Data Manipulation as Crisis

Dr. Pulsinelli is Professor of Economics at Western Kentucky University, Bowling Green. He is the co-author (with Roger Leroy Miller) of Understanding Economics (West Publishing Co.).

By now everyone should know that many politicians and others with a liberal-socialistic bent are fond of discovering crises. A crisis provides an excuse for drastic and immediate action—by the government. Often misinterpretations of economic data (wittingly or not) have been used to transform problems into crises and to expand the role of government.

To understand the “crisis” mentality, consider the following illustration. Suppose you visit your local grocery store and note that there are 300 cans of soup on the shelf. You return one month later and observe that there are still 300 such cans. Can you safely conclude that this store has a “crisis” because it isn’t selling any soup? A moment’s reflection tells you that such a conclusion would be a non sequitur because the present cans may not be the same cans that were there last month. Before drawing any conclusions, you would need more information.

A simple stock-flow model indicates that if the store sells 100 cans of soup every month and replaces those 100 cans monthly, the stock will remain at 300 cans. Clearly, if the inflow of cans exceeds the outflow (sales), then the stock (inventory) rises; if the inflow is less than the outflow, then the stock falls.

While this model is so obvious that it seems trivial, an understanding of this stock-flow model can help put some recently labeled “crises” in perspective, particularly unemployment, health care insurance, and the distribution of income.


Assume that the size of the labor force (the number employed plus the number unemployed) is 100 and assume that every month a survey is taken. The January survey indicates that four people are unemployed; hence, the unemployment rate for January is 4 percent. In February, suppose those four people find jobs, but four others become unemployed; the unemployment rate is again 4 percent. Assume further that every month something similar happened: the four who were previously unemployed find jobs but are replaced by four others who have become unemployed. The unemployment rate remains at 4 percent all year long.

If one merely observed the unemployment rate, one might conclude that a crisis existed in the economy because the unemployment rate remained at 4 percent. Indeed, a problem might exist if the same people were unemployed each month, all year round. It seems to me that before we can talk about an unemployment problem, we need to know the number of heads of households who have been unemployed for longer than one year. Clearly, if the unemployment rate were 6 percent and that number consisted only of heads of households who have been unemployed for three years, a serious problem might well exist. But that is hardly the problem in the United States.

In recent years, it is not unlikely that in any given month millions of people will become unemployed (as job losers, job leavers, re-entrants, and new entrants into the labor force), and millions will find jobs (or leave the labor force). Or, looked at in another way, of the 7.764 million people who were unemployed in December of 1993, 2.764 million had been unemployed for fewer than 5 weeks and only 925 thousand had been unemployed for 52 weeks or longer. The point is that before one can talk about an unemployment problem (much less a crisis) it is important to know how many chronically unemployed there are.

Understanding the simple stock-flow model helps to clear up another issue. Because the stock of unemployed is positive at any given moment, most people came to believe that full employment is never attained. Hence, John Maynard Keynes’ contention that capitalism is associated with chronic unemployment—with the attendant implication that government must create jobs—seems to be vindicated month after month. The fact is that although it is possible for a surplus of labor (a shortage of jobs) to exist for some jobs (because wage rates are set by unions or governments above market-clearing levels), it is extremely unlikely if not impossible, for a general shortage of jobs to exist.

Health Insurance Coverage

It is widely reported that 37 million people in the United States do not have health insurance coverage, and this statistic is said to reveal a crisis that justifies the Clinton health care plan (read socialized medicine). Even if we accept that (dubious) figure of 37 million, it is not indicative of a crisis.

Every month, there are inflows into the stock of uninsured (people who have just lost their jobs, who have just graduated from high school or college and are no longer covered by their parents’ family plan, and so on), and, there are outflows from that stock (people who accept jobs, who became older and decide that it is now worthwhile to par-chase health insurance, who die, and so on).

The crucial information here is the number of people who are chronically uninsured and who want to purchase insurance but truly cannot afford to do so. If this number were large, it might make sense to look for radical solutions; the number of chronically uninsured Americans is probably between 2 and 3 million.

Perhaps because 37 million people did not indicate a sufficient crisis, President Clinton decided to prolong the flow period and reported that 52 million people were without insurance at one time or another during a one-year period.

Distribution of Income

Income inequality, probably more than any other “problem,” has been used to justify government encroachment on private property and liberty. Data indicate (with some variability) that over very long periods in the United States, the lowest 20 percent of the distribution has received 5 percent of total income, and the upper 5 percent has received about 20 percent of total income. Let’s skirt the issue as to whether such inequality is “fair” or “unfair” and merely note that such studies usually measure pre-tax income. By excluding income-in-kind and government transfers (food stamps, rent subsidies, Medicaid, etc.) they overstate inequality.

What is germane here, however, is the stock-flow model implications. Note that the same families do not continue to occupy the same positions in the income distribution; intergenerational social mobility occurs. Furthermore, such data overstate intra-generational income inequalities as well.

Age/Earnings Profiles. An abundance of evidence indicates that most people reach relatively low incomes in their youth, relatively high incomes in their middle years, and relatively low incomes near and after retirement. Consider now a fictitious economy in which the only determinant of income is age; all 17-year-olds earn $5,000 per year, all 25 year-olds earn $12,000 per year, and so on.

Assume that there are people of all different ages in this economy. Data taken at any given time (cross-section data) will indicate a considerable amount of income inequality in this (conjured) economy, whereas zero lifetime inequality exists. In short, income distribution studies that don’t explicitly adjust for age are biased toward inequality because an age/earnings profile exists for most people, and over a lifetime there will be inflows and outflows through each age (and therefore income) bracket.

Transitory Income Effects.

Consider an economy consisting of clones, in which everyone has the same skills and earns the same income. Income differences result only from temporary phenomena (illness, luck, temporary periods of work strike, temporary industry-specific recessions and expansions, and so on). Using Milton Friedman’s terminology, everyone has the same permanent income; measured income differences result only from transitory income differences. In any given year, some people will be pushed below their permanent income levels and others will be pushed above theirs; hence, a certain amount of income inequality exists in’ a given period even if people have identical lifetime incomes. We conclude that unless income distribution studies consider only permanent (or lifetime) income, they will be biased toward measuring inequality; studies analyzing measured income overstate income inequality. Stated in another way, because of the vicissitudes of life, there will be continuous inflows and outflows through each income bracket. But because there will be (a lot of) different people in each bracket, income inequality will be overstated.


Demagogues will always use data for their own purposes. The rest of us need to be aware of how data can be misused and abused. Unfortunately, it is a lot easier (and the rewards are higher) to emote and shout “crisis” or “unfair” than it is to analyze data and put them into perspective for the uninitiated. Perhaps that is one reason the story of the twentieth century is largely about “crises” that induce people to hand power to governments who keep that power long after the alleged crisis has disappeared.

Related Articles


{{}} - {{relArticle.pub_date | date : 'MMMM dd, yyyy'}} {{}} - {{relArticle.pub_date | date : 'MMMM dd, yyyy'}}
{{article.Topic.Topic}} {{article.Topic.Topic}}