multiplied by the value of a statistical life (and divided by a million) to derive a first-year monetary estimate for the range of benefits from lives saved per mattress. Based on the existing literature, a value of a statistical life of five million dollars is assumed (Viscusi, 1993). The estimated reduction in the risk of injury is similarly used to derive the range of first-year monetary benefits from injuries prevented. The benefits from preventing an injury (the cost of an injury) in 2004 are estimated to average about $179,300, based on Miller

et. al.

(1993). The first-year benefits associated with preventing deaths and injuries equal $7.93 to $8.45 for an estimated mattress life of 10 years and $6.11 to $6.51 for an estimated mattress life of 14 years.

Lifetime benefits are derived by projecting annual benefits for the life of the mattress and summing the discounted (at a rate of 3 percent) stream of annual benefits (measured in constant dollars). The number of mattresses in use is projected to grow at a rate of zero to three percent, based on the average growth rate for the 1981-2002 period. Since the number of deaths and injuries are implicitly assumed to remain constant over time, a positive growth rate of mattresses in use implies a declining risk over time. The lower end of the ranges for estimated (10 and 14 years) lifetime benefits correspond to a 3 percent projected growth rate and the lower end of the effectiveness ranges. The upper end of the ranges for estimated (10 and 14 years) lifetime benefits correspond to a zero percent projected growth rate and the upper end of the effectiveness ranges. For an expected mattress life of 10 years, the resulting expected lifetime benefits of saved lives associated with the proposed standard equal $51.70 to $62.22 per mattress. The corresponding benefits of prevented injuries equal $9.93 to $12.03. Hence, for an expected mattress life of 10 years, the expected total lifetime benefits of a compliant mattress equal $61.66 to $74.25. For an expected mattress life of 14 years, total benefits equal $59.88 to $75.71 per mattress. The sensitivity analysis section below examines how the results might change when a discount rate of seven percent is used.

7. Expected Costs of the Proposed Standard

This section presents the expected resource costs associated with the proposed standard. Resource costs are costs that reflect the use of a resource that would have been available for other uses had it not been used in conjunction with the production of mattresses compliant with the proposed standard. These costs include material and labor costs; testing costs; costs to wholesalers, distributors, and retailers; costs of producers' information collection and record keeping; costs of quality control/quality assurance programs; and compliance and enforcement costs. The effect on retail prices will be discussed in Section 8.

Material and Labor Costs

. To comply with the proposed standard, the construction of most mattress/sets will include a barrier technology with improved fire performance. This barrier may be thick (high-loft) or thin (sheet). High-loft barriers are generally used to replace some of the existing non-woven fiber, foam, and/or batting material, leading to a smaller increase in costs than sheet barriers, which constitute an addition to production materials (and costs).

According to several barrier producers and mattress manufacturers, the price of a high-loft barrier that would make a mattress comply with the proposed standard, defined to have a width of 88 to 92 inches, is $3.00 to $5.00 per linear yard. The high-loft barrier replaces the currently-used polyester batting, which costs an average of $0.50 to $1.70 per linear yard. Hence, the net increase in the cost attributed to the use of the high-loft barrier is $1.30 to $4.50 per linear yard, which translates to a net increase in barrier-related manufacturing costs of $7.80 to $27.00 for a queen-size mattress/set.

4

The queen-size is used for all the cost estimates, because it is the mode size, used by 34 percent of consumers in 2002.

4

This calculation is based on the assumption that a queen-size mattress/set requires six linear yards of the barrier mateiral to be used in the two (top and bottom) panels of the mattress and the side panels of both the mattress and foundation. Some producers are able to use less than six linear yards, which reduces their cost per queen mattress/set.

According to several barrier producers and mattress manufacturers, the price of a sheet barrier that would make a mattress comply with the proposed standard is $4.00 to $6.00 per linear yard. Because of its different texture, the sheet barrier would generally not replace any of the materials being used in the construction of the mattress/set.

5

This translates to $24.00 to $36.00 for a queen-size mattress/set. The large difference in the net cost of the two barrier types suggests that if a barrier's fire performance is not a function of its type, most manufacturers will use high-loft barriers, the less costly alternative. A large mattress manufacturer also indicated that mattresses produced with sheet barriers in the top panel of the mattress (as opposed to the side panels) may be less comfortable.

5

The only exception to this might involve using a sheet barrier in the side panel of the mattress and

foundation. Because the existence of cushioning along the side of the mattress and foundation would probably not be noticed or missed by consumers, substitution of the sheet barrier for the material currently being used in the side panel may be implemented to reduce the cost of using the sheet barrier. The side panel is small, relative to the size of the entire surface area of a mattress/set, and its possibly different construction is therefore not included in the cost calculation. This leads to a slight over-estimation of the cost of the sheet barrier and consequently the relative cost of using a sheet instead of a high-loft barrier.

In addition to the increase in material costs due to the use of a barrier, costs will increase due to the use of fire-resistant (FR) thread for tape stitching. According to several thread producers, the cost of FR thread is $0.41 to $0.60 per queen-size mattress/set. Given that the cost of nylon (non-FR) thread is about $0.10 per queen-size mattress/set, the net increase in costs per queen-size mattress/set due to the use of FR thread is $0.31 to $0.50.

Costs may also increase due to slightly reduced labor productivity. Based on industry estimates of an average of two labor hours for the production of a queen-size mattress/set, and a 10 percent reduction in labor productivity and an industry average hourly wage rate of $11.50, the cost increase due to reduced labor productivity is about $2.30.

The increase in the materials and labor costs of a mattress, is thus equal to $10.41 ($7.80 barrier cost + $0.31 thread cost + $2.30 labor cost) to $29.80 ($27 barrier cost + $0.50 thread cost + $2.30 labor cost) for a high-loft barrier and $26.61 ($24.00 barrier cost + $0.31 thread cost + $2.30 labor cost) to $38.80 ($36 barrier cost + $0.50 thread cost + $2.30 additional labor cost) for a sheet barrier.

6

Various types of high-loft and sheet barriers are widely available for sale and therefore it is expected that those whose prices are at the upper end of the range will either not be produced (because mattress manufacturers will not buy them) or their prices will drop (so that they can compete with other barriers available for sale). Hence the total materials and labor costs will most likely be closer to the lower end of the estimated range.

6

Some producers are also using an FR mattress edge binding tape, which costs an average of $2.52 per mattress, while a non-FR tape costs an average of $1.68. This makes the net increase in costs, due to using FR edge binding tape, equal to $0.84. This cost is not added to the total production costs, because it is not required for the mattress to pass the burn test.

