“Civilization is the limitless multiplication of unnecessary necessities.”
Many observers in the run-up to the current financial crisis recognized that the American appetite for multiplying unnecessary necessities—like houses consumers bought but could not afford, and that developers built but could not sell—was creating a financial house of cards that could only spell trouble over the long run. But few predicted how devastating it would be when the cards came tumbling down. Part of the reason for both the scale of the crisis and the immensity of our surprise in the face of it is that the tools market watchers were using to measure and assess financial phenomena were inadequate.
Niall Ferguson has insightfully described our current predicament as the end of “planet finance.”1 By “planet finance” Ferguson means to draw attention to the fact that since roughly 1980 an elaborate virtual world, practically parallel to the one in which we live and work every day, evolved through inscrutable credit and investment mechanisms. “Credit default swaps” and “mortgage-backed securities”, once obscure expressions, have now become household terms, even if their exact nature remains mysterious. One need not be able to explain what they are to understand that financiers and investors were way off target in determining the values and risks of such instruments. The measures and the reality supposedly measured were out of sync.
Many policymakers have sensed the mismatch between reality and measurements of it at least to some degree, but they have preferred to focus their rhetorical fire on calls for enhanced regulation. In the words of French President Nicolas Sarkozy, the world needs a “better-regulated form of capitalism with a greater sense of morality and solidarity.” Just about every major world leader has uttered similar words, though most show their affection for morality and solidarity by haranguing greedy investors rather than either complicit or slumbering politicos and bureaucrats. Moral scolding alone, however, won’t restore confidence in capitalism. As policymakers embark on a new adventure to bring greater regulatory clarity to the markets, they need to remember that success rests in large part on their ability to understand economic reality, and that succeeding in this regard is as much art as it is science. They will eventually have to face an enduring truth: Most of our economic measures are flawed.
Consumers and Consumption
Debates have raged throughout the media about the difference between the financial sector and the “real economy” during this crisis. One thing both have in common is that they are always more complex than the tools we use to measure and understand them. This is perhaps most evident when we consider a key factor that links the real economy and planet finance: consumption.
No normal person wants to be known as a defender of consumerism, but economists are usually more than happy to defend consumption. Consumption is largely how we understand progress: We measure payment for goods and services as a way to understand whether our national life is improving. We call this measurement Gross Domestic Product.
GDP is a measure of all the goods and services produced in a given country, and, along with investment, government spending and net exports, consumption is a major pillar of how we calculate GDP. The problem is that what we often call “consumerism” also worms its way into the measure of consumption. If we understand consumerism to reflect the acquisition of goods and services that people cannot afford and probably do not need, then measuring it as a part of consumption might suggest that avarice is a virtue. It isn’t, of course. We know that living beyond our means is not a good thing, which means that including it in our GDP calculations is also probably not a good thing. And yet we do this as a matter of course, partly because it is not always intuitively obvious when consumption stops and consumerism begins. If millions of families threaten their financial future by financing excessive purchases through home equity loans or overburdened credit cards, GDP goes up anyway.
To make matters more complex, the inverse is also true: GDP increases when goods are produced but sit unsold. For example, before the housing markets began unraveling in 2007, the U.S. inventory of homes that had gone unsold for more than a year had been on the rise. The construction of new homes was included in our GDP calculations even if no one bought them. In fact, any time inventory grows because of weakened demand, regardless of whether the product is a house or a tube of toothpaste, GDP counts it as increased productivity. Therefore, whether consumers leverage their homes to finance excessive consumption or builders produce homes that no one buys, GDP rises.
I don’t mean to suggest that GDP or any similar measure is meaningless or necessarily misleading. To the contrary, investors, executives and policymakers need GDP figures to make informed decisions about a range of issues that ultimately affect real people’s lives, and GDP is generally helpful in this regard. The point is that economic reality and economic measures are invariably partners in a dance in which the former leads and the latter tries to follow. Sometimes the dance works, but other times the partners fall awkwardly out of step. When the latter occurs, it does so because of two main weaknesses that mirror each other: Sometimes they mask positive realities, sometimes negative realities. These weaknesses skew decision-making by policymakers and investors in ways that affect the resources and quality of life of massive numbers of people. This point can be illustrated by looking at some commonly used economic measures in the United States, GDP first and foremost.
