Abstract
This Selected Issues paper on United States 2012 Article IV Consultation discusses rebound of manufacturing production. The U.S. share in global manufacturing production declined through most of the past three decades, but it has stabilized since the Great Recession. It currently represents about 20 percent of global manufacturing value added. Interestingly, after a sharp increase during most of the previous decade, China’s share in global manufacturing has also stabilized since the Great Recession, at a level similar to that of the United States. The notion of a manufacturing renaissance has been fuelled partly by the rebound in production since the end of the Great Recession.
The U.S. Manufacturing Recovery: Uptick or Renaissance?1
A. Introduction
1. A notable rebound of manufacturing production following the Great Recession has generated renewed interest in this sector among analysts and policy makers alike. Amid increasing anecdotes of a “renaissance” in U.S. manufacturing, many commentators have argued that the sector may contribute more significantly to domestic GDP and global industrial output going forward.2 They note that a number of favorable conditions—including a more depreciated exchange rate, lower domestic energy prices, volatile shipping costs, and significant increases in labor costs in emerging markets—could support steady increases in U.S. manufacturing output and employment, beyond those that could be attributed to just a cyclical rebound. The potential for growing demand from booming shale oil and gas activity have also been noted. At the same time, promoting manufacturing as an engine of high-wage jobs and growth is a key part of the U.S. administration’s economic policies.3 Others analysts are more skeptical, and argue that manufacturing output is merely rebounding to its pre-crisis level.4
2. This chapter investigates whether a renaissance is evident in U.S. macroeconomic data, and whether manufacturing could make a first-order contribution to long-term growth. First, it examines current and pre-crisis production levels for sub-sectors, as well as the share of manufacturing in U.S. and global GDP. Second, it documents a number of key structural factors contributing to the profitability of the U.S. manufacturing sector (in particular declining labor and energy costs). Third, it explores whether manufacturing could make a first order contribution to U.S. economic growth in the coming decade—on the back of relative cost advantages and the pull from growing shale oil and gas activity in the U.S.
3. The remainder of the chapter is organized as follows. Section B presents stylized facts on the U.S. manufacturing recovery. Section C presents results from a cross-country panel regression seeking to establish a link between manufacturing production and input costs that appear to be relevant as determinants of activity in the short-to-medium term. Section D discusses the potential rise in demand for U.S. manufactures from the coming boom in the shale energy sector. Section E takes a longer-term perspective, and assesses whether against the backdrop of increased global economic integration, as well as strong growth in emerging market economies, manufacturing can significantly add to U.S. growth in the next decade. The last section concludes with a summary description of the main takeaways.
B. Stylized Facts
4. Although manufacturing has been losing importance as a share of GDP during the last three decades, it remains an important sector of the U.S. economy. It accounts for about 75 percent of private sector R&D investment, represents more than 50 percent of export earnings, and provides for most high-wage jobs, especially for blue-collar workers.
5. Manufacturing jobs have also been on a declining trend, with a large drop in the Great Recession. Manufacturing employment declined by about 19 percent between the start of the 2001 recession and end-2007; it declined another 15 percent during the Great Recession. It has increased by 2 percent since the recovery started in mid-2009. The Great Recession hit employment in large firms the hardest, and somewhat less small and medium-sized firms, which are contributing more to manufacturing employment growth than in the past. Post-recession employment growth has been strongest in durable goods manufacturing (in particular in Computers and Electronics, and Machinery), while employment in the nondurable goods sector has remained stagnant.
6. The U.S. share in global manufacturing production declined through most of the past three decades, but it has stabilized since the Great Recession. It currently represents about 20 percent of global manufacturing value added (in nominal terms). Interestingly, after a sharp increase during most of the previous decade, China’s share in global manufacturing has also stabilized since the Great Recession, at a level similar to that of the United States.
7. The notion of a manufacturing renaissance has been fuelled partly by the rebound in production since the end of the Great Recession. It is important to note that this recovery was preceded by a very strong cyclical decline of about 20 percent during the Great Recession. Indeed, in early 2013, manufacturing production was still 4 percentage points below its pre-recession level.