Costs of Prototype and Confirmation Testing.

Each mattress/set prototype is required to be tested in triplicate for prototype qualification. According to industry representatives, the cost of testing per twin-size mattress/set may be about $500: the sum of the average cost of the materials and shipping ($100) and the cost of the use of the lab ($400). Hence, the cost of testing three mattresses/sets for prototype qualification equals $1500. Additionally, if some mattress/set prototypes do not pass the first time, then the cost will be higher, because additional tests will be done after action is taken to improve the resistance of the prototype. If 10 percent of mattresses are retested, then the average cost of testing a prototype would be 10 percent higher, or $1650. This cost is assumed to be incurred no more than once per establishment for each prototype. It is expected that a qualified prototype will be used to represent a mattress construction (

e.g.

, single-sided pillow top) with all other prototypes using the same construction (with different size and different ticking materials) being based on the qualified prototype. If companies pool their prototype definitions across different establishments or different companies, testing costs would be smaller as all but one of the firms/establishments producing to the specification of a pooled prototype will burn one mattress (for the confirmation test) instead of three (for the prototype test). The probability of a mattress failing a confirmation test is small. Therefore, it is expected that the average cost of testing per mattress will be lower for firms and/or establishments that pool their results than for those that do not.

If manufacturers test every mattress construction (

e.g.

, single-sided pillow top, double-sided pillow-top, tight-top, euro-top, * * * etc.), which is estimated, based on conversations with manufacturers, to average about twenty per manufacturer, for every establishment in a given year, then their average testing cost per mattress would approximately equal 92.5 cents ($1650 × 20 constructions × 639 establishments/22.8 million conventional mattresses) per mattress for the first year of production. If manufacturers use a qualified prototype of the least fire-resistant mattress/set construction (“worst case”) to represent other mattress/set constructions, then the average cost of testing per mattress for the first year of production will be reduced. Pooling testing results across establishments and/or firms will further reduce the average cost of testing per mattress. On an annual basis testing costs will be further reduced because prototypes need only be tested in the year they are first developed.

Cost of Information Collection and Record Keeping.

In addition to prototype testing, the proposed standard will require detailed documentation of all tests performed and their results including video or pictures; prototype or production identification number; date and time of test; and name and location of testing facility; test room conditions; and test data for as long as the prototype is in production and for three years after its production ceases. Manufacturers are also required to keep records of a unique identification number for the qualified prototype and a list of the unique identification numbers of each prototype based on the qualified prototype and a description of the materials substituted and/or the size change. Moreover, they are required to document the name and supplier of each material used in construction of a prototype and keep physical samples of the material. Additionally, they are required to identify the details of the application of any fire retardant treatments and/or inherently fire resistant fibers employed relative to mattress components. This documentation is in addition to documentation already conducted by mattress manufacturers in their efforts to meet the cigarette standard. Detailed testing documentation will be done by the test lab and is included in the estimated cost of testing. Based on CPSC Office of Compliance staff estimates, all requirements of the proposed standard are expected to cost an establishment about 110 minutes, or 1.3 hours, per qualified prototype. Assuming that every establishment will produce 20 different qualified prototypes, the increase in record keeping costs is about $935 (110 minutes × 20 qualified prototypes × $25.50 in average civilian workers' compensation per hour) per establishment per year. (Note that pooling among establishments or using a qualified prototype for longer than one year will reduce this estimate.) This translates to an average cost of 2.6 cents per mattress for an average establishment, with average output of 35,681 conventional mattresses.

Cost of Quality Control/Quality Assurance Programs.

To ensure that all mattresses are produced to the prototype specification across all factories and over the years for which a production line exists, mattress

manufacturers will need a thorough well-documented quality control/assurance program. The top 12 mattress producers (with a market share of almost 80 percent) have existing quality control programs which could be modified to fit the new standard with minimal additional costs. Smaller producers, whose quality control programs are less detailed or non-existent, will incur some incremental costs as a result of the proposed standard. These incremental costs will be small for each manufacturer and less when measured per mattress. (See the section on impact of the proposed standard on small businesses for a description of their cost of quality control and quality assurance programs.)

Additionally, although the proposed standard does not require production testing, it encourages random production testing to assure manufacturers that their mattresses continue to meet the requirements of the rule, as a possible component of the quality control/quality assurance program. Assuming that an average of 3 mattress/foundation constructions will be tested per establishment per year yields an estimated cost of production testing of about $1500. Based on this assumption, the estimated cost of testing mattress/foundation sets for quality assurance purposes, therefore, equals 4.2 cents per mattress ($1500/35,681) for an average establishment.

The labor needed to meet the quality assurance measures required by the standard is estimated by CPSC Office of Compliance staff to be 224 minutes per establishment per prototype per year. Assuming that every establishment will produce 20 qualified prototypes, the increase in labor costs associated with quality assurance requirements of the proposed standard is about $1904 (224 minutes × 20 qualified prototypes × $25.50 average civilian workers' compensation per hour) per establishment per year. (Note that pooling among establishments or using a qualified prototype for longer than one year will reduce this estimate.) This yields an average cost of 5.3 cents per mattress for an average establishment, with average output of 35,681 mattresses. Hence total costs of quality assurance/quality control programs may average about 9.5 cents (4.2 + 5.3) per conventional mattress per year.

Costs to Wholesalers, Distributors, and Retailers.

An added cost of the proposed standard is the increase in costs to wholesalers, distributors, and retailers in the form of additional storage, transportation, and inventory financing costs. Since a mattress complying with the proposed standard will not be bigger than a similar mattress produced before the standard becomes effective, storage and transportation costs are not expected to increase. Inventory financing costs will increase by the average cost of borrowing money, applied to the wholesale price of a mattress over the average inventory holding time period. Since most mattress producers use just-in-time production and have small inventories, this additional cost will probably not exceed ten percent of the increase in production cost (which is the sum of material, labor, testing, record keeping, and quality assurance costs). A ten percent mark-up is, therefore, being used to measure the cost to wholesalers, distributors, and retailers. This yields a resource cost to wholesalers, distributors, and retailers equal to $1.15 to $3.98 per mattress/set. Retail prices may increase by more than the ten percent mark-up. Section 8 discusses the impact of the proposed standard on retail prices of mattresses.

Costs of Compliance and Enforcement.

