Analyzing the statistics:

  Examination of the statistics on wage growth and inequality since 1980 reveals, as might be expected, a mixed bag of statistical error that reduces but does not eliminate the extent of inequality, reveals the varying rates of growth of inequality during the quarter century since 1980, and the various contributing factors of which foreign trade turns out to be a relatively minor component. Lawrence writes prior to the credit crunch, which has undoubtedly at least temporarily substantially altered the extent of inequality.

  The inaccuracy of macroeconomic statistics is both pervasive and notorious, as FUTURECASTS repeatedly reminds its readers. See, Economic Statistics and Macroeconometrics: The Figures Lie. Those who utilize these statistics without analyzing them thereby confess their incompetence, ignorance and/or their elevation of ideological and propaganda interests above the truth.
 &
  Lawrence draws his figures from different sources and applies varying means of analysis in his effort to deal with the unreliability of macroeconomic statistics and mathematical analysis. That's hardly perfect, but it probably is the best that can be done. His findings are strong enough and sufficiently in line with economic logic to support confidence in their general validity.

The picture is considerably different if based on "compensation" rather than "wages" and if a uniform price deflator is used.

Technological change since 1980 appears to be the primary factor involved in the actual growth in inequality, although trade factors do seem to have played some role between 1980 and 1990.

Low skilled compensation seems to have kept pace after 1990.

  Labor productivity growth has been increasing faster than real wages, which would indicate that labor is no longer getting its traditional share of the economic pie. However, Lawrence demonstrates that about 60% of this gap is due to measurement problems involving the failure to include the rapidly rising cost of benefits in the wage calculations and the use of different price deflators for output and real wages. In other words, the picture is considerably different if based on "compensation" rather than "wages" and if a uniform price deflator is used. Another 10% is attributable to advances in white collar wages as technology has transformed the white collar world.
 &
  That leaves a still substantial gap of about 30% of the raw figures. Lawrence attributes that gap to a variety of factors, including gains by skilled workers, the gains of the super rich, and increases in class inequality due primarily to the rising share of corporate profits in national income prior to the credit crunch.
 &
  Technological change since 1980 appears to be the primary factor involved, although trade factors do seem to have played some role between 1980 and 1990. The U.S. is producing much more than it ever did, but technology enables it to produce more with fewer workers. The big surprise for Lawrence was that low skilled compensation seems to have kept pace after 1990. Since 1999, "whether workers are distinguished by skill, education, unionization, occupation, or major sectors," most relative wage and compensation measures show little increase in inequality.
 &

Low skill manufacturing of tradable products took its hits in the 1980s, so there is little of that left to be displaced.

Another factor is the business cycle.

  One explanation is that the U.S. is outsourcing its commodity manufacturing and concentrating on high value-added manufacturing of sophisticated products requiring more skill-intensive and capital intensive production methods. Low skill manufacturing of tradable products took its hits in the 1980s, so there is little of that left to be displaced. Thus, the impacts of trade competition now causes overall change in the nature of domestic manufacturing without impacting relative compensation levels of those still in the manufacturing sector.
 &
  Another factor is the business cycle. There were surges in corporate profits and asset values prior to the dot-com bust and the recent credit crunch. These surges had a big impact on inequality statistics. Over the long run, however, labor's share of the economic pie has been remarkably stable, and it may prove to be back at its traditional levels as the dust settles on the credit crunch.
 &

Inequality:

&

If inequality is to be reduced, it requires policies in the tax transfer area. Constraints on trade would be relatively ineffective and indeed might be counterproductive.

There are huge differences between after-tax income and household income figures and those for individual pretax and transfer income.

Trade would most likely affect factor prices, so Lawrence relies wherever possible on data on hourly earnings.