8. But this aggregate dynamic masks significant differences between the durable and non durable sectors. The post-recession recovery has been driven almost entirely by a rebound in durable goods production. Although the rebound is in part the natural consequence of a stronger cyclical decline for durables, the difference in performance is remarkable: durable goods production surpassed its pre-recession level in 2011:Q3 (one quarter before overall real GDP did). In contrast, nondurable goods production is still about 10 percentage points lower than its pre-recession level (6 percent above its trough). Compared with the recoveries after the previous two U.S. recessions, the increase in durable goods production is markedly stronger during the ongoing recovery (unsurprisingly, given the comparatively steeper decline). By contrast, the rebound in nondurable goods has been weaker than that observed after the 1990 and 2001 recessions (Figure 1).
History of Manufacturing Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics, World Bank Databank, and Fund staff estimates.1/ Values for 2011 and 2012 are staff estimates.2/ For the Dec 07 - Jun 09 recession, t+15 represents January 2013.History of Manufacturing Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics, World Bank Databank, and Fund staff estimates.1/ Values for 2011 and 2012 are staff estimates.2/ For the Dec 07 - Jun 09 recession, t+15 represents January 2013.History of Manufacturing Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics, World Bank Databank, and Fund staff estimates.1/ Values for 2011 and 2012 are staff estimates.2/ For the Dec 07 - Jun 09 recession, t+15 represents January 2013.9. The strength of the durable sector has been concentrated in just a few subsectors. Three (out of ten) subsectors are responsible for the observed rebound in durables: Computer and Electronics, Motor Vehicles, and Machinery. While it is possible to argue that the increase in vehicles and machinery is mostly cyclical, computer and electronics have exhibited a robust positive trend during the past decade, including through the Great Recession (Figure 2). In contrast, most subsectors within nondurable goods have continued to decline or have shown a slow rebound after the Great Recession, including Chemicals, Plastics and Rubbers. The exception has been Petroleum Products, which has recovered to pre-crisis levels.
The Manufacturing Rebound and Natural Gas Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics and the U.S. Energy Information Agency.The Manufacturing Rebound and Natural Gas Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics and the U.S. Energy Information Agency.The Manufacturing Rebound and Natural Gas Production
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Sources: Haver Analytics and the U.S. Energy Information Agency.10. The decline in U.S. manufacturing production during the Great Recession was comparable to that in other G-7 countries but the recovery paths have been different. The rebound has been relatively strong in the United States and Germany, but has been muted in France and Italy. Moreover, although the recovery in U.S. manufacturing was in line with that of other G-7 countries through mid-2011, it has been relatively stronger thereafter (Figure 2).5 Looking at durable and nondurable goods production in G-7 countries provides additional insights. The post-crisis recovery in durable goods production in the United States was stronger than in other G-7 countries (Germany’s recovery had been stronger through mid-2011, but it stagnated thereafter). In contrast, the recovery in U.S. nondurable goods production has been poor also compared to that in other G-7 economies (Figure 3).
Manufacturing Production in Selected Countries
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Source: Haver AnalyticsManufacturing Production in Selected Countries
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Source: Haver AnalyticsManufacturing Production in Selected Countries
Citation: IMF Staff Country Reports 2013, 237; 10.5089/9781484376553.002.A001
Source: Haver AnalyticsC. Drivers of U.S. Manufacturing in the Short Run
11. A number of conditions have been highlighted as drivers of a U.S. manufacturing revival.6 These include:
A more competitive real effective exchange rate (REER); the U.S. REER has depreciated over the last decade, including due to the more advanced cyclical positions of emerging market economies relative to that of the U.S.
A decrease in the relative price of labor in the U.S. vis-à-vis emerging markets; the significant relocation of production to emerging Asia during the 1990s and the 2000s and high unemployment in the aftermath of the Great Recession have resulted in lower wage pressures in the U.S. and favorable changes in unit labor costs (ULC) over the last decade.
A significant reduction in domestic energy prices following technological breakthroughs in the exploitation of shale gas; in particular, recent advancements in drilling technology (including shale gas fracking) resulted in a significant increase in natural gas production in the U.S. and led to a reduction of domestic prices, which are currently about one fourth of those in Asia and Europe (Figure 2). According to projections by the Energy Information Agency (EIA), continued increases in shale gas production in the next few decades should keep prices comparatively low, although the favorable price gap with respect to Europe and Asia is expected to persist for the next few years before gradually diminishing (Chapter 2 of the Selected Issues).