Compliance and enforcement costs refer to the costs incurred by CPSC to ensure that manufacturers are complying with the proposed standard. Based on past experience with the existing mattress standard, the estimated CPSC inspection time spent per location (establishment) equals 33 hours for inspection and 6 hours for sample collection. This yields a cost per inspection of about $1,664.52 (39 hours × $42.68, the average wage rate for CPSC inspectors). Additionally, compliance officers spend an average of 20 hours per case, making their cost equal to $1,032.80 (20 hours × $51.64, the average hourly wage rate for compliance officers). This yields an average compliance and enforcement total labor cost of $2,697.32 per inspected establishment per year.

It should be noted that the expected cost per establishment, if less than one hundred percent of establishments are inspected every year, equals the cost per inspected establishment times the probability that a given establishment will be inspected. Though the probability that a given establishment will be inspected in a given year is not known, assuming that a third of all establishments will be inspected (

i.e.

, about 213 establishments) yields a compliance and enforcement total expected labor cost of $899.11 ($2,697.32 × (

1/3

)) per establishment per year.

In addition to labor costs, CPSC will incur testing costs. It should be noted that the decision to collect samples after an inspection visit is made at the discretion of the investigator and, therefore an accurate assumption about the number of samples collected and sent for a burn test cannot be made. If, based on inspection, samples from 10 percent of all inspected establishments were to be collected and sent to a lab for a burn test, and if samples representing 5 qualified prototypes are taken from each of these establishments, then the total cost of CPSC testing will be $157,500 (5 qualified prototypes × $1,500 (the cost of testing up to 3 mattresses for each qualified prototype) × 21 (10 percent of 213 inspected establishments)). These assumptions about frequency of testing yield an expected cost of testing per establishment of $246.48 ($157,500/639).

Therefore the expected total CPSC wage and testing costs associated with the proposed standard per establishment per year equal $1,145.59 ($899.11 + $246.48). With an average production of 35,681 mattresses per establishment (22.8 million mattresses divided by 639 establishments), the average CPSC wage and testing costs equal 3.2 cents per mattress ($1,145.59/35,681). These costs are expected to decrease over time as manufacturers learn the requirements of the proposed standard.

Total Resource Costs.

Therefore total resource costs (including material costs, labor costs, costs of prototype and confirmation testing, paperwork collection and record keeping costs, costs of quality control/quality assurance programs, production testing costs, costs to wholesalers, distributors, and retailers, and costs of compliance and enforcement) are estimated to range from $12.63 to $43.86 per mattress. This range includes both the high-loft and sheet barriers. The section on the impact of the proposed standard on small businesses and other small entities discusses how costs of testing and quality control/quality assurance programs may differ for small businesses and strategies that small manufacturers might adopt to reduce these costs.

Projected Future Costs.

It is possible that costs associated with the standard will decline over time. A supplier of fire resistant barriers predicts that the price of the barriers will decline by 40 percent in the next two years, due to decreased uncertainty and increased competition. (They have already dropped significantly since TB603 was proposed.) The increase in labor costs due to decreased productivity is expected to be temporary and be reduced when workers get more training and/or the older machines get replaced

with newer machines that are more capable of handling the FR thread and material used in fire resistant barriers. Moreover, as noted above, prototype testing costs are expected to decline after the first year of the standard.

The proposed standard references an effective date of twelve months following publication of a final rule. The costs reported here are based on the assumption that supplier companies will be able to maintain existing capacity. If federal standards for bedclothes and upholstered furniture were mandated at the same time and input producers were not given enough time to increase their capacity, input prices would rise in the short-run because of increased demand for the FR material used by all three industries.

Unquantifiable Costs.

A mattress manufacturer indicated that in response to an FR mattress standard, the number of models/styles produced may be cut by half. If this response is typical, then there may be a reduction in consumers' utility, because of the reduction in mattress types that they would have to choose from. Others indicate that there will be an aversion to producing double-sided mattresses, because it would be harder for them to pass the burn test. Double-sided mattresses possibly have a longer expected life than single-sided ones. To the extent that consumers prefer double-sided mattresses to single-sided mattresses, the shift away from producing double-sided mattresses imposes a non-monetary cost. Though unquantifiable, this reduction in choices of construction type and design is an added cost to consumers of the proposed standard.

Another unquantifiable cost is the possible increase in liability insurance faced by mattress manufacturers. Because the draft proposed standard measures the performance of the entire mattress when exposed to fire, and not its individual components, liability will be shared by input suppliers and mattress manufacturers. Industry representatives expect that manufacturers' liability insurance will increase to reflect the additional possibility of litigation. This increase, however, cannot be quantified because of the novelty of this performance test. Compliance of more mattress firms with the California TB 603 standard may enable us to estimate the additional liability insurance. Notice that any increase in liability insurance faced by FR input suppliers will be included in the price charged for the FR inputs and does not add to the total increase in resource cost that is expected to result from the proposed standard.

8. Benefits and Costs of the Proposed Standard

This section compares benefits and costs of the proposed standard, presents a sensitivity analysis, and highlights the impact of the proposed standard on retail prices, small businesses, children, and the environment. The sensitivity analysis examines the effect of changing some of the assumptions used earlier. The analysis shows that net benefits continue to be positive under a reasonable range of assumptions about the death and injury effectiveness of the proposed standard, the reduction in injuries resulting from the proposed standard, the value of a statistical life estimate, the discount rate, or the expected mattress life.

The expected aggregate lifetime benefits associated with one year's production of mattresses (25.3 million units) using a discount rate of three percent and an expected 10-year mattress life are $1.56 to $1.88 billion ($61.66 to $74.25 per mattress × 25.3 million mattresses). The corresponding expected aggregate costs of the proposed standard are $0.32 to $1.11 billion ($12.63 to $43.86 times 25.3 million). The resulting net aggregate benefits equal $0.45 to $1.56 billion ($17.79 to $61.62 times 25.3 million). For a mattress life of 14 years (and a 3 percent discount rate), aggregate lifetime benefits, costs, and net benefits of the proposed standard associated with one year of production are $1.52 to $1.92, $0.32 to $1.11, and $0.41 to $1.60 billion respectively. The expected benefits of the proposed standard will accrue for a long period of time and discounted net benefits will, therefore, be much greater than net benefits associated with only the mattress production in the first year the standard becomes effective.

Sensitivity Analysis.

The previous analysis compares benefits and costs of the proposed standard using expected mattress lives of 10 and 14 years, a discount rate of 3 percent, an expected effectiveness rate of the proposed standard of 80 to 86 percent of deaths and 86 to 92 percent of injuries, an estimated value of a statistical life of 5 million dollars, and an estimated cost of injury of $179,300. This section examines the effect of changing any of these assumptions on the expected net benefits of the proposed standard.

