RESEARCH The Online Platform Economy through the Pandemic

下图1显示了某一月内获得在线平台经济收入的家庭所占比例，以及该月之前一年内任何时间获得在线平台经济收入的家庭所占比例. Prior to the COVID pandemic, we observed continued growth in participation in the Online Platform Economy, peaking at 2.5 percent monthly, with almost 7.5 percent of households earning platform income within the past year in January of 2020. At the onset of the pandemic, monthly participation fell from its peak to 1.9 percent, a 25 percent decrease in month-to-month participation, and then gradually recovered, nearing its peak of 2.5 percent by June of 2021.

As evident in Figure 1, there is a large difference between the share of families earning platform income in a given month versus over the prior year. 这表明每月活跃用户的数量大大低估了在线平台经济作为一年中收入来源的重要性. Our January 2020 platform participation estimates of roughly 7.5 percent of families in the prior year and 2.5 percent in the prior month fall below estimates of the prevalence of contract or alternative work arrangements from 2018 administrative tax records and the 2017 contingent worker survey but above their estimates of the Online Platform Economy (Garin and Koustas 2021; Department of Labor 2018).

Figure 1: Nearly 8 percent of families earned income from the Online Platform Economy in the year before the pandemic

Figure 2 shows monthly supply-side participation rates by platform sector. At the onset of the pandemic, supply of drivers and lessors fell precipitously, likely due to social distancing measures, travel restrictions, and hesitancy among suppliers about the virus.

Driver participation fell by more than a third from roughly 1.5 percent to below 1 percent in April and May of 2020. 司机参与率的部分下降可能是由于在大流行期间愿意开车的司机数量减少(供给侧)和需要驾驶服务的个人或企业比例下降(需求侧)。. However, drivers transport not just people but also goods, such as prepared food. Further into the pandemic, as restaurants and other businesses adapted to online ordering and delivery, many rideshare drivers switched to delivery (Accenture 2021). 这种驾驶的供应可能也更高，不仅因为需求增加，而且由于与客户接触较少，病毒的风险较低.

Figure 2: Most Online Platform Economy participants work in the transportation sector

Participation in the leasing sector dropped 50 percent, from 0.2 percent to 0.1 percent of the sample, at the onset of the pandemic. Participation gradually increased back to its pre-pandemic peak by June 2021. 由于居家令和旅游的整体减少，短期租赁物业或车辆的需求可能有所下降.

At the same time, growth in selling platforms continued and perhaps even accelerated slightly, consistent with dramatic shift towards online purchasing during the pandemic (Wheat, Mac, and Duguid 2021). With sustained growth in selling platforms, the overall supply-side participation in the Online Platform Economy only slightly decreased.

Next we explore whether changes in participation are caused by a slowdown in new entrants or an acceleration in exits among existing participants. This is an important distinction for several reasons. First, in previous work and as evidenced in Figure 1, we observe a high degree of churn among platform participants, with the majority of workers earning platform income in just a few months out of the year, often when they experience periods of unemployment (Farrell, Greig, and Hamoudi 2018). We want to know whether participants are turning to platform work in a time of need, or exiting from platform work potentially due to a fall in consumer demand, heightened public health risk, or the availability of expanded unemployment benefits.

We find evidence of both. 图3将月参与率分解为前一年已获得平台收入的现有参与者与当月参与在线平台经济且前一年未获得平台收入的新参与者. Focusing on drivers in the top left panel, 我们观察到，司机参与的急剧下降完全是由现有司机退出的加速造成的(实线)。. At the same time, new drivers continued to enter platforms for the first time in a year at the roughly same pace throughout the pandemic (dotted line). We observe a similar pattern for non-transport workers (bottom left panel), who include workers engaged in both in-person and remote services.

In contrast, participation rates of both new and existing lessors fell dramatically at the start of the pandemic. This is likely again due to social distancing regulations and decreased demand for travel and hospitality services. However, the drop in new lessors, in contrast to the steady flow of new drivers, may also be tied to the kind of people who turn to leasing versus transportation platforms. As we outline in Finding 3, lessors tend to have significantly higher incomes than drivers.

Like leasing platforms, the trends in capital selling mirror wider economic shifts in consumer demand following the pandemic. Both new and experienced sellers increased their participation following the pandemic as consumers shifted to online retail. In addition, unlike leasing, the entry costs for online selling are relatively low, and people may have turned to the Online Platform Economy out of necessity to make ends meet or while working from home.

