July 6, 2010 / 8:51 PM / 10 years ago

Special Report: U.S. data dogs on quest for sexier statistics

WASHINGTON (Reuters) - Alan Krueger can wax poetic about data — literally.

The top economic adviser to U.S. Treasury Secretary Timothy Geithner, Krueger quoted poet Carl Sandburg in an 84-page research paper he co-authored proposing a new database to measure how people spend their time in order to understand what makes the economy tick.

"Time is the coin of your life," Sandburg wrote. "It is the only coin you have, and only you can determine how it will be spent. Be careful lest you let other people spend it for you." (here)

Before he was appointed Assistant Secretary for Economic Policy last year, Krueger was a scholar whose passion was time use studies. Now he is in a position to bring this type of information into economic policy decisions in Washington.

The United States is deluged with economic data, yet the figures cannot conclusively answer even the most fundamental questions: Is the recession really over? Are people living better? Is government serving its citizens well?

Inside the corridors of power, the nerds are stirring. A handful of data-loving economists in key positions at Treasury, the Commerce Department and the White House are pushing for alternative measures to provide a clearer picture of how well the economy is working.

The interest is more than just academic. U.S. economic data moves stock markets, drives public policy, and can even swing a presidential election.

No one is talking about jettisoning the classics like gross domestic product, or GDP, the broadest measure of the nation’s economic output. But by combining that information with deeper understanding of how people live, work and feel, officials hope to identify economic trouble spots more quickly and make better policy decisions.

“If you were running a $2 trillion company, would you spend such a small share of your budget collecting data on how it’s working?” Krueger said in an interview in his office, where stacks of economic data sit on bookshelves and tables. “We grossly underspend on statistics.”

In fact, the Agriculture Department spent more than four times as much on research last year as the total budget for the Bureau of Labor Statistics, the group responsible for crunching some of the most critical numbers such as the jobless rate.

Krueger isn’t the only data devotee in high places. Austan Goolsbee, one of President Barack Obama’s economic advisers, called himself a “data dog” in his Senate confirmation hearing last year. And the data dogs may finally have their day.

Two new sets of statistics are due to be launched next year. The Labor Department is working on an enhanced time use study to track what Americans do all day and how they feel about those activities, a project that draws on Krueger's academic research. (www.bls.gov/tus/)

And the Commerce Department is planning a new poverty metric it hopes will provide a more up-to-date measure of which groups are struggling to meet basic needs. (here)


The latest recession proved what we don’t know can hurt us. The Federal Reserve vastly underestimated the destructive power of a U.S. housing slump. One reason the Fed was so off-base is that there was little or no data tracking the trading of complex financial instruments created in the past decade.

Data on economic output also did not tell the whole story of how this downturn hurt living standards, which helps explain why President Barack Obama still gets low marks on his handling of the economy even though it has been growing for a year.

As the debt crisis in Greece calls attention to America’s own wobbly public finances, it is becoming more important to understand how people live to determine where government ought to channel increasingly limited resources.

Rebecca Blank, Undersecretary for Economic Affairs at the Commerce Department, senses an opportunity to persuade Congress to allocate more money for statistics now that lawmakers understand the cost of not knowing. “We have seriously underinvested in data in this country because it’s not sexy,” she said.

Like Krueger, Blank was trumpeting better poverty data long before she took up her Commerce post, and now she finds herself in a position to make it happen. (www.brookings.edu/~/media/Files/rc/papers/2008/12_poverty_measurement_blank/12_poverty_measurement_blank.ashx)

“The problem with data collection is someone has to be interested in it and it costs a little bit of money,” she said.

Obama’s budget request for next year includes $16 million more for Blank’s economics and statistics administration, and an additional $34 million for the Bureau of Labor Statistics.

Both require congressional approval, and while the sums are a tiny portion of the total $3.76 trillion in government outlays, deficit hawks are ascendant in Congress so nothing is guaranteed.


For all its shortcomings, U.S. economic data is arguably the most robust in the world. Compared with Europe, for example, U.S. readings on employment and GDP are released faster and contain more detail. There is, however, a trade-off between speed and accuracy, and U.S. reports are subject to revisions that can be dramatic.

