Bailing outsourcing

Journal of Comparative Economics xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Comparative Economics journal homepage: www.elsevier.com/locate/jce

Bailing outsourcing Travis Ng The Chinese University of Hong Kong, Hong Kong

a r t i c l e

i n f o

Article history: Received 22 January 2013 Revised 18 December 2013 Available online xxxx JEL classification: D40 L15 L51 Keywords: Offshoring Outsourcing Bailouts

a b s t r a c t Ng, Travis—Bailing outsourcing In a model of organizational choice, this paper shows that in face of an increasingly expected bailout from the government, outsourcing input production to an offshore location is more likely an optimal choice for a firm. Such a response is consistent with the three trends in the US manufacturing sector after the crisis: (a) employment keeps declining; (b) massive layoffs have not stopped; and (c) imported intermediate inputs have been gaining importance. Journal of Comparative Economics xxx (xx) (2014) xxx–xxx. The Chinese University of Hong Kong, Hong Kong. Ó 2014 Published by Elsevier Inc. on behalf of Association for Comparative Economic Studies.

1. Introduction This paper presents a model of organizational choice when a firm expects the government to be more likely to bail it out in face of financial distress. Relative to domestic production, outsourcing the production of inputs to an offshore location lowers a firm’s production cost at the expense of more unpredictable product quality. This is due to the difficulty of controlling quality across border: it hinges on the effectiveness of contract enforcement in the offshore location (Lu et al., 2012). Bailouts distort this trade-off by making unpredictable product quality less costly, resulting in a stronger-than-optimal incentive for the firm to outsource to an offshore location. If this mechanism has been what is going on, bailouts would have weakened rather than strengthened domestic job creation. The following 3 trends in the US manufacturing sector seems to support such a possibility: (a) employment growth has been weak; (b) massive layoffs in the manufacturing sector did not stop; and (c) imported intermediate inputs have been gaining importance. In October 2008, the US senate passed a vote on using the Troubled Assets Relief Program (TARP) money for non-financial institutions, officially beginning the unusually controversial auto industry bailouts. The expectation of bailouts has never been higher. Acharya et al. (2013) use the difference between what the large US banks paid their creditors and government debt yield as an estimate.1 They show that the expectation of bailouts in the US has been steadily rising from 2004–2006. It

E-mail address: [email protected] More precisely, their estimate is the difference between what the banks paid creditors for borrowing money and what investors earned from owning US government debt. The authors subtracted that number from the same measurement for smaller banks and, factoring in the differences in risk unrelated to bank size, determined the value of the implicit government subsidy, an estimate of the expectation of bailouts. Such an estimate peaks at 2009.20 basis points between 2002 and 2006, peaking up to over 120 basis points. 1

http://dx.doi.org/10.1016/j.jce.2014.01.002 0147-5967/Ó 2014 Published by Elsevier Inc. on behalf of Association for Comparative Economic Studies.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

