EXHIBIT A Safe Harbor Principles 1. Fixed Base Weights: In constructing the price level component of the Cost-of-Living Allowance ("COLA") index, price relatives will be averaged using a system of fixed expenditure weights. If available, the weights will be drawn from the base area. If base area weights are not available, regional or national weights may be used. Discussion: The Office of Personnel Management ("OPM") will use Bureau of Labor Statistics ("BLS") Consumer Expenditure Survey ("CES") data to compute weights. OPM will not use micro data or data that could not be made available to the public under the Privacy Act or Freedom of Information Act. OPM will use Washington, DC, CES data to the extent practicable and will use nationwide CES Total Complete Reporting data at any level (e.g., the subcategory, sub-subcategory, etc., and item level) for which there are no DC area data. Nationwide data may also be used when the DC-area data appear to be highly variable over time. OPM will not use rolling averages because surveys will be conducted only once every 3 years. If BLS combines Washington and Baltimore for CES reporting purposes and the DC CES data are no longer available, OPM may use Washington/Baltimore CES data. 2. Democratic Weights: To the extent feasible, weights for calculating the price level component will be based on an average, weighted by population, of expenditure shares in the expenditure range typical of Federal employees. 3. Single Income Approach: In producing the price level component of the COLA index, a single set of weights will be used to represent spending patterns as reflected in the CES data used under Safe Harbor Principle ("SHP") 1. Discussion: OPM will use a single set of weights that will be derived as described in the "Discussion" in SHP 1 and in SHP 2. Because CES "pre-pub" data for the DC area are not currently available by income level OPM may have to select another approach in consultation with the Employee Involvement Structure ("EIS") (for the purposes of these Safe Harbor Principles, EIS refers to the entities described in SHP 24A -24C) and the Technical Advisory Committee ("TAC") (described in SHP 25). 4. Market Basket Composition: Limited variations are allowable among the COLA areas in the items to be used for pricing. The base area items will be matched with those used for each allowance area. Discussion: As is its current approach, OPM will determine the significant CES items by selecting relatively well-defined items (e.g., "apples") as compared with groups of miscellaneous items (e.g., "miscellaneous fresh fruits") that have a significant weight in terms of consumer expenditures. Generally, OPM selects two items for each item category (e.g., "apples" and "oranges"). Prior to each survey, OPM will choose appropriate products (types, brands, models, and sizes) that represent the selected CES items. The items and products chosen will be those that are most relevant to the COLA area and also found in the DC area. 5. Quality and Quantity Comparisons: The items used for pricing goods and services in a COLA area will be compared with identical items in the base area insofar as reasonably possible. To the extent that prices are compared for items which are not identical, such items will be of closely similar quality and quantity, with quantity adjustments as necessary. Discussion: This principle affirms OPM's current practice, which OPM will continue. 6. Sale Prices: Price of items, excluding motor vehicles, will be recorded as actually charged at the time surveyed. If the item is on sale, that price will be used, but this does not include market-wide distress sales. Discussion: With two exceptions, OPM will survey the price of the item at the time of the survey. If the item is on sale, the sale price will be used. The only exceptions will be (1) the price of motor vehicles and (2) the price of other items when the price is reduced because of a market-wide distress sale or because the establishment is going out of business. In the latter case, another outlet will be selected, if possible. 7. Outlet Selection: In consultation with the EIS and the TAC, while it exists, OPM will, as warranted, conduct periodic Point of Purchase surveys ("POPS") to determine the outlets or types of outlets in the COLA areas in which to price items. Items will be priced in the same types of outlets in the COLA areas and in the base area to the extent reasonably practicable. OPM will determine the frequency and content of the POPS in consultation with the EIS and TAC. Discussion: This principle allows OPM discretion to conduct POPS in consultation with the EIS and TAC, while it exists, when they are warranted. 8. Item Selection: OPM will consult with the EIS to select representative items (types and brands) to be priced in each COLA area. Items may be selected using judgmental sampling. Then these same items will be surveyed in theCOLA area and the base area. The survey items and replicates (i.e. prices) will, as warranted, be reallocated or the number increased to reduce overall index variance. For example, too many items may now be surveyed under food and automobile maintenance and too few in medical services, clothing, appliances. In general, it is likely to be more important to increase the range of items within a category like men's footwear rather than the number of brands of a given item. For example, rather than add a second brand of men's shoes, it is likely to produce a more stable measure if another type of shoe, say a hiking boot, were added. Discussion: This principle allows OPM discretion to make such changes in consultation with the EIS and TAC, while it exists. OPM will give serious consideration to surveying more items in categories where prices seem highly variable over time. The extent to which OPM can do this will depend on resources and workload. This principle envisions that fewer items could be surveyed in other categories, thereby offsetting at least some of the potential workload increase. 