The U.S. control totals for the Household Personal Consumption Expenditures (PCE) come from the BEA’s National Income and Product Accounts (NIPA) for the current year.  Beginning with the 2017 IMPLAN data, we now also incorporate state-level PCE data from the BEA (which are lagged one year but controlled to the current NIPA totals).  There are about 100 NIPA expenditure categories,  which we distribute to the IMPLAN sectoring scheme using more detailed PCE data from the latest BEA Benchmark I-O.  

The BEA data are for all household income groups combined; to break these out by household income group we turn to the Bureau of Labor Statistics’ Consumer Expenditure Survey (CES) data, which are lagged but provide estimates of expenditures by various income categories.  

These expenditure data are in terms of purchaser prices, so we must margin the data; that is, we must split the purchaser values among the producer price and any transportation, wholesale, and retail margins. This is done using data from the latest BEA Benchmark. These values are then matched to the appropriate producing, transportation, wholesale, and retail sectors, resulting in an allocation of PCE spending across IMPLAN sectors for each IMPLAN household income group.

For each household income category, these national spending values by IMPLAN sector are converted to coefficients (proportions of total PCE). To get state-level and county-level PCE expenditures by IMPLAN sector and household income category, these national coefficients are multiplied by each state and county’s spending by household income category. Therefore, the spending pattern for each income class is constant across the U.S. While the CE data is reported by large region, analysis performed by IMPLAN showed no statistically significant difference between regions.

INTRODUCTION:

Both Margins and Deflators are included in the IMPLAN database. Margins allow for consumer expenditures to be traced though retail, wholesale, and transportation Industries back to the industries who manufactured the product, allowing the appropriate allocation to the producing Industries. Built-in Deflators allow for adjustments of the Dollar Year you enter on the Impacts screen and the Dollar Year you see in your Results. Note: margins and deflators in IMPLAN are not regionally specific.

 

MARGINS:

Most Input-Output models, including IMPLAN, record expenditures in producer prices. This allocates expenditures to the Industries that produce the goods or services. Any Output or sales you want to apply to multipliers that are in purchaser prices (prices paid by final consumers) need to be converted to producer prices or allocated to the producing Industries. Margins enable the move from producer to purchaser prices or vice-versa. Data on margins comes from the BEA Benchmark I-O tables.

Below is an example to show how a purchase is allocated with Margins. Assume that a consumer spends $150 at a retail store. A portion of that price, $20 in this case, is retained by the retailer. Another portion, $30 goes to transportation costs, and $100 goes to the producing Industry that actually made the item.  

 

 

The only Industries that can be margined in IMPLAN are retail and wholesale. Nearly all Commodities, on the other hand, can be margined.  This file shows the 2018 Margins for Industries and Commodities.

 

 

INDUSTRY EVENTS

This feature is only available for retail and wholesale Industry Events. To apply margins for an Industry Event, open the menu. Here you will have the option to choose between Total Revenue (Purchaser Price) and Marginal Revenue (Producer Price). Most times, we only know the Purchaser Price, so leaving Total Revenue selected is the right choice. The Results in this case will only show the retail margin and its impact. Your Direct Effect to the retail Industry will be smaller than the Total Revenue you entered on the Impacts screen. All of other pieces of the value chain are lost (production, transportation, and wholesale).

If Marginal Revenue was chosen, the full $1M would be the Direct Output in your Results. This implies that you are modeling what the retailer is keeping (instead of just a portion of the item cost).

 

 

COMMODITY EVENTS

To apply margins in a Commodity Event, open up the menu icon and select between Total Revenue (Purchaser Price) and Marginal Revenue (Producer Price). Again, we usually know the Total Revenue, so the default selection is fine. The Results in this case will show us the impacts on the entire value chain for this Commodity; production, transportation, wholesale, and retail, with Direct Effects in each (when applicable).

If Marginal Revenue was chosen, you would again expect to see the full $1M applied to the Commodity 3366. The only deduction you might encounter is if some of the product was taken from inventory or produced by the government. 

 

 

 

DEFLATORS:

Deflators are used to adjust for relative price changes over time. Output deflators are Industry specific and are used to adjust Industry Output. GDP deflators are not Industry specific and are used to adjust Final Demand and Value Added. Output and GDP deflators from the BEA are used for all past years, while BLS output deflators are used for future years.

The Bureau of Economic Analysis (BEA) provides historical Output deflators which we use for past to current years. The BEA Output deflators are provided with the BEA Gross Output data.