12. The importance of these factors for manufacturing activity can be assessed through a simple panel regression. The model includes natural gas prices, ULCs, and REERs as possible drivers of manufacturing productions:
where LD(IP) is the log of difference of Industrial Production (IP) in manufacturing, Spread is the difference between the price of natural gas in country i and the world average, LD(REER) is the log difference in real effective exchange rates, LD(ULC) is the log difference of unit labor cost, and RecDummy is a dummy for the Great Recession. The regression results are reported in Table 1. As time series covering a sufficiently long period are only available for developed economies, cross sections only include G-7 countries. Columns (1)-(3) include the main regressors; the spread between U.S. gas prices and the rest of G-7 countries, the log difference of the real effective exchange rate, and log difference of the unit labor cost, one at a time, respectively. In addition, all regressions include one lag of the dependent variable, a Global Recession dummy, and quarterly time fixed effects. The dataset used is based on quarterly observations for G-7 countries from 2001 Q2 to 2013 Q1.7
Panel Regression Estimates for Total Manufacturing Production
Panel Regression Estimates for Total Manufacturing Production
| VARIABLES | (1) | (2) | (3) | (4) | (5) |
|---|---|---|---|---|---|
| Spread | REER | ULC | Spread-REER | All | |
| Spread | 0.000774** | 0.000780** | −0.000225 | ||
| (0.0466) | (0.0341) | (0.653) | |||
| Ind(REER) | −0.170*** | −0.170*** | −0.143*** | ||
| (1.23e-07) | (9.87e-08) | (7.41e-06) | |||
| Ind(UCL) | −0.418*** | −0.352*** | |||
| (1.59e-10) | (4.50e-08) | ||||
| Constant | −0.0240*** | −0.0247*** | 0.0269*** | −0.0253*** | −0.0156* |
| (0.00324) | (0.00152) | (0.00328) | (0.00108) | (0.0754) | |
| Observations | 303 | 303 | 260 | 303 | 260 |
| R-squared | 0.708 | 0.735 | 0.778 | 0.740 | 0.798 |
| Number of Countries | 7 | 7 | 7 | 7 | 7 |
Panel Regression Estimates for Total Manufacturing Production
| VARIABLES | (1) | (2) | (3) | (4) | (5) |
|---|---|---|---|---|---|
| Spread | REER | ULC | Spread-REER | All | |
| Spread | 0.000774** | 0.000780** | −0.000225 | ||
| (0.0466) | (0.0341) | (0.653) | |||
| Ind(REER) | −0.170*** | −0.170*** | −0.143*** | ||
| (1.23e-07) | (9.87e-08) | (7.41e-06) | |||
| Ind(UCL) | −0.418*** | −0.352*** | |||
| (1.59e-10) | (4.50e-08) | ||||
| Constant | −0.0240*** | −0.0247*** | 0.0269*** | −0.0253*** | −0.0156* |
| (0.00324) | (0.00152) | (0.00328) | (0.00108) | (0.0754) | |
| Observations | 303 | 303 | 260 | 303 | 260 |
| R-squared | 0.708 | 0.735 | 0.778 | 0.740 | 0.798 |
| Number of Countries | 7 | 7 | 7 | 7 | 7 |
13. Each of the three potential determinants of manufacturing is found to be highly statistically significant, and with the expected signs. Specifically, the estimated coefficient for the Spread is positive, indicating that the spread between the natural gas price in the U.S and G-7 countries is positively correlated with growth of U.S. manufacturing (and negatively correlated with manufacturing growth in the rest of the G-7 countries). The negative estimate for REER is consistent with the prior that a depreciating currency could boost manufacturing growth, as demand for exports increases. In turn, the negative sign for the ULC coefficient indicates that decreasing labor costs could augment US manufacturing growth.
14. Labor costs seem to be a more robust determinant of manufacturing activity than the other factors. Column (4) considers Spread and REER together in the same regression and shows that both variables remain significant. Column (5) includes all three variables in the same regression and shows that only the coefficient estimates on REER an ULC remain significant while estimates of the Spread variable becomes insignificant. It is worth noting that magnitudes of the significant coefficients indicate that ULC dominates the other two relationships, which could in turn reflect the strong decline of labor costs in the U.S. relative to other G-7 economies. At the same time we have to interpret this result cautiously given the small number of observations and possible noise in the ULC data.8
15. Moreover, Input-Output (I-O) tables suggest that a number of manufacturing sectors would profit from lower energy costs. For instance, a 10 percent decrease in the cost of energy implies a (continuous) 10 percent increase in the gross operating surplus of the Primary Metals sector, a 6 percent increase in Printing and Related Activities, a 5 percent increase in Paper Products, and a 4 percent increase in Chemical Products.
