The U.S. Productivity Miracle (Made in China)

The U.S. "productivity miracle" is almost entirely dependent on growth in U.S. imports and foreign labor outsourcing.

John P. Hussman, Ph.D.
All rights reserved and actively enforced.

Market Climate

The Market Climate for stocks remains characterized by unusually unfavorable valuations and modestly favorable market action. Insider selling has eclipsed even the unusually high levels of recent weeks. The CBOE volatility index near 16% meanwhile suggests substantial complacency. As I've noted before, low VIX levels are not in themselves negative, but they do suggest that risk aversion has reached very low levels. When markets are richly priced, the greatest risk is an abrupt increase in risk aversion. We don't see evidence of that yet, but the possibility that such a shift could be abrupt already places us in a moderately defensive position in stocks, with about 50% of our stock holdings hedged against the impact of market fluctuations. The Market Climate for bonds remains characterized by modestly favorable valuations and modestly favorable market action.

The Hussman Funds paid their annual capital gains distributions last week. Details are on the Fund News section of our Home page.

Long-term productivity growth is not a large number

One of the most deeply held beliefs of investors is the notion that the U.S. has enjoyed a "productivity miracle" in recent years. Computers, the rise of the internet, and the spread of new inventory management techniques are credited for this revolution.

Two questions are worth asking. How much of a boom has U.S. measured productivity actually enjoyed? and, 2) How much of that measured productivity growth is real?

The first question is fairly simple to answer. Since 1947, productivity measured by output per worker has grown at a healthy rate of 2.0% annually. This overall growth rate masks a certain amount of variation. From 1950 to 1980, U.S. productivity grew by 2.1% annually. During the 1980's productivity growth slowed considerably, to a growth rate of just 1.3% annually. From 1990 through the third quarter of 2003, U.S. productivity has grown at a 2.2% annual rate, a full two-tenths of one percent over its long-term average.

Wait. Didn't productivity grow by 8.1% last quarter alone? Well, yes and no. That 8.1% is quarterly growth at an annualized rate. The actual increase of just over 2% was significant, of course, but was the combined result of an enormous program of tax rebates, a blowoff in mortgage refinancings prompted by a brief plunge in long-term interest rates on deflation fears, and a contraction in the U.S. labor market. Since the economy enjoyed higher output with fewer workers, the quarterly productivity figure was extremely strong. Even occasionally large quarterly variations do very little to change the long-term average of about 2% annually.

To an economist, even changes of a fraction of 1% in productivity growth have enormous welfare implications over the long-term. A change from 2% U.S. productivity growth to 2.5% over the long-term would be tremendously welcome, but almost unthinkably optimistic. Unfortunately, the way that productivity growth is often discussed, investors appear to believe that long-term productivity growth in the U.S. has accelerated by several percent, with similar implications for long-term earnings growth. That simply isn't the case.

So even though the implications of the recent "productivity miracle" for long-term earnings growth are marginal at best, 2.2% growth since 1990 is not bad from the standpoint of long-term economic health and social welfare. That is, assuming that this acceleration in measured productivity is both real and sustainable.

Is the increase in productivity growth real?

Productivity is measured as output per worker. Essentially, productivity is the total output of the U.S. (GDP) divided by the number of workers required to produce that output. Increases in output, or decreases in the required labor, both increase measured productivity.

Economists know that increasing the amount of capital that workers have at their disposal ("capital deepening") tends to increase labor productivity. So it is not surprising to see productivity increase as a result of the capital spending boom of the 1990's. The difficulty here is that much of that capital spending boom relied on massive inflows of foreign savings. The unpleasant arithmetic of the "savings investment identity" ensures that foreign savings have to bridge any gap between the amount the U.S. invests (factories, housing, capital spending) and the amount it saves. The amount of savings we import from foreigners is measured by our "current account deficit," which is presently the deepest in history.

Wait. Doesn't the current account deficit essentially measure how much our imports exceed our exports? Yes, but here's how that works. If we import $100 worth of stuff, we have to pay for it by exporting $100 worth of stuff. Suppose we import $100 of goods and services, but only export $70 of goods and services. In that case, we've got to export $30 of something else to foreigners, and that something is U.S. securities. So that $30 gap measures both our current account deficit, as well as the amount of savings we import from foreigners (through the sale of securities) in order to finance our economic activity.

In short, if the U.S. savings rate is very low, it is difficult to run a sustained investment boom unless we import savings from foreigners, which essentially means running the current account to negative levels. Not surprisingly, every sustained economic expansion in U.S. history has begun with a current account surplus, which quickly moved to a deficit as investment boomed. With the U.S. current account already at the deepest deficit in U.S. history, the flexibility to run even deeper deficits is limited. As a result, our ability to run a sustained investment boom (with associated productivity growth) is also limited.