2

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

then shifts gear and rises more rapidly in 2007. The expectation dramatically increases in 2008, and peaks at 2009, and continues to be high in 2010. How would such a dramatic increase in the expectation of bailouts affect the private sector?2 While most recent papers about bailouts focus on financial institutions, this paper focuses on the manufacturing sector instead. A major reason is that the politicians do talk a lot about domestic jobs, even more so when they are talking about bailouts.3 Given the weak employment figures in manufacturing, is it possible that bailouts actually have threatened rather than strengthen domestic job creation? If so, how would the underlying mechanism look like? The true cost of bailouts have just started to be better understood. The literature has been trying to estimate the true cost, including the moral hazard related to financial institutions taking exceedingly more risk. The manufacturing sector is unlikely to be immunized. But exactly how the sector’s moral hazard would pan out remains an open question, because they are unlikely to deal as much with financial derivatives as financial institutions do. Obscure financial transactions and off-the-balance-sheet tricks are also less common among manufacturing firms. I will first talk about the politics that motivate my inquiry of the relation between bailouts and jobs. I then discuss several trends that suggest a possible perverse effects of bailouts on manufacturing firms is to outsource more to offshore locations and thus threatening rather than strengthening domestic job creation. 1.1. The bailout politics In recent politics, both the auto industry bailout and outsourcing are widely discussed. At a Colorado pep rally, President Obama praised his GM bailout as an example for American industry to follow. ‘‘The American automobile industry has come roaring back. . .So now I want to say what we did with the auto industry, we can do it in manufacturing across America. Let’s make sure advanced, high-tech manufacturing jobs take root here, not in China. And that means supporting investment here. Governor Romney. . .invested in companies that were called ‘pioneers’ of outsourcing. I don’t want to outsource. I want to insource.’’4 The auto industry bailout, widely touted during the 2012 Democratic National Convention, costs billions.5 The above speech by President Obama has raised concerns about whether the government would bail out even more distressed firms.6 Such an impression is consistent with the Congressional Oversight Panel’s take. In its 2009 report on using the TARP money on the auto industry, the Panel recommends that ‘‘Treasury provide its legal analysis justifying the use of TARP funds for this purpose. This analysis will inform policymakers’ and taxpayers’ understanding of the potential for Treasury to use its authority to assist other struggling industries.’’ (The Congressional Oversight Panel, 2009) While a major focus of the auto industry bailout is on estimating the direct economic cost of taxpayers, an unintended consequence that attracts little attention is the changed expectation among other firms that would change firms’ strategies.7 The media has been reporting links between bailouts and outsourcing. Wells Fargo, the fourth-largest US bank by assets, received $25 billion under the TARP program, was reported to be looking to move some jobs outside the US, sending work in its retirement division, technology areas and other business lines to India and the Philippines.8 A source reported that the Bank of America, the second largest US bank which received $45 billion bailout funds, announced in the fall of 2011 to lay off 30,000 US workers but has plans to hire more in overseas.9 There are reports that other bailed-out companies had fired some of their US workers while shipping their jobs overseas.10 2 Anginer and Warburton (2013) specifically investigate the effects of Chrysler’s bailout on senior bond performance of industrial firms but not financial firms. Chrysler’s bailout undermined the US bankruptcy law by allowing the United Auto Workers Trust, an unsecured creditor of Chrysler, to cut in the line and subordinate the secured creditors of Chrysler during the process. Such undermining means unionized firms’ senior bond should perform badly. Anginer and Warburton (2013), however, found that there was actually no negative effects when they compare senior bonds across unionized and non-unionized industrial firms, suggesting that the expectation of being bailed out by the government for unionized firms more than compensated the negative effects brought by the disrupted bankruptcy process. 3 A good example is the statement from President Barack Obama released by the White House immediately after the government has sold its remaining stake in General Motors on December 9, 2013. http://www.whitehouse.gov/the-press-office/2013/12/09/statement-president. 4 Obama,. 2012. ‘‘Remarks by the President at Campaign Event – Colorado Springs, CO.’’ The White House Office of the Press Secretary, August 9. http:// www.whitehouse.gov/the-press-office/2012/08/09/remarks-president-campaign-event-colorado-springs-co. 5 The precise figure has been a mystery. See Ramseyer and Rasmusen (2011) for the difficulty of coming up with an accurate estimation due to special tax treatments that allow the new GM to pay substantially less corporate tax a few years down the road by carrying the accumulated losses of the old GM across the future tax years. Special tax treatments were available for Chrysler too. 6 President Obama continues on his remarks on the differential treatment between companies that do and do not invest overseas, ‘‘Let’s reform our tax code and let’s make it simpler. And let’s make sure that we’re providing tax breaks to companies that are investing here in Colorado Springs, here in Colorado – not overseas. (Applause.) They’re the ones who need tax breaks. Let’s give tax breaks to companies that are investing here. It’s the right thing to do.’’ Reforming tax code is not bailout, but reforming it with differential treatment on companies of different oversea strategies can be viewed as part of the integrated bailout packages to selected companies. 7 See Table 1 of Reinhart (2011) for a comprehensive time-line of the major events surrounding the financial crisis in 2008. 8 Rothacker. 2012. ‘‘Wells Fargo looking to move some jobs to India, Philippines.’’ Reuters, June 20. http://www.reuters.com/article/2012/06/21/uswellsfargo-offshoring-idUSBRE85K05A20120621. and Campbell, 2012. ‘‘Wells Fargo May Send Some Jobs To India, Philippines.’’ Bloomberg, June 21. http:// www.bloomberg.com/news/2012-06-20/wells-fargo-may-send-some-jobs-to-india-philippines.html. 9 Harkinson, Josh. 2012. ‘‘3 Years After Taxpayer Bailout, Bank of America Ships Jobs Overseas.’’ Mother Jones, May 29. http://www.motherjones.com/politics/ 2012/05/bank-of-america-outsourcing-call-center-philippines. 10 According to a news source, GM actually outsources quite a lot. GM outsources almost two thirds of its jobs overseas. Less than one in five GM vehicles are manufactured in the United States. After the bailout, GM’s 2011 annual report shows General Motors of North America accounting for 98,000 of the 207,000 GM jobs worldwide. But 12,000 of these jobs are in Canada and 11,500 are in Mexico. Accordingly, GM has 74,500 jobs in the United States and 122,500 abroad. Almost two thirds of GM’s jobs are in other countries. Source: Gregory, 2012. ‘‘Outsourcer-In-Chief: Obama Of General Motors.’’ Forbes, August 12.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

3

2000

4000

6000

8000

10000

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

2002

2004

2006

2008

2010

2012

Year Mass Layoff Events

Extended Mass Layoff Events

Fig. 1. Mass layoff data for the US manufacturing sector. Source: the US Bureau of Labor Statistics.

One may wonder why the money spent on saving the US economy through the TARP does not come with explicit constraints that prevent those companies that took the money from shipping jobs overseas. There is indeed a regulatory agency called the Special Inspector General for TARP (SIGTARP) set up alongside the TARP. Its main duty is to act as an oversight mechanism, making sure the money taken from the TARP fund is not abused or gamed by the financial institutions. The money going to the auto industry bailout comes from the TARP too. But the TARP is a program tailor-made for the financial institutions to deal with the ‘‘toxic’’ assets related to real-estate related securities. When the US government took action to bailout the Genearl Motors, the TARP money has never gone to any non-financial institution.11 To the best of my knowledge, there is no document clarifying whether there is a role played by the SIGTARP to oversee the outsourcing activities of participants of the TARP. The Congressional Oversight Panel that overlooks the TARP money on the auto bailouts also did not mention about offshore outsourcing (The Congressional Oversight Panel, 2009, 2011). Calculating the gains and losses of spending TARP money on the auto industry has never been easy. On December 9, 2013, the US government sold its remaining shares of GM (it once owned 60.8% using $49.5 billion earmarked for GM under the TARP). Treasury officials claimed that the government has recovered about $39.9 billion, making the total losses from the auto industry bailout at around $15 billion. In a follow-up statement released by the White House, President Obama claimed that the industry has added more than 372,000 new jobs.12 But he did not specifically mentioned when the counting starts and ends, and whether this has anything to do with the bailout. These types of confusion beg the question as to whether bailouts really help or hurt domestic job creation. 1.2. Recent trends in the US manufacturing sector The organizational choice model that depicts the way bailout distorts organizational strategies is motivated by recent trends in the US manufacturing sector. The US overall employment figures have been weak since the recovery officially started in the third quarter of 2009.13 In the manufacturing sector, not only has job creation been slow, mass layoffs events are numerous. Fig. 1 shows the number of events of mass layoffs and extended mass layoffs of the US manufacturing sector.14 It shows that the manufacturing sector has been losing jobs quite rapidly since 2007. Whereas the number of mass layoff events falls back to the pre-recession level in 2010, the manufacturing sector employment has been smaller in 2010 than in 2007, as shown in Fig. 2. Fig. 2 also shows that the industrial production index weakens during the recession but climbs back up after the recession. In contrast, employment in the manufacturing sector stays weak. Houseman et al. (2011) has documented such 11