16. The model suggests that U.S. manufacturing could benefit from a continuation of recent trends in key costs. For example, a 1 percent decrease in U.S. ULC vis-à-vis other G-7 economies would result in an increase in U.S. industrial production of about 0.8 percent; similarly, a 1 percent REER depreciation would boost production by 0.2 percent. In turn, if the natural gas price gap between the U.S. and other G-7 economies would double, this would provide an additional stimulus to manufacturing production equivalent to 1.5 percent. Given the elevated degree of slack in the U.S. labor market, unit labor costs are likely to decline further in the next few years (before recovering) while the favorable cost advantage in natural gas is also likely to last for a few more years, thus supporting manufacturing activity in the U.S.
D. The Energy Boom—How Much of a Pull for U.S. Manufacturing?
17. The increasing U.S. production of oil and gas through unconventional extraction techniques can result in positive spillovers for manufacturing. In their baseline projections, the Energy Information Agency (EIA) suggest that total domestic production of oil and gas could increase by 10–15 percent through the end of the decade, with their upside risk scenarios pointing to increases of 30–50 percent. These scenarios factor in decreases in conventional production as well as significant increases of shale gas and tight oil production volumes. In this connection, tight oil is projected to increase by 40–120 percent through 2020, while shale gas growth is projected to fall in the 35–60 percent range. Higher production will necessitate higher inputs, including from the manufacturing sector.
18. However, basic analysis using I-O accounts suggest that the ‘pull’ from the energy boom to manufacturing would be limited. The additional production in the oil and gas industry brought about by the energy boom would result in a positive contribution to manufacturing growth of around 0.1–0.3 pp per year through the end of the decade.9 The increase would be larger for nondurable goods manufacturing (between 0.2 and 0.3 pp per year), as this includes the production of refined products. The industries that would benefit the most include chemical products, primary metals, fabricated metal products, and machinery. Interestingly, a number of the sectors that would be most benefitted have not been part of the manufacturing recovery to date, so the new source of demand for these sectors marks a positive development.
E. Long-Term Impact of U.S. Manufacturing
19. The evidence presented so far suggests that cyclical factors and favorable relative costs conditions may have been behind the observed recovery in U.S. manufacturing. In particular, the regressions in Section C provide evidence that cost reductions and a more depreciated REER have underpinned the recovery in U.S. manufacturing, and that these factors affected more strongly durable goods production.
20. Going forward, the question is whether a U.S. manufacturing revival could make a first-order (and sustained) difference to long-run growth. To approach this question, this section analyzes long-term cross country evidence on the determinants of the share of manufacturing in output, as well as long-term trends in U.S. manufacturing exports.
21. As emerging markets grow richer, the difference in their manufacturing-to-output ratio relative to that in the U.S. should decrease. Economic development models suggest that the manufacturing-to-output ratio increases at early stages of development, but that it then peaks and decreases as real per-capita income reaches relatively higher levels (Herrendorf et. al., forthcoming). Part of the reason could be that as an economy develops its currency tends to appreciate, losing part of its cost advantage to other less-developed economies, in particular for low-technology, more labor-intensive, industries. Cross-country evidence provides support to this claim. When their per capita income is compared to that in the U.S., countries cluster in two groups: One group (that includes most emerging and developing economies) is concentrated around per capita income levels which are about 20 percent of those the U.S., while the other group (that includes most developed economies) is concentrated around per capita income levels of about 80 percent of that in the U.S. Countries in the first group have manufacturing-to-GDP ratios that are larger than in the U.S., while for the second group, the ratios are similar to the U.S.. Real per-capita income and manufacturing-to-GDP ratios across G-20 countries follow the same bi-modal distribution.