9. Utility Function: A function (rather than a good) will be adopted for utilities. The function will be maintenance of a standard ambient temperature of a standard (e.g., 2000 square foot living space) house in each area. A survey of employees will be used, as warranted, to determine heating/cooling energy sources and to arrive at a weighted cost. Information on typical energy requirements will be obtained from appropriate local sources. Other energy requirements for the standard house can also be considered. Discussion: Currently, OPM knows of no alternate acceptable data sources and believes an employee survey is the only way that such data could be obtained. The principle does not exclude the use of other data sources. For example, if the Department of Housing and Urban Development has data appropriate for this application, OPM should be able to use that data in place of surveying Federal employees. In applying this principle, the Government anticipates that OPM will use only new home construction in each area; otherwise engineering estimates could vary greatly. Although energy costs will be estimated for the same standard home (i.e., same size and room count) in each area, the construction (number of floors, materials, insulation, energy sources) and climate will vary by area. 10. Employee Surveys: Periodic employee surveys will be conducted, as warranted, to get specific information on current issues as recommended by the TAC, while it exists, and the EIS. Discussion: This principle reflects OPM's current practice. As it has in the past, OPM is willing to conduct employee surveys as the need arises and as resources permit. This principle makes it clear that such surveys are not mandatory. 11. Out-of-Area Prices: In consultation with the EIS, OPM may use prices for categories other than housing that are gathered in one area to estimate (with appropriate adjustments) prices in other areas within the same region. This does not preclude adoption of adjacent area prices where local prices are unavailable for individual items. Discussion: The principle acknowledges that OPM has authority to decide whether to make expanded use of out-of-area prices (e.g., Maui's Food-at-Home prices for Kauai). Currently, except for catalog pricing, OPM uses out-of-area prices in only two situations: The first is when the item (e.g., a specific brand and model) cannot be found within an area at the time of the survey. In that case, OPM normally does one of two things. If the relative prices for the other items within the subcategory appear to be sound, OPM foregoes the price and increases the weight of the other items in the subcategory. Otherwise, OPM uses the price from another retailer in a neighboring area (e.g., it uses a Kailua Kona price for Hilo) or uses the regional average price (e.g., the average of the price in Anchorage and Fairbanks for Juneau). The second is when the item is not sold within an area (e.g., new car purchase in Nome), in which case OPM obtains the price in an appropriate area and adds the cost of shipping and handling costs to get the item to the survey area. This principle allows expanded use of out-of-area prices. Although this could potentially reduce OPM's workload, research is necessary to see if and where it might be appropriate to adopt such a methodology. Until that research is completed, OPM will not apply the portion of the principle that addresses expanded use of out-of-area prices. 12. Elimination of Components and Cost Buildup: OPM will use the categorization that BLS uses in BLS's publication of CES results, and OPM will use a price relative approach, rather than a cost-buildup approach, to the extent practicable. Discussion: Current regulations specify four components: Goods and Services, Housing, Transportation, and Miscellaneous. Runzheimer, in developing the original model, used a cost-buildup approach for housing and transportation. The components and cost-buildup approach are unnecessary. In their place, OPM will use the aggregation levels reported in the CES and use a straight index approach rather than the cost-buildup approach, to the extent practicable. 13. Elimination of the Survey of Rest of the State of Alaska: Rather than conducting an onsite survey of one or more places in the Rest of the State of Alaska COLA area, OPM may use alternative sources of information to set the COLA rate for this area; provided that if the use of such information indicates that the COLA rate for that area should be reduced, OPM shall re-institute the onsite surveys on one or more locations and publish the results prior to implementing a COLA reduction, if the surveys confirm a reduction is warranted. Discussion: OPM will stop surveying this COLA area and instead will use other available data, e.g., that published by the University of Alaska, which compares many prices and costs in towns and villages in this area to prices and costs in Anchorage, Fairbanks, and Juneau. If in using this methodology the index for this allowance area ever falls below 125, OPM will conduct one or more onsite surveys to verify whether the reductions are warranted. 