Comparing expected benefits and costs of the proposed standard, it is clear that net benefits are expected to be positive (

i.e.

, expected total benefits exceed expected costs) for an average mattress life of 10 or 14 years. Though increasing the expected mattress life from 10 to 14 years, while using the 3 percent discount rate, expands the positive range of net benefits, it does not affect the conclusion regarding net benefits. A further increase of the expected life of a mattress similarly would not affect the estimate of net benefits. For example, using the Product Population Model estimate of the number of mattresses in use based on an expected mattress life of 18 years (equal to 367.1 million mattresses) yields net benefits of $14.42 to $64.49 per mattress, using a discount rate of 3 percent.

Net benefits are also positive using discount rates of 3 and 7 percent. Using a 3 percent discount rate, net benefits per mattress equal $17.79 to $61.62 for an average life of 10 years and $16.01 to $63.08 for an average life of 14 years. Using a 7 percent discount rate, net benefits per mattress equal $9.36 to $50.88 for an average life of 10 years and $5.15 to $48.26 for an average life of 14 years. Assuming a larger discount rate reduces net benefits, because future benefits reaped over the life of the mattress contribute less to total benefits.

Net benefits are based on an estimated value of a statistical life equal to $5 million. Changing the estimate used for the value of a statistical life does not have a major impact on the results. For example, if $3 million, the lower bound estimate in Viscusi (1993), is used as an estimate of the value of a statistical life, net benefits become -$2.90 to $36.73 per mattress (using a 3 percent discount rate and an estimated mattress life of 10 years).

7

Alternatively, a $7 million estimate, the higher bound estimate in Viscusi (1993), yields net benefits equal to $38.48 to $86.51 per mattress (using a 3 percent discount rate and an estimated mattress life of 10 years).

7

The range for net benefits was derived by subtracting the upper end of the cost range from the lower end of the benefits range to get the lower end of the range for net benefits and subtracting the lower end of the cost range from the higher end of the benefits range to get the higher end of the range for net benefits. Because of this method, both ends of the range for net benefits are a very unlikely occurrence.

Changing the estimate used for the cost of injury will have minimal impact on the results, because the share of benefits from reduced injuries is 16 percent of total benefits. Hence, even if there were no reduction in injuries from the proposed standard, the net benefits would be $7.86 to $49.59 per mattress (using a mattress life of 10 years and a 3 percent discount rate).

The analysis assumes that the effectiveness of the proposed standard ranges from 80 to 86 percent for deaths and 86 to 92 percent for injuries. The

net benefits will remain positive, with a lower effectiveness rate. For example, assuming an effectiveness rate of preventing death of only 55 percent yields net benefits of $1.86 to $39.84 per mattress and aggregate net benefits of 50 million to 1.01 billion dollars from all mattresses produced the first year the proposed standard is mandated (using a mattress life of 10 years, a 3 percent discount rate, and the same effectiveness for injuries as used in the baseline analysis). Also, assuming a smaller number of deaths and injuries before the proposed standard is mandated (a smaller baseline risk) would still result in positive net benefits. A 25 percent reduction in baseline death and injury risks yields net benefits of $2.38 to $43.06 per mattress and aggregate net benefits of $60 million to $1.09 billion from all mattresses produced the first year the mattress standard is mandated (using a mattress life of 10 years, a 3 percent discount rate, and the estimated effectiveness measures used in the baseline analysis).

Impact on Retail Prices.

One of the top four mattress manufacturers in the industry has re-merchandised its product lines to lower the costs of other materials so that total costs (and prices) are the same as they were before the production of mattresses that comply with TB603. Other manufacturers have indicated that they will have to increase their price which, according to some manufacturers and based on reported traditional industry mark-ups, might translate to an increase in the retail price to consumers that could reach approximately four-fold the increase in manufacturer's costs. Hence the average increase in the price at which mattress manufacturers are willing to sell their products (supply price) will be anywhere between the price of a similar mattress without FR material and that price plus four times the increase in the costs of production. Given the presence of at least one company that will not increase the price, it is unlikely that the new average price will be close to the higher end of the range because of competition for market share among manufacturers.

The market (equilibrium) price is determined by the intersection of consumers' willingness to buy and producers' willingness to sell the product at different prices. The value the equilibrium price will take (relative to the price before the introduction of fire resistant mattress/sets) will be affected by the change in the demand and supply curves for fire resistant mattress/sets and their relative elasticities. Assuming that the demand curve is unaffected, the equilibrium price will reflect the price elasticity of demand (

i.e.

the sensitivity of the change in the quantity demanded to the change in price) as well as the shift in supply. In the short-run, consumers have a relatively elastic demand curve, because they can always postpone the purchase of a durable good, and therefore the increase in the equilibrium price is expected to be much lower than the increase in the supply price (what producers would want to sell the same number of mattress/sets for). Because of the relatively high elasticity of demand, sales are likely to decrease in the short-run. In the long-run, the demand curve is less elastic, and therefore the equilibrium price and quantity (sales) will be higher than the short-run price and quantity.

Given the availability of mattresses whose retail prices will not increase and the competitive nature of the industry, it is possible that, on average, prices will rise by about twice the costs associated with the standard (

i.e.

, retail price mark-up will average about twice the increase in manufacturing costs). Under this assumption, consumers would pay an additional $22.91 ($11.46 × 2) to $79.69 ($39.85 × 2) per mattress/set (compared to the price they would have paid for a current mattress that does not comply with the proposed standard.

8

Assuming that the demand curve for mattresses is unaffected by the draft proposed standard, some consumers will choose not to purchase (or at least delay the purchase of) a new mattress/set. These consumers who delay or choose not to purchase a new set will not be getting the value (or benefits) that they would have gained from purchasing a new set. This loss, though difficult to quantify, is sometimes measured as a loss in consumer surplus (McCloskey, 1982).

8

These cost figures include labor and material costs; testing costs; record-keeping costs; and quality assurance program costs. They do not include the costs to wholesalers, distributors, and retailers or compliance costs because they are not incurred by the manufacturers.

It is unlikely, however, that the post-standard demand curve for mattresses will be the same as the current demand. Early 2004 market observations indicate consumer and retail enthusiasm about the fire resistant mattresses already available for sale (Furniture Today, April 26th, 2004.) If this enthusiasm generally reflects consumers' preferences, then the demand for mattresses may increase. This would tend to offset any reduction in mattress sales and possible losses in consumer surplus.