Gross Domestic Product: GDP is beset by two key imperfections beyond those already suggested. The first was noted by Adam Smith, the patron saint of free-market capitalism, long before GDP was invented. As Smith wrote in The Wealth of Nations, “There is one sort of labour which adds to the value of the subject upon which it is bestowed: there is another which has no such effect.” He called the former productive labor, the latter unproductive labor. His first example of unproductive labor is a “menial servant”, or someone whose work loses its value the moment it is performed. He notes that the “labour of some of the most respectable orders in the society is, like that of menial servants, unproductive of any value.”2
Smith’s insight is old but not obsolete. We might wonder about, say, the productive value of auto body shops or medical malpractice lawsuits. We need to fix our cars after an accident, and we need justice when doctors are careless, but neither adds much to a society’s productivity, however much employment or income it provides. Or we might wonder about the $5 billion spent every year on U.S. tax-preparation services: The more complicated the U.S. tax code becomes, the more tax-preparation services grow, but without adding anything to the nation’s productivity. Robert Kennedy once said that GDP “counts air pollution and cigarette advertising, and ambulances to clear our highways of carnage . . . [I]t measures everything, in short, except that which makes life worthwhile.” Kennedy’s remark, though delivered in characteristic political hyperbole, makes the compelling point that just because people pay others for some good or service does not mean that it contributes to a society’s well-being. Measuring in such a way that all spending connotes progress toward a broader and deeper prosperity is misleading.
The second imperfection with GDP is that it not only counts non-productive work as a good thing, it sometimes counts counterproductive activity as a good thing. Growing inventories, as cited earlier, are included in GDP, even though they may signal weakened demand. In the last quarter of 2008, growing inventories in the United States added 1.32 percentage points to GDP at a time when demand was falling precipitously. Final sales during that quarter fell by nearly 5 percent, the largest quarterly drop in 28 years.
Public policy may also drive up the cost of doing business in ways that increase prices while adding little by way of productive value. Housing regulations, for instance, have been on a steep rise over the past thirty years, and by the time the housing market began crumbling in 2007 they likely accounted for between $48,000 and $84,000 per median-priced home in the United States.3 Now, many housing regulations are necessary for safety and environmental reasons, but many unnecessarily raise the price of a home. Overzealous land-use planners also make homes more expensive and contribute to hollow increases in GDP in the process.
Savings Rate: The savings rate is perhaps the clearest example of a calculation that differs from the economic reality it is meant to describe. Simply put, the official U.S. savings rate is calculated by subtracting the amount people spend from their total disposable income. Families who plan for the future by investing money in 529 college accounts for their children and in 401(k) accounts for their retirement would be surprised to learn that the extra money they put into those accounts actually reduces their personal savings rate, as it is officially calculated.
The same goes for a mortgage. Most people do not think of the equity they accrue through their monthly mortgage payment in the same way they think of buying groceries or paying off their car. And they’re right, but the U.S. government does not see it that way. It is an odd measure indeed that regards money stuffed in a mattress as savings, but not home equity. Clearly, it is good to know whether people are spending more than they earn each month, but it is arguably more important to know the ratio of their debts to their assets.
Since the U.S. Department of Commerce began using it in 1947, the personal savings rate has perhaps been most useful as one datum among many in the National Income and Product Accounts to aid economists and analysts in their efforts to understand economic trends and formulate forecasts. But as a standalone indicator of the well-being of American households, it is misleading. Rather than providing answers to questions about how well Americans are preparing for the future, it is most helpful in the questions it raises: If the savings rate drops to zero or turns negative, how much of this is a result of increased household debt, and what are the implications for families’ abilities to fund their own retirements? How much are families investing each month, and does the resulting trade-off in lower savings levels matter in the long run? And so on.
Similar issues bedevil the savings rate globally, as well. The Organisation for Economic Cooperation and Development (OECD) compares the savings rates among its member nations but ultimately concludes that in the absence of a universal measure, the savings rate of countries will be different because of the variety of definitions they employ.4 Once again, this means that the measure is most useful for the questions it raises about household savings and investment patterns among nations. Analysts require multiple data sources beyond the savings rate to answer those questions.
Consumer Price Index: The CPI is used as a tool for understanding the cost of living. It measures the changes in the prices of goods that ordinary people buy year-by-year. The most common critique of the CPI is that it has trouble capturing the cost-of-living gains through innovation in products and entire economic sectors, or through changing consumer behavior. For instance, people may start eating more chicken than beef because of fears about mad cow disease or surprising improvements in the taste of chicken, but the CPI uses a fixed “basket of goods” approach for tracking prices each year that does not adjust to such changes in consumer behavior. If that basket were to include beef but not chicken, the CPI would record a drop in beef prices yet fail to measure the subsequent rise in chicken prices.
To further complicate matters, inflation is measured both in its “core” and “headline” instantiations. Core CPI assesses inflation rates without including food and energy prices, which are volatile and subject to shocks, while headline CPI includes them as a way to assess inflation in a manner that matters most to the average household. There is evidence that core CPI does a good job of predicting headline CPI, and yet shocks in, say, crude oil prices can create significantly higher headline CPI than core CPI. This complicates questions about which CPI version policymakers should use, when, and for which purposes.5
Beyond pure methodological challenges, innovations in computer technology, Internet purchases and other advances that create value by saving people time and money in other areas of their lives are difficult for the CPI to capture. It’s not that people are getting literally more for their money; it’s a case of their getting more from the way they spend their money. The CPI is powerless, as currently constructed, to measure this.
