Using the COVID 2020 Quarterly Data


The COVID 2020 Q2 Data and COVID 2020 Q3 Data are available in the IMPLAN application as Data Years COVID 2020-Q2 and COVID 2020-Q3. This article will cover selecting the new dataset and considerations for using it in an economic study. 


First, sign into the IMPLAN account and go to the project tab. From the project tile, click New Project in the top right of the screen.

A popup window will appear to create a new project. The Industry Set should be set to the 546 Industries. This Industry Set is specifically for the U.S. and U.S. Territories data. After the correct Industry Set is selected, name the project, then click Create Project.

After creating the Project, the Regions screen will appear in the Map View of the U.S. Use the Data Year dropdown list to select COVID 2020-Q2 or COVID 2020-Q3. Then, select a Region by clicking on the map, using the Search Bar, or from the List View. 


So we know that 2020 wasn’t a typical year. The data for 2020 is no exception. Our Data Team put together an annualized dataset based on the second and third quarters of 2020: when COVID hit the world. So the entirety of the dataset looks at how the whole year would look based on these two quarters of the year. 

The data is seasonally adjusted. Industries that employ more folks in the summer months are better represented in their employment for 2020. We also know that some of these Industries were the hardest hit with layoffs and furloughs: restaurants, airlines, and other tourism related fields, for example.

Due to the nature of quarterly data releases, outlined in their Release Notes, the estimates for the COVID 2020-Q2 and Q3 data are not as reliable as those of the annual IMPLAN datasets. The values for Employee Compensation, Employment, Personal Income and Household Personal Consumption are based on detailed source data. The estimates for Output, Intermediate Inputs, and TOPI (see note below on taxes) are extrapolated from employment and productivity data, whereas OPI is estimated solely as a residual; accordingly, they are less reliable than the estimates of employment, compensation, and personal consumption. The values for OPI are the weakest in the data.

The good news is that the composition of household spending did change to reflect the adjusted spending behavior. This will affect the Induced Effects. Both the amount of spending and the items purchased are different in this dataset. Industry Spending Patterns and Institutional Spending Patterns also changed based on Output. For example, if there is less Output in restaurants, there will be less purchasing from restaurants by Industries and Institutions. 


There were two main forms of government stimulus in 2020. The first was the stimulus checks sent to individuals and families coupled with expanded unemployment benefits. Some of this money was spent and some was saved. The levels and types of spending and saving in the Household Social Accounting Matrix column will include what Households did with the stimulus and unemployment benefits money. Changes in the relative amounts that Households spend among the various commodities is also captured in the COVID 2020-Q2 and Q3 data (e.g., more on grocery stores, less on restaurants). 

While gains in Household Income due to stimulus checks and unemployment benefits will be captured in the underlying data in the new dataset and are accounted for in the relationship of spending per dollar of new income, IMPLAN will not assume households receive more payments from these sources in the analysis because government payments are not by default internalized in the IMPLAN multipliers. If additional gains in Household Income are expected, these gains can be analyzed using a Household Income Event or by modeling the specific spending as individual Events.

Note, even with the fiscal efforts to support household spending, there were exceptionally high personal savings rates in 2020 Q2 and Q3 as people spent less money due to quarantine and concern over less economic security. The increased savings rates will be applied to income analyzed in IMPLAN. So, all else equal, this will lead to lower Induced Effects.


The second government initiative was the Paycheck Protection Program (PPP). As this money was treated as subsidies for Industries, it shows up as reductions in TOPI. Remember, TOPI is reported as a net of subsidies. The decreases in TOPI were offset by an equal increase in OPI, which explains why some Industries see losses across all pieces of Value Added except for OPI. This OPI is the infusion of the government money that is to be used to cover operations and maintain employment levels. All else equal, when a firm gets a subsidy, it has a negative effect on net TOPI and a positive effect on OPI. For more information on this, check out the BEA. Again, read the cautions below on reporting tax results.


For some Industries, there are very high losses evident from anecdotal evidence as well as the COVID 2020-Q2 and Q3 data. We expect to see decreases in restaurants, travel, and amusement activities. Decreases in Output, Value Added and its components, and spending on Intermediate Inputs frame the story for each Industry. Many Industries were highly subsidized already, and then added additional monies through the PPP. Many Industries had to shut down (at least temporarily), decrease workforces, or pivot their entire business model. Many spent less money overall on Intermediate Inputs and a few saw international export markets dry up. All of these moving pieces need to be brought to mind when examining the losses seen in the data.

Regions and states that put stay-at-home orders in place earliest will show the largest losses. So if the county or state was quick to shut down, the Results will be more dramatic than in other areas.