The BEA also has historical GDP deflators which we use for past to current years. The BEA GDP deflators come from NIPA Table 1.1.9 – Implicit Price Deflators for Gross Domestic Product. The BLS produces time-series of Output estimates for its Employment Growth Model. The Outputs are projected in real and constant dollars. This gives implicit price index projections which are the basis for projections of the IMPLAN deflators.

Both the BEA and BLS deflator data have fewer Industries than the IMPLAN Industry scheme; therefore, all IMPLAN Industries within a single BEA or BLS Industry will have the same deflator. This file shows the 2018 Deflators for Industries and Commodities.

 

DETAILED INFORMATION:

The purchasing power of a dollar changes over time (typically decreasing) due to inflation, a cyclical phenomenon by which prices of goods and services increase[1], which spurs workers to demand higher wages, which in turn increases demand for goods and services, thereby spurring additional price increases, and so on.  Due to inflation, a dollar in 2017 cannot purchase as much as did a dollar in 2001, for example; as such, a 2017 dollar is not the same thing as a 2001 dollar.  IMPLAN’s deflators are indexes of inflation, with the deflator for the model data year set at 1.00. 

The deflators are not used to create the social accounts or multipliers but are necessary for impact analysis whenever the Dollar Year of the event differs from the Data Year being used. The same model year multipliers are used regardless of the Dollar Year of the event; it is the value applied to those multipliers that changes when the Dollar Year of the event differs from the Data Year.

All the relationships in the multipliers are based on model year prices, so the Direct Effects applied to those multipliers need to also be in the correct Dollar Year – this is accomplished via the deflators. The value applied to the multipliers is the user-entered value divided by the deflator. The deflators also allow impact results to be viewed in years other than the model year, regardless of whether or not the Dollar Year of the event differs from the Data Year. 

While the Event values and/or result values can be inflated or deflated, depending on whether the index value being applied is less than 1.00 or greater than 1.00 (i.e., depending on the Industry or Commodity and whether one is adjusting to a future or past value), we use a single term – deflators – to refer to all of these index values.

The Output deflators are specific to the Industry or Commodity and are applied to the Output value, while GDP deflators are the same for every Industry and Commodity and are applied to all of the value-added components.

Margins are derived from the Bureau of Economic Analysis Input-Output tables.  Margins are particularly important for Personal Consumption Expenditures (PCE) values as nearly all household purchases of goods are through a retail Industry. The Margins used to form the PCE data elements are compiled from the BEA Detailed Benchmark tables. This data provides the Margins associated with each of the different Personal Consumption categories. These PCE categories are modified to fit IMPLAN Industry definitions.

 

DOLLAR YEAR & DATA YEAR:

By default, impacts will be reported in current year dollars; however, because IMPLAN data are typically lagged a year (i.e., 2018 data were released in 2019), it is handy to be able to report the results in current year dollars. This can be achieved by changing the Dollar Year on the Results screen. This is just an option, and is not the same thing as changing the Dollar Year on the Impacts screen. The Dollar Year must match the year that the dollars represent – this ensures that the correct value is applied to the multipliers.  

Suppose you are going to model the impact of the 500,000 visitors that came to your tourist attraction in 2020.  Also suppose that you didn’t conduct your own visitor expenditures survey and are thus borrowing a survey that was conducted on a similar tourist attraction in a similar region but way back in 2010.  That survey gives you the per-tourist expenditures on things like lodging, food, transportation, and entertainment.  For example, each tourist spent $200 on lodging during their 3-night trip to that attraction. If you were to set Dollar Year to 2020 and put in $200 you would be understating your impact because 3 nights at a similar hotel would cost more than $200 in 2020 due to inflation!  So you’d want to set the Dollar Year to 2010, since that is the year that those $200 represent.  IMPLAN will inflate accordingly and apply the value to the multipliers. 

More information on on this can be found in the article Dollar Year & Data Year.

 

2018 MARGINS Margins represent the value of the wholesale and retail trade services provided in delivering commodities from producers’ establishments to purchasers. This file contains the four components of the Value Chain: retail, wholesale, and transportation margins for each Industry along with the Producer value.	Download
2018 DEFLATORS Deflators are used by the software whenever the Event Year is set to a year that differs from the model Data Year. This file has the deflators/inflators for 1997-2060.	Download

Databases are available for the U.S., all 51 states (includes Washington, D.C.), 5 US Island area territories, and all 3,000+ counties in the United States. 30,000+ zip code level files are available. We also have a number of international data sets. You can view information on data selection on our website.