As I've noted before, I expect that any surge in U.S. capital spending will likely be financed through a reduction in U.S. housing investment. Given that the mortgage refinancing index peaked several months ago, it is likely that the housing market is in the process of peaking as well. In other words, we may very well see increases in some areas of U.S. investment, but they will be accompanied by decreases in other areas. As a result, we're likely to see relatively flat growth in gross domestic investment, rather than the broad increase in gross investment that typically emerges in strong and sustained economic expansions.

The investment boom of the past decade does help to explain the faster U.S. productivity growth since 1990, but again, the actual acceleration is only a few tenths of a percent above the long-term 2% trend for U.S. productivity growth, and the nation's deep current account deficit puts the sustainability of capital spending growth in question.

Lies, damned lies, and statistics

Unfortunately, the "productivity miracle" is also partly a statistical quirk. A significant portion of what we "import" from foreign countries actually represents intermediate goods produced by foreign subsidiaries of U.S. companies, or companies formed by direct U.S. investment into those countries. Think about this for a second. We import the same intermediate goods we would have manufactured at home, but at cheaper prices (meaning that the deduction to GDP on account of imports will generally be proportionately smaller than the corresponding loss in U.S. employment). Yet when we turn around to calculate productivity, we don't count the foreign jobs used to produce that output. As Fabrizio Galimberti has noted in the Economist, foreign outsourcing has the effect of artificially raising productivity figures because subtracting imports from GDP does not adequately correct for their impact on final output, yet foreign labor is not counted, so measured output per worker increases.

The net effect of all this is that the U.S. "productivity miracle" is almost entirely dependent on growth in U.S. imports and foreign labor outsourcing.

I had a lot of fun putting the following chart together (any use of this should include "Source: Hussman Funds", thanks). Since productivity is cyclical, I used the 5-year growth rate of non-farm productivity, which smooths out short-term fluctuations. To get at the idea of import growth over and above the growth in overall consumption, the other line is the 5-year U.S. import growth rate over and above the growth rate of personal consumption expenditures.

Very simply, import growth captures both the "true" part of productivity growth (since increased capital investment typically requires an expanding current account deficit) as well as the illusory part of productivity growth (resulting from the failure to account for foreign labor input in the productivity numbers). In both cases, it is misplaced optimism to expect rapid and sustained growth in U.S. productivity when the U.S. current account is already at a record deficit.

No Lucy, you can't do the show

One of the most famous arguments for free trade was developed over a century ago by David Ricardo. Right. David. I always get that wrong. Ricky was the one on "I Love Lucy."

Consider two countries. In TVland, an hour of work can make 3 TVs or 6 loaves of bread. In Doughnia, an hour of work can make just 1 TV or just 4 loaves of bread. Clearly, TVland has "absolute advantage" in both TVs and bread. But notice that by giving up 1 TV, TVland can produce only 2 loaves of bread, while Doughnia can give up 1 TV and produce 4 loaves of bread. So though Doughnia has no absolute advantages, it does have a "comparative advantage" in making bread.

Ricardo's solution was to encourage each country to specialize their production to those goods in which they had comparative advantage. So TVland focuses only on TVs, producing 3 for each hour of work. Meanwhile, Doughnia focuses only on bread, and produces 4 loaves.

Now here's the neat part. As long as the "terms of trade" are somewhere between what the countries could produce without trade (i.e. between TVland's 1 TV for 2 loaves and Doughnia's 1 TV for 4 loaves), everybody benefits from trade.

So suppose TVland trades 1 of its TVs for 3 of Doughnia's loaves. After this trade, TVland has 2 TVs and 3 loaves of bread, which is better than it could have done by producing both. Meanwhile, Doughnia has 1 TV and 1 loaf of bread, which is also better than it could have done on its own.

That's the standard argument for free trade. It does, however, assume a few things. First, notice that all the bread bakers in TVland have become unemployed, and now have to learn how to weld transistors. Meanwhile, Doughnian tech workers now have to learn baking. In the real world, these dislocations are painful. While there are various theorems in economics that assure that it is always possible to compensate the losers from free trade in a way that the nation still benefits, in practice, that compensation often does not occur.

Messing with the terms of trade

Let's take this to the U.S. economy. There is no question that the U.S., and especially its consumers, benefit from the availability of inexpensive imports. But there is also a cost in terms of displaced manufacturing jobs. This is not an argument against free trade, but it is an argument against taking these displacements glibly or without other policies to address them.

A second issue is the terms of trade. One of the arguments against free trade is that it will tend to drive U.S. manufacturing wages down to the impoverished levels of China, Mexico and India. Clearly, this overlooks the distinction between comparative advantage and absolute advantage. Free trade benefits both countries even if one country is at a profound disadvantage in absolute terms. There is no reason for wages to converge if the absolute productivity of workers differs across countries.