That was the reason why in October 2008, the Senate has to officially vote for a go-ahead for widening TARP’s beneficiary pool. http://www.whitehouse.gov/the-press-office/2013/12/09/statement-president. Feng and Hu (2013) estimated that after correcting for mis-classification errors in the Current Population Survey, the unemployment rate stood at 11.5% between December 2007 and August 2011 (the officially reported 8.1%). The corresponding figure for November 2001–November 2007 was 6.9% (the official one was 5.1%). Between December 2007 and August 2011, non-white and those below 40 years old experienced the toughest time in terms of facing distressing unemployment. The corrected unemployment rates (official rate) for the following demographic groups are: Male/White/age40: 14.5(10.1); Male/Nonwhite/ age40: 19.3(16.0); Male/Nonwhite/age > 40: 13.9(9.6); and Female/Nonwhite/age40: 19.8(13.4). 14 The Bureau of Labor Statistics explains that ‘‘monthly mass layoff numbers are from establishments which have at least 50 initial claims for unemployment insurance (UI) filed against them during a 5-week period. Extended mass layoff numbers (issued quarterly) are from a subset of such establishments where private sector nonfarm employers indicate that 50 or more workers were separated from their jobs for at least 31 days.’’ An event can lead to more than 1 workers being laid off. Therefore, the number of separations (between an employee and an employer) is also reported. But the time trend is similar to those of the number of events and thus is omitted. 12 13

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

16000 14000

95

12000

90

(Index 2007=100)

10000

85 2000m1

(Thousands of persons)

18000

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

100

4

2002m1

2004m1

2006m1

2008m1

2010m1

2012m1

Time Industrial Production Index

Manufacturing Employment

Fig. 2. Industrial production index and the employment of the US manufacturing sector. Source: St. Louis Fed FRED).

a divergence: that the US manufacturing sector employment has been declining over time while manufacturing labor productivity rapidly grows over time. Another important trend in the US manufacturing sector that has caught much attention in the literature is that it has been using increasingly more imported intermediate goods and services (Antràs and Staiger, 2012, Feenstra and Jensen, 2012, Houseman et al., 2011). Fig. 1 of Eldridge and Harper (2012) shows the relative importance of the different types of input factors. Their data comes from the import matrix supplementary tables to the annual input–output table of the Bureau of Economic Analysis (BEA).15 Eldridge and Harper (2010, 2012) compute that imported inputs as a share of the value of manufacturing production grew from 13.1% in 1997 to a peak of 21.6% in 2008. During the same period of time, the corresponding share of domestic inputs declined from 38.9% to 35.5%, and that of labor declined from 30.1% to 25.7%. The financial crisis brought down the imported inputs share to 16.7 percent in 2009, but it quickly bounced back to 19.7% in 2010. Capital’s shares seem to be showing signs of recovery, while the share of domestic intermediate inputs continues to decline as a share of the value of manufacturing production coming out of the recession. The quick rebound of the share of imported intermediate inputs after 2009 seemingly comes at the expense of the continuing falling of the shares for domestic intermediates and labor. Eldridge and Harper (2010) and Houseman et al. (2011) estimate that the shift to imported, rather than domestic, inputs accounts for a significant portion of US manufacturing productivity gains in the past decades. 1.3. Literature review Research on the perverse incentives of bailouts has a long history (for example, Dewatripont and Maskin (1995) is the seminal work). Important work by Goldfeld and Quandti (1988, 1992) formalize the channels through which bailouts induce firms to hire more lobbyists and invest more in building political relationship. Almost all recent research on bailouts relate to the financial institutions since before the 2008 financial crisis, all firms who were bailed out were financial institutions. Those financial firms that are bailed out during the 2008 financial crisis all had some exposure to the ‘‘toxic’’ real-estate market related securities. Excessive risk-taking behaviors by financial institutions has been formalized by Panageas (2010) with supportive empirical evidence featuring German banks by Dam and Koetter (2012). This paper instead focuses on the manufacturing sector for three reasons. First, while the number of manufacturing firms that were actually bailed out is small, the number of firms that were triggered by the auto industry bailout to change their bailout expectation can be large and affect the whole sector. Second, the large-scale auto industry bailout remains a hot subject in recent US politics. Third, the US media has been covering news about the outsourcing behaviors of those firms that were bailed out. Offshore outsourcing has been a significant phenomenon; a huge literature has examined the effects of offshore outsourcing on labor market in terms of wages, inequality, employment, etc. (Antràs and Staiger, 2012).16 In contrast, this paper focuses on the role of domestic policy uncertainty as a catalyst of offshore outsourcing. As such, I focus on the causes of offshore outsourcing rather than the consequences. 15 Alternative data sources that do not rely on input–output table identify imported intermediate inputs using end-use classifications. First, the OECD STAN Bilateral Trade Database by Industry and End-use category offers such classification (Zhu et al., 2011). Second, the US Bureau of Census US Imports by 5-digit End-Use Code can be used too (Feenstra and Jensen, 2012). 16 See Antràs and Staiger (2012) and its brief literature review, and the two influential lectures by Feenstra (2008, 2011) for more comprehensive review of the literature.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

5

Fig. 3. The time-line.

