ACCOUNTING FOR CHANGES IN SOURCE DATA: NOAA FISH LANDINGS
- We use NOAA’s landings data for initial estimates of commercial fishing output for those states covered by NOAA (coastal states and the states surrounding the Great Lakes). However, starting in calendar year 2020, NOAA changed their data provider for the Gulf States (AL, LA, MS, TX); NO to GulfFIN, and according to our contact at NOAA, GufFIN redacts their confidential data instead of including it as some form of aggregate. As such, we no longer use the current NOAA data for these states; instead, we project the 2017 NOAA values for these states using the change in Employment and Employee Compensation in this sector in these states.
- We had been relying on BLS CEW data to a) estimate commercial fishing output for states not covered by NOAA and b) project the lagged output estimates to the current IMPLAN year. However, since a) CEW does not fully cover the Commercial Fishing sector, b) the degree of coverage very likely varies by state, and c) CEW data do not include proprietors, we now use Employment and Employee Compensation from later stages in the data production process to account for proprietors and adjustments for such under-coverage.
- NOAA’s new species-specific data (and thus also their aggregations across species, by state) include numerous instances of pounds reported with no accompanying dollar value. We assume these are a mixture of onsite consumption (similar to on-farm consumption of production) and by-catch (fish caught as a byproduct and either sold but not reported or thrown back). We estimate an average dollar per pound for these species across all states and apply that value to those pounds missing a dollar value. Any remaining instances lacking dollar values are assumed to have been catches thrown back (i.e., true NULLS) and are not further adjusted.
IMPROVED STATE-LEVEL DISTRIBUTION OF PERSONAL PROPERTY TAXES
Personal property taxes are not distinguished from other types of property taxes in the Annual Survey of State and Local Government Finances (including in Census years) or the Annual Survey of State Government Tax Collections Data. We had been using this aggregate property tax to distribute state-level personal property tax, but beginning with the 2019 IMPLAN data set, we now use more-detailed data from the BEA REA SA50 table for this initial state-level distribution.
IMPROVED STATE-LEVEL DISTRIBUTION OF PERSONAL CAPITAL INCOME
Beginning with the 2019 IMPLAN data set, national values for personal capital income, consisting of dividend income, interest income, and rental income of persons, is distributed using data from BEA REA table SA40, which reports each of these items for states. Previously, we had used a single, state-level, aggregate value to distribute those three components of personal income.
IMPROVED OUTPUT ESTIMATES FOR OIL & GAS EXTRACTION
Because the price of crude oil is so volatile, growth in Employment and Labor Income in this industry does not always reflect growth in Output as closely as in other industries whose products do not experience such price volatility. Thus, using growth rates based on Employment or Labor Income to project the lagged BEA Output values is not ideal for these sectors. Therefore, starting with the 2019 IMPLAN data, we will be using the Energy Information Administration (EIA) data on the prices and physical production of crude oil and natural gas to project the lagged BEA Output value for the oil and gas extraction sector.
The benefits of this change are two-fold: in addition to improving the Output estimates for the oil and gas extraction industry itself, these improved Output estimates will result in more accurate estimates of the petroleum refinery industry's purchases of crude oil, thereby yielding improved impact estimates.
NO LONGER ESTIMATING NON-DISCLOSED CBP VALUES
The Census Bureau has adopted a new policy under which they no longer provide establishment counts for cases in which the establishment count is less than three, as the number of establishments is now considered sensitive. This new decision to omit from their tables all records with fewer than three establishments was a new policy that was set in place to protect the confidentiality of businesses.
This new practice makes a missing record of 1-2 establishments (an existing industry) indistinguishable from a missing record of zero establishments (a non-existent industry). The omission of records with less than three establishments, in addition to making it impossible to estimate their employment and income values, also makes it impossible for us to obtain high quality estimates for the non-disclosed records with three or more establishments, since we are not able to roll up across NAICS levels without having values for sibling sectors.
Therefore, beginning with the 2019 IMPLAN data set, we no longer create estimates for non-disclosed CBP data; rather, we only use the disclosed data. This will affect our estimates of non-disclosed BLS CEW values in those cases in which we need to turn to CBP (lack of disclosed raw CEW from previous year) and in which the CBP value is not disclosed; in those cases, we will now move right to using a projected CEW value from two years back, if possible, followed by the use of state-level or U.S.-level ratios as the second and third options for obtaining a first estimate for CEW, rather than having CBP as a second option.
NO LONGER USING ANNUALLY-UPDATED ZBP VALUES
The changes to the Census Bureau's disclosure practices described above also affect the Census ZIP Code Business Patterns (ZBP) data, beginning with the 2017 data. Thus, the most recent set of ZBP data that are suitable for use in IMPLAN is the 2016 ZBP data. As described in this article, the ZBP data, along with zip-code-level demographic data, are used to split county-level data into zip-code-level data. The zip-code data are then aggregated into Congressional District data. An unfortunate consequence of these changes, therefore, is that the zip-code and Congressional District data will rely on increasingly lagged ZBP data for distributing county-level industry values to zip-codes.