In addition, shifts in entire marketplaces affect the CPI. Wal-Mart, because of its ubiquity and well-oiled capacity for crushing competition, offers food at prices 10–25 percent lower than its largest competitors and enjoys a market share near 20 percent. Yet, the CPI methodology for pricing goods does not account for the behemoth food retailer, and as a result, it is probably overestimating food inflation by .32 to .42 percentage points, which translates into a significant upward bias in the inflation rate.6
There may also be, however, an offsetting downward bias. More than a decade ago, a U.S. government commission—the highly influential Boskin Commission, as it became known—estimated that the CPI overstated changes in the cost of living by 1.1 percentage points a year. This rather small-sounding figure becomes elephantine when considering that the cumulative additional national debt over a dozen years amounts to more than $1 trillion because of how the CPI is used in debt calculation. Because the CPI is also used in how inputs are included in national income accounts, an overstated CPI means that GDP is likely understated.7 The Boskin Commission has been criticized for basing its calculations on ad hoc reasoning more than empirical conclusions in a politically motivated effort to reduce Social Security payments, which are indexed to the CPI. The Commission was clearly under congressional pressure to cast its findings in the context of a larger public debate about entitlements. The most interesting aspect of the commission that critics discuss, though, is not its politicization but its effort to ascribe too much specificity to its most famous conclusion, namely the 1.1 percent upward inflation bias. For instance, the commission has been criticized for incorrectly estimating the economic effects of improvements in product quality and thereby missing the downward effect that such improvements would have on inflation.
The exact nature of biases in the CPI remains the subject of ongoing debate. Meanwhile, the commission’s conclusions about the upward inflation bias have proven influential globally as other nations adopt it in their own policymaking contexts.8 Decisions affecting billions, if not trillions, of dollars are based on a measure about which experts agree perhaps on one thing only: Consensus about how to calculate inflation in prices is inherently difficult, if not impossible, to achieve.
Poverty Rate: Most observers recognize that the official U.S. poverty rate is flawed, but we seem unable to agree on how to improve it. So we go on faithfully employing what we have in the spirit of the adage that, “If you cannot do something well, learn to enjoy doing it poorly.”
This really won’t do, however. Billions of taxpayer dollars are spent each year through the application of a measure that was calculated in the 1960s based on estimates of a subsistence diet. Increases in education levels since the 1960s, efficiencies gained through technology, cheaper household appliances—none of these factors are taken into calculating the poverty rate. Today, one can be officially poor in the United States and be overweight, have an iPod, an LCD television, air conditioning, cable TV and dozens of other amenities that the architects of the War on Poverty would have thought signalled victory.
Meanwhile, education levels and access to quality education, undeniably among the most significant determinants of a person’s future poverty or prosperity, play no role whatsoever in helping policymakers use taxpayer resources to “fight poverty.” The earnings gap between the educated and uneducated has reached levels not seen since 1915, when higher education was reserved for a small, elite population, but the U.S. government instead relies upon measures rooted in a 1960s understanding of how much food a family needs.9 The poverty rate is of little help in distinguishing between a retiree on a fixed income with a large retirement fund whose house is paid off and a single mother who never finished high school, lives in public housing and works for minimum wage. The economic reality of poverty and how we measure it are ridiculously unhinged from one another. How can we ever effectively ameliorate that which we seem determined to mismeasure and misunderstand?
Moving the Measures
If GDP’s main shortcoming is its tendency to include factors it should not count, such as non-productive or counterproductive activity, the savings rate misfires in its tendency to exclude factors it should count, such as equity or investments. The CPI tends to understate inflation in some ways for political reasons but to overstate it in others because it is too inflexible to account for dynamism in the marketplace, and the poverty rate miscalculates deprivation because it largely counts the wrong things. Economic measures both overstate and underestimate, and they always will. Yet they will always be with us, because we need them to make decisions about mass resources, whether taxpayer dollars or investments.
Given both the imperfection and inevitability of economic measures, we have two options. We can review the actual measures we use and try to refine them, or we can do a better job of aggregating essential data and information to supplement key measures in areas where they fail or are weak. The first will likely prove difficult. Despite widespread agreement that the poverty rate is inherently flawed, for instance, agreement on a replacement is nowhere in sight. And the CPI’s twin flaws create a result that is close enough for governmental work, as they say, so it will probably not be measurably improved. The best hope, therefore, lies in the second option.