  How accurate is this statistical picture? What role has globalization played in this result? Trade has increased from about 20% of GDP in 1980 to about 28% in 2005,with recent import growth concentrated in goods and services from developing countries, Lawrence points out. Has trade put downward pressure on the wages of low income workers? Critics like the prominent economist and columnist Paul Krugman have answered this question affirmatively, Lawrence points out. Public support for free trade is waning.
 &
  Lawrence emphasizes that many factors besides trade are involved. He points to the impacts of technological changes, the drastic decline in private sector unionization, changing social norms, modifications in corporate governance, rising asset prices prior to the credit crunch,  industrial decline in many sectors, and immigration. He also emphasizes the major portions of the problem that are artifacts of statistical measurement and technical factors. He does not contend that trade has had no influence. His conclusion is that if inequality is to be reduced, it requires policies in the tax transfer area. Constraints on trade would be relatively ineffective and indeed might be counterproductive.
 &
  Lawrence divides his analysis into three parts. One involves wage inequality within worker classifications, another concentrates on the gains of the super rich, and a third on the gains in asset wealth and income from the capital of the upper classes.
 &
  He confines his analysis to pretax income and transfer inequality, and does not cover the great increase in wealth inequality. Wealth inequality is naturally increased during periods of rapid asset appreciation such as occurred throughout the "Great Moderation" from 1982 until the current credit crunch. He also concentrates on individual pretax and transfer incomes. There are huge differences between after-tax income and household income figures and those for individual pretax and transfer income. Even this narrow focus is complicated by worker decisions on how much to work and thus does not accurately reflect hourly wages. Trade would most likely affect factor prices, so Lawrence relies wherever possible on data on hourly earnings.

  "For much of the study I adopt the perspective of a typical blue-collar US worker. The central question I ask is, how much better off would this worker have been had income inequality not increased? Answering this  question requires not only comparing the relative growth of this worker's earnings to those of richer workers but also taking account of how much additional income would have been available to individual blue-collar workers had inequality not increased. In other words, what is important is not simply the percentage increase in the relative earnings of one group compared with another, but what share of total income this increase constitutes."

There has been a "relatively strong rise in real incomes for households with children in the lowest quartile." This contradicts claims that immigration and low-wage international competition is causing impoverishment.

Increases in "super rich" top 1% inequality have occurred in spurts in line with the business cycle

  Of course, increases in inequality do not mean an increase in poverty. There has been a "relatively strong rise in real incomes for households with children in the lowest quartile." This contradicts claims that immigration and low-wage international competition is causing impoverishment.
 &
   In the 1980s, wage inequality increased primarily because of the increasing rewards for skills. In the 1990s, wages at the bottom levels actually kept pace while median levels failed to keep up. Since 2000, all wage levels have done poorly except for those at the very top that surged ahead. Increases in class inequality appeared only after 2000, while increases in "super rich" top 1% inequality have occurred in spurts in line with the business cycle between 1985 and 1988 and in the late 1990s.
 &
  Most surprising, Lawrence point out:

  [Between 1999 and 2006] US relative wage and compensation measures indicate very little evidence of increased inequality by skill, education, unionization, or occupations and sector---indeed, if anything, compensation in manufacturing increased relatively rapidly. Apparently, neither trade nor technological change -- nor anything else -- has continued to increase conventional wage inequality."

The particular labor inputs in imports from low wage countries are qualitatively different from domestic labor inputs. They thus reflect increased specialization rather than direct competition.

Evaluation of developments between 2000 and 2006 must be tentative. Much of the disparity in wealth growth looks cyclical.

Most of the increase in wage inequality prior to the credit crunch was due to expansion and profit growth in financial services.

Manufacturing compensation has increased relatively more rapidly than compensation in general.