2. The model I first describe the timeline shown in Fig. 3. A firm operates for at most two periods. Before its operation (period 0), it decides whether to outsource the production of a key component to an offshore supplier or to produce it in-house. The key component can be of either high or low quality. Following Economides (1999), assume that the final product is of high-quality only if the key component turns out to be of high-quality. Whereas the quality level has been realized in both period 1 and 2, the organizational choice made in period 0 determines the probability of realizing high quality level. In period 1, the firm generates revenue from selling its product. Before heading to period 2, however, the firm faces a liquidity shock. Insufficient cash-holding bankrupts the firm. If the firm does not bankrupt, the firm sells its product again in period 2. The product quality determines the firm value. First, conditional on survival, a high-quality product generates more profit than a low-quality product. Second, the firm with a high-quality product is more likely to survive through the liquidity shock than if it carries a lowquality product. Denote the probabilities of surviving in period 2 for a firm of high-quality and of low-quality product as qH and qL , respectively. Assume qH > qL .17 This is consistent with the notion that the market’s competitive force works in weeding out firms of low-quality products faster than those of high-quality products. In Section 4, the possibility of being bailed out distorts these survival rates. Conditional on survival, the firm faces a demand such that a high-quality product generates a per-period profit of 1, while that of a low-quality product is 1  a, with a 2 ð0; 1Þ.18 Therefore the firm value is19

v H ¼ 1 þ qH 1 v L ¼ ð1  aÞ þ qL ð1  aÞ

if the product turns out to be of high-quality; if the product turns out to be of high-quality:

ð1Þ

2.1. The organizational choice in Period 0 Period 0 has the following time-line. At stage 1, the firm decides whether to outsource to an offshore supplier or to produce the key component in-house. At stage 2, the component maker (either an internal division of the firm or an offshore supplier) takes precautions x 2 ½0; 1. Taking precautions x means the probability of having a high-quality component is x. At stage 3, the component is delivered. In the case of in-house production, the transaction is carried out without any friction, and the game ends. If the firm outsources to an offshore supplier, the two signs a contract stipulating a payment T > 0 for the firm to the supplier for the delivery of a high-quality component. The quality of the component is observable upon delivery by both parties, but it is not perfectly verifiable by a third party such as the court. The value of T is a function of the outside option R of the offshore supplier. Following Williamson’s notion of ex post lock-in, assume that ex ante, there are many potential offshore suppliers to choose from. As such, T is negotiated such that the supplier’s ex ante expected payoff equals its outside option.

17 A firm’s survival probability equals the chance that the size of its liquidity shock being smaller than its cash flow. Debt-financing before period 0 when the firm is first set up can lead to such a liquidity shock. For simplicity, however, the firm’s capital structure decision is not modeled. It does not mean that a firm’s capital decision is orthogonal to its organizational structure. The condition qH > qL amounts to the assumption that it is costly to structure the capital ex ante to fully hedge the risk of low product quality on the survival of a firm. Instead of using the number of competitors or the market concentration to proxy the intensity of competition, q embeds the notion of competition in the spirit of Alchian (1950), Stigler (1958) and Baker and Kennedy (2002): that competition improves efficiency through natural selection and that tougher competition weeds out incompetent firms more effectively. 18 This is to assume that the firm with a low-quality product cannot fully mimic the pricing strategy if it were to have a high-quality product to sell. This can be due to the product being search good, or that it can be an experience good but there is a small fraction of the informed customers who can tell whether the product is of high- or low-quality. 19 Note that for simplicity, I assume no discounting.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

6

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

However, ex post lock-in would lead to haggling and opportunism. Because lowering production cost motivates offshore outsourcing in reality, I assume that the outside option of the offshore supplier to be not too attractive. This leads to a reasonably low T.20 With no disagreement on the quality, payment would be made according to the contract; otherwise, the two may go to the court in the offshore location to resolve their dispute. The quality of the court in the offshore location is modeled as the   probability of coming up with the correct ruling, denoted h 2 12 ; 1 . If h < 1, contract enforcement with the offshore supplier 21 is imperfect. If a supplier fails to deliver a high-quality component and the firm sues it in the court, there is only a probability of h that the court rules against the supplier. Similarly, if the supplier indeed delivers a high-quality component but the firm claims otherwise, there is only a probability of h that the court rules against the firm. If the court rules against the supplier, the supplier needs to pay a damage of v H  v L to the firm, putting the firm in the same position as if the supplier has delivered a high-quality component.22 Without loss of generality, the cost of litigation is assumed to be the same for both parties, and is denoted by k. With no information asymmetry, the two parties will renegotiate the terms of the transaction using the court litigation as the default, rather than settle their dispute in court. Closing the model requires the specification of the production cost. Denote cðxÞ as the cost of taking precautions.23 For simplicity, I do not assume that the offshore supplier has an inferior technology, i.e., the same precautions-taking technology cðxÞ applies to both in-house production and the offshore supplier. As such, the results in the paper cannot be driven by the differences in the technology across countries. Denote the fixed cost of in-house production F I and that for the offshore supplier as F O . 2.2. Equilibrium product quality In the case of in-house production, the firm chooses the level of precautions x to maximize the expected profit, i.e.,

max xv H þ ð1  xÞv L  cðxÞ  F I :

