I don’t think I was quite precise enough in my last article.

Econopundit posted a few remarks on it, and pointed out a helpful paper published by the Bureau of Labor Statistics discussing the different series and their methodologies and reasons why they produce different measurements. Econopundit summarized his analysis as follows:

Viewed just as a series, differences between the two surveys are actually vanishing on a long-term basis, with a possible interruption in this trend following the last recession. The current uptick is sharp, to be sure, but we've seen at least one downtick (roughly '64-70) apparently just as sharp, and similar but smaller upticks following other recessions.

In short: neither data nor method seem questionable. If the two surveys correlated perfectly it would be a waste of tax dollars to collect them both. These are two useful and perfectly valid methods of measuring the same thing. Neither is "right." Neither is "wrong." The complement and validate each other.

This is true, but ignores the fact that there is also a derivative data series which is politically at play, which is not so much the absolute number of jobs, as the change in employment year-over-year (or administration-over-administration). The manner in which the separation between the two surveys series has recently increased (with the two data series heading in opposite directions) is not irrelevant.

Mathematically speaking: the two main data series (employment in the households survey, and employment in the payrolls survey) over time are themselves closely positively correlated with each other. A look at the graph which Econopundit posted over the weekend (and which I referred to yesterday) suggests something like a +0.8 correlation (just estimating by eye), which of course is quite high. The separation between them has sometimes increased, and sometimes decreased, but the two series have generally moved in the same direction and are highly positively correlated.

But recently the separation between them is increasing because one derivative series (year-over-year change in employment measured by the households survey) is positive, while the other derivative series (year-over-year change in the payrolls survey) is negative. This means that for the politically sensitive derivative measure of year-over-year change in employment, there is negative correlation between the two series. And although it is true that there have been sharp upticks in the separation between the other two series before, it is not quite accurate to say that those previous periods (eg, 1980-1985) are similar: while the separation between the surveys was increasing then, both series were increasing so the derivative series remained positively correlated. I maintain that the recent four-year period 2000-2003 has been unique due to the negative correlation of the two derivative data series, and that this shows some new and fundamental developments in the way work gets done and people get hired.

I actually agree with Econopundit’s last paragraph entirely. My main point in all this is not that the payrolls survey is wrong or should be ignored, but that the overall employment situation has evolved in some subtle ways not entirely or even closely summarized by merely quoting only the payrolls survey. It continues to be a disappointment that the drop in employment from the payrolls survey is all one hears about in the mainstream press—which, after all, is where the main body of voters will get their information when assessing George Bush’s performance in office. The households survey suggests other conclusions than the payrolls survey, and both need to be considered, which politicians don’t seem to be doing to any real extent.

I've spent a bit of time looking around for someone else commenting on this business of net job creation during the Bush presidency, and I haven't found much of it. I have occasionally found a mention that acknowledges the increase in the employment figures reported by the household survey, but generally there is a pooh-poohing that it's not really proper to say that one data series is better than the other. This, generally, is true; as I pointed out in the last article, it's wrong to ignore one in favor of the other when one needs to consider both of them to gain a comprehensive perspective. But there's something a bit unusual going on in the economy right now that is illustrated only by the household survey of employment.

There is a neat graph in an article posted over the weekend over at Econopundit, which shows a graph over time of three data series: the payrolls survey; the household survey; and a Yale model which is a predictor of total employment. All three give different numbers, which isn't really the point. Econopundit picks up on one of the key points shown on the graph, but misses on another.

First, as Econopundit points out, the Yale model gives different absolute numbers than measured by either survey series, but as his graph shows, it does track very closely and has a highly positive correlation with actual employment. When the model says employment should go up or down, it generally does, and generally in proportion to the quantity predicted. And right now, under current tax policy, the Yale model predicts 10 million new jobs over the next four years. It probably isn't a coincidence that John Kerry has positioned himself to take credit for this development, by promising minor tinkering with the tax code which, he says, will create--ahem--10 million new jobs over the next four years.

The other point which I think is key is that, as Econopundit's graph shows, while the spread between the payroll and household surveys has evolved over time, nowhere else on the graph but presently do we see the two data series moving in completely opposite directions for a four-year period such as 2000-2003. The mere fact that such an unprecedented divergence of the two series has been going on for so long demands our attention, and points to some highly unusual employment market developments.