Impact on Small Businesses and Other Small Entities.

The increase in material and labor costs to meet the proposed standard is not likely to be dependent on a firm's size and will therefore not disproportionately affect small businesses. The cost imposed disproportionately (per unit produced) on small businesses will be the cost of testing, information collection and record keeping and quality control/quality assurance programs. While these costs are estimated to be a little over one dollar per mattress per year for average-sized establishments, they could be substantially higher for small mattress manufacturers. The proposed rule includes measures that these manufacturers can use to minimize the testing burden. Furthermore, firms with more than one establishment, or different firms, may be able to reduce these costs by pooling their testing and quality control programs over all establishments or firms.

Use of pooling across establishments and firms would ameliorate the impact of the proposed standard on small businesses. By getting together across different states and regions, small manufacturers who do not share a common market (and therefore do not compete with each other) can resemble a large producer in their testing and quality control/quality assurance efforts and therefore reduce their costs per mattress. It is also expected that some barrier suppliers would be willing to do the testing and quality control/assurance programs for small manufacturers in exchange for a small charge, which will be similar to the average cost per mattress for large businesses, because the volume of output will be large.

Impact on the Environment.

The extraction, processing, refinement, and conversion of raw materials to meet the proposed standard involve energy consumption, labor, and the use of potentially toxic chemicals. Most manufacturing has some impact on the environment, and manufacturing fire resistant mattresses is no exception. Because the proposed standard is a performance standard, it does not restrict manufacturers' choice of fire resistant materials and methods that could be used in the production of mattresses. There appear to be several economically viable options to meet the standard that, based on available information, do not impose health risks to consumers or significantly affect the environment. (See discussion at Section N of this preamble.)

Impact on Children.

Deaths and injuries among children constitute a substantial proportion of mattress-

related fire losses, and of the potential benefits of the proposed standard. A CPSC staff report, based on a field investigation study in 1995 to learn more about cigarette-ignited fires and open-flame fires, found that 70 percent of open-flame fires involved child play and that child play was involved in 83 percent of the 150 deaths of children less than five years of age. A National Association of State Fire Marshals 1997 study also indicated that 66 percent of the small open-flame ignitions were reportedly started by children under the age of 15 (21 percent by children under 5).

For virtually all of the fires started by children less than 15 years of age, the ignition was not witnessed by an adult (Boudreault and Smith, 1997). Reducing the likelihood of flashover in the first 30 minutes of the fire may therefore benefit children disproportionately, as it allows enough time for adults to detect the fire and save young children in close proximity to the fire. Also children between 5 and 9 who sometimes do not cooperate with adults and run away from adults to other parts of the occupancy will have enough time to be found and rescued by an adult.

The Epidemiology staff's memorandum shows that, based on national fire estimates for the years 1995-1999, children younger than 15 accounted for 27 percent of addressable deaths and 23 percent of addressable injuries. They also indicate that the proposed standard would reduce deaths and injuries to children ages 5 and younger by 85 to 92 percent and 80 to 87 percent respectively. Deaths and injuries to children ages 5 to 14 were estimated to be reduced by 94 to 97 percent and 88 to 94 percent respectively. This represents a total of 100 to 110 deaths of children less than 15 years of age per year for the 1995-1999 period. It also represents 410 to 440 injuries to children less than 15 years of age for the same period.

9. Alternatives to the Proposed Standard Alternative Maximum Peak Heat Release Rate (PHRR) and Test Duration.

The initial California TB 603 proposal required the duration of the test to last 60 minutes with a maximum PHRR of 150kW. After receiving comments on this proposal, the California Bureau of Home Furnishings and Thermal Insulation changed the criterion to a maximum of 200 kW PHRR in the first 30 minutes, the requirement for both the federal proposed standard and the current TB 603.

Increasing the duration of the test and reducing the PHRR would, according to several input suppliers, increase the resource costs to manufacturers of a queen mattress/foundation set by $15.06 to $50.65 compared to non-complying products (

i.e.

, those not conforming to the proposed standard.)

9

Adding the costs to wholesalers, distributors, and retailers, the costs of testing, quality control/assurance programs, record-keeping, and CPSC compliance efforts, yields a total resource cost of the stricter standard (150 kW and 60 minutes) of $16.59 to $55.74 (costs to manufacturers {$15.06 to $50.65} + cost to wholesalers, distributors, and retailers, equal to 10% of costs to manufacturers {$1.51 to $5.06} + 3.2 cents CPSC compliance costs) per mattress. This represents a marginal increase in costs of $3.96 ($16.59−$12.63) to $11.88 ($55.74−$43.86) over the costs associated with the proposed standard.

10

9

The lower end of the range is based on barrier price of one supplier, whose capacity is expected to meet 25 to 30 percent of the whole market demand in the short run. The next cheapest alternative costs $24 for the barrier material alone.

10

These cost estimates (and the resulting marginal increase) should be viewed as approximate since no extensive tests of the barriers have been conducted for 60 minutes, as most manufacturers are focused on meeting the less strict requirements. Input suppliers generally do not assemble and test large numbers of mattresses, and may therefore underestimate reduced labor productivity and/or reduced output per machine (compared to a maximum PHRR of 200 kW for a 30-minute test) due to handling the thicker denser barrier. A number of mattress producers estimate that to meet the stricter standard, manufacturing costs would increase (over those of non-compliant mattresses) by $50 to $70 for a queen-sized set (Furniture/Today, July 21, 2004).

Such increase in costs would likely result in consumers facing higher mattress prices. Based on traditional industry mark-ups, the new price may reflect a two- to four-fold increase over the increase in production costs, depending on the relative elasticity of demand and supply for mattresses. This yields a total increase in the average price of a queen mattress/set of $30.11 (2 × $15.06) to $202.58 (4 × $50.65). Potential benefits of the stricter standard could be higher than the proposed standard, but the extent is uncertain. Given an effectiveness rate of greater than 80 percent of the proposed standard, the additional benefits of stricter test requirements are limited. Assuming that the stricter standard could eliminate 50 percent of the remaining deaths and injuries (

i.e.

, it could save 39 additional lives and prevent 136 additional injuries), then an additional benefit of about $7.66 per mattress would be expected. This additional benefit, however, would come with additional costs (discussed above) and therefore may reduce net benefits. Moreover, a small increase in net benefits may not justify the large increase in retail price that would result from a stricter standard.