14. Elimination of Commissary/Exchange COLA Rate: OPM will not make an offset to COLA based on any special purchasing privileges offered Federal employees in the COLA areas. Discussion: OPM will recommend that Executive Order 10000 be changed so that OPM no longer has to make special adjustments when Government quarters or commissary and exchange shopping privileges are provided. The rent for quarters is set according to Office of Management and Budget Circular A-47 at prevailing rates, so no COLA offset should be necessary. Likewise, when commissary and exchange shopping privileges are offered as a recruitment incentive, an offset essentially undoes the incentive. 15. Allowance and Survey Area Definitions: The current allowance and survey area definitions will remain in place until changes appear warranted, at which time OPM will consult with the EIS before making a final determination regarding such changes. Discussion: The principle makes it clear that the current allowance and survey area boundary definitions have been and are appropriate and that OPM does not need to change these definitions. In the future, however, OPM may change the boundary definitions as warranted to meet changing circumstances provided that OPM first seeks the advice of the EIS. 16. Base Area: Base area for living cost and other allowance determinations (but not necessarily weight determinations) will be the Washington, DC metropolitan area. Until the issue is reexamined through an employee survey, the "metropolitan area" can continue to be the current OPM survey area. OPM will consult with the EIS before making changes in the definition of the area. Discussion: This principle affirms OPM's current approach, which OPM will continue. Current regulations, however, are out of date as a result of the Office of Management and Budget redefinition of the Washington, DC, metropolitan statistical area. Therefore, this principle makes it clear which geographic areas are covered in the DC area survey. 17. Interim Adjustments: Prices for housing, goods and services will be collected once every 3 years and rotated by region. Interim annual adjustments will use the existing Consumer Price Index - Urban ("CPI-U") data (relative to base area) as available. OPM will use the Honolulu index for all of Hawaii and Guam; the Anchorage index for all of Alaska; and the Puerto Rico index for the Caribbean. OPM will consult with the EIS prior to making changes to the definition of regions or using different indexes. 18. Hedonic Housing Model and Rental Equivalence: Shelter price relatives will be estimated for owners and renters from the triennial regional sample. The sample for the region will be pooled with the comparison sample from the base area and price relatives for the COLA areas will be estimated using hedonic regression models to adjust for quality differences. Discussion: OPM will adopt a rental-equivalence approach to estimate shelter costs and a hedonic regression approach to compare housing of similar quality. To identify the living communities to be surveyed, OPM will use the results of the 1992/93 employees survey, JPC's survey, and/or other appropriate information. How the housing data will be collected is not known or stipulated. OPM may survey Federal employees, collect the data on its own or through a contractor, enter into an interagency agreement with another Federal agency (e.g., the Department of Interior), or use some other appropriate approach. 19. Natural Disasters: OPM will have flexibility in application of the program in the event of natural disasters. OPM and the EIS will review natural disasters in COLA areas on a case-by-case basis, and OPM may make temporary COLA rate adjustments, changes in methodology, survey timing changes, or other appropriate changes. Discussion: OPM currently has this flexibility. If a natural disaster significantly affects living costs for a substantial period of time (e.g., for a year), OPM will be able to implement an interim rule to conduct an off-cycle, on-site price survey or make some other appropriate change or adjustment. This principle does not require OPM to take any action except as OPM determines is appropriate. 20. Rounding COLA Rates, Decrement, and Threshold: COLA rates shall be calculated to the nearest whole percentage point and shall not be diminished by any decrement greater than 1 percentage point per year. The threshold level will be a COLA index (i.e., the price index plus the adjustment included pursuant to SHP 23) of 102.00. Discussion: OPM will adopt the practice of rounding COLA rates to the nearest whole percentage and reducing COLA by no more than 1 percentage point per year. The limitation on reductions will apply whether the reduction is due to a change in the methodology or relative change in living cost. If survey results cause a COLA index to go below the threshold, the 1-point decrease limitation will continue until the rate is eliminated or the index equals or exceeds the threshold. (For example, if an existing COLA rate is 2 percent and the new index does not equal or exceed 102.00, the 1 percentage point decrement would apply (i.e., the new COLA rate would be 1 percent). The next year, if the index still did not reach or exceed 102.00, the COLA rate would be zero.) 21. Natural Disaster Insurance: Losses due to natural disasters will be included in the price index to the extent they are reflected in insurance costs. Discussion: Homeowner natural disaster insurance is a cost born by the owner and passed on to the renter as part of the rent. Therefore, consistent with the rental equivalence approach, OPM will not survey the price of homeowner insurance. OPM will continue its practice of surveying the price of renter insurance. To the extent that companies in the allowance or DC areas generally offer additional riders to renter insurance policies to cover disaster losses, OPM will include the price of such riders in the price of the renter's insurance it surveys. 