In our current global economic landscape, there are two main areas in which improvements in economic measurement can be aided by aggregating data from different sources. The first is the latent assets—be they financial, human or social capital—that have been empirically shown to serve as catalysts for economic growth or buffers against economic hardship. Equity in homes, cited earlier, is an example of an asset missed by the savings rate that should be considered before conclusions are drawn about household savings. Equally important are non-economic sources of socioeconomic strength such as social trust and stable political institutions.
The Legatum Prosperity Index, to take one example, attempts to provide measures of “social well-being” that account for such important phenomena. Thus, at the beginning of the financial crisis in 2008, Slovakia, Bulgaria and Singapore were the fastest growing nations in the previous five years, according to key indicators of economic growth. But Slovakia and Bulgaria suffer from highly negative ratings for the quality of their social capital, freedom of choice and equality of opportunity. Singapore, on the other hand, scores slightly above average in all three. It also has high governance ratings compared to Slovakia’s somewhat lower and Bulgaria’s much lower ratings. Even before the global economic crisis began to take its toll on Eastern Europe, it was clear by using these comprehensive measures of socioeconomic strength that growth in Slovakia and Bulgaria was fragile compared to growth in Singapore. The Legatum Prosperity Index is one aggregator of important data, but there is still plenty of room for experimentation by the world’s leading governments and institutions to bring together useful indicators of overall socioeconomic well-being as a supplement to mainline economic data.
The second area for improvement in economic measures concerns risk. As we have learned in the current global crisis, failing to accurately understand risk can lead to massive problems. GDP can mask weakened demand. The poverty rate provides little help in understanding whether America is at risk of greater social unrest or requires greater social welfare spending, since it defines poverty in a way that is at odds with the nature of poverty. Most of our standard official measures in effect ignore risk altogether.
Investors and businessmen understand the role risk plays in decision-making and regularly retain the services of private intelligence aggregation firms such as Gerson Lehrman Group, Eurasia, Strafor, Aegis, Oxford Analytica and others to help them grasp political, economic and social risks that standard economic assessments miss. But risk factors should be more publicly accessible in commonly accepted economic measures. Transparency International, Global Integrity and the Heritage Foundation’s Index of Economic Freedom all attempt in different ways to account for risk in their country ratings. The Legatum Prosperity Index utilizes these and a variety of other indicators to establish governance risks inherent within countries to help make better sense of purely economic indicators, but much more remains to be done. There still exists no reliable single aggregator or index of risks that successfully adjusts economic measures accordingly, which is why companies currently pay significantly to fill the gap.
The temptation to doubt all economic indicators because of the current crisis must be resisted, and so must we resist the illusion that by providing new and better indicators we can solve all the problems before us. Measurements cannot change human nature, and it is human nature and behavior that cause economic dislocations as well as economic advance. Economic measures will always be imperfect, and they will always be needed. The task for scholars, policymakers, investors and foundations is to find new ways to account for latent assets and to better measure risk, so that we may live in a world whose measures more closely match its reality. People will still occasionally screw up, of course, but doing so will be ever so slightly more difficult.
1Ferguson, “Wall Street Lays Another Egg”, Vanity Fair (December 2008).
2Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, vol. 1, book 2, chapter 3 (University of Chicago Press, 1976), pp. 351–2.3Median home prices are based on the January 2007 figures published by the U.S. Census Bureau and the U.S. Department of Housing and Urban Development. Other figures cited are based on the estimate of a 1991 Federal advisory commission report, “Not In My Back Yard: Removing Barriers to Affordable Housing”, Advisory Commission on Regulatory Barriers to Affordable Housing (U.S. Department of Housing and Urban Development, 1991), p. 4. If anything, regulatory costs have risen as a percentage of sales prices since this report was published.
4Cedric Audenis, Stephane Gregoir and Claudie Louvot, “The Various Measures of the Saving Rate and their Interpretation”, Organisation for Economic Cooperation and Development (2002), p. 4.
5See Daniel Thornton, “Measure for Measure: Headline versus Core Inflation”, Monetary Trends, St. Louis Federal Reserve (2007).
6Jerry Hausman and Ephraim Leibtag, “CPI Bias from Supercenters: Does BLS Know that Wal-Mart Exists?” National Bureau of Economic Research Working Paper No. 10712 (August 2004).
7See Michael Boskin et al., “Consumer Prices, The Consumer Price Index, and the Cost of Living”, Journal of Economic Perspectives (Winter 1998).
8Jack E. Triplett, “The Boskin Commission Report After a Decade”, International Productivity Monitor, vol. 12 (2006), pp. 42–60.
9See, for instance, Claudia Goldin and Lawrence Katz, “The Race Between Technology and Education: The Evolution of U.S. Educational Differentials, 1890 to 2005”, National Bureau of Economic Research and Harvard University (2007).