A bedding official estimated that such price increases may result in reduction in sales of 25 percent or more (Furniture/Today, July 21, 2004). The larger increase in prices (compared to the less strict test requirements) and the resulting reduction in sales could drive some of the smaller manufacturers out of business. (The stricter standard is more likely to require replacing some existing machines to accommodate the denser barrier material, which would be disproportionately more costly for smaller firms whose machinery is older and less sophisticated.) Since mattresses are durable goods, one would expect a larger drop in sales in the short-run, as consumers choose to keep their old mattresses longer than before. This would make the reduction in sales more pronounced in the short-run, increasing the likelihood that some firms may exit the market. Moreover, if a large number of consumers choose to extend the life of their mattresses for a longer time period, it will take longer to achieve the benefits expected to be associated with the safer mattresses.

Alternative Total Heat Released in the First Part of the Test.

TB 603 limits the total heat released during the first 10 minutes of the test to 25 MJ. The proposed standard's stricter limit (15 MJ in the first 10 minutes) reduces the expected size of the initial fire and hence allows consumers a greater chance to escape the fire and get out of the room, even if the room never reaches flashover. The effectiveness rates presented in the analysis are based on the stricter criterion. Using the TB 603 criterion (25 MJ in the first 10 minutes) would likely reduce estimated benefits (the estimated reductions in deaths and injuries), without having any significant effect on costs. According to several producers, mattresses that use existing barrier technology release total heat that is far below the 25 MJ requirement of TB 603. Therefore, using the TB 603 criterion for the total heat released would not change costs but could potentially reduce the benefits and, hence, the net benefits of the proposed standard.

Moreover, because of the small fuel load of ticking materials currently being used, the lower total heat release requirement allows the production of mattress/sets based on a prototype that has not been tested so long as it differs from the qualified prototype only with respect to ticking and the ticking material is not part of the fire resistance solution. Requiring a test for every prototype with a different ticking was

rejected because of the magnitude of the burden it would impose on small manufacturers who do not produce large numbers of any one prototype and who would have been disproportionately adversely affected by these requirements.

Alternative Testing Requirements.

The proposed standard requires prototype testing (of three mattress/sets) before a manufacturer starts production of a given mattress design and a confirmatory test of one mattress if more than one establishment or firm are pooling their results. Manufacturers may sell a mattress/set based on a prototype that has not been tested if that prototype differs from the qualified prototype only with respect to: (1) Mattress/foundation size; (2) ticking, unless the ticking of the qualified prototype has characteristics designed to improve the performance on the burn test; and/or (3) any component, material, or method of construction that the manufacturer can demonstrate on an objectively reasonable basis will not cause the prototype to exceed the test criteria. Though production testing (

i.e.

, random burning of mattress/sets to ensure that all production units meet the standard) is encouraged by the proposed standard under quality assurance program requirements, it is not required. The individual manufacturer's decision on whether to conduct production testing (and if so, at what frequency) will clearly depend on the efficacy of his/her quality assurance/control efforts.

As an alternative, the proposed Federal standard could, like TB 603, omit testing or prototype definition requirements. Without testing, however, it might be difficult for manufacturers to know whether their mattresses will comply with the standard. Alternatively, the standard could require production testing with a specified frequency. This specification, however, could result in unnecessary costs if they are not justified given the quality control measures generally undertaken by manufacturers in the absence of the proposed standard. Requiring more tests per establishment, prototype, or enterprise will increase the estimated costs per mattress and could reduce net benefits.

Alternative Effective Date.

The proposed effective date is twelve months from the date of publication of the final rule in the

Federal Register

. Given the length of time needed to ensure the availability of inputs for the production of barrier materials, availability of barriers for mattress producers, and a sufficient volume of inventories at retailers' showrooms, an earlier effective date may result in higher input costs to manufacturers. More importantly, it is expected that smaller manufacturers will be disproportionately affected, as they are more likely to wait to invest in development efforts until the technology is developed by larger firms, or until the proposed standard becomes effective. A later effective date (longer than twelve months) could reduce expected net benefits as more fires, deaths, and injuries associated with mattresses would occur between the date of publication in the

Federal Register

and the date the standard becomes effective. The staff is unaware of evidence that small manufacturers would be negatively impacted by a twelve months period relative to a longer period, such as eighteen or twenty-four months. The Commission is requesting comments from small businesses on the expected economic impact of the effective date and other requirements of the proposed rule (

see

section M of this document below).

Taking No Action or Relying on a Voluntary Standard.

If the Commission chose to take no action, California may attempt to enforce its standard despite the Commission General Counsel's position on preemption. Larger producers are already moving to comply with California's standard. They also want to avoid product liability claims associated with selling mattresses with different fire resistance in other areas of the country. Three of the largest four producers plan to meet TB 603 nationwide by the end of 2005. Some small California manufacturers may have a smaller incentive to meet TB 603 than a Federal standard. Small manufacturers who do not sell in California may similarly have no incentive to sell mattresses that meet TB 603 requirements in other parts of the country. Hence, expected aggregate net benefits associated with the draft proposed standard are higher than the net benefits that might result under California TB 603 even if it could be enforced in the face of preemption concerns.

No effort has been undertaken to develop a voluntary standard. Furthermore, industry representatives support a mandatory standard to level the playing field among domestic producers (large and small) and importers. If a voluntary standard were developed, the economic burden would fall primarily on the larger firms (who would likely be the first to comply), their market shares could be reduced and benefits to consumers (in terms of reduced deaths and injuries) would likely decline accordingly.

Labeling Requirements.

The Commission could require labeling on mattresses to warn consumers in lieu of a standard. Labeling is not considered an effective option for reducing the risk of fires. Since mattress labels are usually covered by bedclothes and may not be seen by the mattress users, labeling mattresses is likely to be an ineffective means of warning consumers. Moreover, fires started by children who cannot read or do not change the bed sheets will not be reduced by a labeling requirement. Hence, while labeling costs are probably negligible, labels are unlikely to reduce mattress fires.

Labeling of chemically treated components has been suggested as a possible requirement of the draft standard, to inform consumers of the materials used. The costs of such labeling would also be negligible, since existing mattresses have labels and producers could probably add a description of the chemical treatment (if any) to the existing label. Labeling of chemically treated components could provide small unquantifiable benefits to consumers as it would provide some additional information. However, because a label would only provide the name of any chemical treatment without any information about whether the treatment has any potential health effects, it would be of little practical use for the consumer. Information on the use of chemically treated components is, however, required as part of the record keeping requirements of the standard.