  Lawrence offers some general explanations for this development. Imports may have put downward pressure on wages generally but competition from low-wage countries could not disproportionately impact blue collar wages in the high value-added sophisticated manufacturing that continues in the U.S. Today, blue collar manufacturing jobs in the U.S. are held by relatively skilled workers. What the U.S. imports today is no long produced in the U.S., so declining import prices provide benefits for consumers.  (It thus stimulates domestic as well as import commerce). Also, competition from labor intensive production abroad is often overcome by capital and skilled labor-intensive production domestically.
 &
  The figures show that competition from manufactured imports is now concentrated "in U.S. manufacturing sectors that pay significantly higher than average U.S. wages." Thus, adjustments forced by import competition do not "fall disproportionately on less-skilled workers" and so does not increase inequality. Moreover, the particular labor inputs in imports from low wage countries are qualitatively different from domestic labor inputs. They thus reflect increased specialization rather than direct competition.
 &
  Lawrence wisely cautions that evaluation of developments between 2000 and 2006 must be tentative. Much of the disparity in wealth growth looks cyclical. (Indeed, the credit crunch must have altered this picture considerably.) He also notes that most of the increase in wage inequality prior to the credit crunch was due to expansion and profit growth in financial services. Offshoring has been a very minor factor and there has been little profit growth in manufacturing of traded goods. From the point of view of "compensation," he points out that the share of compensation in manufacturing or even in traded goods has not declined more rapidly that the rest of private industry, "and manufacturing compensation has actually increased relatively more rapidly than compensation in general."
 &
  Globalization has undoubtedly had some impact on the surging compensation of the "super rich," but this "is also being driven by technological changes, institutional developments such as financial deregulation, changes in US corporate governance practices, and rising asset markets, most of which have domestic origins."
 &

The gap between wage and productivity growth rates:

What happened to the other 65% of productivity growth?

Social Security, life insurance, health care and retirement benefits all ultimately come out of the pool of funds available for wages.

About 20% of the growth in inequality between blue collar and white collar workers disappears when total compensation is used as the standard.

  Benefits ate up a big chunk of this. Social Security contributions, life insurance, health care and retirement benefits all ultimately come out of the pool of funds available for wages.

  "The practice of referring to hourly wages -- or take home incomes -- without accounting for these benefits is a serious omission, not only for measures of take-home pay such as average hourly earnings but also for many of the household income measures that are frequently taken as indicating trends both in real incomes and inequality."

  The rapid increase in the costs of benefits accounts for about 25% of the gap between productivity growth and real wages. Taking account of benefits also materially reduces the growth in inequality.  About 20% of the growth in inequality between blue collar and white collar workers disappears when total compensation - including benefits - is used as the standard.
 &

The prices of goods and services produced in the business sector have risen more slowly than the goods and services workers consume, and this difference is reflected in the two different deflators. Thus, comparing real productivity growth statistics with real wage growth statistics is like comparing apples and oranges.

If the business sector price deflator is used for both wages and productivity growth, not only does 35% of the gap disappear, but wage growth suddenly shows a healthy average 1.5% real per  year increase.

  The use of different deflators for measuring real wages and for measuring productivity accounts for an even bigger chunk -  about 35% - of the gap. The prices of goods and services produced in the business sector have risen more slowly than the goods and services workers consume, and this difference is reflected in the two different deflators. Thus, comparing real productivity growth statistics with real wage growth statistics is like comparing apples and oranges.

  "In particular, the [business sector price deflator] has a higher weight for investment goods -- such as computers and machinery, whose prices have risen slowly---or even declined -- while the consumer price index gives a larger weight to housing and imports -- such as petroleum -- whose prices have increased more rapidly."

  If the business sector price deflator is used for both wages and productivity growth, not only does 35% of the gap disappear, but wage growth suddenly shows a healthy average 1.5% real per  year increase.

  This conforms well with survey data showing appreciable increases in household living standards during this quarter century. Of course, these statistics are misleading in another way. Most people in the lowest 20% of wage earners are no longer there ten years later. People do generally succeed in moving up the wage ladder as they mature and become more productive.

  Lawrence estimates that only about 15% of the statistical gap between productivity growth and wage growth is attributable to rising inequality of wage growth. The rest is attributable to about a 10% rise in class or profits inequality and a 5% rise in compensation inequality with the super rich. However, these influences waxed and waned at different times during this quarter century. This indicates that different factors were at work at different times. The rise and fall of the business cycle was more influential than steadier influences like trade.
 &

While there would appear to be some correlation between conventional theory and rising wage inequality in the 1980s, "the correlation between increased trade and increased wage inequality" breaks down when applied to the quarter century as a whole and especially to the last 15 years.

That wage inequality has not increased appreciably since 2000 is the most surprising finding for Lawrence. This contradicts conventional trade theory and econometric models.

