Measuring What Matters: New Approaches for Assessing Economic & Social Impacts in the Wake of COVID-19

Written by Shuprotim Bhaumik and Kristina Pecorelli
 
With every passing day, it becomes clearer that the impacts of COVID-19 are exacerbating underlying inequities in American society. Understanding exactly which populations are most at risk in which localities will be important for fiscally constrained states and municipalities that need to target scarce resources equitably and efficiently. Rethinking how we approach economic impact assessment is critical to gaining that understanding.
 
Economic impact assessments (EIAs) estimate the impacts of projects, programs, and policies to support requests for investment. Results are typically expressed in terms of spending, jobs, and wages, and rely on multipliers that are based on observed economic relationships made over a period of time to show how changes in one segment of the economy reverberate throughout the rest of the economy. For instance, assume a jurisdiction proposes to use a tax abatement to secure the relocation of a back-office of a financial services firm: the analyst ascertains the number and types of jobs the firm proposes to bring to the location (“direct jobs,” e.g. accountants), then employs a model to estimate the number and types of jobs that the firm will create through its spending (“indirect jobs,” e.g. legal services) and that the employees of the firm will create through their own spending (“induced jobs,” e.g. baristas).
 
In the COVID-19 era, EIAs can help civic leaders accurately assess the damage, make the case for funding, and ensure that recovery funds are spent efficiently and equitably. The challenges of impact assessments during COVID stem from the peculiar nature of the pandemic, which is rewriting the fundamental patterns of local economies in real time. The multipliers contained in the traditional models are likely no longer accurate: amidst select business closures, phased re-openings, and massive shifts in consumer demand, impacts will vary widely by type of business and even the nature of work within businesses. Some impacts may become permanent as the crisis reshapes how certain sectors operate in response to the global pandemic.
 
The American Reinvestment and Recovery Act, the federal response to the Great Recession of a decade or so ago, resolved this concern, which we believe to be far more fundamental now than it was then, by refusing to consider estimation of anything more than direct jobs in the EIAs it mandated to secure stimulus funds. We believe a more nuanced approach is possible, indeed required.
 
HR&A is currently working with Nassau and Suffolk Counties in New York to develop an approach to measuring the real-time impacts of COVID-19 on Long Island. The region has rapidly transitioned from the nation’s first postwar suburban district to more than a dozen distinctly different cities and townships with a population of 2.8 million. Long Island is also a more diverse place than it was decades ago, driven in part by a growing immigrant and non-white population. The jobs-housing balance has shifted to more job-intensity, and emerging health care, agritourism, biotech, and clean energy industries have replaced what was once a manufacturing-dominant region. Less positively, racial and income disparities and segregation in housing and schools have persisted. A lack of affordable housing is reinforcing a brain drain; young Long Islanders are leaving the region.
 
Working with leaders from both counties, we’ve developed an analytic framework that considers the initial economic shock from COVID-19, reflecting the impact of business shutdowns and subsequent reductions in consumer demand; differential rates of recovery, depending on when businesses are allowed to reopen and how well-positioned they are to resume full operations; and “new normal” levels of economic activity, which may be at, above, or below the pre-COVID levels depending on economic sector.
 
To apply this framework, we categorized all jobs within the regional economy according to unemployment risk. Expanding upon methods developed by the St. Louis Fed and UChicago Becker Friedman Institute, risk levels were assessed at the industry and occupation level, reflecting whether jobs were classified by public authorities as “essential” to public health and safety and thus “permitted” to remain open, can be performed off-site, and are salaried or wage-paying. For example, cashiers are typically considered an occupation at high risk of unemployment during an economic shock owing to non-essential status, inability of the job to be performed remotely, and hourly wage pay basis. However, this assumption does not hold true for grocery store cashiers in this (or most other) shocks. Hence, more careful examination of the effects of the pandemic is warranted to guide efficient, equitable policy response.
 
The table below shows the distribution of jobs on Long Island by high-, medium-, and low-occupational risk of loss and essential versus non-essential industry, a distribution we developed via consultation of cumulative weekly unemployment insurance claims, documented changes in monthly employment by sector available as of this writing, local and national survey data, and other sources portraying consumer behavior and demand responses under future public health scenarios.
 
By examining the characteristics of workers within occupations or industries at greatest risk of unemployment, we can also understand who is most likely to suffer economic losses (possibly in conjunction with poor health outcomes). As shown below, low-income, low-skilled workers, and to a lesser but still significant extent, Hispanic or Latino workers, are disproportionately affected by the economic shock of COVID-19 on Long Island. This is largely a function of the types of businesses that employ these populations—namely Retail, Accommodations and Food Services, and Arts, Entertainment, and Recreation—being among the most severely impacted by the initial economic shock of COVID-19. (If risk were equally—which would not equate to equitably—shared, the percentages across any given row should be equal or similar to the percentage for all local workers shown in the final “Total Long Island” column. Differences suggest areas for policy focus.)
 
While the unfairness of the virus’s effects—both physical and economic—are well-known, that a disproportionate share of jobs at high risk for unemployment are occupied by workers without a post-secondary education, while jobs with the lowest risk of unemployment are disproportionately occupied by workers with a bachelor’s degree or above, suggests that those at highest risk are also those most limited in their ability to find alternative job opportunities given their existing skill set. This suggests that retraining or reskilling initiatives need to be central to future recovery efforts on Long Island.
 

 
This exercise will undoubtedly prompt uncomfortable but crucial questions for recovery. For example, should we prioritize aid to sectors that are disproportionately impacted, but which tend to be lower-paying and offer fewer opportunities for economic mobility? Or should we double down on sectors with the highest economic multiplier effects, and focus instead on addressing the barriers to entry that have historically prohibited access by currently low-skilled workers? Should aid be based on the size of business or those that employ a disproportionate share of high-risk workers? Do concentrations of impact among low-wage workers suggest there should be a focus on housing, transit, and other cost factors that influence economic survival during a long-term recovery? What sorts of retraining or reskilling efforts will be required for jobs permanently displaced by the crisis, and which organizations are best equipped to run them based on the populations affected?
 
All of these questions are critical to an economic recovery plan. Answering them begins with a robust understanding of impact and its distribution. The impacts, their distributions, and therefore the answers that inform recovery will all vary by location.