22. The following panel regression explores the association between the U.S. manufacturing-to-output ratio (relative to that in other countries) with relative costs and real per-capita income levels. Concretely,
where MRt,i denotes the ratio of the share of manufacturing in output in country i to that of the U.S. (both in nominal terms) at time t, YRt,i denotes the ratio of real per capita income in country i to that of the U.S. at time t, YRSQt,i denotes the same variable but raised to the square power, and PRt,i represents the manufacturing deflator (calculated on US$ values) in country i to that of the U.S., at time t; the latter variable serves as a proxy for relative currency appreciations. For robustness considerations, the regression in (2) was estimated for 20-year rolling periods (beginning in 1970), and for a number of alternative country groupings including developed, emerging markets, and developing economies.
23. Results suggest that it is unlikely that the manufacturing-to-output ratio in emerging market economies will lose significant ground vis-à-vis the U.S. in the next few years. Results seem to suggest that the share of manufacturing in output vis-à-vis the U.S. begins to decrease for countries that reach real per capita incomes between 40 and 70 percent of those in the U.S. (i.e., at about those levels of relative real per capita income the quadratic term in equation (2) offsets the linear term). The effect of relative currency appreciations is less robust, but results for country samples including developed and developing countries suggest a negative effect of up to 0.7 percentage points in the relative ratio of manufacturing to output, for a 10 percent currency appreciation; this effect is much stronger (and statistically significant) for G-7 countries. (Table 2).10
Panel Regression Estimates for Long-Term Manufacturing Model
Panel Regression Estimates for Long-Term Manufacturing Model
| VARIABLES | G - 20 | G - 20 (ex-CHN) | G - 7 | S - 1 | All | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Constant | 3.74 | 26.00 | 331.02 | ** | 33.55 | * | 49.24 | *** | ||
| Real Per Capita Income | 6.08 | *** | 4.70 | *** | −3.61 | 3.61 | * | 2.67 | *** | |
| Real Per Capita Income (squared) | −0.04 | *** | −0.03 | *** | 0.01 | −0.02 | * | −0.01 | *** | |
| Manufacturing Deflator | 0.03 | 0.00 | −0.25 | *** | 0.07 | 0.06 | ||||
| R-squared (adjusted) | 0.81 | 0.77 | 0.81 | 0.75 | 0.78 | |||||
| Number of Observations | 5 | 5 | 5 | 5 | 5 | |||||
| Number of Cross Sections | 18 | 17 | 6 | 30 | 49 | |||||
| Number of total Observations | 90 | 85 | 30 | 150 | 245 |
Panel Regression Estimates for Long-Term Manufacturing Model
| VARIABLES | G - 20 | G - 20 (ex-CHN) | G - 7 | S - 1 | All | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Constant | 3.74 | 26.00 | 331.02 | ** | 33.55 | * | 49.24 | *** | ||
| Real Per Capita Income | 6.08 | *** | 4.70 | *** | −3.61 | 3.61 | * | 2.67 | *** | |
| Real Per Capita Income (squared) | −0.04 | *** | −0.03 | *** | 0.01 | −0.02 | * | −0.01 | *** | |
| Manufacturing Deflator | 0.03 | 0.00 | −0.25 | *** | 0.07 | 0.06 | ||||
| R-squared (adjusted) | 0.81 | 0.77 | 0.81 | 0.75 | 0.78 | |||||
| Number of Observations | 5 | 5 | 5 | 5 | 5 | |||||
| Number of Cross Sections | 18 | 17 | 6 | 30 | 49 | |||||
| Number of total Observations | 90 | 85 | 30 | 150 | 245 |
24. However, U.S. manufacturing exports could provide an outlet for stronger growth in the sector during the coming years. A look at the last few decades suggests that U.S. manufacturing exports have been more resilient than total manufacturing. While manufacturing has decreased as a share of GDP, manufacturing exports have remained about constant, and have come to constitute a larger share of total manufacturing output. However, in spite of this resilience, U.S. manufacturing exports have lost market share during the past few decades, in particular with respect to China and other emerging market economies. The application of a ‘Constant Market Share’ analysis to U.S. manufacturing exports suggests that the loss in market share is in part explained by the fact that U.S. exports have been primarily directed towards less dynamic regions (e.g., North America).11