ð2Þ

x2½0;1

and the associated first-order condition is

v H  v L ¼ c0

  xI ;

ð3Þ

where xI is the optimal level of precautions under in-house production. For the equilibrium product quality under outsourcing, consider the subgame in which the supplier delivers a low-quality component and the firm sues it in the court. With probability h, the court will rule against the supplier, under which the supplier is required to pay the damage v H  v L ; otherwise, the supplier does not need to pay any damages. The expected payoff for the supplier is T  hðv H  v L Þ  k, and that for the firm is v L  T þ hðv H  v L Þ  k. With no asymmetric information on the expected litigation outcome and costless renegotiation, they will settle their dispute without actually going to the court. Take ‘‘Nash-bargaining’’ in renegotiation with equal bargaining powers. Denote pl as the settlement price that the supplier pays to the firm. The payoffs for the supplier and the firm are T  pl and v L  T þ pl , respectively. The settlement price pl is the solution to the following optimization problem 1

1

max f½T  pl   ½T  hðv H  v L Þ  kg2  f½v L  T þ pl   ½v L  T þ hðv H  v L Þ  kg2 : pl

The corresponding solution is

pl ¼ hðv H  v L Þ:

ð4Þ

Consider now the subgame in which the supplier delivers a high-quality component but the firm claims otherwise. With probability 1  h, the court will rule against the supplier, under which it is required to pay the damage v H  v L to the firm; otherwise, the supplier does not need to pay any damages. The expected payoff for the supplier is T  ð1  hÞðv H  v L Þ  k, and that for the firm is v H  T þ ð1  hÞðv H  v L Þ  k. Again, with no asymmetric information on the litigation outcome and costless renegotiation, they will settle their dispute without actually going to court. Denote ph as the settlement price. The expected payoffs for the supplier and the firm are T  ph and v H  T þ ph , respectively. The settlement price ph is the solution to the following optimization problem 1

1

max f½T  ph   ½T  ð1  hÞðv H  v L Þ  kg2  f½v H  T þ ph   ½v H  T þ ð1  hÞðv H  v L Þ  kg2 : ph

20

I will show that T is a function of R, where R is exogenous and dependent on the wages and general economic environment of the host country. This can happen, for instance, when the members in the jury do not have the expertise to evaluate the component’s quality that requires special industry knowledge. 22 The rule of expectation damages is used here. The math will change slightly without altering the final results of the paper if opportunity-cost damages or reliance damages are used. 23 The appendix details the technical assumptions of this cost function. Assuming an interior solution requires cðxÞ to be convex, continuous, and twice differentiable. Assuming quality matters more makes outsourcing a less attractive option requires a bit more technical assumption. 21

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

7

The corresponding solution is

ph ¼ ð1  hÞðv H  v L Þ:

ð5Þ

Rolling back to Stage 2, the supplier chooses x to maximize its expected profit, i.e.,

max T  xph  ð1  xÞpl  cðxÞ  F O ;

ð6Þ

x2½0;1

and the associated first-order condition is

  ð2h  1Þðv H  v L Þ ¼ c0 xO :

ð7Þ

where xO is the optimal investment level under outsourcing. Comparing (3) with (7) allows us to conclude that xO 6 xI because c00 ðxÞ > 0 and h 6 1. The quality of the component under outsourcing is lower than that under vertical integration, more so as the quality of contract enforcement deteriorates (i.e., hdecreases). If contract enforcement were perfect ðh ¼ 1Þ, there would have been no difference in the component quality between outsourcing and vertical integration. Since the quality of the component determines the quality of the final product, Proposition 1 of Lu et al. (2012) can be derived as: Proposition 1. Product quality is lower under outsourcing compared to that under vertical integration. However, the negative impact of outsourcing on product quality is mitigated by the effectiveness of contract enforcement.

3. Equilibrium organizational choice Rolling back to stage 1, the firm chooses whether  to outsource   to an offshore supplier or to produce in-house. The expected profit of producing in-house is pI ¼ xI v H þ 1  xI v L  c xI  F I . Rearranging terms, we can express it as

  



pI ¼ v L þ xI ðv H  v L Þ  c xI þ F I : The expected profit of outsourcing is we can express it as

ð8Þ

pO ¼ xO ðv H þ ð1  hÞðv H





   v L ÞÞ þ 1  xO ðv L þ hðv H  v L ÞÞ  T. Rearranging terms,

pO ¼ v L þ xO ðv H  v L Þ þ xO ð1  hÞ þ ð1  xO Þh ðv H  v L Þ  T

ð9Þ

The payment T can now be pinned down by the outside option of the offshore supplier. Recall from (6) that the expected  profit of the supplier is T  xO ph  1  xO pl  c xO  F O , which has to be no smaller than its outside option R. In equilibrium, with many potential suppliers, the offshore supplier’s expected payoff would be equal to R, i.e.,

    T  xO ph  1  xO pl  c xO  F O ¼ R;

ð10Þ

Substituting in (4) and (5) and rearranging terms, we have

      T  xO ð1  hÞ þ 1  xO h ðv H  v L Þ ¼ R þ F O þ c xO :

