U.S. equity ETFs snapped up $3.1 billion in U.S. stocks in February, the second month of being net buyers on the market. They bought $1.8 billion in financials and $878 million in industrials, the two sectors with largest ETF buying. On the other hand, selling mostly concentrated in IT companies, which ETFs sold $744 million in stocks. U.S. ETFs sold $368 million in Apple's (AAPL) shares alone, the largest ETF share sale in a company in February.
Most noticeably, U.S. ETFs bought $211 million worth of Facebook's (FB) shares in the past month. Most transactions happened on two days - Feb. 12 ($160 million) and Feb. 27 ($62 million), so it also means ETFs were net sellers of $11 million in the rest of the month. The number of ETFs holding FB dropped in February from 50 to 48, so I believe the inflow was reflecting a weight increase in ETFs. After all, FB has rebounded over 50% since its lows in September. Notice the two spikes of FB volume in February - one occurred on the same day with the ETF flow (Feb. 12) and the other one happened one day after the ETF flow (2/28).
I noticed there were other cases when a spike in ETF flow was followed by a spike of the stock's volume the next day. For example, Corrections Corp of America (CXW) received the largest ETF flow after Facebook in February. The inflow came on Feb 27 as ETFs increased their holdings of CXW by 77.5% on that day, or $135 million. CXW's trading volume surged on Feb 28 to eight times the 30 day average volume.
I went on to examine all 236 cases in February in which transactions by ETFs caused an over 10% change of the total value of the stock held by ETFs. (e.g. if ETFs hold $10 million worth of ABC's stocks, and on one day they bought more than $1 million in ABC, that's an over 10% change.) It doesn't matter if the transactions were buying or selling, and I excluded a few cases that are caused by acquisition, as ETFs tend to sell a company's stocks before the company is acquired.
What I found is interesting. In those 236 cases, if the ETF holding's value is bigger than 5% of the company's market cap, on average the trading volume would be 2.14 times higher than the 30-day average volume one day after the spike in ETF flow. If the ETF ownership is between 3% and 5% of market cap, the next day volume would be 1.88 times higher than the 30-day volume. If the ETF ownership is between 1% and 3%, the ratio is 1.34 time. And for companies that have less than 1% of market cap in ETF, the next day volume would be only 1.10% times higher than the 30-day volume.
It makes sense that if the ETF holding is a big portion of a company's market cap, then a significant change of the value of the ETF holding should trigger a higher volume. While I'm not saying the finding here proves a relationship of statistical significance, I can at least suggest that it's more likely you'll see a higher volume on a stock if you know that ETFs hold more than 5% of its market cap and you just saw a big ETF flow in or out of that stock.
Our analysis of the ETF holdings data showed that over 900 companies have more than 5% of their market cap held by U.S. equity ETFs, and the median number is 3.6%. About two-thirds of these companies are small-caps with market cap below $2 billion, and about 100 companies are over $5 billion. Depending on the niche of the market, I think any investor/trader can use such a list of 30 to 100 stocks on a daily basis, in case the ETF flow data raise a flag on any given day.