Figure 3: Experienced drivers stopped using online platforms following the pandemic but new drivers continued to take up platform work

Having documented that many existing participants left the Online Platform Economy, we next examine the extent to which platform participants received unemployment insurance. Under the Pandemic Unemployment Assistance program, platform workers could have become eligible for unemployment benefits due to a loss of platform income. Additionally, platform workers or their family members could have received unemployment benefits as a result of losing a non-platform job.

Recognizing that job losses during the pandemic were concentrated in low-income sectors, we start by documenting pre-pandemic total income in Figure 4. We find that baseline income of platform participants varies considerably by sector. In 2019, prior to the pandemic, median total take-home income was roughly $49,000 for drivers, $59,000 for non-transport workers, $64,000 for sellers, and $102,000 for lessors. Thus, intuitively, drivers tend to be lower-income workers, while lessors tend to have the highest incomes. Meanwhile, people with no platform income in 2019 had a median total income of $55,000, slightly higher than platform transport workers.

Figure 4: Transportation workers have incomes slightly lower than the general population while capital lessors have significantly higher incomes

2019年收入的这些鲜明模式表明，大流行的经济冲击可能对不同行业的平台参与者产生了非常不同的影响，因为不仅他们的平台工作性质不同，他们的非平台工作性质也不同. To explore this, 我们查看了2019年所有有平台收入的家庭，并计算了有多少人在疫情期间获得了失业保险. These UI rates are plotted in Figure 5, alongside the UI rates of families who received no platform income in 2019. In this sample, we are unable to distinguish between people who received traditional UI and those who received PUA.

Figure 5: Many more Online Platform Economy workers received unemployment insurance than the general population

Overall, we see significantly higher UI rates among platform participants relative to the non-platform group. At its peak, the UI receipt rate of drivers was just under 19 percent, over twice the rate of the non-platform group. UI receipt rate among platform participants in other sectors are also elevated, peaking between 13 and 15 percent.

It is important to note that the demographics of platform participants may be different from the non-platform population, and this may in turn be driving some of the differences in UI rates. For example, if platform drivers are younger than the general work force and all young workers were hit relatively hard by the pandemic recession, that may explain why the platform drivers have such high UI rates. To explore how much demographics are driving the difference in UI rates, we re-weight the non-platform sample so that it has the same joint age and income distribution as the platform sample. The dashed gray line in Figure 5 shows the UI rate of this re-weighted sample.

重新调整非平台样本的权重，使其具有与平台参与者相同的年龄和收入特征，从而缩小了两组用户界面收入差距的一小部分.

A number of other factors could be driving these differences, such as industry, geography, or the nature of the workers themselves. That is, workers with non-platform jobs that were more irregular or directly impacted by the pandemic might be more likely to moonlight as platform drivers, workers, and sellers. As such, the pandemic-induced recession, which disproportionately caused job losses among lower-income workers, likely affected platform participants more than the general population. It is also possible that platform workers, facing a loss of consumer demand and a public health risk, applied for and received Pandemic Unemployment Assistance (PUA) which expanded UI eligibility to self-employed and contingent workers.

Next we return to active platform participants and document trends in monthly platform revenues since 2018. Figure 6 shows median platform income among families who earned any income in the given platform sector in the current month. It is important to note these earnings represent revenues to participating families, and not profits net of expenses. Additionally, given the dynamic pricing models commonly deployed by platforms, a change in revenues could reflect a change in prices, units sold, or both.

Median monthly platform income for lessors was roughly $1,300 or at least three times higher than monthly revenues in the other three platform sectors. Lessor revenues were steady leading up to the pandemic, fell briefly at the onset of the pandemic, and then rose sharply. The subsequent rise in earnings is likely due to a change in who is actively participating in the online platforms. Recall that Figure 3 shows a steep drop in both new and experienced lessors. Families more involved in platform leasing, like those with more units to lease, derive more total revenue from leasing and would be more likely to remain in the sample or be the quickest to rebound. In addition, it could reflect an increase in prices or longer-term rentals as families adjusted to pandemic lifestyle.

For drivers, the trend in monthly platform revenues is also generally flat before the pandemic, but active drivers saw a 40 percent drop in platform revenues in April 2020. Given the drop in the supply of drivers evident in Figure 2, the fall in revenues is evidence of a fall in consumer demand for transportation services.  Revenues among active drivers quickly rebounded in June 2020 to near pre-pandemic levels, even as driver participation rates were still depressed. Insofar as Figure 6 documents total platform-derived income, the product of hourly earnings and total hours worked, it cannot inform us about whether drivers are working longer hours earning less per trip in order to sustain revenues.

