FOR DISCUSSION ONLY

 

 

 

 

 

Developing Current Cost of Living Indices for Southern Regional Educational Board States

 

 

 

 

 

 

 

 

 

 

 

 

 

Prepared for

 The Division of Administration

Education Estimating Conference

 

 

 

Prepared by

Raymond J. Brady. D.Sc.

Systems Solutions Consulting

July 2001

 

 

Objective:

 

This research has two objectives: (a) to update the 1996 AFT (American Federation of Teachers) Interstate Cost of Living Index used as one variable to develop a weighted average teacher salary for Southern Regional Educational Board (SREB) states, and (b) to develop a methodology to continuously update these indices.

 

Introduction:

 

Currently, no official state-to-state or MSA-to-MSA cost-of-living indices exist in the United States. The Bureau of Labor Statistics prepares a Consumer Price Index (CPI) for the United States and select areas, but the CPI, per se, is not a cost-of-living index. Rather, it is a time series index that measures the change in the value of a basket of goods and services, and is a measure of inflation. A 1999 cost-of-living index is available from the AFT based upon data gathered from the American Chamber of Commerce Association, but the selection of areas covered is insufficient to develop state level indices necessary to calculate the cost-of-living differences between SREB states. The indices developed in this report are specific to SREB states, and MSA’s within that region. The data presented covers parts of the year 2000 and the first quarter of 2001.

 

Accuracy:

 

These indices were developed from secondary data gathered by organizations in the business of developing comparative measures of living costs. Although the data comes from three different web sites, actual data comes from two sources: Dowden and Company and the Center of Mobility Resources. To maximize the statistical usefulness of these data sources, a composite index was developed which averages all the sources.  A simple correlation analysis was run on the three data sets to identify co-linearity, and hence, whether one data set was in fact a sub-set of another. The results are:  (a) column 1 data set has a correlation of 0.54 with column data set 2 and a correlation of 0.38 with column 3 data set , and (b) column 2 data set has a correlation of 0.45 with column 3 data set. What does this mean? Obviously, the lower the correlation the greater the independence of the two data sets. Columns 1 and 2 have the highest correlation. Since we do not know the accuracy of any one data set, pooling data seeks to minimize errors in any one data set.  

 

Methodology:

 

A) Sample Size

 

Data were collected from approximately one-hundred-seven-teen Metropolitan Statistical Areas (MSA) that comprise the sixteen SREB states. Table 1 shows the percent of population within each SREB state that is concentrated in MSA’s. Independent of primary data collection issues, it is important to assure that the sample size within each state is reflective as nearly as possible of the state as a whole, since the objective is to develop state level cost-of-living comparative indices. Overall, the fraction of total state population residing in the MSA’s of each SREB  ranges from a high of 93% in Delaware to a low of 21% in Mississippi. If the data gathered by the various research organizations reasonably reflect of the cost-of-living in the MSA’s, it appears that the samples, when developed as a composite index, would likely be representative of the cost-of-living for the SREB states.

 

A composite index for Mississippi would likely have the weakest representation for the cost-of-living for that state, but this assessment is not conclusive. Rather, it reflects the assumption that the lower the fraction of the state’s total population in an MSA, the lower the probability that the sum of the MSA’s cost-of-living indices could be extrapolated (correctly) to a statewide average.  

 

Table 1

Fraction of Estimated State Population Residing in Metropolitan Statistical

 Areas, 1999

 

SREB State                                 MSA Fraction of Total State Population

 

Alabama                                                            69%

Arkansas                                                            47%

Delaware                                                            92%

Florida                                                               93%

Georgia                                                             70%

Kentucky                                                           44%

Louisiana                                                           75%

Mississippi                                                          21%

North Carolina                                                   69%

Oklahoma                                                          59%

South Carolina                                                    62%

Tennessee                                                          72%

Texas                                                                85%

Virginia                                                             47%

West Virginia                                                      37%

 

Source: Bureau of Census; Systems Solutions Consulting, July 2001

 

 

B) Secondary Data Collection

 

The first step in developing state level cost-of-living indices is to develop MSA comparisons. This requires the identification of a benchmark MSA to compare all other MSA’s in the sample. Baton Rouge MSA is used as the benchmark. Therefore, every other MSA in the sample is compared against Baton Rouge. The cost-of-living in Baton Rouge is set at $1000 and that variable cost is compared against all other MSA’s. Data are collected from each of the sources and stored in a spreadsheet format.

