Wow, there is nothing like searching through your archives for an anecdote you want to recycle only to discover that the post you used it in pre-debunked the very same paper that you are now girding up to debunk.
A whole bunch of readers sent me either links to a paper “The Network for Global Control” by physicists Stefania Vitale, Stefano Battiston and James Glattfelder, or to reports on it (for instance, at PhysOrg, which I put on a black list because it has embarrassed me too many times before readers by reporting reverently on dubious studies). I pre-debunked a report on an earlier version of this paper, as reported in Inside Science, by Battiston and Glattfelder.
Let’s look at a breathless summary per PhysOrg:
Using data obtained (circa 2007) from the Orbis database (a global database containing financial information on public and private companies) the team, in what is being heralded as the first of its kind, analyzed data from over 43,000 corporations, looking at both upstream and downstream connections between them all and found that when graphed, the data represented a bowtie of sorts, with the knot, or core representing just 147 entities who control nearly 40 percent of all of monetary value of transnational corporations (TNCs).
In this analysis the focus was on corporations that have ownership in their own assets as well as those of other institutions and who exert influence via ownership in second, third, fourth, etc. tier entities that hold influence over others in the web, as they call it; the interconnecting network of TNCs that together make up the whole of the largest corporations in the world. In analyzing the data they found, and then in building the network maps, the authors of the report sought to uncover the structure and control mechanisms that make up the murky world of corporate finance and ownership.
This paper is a garbage-in, garbage out analysis. It does a network analysis based on share holdings at an institutional level. The problem is that the authors never bothered to understand how shares are held and how voting behavior varies based on institutional arrangements. To put it another way, the problem with the study is its failure to understand ownership, control and voting behavior in the institutional funds management arena.
As our Richard Smith tartly noted, “This looks like a list of people who *aren’t* in control. And since they own so much equity worldwide, that is interesting too.”
Let’s raise a very basic issue: indexing. In the US, passive management or indexing has become very popular. Crudely speaking, index funds compete on an institutional level by how closely they track the index and how low they keep their costs. An index fund manage does not care a whit about how the management of the companies in his fund are doing. The investors in his fund would be upset to learn if the fund manager was wasting his time thinking about how to vote the shares in their funds at the annual shareholders meeting.
But the authors didn’t bother thinking about how index funds complicate their tidy story of big powerful institutions controlling the world.
Another gaping omission is shares in brokerage accounts held in street name. The owner is the individual investor but again, this study would treat it as owned by the broker, say, Merrill Lynch.
Richard added to the list of significant oversights in the study:
· Custodians only vote as instructed by their clients. The paper doesn’t distinguish shares held at custodians from shares in the name of the underlying investor. Custodians’ infrastructure to support voting international shares is really poor which tells you, all by itself, how much that’s done. And all the voting processes and rules vary from country to country. So internationally held shares may tend not to get voted either.
· Passive indexers don’t vote
· Many so-called ‘active’ funds, that are actually closet indexers, may not do so either.
· IB inventory is hard to vote: because of the potential conflicts of interest that arise there tend to be in-house lawyers in the loop. That does not streamline the process…
· Insurers and pension funds should vote but often don’t.
· Where is Blackrock? Where are the SWFs? Rolled into custodian totals? Or too small to show? We can’t tell.
As we indicated above, we saw this all coming in the beta version of this study. But since it hews with conventional wisdom, no doubt it will be touted despite being abjectly wrongheaded. Our August 2009 post:
Conspiracy theorists will have to wait until the article described in Inside Science is published to determine whether it delivers on its claims. It purports to analyze stock holding across 48 countries and alleges they are held in very few hands. But the work was done by physicists, which means they may not have understood the limits of the data they were working with.
I suspect this will wind up resembling a paper a friend studied in his graduate level statistical methods course over two decades ago (he has since gone on to a successful career in academia). Everyone in the seminar was assigned a single paper and told to analyze the techniques used and to present their findings to the class. This was the sole basis for the grade.
The paper my buddy got had already created a bit of a stir, although it had not yet been published. The author had looked at the prices at which the Fed did its daily operations (then the famed “noon buying rate”) and compared it to the results of Treasury auctions. The paper concluded the Treasury was doing a terrible job, as demonstrated in all sorts of analyses.
When my friend’s day to present came, he stood up and said, “I have only one comment to make. The Fed conducts its daily operations in transaction sizes ranging in the millions. Treasury auctions are in the billions. The Fed data is irrelevant to the Treasury analysis,” and sat down.
He received an A.
In this case, an obvious fly in ointment is many (most?) stocks are held in street name, meaning in the name of the brokerage firm or fund, not the ultimate owner. I presume it is impossible to segregate accounts where the broker has discretion to trade versus those where the clients simply trades through the securities firm.
But even if the analysis is flawed, it might stir up some interesting discussion.
From Inside Science (hat tip reader John D):
A recent analysis of the 2007 financial markets of 48 countries has revealed that the world’s finances are in the hands of just a few mutual funds, banks, and corporations. This is the first clear picture of the global concentration of financial power, and point out the worldwide financial system’s vulnerability as it stood on the brink of the current economic crisis.
A pair of physicists at the Swiss Federal Institute of Technology in Zurich did a physics-based analysis of the world economy as it looked in early 2007. Stefano Battiston and James Glattfelder extracted the information from the tangled yarn that links 24,877 stocks and 106,141 shareholding entities in 48 countries, revealing what they called the “backbone” of each country’s financial market. These backbones represented the owners of 80 percent of a country’s market capital, yet consisted of remarkably few shareholders.
“You start off with these huge national networks that are really big, quite dense,” Glattfelder said. “From that you’re able to … unveil the important structure in this original big network. You then realize most of the network isn’t at all important.”
The most pared-down backbones exist in Anglo-Saxon countries, including the U.S., Australia, and the U.K. Paradoxically; these same countries are considered by economists to have the most widely-held stocks in the world, with ownership of companies tending to be spread out among many investors. But while each American company may link to many owners, Glattfelder and Battiston’s analysis found that the owners varied little from stock to stock, meaning that comparatively few hands are holding the reins of the entire market.
“If you would look at this locally, it’s always distributed,” Glattfelder said. “If you then look at who is at the end of these links, you find that it’s the same guys, [which] is not something you’d expect from the local view.”
Matthew Jackson, an economist from Stanford University in Calif. who studies social and economic networks, said that Glattfelder and Battiston’s approach could be used to answer more pointed questions about corporate control and how companies interact….
Based on their analysis, Glattfelder and Battiston identified the ten investment entities who are “big fish” in the most countries. The biggest fish was the Capital Group Companies, with major stakes in 36 of the 48 countries studied. In identifying these major players, the physicists accounted for secondary ownership — owning stock in companies who then owned stock in another company — in an attempt to quantify the potential control a given agent might have in a market….
Glattfelder added that the internationalism of these powerful companies makes it difficult to gauge their economic influence. “[With] new company structures which are so big and spanning the globe, it’s hard to see what they’re up to and what they’re doing,” he said. Large, sparse networks dominated by a few major companies could also be more vulnerable, he said. “In network speak, if those nodes fail, that has a big effect on the network.”
The results will be published in an upcoming issue of the journal Physical Review E.