L. Paperwork Reduction Act

The proposed standard will require manufacturers (including importers) of mattresses/sets to perform testing and maintain records of their testing and quality assurance efforts. For this reason, the rule proposed below contains “collection of information requirements,” as that term is used in the Paperwork Reduction Act, 44 U.S.C. 3501-3520. Therefore, the proposed rule is being submitted to the Office of Management and Budget (“OMB”) in accordance with 44 U.S.C. 3507(d) and implementing regulations codified at 5 CFR 1320.11. The estimated costs of these requirements are discussed below.

Costs of Prototype and Confirmation Testing.

According to industry representatives, the cost of testing per twin-size mattress/set may be about $500: the sum of the average cost of the materials and shipping ($100) and the cost of the use of the lab ($400). Hence, the cost for testing three specimens as required by the proposed rule mattresses equals $1500. This cost is assumed to be incurred no more than

once per establishment for each prototype. If manufacturers test every mattress construction (

e.g.

, single-sided pillow top, double-sided pillow-top, tight-top, euro-top, * * * etc.), which is estimated, based on conversations with manufacturers, to average about 20 per manufacturer, for every establishment in a given year, then the estimated industry testing cost per establishment per year would approximately equal $30,000. The total number of establishments producing conventional mattresses in 2001 was 639. Using an estimated number of 739 producers of both conventional and unconventional mattresses, the annual cost to all (conventional and unconventional) mattress producers is $22.17 million.

Cost of information collection and recordkeeping.

In addition to prototype testing, the proposed standard will require detailed documentation of prototype identification and testing records, model and prototype specifications, inputs used, name and location of suppliers, and confirmation test record, if establishments choose to pool a prototype. This documentation is in addition to documentation already conducted by mattress manufacturers in their efforts to meet the cigarette standard. Detailed testing documentation will be done by the test lab and is included in the estimated cost of testing. Based on CPSC Office of Compliance staff estimates, all requirements of the proposed standard are expected to cost an establishment about 110 minutes, or 1.3 hours, per qualified prototype. Assuming that every establishment will produce 20 different qualified prototypes, the increase in record keeping costs is about $935 (110 minutes × 20 prototypes × $25.50 in average civilian workers' compensation per hour) per establishment per year. (Note that pooling among establishments or using a prototype qualification for longer than one year will reduce this estimate.) This translates to an annual cost to all (conventional and unconventional) mattress producers of $690,965 ($935 × 739).

Cost of quality control/quality assurance programs.

To ensure that all mattresses are produced to the prototype specification across all factories and over the years for which a production line exists, mattress manufacturers will need a thorough well-documented quality control/assurance program. The top 12 mattress producers (with a market share of almost 80 percent) have a existing quality control programs which could be modified to fit the new standard with minimal additional costs. Smaller producers, whose quality control program is less detailed or non-existent, will incur some incremental costs as a result of the proposed standard. These incremental costs will be small for each manufacturer and less when measured per mattress. (See the section on impact of the proposed standard on small businesses for a description of their cost of quality control and quality assurance programs.)

Additionally, the proposed standard encourages random production testing to assure manufacturers that their mattresses continue to meet the requirements of the rule. Assuming that an average of 3 mattress/set constructions will be tested per establishment per year yields an estimated cost of production testing of about $1500. The labor needed to meet the quality assurance measures required by the standard is estimated by CPSC Office of Compliance staff to be 224 minutes per establishment per qualified prototype per year. Assuming that every establishment will produce twenty different qualified prototypes, the increase in labor costs associated with quality assurance requirements of the draft proposed standard is about $1904 (224 minutes × 20 qualified prototypes × $25.50 average civilian workers' compensation per hour) per establishment per year. Hence total costs of quality assurance/quality control programs may average about $3,404 ($1500+1904) per establishment per year. This translates to an annual cost to all (conventional and unconventional) mattress producers of $2,515,556 ($3,404 × 739).

M. Initial Regulatory Flexibility Analysis

1. Introduction

The Regulatory Flexibility Act (“RFA”) generally requires that agencies review proposed rules for their potential economic impact on small entities, including small businesses. Section 603 of the RFA calls for agencies to prepare and make available for public comment an initial regulatory flexibility analysis describing the impact of the proposed rule on small entities and identifying impact-reducing alternatives. Accordingly, staff prepared an initial regulatory flexibility analysis for the mattress proposed rule.

9

A summary of that analysis follows.

2. Impact on Small Businesses and Other Small Entities

Summary of proposed requirements.

The proposed standard will apply to all mattresses and mattress and foundation sets, as discussed earlier in this document (see section G.2. above). Options that manufacturers may choose in order to meet the proposed standard include one or a combination of the following: fire resistant ticking; chemically treated or otherwise fire resistant filling products; or a fire blocking barrier (either a sheet style barrier or a high-loft barrier, sometimes called a fiber barrier).

For each qualified prototype, three mattresses/sets must be tested and must pass the test requirements. To obtain a passing result, each mattress/set must pass a 30 minute test, where the PHRR does not exceed 200 kW and the total heat release does not exceed 15 MJ in the first 10 minutes of the test. A failure of any of the sets would require that the problem be corrected and the prototype be retested and pass the test (in triplicate). Manufacturers may sell any mattress/foundation set based on a qualified prototype. Manufacturers may also sell a mattress/set based on a prototype that has not been tested if that prototype differs from a qualified prototype only with respect to (1) mattress/foundation size; (2) ticking, unless the ticking of the qualified prototype has characteristics designed to improve performance on the burn test; and/or (3) any component, material, or method of construction that the manufacturer can demonstrate, based on an objectively reasonable basis, will not cause the prototype to exceed the test criteria specified above.

If one or more establishments (plants within the same firm) or independent firms choose to “pool” prototypes, then each pooling plant or firm is required to conduct a confirmation test for one mattress/set it produces locally. If that set fails, then its producer cannot sell mattresses based on that prototype unless it successfully tests another mattress/set after correcting its production to make sure that it is identical to the original prototype. A pooling firm may sell other mattresses that have not been tested by the pooling firm if they differ from the pooled prototype only with respect to (1) mattress/foundation size; (2) ticking, unless the ticking of the qualified prototype has characteristics designed to improve performance on the burn test; and/or (3) any component, material, or method of construction that the manufacturer can demonstrate, based on an objectively reasonable basis, will not cause the prototype to exceed the test criteria specified above.