  Most of the wage inequality occurred in the 1980s, a period that included the 1980-1982 depression followed by a period of high real interest rates and slow growth. The period 1995 to 2000 saw another surge in wage inequality due to the tremendous gains of the super rich during the dot-com boom. The super rich had lost ground during the prior five years that included a recession and a period of slow growth. Corporate profits and other factors of class inequality rose throughout the quarter century except during the dot-com boom, which was based on prospective rather than current profits.
 &
  That wage inequality has not increased appreciably since 2000 is the most surprising finding for Lawrence. This contradicts conventional trade theory and econometric models. According to theory, "opening up to trade should raise the relative price of skilled-labor wages and reduce unskilled-labor wages." Skilled labor would gain from increased demand for their goods in new developing country markets while unskilled labor would suffer from increased competition for their goods from those new markets. If these theories were correct, increased inequality along skill lines would be expected from expansion of trade with low wage developing nations.
 &
  There are other theories. Labor economists try to explain relative wages on the basis of changes in relative labor supplies within individual countries, and a "specific factors model of trade" that assumes a lack of internal mobility of factors confines the impacts to export and import sectors. "Specific factors" theories would not explain the general increase in inequality during the quarter century. While there would appear to be some correlation between conventional theory and rising wage inequality in the 1980s, "the correlation between increased trade and increased wage inequality" breaks down when applied to the quarter century as a whole and especially to the last 15 years.
 &

"[In the 1990s], perhaps because of a failure to raise minimum wages, the weakest wage performance was in retail trade---the quintessential nontraded goods sector."

  There thus seems to be three distinct phases during this quarter century whether blue collar compensation is compared with output per worker or white collar compensation or judged by the gap between high school and collage graduates. Inequality increased substantially across the board in the 1980s, only about half as much in the 1990s with the increase in inequality generally confined to the surge in the earnings at the top, and there was practically no increase in inequality since 2000. "The 1990s issue is thus about the increases at the top end rather than inequality across the board." Manufacturing compensation has risen significantly since 2000.

  "[In the 1990s], the raises received by blue-collar manufacturing workers were similar to those received by workers in health, education, and financial services---all of whom also enjoyed rapid increases. By contrast, perhaps because of a failure to raise minimum wages, the weakest wage performance was in retail trade---the quintessential nontraded goods sector."

  Lawrence candidly acknowledges that impacts with long lags complicate this analysis. How much of the increase in inequality of the 1980s was due to events in the 1970s during the Great Inflation? How much of an impact after 2006 will trade flows from before 2006 have? (The factors that contributed to the credit crunch can be traced back through the entire quarter century.)
 &

Globalization:

In the 1970s there was no increase in wage inequality, and wage growth was strong towards the end of the 1970s, particularly for union labor.

Import penetration was far less in the 1980s than in the 1970s or even in the 1990s, yet the 1980s are the decade when most of the increase in wage inequality occurred.

The share of imports from developing countries grew rapidly after 2000, "while the share of imports from developed countries actually declined." Yet there was very little additional inequality.

  It doubled to just about 20% by the end of the 1970s. The dollar devaluation and oil price surges account for much of this increase. During the 1970s, the relative prices of textiles declined by 30%. Textiles are "the paradigmatic unskilled-labor goods." However, there was no increase in wage inequality, and wage growth was strong towards the end of the 1970s, particularly for union labor.
 &
  While trade shares of GDP did not increase in the 1980s, its composition changed. Oil prices plummeted, to be displaced by a rapid increase in the value of manufactured imports. However, about half of the increase in manufactured imports came from developed countries. Import penetration was far less in the 1980s than in the 1970s or even in the 1990s, yet the 1980s are the decade when most of the increase in wage inequality occurred.

  "One would have to believe that there are substantial and variable time lags in the responses of wages to prices to claim that these data support the proposition that trade prices have significantly affected relative wages."