22A. Survey Plans and Methodology: Prior to the conduct of each living-cost survey, OPM will consult with the EIS on the details of the survey plans and methodology. Also, OPM will publish for notice and comment any change in the survey methodology that requires a change in regulations. Where, however, no change in the regulations is required, it will not be necessary for OPM to publish for notice and comment the survey plans and methodology in advance of the surveys. 22B. Survey Reports: After each living-cost survey, OPM will continue to publish living-cost survey reports for notice and comment in generally the same manner it currently publishes them. It will not be necessary for OPM to publish survey reports in greater detail or to use some other approach or tabulation. Discussion: OPM publishes the results of its living-cost surveys as a notice in the Federal Register. The publication strives to provide an overview of the survey results with a meaningful level of detail. OPM takes into consideration the comments it receives in two ways. First, as appropriate, OPM makes changes in the way the data it collected are used and analyzed. Second, as appropriate, OPM makes changes in the design of prospective surveys, analytical techniques, and other policies and practices. 23. Adjustment Factor: OPM will adopt the following adjustment factors to reflect differences in need, availability of and access to goods and services, and quality of life. These factors will be added to the price index. Allowance Area Amount Added to Price Index Anchorage, AK 7.0 Fairbanks, AK 9.0 Juneau, AK 9.0 Rest of the State of Alaska 9.0 Honolulu County, HI 5.0 Hawaii County, HI 7.0 Kauai County, HI 7.0 Maui County, HI 7.0 Guam and CNMI* 9.0 Puerto Rico 7.0 U.S. Virgin Islands 9.0 *Commonwealth of the Northern Mariana Islands 24. Employee Involvement Structure ("EIS"): In OPM's work with the EIS, the TAC and/or other committees in all phases of implementation, the parties recognize that OPM must comply with the Federal Advisory Committee Act (5 U.S.C. app. 2) and other relevant law. 24A. Employee Involvement (First Phase): During the period from the date the settlement agreement is approved by the appropriate court(s) until the date OPM issues final regulations to implement the settlement, the Survey Implementation Committee ("SIC") and the TAC will advise OPM as it prepares both its proposed and final regulations. Funding relating to the activities of the SIC and TAC will be provided in the manner set forth in Exhibit D. 24B. Employee Involvement (Second Phase): The SIC will continue to exist during the period from the date OPM issues final regulations to implement the settlement to the end of the first survey cycle in all COLA areas (i.e., during the first 3 years of implementation of the new regulations). Funding for the expenses related to the meetings of the SIC will be provided in the manner set forth in Exhibit D. Prior to each survey conducted in this cycle, the SIC will review the plans and methodology for the survey and provide to the appropriate OPM management official(s) advice or comments. Following each survey, the SIC will again meet to review the analysis of the results of the COLA surveys. During this period, OPM may establish other committees, which may be comprised of agency representatives, collective bargaining representatives, and others, including representatives of the COLA Defense Corporations. Such committees may be established, possibly on a test or pilot basis, pursuant to regulations issued during the first or second phase and may advise OPM concerning any aspect of the COLA program, including the planning and conduct of the survey and analysis of the results. The SIC may discuss the development of regulations providing for ongoing involvement of agency and collective bargaining representatives in the administration of the COLA program beyond the second phase. At the end of the second phase, the SIC will dissolve, and OPM will determine the nature and extent of prospective agency and collective bargaining representatives' involvement in the COLA program by issuing regulations. 24C. Employee Involvement (Third Phase): After the end of the first survey cycle in all COLA areas, OPM may involve agency, collective bargaining, and other representatives in the COLA program and in accordance with any regulations issued during or after the Second Phase. 25. Expert Consultants: A one to three member TAC will be established to advise the SIC and the appropriate OPM management official(s) during the First and Second Phases as needed on economic and statistical issues relating to the COLA program. OPM or the SIC may refer questions involving the application of the settlement agreement, including the Safe Harbor principles and implementing regulations, to the TAC. Funding for the necessary expenses of the TAC will be provided in the manner set forth in Exhibit D. At the end of the Second Phase, the TAC will dissolve. 26. Access to Data: Employees shall be afforded access to survey data within the limits of law. Generally, OPM will consult with the EIS before deciding or agreeing with a third party to keep categories or levels of data confidential. Discussion: This is OPM's current policy. The policy generally favors disclosure.