ð11Þ

Therefore, the firm’s expected profit of outsourcing can be rewritten as functions of the supplier’s cost of production and its outside option as



 

pO ¼ v L þ xO ðv H  v L Þ  R þ F O þ c xO :

ð12Þ

Comparing (8) with the trade-off of outsourcing. Outsourcing compromises product quality, reducing the gross  (12) gives  expected profit by xI  xO ðv H  v L Þ. This is the cost the firm has to pay for outsourcing. Weakening the quality of contract enforcement raises this cost. The associated benefit,   however, equals the saving in the production   cost, expressed as the difference between in-house production cost of c xI þ F I and the outsourcing cost of R þ F O þ c xO . The firm chooses to outsource if and only if the benefit is larger than the cost, i.e.,

         c xI þ F I  R þ F O þ c xO P xI  xO ðv H  v L Þ:

ð13Þ

The following proposition summarizes the result: Proposition 2. The firm’s equilibrium organizational mode is to outsource the key component if and only if the following inequality holds:

         c xI þ F I  R þ F O þ c xO P xI  xO ðv H  v L Þ:

ð14Þ

Inequality (14) is more likely to hold if (a) cðxÞ increases sharply (i.e., more convex), (b) F I is large, (c) F O is small, (d) R is small, (e) xI  xO is small, or (f) v H  v L is small. Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

8

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

In words, everything else equal, outsourcing is more likely the equilibrium choice if. (a) (b) (c) (d) (e) (f)

the cost of taking precautions increases very quickly; the fixed cost of domestic production is high; the fixed cost of offshore production is low; the outside option of any offshore suppliers is low; contract enforcement is strong (i.e., h is closer to 1) in the offshore location; and product quality does not make that much of a difference to the firm value.

4. The incentive effect of raising the expectation of a bailout Cochrane (2009) refers to bailouts as creating an expectation among firms that traps the government. In this model, bailouts create a kind of policy uncertainty: it used to be certain that when a firm fails, it fails. The usual bankruptcy process kicks in. It is not the case anymore, however, when firms start expecting that they can survive through their financial distress because there is a certain chance, denoted g 2 ð0; 1, that the government would bail them out. A firm of product quality s ¼ H; L is under financial distress between periods 1 and 2 with probability 1  qs . With g > 0, the firm would be bailed out with probability ð1  qs Þg.24 As visually depicted in Fig. 4, therefore, the overall survival rate of a firm of product quality s ¼ H; L, becomes

c qs ðgÞ ¼ qs þ ð1  qs Þg:

ð15Þ

Recall that product quality affects the firm value through two channels. First, a low-quality product generates a lower per-period profit (conditional on survival, a high and low-quality product generates a per-period profit of 1 and 1  a, respectively). Second, it affects the firm’s probability to survive through the intermediate liquidity shock. The bailout probability g would now enter the firm value in (1) as

v H ¼ 1 þ qcH ðgÞ1 v L ¼ ð1  aÞ þ qcL ðgÞð1  aÞ Consequently, their difference,

if the product turns out to be of high-quality; if the product turns out to be of high-quality:

ð16Þ

v H  v L , also depends on g:

v H  v L ¼ a þ ½ qcH ðgÞ  ð1  aÞ qcL ðgÞ:

ð17Þ

The first term is the difference in firm value from period 1; the square bracket represents the difference in firm value from period 2. How does the change in the expectation of bailouts affect this value differential? Substituting (15) into this difference and taking the derivative with respect to g gives

@ðv H  v L Þ ¼ ð1  qH Þ  ð1  aÞð1  qL Þ: @g

ð18Þ

If expecting a bailout to be more likely converges the firm values (i.e., @ðv H@gv L Þ < 0), making product quality less relevant, then inequality (14) is more likely to hold, making outsourcing the more likely organizational mode. Fig. 5 shows visually that an increasingly expected bailout distort the incentive to outsource by distorting the market’s competitive force in weeding out incompetent firms. The shaded area represents the set of survival rates (or using the terminology of Baker and Kennedy (2002), the set of ‘‘economic grim reapers’’) such that the market’s competitive force is quite effective in weeding out firms of lower-quality products. Under this set, the partial derivative is negative (i.e., @ðv H@gv L Þ < 0). If the market’s competitive force becomes stronger (i.e., qH increases (closer to 1) while qL decreases (closer to 0)), the partial derivative is large in absolute size. Increasingly expected bailouts distort the market’s competitive force in weeding out firms of low-quality products, in turn distorting firms’ incentive to outsource. It is more likely to encourage firms to outsource more when the market’s disciplinary power is greater. The following proposition summarizes the result: Proposition 3. If the survival probabilities and the per-period profit difference between high- and low-quality product are such that

ð1  qH Þ  ð1  aÞð1  qL Þ < 0;

ð19Þ

the firm is more likely to choose outsourcing as the equilibrium organizational mode when it expects a bailout to be more likely. If the market’s competitive force weeds out firms of low-quality products more effectively, such an outsourcing incentive becomes even stronger. 24 Bailouts are modeled in a way similar to Panageas (2010). Specifically, if the firm is threatened by imminent default, the firm expects that with some probability the government makes a transfer large enough to prevent it from defaulting. Similar to Panageas (2010), this paper does not model the more complicated bailout strategies where the government either obtains a fraction of firm’s stock, or a fraction of the firm’s assets (or both) in exchange for providing the associated transfers.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

9

Fig. 4. The intermediate liquidity shock with a potential bailout.

Fig. 5. The shaded area would have bailouts encouraging outsourcing.