Employers right now, buoyed by consistent productivity gains, simply don't want to hire new people even though business has picked up. Overtime and temporary labor and outsourcing to independent contractors has supplied the difference. This has been good for the companies in question, who retain far greater flexibility in staffing and manpower levels; it has also been good for the independent contractors themselves, who now find themselves working for wages and profits, not just wages. It's a rather new trend, which deserves much more consideration and discussion than it has generally received. Demagoguing over jobs lost when measured by just one data series does not serve the country or the economy very well.

Since the election of the Leader of the Free World will evidently hinge primarily on whether the nation has more or fewer jobs come November 2004 than it had under Bill Clinton, a close examination of the surrounding facts is in order.

Economic analysis isn't the easiest thing in the world, and it requires one to pay attention for an extended period to some subtle concepts, which much of the American public isn't necessarily good at. But in this case the stakes are sufficiently high in the coming election that we should try really hard to have an informed national debate on the subject. Soundbyte politics is easy; substance is hard. This is more technical than most of what I post, but for serious discussion of the facts this simply can't be avoided. I will be happy to elaborate on anything which I've not sufficiently explained; just leave a comment or send an email. This is my opening contribution to the serious national debate about economics which we truly need to have.

The Democrats will have us believe that the economy is really bad right now. The Republicans point to fantastic GDP growth, low interest rates, and low inflation and say that things are pretty swell. The Democrats counter that for those out of work, low prices and high GDP don't matter, and that there are too many of this sort out there. And that the economy is really, really bad unless you are a fat cat who works for Halliburton.

It's apparently all about jobs. Jobs, jobs, jobs. I keep hearing ceaselessly about jobs, and in particular the absolute number of jobs gained or lost under George Bush, as if that single number represents a competent referendum on his performance as President. The obsession with this particular data series strikes me as being a new and somewhat unusual target of economic fixation; during the "it's the economy, stupid" run-up to the 1992 election, for example, it was the unemployment rate that was the important number. During the 1970s, truly a dark economic time in America, the misery index, gained by adding the unemployment rate to the inflation rate, was frequently the main subject of complaint.

Now it's the absolute number of jobs gained or lost, presumably because the opposition party doesn't really think that complaints based on the unemployment rate or the misery index will stick. And they're probably right: the unemployment rate in January 2004 was 5.6% (compared to 7.3% at the time of the 1992 election), which is pretty close to the 5.0% unemployment rate which is traditionally considered "full employment." The misery index in January 2004 was a mere 7.5% (5.6% for unemployment, plus the 1.9% increase in CPI from January 2003 to January 2004; this figure compares to the 20.1% misery index in October 1980, at the end of Jimmy Carter's Reign of Terror). So, using the various historical shorthand forms, present conditions look reasonably good despite John Kerry's rather fatuous bellowing of the worst economic performance since Herbert Hoover.

Now this business of absolute numbers of jobs is a shifty target. We have been hearing for months of the mythical "2.2 million jobs lost under This President," which is presumably the numerical basis for the Herbert Hoover comparison (though naturally this disregards the fantastically larger overlying civilian workforce in 2001-2004, so that 2.2 million jobs lost under Hoover would have represented a far larger portion of the workforce suddenly out of work). At the same time we are told of a near-apocalyptic number of jobs being shipped overseas, which complaint we've heard over and over at least since NAFTA and presumably before. It's nothing new, but it's tempting to draw a correlation between the mythical 2.2 million jobs lost and this international outsourcing and blame the latter, for which protectionist policies are the natural solution.

So, before we go looking for the culprits of the problem, let's first examine what the numbers themselves specifically are and determine whether the mythical 2.2 million jobs lost is mythical after all, or real. Now would be a good time for you to run off to get the actual numbers in front of you, so we can all argue from the facts. Virtually all the numbers we'll use from here out are conveniently summarized in a document called Economic Indicators January 2004. This document is published monthly by the Government Printing Office and is a compilation of economic statistics gathered into one handy spot from various government sources.

Job growth (as reported by the Department of Labor's payroll survey) has, in fact, lately been disappointing. This is the data series Kerry has focused on, and it's here that he gets his claim of 2.2 million jobs lost. The number series, in summary, looks like this: (this is the table at the bottom of page 14 in the Economic Indicators. Job counts are given in thousands of jobs, and are yearlong averages.)