25. These elements allow exploring the possible contribution of manufacturing to medium-term U.S. GDP growth. As indicated above, it is unlikely that convergence in per-capita income during the next few years will be enough for the manufacturing-to-output ratio in emerging market economies to begin losing ground vis-à-vis that in the U.S. However, this does not mean, that the manufacturing share in U.S. GDP will necessarily decrease.12 Emerging market economies are projected to continue growing at faster rates than developed economies and to increasingly contribute to global trade (including of manufacturing goods), which will continue to grow faster than world GDP.13 If the U.S. keeps its share in G-20 manufacturing exports through the end of the decade at the level observed in 2011, manufacturing could add up to 0.4 percentage points to growth per year through 2020. Further, an increase of 1 percentage point in the U.S. share in total G-20 manufacturing exports by 2020 would result in an additional 0.2 percentage points of growth per year. Moreover, the results from estimating (2) suggest that a one standard deviation increase in currency appreciation (namely a 25 percent increase in US$ manufacturing prices in comparator countries) through 2020 could add up to about 0.5 percentage points of additional GDP growth per year. Contributions to growth of such magnitudes are significant, given that manufacturing contributed less than 0.2 percentage points to growth per year (on average) during the first decade of the century.14
26. For manufacturing to have a first-order impact on growth during the next few years, the U.S. will have to diversify its export base. In order to keep its share in international markets, the U.S. will have to diversify its manufacturing exports base to more dynamic regions (e.g., Advanced and Emerging Asia). COMTRADE data compiled by the United Nations suggest that the share of U.S. manufacturing exports to the world’s dynamic regions remains low, but that has grown significantly during the past decade. In this connection, a look at the product composition of exports suggests that the external sales of chemical and plastic materials (all of which use energy intensively) have broadly outperformed the external sales of other sectors during the past decade. The growth rate of exports of machinery (electrical ant other), and transport equipment (which together account to about 40 percent of exports), has been less impressive, but the lower-than-average growth rates masks a change in the direction of exports, with sales to more dynamic regions (most notably Emerging Asia) increasing, and those to more mature markets stabilizing or decreasing during the past few years. The more attractive input costs vis-à-vis other G-7 economies (as pointed out in Section C) also bodes well for the future performance of U.S. exports.
F. Concluding Thoughts
27. While excessively optimistic claims of a U.S. manufacturing renaissance seem unwarranted, some sectors have indeed rebounded strongly following the Great Recession. Available data suggests that there are sectors within the durable goods category that have withstood the Great Recession well or have strongly rebounded thereafter, providing hope for a strong manufacturing presence in the U.S. and the global marketplace in some subsectors.
28. A number of ongoing factors are likely to positively impact the profitability of the U.S. manufacturing sector in the near-to-medium term. First, a combination of declining production costs—falling natural gas prices and ULC, and some real depreciation of the U.S. dollar—could catalyze new investment in the manufacturing sector, providing a boost to growth. Second, expanding shale oil and gas activity will create new demand for U.S. manufacturing output going forward.
29. The contribution of manufacturing exports to growth could exceed those of the recent past, fueled by rising global trade. U.S. manufacturing exports have proven resilient during the crisis. Further increases will require that the U.S. diversify further its export base towards the more dynamic world regions. In the long term, it is likely that a U.S. REER depreciation and the convergence of real per capita income in fast growing emerging market economies would result in a gradual increase in the manufacturing-to-output ratio in the U.S. vis-à-vis such economies.
References
The Atlantic, 2012, http://www.theatlantic.com/magazine/archive/2012/12/the-insourcing-boom/309166/?single_page=true (November).
Citi, 2012, “Is a Renaissance in U.S. Manufacturing Forthcoming?” Citi Research (May 31).
Financial Times, 2012, http://www.ft.com/intl/cms/s/0/65230da8-0317-11e2-a484-00144feabdc0.html#axzz2EnpQZY00 (September 21).
Goldman Sachs, 2013, “The US Manufacturing Renaissance: Fact or Fiction?” U.S. Economic Analyst, Issue No: 13/12 (March 22).
Herrendorf B., R. Rogerson, and A. Valentinyi, forthcoming, “Growth and Structural Transformation,” Handbook of Economic Growth.
Helper, S., T. Krueger and H. Wial, 2012, “Why Does Manufacturing Matter? Which Manufacturing Matters? A Policy Framework” Metropolitan Policy Program, Brookings Institution, February, Washington D.C.
International Monetary Fund, 2013, “World Economic Outlook,” (April).