4.1. Implications The direct cost of bailouts is that taxpayers’ money is being transferred to some incompetent firms, forgoing the chance of freeing up the economic resources controlled by these firms through normal bankruptcy processes. An unintended consequence the literature has not explored is that bailouts also indirectly make firms less competent by distorting their organizational strategies. This distorted organizational strategy has a few implications. First, firms are less likely to produce their inputs domestically. This can contribute to the continuous employment decline in the US manufacturing sector after the crisis. It is also consistent with the mass layoffs data for the manufacturing sector. Second, firms are more likely to use imported intermediate inputs. This is consistent with the manufacturing sector using increasingly more imported intermediate inputs, as well as the quick rebound of the importance of imported intermediate inputs after 2009 as Fig. 1 of Eldridge and Harper (2012) shows. Third, there is no obvious reason to expect that employment shrinkage would be associated with significant output decline. This is consistent with the trend that employment has been declining in the manufacturing sector but with little sign of decline in its output. 5. Conclusion Constructing an index of policy uncertainty, Baker et al. (2013) show policy has been especially uncertain during the period of Lehman’s collapse, and the introduction of the TARP, both relate to bailouts.25 Their empirical results are consistent with the view that policy uncertainty has been hampering employment growth.26 One plausible underlying mechanism is that increased uncertainty raises the strategic value of ‘‘wait-and-see’’ or ‘‘holding back.’’ It may well explain why investment and job creation by the private sector has been weak, but less so for substantial job loss and an increase of the use of imported intermediate inputs in the post-crisis period. 25 Zingales (2012, pp. 89) gives an account of the spike in the policy uncertainty index around this period of time, ‘‘the government not only took no action to reduce the risk of a crisis but contributed to precipitating it via inconsistent policy choices: bailing out creditors but not shareholders in Bear Stearns, Fannie Mac, and Freddie Mac; wiping out both creditors and shareholders in Lehman Brothers and Washington Mutual; and bailout out both in AIG.’’ 26 Addressing to the Committee on the Judiciary, Taylor (2012) also argues that ‘‘the delayed recovery is due to poor government policies, of which regulatory expansion and policy uncertainty are a substantial part.’’

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

10

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

This paper supplements their finding by offering an alternative underlying mechanism through which policy uncertainty hinders employment growth. Specifically, the policy uncertainty this paper focuses on is that rather than a certain no-bailout policy, the possibility of being bailed out represents a lack of government commitment and thereby raises policy uncertainty (Du and Li, 2007). When a firm is triggered to expect that the government would bail it out if it goes into trouble, its incentive to outsource to an offshore location would be distorted. When product quality makes a meaningful difference to the survival of the firm, the distortion is to encourage the firm to outsource more to an offshore location. Such a distortion is consistent in spirit with the distortionary effect of bailouts on a firm’s input mix shown in Goldfeld and Quandti (1992). Acknowledgement I grateful acknowledge the finanical support of the Research Grant Council (GRF Projects no.: CUHK 4492/09H). Appendix A. The technical assumption The condition for offshore outsourcing is:

         c xI þ F I  R þ F O þ c xO P xI  xO ðv H  v L Þ:

ð20Þ

Technically, we need to analyze how a change in v H  v L affects all the components in the inequality. I assume that when quality matters more, offshore outsourcing becomes a less attractive option. This statement is correct only if this inequality is less likely to satisfy when v H  v L increases. Rearranging terms gives us:

       F I  ðR þ F O Þ P xI  xO ðv H  v L Þ  c xI  c xO :

ð21Þ 



The LHS, FI  ðR þ F O Þ, is exogenous. The RHS depends on v H  v L . Recall that ð2h  1Þðv H  v L Þ ¼ c and xI . Define c ¼ 2h  1; because imperfect contract enforcement in the offshore court requires h to be smaller     than 1, c < 1. Define k ¼ v H  v L . Define f ðxÞ ¼ c0 ðxÞ. Then we can write ck ¼ f xO and k ¼ f xI . Then the RHS can be expressed as functions of f ðxÞ and k: 0

v H  v L ¼ c0

ðf 1 ðkÞ  f 1 ðckÞÞk 

Z

xO

f 1 ðkÞ

f ðxÞdx:

ð22Þ

f 1 ðckÞ

If quality matters more makes offshore outsourcing a less attractive option, this expression should increase when k increases. Drawing the curve f ðxÞ; ðf 1 ðkÞ  f 1 ðckÞÞk is the area of a rectangle with height k in-between f 1 ðkÞ and f 1 ðckÞ, R f 1 ðkÞ and f 1 ðckÞ f ðxÞdx is the area under the curve f ðxÞ in-between f 1 ðkÞ and f 1 ðckÞ. The expression is thus the difference of the two areas. With this graph in mind, redefine gðxÞ ¼ f 1 ðxÞ. The RHS can then be expressed as

GðkÞ ¼

Z

k

gðxÞdx  ðk  ckÞgðckÞ:

