Some items on our site have recently moved. Visit our News Hub for selected articles, special reports, podcasts and other resources.
BlackRock challenges 'common ownership' paper on airlines with fresh data
30 January 2019 20:51
BlackRock, the world’s biggest asset manager, has amassed hard data casting doubt on an academic paper central to the now-popular hypothesis that index funds may be harming competition by owning stakes in rival companies.
The BlackRock analysis, seen by MLex, challenges research that drew a connection between this “common ownership” of shares and increased airline ticket prices.
The data underline a simple reality: During the years covered by the study, five of the airlines filed for bankruptcy protection, meaning their shares were delisted from stock-exchange trading and dropped from indexes such as the S&P 500 for stretches of the study period.
Funds that passively track those indexes would then have sold the shares and stayed away from them until the airlines exited bankruptcy and were added back into the index. US Airways, which entered bankruptcy twice during the study period, was not represented in the index for more than four years, the analysis says.
“Assuming asset managers ‘control’ companies they don’t even own is clearly without merit,” it says.
— Airline paper —
The original research by José Azar, Martin C. Schmalz and Isabel Tecu showed that asset managers BlackRock, Vanguard and State Street held stakes in competing US airlines.
The top seven shareholders in American Airlines, for example, jointly controlled 49.55 percent of the carrier’s shares and were also among the top 10 investors in Southwest Airlines and other competitors, the academics said. At Southwest, the top six shareholders were also top-10 shareholders of American and Delta.
The implication was that such investors might have an interest in curbing the competitive instincts of the companies involved. If you own shares in two airlines, there would be no benefit in one cutting its ticket prices to take a bite out of the other, the theory goes.
The phenomenon left airline ticket prices much higher than they might be otherwise, the research estimated.
— Gaps —
The authors based their findings on the investors’ quarterly “13F” disclosures of their holdings to the US Securities and Exchange Commission. Their paper did acknowledge that this dataset presented a gap.
“Holdings are not observed during bankruptcy periods,” they said. Their solution: “During the bankruptcies of American Airlines, Delta Airlines, Northwest Airlines, United Airlines, and US Airways, we repeat the last observed value for percentage of shares owned.”
BlackRock’s new analysis faults this approach.
“The authors noticed that the data did not show holdings in affected airlines during bankruptcy periods,” the study says. “So they chose to override the data from the filings with an erroneous assumption that holdings during bankruptcy are the same as those found in the 13F data for the quarter preceding the bankruptcy.”
It’s true, of course, that asset managers may have actively managed funds as well as index funds, meaning they might have continued to hold some airline shares during the bankruptcy period. So BlackRock ran the data to produce a chart showing how many shares it held in American Airlines from the time the airline filed for bankruptcy protection on Nov. 29, 2011, to its exit roughly two years later.
The fever line looks like a crooked U, with BlackRock’s holdings plunging from 14 million shares to almost zero after American Airlines entered bankruptcy and staying there until shortly before the carrier exited.
The chart “illustrates the magnitude of the discrepancy between the actual holdings of American Airlines stock in BlackRock-managed portfolios and the holdings that were used in the ‘airlines paper’ based on this error in the authors’ data compilation,” the analysis says.
“The discrepancy is in the order of millions of shares which reflects an actual ownership under 0.09 percent versus the authors’ assumption of 4.35 percent,” BlackRock said. “This is a significant error that affected 28 out of 56 quarters in the study period.”
No results found