McKinsey Global Institute, 2012, “Manufacturing the Future: The Next Era of Global Growth and Innovation” (November).
Morgan Stanley, 2013, “US Manufacturing Renaissance Is It a Masterpiece or a (Head) Fake?” (April 29).
New York Times, 2012, http://www.nytimes.com/2012/11/25/magazine/skills-dont-pay-the-bills.html?ref=itstheeconomy (November 25).
White House, State of the Union Address, 2013, http://www.whitehouse.gov/the-press-office/2013/02/12/remarks-president-state-union-address (February 12).
This chapter was prepared by Oya Celasun, Gabriel Di Bella, Tim Mahedy, and Chris Papageorgiou. It has benefited from discussions with Tam Bayoumi, Olivier Blanchard, Ben Bridgman, Craig Burnside, Roberto Cardarelli, Thomas Glaessner, Chang-Tai Hsieh, Deniz Igan, Simon Johnson, Gian Maria Milesi-Ferretti, Catherine Pattillo, Dani Rodrik, Nikola Spatafora, Martin Sommer, and Egon Zakrajsek.
For instance, recent public media articles and reports (e.g., Financial Times, September 21, 2012; New York Times, November 25, 2012; Citi Research, May 31, 2013).
Issues and policies related to the U.S. Manufacturing had a prominent place in the last State of the Union address (February 12, 2013).
This is illustrated by relatively high correlation coefficients (in excess of 0.8) between quarterly growth rates in U.S. manufacturing and those of comparator countries through mid-2011 (with the exception of Japan, which was affected by an earthquake in 2011), and reduced correlation coefficients thereafter (Figure 3). Interestingly, China’s manufacturing output (as measured by Purchasing Managers Index, PMI) seems to broadly lead output changes in U.S. manufacturing through mid-2011, though it seems to stagnate thereafter.
The ULC series ends at 2011 Q3.
Splitting the dependent variable into durable and nondurable manufacturing production provides a robustness check. The positive correlation between Spread and manufacturing growth is mainly driven by durable goods production; the same applies for both REER and ULC, as the regression for nondurable goods production results in significant coefficients for both REER and ULC, but of magnitudes that are four and six times smaller than the coefficients obtained for the durable goods regression. Other robustness checks also underline the dominating effect of durable goods in total manufacturing production trends. Further robustness checks (including using more lags of key regressors, using different sub-periods, and including additional regressors) did not qualitatively alter the results.
The gross increase in production (by type of activity) is given by X=T-1* D, where X is a 65-row vector consisting of the value of inputs used in production, T-1 is the inverse of the transformation matrix (which includes the intermediate input coefficients used by all production sectors), and D is a “shock” vector, with zeros in all lines except in those corresponding to oil and gas extraction and oil and coal products. The latter two were filled alternatively by the projected increase in production in tight oil and shale gas from 2012 to 2020 as included in the EIA reference and high-resource cases corresponding to the 2013 Annual Energy Outlook. Increases in durable and non-durable goods production was converted to value added using the aggregate ratios of value added-to-gross output for 2011.
The source for the data is the United Nations Database. Results reported in Table 2 correspond to fixed-effect regressions for 1990–2010. Details on the results for non-fixed effect regressions, other time periods, and other country groupings, as well as databases, are available from the authors upon request.
The Constant Market Share Analysis is based on the idea that the product and geographical structure of a country’s exports can affect is export growth. It constitutes a common method of analysis in international trade.
The model does not allow establishing to what extent changes in the relative ratio of manufacturing-to-output between two countries will occur through changes in the numerator, the denominator or both. Evidence for the past decade suggests that the U.S. manufacturing-to-output share, as well as the relative price of manufacturing, have both stabilized. Calculations on the potential impact of manufacturing in GDP growth assume that the relative price of manufacturing remains stable vis-à-vis GDP deflator. Moreover, it is assumed that the share of manufacturing exports in total manufacturing remains unchanged.
Growth and trade projections used in the analysis are those in World Economic Outlook (2013).
Manufacturing contributed about 0.7 percentage points (pp) per year to growth in the 1970s, and about 0.5 pp in the 1980s and the 1990s. The contribution to growth was larger in the 1960s (about 1.3 pp/year) and the 1950s (1.1 pp/year), on average.
