A Real Issue with Data – How to Measure the GDP or is it GDI

From the Wall Street Journal:

. . . . To be precise, they (the Fed) wonder if the U.S. economy grew 3.1% in the fourth quarter last year from a year earlier, as the government's gross-domestic-product data suggest, or a more robust 4.1%, as measured by the lesser-known gross domestic income.

It isn't an idle exercise. If the stronger number is right, it would help explain why unemployment has remained so low. It would also suggest that productivity growth has remained relatively robust rather than slowing sharply, as the regular data imply. Strong productivity growth means the economy can grow faster without pushing up inflation.

Unfortunately, a closer examination of the data suggests the stronger number is misleading. The reason: Employees are reaping big gains on stock options, but the cost of those options probably isn't fully reflected in corporate profits.

The Commerce Department's Bureau of Economic Analysis measures the nation's economic output two ways: total spending, or GDP, and total income, or GDI. In theory the two should be equal, but in practice they seldom are, because they are drawn from different sources and are subject to differing sampling and measurement challenges.

Occasionally, the divergence signals something important. In the 1990s, former Fed Chairman Alan Greenspan posited that the fact GDI was growing faster than GDP was a clue that productivity, or output per worker, was growing faster than official data suggested. People were producing and earning more, he figured, but GDP wasn't capturing it. More recently, a study by Fed staff economist Jeremy Nalewaik argues that GDI is better than GDP at identifying recessions, and "an increased focus on GDI may be useful in assessing the current state of the economy.". . . .

. . . . Because measuring both GDP and GDI is so complex, it is difficult to pinpoint the source of the discrepancy, and it may never be resolved. One possible explanation is that GDP has been underestimated and will be revised higher; that would help explain why the U.S. job market is so strong. But a more likely explanation is that GDI has been overestimated, and may be revised lower.

Here's why. Stock options are a growing and highly volatile part of employee compensation. Suppose an employee has an option to buy his company's stock at $20, and the stock is trading at $30. If the employee exercises that option and buys the stock, the $10 profit -- even if the stock isn't sold -- is treated as income both by the Internal Revenue Service and the BEA, and as an expense to the company.

Option income, however, is hard to measure. It's not included in the Bureau of Labor Statistics' monthly wage figures. The BEA has to estimate it, and it may undercount it when a rising stock market increases the profits from exercised options.

Four months after the end of each quarter, the BEA does get more comprehensive labor-income figures that include options. A year ago, that led to a big upward revisions in first-quarter income; last week, it led to a similar upward revision in fourth-quarter income.

The data don't tell the BEA where the extra income came from; the evidence suggests options: Figures from Thomson Financial suggest those were the first and fifth most lucrative quarters ever for corporate officers, directors and other "insiders" cashing in options.

In theory, that added option income should have reduced corporate profits by an equivalent amount. But quarterly shareholder reports, an important source for the BEA's quarterly profit calculations, deduct far less than the entire value of an exercised option as an expense. The BEA gets its first look at that value a year later when companies file their annual reports, and it takes up to another year for the final figures to come from the IRS. There's a good chance that later this year, the BEA will revise down last year's profit data, shrinking GDI. Until that happens, the additional options income is, in effect, double-counted: in employee income, and in artificially high profits.

There is a good chance that later this year, the BEA will revise down last year's profit data, shrinking GDI. Until that happens, options income is, in effect, double-counted: in employee income, and in artificially high profits.

A BEA official concedes such double-counting is a "possibility," while adding, "We've learned from experience it's difficult to understand all of the factors that affect corporate profits."