After the U.S. lost 22.2 million jobs in March and April this year, 9.2 million jobs were created in May, June, and July. In August, the pace of monthly job growth fell to its slowest pace, with 1.4 million jobs created. In some sense, any job growth at all in August was a relief. In the beginning of the month, the first gash of austerity hit the economic recovery, with the enhanced $600 weekly unemployment insurance (UI) benefits cutting off. Data released today by the Bureau of Economic Analysis (BEA) show that this cut-off of enhanced UI benefits (the Pandemic Unemployment Compensation, or PUC) pulled $667 billion of purchasing power out of the U.S. economy in the month of August alone (expressed as an annualized amount). It would have been more, but some of the enhanced $600 payments spilled over into August data.

One way to scale this impact is to express it as the equivalent of an across-the-board pay cut for all U.S. workers—in these terms it can be thought of as an economywide 7.1% pay cut. In September’s data, when the full $600 is completely gone from personal income data, this will rise to closer to 10%.

Figure A below shows UI benefits as a share of total wage and salary income in the U.S. economy over time, highlighting the extraordinary boost to incomes provided by the enhanced UI benefits in response to the COVID-19 shock at least until August, when the expiration of PUC caused UI benefits to fall sharply.

In any normal month, imposing austerity this sudden and this large would cause a recession that would play out as a sharp contraction of economic output and, very shortly thereafter (maybe even in the same month), employment would also fall. This raises the obvious question of how employment increased by 1.4 million in August, and it will make any job gains at all in September a bit of a puzzle, on first glance.

The answer to the puzzle is clear: These are obviously not normal times. More specifically, there are two strong forces pushing the economy in different directions at the moment. Call them the reopening effect and the income effect.

The reopening effect is straightforward to explain. When large swaths of the economy shut down in March and April, the decline in output—or gross domestic product (GDP)—and employment was enormous. Since April, the pandemic-induced shutdowns of economic activity have begun reversing. Given how utterly enormous and widespread these shutdowns were, it is unsurprising that even partial reopenings have led to very large increases in GDP and employment from the COVID-19 trough of economic activity.

Figure B below shows data from the Federal Reserve Bank of Dallas’s Economic Mobility and Engagement Index (MEI). The MEI combines a number of data points on household mobility and trips away from home in recent months. It is an excellent summary measure of the extent of social distancing happening in the economy relative to pre-coronavirus trends. In the figure below, the index is set to average zero in February, and the lowest level of mobility (reached during March) is set to -100. As the graph shows, from the largest effect of social distancing seen in March, the economy has recovered well over half of its post-coronavirus fall in mobility by July, with a much slower pace of improvement thereafter. What this means is that recovery has indeed so far been driven by a very strong reopening effect. But more than half of this effect has already been “spent,” and its force in recent months is clearly waning. The waning force of the reopening effect in July and August is likely driven in large part by widespread resurgence of COVID-19 cases. This highlights yet again that public health policy is economic policy and that recovery will only proceed free of this constraint if a vaccine or effective treatment is found, or if far better public health containment measures are adopted.

Perhaps the most striking thing about the months of coronavirus-induced shutdowns, however, was that while output and employment cratered, personal income actually rose, as the Coronavirus Aid, Relief and Economic Security (CARES) Act provided enhanced UI benefits and a nearly across-the-board Economic Impact Payment (EIP) to U.S. families. In a sense, by borrowing from the future to support the incomes of workers who had been laid off due to the coronavirus, the CARES Act erected a firewall between sectors directly forced to shut down due to COVID-19 (think restaurants, hotels, and airlines) and those sectors that could continue to work (think grocery stores and online retail). If the 22.2 million workers who had been laid off in March and April had received no additional benefits from CARES, they would have been forced to radically reduce their consumption spending as their incomes collapsed, and this would have led the economic shock to spill over more widely across the economy. With the enhanced UI benefits cut off and no further EIP, we have seen the first large gash of austerity hit the recovery, which will work through the income effect to slow the recovery greatly in coming months.

There is one small caveat to the influence of the income effect—the rise in personal savings during the months of shutdown. While personal incomes did not fall during the months of shutdown, there was a great reduction in what this income could be spent on. Face-to-face services were shuttered, and so essentially no money was spent on them. The combination of stable (or even slightly rising) personal income and a reduced ability to spend meant that personal savings skyrocketed in those months. This increase in savings translated into purchasing power ready to spend once reopening began. This spending out of the accumulation of savings can delay the reckoning of the income effect for a while, but a few months of extra savings will not fully neutralize the negative income effect stemming from millions of unemployed workers collecting only stingy UI benefits (or none at all once durations expire and the enhanced eligibility requirements included in the CARES Act expire in December). Further, the increased spending out of accumulated personal savings will likely happen only slowly in coming months, and the pace of this will be affected by the same influences—public health policy choices—that dictate the pace of the reopening effect.

As we enter October, the reopening effect will likely continue to provide a mild boost to the economy’s growth, but a massive resurgence of the virus could spark reversal on this front. The stock of accumulated savings from the shutdown months will also provide a tailwind to consumption spending in coming months, but the massive negative effect on incomes stemming from the expiration of the enhanced UI benefits will weigh heavily on growth.

It is theoretically possible that the first two influences “win” over the third and a full economic recovery happens without the U.S. economy falling back into recession. But this is an extremely unlikely scenario. The first gash of austerity—the expiration of enhanced UI benefits—is almost enough by itself to slow the economy’s growth to zero in coming months. Layer on top of this the second big gash of coming austerity—state and local government cutbacks in response to the revenue collapse that occurred during the months of shutdown—and it seems nearly impossible to avoid a resurgence of recession before January, unless policymakers radically change course. A good place to start is a reestablishment of enhanced UI benefits and large-scale federal fiscal aid to state and local governments.