  In the 1990s, imports as a share of GDP increased about 45%. This was not as much as in the 1970s, but it was concentrated in imports from developing countries. Merchandise imports from developing countries increased from 2.8% to 5.3% of GDP. Yet the growth in inequality was mostly due to the surge in top earner compensation towards the end of the decade. Any correlation between import price pressure and increased trade inequality is "extremely weak."
 &
  Even not counting oil imports, the share of imports from developing countries grew rapidly after 2000, "while the share of imports from developed countries actually declined." Yet there was very little additional inequality.
 &

"I can conclude that without the impact of trade on wage inequality between 1981 and 2006, the wages of blue-collar workers would have been 1.4% higher than they were in 2006 and that almost all of this took place before 2000." 

As the world reacts to trade by specialization, import competition declines and instead imports increasingly compliment domestic production.

  Factors other than trade, such as technological changes, relative supplies of skilled and unskilled workers and changes in final product demand, must be examined to explain these results. A variety of methods have been used to compare the role of trade with such other factors, and while some role has been recognized for trade, mostly in the 1980s, trade never amounts to a dominant factor. Skill based technological change generally turns out to be the dominant factor.

  "I can conclude that without the impact of trade on wage inequality between 1981 and 2006, the wages of blue-collar workers would have been 1.4% higher than they were in 2006 and that almost all of this took place before 2000." (emphasis in original)

  To reduce that 1.4% of wage inequality by constraining trade would have reduced compensation across the board - and by considerably more than 1.4%..

  Lawrence speculates that the impact of globalization on wages was greatest at the beginning of the globalization process because imports competed with domestic production. However, as the world reacts to trade by specialization, import competition declines and instead imports increasingly compliment domestic production. That appears to be what happened during this quarter century.

  "Technically, the economy moves out of the cone of diversification and no longer produces unskilled-labor-intensive goods. The strong link between trade prices and factor prices is thus broken. If there are additional declines in relative prices of imported unskilled -labor-intensive goods and services, US consumers gain, but US relative factor prices are unaffected."

  It is impossible to support high living standards on low value-added commodity manufacturing. To maintain its high living standards, the U.S. had to concentrate on high value-added manufacturing.

Trade is now complementary rather than competitive, and actually improves the competitiveness of domestic production in both domestic and in export markets.

Developing countries like China and Mexico also exhibit substantial wage benefits for those in industries that compete in international markets.

  Specialization in the U.S. has increased the skilled production that now predominates and has reduced the impact of trade on inequality rates. There are fewer unskilled-labor-intensive activities capable of moving abroad. Indeed, domestic production that competes with low wage imports generally does so with capital-intensive high-skilled, high productivity methods. Further, much domestic production imports low-skilled components for capital intensive high value added domestic fabrication. This trade is complementary rather than competitive, and actually improves the competitiveness of domestic production in both domestic and in export markets.
 &
  Examination of wage levels in manufacturing industries engaged in world markets and/or competing with imports indicates that such competition now involves higher skill levels here and abroad, and higher wage levels across the board with the benefits the greatest at the lowest levels. Developing countries like China and Mexico also exhibit substantial wage benefits for those in industries that compete in international markets.

  "This means that on average the US manufacturing workers employed in industries that are most engaged in international competition earn relatively high wages at all percentiles, with the highest percentage differences at the lowest percentiles."

Disaggregating the statistics:

Even the broadest study indicates job loss at less than 2% of the workforce.

The macroeconomic data produced by high levels of aggregation may be deceptive. They "could miss some part of the story."

Where products are both imported and produced domestically, the two groups may be qualitatively different and thus not really in competition.

Studies of disaggregated data demonstrate higher levels of specialization and factor-intensity variances.

  The figures also vary over time. The composition of the wage labor displacement in the 1980s was concentrated among the less skilled while in 2000-2004, most of the displacement was in high end jobs.

  "[In the 1980s], displacement due to trade was relatively concentrated among unskilled workers, which led to numerous studies that argued that trade had contributed to wage inequality [cites]. But in the 1990s and after 2000, this has not been the case. Using US input-output coefficients, the net factor content of trade looks increasingly like the factor content of the US economy in general."