The first consideration when using the COVID 2020-Q2 or Q3 Data is to remember that this is annualized based on what happened in the second or third quarter of 2020. The relatively normal first quarter is not represented in this dataset; it’s truly a look at the start (and hopefully worst) of the pandemic.

When comparing the 2019 data to the COVID 2020-Q2 or Q3 Data, there will be vast changes. Remember we are looking at a full year of complete data in 2019 compared with only the annualized activity from the second or third quarters of 2020. One cannot attribute all of the differences between the COVID 2020-Q2 and Q3 datasets and the 2019 dataset to the pandemic and associated government responses. 


Using the Annual IMPLAN datasets might still be the best option when trying to model anything pre or post-pandemic. We recommend using the Data Year that looks most like the year of the analysis. Include a footnote to the report or presentation that notes not only which dataset was utilized, but also includes a disclaimer about the potential implications of using it.

When using the COVID 2020-Q2 or Q3 Data Years, always add as much information (Output, Employee Compensation, Proprietor Income, and Employment) as possible to the Event in the Advanced Fields. Referencing an Industry’s Leontief Production Function (LPF) from 2019 can be helpful if an Industry’s production decreased during the quarter and their LPF changed in a way that does not reflect their operational LPF. For example, if Employment has decreased in an Industry, particularly among their lower wage employees, the Industry’s Average Employee Compensation (EC) likely increased and will potentially lead to an overestimation of Direct EC in an impact analysis. In reality, that Industry will probably need to hire for the positions lost to ramp up operations again.

Since the Direct Effects and the first round of Indirect Effects are the most significant in an impact analysis, modifying them in an analysis will help to better reflect the current state of the economy. The Industry Impact Analysis (Detailed) is the most appropriate Event Type to use in those scenarios, as it allows users to specify any known adjustments to the business or Industry operations. 

Also consider running the analysis using the 2019 Data Year and the COVID 2020-Q2 or Q3 Data Years to see a range of the potential economic impacts. If the economy shifts back to a pre-pandemic state, this may help estimate future impacts. Although, our economy may never return to the “old normal” and it is unclear how the recovery will look. Brookings outlined some potential options that are worth considering when modeling the impact.

Finally, if it becomes necessary to model a specific scenario, edits can be made to Industry levels and per-worker values in the Study Area data by customizing a Region


Running an Industry Contribution Analysis using the COVID 2020-Q2 or Q3 data is a great way to see what the effects of a business or Industry are in the COVID economy.  Run the analysis at a 100% contribution to estimate the effects supported by an entire Industry. If the analysis is for a single firm or business, use the current level of Output they produce. 

Again, comparing a COVID 2020-Q2 or Q3 Industry Contribution to a 2019 Industry Contribution is recommended for estimating the before and after effects of COVID-19 (keeping in mind it's unknown what changes occurred in 2019). However, remember to note that the changes in taxes between these two Data Years will be dramatic because of the large subsidies in the COVID 2020-Q2 and Q3 data.



Users should exercise caution when using IMPLAN Occupation Data tied to the Covid 2020 Q2 and Q3 Data. This is because there is no update to the most recent Occupation Data. The 2020 Occupation Data will not reflect some of the dramatic changes that have occurred due to the pandemic during these specific quarters. 

One notable example is the Industry 502 - Amusement parks and arcades. The source data indicate that employment declined by 44% and earnings declined by 2%. This is likely due to a reduction in seasonal and hourly employees while full-time administrative and executive employment was maintained. The smaller decrease in earnings is because the year round employees earn higher incomes. However, when applying Occupation Data to the COVID 2020-Q2 Data, this change in the composition of employment will not be reflected. So, it would appear as if everyone in the Industry saw pay rates nearly double, which is almost certainly not the case given what we know about how the pandemic has affected this Industry. 


Because TOPI is net of subsidies by definition and the second quarter of 2020 saw excessive subsidies due to the CARES act, tax results are overestimated at the local and state levels. This can be seen when examining the tax detail and seeing that all the growth is in TOPI, not in personal taxes. While local and state TOPI actually stayed relatively stable, the share of these of total TOPI looks to increase in the data. Therefore, we recommend using known tax estimates for the local and state levels or using tax estimates based on the 2019 Data Year in lieu of the COVID 2020-Q2 or Q3 Data Years.


Evolving Economy - COVID 2020 Q2 Release Notes

Evolving Economy - COVID 2020 Q3 Release Notes

Pandemic: Analyzing the Economic Impacts of the Coronavirus

Pandemic: Additional Considerations when Modeling the Coronavirus


Written August 30, 2023

Was this article helpful?
1 out of 1 found this helpful