That said, it is quite possible to skew the terms of trade so that U.S. workers lose jobs and yet foreign workers enjoy little benefit. This is particularly true in the case of China. Part of the flap over the Chinese yuan is that it is currently undervalued (while I think a gradual adjustment would be useful, an abrupt adjustment would be near-catastrophic). A cheap yuan encourages companies to substitute U.S. labor with foreign labor.

Unfortunately, the often repressive conditions in China also make it possible that Chinese workers get very little of the benefit from free trade. So while free trade has clear gains for U.S. consumers, and also benefits companies that shift jobs overseas, there is a strong possibility that workers in both countries receive an inequitable share of these gains.

Labor Ethics 101

Case in point is a recent proxy we received from Sun Microsystems. Among the items was a strikingly modest proposal regarding rights of workers in China. This is an interesting issue, of course, since the U.S. has an enormous trade deficit with China, and has lost a great deal of manufacturing jobs there. While trade in itself has the potential to benefit both countries - to the U.S. through greater access of consumers to inexpensive imports and to China through employment for their impoverished people - the actual conduct of American companies in China has a great deal to do with whether these benefits actually arise. There is a difference between inexpensive labor and exploitation, and the distinction is not particularly subtle.

The elements of the proposal in Sun's proxy are summarized below:

-------------------------------------------------------------------------------------------------------

•  No goods produced by the company or its suppliers shall be produced by bonded labor, forced labor or within prison camps.

•  Wage and hour guidelines should adhere at least to those provided by China's national labor laws.

•  Facilities and suppliers shall prohibit the use of corporal punishment and physical, sexual or verbal abuse.

•  Facilities and suppliers shall use production methods that do not endanger workers safety or health.

•  Facilities and suppliers shall not call on police or military to enter their premises to suppress workers rights.

•  Employees shall have the freedom of association, assembly, of forming unions and bargaining collectively, of expression, and freedom from arbitrary arrest or detention. (This sort of clause always has to be read carefully, because sometimes human rights proposals are really mandatory unionization proposals in drag. It's completely appropriate for employees to have the right to unionize and bargain collectively, so long as a company is not prohibited from hiring non-union workers if reasonable agreements can't be negotiated. There's no mandatory unionization in this policy).

•  Employees shall not face discrimination on the basis of age, gender, marital status, political or religious activity, or arrest for peaceful protest.

•  Facilities and suppliers shall use environmentally responsible methods of production.

•  Facilities and suppliers shall prohibit child labor, or at a minimum comply with guidelines on minimum age for employment within China's national labor laws. (Again, this one is not as obvious as it might seem. The absence of child labor is in some sense a luxury of prosperity, and in many underdeveloped countries, child labor is a factor in the survival of the family, as it was during the early agricultural development of the United States. Even so, child labor is largely the result of the unavailability of living wages for adult workers. American companies should not allow this situation within the scope of their foreign operations).

•  The company will not provide products in China that can be used to commit human rights violations.

•  The company will issue annual statements detailing its efforts to uphold these principles.

-------------------------------------------------------------------------------------------------

Here is Sun's recommendation to shareholders:

Our Board unanimously recommends a vote 'AGAINST' the proposal for the following reasons:

•  More than eighty-five percent of shares voted on a substantially identical proposal at our last two annual meetings were voted against the proposal. (Well, shame on them.)

•  It would limit Sun's ability to manage complex and sensitive issues related to our operations in China.

•  Compliance would be difficult to measure, time-consuming, and costly, and would result in a diversion of resources from other equally important issues.

The proxy goes on to discuss Sun's commitment to the general intent of these principles, and the belief that management, rather than "broad and sweeping" policies, are best suited to address these issues.

I have no information as to whether Sun or its suppliers have had any direct or indirect involvement in these activities at all, and these comments should not be taken as any suggestion otherwise. But my view is simple. Some practices are so clearly unethical that they should be rejected out of hand, and ruled out formally, vocally, and without shades of grey, even if that rejection might increase the costs or reduce the business of a company. Sun offered no firm alternative policy for vote on these issues, just a statement that the issues were "complex and sensitive." Had Sun responded with some sort of clear, even reduced alternative, a no vote might have been reasonable. Sun did not provide this choice. We voted in favor of the proposal, though with little expectation of seeing it adopted.

In many cases, we believe that it can be useful to hold shares in an otherwise strong company and attempt to induce changes from within. This is not true for companies that are already marginal holdings for us. In Sun's case, we observe competitive pressures from the Windows/Intel platform, and increasing adoption of open systems such as Linux.