FROM THE WALL STREET JOURNAL BY JOSEPH STERNBERG| FEBRUARY 18, 2014
China is in an economic swoon at the moment, if you believe some recent readings suggesting that industrial activity is slowing. Investors in Japan were spooked last month by a record trade deficit in 2013. Closer to home, economists debate quarterly GDP numbers in the U.S. to determine how the current recovery compares with those in the past.
In these and countless other ways, economic data shape our political debates, our business decisions and our investment choices. Yet, as Zachary Karabell points out in "The Leading Indicators," our fetish for economic data in general, and for certain headline numbers in particular, is a relatively recent phenomenon. And too often we end up hostage to the data we ourselves have created.
Attempts to count on a national scale trace back to the Bible or before, and Mr. Karabell begins his story with William the Conqueror's "Domesday Book" census of 1086. But the Enlightenment saw the first serious attempts to measure national resources in a precise way, in service of more scientific government.
One early enthusiast for "political arithmetick" was 17th-century Englishman William Petty. Motivated by a conviction that national wealth equated to national power, Petty set about counting England's assets. Petty's work was more guess than measurement, but Mr. Karabell notes that he nevertheless succeeded in "establishing the principle that sound policy could be made only if rulers could gauge the actual wealth of the nation."
The word "statistic" itself traces to a German, Gottfried Achenwall, in 1749 and is a portmanteau of Latin and Italian words meaning roughly a description of the state. Theorists like Petty and Achenwall saw a direct link between data collection and governance. So did the American Founders, in a way, when they mandated a decennial census in the conviction that accurate population measurement would facilitate effective representation and good government.
Initial efforts to conduct anything other than bare-bones population counts were hobbled by the practical difficulty of collecting large amounts of data and the inability to analyze data absent modern statistical methods. The second half of the 19th century saw advances in solving both of those problems and also the rise of a progressive philosophy that demanded leading indicators as inputs for the expanding machinery of more activist—and, progressives claimed, more scientific—governance.
The author highlights the extent to which progressive ambition fueled the creation of new economic measurements. In a congressional hearing addressing the issue in 1931, Frederick Dewhurst of the Commerce Department testified that he was "impressed constantly with the requests for information that we get in the Department of Commerce, so many of which have to be answered to the effect that we do not know." Sen. Robert La Follette Jr. of Wisconsin sought to address this shortcoming by introducing legislation to create new accounting for national income in 1932 to help "Congress to determine policies."
The leading indicators that Mr. Karabell describes are central to modern governance. The measurement of GDP in theory allows one to gauge how far actual output deviates from "potential" output—a foggy calculation on which hundred-billion-dollar stimulus plans are founded. Consumer-price inflation data create the illusion that a fiat currency can provide the same price stability that a metallic standard would by giving central bankers a target at which to shoot. Data also help politicians build support for their policies, a technique adopted by Franklin D. Roosevelt in his 1936 campaign when he defended his deficit-spending binge by pointing to supposed increases in national income.
How scary, then, to read Mr. Karabell's recitation of the flaws with all these data. Washington's GDP statistic until recently didn't include the value of much intellectual property created on America's shores. The unemployment rate doesn't include unemployed individuals who have given up looking for work or who are underemployed. And inflation indexes are prone to manipulation both in what they exclude—think of the "core" consumer-price index, which doesn't include the things like food and gas that people spend so much of their money on—and in how they measure what they do cover.
Mr. Karabell offers an engaging account of the history of these indicators, and his explanation of their flaws is both readable and useful for non-economists trying to make sense of the barrage of numbers with which they're pelted on a regular basis. No one can argue that we need to be more informed, and realistic, about what and how big-picture data measure.
But how then to answer our modern demand for data? Here Mr. Karabell's solution is less compelling. He proposes "bespoke indicators" measuring micro trends of relevance to specific individuals or companies, or data aimed at answering specific policy questions, to free us from outdated, one-size-fits-all numbers. This advice is best suited to businesses, an increasing number of which are already doing precisely this.
But concerning governance, where leading indicators arguably wreak the most havoc, the fundamental problem isn't that the data we use today are flawed but rather that any metric inevitably misrepresents a system as complex as a modern economy. Look at how challenging it has been to get an accurate count even of the number of uninsured Americans pre-ObamaCare.
There is another way. We can return to two economic data points that Mr. Karabell mentions either only in passing or not at all: votes and prices. Popular assent to rulers has often provided a guide to economic conditions that is at least as reliable as any statistic an economist can concoct. Call it the original measure of "gross national happiness." Meanwhile, prices for capital, inputs and outputs provide better guides to entrepreneurs and managers than any leading indicator. This is, after all, how entrepreneurs made decisions before the advent of leading indicators.
A tyranny lies behind our reliance on leading indicators: the conceit that data-driven technocrats can know the everyday business of our lives better in aggregate than we do individually. The rolling ObamaCare debacle is bringing to the fore a deepening skepticism about the competence of data-driven progressivism. Not a moment too soon.