Between the Labor and Commerce departments, there are dozens of sets of statistics released each month, and that doesn’t even count figures from the Federal Reserve and Treasury. In addition, there are various surveys provided by private sector groups such as ADP Employers Services which provides employment data from private firms, as well as the National Association of Realtors which supplies housing reports.

But they cannot measure everything, and the holes tend to be more noticeable in times of crisis.

The Commerce Department can tell you that Americans spent $714 million at men’s clothing stores in April but there is no tally of how many families missed meals even though their income exceeded the cutoff to qualify for food stamps.

Labor Department databases can tell you how many people work at museums, parks and zoos, but cannot accurately gauge how many stay on jobless benefits because it makes more economic sense than taking a low-paying position and spending half of every paycheck on childcare.

Just how imperfect economic data can be was laid bare during the past two years. The Labor Department underestimated the number of unemployed by nearly 1.4 million during the recession because of a failure with the economic model it used to estimate how many companies opened or closed in a given month.

That meant the job market was far weaker than thought when the Obama administration took office. Better information would not have altered the policy approach, Treasury’s Krueger said, but it might have made it easier for the White House to make the case for a massive stimulus package.

“The public would have been more confident that the medicine we were applying was working,” he said. “Maybe it would have generated more support for more action or swifter action, but I don’t think it would have fundamentally changed the direction of the administration’s policy.”


The Labor Department isn’t the only agency to get the numbers wrong occasionally. Douglas Holtz-Eakin, former head of the Congressional Budget Office which crunches numbers for Congress, recalled he once missed a budget forecast by $100 billion because he had to rely on outdated statistics.

“It’s embarrassing to miss by that much,” he said.

Holtz-Eakin said it was a “constant struggle” to get reliable, real-time data, and simply spending more money on data gathering isn’t enough. He said businesses object to the added paperwork burden of complying with government data requests, and individuals have privacy concerns.

Randall Kroszner, who served on the White House’s Council of Economic Advisers under President George W. Bush, said he also pushed for better data but with little success.

The current head of the CEA, Christina Romer, said her chief forecaster has to remind her not to get too caught up in any single piece of data because it could prove to be anomalous or get sharply revised the next month.

“They’re state-of-the-art,” she said of U.S. economic data. “They’re still imperfect.”


The quest for better data goes back generations.

Gross domestic product, the best known and still the most important gauge of the economy, traces its roots to the Great Depression, when Congress ordered a report on national income.

The New York Times carried a two-paragraph Associated Press story about it on June 14, 1932, placing it just above a brief report about a furrier who committed suicide by drinking a bottle of poison because of “recent financial reverses.”

The task of drawing up the national accounts fell to a young economist named Simon Kuznets who was working at the National Bureau of Economic Research in New York. His report to Congress in 1934 was a bombshell. National income had dropped 40 percent in four years. A measure of inflation tumbled 32 percent. (here/SOD&CISOPTR=888&REC=2)

Kuznets, who was awarded the Nobel prize in economics in 1971, set up a framework for measuring economic output that still forms the basis of today’s GDP.

Some economists say the measure looks outmoded. Today’s economy is very different from the 1930s. Kuznets broke out 12 major industrial categories, starting with agriculture, which in 1930 employed 21.5 percent of the U.S. work force.

Services, now the largest component of the economy, was listed eleventh, just ahead of “miscellaneous.”


Kuznets devoted part of the first chapter to spelling out the limitations of the data. Some of the difficulties he identified still frustrate economists today.

In its early days, the national income data was used to help figure out how quickly the United States could ramp up production for World War Two, and some economists say it still has a war-time feel. Kuznets himself was keenly aware the data did not and could not measure well-being. He wrote about the “reverse side of income, that is, the intensity and unpleasantness of effort going into the earning of income.”

Nearly eight decades later, economists are still trying to figure out how best to put a number on well-being. Krueger sees time use studies as a step in that direction. GDP “doesn’t include the joy we get from walks, or the pleasure we get from a game of chess,” he said.

The biggest hole in Kuznets’s national income report may be what he called the “services of housewives and other members of the family,” which includes cooking, cleaning and childcare.

