dusselpeters.com · 2020. 7. 17. · Created Date: 20200620052100Z

Temple University Press

Chapter Title: Learning and the Limits of Foreign Partners as TeachersChapter Author(s): Enrique Dussel Peters, Clemente Ruiz Durán and Michael J. Piore

Book Title: Free Trade & Uneven DevelopmentBook Subtitle: North American Apparel Industry After NaftaBook Editor(s): Gary Gereffi, David Spener, Jennifer BairPublished by: Temple University Press. (2002)Stable URL: https://www.jstor.org/stable/j.ctt14bt2r1.16

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

https://about.jstor.org/terms

Temple University Press is collaborating with JSTOR to digitize, preserve and extend accessto Free Trade & Uneven Development

This content downloaded from 201.137.202.220 on Sat, 20 Jun 2020 05:21:00 UTCAll use subject to https://about.jstor.org/terms

Enrique Dussel Peters, Clemente Ruiz Durán,and Michael J. Piore

Learning and the Limits of

Foreign Partners as Teachers

Recent Economic Trends

The garment industry is being hailed as theoutstanding success of the North AmericanFree Trade Agreement (NAFTA), at least fromthe Mexican point of view. Garment exportsto the United States have expanded from lessthan $ million in to $. billion in. Moreover, since , when the agree-ment actually went into effect, that rate hascontinued to increase as more and more pro-ducers move facilities from other parts ofNorth America and the Caribbean Basin toMexico. But NAFTA is the culmination of theprocess of opening the Mexican economy totrade, a process that began in the mid-s,and the increase in imports from Mexico asso-ciated with that process has also been dramatic.As shown in Table ., in the period leadingup to NAFTA (‒) the annual increasein real imports averaged . percent. Tables.‒. additionally reflect that maquiladoraexports have been the driving force in Mex-ico’s garment industry. Specifically, temporaryimports to be reexported (i.e., imports that aretransformed temporarily, without payment oftariffs or taxes and without value added,through programs such as the maquiladora

program) remain the core of garment exports(Alvarez Galván and Dussel Peters ).

Independent of the recession in Mexicanexports since , the import figures reflectin part that the Mexican garment industry isincreasingly a subcontracting operation, anextension of the pattern of development ini-tiated under the maquiladora program whereaccess to U.S. markets is mediated by foreigncompanies that design the product, supply thematerials (in garments, often in the form ofcut pieces), specify the production process, andthen take over the final output for sale abroad.The annual increase in imports for plants oper-ating under this program in ‒ averaged percent.

But the import figures also reflect a darkerside of the structural changes occurring in theMexican economy. The opening has had a dev-astating impact on traditional producers; thecountry has increasingly lost its domestic mar-ket to imported foreign goods. It is hard toidentify this loss precisely, because figures forthe industry as a whole mask the division be-tween the expanding and contracting sectors,and so many of the losses have been in smallfirms in the informal sector that the officialfigures do not capture at all. The magnitude of


..

Gen

eral

Dat

a on

Gar

men

t Ind

ustr

y,

‒

a

Cum

ulat

ive

Cha

nge

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

1988

–99

1988

–93

1994

–99

GD

P (s

hare

ove

r to

tal)

0.86

0.75

0.71

0.70

0.69

0.64

0.61

0.60

0.65

0.70

0.69

0.70

0.68

0.70

0.67

GD

P gr

owth

(198

8 �

100)

b10

0.0

103.

611

4.5

120.

012

4.2

121.

412

4.3

116.

613

6.2

147.

215

3.9

162.

94.

54.

06.

1

GD

P gr

owth

(198

8 �

100)

, to

tal e

cono

myb

100.

010

4.2

109.

511

4.1

118.

212

0.5

125.

911

8.1

124.

213

2.6

139.

314

4.5

3.4

3.8

3.7

Em

ploy

men

t (sh

are

over

tota

l)0.

910.

900.

910.

920.

920.

960.

960.

971.

151.

341.

471.

640.

950.

851.

03

Em

ploy

men

t gro

wth

(198

8 �

100)

b10

0.0

99.9

100.

910

1.5

101.

210

5.8

106.

110

6.9

126.

714

7.8

161.

918

0.9

5.5

0.5

11.3

Em

ploy

men

t gro

wth

(198

8 �

100)

, to

tal e

cono

myb

100.