INCLUSION OF 2017 CFS DETAILED DATA
The Census releases new Commodity Flow Survey (CFS) data every five years, first as a small preliminary data release and then later as a larger detailed data set. IMPLAN incorporates the CFS data over two years. The 2018 IMPLAN data incorporated the Preliminary 2017 CFS data on commodity shipment characteristics. The preliminary data are used to revise estimates of average miles shipped for shippable commodities, which are used as a calibrator for the gravity model, and estimates of shipment-modes by industry. The new 2019 IMPLAN data incorporate the newly released detailed CFS data (released in July 2020), which include greater commodity specificity on average miles shipped and shipment-modes. These data are used to further revise both sets of estimates.
NO LONGER USING CENSUS COUNTY BUSINESS PATTERNS BY LEGAL FORM OF ORGANIZATION DATA
Since the 2013 IMPLAN data year, we had been using these data in combination with BLS CEW data to distribute BEA REA proprietor employment and income, which are in a more aggregate sectoring scheme, among the IMPLAN sectors that map to each BEA REA sector. Due to the Census Bureau’s new practice of omitting records with fewer than three establishments (described in more detail above) and the fact that state-level data by industry and legal form of organization often have records with fewer than three establishments, these data are not sufficiently reliable in their current form. Therefore, we have returned to the previous method of relying solely on FLS CEW data for the distribution.
IMPROVED STATE AND LOCAL GOVERNMENT SALES BEHAVIOR
Prior to the 2019 IMPLAN data year, the only State and Local Government institution that had sales was State and Local Government Non-Education (with State and Local Government Education getting its funding through State and Local Government Non-Education). With the incorporation of the new 2012 BEA Benchmark tables, beginning with the 2018 IMPLAN data set, we now have an additional State and Local Government institution: State and Local Government Hospitals and Health Services. Note: this split out is not available in IMPLAN Pro which only contains 544 Industries.
With the addition of the new Hospitals and Health Services institution, it should ideally be this new institution, rather than the State and Local Government Other Services institution, which sells the following four commodities: Outpatient care centers (3486), Home health care services (3488), Other ambulatory health care services (3489), and Hospital services (3490). Starting with the 2019 IMPLAN data set, this is now the case.
In addition to providing information on the goods and services sold by the State and Local Government Hospitals and Health Services institution, the new 2012 BEA Benchmark also provides information on the goods and services sold by the State and Local Government Education institution. Therefore, it should ideally be this institution, rather than the State and Local Government Other Services institution, which sells the following 4 commodities: News syndicates, libraries, archives and all other information services (3437), Elementary and secondary schools (3480), Junior colleges, colleges, universities, and professional schools (3481), and Other educational services (3482). Starting with the 2019 IMPLAN data set, this is now the case.
Finally, we also added NIPA control totals for the sum of the four commodities sold by each of these institutions. These changes will have no effect on impact analyses run from an industry basis. They will have very minor, if any, effect on commodity-based impact analyses (it will be case-dependent), but results (and model data) will more accurately reflect the actors involved in commodity production.
NEW OTHER VALUE-ADDED PROCEDURES
Prior to the 2019 IMPLAN data year, the BEA data we used for state-level industry-specific Gross Operating Surplus, Gross Taxes on Production and Imports, and Subsidies were lagged one year relative to the IMPLAN data. As such, we used other data sources to project the BEA data, as described in the article Estimating and Distributing Value Added. Beginning with the 2019 IMPLAN data, the BEA is now publishing these data in a much more timely fashion that allows us to use data that are in the same year as the latest IMPLAN data being produced, thereby doing away with the need for projection.
INCORPORATION OF THE 116TH CONGRESSIONAL DISTRICTS
A new relationship file between Zip Code Tabulation Areas (ZCTAs) and Congressional Districts has been published by the Census Bureau. This new relationship file has been incorporated into the development of IMPLAN’s 2019 Congressional District data.
INCORPORATION OF THE 2017 CENSUS OF GOVERNMENTS
The Census of Governments is conducted every 5 years, with the latest one being conducted in 2017 and published in 2020. We use the Census of Government data on State and Local Government Finances to supplement the Annual Survey of State and Local Government Finances, the latter of which is released two years lagged relative to the IMPLAN data. We start with values from the latest Census of Government (inflated to the same year as the annual survey data. We then replace any Census values with values from the Survey, where available. These values are used to distribute U.S. control values to the states and counties. The incorporation of the new Census data may cause some significant changes to tax impacts.
The Occupation Data for 2019 was updated February 10, 2021.
Written November 11, 2020
Updated February 11, 2021