Table 1.
        Nonagricultural
Year      Employment
2000        131,785
2001        131,826
2002        130,341
2003        129,932

So this data series shows a loss of roughly 1.85 million jobs between the average for 2000 and the average for 2003. What if we look at the actual yearend number for 2000, just days before Bush took office (instead of the yearlong average) and compare it to the actual latest month, January 2004?

Table 2.
            Nonagricultural
Year          Employment
Dec 2000        131,953
Jan 2004        130,155

Still about 1.8 million jobs lost according to this figuring of the data. But the lowest monthly figure for 2003 (according to Economic Indicators January 2004) was 129,789 in August; this number actually reflects a decrease in the number of jobs of 2.16 million, compared to the December 2000 figure above, so at least we can confirm Kerry's 2.2 million jobs claim and we know how his numbers were derived. And even according to the data series Kerry evidently is referring to, it now would be more correct to speak of a decrease of 1.8 million jobs, not 2.2 million, which still seems to be the number I keep hearing.

Such monolithic reliance on a single data series is something that the astute listener should regard with some distrust. There is no such thing as a perfectly designed or totally comprehensive data series, and this one (Nonagricultural Employment) is no exception. The important thing is that whatever each series' shortcomings are, they are repeated each month and their values are therefore comparable over time. But the Department of Labor collects other data series as well, such as the Unemployment Rate, to provide collectively a more comprehensive treatment of our employment situation. The other data series we will consider in detail is Status of the Labor Force, and this is presented on page 11 of Economic Indicators January 2004.

The same time period considered above, 2000-2003, looks rather different in the Status of the Labor Force report, summarized as follows:

Table 3.
        Total Civilian
Year      Employment
2000        136,891
2001        136,933
2002        136,485
2003        137,736

This series shows, rather surprisingly, an increase in total civilian employment of 845,000 jobs from 2000 to 2003. Where does the difference between the two series come from?

It helps to know a bit about the methodology of the two series. The Nonagricultural Employment report comes from an employers' survey. Companies with employees and payrolls are the only contributors to this number. The Status of the Labor Force report comes from a households survey, where respondants are asked to characterize themselves as "working" or "not working." So what's the difference?

Footnote 1 on page 14 of our January 2004 report spells out what John Kerry's data series doesn't consider:

Excludes proprietors, self-employed persons, unpaid family workers, and private household workers. Data from the household survey [Status of the Labor Force] shown on p. 11 include those workers and also count persons as employed when they are not at work because of industrial disputes, bad weather, etc., even if they are not paid for the time off. In the series shown here [Nonagricultural Employment], persons who work at more than one job are counted each time they appear on a payroll, in contrast to the series shown on p. 11 where persons are counted only once--as employed, unemployed, or not in the labor force.

So anyone who quits, or is laid off, and then finds identical or other work as a self-employed independent contractor is counted as unemployed in John Kerry's data, even if they are in fact working and even if they are earning more than previously since they are working for profits and not wages. And, according to the household survey, more people themselves say they are working now than in Bill Clinton's last year in office.

So where have all those 2.2 million jobs (really 1.8 million) gone? Nowhere. They're mainly just working for themselves as proprietors now instead of for a corporate employer. It's just a different type of employment arrangement. It's worth stating again, since you hear so little of it:

More people are working now than were before George Bush took office. We have numbers to prove it.

Yes, yes, I hear John Kerry harrumphing, but those people now are uninsured and don't have unemployment insurance, and aren't fully protected by Workers' Compensation, etc. Maybe so. But sole proprietors can pay for unemployment insurance for themselves if they choose to pay for it (at least in Ohio, the laws of which I'm familiar with). They should pay into the Workers' Compensation fund for themselves, for their own protection, though they aren't legally required to. And as a small business owner myself, I know that it is possible to buy health insurance for a small company, though it can be expensive. It's a riskier way to make a living than by showing up for a wage-paying position with a company, but the returns tend to have more upside potential since the proprietor keeps all the profits (after tax, of course) and these tend to be higher than the wage value of a job.