010

2.9

107.

811

1.0

112.

811

4.1

117.

011

3.6

117.

512

1.9

127.

313

0.5

3.6

0.8

6.2

Prod

uctiv

ity g

row

th (1

988

�10

0)b

100.

010

3.7

113.

511

8.3

122.

711

4.7

117.

110

9.1

107.

599

.695

.090

.1�

0.9

2.8

�5.

1

Prod

uctiv

ity g

row

th (1

988

�10

0),

tota

l eco

nom

yb10

0.0

101.

310

1.5

102.

810

4.8

105.

610

7.6

104.

010

5.7

108.

810

9.4

110.

70.

91.

10.

6

Rea

l wag

e gr

owth

(198

8 �

100)

b10

0.0

101.

099

.510

3.0

107.

910

8.1

108.

588

.080

.583

.685

.490

.2�

0.9

1.6

�3.

6

Rea

l wag

e gr

owth

(198

8 �

100)

, to

tal e

cono

myb

100.

010

0.7

101.

110

7.0

116.

312

3.2

127.

510

9.1

103.

510

9.8

113.

011

6.8

1.4

4.3

�1.

7

Exp

orts

(sha

re o

ver

tota

l)0.

400.

380.

340.

400.

430.

450.

510.

690.

851.

121.

471.

451.

050.

401.

16

Exp

ort g

row

th (1

988

�10

0)b

100.

011

7.7

124.

514

1.8

130.

014

0.4

180.

233

9.2

468.

157

1.1

754.

184

2.3

21.4

7.0

36.1

Exp

ort g

row

th (1

988

�10

0),

tota

l eco

nom

yb10

0.0

101.

510

6.3

114.

211

5.6

124.

714

2.1

197.

523

3.8

252.

726

7.2

290.

110

.24.

515

.4

Impo

rts

(sha

re o

ver

tota

l)0.

611.

111.

201.

231.

511.

531.

110.

720.

610.

761.

010.

930.

941.

310.

87

Impo

rt g

row

th (1

988

�10

0)b

100.

021

9.7

301.

937

4.2

577.

759

5.1

527.

025

0.0

268.

942

0.3

599.

158

9.7

17.5

42.9

2.3

Impo

rt g

row

th (1

988

�10

0),

tota

l eco

nom

yb10

0.0

121.

314

9.2

181.

122

5.0

223.

026

9.0

195.

824

5.4

311.

635

9.2

396.

813

.317

.48.

1

Tra

de b

alan

ce/G

DP

�3.

35�

13.6

0�

17.9

3�

19.0

7�

28.0

5�

27.4

4�

21.2

6�

0.99

6.17

5.64

2.30

5.12

�9.

4�

18.2

�0.

5

Tra

de b

alan

ce/G

DP,

to

tal e

cono

my

�0.

52�

2.02

�2.

62�

4.64

�6.

91�

5.76

�6.

940.

04�

0.87

�3.

27�

5.88

�5.

35�

3.7

�3.

7�

3.7

Sour

ce:

Aut

hors

’ est

imat

es b

ased

on

data

obt

aine

d di

rect

ly fr

om I

NE

GI

(Ins

titut

o N

acio

nal d

e E

stad

ístic

a, G

eogr

afía

e I

nfor

mát

ica)

.a R

efer

s to

Bra

nch

(G

arm

ents

) in

the

Mex

ican

Nat

iona

l Acc

ount

ing

Syst

em; d

oes

not i

nclu

de th

e m

aqui

lado

ra in

dust

ry.

b The

per

iods

‒

,

‒

,an

d

‒

re

fer

to th

e av

erag

e an

nual

gro

wth

rat

e.


..