“It was considered best to omit this large group of services from national income, especially since no reliable basis is available for estimating their value,” the 1934 report to Congress says.

Nancy Folbre, an economist who teaches at the University of Massachusetts, is still fuming over that line. “Many people have a very forgiving attitude toward Kuznets,” she said. “I’m not convinced. Kuznets was pretty dismissive of the issue of household production.”

She calls herself the “token spokesperson” for the cause of including household labor in measures of economic output, and spoke about it at a Sarkozy Commission conference last fall looking at the adequacy of economic indicators.

The commission, created by French President Nicolas Sarkozy in 2008, includes Nobel laureates Joseph Stiglitz and Amartya Sen. In a report released last September, they called for broader measures of the economy that go beyond just growth. (here)

Ignoring the value of household labor can skew readings on well-being, poverty and inequality. Imagine two families, each with two parents and two children, each earning a household income of $50,000. One has a single bread-winner earning that full amount, while the other has two working parents, each earning $25,000.

In the first family, the parent who doesn’t work outside the home has time to care for children, make meals or clean, services that the second family might have to pay someone else to provide.

Looking solely at household income, the two families appear similarly situated. Add in the value of household labor and one family looks much poorer.

Folbre wants satellite accounts that would include the value of this so-called non-market labor. The Commerce Department is working on that, but Folbre is tired of waiting.

“Let’s stop talking about the damn satellites,” she said. “Put the satellites into orbit.”


Krueger’s time use study can shed light on how much time is spent on housework, if not the value of it.

The Labor Department began collecting this information in 2003 and the first estimates were released the following year. The agency conducts telephone surveys and asks people how they spent their time over the previous 24 hours.

Starting this year, Labor is adding a new component that will ask people a subjective question — how they feel during various activities.

To the untrained ear, that may sound as mundane as telling a neighbor to have a nice day. But it can be a vital addition for policy decisions. For example, Krueger said the data shows people spending more time commuting, and it’s an activity they loathe, which suggests government ought to direct more money to improving infrastructure and transit.

Roughly 50 other countries are collecting similar data, including Germany, Australia, Canada, New Zealand and South Africa, the U.S. Labor Department said. To keep the results consistent, there is a coding lexicon to ensure interviewers are properly classifying each activity.

Although the responses are officially listed only under broad categories such as sleeping or housework, the level of detail in the lexicon is astonishing — and perhaps a bit unsettling to privacy advocates worried about Uncle Sam peering a bit too closely into people’s lives.

Under “grooming” it lists among the possible responses everything from gargling mouthwash to brushing lint off clothing. It differentiates between religious and nonreligious meditation (the former goes under “religious and spiritual activities”; the latter “health-related self care”).


The Labor Department did not have the funding for this enhanced time use project, so Krueger found an unlikely benefactor at the National Institutes of Health.

Richard Suzman is director of behavioral and social research at the National Institute on Aging, part of the NIH, and a data fan himself. He has been pushing for better health-related economic research since he joined the agency in 1983.

“There were some people who said ‘We don’t do economics at NIH’,” he said. “I did some of it below the radar for a while.”

Now, he said his agency is “probably the largest funder” of such research, although that is primarily because there isn’t much funding for it.

The link between health and wealth goes both ways. Poor health can hurt finances, and poverty is often associated with higher incidence of illness.

Suzman is interested in how older people spend their time. What do retirees do with themselves? How much exercise are they getting? How does time use change with illness or disability? Do they spend much time alone?

The answers to all those questions can influence health and well-being. That in turn has an impact on government spending on Medicare, the healthcare program for the elderly and one of the biggest future strains on U.S. public finances.


The Commerce Department has no other funding source lined up, Blank said, so if Congress denies the Obama administration’s request for extra money the new poverty data won’t be produced.

Even if it does prevail, the new data won’t replace the current poverty measure, which will still be used to determine eligibility for various government aid programs. Instead, the new report is designed as a “supplemental” data set that better reflects 21st century realities of poverty.

The existing measure, which dates back to 1964, has its critics from both sides of the political spectrum.

Liberals have long complained that it doesn’t take into consideration things like childcare costs, which were less of an issue 50 years ago when most mothers stayed home.

