The Economy's Missing Metrics

Technology has changed our lives in so many ways that simply cataloguing those differences is a daunting task. But it is important to do so because of the imperative to measure how we are doing in order to more effectively allocate resources.

It has been apparent for some time that traditional economic metrics and financial reports have failed to convey the evolution/revolution in the way people live their lives, conduct their business and interact with the world around them. Intangible assets like innovation, brand and law as drivers of growth have surpassed tangibles like property, plant and equipment as useful measures of success or failure.

Similarly, the benefits conveyed by services like Google search or Apple music or TripAdvisor directions have yet to be factored into the equations determining whether people are doing better or worse (themselves, open to interpretation) - and how that might be evaluated.

This should not be viewed as an indictment but as an opportunity as the following article explains. With bigger and bigger data combined with ever more powerful computational devices and techniques, we are limited only by our own capacity to imagine what might be. JL

Adam Davidson reports in the New York Times:

Economic statistics should be an essential part of a democracy: they could hold our leaders accountable by demonstrating whether the government was making life better or worse. Because they rely on fixed definitions — created decades ago — of the phenomena they’re charged with measuring, they do a poor job of capturing the ways in which people’s economic lives are changing.

Nearly every day, a few hundred federal workers, carefully spread around the United States by a rigorous statistical model, carry out a peculiar task on a special, secure tablet computer. Throughout the day, they are directed to visit specific stores in search of specific items — say, organic romaine lettuce hearts; or a 2015 Hyundai Sonata Sport with the premium package; or a men’s long-sleeve button-down shirt, blue, size XL and made of 80 percent cotton and 20 percent polyester. Over the course of the month, 80,000 prices are entered into tablets throughout the country, and the data flow to Washington for processing. Parsed and analyzed by economists, that information determines the official United States government inflation rate: arguably the most influential bit of data in the world, determining whether new factories are built, new employees hired.

Inflation is just one of the many economic statistics that the federal government creates. There’s unemployment, trade in goods and services, personal income, consumer spending, rental vacancy, new home construction and thousands more. These statistics are not objective facts of the world, like the standard boiling point of water or the mass of a stone; instead, their very measurement creates them. Slight shifts in the design of surveys or the assumptions about the answers can lead to quite different results.Because these assumptions are inherently debatable, the government often winds up publishing multiple versions of metrics. (Inflation has two distinct measures: personal consumption expenditures, known as P.C.E., and the Consumer Price Index, known as C.P.I. Unemployment has six, numbered U1 through U6, to reflect different interpretations of what ‘‘unemployed’’ actually means.) Our federal economic statistics are a remarkable resource: enormous, rigorous, essential. Together, they help us understand how the economy is working and whom it is, and isn’t, working for.

But in the decades to come, as we try to understand whether the American dream is deceased or merely slumbering, our statistics could get so much better. Because they rely on fixed definitions — created decades ago — of the phenomena they’re charged with measuring, they do a poor job of capturing the ways in which people’s economic lives are changing. The statistics are all but useless at measuring the change in general welfare created by new technologies, like Google Search, that make once tiresome tasks far easier to complete (at the cost of adding a whole universe of time-wasting distractions). In measuring employment, the stats are built around a model of full-time, fixed jobs in fixed locations; they struggle to keep track of Uber-like companies that employ people for brief gigs with no central workplace. Entrepreneurship, too, is measured quite crudely: It’s impossible, just looking at the new-firm stats, to distinguish the creation of Facebook from the opening of a small deli in Dubuque. Definitions of occupations are rigid and often archaic. Manufacturing work, for example, is broken down into dozens of discrete subcategories — tool grinders, sewing-machine operators, tire builders, woodworking-machine setters — while all ‘‘software developers’’ and ‘‘web developers’’ are organized into single catchall groups. You could learn a lot about the wages and employment patterns of adhesive-bonding-machine operators (18,210 workers in 2014, median wage $16.28 per hour). But you’d be hard-pressed to find any guidance on whether you’d make more money learning the Ruby on Rails computer-programming framework instead of developing your graphic-design skills.

Most of our economic statistics date to World War II. When Franklin D. Roosevelt first assumed the presidency, he asked his advisers how the economy was doing, and they responded with anecdotes: rail cars running half empty in Chicago, say, or department stores lagging in customers. Other than inflation, which had been measured since 1913, there were no reliable ways to judge economic strength over time. Spurred by F.D.R., the federal government sought to quantify the economy, creating a dozen different offices of economists and statisticians in the Departments of Labor, Commerce and Agriculture; the Federal Reserve; and several other agencies. The politicians wanted simple, straightforward answers to their economic questions. But the economists knew their answers were inherently subjective. They debated what the statistics should measure — arguments that weren’t resolvable with data because the data expressed the values and assumptions used in gathering them.

The most important thinker in these debates was Simon Kuznets, an immigrant from Russia. He was an economist at the National Bureau of Economic Research, an independent nonprofit that has served, since 1920, as the semiofficial host of fundamental discussion about the best way to measure an economy. Kuznets believed that economic statistics should be an essential part of a democracy: that they could hold our leaders accountable, by demonstrating whether the government was making life better or worse. Kuznets argued — as, incidentally, most economists do — that measures like employment and inflation are substitutes for what really matters: our quality of life, as each person defines it for himself. We measure money and other practical things because we don’t know how to measure happiness or fulfillment precisely.

In the 1930s, at the N.B.E.R., Kuznets amassed whatever data he could from industry sources and others and built the first comprehensive statistical model of an economy. Kuznets is seen around the world as the founding father of gross domestic product, and of national accounts more generally. But he didn’t like that honorific. In his 1971 Nobel Prize acceptance speech, Kuznets denigrated his accomplishment. He pointed out that government data doesn’t measure what really matters. It has no metric for ‘‘pollution and other negative results of mass production.’’ It doesn’t tell us if the country’s people have more mobility and freedom, more time to spend with family or to pursue pleasures or passions that don’t generate income. Throughout his career, Kuznets argued that military armaments should be heavily discounted in G.D.P. measures, because, by design, they destroy the world rather than build it up.

He sounded, at times, like a starry-eyed hippie. And that is certainly how he was viewed by the bureaucrats who set up our national accounting systems during the 1940s. They didn’t know how to value a mother’s ability to raise her children or what price to put on a pristine river or a mountaintop. They built our statistics around numbers they could gather, like the scale of industrial output, or the number of hours that a sample of American workers had spent on the job. Kuznets won the intellectual war, but he lost the practical battle. I’m fairly sure most economists, today, would prefer economic statistics that capture more fundamental measures of well-being. Instead, the government measures the numbers it always has instead of the ones that matter most to us.

It’s hard to blame the 1940s bureaucrats, who didn’t have electronic calculators, let alone computers. They could accommodate only so much data, so they built a system based on sporadic sampling and crude approximations of human happiness. Now, though, we have the potential to build a truly Kuznetsian system. Americans have access to devices that can know where they are at all times, what they’re doing, how fast their hearts are beating, how much sleep they’re getting, how much time they’re spending at work or with family or out on a lake. This sort of data is being collected — with or without our knowledge — by tech companies: Google, Fitbit, Withings and countless more. These companies want to profit from this information. But an anonymized sample of all these kinds of data, combined with other, more traditional economic metrics, could create a whole new public system for economic statistics.

The cost would be trivial in the context of an economy where a single statistic can cause billions to change hands in an instant. It is hard to think of any government investment that would have a greater impact than creating robust ways to measure the quality of our lives. We are in the early stages of an election period in which there will be significant, competing claims about the best ways to manage a modern economy. The only way to judge those claims is with data — the richer and more human the better.