Mex

ico:

Exp

orts

, Im

port

s, a

nd T

rade

Bal

ance

,

‒

a

Cum

ulat

ive

Cha

nge

1990

–19

94–

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2001

1990

–93

2001

U.S

.$ M

illi

ons

Gar

men

tsb

Exp

orts

7852

482

299

91,

500

2,52

03,

557

5,41

76,

430

7,56

38,

427

7,83

645

,674

2,42

443

,250

Impo

rts

361

667

1,06

21,

124

1,47

41,

737

2,30

93,

208

3,62

53,

517

3,47

23,

323

25,8

783,

213

22,6

64T

rade

Bal

ance

�28

2�

142

�24

0�

126

2678

31,

248

2,20

92,

805

4,04

64,

955

4,51

319

,796

�79

020

,586

Tot

al E

cono

my

Exp

orts

26,8

3842

,687

46,1

9551

,832

60,8

3379

,823

96,0

0011

0,38

011

7,50

013

6,70

316

6,42

415

8,54

21,

093,

757

167,

552

926,

205

Impo

rts

32,8

0251

,724

64,2

1367

,548

79,3

7472

,475

89,4

6910

9,79

812

5,24

614

2,06

317

4,47

316

8,27

51,

177,

460

216,

287

961,

173

Tra

de B

alan

ce�

5,96

4�

9,03

7�

18,0

18�

15,7

16�

18,5

417,

348

6,53

158

2�

7,74

6�

5,36

0�

8,04

9�

9,73

3�

83,7

03�

48,7

35�

34,9

68

Perc

enta

ge (O

ver

Res

pect

ive

Tot

al)

Gar

men

tsb

Exp

orts

0.29

1.23

1.78

1.93

2.47

3.16

3.71

4.91

5.47

5.53

5.06

4.94

4.18

1.45

4.67

Impo

rts

1.10

1.29

1.65

1.66

1.86

2.40

2.58

2.92

2.89

2.48

1.99

1.97

2.20

1.49

2.36

Gro

wth

Rat

e (P

erce

nt)

Gar

men

tsb

Exp

orts

—56

9.5

56.7

21.5

50.2

68.0

41.1

52.3

18.7

17.6

11.4

�7.

052

.066

.426

.6Im

port

s—

84.9

59.3

5.9

31.1

17.9

32.9

39.0

13.0

�3.

0�

1.3

�4.

322

.425

.512

.3

Tot

al E

cono

my

Exp

orts

—59

.18.

212

.217

.431

.220

.315

.06.

516

.321

.7�

4.7

17.5

14.1

14.7

Impo

rts

—57

.724

.15.

217

.5�

8.7

23.4

22.7

14.1

13.4

22.8

�3.

616

.015

.511

.3

Sour

ce:

Aut

hors

’ est

imat

es b

ased

on

Ban

com

ext (

).

a Incl

udes

maq

uila

dora

act

iviti

es.

b Ref

ers

to c

hapt

ers

(a

rtic

les

ofap

pare

l and

clo

thin

g ac

cess

orie

s, k

nitt

ed o

r cr

oche

ted)

and

(art

icle

s of

appa

rel a

nd c

loth

ing

acce

ssor

ies,

not

kni

tted

or

croc

hete

d) o

fth

e H

arm

oniz

ed T

ariff

Sche

dule

.


this effect is suggested by one estimate for ‒

, when official imports in garments, not in-cluding maquiladoras, rose percent; whenused and contraband clothing are included, theincrease was percent (according to infor-mation provided by one firm interviewed forthis study). In real terms the value of produc-tion in garments increased by only . percentover the period.

These figures changed dramatically after thedevaluation of the peso in December . In imports of garments, excluding maquila-doras, declined by a startling percent. But agood part of that decline reflects the suppres-sion of Mexican domestic demand and cannot

be sustained over the long run. In fact, in

imports of garments began to rise again—by percent—wiping out over percent of theimport decline in the previous two years. Thelosses in the domestic market to imports areparticularly surprising given that Mexico’scomparative advantage should lie precisely inthese low-wage, labor-intensive industries.Considerable adjustment is to be expected inthe face of newly emergent foreign competi-tors. It is not clear, however, why that adjust-ment should involve a loss of the domestic mar-ket. In principle, if Mexico can be competitiveon the international front, it should be able tocompete on the domestic front at least as well.

This chapter reports the findings of a studydesigned to explore why comparable competi-tion on the international and domestic frontshas not been the case in Mexico. The findingsare based on material gathered in the periodfrom to as part of a larger projectstill continuing on the adjustment of Mexicanfirms to the opening of the economy to trade.While the focus here is on the clothing indus-try, the study on which it draws is focused ontraditional industries more broadly, and mate-rial from shoes, furniture, and ceramics sup-plements that drawn directly from the cloth-ing industry in developing the argument.

