In the four weeks between March 15 and April 11, more than 20 million workers applied for unemployment insurance (UI) benefits. This is more than five times the worst four-week stretch of the Great Recession, which, at the time, was the worst recession the United States had seen since the Great Depression. What are we really up against here?

The situation is unfolding quickly, but one helpful way to understand it is to put the UI numbers in the context of the unemployment rate. When we do that, we find the jump in jobless claims over the last four weeks would have increased the unemployment rate to 15.7%—if everything else (like the rate of hiring and the rate of people voluntarily quitting their jobs) stayed the same—and all the workers who filed for unemployment benefits were counted as unemployed. But everything else definitely did not stay the same, and it is unlikely that all the workers applying for UI were counted as unemployed.

• Everything else did not stay the same. The overall change in employment over any period is equal to the number of hires over the period minus the number of job “separations” over the period (in the pre-virus period, there were around 5.9 million hires and 5.7 million separations every month). For one, hires drop dramatically in recessions. And secondly, separations are not just job losses where people filed for UI. They can also be job losses where people didn’t file for UI (or were frozen out of the system), or voluntary quits, retirements, or worker deaths. All these parts are moving right now. In other words, there are important elements aside from what shows up in the UI data that go into determining the overall change in employment. As a result, it remains to be seen if the overall employment decline related to the coronavirus will be greater or less than the increase in coronavirus-related jobless claims.
• Not all workers who filed for unemployment benefits will be counted as unemployed. A worker is counted as unemployed in the monthly unemployment numbers if they are on furlough (i.e., on temporary layoff), or if they don’t have a job but are available to work and are actively seeking work. So, someone who lost their job but is not actively seeking work because the virus makes job search impossible will not be counted as unemployed. And someone who lost their job because they have to care for a child whose school or day care closed would not be counted as unemployed because they are not available to work. Instead, these workers would be counted as dropping out of the labor force. It is difficult to know what share of workers who applied for unemployment insurance benefits will be counted as unemployed, but in the March unemployment data, which showed the leading edge of the impact of the virus, only about half of the drop in employment showed up as an increase in unemployment, and the rest showed up as a drop in the labor force participation rate. If that ratio holds, the unemployment rate will rise only half as much as it “should” in this crisis. (This is why it would be useful to use the employment rate as the metric when determining when pandemic-related relief provisions should trigger off, instead of solely the unemployment rate.)

Who is getting hit the hardest by job loss? Many of the jobs in at-risk sectors are low-wage jobs, like those in restaurants and bars, hotels, personal services, and brick-and-mortar retail. That means low-wage workers are seeing disproportionate job loss. We also have initial data on job loss by gender, and it shows that women have been hit harder by job loss than men. We don’t yet have solid job loss numbers by race and ethnicity, but because black and Hispanic workers are more concentrated in front-line service jobs that have been hit hard by social distancing, black and Hispanic people are likely experiencing greater job loss. It is worth noting that because low-wage workers are more likely to face job loss in this pandemic, wage growth numbers during this crisis should be interpreted with caution. With millions of disproportionately low-wage workers losing their jobs, average wages will likely increase sharply. However, this will not be because worker are getting meaningful raises, it will be because a large share of low-wage workers were dropped from the calculation because their jobs disappeared.

Further, job losses are ongoing. The decline in employment could exceed 30 million jobs in the near term. If all the workers represented by an employment decline of 30 million were counted as unemployed, that would increase the unemployment rate to 21.7%. But because it is likely that many workers who are out of work because of the virus will not be counted as unemployed, the unemployment rate will likely not rise that high. That will not, however, reduce the pain the employment losses cause. Further, because our system ties health insurance to work, it is likely that more than nine million workers lost their health insurance in the last month, and millions more will soon follow.

The CARES Act, while large, is no match for the damage the economy is facing, and federal policymakers need to do more. The next relief and recovery package should provide aid to state and local governments, extend unemployment insurance benefits, provide better protections for workers and jobs, and include funding to safeguard our democracy. And, importantly, we cannot turn off federal government relief too soon. It is likely that even under the best-case scenario—a rapid bounce-back in the second half of the year—the unemployment rate will be close to 8% a year from now. Congress has the unique power to greatly mitigate the economic suffering associated with the coronavirus shock, and it needs to act.