  Lawrence emphasizes that the macroeconomic data produced by high levels of aggregation may be deceptive. They "could miss some part of the story." (They almost always do!) For example, error is introduced if many of the imported products are no longer produced domestically. Where products are both imported and produced domestically, the two groups may be qualitatively different and thus not really in competition.
 &
  Lawrence notes that studies of disaggregated data demonstrate higher levels of specialization and factor-intensity variances. Wages earned in domestic industries that compete with imports from low-wage developing nations are well above manufacturing averages except for those in competition with imports form China.

  "This analysis not only helps to explain why the rapid import penetration by developing countries has not been associated with unusually weak wages among the least skilled Americans but also suggests that, unlike the earlier period, trade has not been reinforcing the effects of immigration on these workers' wages."

The failure to raise minimum wages in the 1980s seems to provide the strongest explanation for the pattern of wage inequality at the lowest levels.

The displacement due to offshoring is insufficient to be a major factor.

  The spread between the most unskilled and the median wage has not widened since 1990. Thus the recent surge in immigration has not had the feared effect of widening wage inequality for unskilled workers. Also, "simplistic application of the technology explanation for inequality" has failed to comport with this new pattern. The failure to raise minimum wages in the 1980s seems to provide a stronger explanation for the pattern of wage inequality at the lowest levels than changes in information technology and other skill-based technology.
 &
  Instead, skill-based technology developments may have had their biggest impact on inequality between the middle incomes and top incomes. Trade, too, may reward those at the top while offshoring hurts those in the middle. However, Lawrence finds that the displacement due to offshoring is insufficient to be a major factor.
 &

Profits:

&

Similar divergences of wage and profit shares occurred during prior business cycles. Further data from the late stages of the business cycle upsurge and its demise will clarify whether this is indeed a cyclical phenomenon.

  However, the business cycle seems to be responsible for much of this growth of profits. The business cycle is far more apparent in the financial sector - a prime beneficiary of the last cyclical upsurge - than in manufacturing. Lawrence shows similar divergences of wage and profit shares during prior business cycles. Further data from the late stages of the business cycle upsurge and its demise will clarify whether this is indeed a cyclical phenomenon.
 &
  The 4% rise in the share of corporate profits in national income between 2000 and 2006 came more at the expense of net interest payments and rental incomes than compensation. This was a period of unusually low interest rates (a factor that fueled the financial bubbles that burst during the credit crunch). The impact of surging health care benefit costs on wages since 2000 is far more important than the cyclical rise in profits. "The strong growth of compensation among unionized workers is not due to strong wage growth but to the fact that health care benefits have become more costly."
 &

Superstar inequality:

&

  Given the role of stock options in the pay of CEOs, it is not surprising that studies of the rise and fall of CEO pay compared to average worker pay tracks closely the rise and fall of the stock market averages running back to 1970. However, stock options became an increasingly important component of CEO pay only with the tax law changes of 1994. Congress, in its infinite wisdom, had denied the deductibility of wages above $1 million but allowed full deductibility of "performance based compensation." As always, unintended consequences attended this ham handed Congressional effort at economic micromanagement. Stock options were also an increasingly attractive method for compensating employees in the dot-com startups.
 &
  Were the CEOs worth their increased compensation, or were they simply getting better at "skimming" from corporate wealth? Some argue that the six-fold increase in CEO pay matches very well the six-fold increase in market capitalization between 1980 and 2003. With the great size and complexity of modern multinational corporations, the skills of the CEOs make a huge difference in outcomes.
 &

  Trade has not had any apparent role in this. The proportion of overseas operations in the activities of domestic corporations was about the same in 1980 as in 2000. "On average, US multinational firms were basically no more globalized in 2000 than they were in 1980!" The share of US corporate profits from abroad was similarly the same - about 15% - in 2000 as in 1980, although that share did spike higher significantly thereafter for a couple of years.

  "If CEO pay had been tied to stock market values, and these values related to firm output and/or profits, then the role played in pay increases by global activities would have simply been proportional to their 1980 share."

Globalization has certainly broadened the markets and opportunities for superstars in many fields. However, domestic markets remain the predominant source of income.