Manufacturers are required to keep records of all tests performed and their results, including video or pictures, prototype identification number, date and time of test, name and location or

testing facility for as long as the prototype is in production and for three years after its production ceases. Manufacturers are also required to keep records of a unique identification number for the qualified prototype and a list of the unique identification numbers of all other prototypes based on the qualified prototype, together with a description of the material substituted and/or size change. Moreover, they are required to document the name and supplier of each material used in construction and keep physical samples of the material. Additionally, they are required to identify the details of the application of any flame retardant treatments and/or inherently flame retardant fibers employed relative to mattress components. Finally, they are required to have an adequate quality assurance program in place.

Impact on small businesses.

The proposed standard covers manufacturers and importers of mattresses. There were 557 mattress firms and 639 mattress establishments in 2001, according to the Statistics of U.S. businesses, Census Bureau data. All but the largest twelve firms had less than 500 employees. The U.S. Small Business Administration's Office of Advocacy defines a small business as one that is independently owned and operated and not dominant in its fields. A definition that is used frequently and is less subject to interpretation is a firm with fewer than 500 employees. The latter definition classifies 97.8 percent ((557 − 12)/557) of all mattress firms as small businesses.

Average employment per firm for the whole industry is 45.8 employees. Average employment for the 1-4 employees per enterprise group, which represents 22.98 percent of all firms, is 2.4 employees. Average employment for the fewer than twenty employees per enterprise group, which represents 61.22 percent of all firms, is 6.2 employees. Hence more than half of mattress firms have fewer than twenty employees.

In addition to domestic producers, importers will be affected by the proposed standard. Imported mattresses represent less than two percent of total U.S. shipments.

The increase in material and labor costs of the proposed standard (estimated to be $10.41 to $38.80 per mattress) is not likely to be dependent on a firm's size and will therefore not adversely affect small businesses. Larger firms are bearing all the capital investment costs of research and development, sharing some of these costs with input suppliers. Most smaller firms are waiting to buy from the suppliers a barrier solution, which has been tested extensively and is known to meet the standard. The price smaller firms pay to cover the development and testing costs borne by the supplier will not disproportionately impact them, because it is not measured relative to their small output, but relative to the supplier's output. Other smaller firms may combine their development efforts to be able to benefit from dividing the costs over a larger number of firms. Finally, small mattress producers who do not assemble the mattress panels, but buy them from a panel supplier are effectively acting as a large producer by combining all their output. This is because the panel supplier will be responsible for including a barrier in the panel assembly and will pass that cost on to the mattress producers, again not disproportionately impacting the small producers who buy the already assembled panels.

The cost imposed disproportionately (per unit produced) on small businesses will be the cost of testing, information collection and record keeping, and quality control/quality assurance programs. While the regulatory analysis estimates these costs to be a little over one dollar per mattress per year for average-sized establishments, they could be substantially higher for small mattress producers. If manufacturers use a prototype qualification to produce mattress/set constructions for longer than a year, or if they use a worst-case prototype to represent other mattress constructions, these costs will be lower. Furthermore, firms with more than one establishment may be able to reduce these costs by pooling their testing and quality control programs over all establishments. Small independent firms could also pool their testing to reduce their costs per mattress.

Use of pooling across establishments and firms would ameliorate the impact of the proposed standard on small businesses. By getting together across different states and regions, small manufacturers who do not share a common market (and therefore do not compete with each other) can resemble a large manufacturer in their testing and quality control/quality assurance efforts and therefore reduce their costs per mattress. It is also expected that some barrier suppliers would be willing to do the testing and quality control/assurance programs for small manufacturers in exchange for a small charge, which will be similar to the average cost per mattress for large businesses, because the volume of output will be large.

3. Alternatives and Their Possible Effect on Small Businesses

Alternatives considered by the Commission are discussed in the Preliminary Regulatory Analysis section of this preamble (Section K). As discussed therein, increasing the duration of the test and reducing the PHRR would increase costs without necessarily increasing benefits. This would necessitate an increase in costs. Staff estimates the marginal increase in costs to be $3.96 to $11.88 over the costs of the proposed standard. Although a stricter standard might increase benefits, any increase is likely to be small since the proposed standard has an effectiveness rate of 80 percent.

An increase in costs would likely result in an increase in mattress prices. A bedding official estimated that such price increases may result in reduction in sales of 25% or more (Furniture/Today, July 21, 2004). The larger increase in prices (compared to the less strict test) and the resulting reduction in sales could drive some of the smaller producers out of business. (A stricter standard would be more likely to require replacing some existing machines, to accommodate the denser barrier material, which would be disproportionately more costly for smaller firms, whose machinery is older and less sophisticated.) Since mattresses are durable goods, one would expect a larger drop in sales in the short-run than in the long-run, as consumers choose to keep their old mattresses longer than before. This would make the reduction in sales more pronounced in the short-run, increasing the likelihood that some firms may exit the market. Moreover, if a large number of consumers choose to extend the life of their mattresses for a longer time period, it will take longer to achieve the benefits expected to be associated with the safer mattresses.

As discussed in the preliminary regulatory analysis, the Commission also considered a different criterion for the total heat released during the first 10 minutes of the test,

i.e.

, 25 MJ instead of the 15 MJ the Commission is proposing. Using the 25 MJ criterion would likely reduce estimated benefits (the estimated reductions in deaths and injuries), without having any significant effect on costs. According to several producers, mattresses that use existing barrier technology release total heat that is far below the 25 MJ level.

Moreover, because of the small fuel load of ticking materials currently being used, the lower total heat release requirement allows the production of mattress/sets based on a prototype that has not been tested so long as it differs

from the qualified prototype only with respect to ticking and the ticking is not part of the fire resistance solution. Requiring a test for every prototype with a different ticking was rejected because of the magnitude of the burden it would impose on small producers who do not produce large numbers of any one prototype and would have been adversely affected by these requirements.

The Commission also considered alternative testing requirements. The proposed standard requires prototype testing (of three specimens of mattress/sets) before a manufacturer starts production of a given mattress design and a confirmatory test of one mattress if a firm is producing a mattress based on a prototype produced by another manufacturer in a pooling arrangement. Although production testing (

i.e.

burning mattress/sets to ensure that production units meet the standard) is encouraged by the proposed standard under quality assurance program requirements, it is not required. The individual manufacturer's decision on the need for and frequency of production testing will clearly depend on the efficacy of its quality assurance/control efforts.

As an alternative, the Federal standard could, like TB 603, omit testing requirements. However, without testing, it might be difficult for manufacturers to know whether their mattresses will comply with the standard. Alternatively, the standard could require production testing with a specified frequency. This specification, however, could result in unnecessary costs if they are not justified given the quality co

This text is long and has been trimmed here. Open the source document for the complete record.