All IMPLAN datasets come in .ODF format and include a set of national I-O matrices and tables that allow the software to create a complete set of regional social accounts. Each U.S. data file – whether you are working at the zip code, territory, state, or national level – includes economic and demographic variables at a 536* industry/commodity sector level (defined by NAICS). Data variables include employment, income, taxes, and transactions by governments and households.

IMPLAN databases are constructed exclusively by IMPLAN Group LLC. IMPLAN has U.S. data available from 1996 forward. For trade flow estimate, we provide a National Trade Model for 2007 data sets and newer for US states and counties. Previous data years, US territories, and zip code files require use of the Econometric RPC trade flows method. National models (US and foreign countries) have foreign imports and exports pre-defined as part of the data base.

Additionally, all IMPLAN data sets include the software and data to build complete Regional SAM Multipliers. The software also provides a framework to easily perform impact analysis and creates reports for analyses and virtually all data elements.

For 2007 and later data, Trade Flow data are included with your annual database (in the Version 3.0 system only). The Trade Flow data available for county level and higher data sets provides all the tools necessary to build Multi-Regional I-O (MRIO) models, whether you are looking at flows between counties, between states, or between aggregated model regions. Currently, the MRIO functionality is only available in IMPLAN Pro. 

 

*Number of sectors depends on the data year. For US files 2013 (and later), data sets are 536 sector; 2007-2012 are 440 sector; 2001-2006 are 509 sector and 1996-2000 are 528 sector.

IMPLAN employment includes both wage and salary employees and self-employed persons in a region. Full-time, part-time and seasonal workers are measured to create an estimate of annual average jobs.

BLS Covered Employment and Wages (CEW) data, BEA Regional Economic Accounts (REA) data, and County Business Patterns (CBP) data are used in conjunction to create IMPLAN data because no one dataset provides enough information to create a complete IMPLAN database.¹ In general, CEW data provide the county level industry structure for IMPLAN, while CBP data are used to make non-disclosure adjustments to CEW data. REA data are used as controls for data not covered by CEW and proprietors.

CENSUS COUNTY BUSINESS PATTERNS (CBP)

County Business Patterns (CBP) is a program run by the U.S. Department of Census. Employment numbers are a count of employees during the week of March 12. This is a point-in-time estimate and not an annual average.

Data at the 6-digit NAICS level of detail include: total number of establishments, total first quarter employment, first quarter current year and total annual payroll, and a breakdown of the number of firms for 12 different employment size classes. As might be expected with 6-digit level specification, there are significant disclosure problems. However, even when the sector data are non-disclosed, CBP provides the number of firms by employee size class. The CBP also excludes most government employees and farm sectors.

The CBP data give a picture of the industrial structure of a region and are used to adjust the CEW data for non-disclosure. There is a time lag, generally one year, between the current year and the most recent CBP data, but an industrial structure generally changes slowly over time. There are virtually no disclosure problems with the national-level CBP data.

BLS COVERED EMPLOYMENT AND WAGES (CEW)

The CEW dataset is one of the most important datasets used in IMPLAN database development. These data provide the industry structure for the states and counties and the “ground truth” for IMPLAN data. There can be many differences between CBP and CEW, but in the end, if wage and salary employment doesn’t exist in CEW, it won’t exist in IMPLAN data sets. The data are provided by the U.S. Department of Labor as part of the Unemployment Insurance Covered Employment and Wages Program.

The CEW dataset provides annual average wage and salary establishment counts, employment counts, and payrolls by county at the 6-digit NAICS code level. These data are collected from a federal/state partnership program. State employment services departments, as part of the Unemployment Insurance Program, collect the data and pass it to the U.S. Department of Labor. As a result, only establishments that pay Unemployment Insurance are captured, hence the name “Covered Employment”. Since these data only capture covered employees, the data set cannot capture self-employed persons, railway employment, religious organizations, military, elected officials, or any other establishments that have their own social insurance program and/or do not pay into the Unemployment Insurance program. Since most farm employment is self employment, CEW data misses much of the farm data. Farm data is supplemented with REA data.

BEA REGIONAL ECONOMIC ACCOUNTS (REA)

The final set of employment and income information is the Bureau of Economic Analysis’s (BEA) Regional Economic Accounts (REA) data. This dataset is the most inclusive available and provides information on sectors such as agriculture, construction, and railroads not directly available through CBP or CEW.