  However, CEOs and other top corporate executives are only a small fraction - only about 2% - of the top 0.5% of top earners. About half are financial executives, investment bankers, asset managers and hedge fund operators. Others include trial lawyers, the independently wealthy, top doctors, sports stars, movie producers and entertainers, and entrepreneurs, few of whom face import competition. "In 2004, only 14.2% of the top 1% was involved in tradable goods production."
 &
  For a century, it has been recognized that market-broadening technology - in transportation and communications - increases competitive pressures on everyone, but opens up vastly increased opportunities for the superstars. It's a "winner take all" phenomenon in many fields where the premium on offer for the superstars rises exponentially. Where people want and can purchase the very best, close substitutes won't do. Entertainers and financial experts serve vast numbers at little extra cost for each additional customer. This has little to do with trade. Legal and medical professionals reap the preponderance of their income from local activities. Almost all of the major financial firms earn about 80% or more of their income in the U.S.
 &
  Globalization has certainly broadened the markets and opportunities for superstars in many fields. However, domestic markets remain the predominant source of income. "While globalization's marginal contribution is impossible to quantify, a reasonable statement would be that globalization is not a disproportionately important source of the recent gains of the top US wage earners."
 &

Job dislocation:

The job loss rate for workers in tradable services is about 75% higher than for those in non-tradable services.

  However, 8.1 million workers could be classified as "displaced" during the three prosperous years through 2005. Less than half of these workers had held their jobs for more than 3 years.
 &
  Displacement from manufacturing jobs is costly. Income loss even for those that find new jobs is about 17%. Results are worse during business contractions and better during business expansions. The greatest losses are realized by the older or less skilled or those with considerable seniority.
 &
  Job security is an increasing concern, especially for those working in tradable goods industries. The statistics are admittedly inconclusive. It is difficult to disentangle domestic from trade influences. However, the job loss rate for workers in tradable services is about 75% higher than for those in non-tradable services.
 &

Even if the trade deficit were to disappear, the job loss in manufacturing would continue.

Outsourcing displacement is likely to remain just a minor factor in the massive churning of jobs each year.

  The dot-com bust and the period of slow job  growth during the next few years had a profound impact. The decline in manufacturing jobs accelerated - and manufacturing jobs continued to drift down even after recovery began. The figures provided by Lawrence are 17.3 million manufacturing jobs in mid-2000, reduced to 14.3 million by the end of 2003, and 14.1 million in 2007.
 &
  Very weak export growth was an important factor, along with continued productivity growth. Estimates of the impact of trade on these figures vary widely, from 12% to 33% of the job loss. Even if the trade deficit were to disappear, the job loss in manufacturing would continue.
 &
  Those with more education were more heavily impacted by unemployment than in the past. During the three years through 2003, almost 10% of college degree holding workers suffered displacement. Long term unemployment also surged upwards. Notoriously, advances in information technology has widened the scope of outsourcing of services for skilled and unskilled workers alike. Nevertheless, all the statistics show that this is still a minor trend. It is likely to remain just a minor factor in the massive churning of jobs each year.
 &

Lawrence concludes that there is no longer any way that the U.S. can protect itself from the impacts of global competition, so it is far better to concentrate on the expanding opportunities of globalization.

  Borders "still matter a lot, and adjustments across borders are slow." Distance, too, still matters a lot. Price differences of similar goods can be substantial across borders without being arbitraged away. The U.S. average with Canada is about 20%, with Europe about 30%, and with Japan about 50%. Differences in laws, customs, standards, regulations and policies account for much of these pricing variances.. Business relationships remain difficult to develop across borders.

  "International trade is one of the factors causing the dislocation of American workers, but it is considerably less important than the business cycle, technology, productivity growth, and competition between domestic firms. Education is important in limiting the costs and reducing the incidence of dislocation, but it is no panacea. Accordingly, the challenge for policy is helping workers undertake these adjustments, regardless of their source."

  Globalization conservatively contributes about 10% to U.S. GDP. Protectionism could greatly reduce the nation's economic performance and would anyway be ineffective at reducing inequality. Lawrence concludes that there is no longer any way that the U.S. can protect itself from the impacts of global competition, so it is far better to concentrate on the expanding opportunities of globalization.

  You can't carve greater shares of social justice out of a diminished economic pie.