New Index Provides More Timely Information
by Dave Senf
May 2010
Whoever said “you can’t teach old dogs new tricks” might want to reconsider after reading this article about the replacement of the Minnesota Labor Market Index (LMI) with the Minnesota Coincident Index (Minnesota Index).
For more than two decades the LMI has been around in one form or another, first appearing in a stand-alone publication and then in Minnesota Economic Trends before landing as a monthly segment in the Minnesota Employment Review. The LMI, developed by economists at the University of Minnesota for the Minnesota Department of Jobs and Training [a forerunner of the Minnesota Department of Employment and Economic Development (DEED)], has provided a single-summary metric aimed at answering the question, How is Minnesota’s economy doing[1]
DEED reports on a variety of monthly economic indicators such as unemployment rate, wage-and-salary employment, and initial claims for unemployment benefits that can be used to track and assess the direction and strength of the state’s economy. These various monthly economic statistics, however, often give mixed signals with some indicators hinting at an expanding economy while other indicators suggest a stagnant or slumping economy. For example, unemployment can go up during a month when total wage- and-salary employment climbs and initial claims for unemployment benefits decline. Combining several economic measures into a composite index of economic activity increases the amount of information being utilized while at the same time minimizing conflicting signals. A composite index cuts through the noise, providing a gauge of economic activity that is easier to interpret and more reliable.
Minnesota’s LMI was patterned after the national Conference Board coincident index, which is released monthly along with the widely watched leading economic indicator. The LMI, originally called the Minnesota Economic Index of Coincident Indicators, included seasonally adjusted wage-and-salary employment, seasonally adjusted total weekly manufacturing hours and seasonally adjusted retail trade sales. The publication of retail sales data for Minnesota was discontinued by the U.S. Census in 1997, forcing a revamp of the .[2] Seasonally adjusted average weekly initial claims for unemployment benefits replaced retail sales as an index component, and the index was renamed the Minnesota Labor Market Index as all three components were labor-market-related metrics.
The LMI seemed to be doing its job over the years tracking the state’s economic growth through recessions and expansions until a few years ago when manufacturing employment peaked in Minnesota. Total weekly manufacturing hours (the product of the average number of hours worked by production workers per week multiplied by the seasonally adjusted monthly number of manufacturing jobs) has been on a downward slide as the state’s manufacturing workforce has shrunk. Since total weekly manufacturing hours, not average weekly manufacturing hours, is one of the three components of the LMI, the downward trend has caused the index to underestimate the level of economic activity in the state over the last few years.
The LMI’s growing underestimation problem is apparent when annual changes in the index are compared to Minnesota’s gross state product (GSP) annual changes. The state’s GSP is the broadest measure of the state’s economic activity and is published with a year lag on an annual basis. Since the LMI was designed to be a monthly measure of Minnesota’s economic activity, annual changes in the LMI should closely track GSP changes. The LMI’s track record with respect to GSP growth was respectable through the 1990s but has slipped over the last 10 years. Minnesota’s economy as measured by GSP increased 15 percent between 2000 and 2007 but decreased by 1.7 percent according to the LMI. The LMI has become outdated and in need of an overhaul if it is going to keep pace with Minnesota’s changing economy.
After a little research, it quickly became clear that there was no need to “reinvent the wheel” since the Philadelphia Federal Reserve Bank was already publishing a cutting edge coincident index for every state. The monthly state indices produced by the Philadelphia Fed are based on the state-of-the-art statistical techniques that have supplanted the methodology of earlier sub-national coincident indices 3] The state indices are single measures that capture the common movements of four state-level indicators:
- Nonfarm payroll employment
- Average weekly hours worked in manufacturing
- Unemployment rate
- Inflation-adjusted wage-and-salary disbursements
The common movements of the four components are used in a system of equations to estimate an unobservable variable, which serves as a proxy for the state of each state’s economy. Each state’s index is set to the state’s long-term GSP trend ensuring that the long-term growth in a state’s index tracks long-term GSP growth.
The Philadelphia Fed’s Minnesota coincident index, which is referred to as the Minnesota Index, has several advantages over the old Minnesota Labor Market Index. The use of average weekly manufacturing hours instead of total weekly manufacturing hours (the product of average weekly manufacturing hours multiplied by monthly manufacturing employment) reduces the weight that Minnesota’s manufacturing sector has on the index. A change in the average manufacturing workweek, say from 41 hours per week to 40.5 hours, will still drag the index down for the month but will not push the index downward over the long term just because manufacturing employment is declining. This is a key upgrade as the state continues to move toward an economy dominated by service-providing activity rather than goods-producing activity.
The inclusion of the wage-and-salary disbursement component is also an improvement since additional information about the economy is being incorporated. Another advantage of the Philadelphia Fed’s index is that changes in Minnesota’s index can easily be compared to other states since each index is calculated with the same statistical technique using the same four economic indicators. Minnesota’s recessions and expansions in terms of depth, duration, and dates can quickly be compared to any other state’s business swings.
Figure 1 compares the old Labor Market Index and the new Minnesota Index. The good news is that both indices bottomed out last fall and have been creeping upward, implying that Minnesota’s economy has been on the mend since last September or October. The shortcoming of the LMI as a reliable monitor of Minnesota’s economic activity in recent years is apparent by the index’s path since peaking in 2000. While Minnesota’s economy contracted during the 2001 recession and struggled during the two years of jobless recovery that followed, the state’s economy did eventually expand beyond the 2000 level. Inflation-adjusted personal income, real GSP, and total employment were all higher in 2007 than in 2000.
The LMI never rebounded above the 2000 peak, giving a false impression that Minnesota’s economy was smaller in 2007 than in 2000. The new Minnesota Index shows 13.2 percent growth between 2000 and 2007, which is in line with the 15 percent growth reported by the state’s GSP estimates. The LMI was a first generation coincident index, useful during the ‘80s and ‘90s, that needed to be replaced in the new century by a second generation index, the Minnesota Index.
1]The original LMI research is reported in Minnesota Economic Indicators: Part I Purpose and Precedent and Part II Methodology and Findings, Maki, Wilbur R. and Peter Stenberg. St. Paul, MN: University of Minnesota, Department of Agricultural and Applied Economics, 1988. (Staff paper P88-36) at http://ageconsearch.umn.edu/handle/14014 and http://ageconsearch.umn.edu/handle/14255.
2]The switch to the current version of the Minnesota Labor Market Index was covered in the June 1999 issue of Minnesota Employment Review. The article is currently unavailable online because of Web page redesign but is available via e-mail by calling the Minnesota LMI help desk at 651-259-7384 (TTY-1-800-657-3973).
3]The Philadelphia Federal Bank’s monthly state coincident indices are available at
http://philadelphiafed.org/research-and-data/regional-economy/indexes/coincident/
Top