 

The second step requires adjusting the “raw” data collected from the various sources, since the objective is not to establish MSA to MSA cost-of-living comparisons, but state-to-state cost-of-living comparisons. Within Louisiana, Baton Rouge MSA is compared against each other MSA within Louisiana. The ratio of the Baton Rouge value to the weighted sum (weighted by the population within each Louisiana MSA) of the Louisiana MSA’s minus Baton Rouge MSA is used a as weigh to adjust all other non-Louisiana MSA’s for each data set. Why? Because spatially, we need a index reflecting Louisiana’s cost-of-living vis-à-vis all the non-Louisiana MSA’s in the sample. So now, instead of comparing Baton Rouge MSA to all other MSA’s, we have created an index that compares Louisiana against all sample MSA’s.

 

For example, using index 1 in which we compare Louisiana against Birmingham MSA, one can see that if the cost-of-living in Louisiana is $1,000; it is $944 in Birmingham. That is, it would require the person to earn only $944 in the Birmingham MSA to get the same “standard of living” that would require a $1000 in Louisiana.

 

c) State Cost-of-Living Index and Weighted Index Deflator

 

The third  step is to create a cost-of-living index for each state by multiplying the population of each MSA by the cost-of-living index for each MSA,  summing and dividing that sum by the total MSA population in that state to derive a state level cost-of-living index. Finally, the composite value is divided by 10, resulting in an index value based around 100. Since this index measures how much more or less expensive it is to live in one of the SREB states and MSA’s within those states vis-à-vis Louisiana as a whole, we need to develop a deflator is trying to adjust values of the SREB to Louisiana. The formula to do so is quite simple:

 

                                    (1/Cost-of-Living Index (i)) * 100 = weighted index deflator.

 

Table 2 Identifies the State Level Cost-of-Living Indices and Weighted Index Deflator for Each of the SREB States

  

 

Table 2

State Level Cost-of-Living and Weighted Cost Deflator Indices For Each of the SREB States for 2001

 

State                          Cost of Living Index         Deflator Index

 

Louisiana                                               100                                          100

 

Alabama                                                101.5                                       0.985

Arkansas                                                  95.9                                       1.043

Delaware                                                109.7                                       0.912

Florida                                                   105.7                                       0.948

Georgia                                                 106.1                                       0.943

Kentucky                                                 97.1                                       1.030

Maryland                                               112.3                                       0.890

Mississippi                                                99.0                                       1.010

North Carolina                                       104.2                                       0.960

Oklahoma                                                95.8                                       1.044

South Carolina                                        104.3                                       0.959

Tennessee                                                98.5                                       1.015

Texas                                                      98.9                                       1.011

Virginia                                                 102.6                                       0.975

West Virginia                                            96.8                                       1.033

 

 

Source: Systems Solutions Consulting, July 2001               

 

 

How would you apply the weighted deflator to determine the impact of the cost-of-living on teachers salaries? Let’s create an example.

In 1999-2000, the average teacher salary in Louisiana was $33,109. In Alabama, it was reported at $36,689 and in Arkansas at $33,386. Louisiana’s teacher salary is multiplied by 100, since we are concerned about the relative value of the average teacher salary vis-à-vis competing states in current dollars. In the case of Alabama, multiply $36,689 X 0.985 = $36,138. This is the relative purchasing power of a teacher salary in Alabama vis-à-vis the $33,109 for Louisiana or approximately $550 less. In the case of Arkansas, the cost-of-living is substantial less than Louisiana. Therefore, $ 33,386 X 1.043 = $34,821. In short, the purchasing power of a $33,386 is approximately $1,400 greater. Put another way, a teacher need only earn $31,751 in Arkansas to maintain the same standard of living of a Louisiana teacher earning $33,109.

 

 

WHY USE COST-OF LIVING MEASURES?                                                                                      

Two fundamental reasons exist to use cost of living measures. First, they reflect the geographical differences in the relative purchasing power of income. That is, they explicitly adjust current values by the relative cost in each area measured. This provides a more realistic understanding of the comparative value of two incomes in two geographical areas. Second, if one has a public policy that seeks to develop income parity with other states or regions on salaries, it is important to use relative cost indices because they reflect the true differences in income among competing geographies.

 

Finally, adjusting values for the relative cost of one geography vis-à-vis another is good public policy because it gives policy makers a more realistic understanding of what public resources would be required for parity or whatever target is agreed upon.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APPENDIX

 

COST OF LIVING INDICES FOR SREB STATES AND MSA’s