Figure 6: When engaging in online platform work, drivers earn about $400 per month

We can, however, 通过计算平台收入占家庭收入的比例来衡量平台收入对家庭整体财务健康的重要性. We do this in Figure 7, 它将每个家庭的平台行业收入占家庭总收入的比例，并绘制出在给定月份参与特定平台行业的家庭之间这些比例的中位数.

Figure 7: The median driver earns roughly 10 percent of their family’s total income from transportation platforms.

The median lessor derived between 15 and 20 percent of their total family income through leasing platforms prior to the pandemic. This share rose slightly since the onset of the pandemic to closer to 20 percent. This increase in platform income as a share of total family income is again probably due to a change in who is participating, with fewer lessors entering the sector following the pandemic. But even before the post-pandemic increase, lessors had the largest platform income share across all platform sectors; that is, lessors earn a larger portion of their total family income via leasing platform than participants in any other platform sector.

The sectors in which platform income contributes the next highest share to family income are transportation and non-transport work. Before the pandemic, platform income ranged between 10 and 15 percent of total family income for drivers and 9 percent for non-transport workers. As shown in Figure 4, drivers and non-transport workers tend to have the lowest incomes of all platform workers. 这表明，从家庭福利和生活质量的角度来看，在线平台的可用性对司机和非运输工人比对其他参与者重要得多.

After the onset of the pandemic, platform income share for drivers dipped slightly, hovering more consistently around 10 percent. Non-transport workers follow a similar pattern over time, platform income share dropping from 9 percent to 7 percent by June 2021. 这些趋势可能反映了司机和非运输工人家庭收入状况的变化，因为有经验的参与者离开了，新的参与者来到了, as we saw in Figure 3. Additionally, as we have documented elsewhere, lower-income families saw an increase in their total family income during the pandemic due to government income supports, including both unemployment insurance and stimulus (Greig, Deadman, and Noel 2021). 因此，在司机和非运输工人中，平台收入占家庭总收入的比例下降，可能只是反映了家庭总收入的增加. This is particularly likely, given the high rates of UI receipt documented in Figure 7 and the presence UI supplements, which caused many low-income workers to receive more in UI benefits than their prior income (Ganong, Noel, and Vavra 2020).

By contrast, platform sellers make relatively little of their income via online platforms, with the median share ranging between 1 and 2 percent. This, along with the relatively higher total incomes of platform sellers, suggests that online selling is a secondary source of income that is less critical to the financial health of these families.

The Online Platform Economy is crucial source of income for many families even after the shock of the pandemic. At its peak, almost 8 percent of families earned platform income in any 12 month window. This is especially true for drivers, who represent the lion’s share of supply-side platform participants and have the smallest total family incomes. Additionally, lessors derive the highest revenues and the largest share of their total family income from platforms.

During the pandemic the Online Platform Economy served as not just a fallback source of income in a time of potential need but also, in its own right, a source of potential risk. New workers continued to start platform work during the pandemic, allowing them to remain attached to the workforce.  Many others exited platform work potentially due to a fall in consumer demand, heightened public health risk, or the availability of expanded unemployment benefits.

Indeed, 网络平台经济工作者在疫情期间失业保险领取率极高，似乎特别容易受到经济冲击的影响. Almost one in five drivers in 2019 was receiving unemployment insurance at the beginning of the pandemic. Of all platform workers, drivers appear to be the group of biggest concern for policymakers from a welfare perspective. They are the most numerous group, have the lowest family incomes, were the most likely to have received unemployment insurance during 2020.

网络平台经济的持续增长凸显了加强临时工社会保障网络和减少与平台收入相关的行政负担的重要性. The Pandemic Unemployment Assistance program was an important expansion of unemployment insurance during COVID-19 pandemic, representing roughly 40 percent of UI claims before it was ended in September 2021. However, recipients of the Pandemic Unemployment Assistance program faced significant delays in receiving unemployment benefits (Greig et al. 2021), and families who had to wait longer for jobless benefits cut their consumption more dramatically (Farrell et al. 2020). Reducing administrative hassles associated with verifying platform income for the purposes of filing taxes, qualifying for social safety net programs, or gaining access to credit, could materially improve and simplify the financial lives of platform workers.