How does all this play out in balance? According to the Department of Commerce's National Income report (top of page 4 in Economic Indicators January 2004), two components of national income--wages and nonfarm proprietors' income--summarize as follows: (all figures are in Billions of nominal dollars)

Table 4.
                      Nonfarm
                    Proprietor's
Year     Wages        Income         Total Income
2000    $5,782.7      $705.7           $6,488.4
2001     5,940.4       745.6            6,686.0
2002     6,019.1       783.4            6,802.5
2003     6,185.6       827.2            7,012.8

So wages are up, proprietors' income is up, and obviously total income earned by wage earners and proprietors combined (the group identified as "working" in the Status of the Labor Force report) is up during Bush's term. This is in nominal dollars, however, so we need to adjust for inflation, and report everything in "2000 dollars," which is simply an acknowledgement that the same dollar earned today has less purchasing power than one earned in a prior year. The baseline year can be any that is convenient, and we will adjust our 2001-2003 figures downward to use values equal to the value of a dollar in the year 2000. There are a large number of inflation measures available, reaching their results in different ways and giving (not incidentally) somewhat different answers.

It is my opinion that, as a measure of inflation, the Implicit GDP Deflator is much superior to the Consumer Price Index (though I'll give results according to both so it won't be supposed that I'm cherry picking only the favorable measures). The Personal Consumption GDP Deflator, applied to the income figures above, is as follows: (Nominal income is in Billions of current-year dollars; real income is in Billions of 2000 dollars.)

Table 5.
        Nominal      GDP         Real
Year    Income     Deflator     Income
2000   $6,488.4    100.000     $6,488.4
2001    6,686.0    102.038      6,552.5
2002    6,802.5    103.429      6,577.0
2003    7,012.8    105.298      6,660.0

Notes to Table 5:
1. Nominal Income is as calculated in Table 4 of this document.
2. GDP deflator figures come from the table Implicit Price Deflators for Gross Domestic Product on page 2 of Economic Indicators January 2004; the column used is for Total Personal Consumption Expenditures.
3. Real income is calculated as Nominal income divided by the GDP Deflator x 100.
4. The CPI index figures, where the years 1982-84=100.0, are as follows for the above period: 2000=172.2; 2001=177.1; 2002=179.9; 2003=184.0. Refer to the Bureau of Labor Statistics CPI reports, check the first box for US All Items, 1982-84=100 and press Retrieve Data.
5. Real income for the above years, in 2000 dollars deflated by CPI instead of the GDP deflator, is: 2000=$6,488.4; 2001=$6,501.0; 2002=$6,511.3; 2003=$6,563.1.

So even adjusted for inflation we are earning more income among our workers and proprietors now than before George Bush took office. The last question is to consider whether, when this increased income is divided among more people who report that they are working, all workers collectively are better or worse off now compared to 2000. (Income figures are in Billions of 2000 dollars; Civilian employment is in thousands of workers.)

Table 6.
         Real       Civilian       Real Income per
Year    Income     Employment      Employed Person
2000   $6,488.4     136,891            $47,398
2001    6,552.5     136,933             47,852
2002    6,577.0     136,485             48,188
2003    6,660.0     137,736             48,353

Notes to Table 6:
1. Real income is as calculated in Table 5 of this document.
2. Civilian employment is from Table 3 of this document.
3. Real income per employed person is real income divided by the number in total civilian employment. Note that, as Real Income (RI) is in $Billions and Civilian Employment (CE) is in thousands, it is necessary to take (RI/CE) x 1000000 to find Real Income per Employed Person.
4. For those interested in deflating nominal income by CPI, Real Income per Employed Person would be as follows: 2000=$47,398; 2001=$47,476; 2002=$47,707; 2003=$47,650.

You read it here first: more people are now working, and on average each of them is now earning more real income than before Bush took office.

I think it should be obvious that our true economic situation is far too complex for John Kerry's soundbytes about 2.2 million jobs lost and the worst performance against that measure since Herbert Hoover. I further posit that a close analysis of the numbers reveals that our economy now is, at minimum, no worse than before Bush took office--though I'd want to see the unemployment rate drop from 5.6% to the statistically "full employment" level of 5.0% before I suggested that everything was just swimmingly grand. But when John Kerry suggests unusual, European-sounding measures to halt a putative loss of jobs, an astute listener should be aware that the employment situation is by no means so simple or so bleak as Kerry claims.