The REA data series also provides information on self-employment and proprietor income. The major drawback to these data is that they are only available at the 3-digit NAICS level for state and county income, and the 3-digit and 2-digit level for state and county employment. These data provide a means to estimate proprietor employment and income, allowing for completion of the IMPLAN labor income data. The information used in developing IMPLAN data in this section is the following:

• 3-digit State level wage and salary income – SA7 tables
• 3-digit State level wage and salary employment – SA27 tables
• 3-digit State level employee compensation (wage and salary plus other labor income and benefits) – SA06 tables
• 3-digit State level total income (wage and salary and self-employment) – SA05 tables
• 3-digit State level total employment (wage and salary and self-employment) – SA25 tables
• 3-digit County level total income (wage and salary and self-employment) – CA05 tables
• 3-digit County level employee compensation (wage and salary plus other labor income and benefits) – CA06 tables
• 2-digit County level total employment (wage and salary and self-employment) – CA25 tables
• 6-digit disclosed CEW state and county employment and income data aggregated to the 3-digit NAICS, BEA sectoring scheme (used to project the REA data to the current data year).

BEA employment and income data are also subject to non-disclosure rules; therefore, estimates are made for non-disclosed values.

SEPARATING COUNTIES FROM INDEPENDENT CITIES

Unlike the CEW and CBP data, which give information on all counties and independent cities in the U.S., the BEA has combined independent cities with their neighboring counties in their REA data series. In Virginia, there are currently 24 such combinations. In 1994 and earlier datasets, WI also had one of these combined regions. These regions require special processes to split into separate Federal Information Processing Standards (FIPS) counties.

Names of REA Counties and their Combined Cities and City ID

VIRGINIA
ALBEMARLE; CHARLOTTESVILLE 901
ALLEGHANY;CLIFTON FORGE; COVINGTON 903
AUGUSTA; STAUNTON; WAYNESBORO 907
CAMPBELL; LYNCHBURG 911
CARROLL; GALAX 913
DINWIDDIE; COLONIAL HEIGHTS; PETERSBURG 918
FAIRFAX; FAIRFAX CITY; FALLS CHURCH 919
FREDERICK; WINCHESTER 921
GREENSVILLE; EMPORIA 923
HENRY; MARTINSVILLE 929
JAMES CITY; WILLIAMSBURG 931
MONTGOMERY; RADFORD 933
PITTSYLVANIA; DANVILLE 939
PRINCE GEORGE; HOPEWELL 941
PRINCE WILLIAM; MANASSAS; MANASSAS PARK 942
ROANOKE; SALEM 944
ROCKBRIDGE; BUENA VISTA; LEXINGTON 945
ROCKINGHAM; HARRISONBURG 947
SOUTHAMPTON; FRANKLIN 949
SPOTSYLVANIA; FREDERICKSBURG 951
WASHINGTON; BRISTOL 953
WISE; NORTON 955
YORK; POQUOSON 958

The 3-digit CEW employment and income data are used to proportion the REA data into its component counties with alternative proxies being used where CEW data are unavailable or incomplete.

 

---

¹ Other data sources are used to augment the above-listed sources for some specific sectors that are not fully covered by the above-listed sources or where more current data or data with more geographic or sectoral specificity are available.

Validating the database is an important final step in the data development process.

VALIDATION PROCESS

Once the national model is complete and balanced, it is checked thoroughly for errors. Models are built and multipliers generated. All values are distributed to the states and counties based on the procedures outlined here. Once the data have been distributed to the states and counties, an extensive validation process takes place. State and county models are built and evaluated. The data are also passed through a program that calculates ratios on every value in the database. Any outliers are examined and either documented or fixed if a program or data bug is the cause. Once this process is complete, the databases are released to the public. Users should always examine their study area and make changes if required. 

FORCE ACCOUNT ADJUSTMENT

To conform to I-O accounting definitions, IMPLAN industry data are subjected to an adjustment. This adjustment is called the force account adjustment and is an attempt to keep production activities consistent across sectors. For example, some non-construction industries, such as mining, have a large construction component. The force account construction adjustment moves the construction activity from the mining sector to the appropriate construction sector. The net result is that mining values may be slightly smaller than published estimates and the construction values may be slightly higher. Construction sectors can have production functions very different than the Industry using the constructed elements. Another highly affected sector is Hotels and Lodging. Restaurant, retail, and gaming activities that take place in hotels are moved to their respective sectors, thereby reducing the Hotels and Lodging sector to ~70% of its original size.

By keeping the activities separate, expenditures made during production will be more accurately allocated.