ð23Þ

ck 0

Using Leibniz’s rule, @GðkÞ ¼ gðkÞ  gðckÞ  cð1  cÞkg ðckÞ. @k The sign of @GðkÞ concerns the curvature of the function gðxÞ. If gðxÞ is a linear function or convex function (meaning if f ðxÞ is @k 0 linear or concave), then by Jensen’s inequality, gðkÞ  gðckÞ > ð1  cÞkg ðckÞ. Since c < 1, therefore, 0 @GðkÞ gðkÞ  gðckÞ > cð1  cÞkg ðckÞ and @k > 0. Going back to our assumption of the cost function, therefore, a sufficient condition for the statement that ‘‘quality matters more makes offshore outsourcing a less attractive option’’ is such that c0 ðxÞ is either linear or concave. That means cðxÞ must be three-times differentiable, convex (i.e., c0 ðxÞ > 0 and c00 ðxÞ > 0) and c000 ðxÞ 6 0. References Acharya, Viral V., Anginer, Deniz, Warburton, A. Joseph, 2013. The End of Market Discipline? Investor Expectations of Implicit State Guarantees. Working Paper. Alchian, Armen A., 1950. Uncertainty, evolution, and economic theory. Journal of Political Economy 58 (3), 211–221. Anginer, Deniz, Warburton, Joseph A. 2013. The Chrysler Effect: The Impact of Government Intervention on Borrowing Costs. J. Banking Finan. doi: http:// dx.doi.org/10.1016/j.jbankfin.2013.11.006. Antràs, Pol, Staiger, Robert W., 2012. Offshoring and the role of trade agreements. American Economic Review 102 (7), 3140–3183. Baker, George P., Kennedy, Robert E., 2002. Survivorship and the economic grim reaper. Journal of Law, Economics, and Organization 18 (2), 324–361. Baker, Scott, R., Bloom, Nicholas, Davis, Steven J. 2013. Measuring Economic Policy Uncertainty. Chicago Booth Research Paper No. 13-02. . Cochrane, John H., 2009. Lessons from the financial crisis. Regulation 32 (4), 34–37. Dam, Lammertjan, Koetter, Michael, 2012. Bank bailouts and moral hazard: evidence from Germany. Review of Financial Studies 25 (8), 2343–2380. Dewatripont, Mathias, Maskin, Eric S., 1995. Credit and efficiency in centralized and decentralized economies. Review of Economic Studies 62 (4), 541–555. Du, Julan, Li, David D., 2007. The soft budget constraint of banks. Journal of Comparative Economics 35 (1), 108–135. Economides, Nicholas, 1999. Quality choice and vertical integration. International Journal of Industrial Organization 17 (6), 903–914. Eldridge, Lucy P., Harper, Michael J., 2010. Effects of imported intermediate inputs on productivity. Monthly Labor Review, June, 3–15. Eldridge, Lucy P., Harper, Michael J. 2012. Substitution of Imports for Domestic Labor, Capital and Materials in US Manufacturing. BLS Working Paper. .

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002

T. Ng / Journal of Comparative Economics xxx (2014) xxx–xxx

11

Feenstra, Robert C. 2008. Offshoring in the Global Economy. In: The Ohlin Lectures, 2008 at the Stockholm School of Economics, September 17–18, 2008. Feenstra, Robert C. 2011. Offshoring to China: The Local and Global Impacts of Processing Trade. In: Open Economy Lecture at the University of International Business and Economics, April 18–20, 2011. Feenstra, Robert C., Jensen, J. Bradford, 2012. Evaluating estimates of materials Offshoring from US manufacturing. Economics Letters 117 (1), 170–173. Feng, Shuaizhang, Hu, Yingyao, 2013. Misclassification errors and the underestimation of the US unemployment rate. Am. Econ. Rev. 103 (2), 1054–1070. Goldfeld, Stephen M., Quandti, Richard E., 1988. Budget constraints, bailouts, and the firm under central planning. Journal of Comparative Economics 12 (4), 502–520. Goldfeld, Stephen M., Quandti, Richard E., 1992. Effects of bailouts, taxes, and risk-aversion on the enterprise. Journal of Comparative Economics 16 (1), 150–167. Houseman, Susan, Kurz, Christopher, Lengermann, Paul, Mandel, Benjamin, 2011. Offshoring bias in US manufacturing. Journal of Economic Perspectives 25 (2), 111–132. Lu, Yi., Travis, Ng., Zhigang, Tao., 2012. Outsourcing, product quality, and contract enforcement. Journal of Economics and Management Strategy 21 (1), 1– 30. Panageas, Stavros., 2010. Bailouts, the incentive to manage risk, and financial crises. Journal of Financial Economics 95 (3), 296–311. Ramseyer, J. Mark., Rasmusen, Eric B., 2011. Can the treasury exempt its own companies from tax? The $45 billion GM net operating loss carryforward. Cato Papers on Public Policy 1, 1–53. Reinhart, Vincent., 2011. A year of living dangerously: the management of the financial crisis in 2008. Journal of Economic Perspectives 25 (1), 71–90. Stigler, George J., 1958. The economies of scale. Journal of Law and Economics 1 (October), 54–71. Taylor, John B. 2012. Government Regulatory Policies and the Delayed Economic Recovery. Testimony before the Committee on the Judiciary United States House of Representatives, September 20, 2012. The Congressional Oversight Panel, 2009. September Oversight Report: The Use of TARP Funds in the Support and Reorganization of the Domestic Automotive Industry. The Congressional Oversight Panel, 2011. January Oversight Report: An Update on TARP Support for the Domestic Automotive Industry. Zhu, Shiguang, Yamano Norihiko, Cimper Agnés, 2011. Compilation of bilateral trade database by industry and end-use category. In: OECD Science, Technology and Industry Working Papers, 2011/06, OECD Publishing. Zingales, Luigi., 2012. A Capitalism for the People. Basic Books, New York.

Please cite this article in press as: Ng, T. Bailing outsourcing. Journal of Comparative Economics (2014), http://dx.doi.org/10.1016/ j.jce.2014.01.002