Category Archives: Statistics

Historians and their Index Numbers

John Steele Gordon argues—over on Bloomberg’s recently revamped “echoes” blog—that historians of the US stock market in the mid-twentieth century has been misled by that market’s most prominent index. The handiwork of a publisher (Dow) and a statistician (Jones), the Dow-Jones Industrials evolved from a series of focused indexes into a single number meant to represent the entire NY exchange, and by proxy the American economy.

But for all the power and influence this number has had, Gordon shows how dependent it is on basic assumptions. Swap out AT&T for IBM in the Depression years and the market recovery comes years before we have generally thought.

For our purposes, the Dow, its development, and public understandings of stock indexes strike me as topics awaiting a historian of science’s analysis. I would read that book.


If you haven’t seen the new “echoes” blog—edited by Stephen Mihm, the UGA historian of capitalism in the US, it’s worth a peek.

Also, as long as we’re talking about American science and index numbers, here’s a shout-out to Tom Stapleford’s recent history of the Consumer Price Index, which also happens to be a fascinating history of American statistics generally.

Statistical Infrastructure

It has little to do with America directly, but I am fascinated by the New York Times‘ coverage of India’s new nation-wise statistical and biometric registry: “Aadhaar”—which translates, in an Asimovian twist, to “foundation.”

The project aims to assign a 12-digit ID to every Indian—that’s 1.2 billion IDs—and link those IDs to names, fingerprints, and iris-scans. As Lydia Polgreen, the Times reporter, notes: “It is a project of epic proportions.” It also promises to make the Indian government into the world’s most important aggregator of biometric data, surpassing the US-Visit program by an order of magnitude.

Nandan M. Nilekani, the former chairman of Infosys and Aadhaar’s head, explained the necessity of the system in terms that made it sound like a natural governmental activity: “What we are creating is as important as a road.” It is, in other words, a kind of infrastructure: statistical infrastructure. That’s a phrase I use quite a bit in my own work as I trace the ways that different systems for gathering data about individuals developed in the nineteenth and early twentieth centuries around the life insurance industry. In that story, private and public actors worked in parallel and sometimes together to improve the nation’s system of vital statistical registration, to discipline doctors and nurses, and to build special biometric (of sorts) databases that could help assess each individual’s risk. Yet the United States’ own giant leap in gathering data (Social Security) created a national identity database only as an after-thought and had no thought of including biometric data.

That’s the most intriguing thing about Aadhaar, as viewed through Polgreen’s reporting. Identity sits at the center of the project. Polgreen begins with Ankaji Bhai Gangar volunteering to be IDed with hopes of getting “the first official proof that he exists.” She ends with Mohammed Jalil pointing to the biometric station and saying “This will give me an identity….It will show that I am a human being, that I am alive, that I live on this planet. It will prove I am an Indian.”

I’m wary of Polgreen’s enthusiasm. She brushes aside concerns of “privacy watchdogs” effortlessly. She thrills at the possibililies of overcoming corruption on the local level and getting around the “crippling bureaucracy that is a legacy of [India’s] socialist past.” Aadhaar, we learn, will increase worker mobility and allow for greater agricultural modernization—these are both, we are made to understand, necessarily good things.

I’m all in favor of reducing corruption, improving the distribution of poor relief and welfare benefits. I think the poor ought to have access to savings, credit, cell phones, and teachers who show up to work. Who doesn’t? But will a centralized, national system of identification really do that? Does bypassing local government—rather than, say, fixing it—solve that problem? I can’t pretend to know, but I think there’s reason to be skeptical with any theory of improving governance that tries to bypass local institutions. I do hope my fears prove entirely unfounded.