Conservatives say it overstates poverty because it excludes government benefits such as the earned income tax credit, food stamps and housing assistance.

The supplemental data is trying to address both sets of concerns.

Bruce Meyer, an economist who teaches at the University of Chicago, said including the earned income tax credit would lift six million people above the poverty line, and he wants the data to measure what people consume rather than their income.

He has a broader concern about the quality of government data collection. Normally, government data agencies rely on surveys of businesses and consumers to get numbers on employment, retail sales, and a host of other indicators.

What they cannot accurately count, they impute. As the Labor Department’s miscount of unemployment shows, imputing can be error-prone, but the government is having to do more of that because fewer people are willing to respond to surveys.

“People are over-surveyed,” he said. “It’s no longer a novelty to be asked your opinions.”


The private sector is wading in where the public sector falls short. The University of Michigan has been polling consumers for decades, but more recent entrants include credit card issuers, paycheck processors, and even Google.

Hal Varian, Google's chief economist, said he has spoken to a number of government officials at agencies he did not want to name about what sort of information Google can compile. By tracking search terms such as "Where is the unemployment office" Google is trying to predict how many people will apply for initial jobless benefits in a given week. (here)

Reuters, Bloomberg and other financial news services poll economists for their forecasts on such data every week. Varian said some weeks, Google’s jobless benefits tracker predicted that number more accurately than the economists.

“This fully automated solution is doing at least as good a job as 30 people out there poring over the tea leaves,” he said.

Google is also producing predictions for other indicators including retail sales. Varian said the company is thinking about branching out into politics, predicting election results based on its search data.

In theory, data collection ought to be easier in the electronic age. Companies such as retail behemoth Wal-Mart can track sales the second a credit card is swiped. It may be only a matter of time before GDP can be calculated instantly too.

Krueger, Treasury’s data dog, wants to create a “rapid response” data gathering program that could quickly crunch numbers on how a new development like the swine flu outbreak or the BP oil spill might affect the economy. “It is as if you only had a speedometer that told you your speed five minutes ago,” he said in a speech to economists back in October.


The better data movement has been somewhat subsumed by the Sarkozy Commission and the “economics of happiness” trend which aims to measure well-being. Some economists cringe at that phrase because it sounds trivial.

Carol Graham, a senior fellow at the Brookings Institution think tank in Washington, said when she wrote a book on that topic 10 years ago, her boss told her to take “happiness” out of the title or no one would take it seriously.

“Needless to say, I refused and the book didn’t do all that well either,” she said.

Graham has no trouble getting people to take her happiness research seriously now. In fact, The Economist magazine held a debate in late April on whether GDP or happiness measures were better at gauging living standards.

Keith Hennessey, a top economic adviser in the last Bush administration, argued in favor of GDP even as he acknowledged its limitations.

“Money cannot buy happiness, and GDP cannot measure it,” he said, adding that GDP was adequate and superior to gauges “based on someone’s subjective decision about how you should measure your happiness.”

The United States is not going the way of Bhutan and its Gross National Happiness indicator, but officials are paying closer attention to how people feel — particularly now when the population is awfully unhappy despite a growing economy.

Obama says frequently that the recession isn’t over for those still out of work. But that doesn’t mean he and his team of economic advisers aren’t watching every GDP report, and smiling when the numbers go up.

Romer, the head of the Council of Economic Advisers, said well-being measures have their use, value and place, but not in the GDP tables that Kuznets pioneered.

“GDP is doing pretty well at what it set out to do, which is measure the total amount of everything we produce,” Romer said. “I am very happy to look at other indicators of well-being, but I would rather not muck up GDP with lots of other pieces.”

Editing by Jim Impoco and Clive McKeef

0 : 0
  • narrow-browser-and-phone
  • medium-browser-and-portrait-tablet
  • landscape-tablet
  • medium-wide-browser
  • wide-browser-and-larger
  • medium-browser-and-landscape-tablet
  • medium-wide-browser-and-larger
  • above-phone
  • portrait-tablet-and-above
  • above-portrait-tablet
  • landscape-tablet-and-above
  • landscape-tablet-and-medium-wide-browser
  • portrait-tablet-and-below
  • landscape-tablet-and-below