Paychecks, paydays, and the online platform economy: big data on income volatility

Abstract

In this report the JPMorgan Chase Institute leverages de-identified administrative data on Chase customers between October 2012 and September 2015 to describe the key sources of income volatility among U.S. individuals and the size and growth of the Online Platform Economy. Our findings on income volatility are fourfold. First, while income volatility was prevalent across the board, it was most marked among the young, those in the bottom income quintile, and those living in the West. Second, median income individuals experienced nearly $500 in labor income fluctuations across months, with spikes in earnings larger but less frequent than dips. Third, most of the month-to-month volatility in take-home pay (86 percent) came from individuals staying in the same jobs. Finally, four in 10 individuals experienced a job transition in a given year, contributing 14 percent of the month-to-month volatility in labor income.To examine the Online Platform Economy, we assembled the largest sample of platform workers to date — a dataset of over 260,000 individuals who have offered goods or services on one of 30 distinct platforms. We provide several key insights. First, although only 1 percent of adults earned income from the Online Platform Economy in a given month, more than 4 percent participated from October 2012 to September 2015. We distinguish between labor platforms and capital platforms and find that, although labor platforms grew more rapidly than capital platforms, participation on capital platforms was more than 60 percent higher than participation on labor platforms in every month. Second, the platform economy was a secondary source of income for most people, and reliance on platform earnings did not increase for individuals over time. Third, earnings from labor platforms offset dips in non-platform income, but earnings from capital platforms supplemented non-platform income.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 125 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.