"Reflections on the Service-sector Expansion and Non-Residential Construction Initiative in the PPI: Progress to date and where do we want to go in the future?"
May 2006 Monthly Luncheon
Speaker: Michael W. Horrigan, Assistant Commissioner
Affiliation: Division of Industrial Prices and Price Indexes, Bureau of Labor Statistics
Rapporteur: Anthony Nunes
In his presentation, Mr. Horrigan focused on the recent history and future direction of the Producer Price Index (PPI). He started out by providing a brief overview of the PPI program and the recent expansion in its coverage of service and non-residential construction industries. Horrigan also discussed the importance of finding balances between four major PPI objectives in any future planning: 1) to modernize PPI infrastructure through the JANUS program; 2) to collect more data and produce more indexes; 3) to increase the quality of PPI data; and 4) to make better use of PPI data.
The PPI measures selling prices received by domestic producers for their output. The PPI first started measuring service sector industries such as railroad operations and petroleum pipelines in the mid 1980s. With few resources, the PPI program had to choose services that were relatively easy to define and quantify. Nevertheless, by 2000 the PPI program was capturing roughly half of the in-scope service sector output. Then, in the fiscal year of 2001, the BLS received resources to widen coverage to the construction sector for the U.S. economy and to enhance the ongoing expansion of PPI coverage of the service sector.
With the additional resources, the service sectors covered by the PPI program increased to 58.8 percent in 2004. With the introduction of two banking industry indexes in January 2005, the percentage covered by the PPI program increased to 64.5 percent. In July 2005, the PPI program added several new areas including merchant wholesalers; fitness and recreational sports centers, security and patrol, and internet service providers. These new service areas dramatically increased the in-scope service sector to 76.5 percent. In 2006 the PPI staff thoroughly researched and added amusement parks and golf courses. For next year, the program is adding management consulting services, and in 2008 it will include blood banks, data processing and repair shops. These additions should increase coverage to 77.4 percent.
Another major initiative in the PPI program is the Janus project: a major multiyear project to convert all existing mainframe applications to client server environments. This requires substantial research to develop the software protocol rules, estimation formulas, and seasonal adjustments as well as in-depth analysis of price indexes based on item-level data. As a result of the Janus project, PPI analysts will be able to align individual products with multiple indexes providing much needed flexibility in the construction of alternative aggregation indices.
In order to improve data quality, the PPI program has introduced adjustments for quality change, either through direct measures of the input costs of quality change or, as needed, through the use of hedonic models. For example, a desktop computer may cost a consumer $100 more than the same model the previous year; however quality improvement such as greater processing power may offset, partially or wholly, the price inflation. In order to account for quality changes to the fixed market basket of goods and services being repriced, PPI Industry analysts must stay up-to-date on industry trends. In the case of the automobile industry, for example, analysts travel to Detroit to talk to manufacturers about their last product changes.
Horrigan also described another PPI improvement initiative designed to reduce price index bias owing to the failure to capture rapidly evolving product lines entering the market. With the rapid entry of new products and services, some items might be left out of the PPI sample even if it’s a direct substitute of another in-sample product. By augmenting an existing sample to include new products and producers that did not exist when the original sample was taken, the PPI program reduces the risk of new technologies’ being left out of their pricing calibrations. For example, with sample augmentation, a new generic pharmaceutical can now be captured by PPI without having to wait until the next sample is taken.
Horrigan then turned to the future of the PPI program. One of challenges facing the PPI program is the impact of globalization on the classification of manufacturing industries. In our modern economy, some firms that before were simply manufacturing goods may evolve into wholesalers though the outsourcing of production processes to other countries. Capturing these changes is often difficult.
Horrigan went on to discuss significant remaining gaps in industry coverage by the PPI program. Once the service sector and non-residential construction initiative has been completed, the PPI program will have 82 percent of all U.S. production covered. In evaluating a data-driven approach to future PPI improvements, the emphasis should be on customer needs. User demand from other federal and state agencies, trade associations and academia is invaluable to prioritizing future PPI quality improvements.
Horrigan gave a few examples of the feedback that he has received from end-users. Former Chief Economist for the Bureau of Economic Analysis Jack Triplett believes that educational and business services are significant coverage gaps, while others point to a need for more detail in the wholesale trade and health services.
Horrigan closed his presentation with a few final thoughts on the long-term future of the PPI program, stating that PPI is going forward with a “balanced strategy of adding easier-to-calculate industries in which output and quality adjustments are harder to conceptualize and operationally measure.” He is confident that ongoing improvements will ensure that the PPI program will continue to meet end-user needs.
Back to Lunch Archives