COVID-19 Graphical Analysis

View September 8 data for:

Ask our experts.

AEG continues to track the progression of COVID-19 across the US & around the world. Our analyses help businesses, trade associations, and public institutions to:

  1. Assess the pandemic trajectory using advanced modeling that naturally captures the empirically observed asymmetry in cases over time;
  2. Notice later-occurring (second wave) cases that arise from new infections and variations in institutional records management (such as expanded testing, reclassification, double-counting of test results, expansion to include probable cases, and other factors); and
  3. Understand delays caused by distance, especially in large areas with multiple population centers.

September COVID-19 Commentary:
Early Voters Consider Their Options with Eyes on COVID-19 Cases

Voters in eight states have the option to vote absentee or in-person in September

September 10, 2020–Chicago, IL: As the nation moves closer to election day on November 3, voters in eight states will have the option to vote early in September. Some of these voters will be closely monitoring the spread of COVID-19 as they consider when and how to vote. New analysis from Anderson Economic Group shows that the spread of COVID-19 is slowing in most of these states.

Early voting for the general election will begin September 18 in Minnesota, South Dakota, and Wyoming. By the end of the month, New Jersey, Virginia, Michigan, Vermont, and Illinois will also be open for early voting. All of these states will offer early voting by mail and in-person absentee voting, with New Jersey mailing ballots to all registered voters.

Brian Peterson, the firm’s director of public policy and economic analysis, notes that infection counts in seven of the eight early voting states–all but South Dakota–appear to be decreasing. “Our latest COVID-19 projections show that most of these states are in the downward sloping phase of a second wave of infections,” he said.

Peterson pointed to Illinois, Michigan, Vermont, and Virginia as examples of states where initial infections peaked near the beginning of May, followed by a decline in cases and a subsequent second wave of infections. “Decreasing case counts is a welcome sign, especially as these states gear up for the election and voters venture into public spaces.”

In other early voting states, COVID-19 outbreaks have followed more unique paths. New Jersey, for example, experienced a large peak in infections in April, followed by a subdued second wave. Meanwhile, infections in Wyoming only peaked for the first time in early August. Peterson said that, despite the unique path of the epidemic in each state, the current data point to continued declines in new cases through the end of September.

South Dakota stands out among all states with a significant spike in cases in recent weeks, making it difficult to reliably predict the path of the virus in the near term. In the last month, South Dakota hosted the Sturgis Motorcycle rally, which attracted over 350,000 visitors, and the South Dakota State Fair, which attracted over 200,000 visitors.

Peterson also notes that, while new cases appear to be declining in many areas, voters, election officials, candidates, and campaign staff should exercise caution and consult real data when in doubt. “We have made it through six months of the pandemic,” he explains. “We now have much more information about how the virus spreads and who is at risk. Because of this, making an informed decision about our desired level of interaction with others is now easier.”

About Anderson Economic Group

Anderson Economic Group, LLC, is a US-based research and consulting firm that specializes in economics, public policy, commercial damages, market analysis, and tax and regulatory policy. The firm, founded in 1996, is one of the most recognized boutique consulting firms in the US.

With over 20 years of experience in economic impact modeling, Anderson Economic Group is one of the original pioneers in assessing the economic impacts of higher education. To date, we have completed over 40 such analyses for small and large institutions across the country, including Big Ten institutions such as University of Michigan and Michigan State University.

Also available:The Costs & Consequences of the Coronavirus Pandemic: Three Lessons” presented to the Western Economics Association, June 26 Conference. Patrick Anderson’s analysis considers the dominance of empirical over simulation models, the huge cost of stay-at-home orders, and a discussion of the future of empirical models that includes a focus on AEG’s innovative two-wave models.

See the notes at bottom and the associated memorandum regarding a significant update to our forecasting methodology.

Click Images to Enlarge

States with September Voting

Modeling the COVID-19 Pandemic

Since February, Anderson Economic Group LLC, together with the data analytics firm Supported Intelligence, has been gathering data from multiple countries and U.S. states to model the path of the COVID-19 pandemic. We identified the situation as a crisis as early as our February 10 release, and began to release selected regional analyses in late March. Since then, we have implemented the following practices in our improved model:

  1. We use a two-wave generalized logistic model in conjunction with non-linear curve fitting.
  • We use an improved version of the same mathematical function used for disease growth (including but not limited to the SIR model).
  • Our improvements to the function include but are not limited to the ability to measure asymmetric growth, as well as secondary waves.
  • We compare results across multiple states and countries.
  1. We show the actual data.
  • We focus on a tangible concept: the number of people cases with the disease.
  • We show uncertainty, units, dates, and past prediction errors.
  • We use data compiled by the New York Times for most of the United States, and World-o-Meter for much of the rest of the world.
  1. We describe our methodology, data sources, and limitations.

See the memorandum describing the methods, data, limitations, and Terms of Use for this information.