How cautious did retail investors become?

  • From 22 October to October 28, net retail inflows were $3.3 billion, just half the weekly average of $6.6 billion over the past 12 months. ETFs attracted $2.9 billion, while individual stocks received just $400 million.
  • Overall retail activity measured by JPMorgan was around the 5th percentile, while ETF activity was even lower, at around the 0.4th percentile. Ahead of the earnings-season peak, retail investors took limited positions in companies due to report.
  • Meanwhile, crowding in the low-volatility factor across the broader market rose from near the 0th percentile in June to around the 18th percentile, reflecting an increase in defensive positioning. This measure covers the entire market, not just retail investors.

Tech remains favored, so why does buying look hesitant?

  • Retail investors continued to buy tech stocks over the past few days, but at a slower pace than the previous week; on 28 October, they became net sellers of tech stocks for the first time in a month. Flows favored semiconductors and hardware over software.
  • For the full week, retail investors net bought $585 million of Nvidia and $405 million of Alphabet. Outside the Magnificent Seven, net purchases in the technology and communications sectors totaled $187 million and $118 million, respectively, while net sales in healthcare and financials reached $338 million and $302 million, respectively.
  • Caution in the cash market did not extend to all trading: JPMorgan's model shows the rolling one-month average of retail investors' share of options trading at around the 86th percentile, although call-option volume declined, driven mainly by technology and consumer discretionary.

What different signals did big tech earnings send?

  • Alphabet's quarterly sales and profit fell short of expectations, and its shares dropped about 11% after hours; cloud revenue of $30.7 billion, however, was slightly above consensus. Third-quarter capital expenditure was $51.4 billion, up 114% year over year, primarily directed toward AI-supporting technical infrastructure.
  • Microsoft beat earnings expectations, with Azure revenue up 48% year over year and overall cloud revenue up 30%; its shares rose about 4% after hours. The company guided for capital expenditure of $50 billion in the next quarter, bringing calendar-year spending to $170 billion, and expects spending to accelerate further in 2027.
  • Retail reactions to other AI-related companies also diverged: SK Hynix's results comfortably exceeded the upper end of buy-side expectations, but retail investors took profits on Wednesday. Oracle forecast $30 billion in free cash flow for the fiscal year ending in May 2029, above analysts' expectations and easing concerns about debt-financed construction of AI facilities; retail investors nevertheless remained net sellers that day.

Why did Brazil's sharp fall draw retail investors in?

  • The report said Lula won Brazil's presidential runoff by a significant margin, a result that differed from polls and prediction-market pricing. Assets subsequently repriced, and the MSCI Brazil Index has fallen about 20% since Monday.
  • Retail investors bought the dip heavily in Brazilian ETF EWZ on Monday, with buying intensity reaching 13 standard deviations. ADRs of Brazilian companies were also mostly bought; for example, net purchases of Nubank totaled $65.2 million that day, while XP saw $16.7 million.

How do Fed signals affect the backdrop for flows?

  • Retail flows were subdued across nearly all ETF categories this week. Fixed income was particularly weak, with flow intensity in long-term US Treasury ETFs and investment-grade bond ETFs at negative 1.8 and negative 2.1 standard deviations, respectively.
  • The Federal Reserve held rates steady as expected, but added language emphasizing that it would gather more information to determine the policy stance needed to return inflation to its 2% target in a timely manner. JPMorgan economists viewed this as a hawkish signal and continued to expect a rate hike in December.