The findings moreover have potential im-plications extending beyond these industriesto the Mexican manufacturing sector as awhole. The dichotomy we observe in the gar-ment industry between the larger, more capital-intensive firms that are prospering under thenew trading regime and the smaller, more labor-intensive firms that are not replicates a patternreflected in the broader aggregates for Mexi-can manufacturing. Indeed, the most success-ful Mexican industries in recent years havenot been those where one would have expectedthe country’s comparative advantage to lie butrather capital- and skill-intensive industries as-sociated with relatively advanced technologies

.. Mexico: Export Structure,‒a

1998 1999 2000 2001

U.S.$ MillionsGarmentsb

Total 6,404 7,554 8,427 7,831Temporary 6,090 7,318 8,196 7,625Definitive 313 236 232 206

TotalTotal 117,442 136,703 166,424 158,547Temporary 97,518 114,814 137,251 131,429Definitive 19,924 21,889 29,173 27,118

Percentage (Garment Total � 100)Garmentsb

Total 100.00 100.00 100.00 100.00Temporary 95.11 96.87 97.25 97.37Definitive 4.89 3.13 2.75 2.63

Percentage (Over Respective Total)Garmentsb

Total 5.45 5.53 5.06 4.94Temporary 6.25 6.37 5.97 5.80Definitive 1.57 1.08 0.79 0.76

Source: Authors’ estimates based on Bancomext ().

aIncludes maquiladora activities.

bRefers to chapters (articles of apparel and clothing acces-sories, knitted or crocheted) and (articles of apparel andclothing accessories, not knitted or crocheted) of the Harmo-nized Tariff Schedule.


..

Maq

uila

dora

Exp

orts

of

Gar

men

ts in

Mex

ico,

‒

a

Cum

ulat

ive

Cha

nge

1990

–19

94–

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2001

1990

–93

2001

Gar

men

tsF

irm

s29

035

039

339

941

252

265

078

690

710

3511

1995

87,

821

1,43

26,

389

Em

ploy

men

t42

,677

48,7

5957

,972

65,9

7382

,513

107,

015

147,

196

183,

241

219,

079

262,

994

286,

584

234,

800

1,73

8,80

321

5,38

11,

523,

422

Tot

al Fir

ms

1,78

92,

013

2,12

92,

143

2,06

42,

267

2,55

32,

867

3,13

03,

436

3,70

33,

450

31,5

448,

074

23,4

70E

mpl

oym

ent

439,

474

486,

146

510,

035

546,

588

600,

585

681,

251

799,

347

936,

825

1,04

3,48

31,

195,

371

1,30

7,98

21,

081,

526

9,62

8,61

31,

982,

243

7,64

6,37

0

Perc

enta

ge O

ver

Res

pect

ive

Tot

alG

arm

ents

Fir

ms

16.2

117

.39

18.4

618

.62

19.9

623

.03

25.4

627

.42

28.9

830

.12

30.2

227

.77

24.7

917

.74

27.2

2E

mpl

oym

ent

9.71

10.0

311

.37

12.0

713

.74

15.7

118

.41

19.5

620

.99

22.0

021

.91

21.7

118

.06

10.8

719

.92

Gro

wth

Rat

e (P

erce

nt)

Gar

men

tsF

irm

s—

20.7

12.3

1.5

3.3

26.7

24.5

20.9

15.4

14.1

8.1

�14

.411

.511

.212

.8E

mpl

oym

ent

—14

.318

.913

.825

.129

.737

.524

.519

.620

.09.

0�

18.1

16.8

15.6

16.1

Tot

al Fir

ms

—12

.55.

80.

7�

3.7

9.8

12.6

12.3

9.2

9.8

7.8

�6.

86.

26.

27.

6E

mpl

oym

ent

—10

.64.

97.

29.

913

.417

.317

.211

.414

.69.

4�

17.3

8.5

7.5

8.8

Sour

ce:

Aut

hors

’ est

imat

es b

ased

on

INE

GI:

<ht

tp:/

/ww

w.in

egi.g

ob.m

x>.

a Incl

udes

maq

uila

dora

act

iviti

es.


such as automobiles and electronics. Exportsare furthermore concentrated in a relativelyfew large firms. Illustrative of this pattern, for‒ the principal three hundred exportingfirms and maquiladoras were responsible, onaverage, for percent of Mexican exports.1

The pattern creates two fundamental prob-lems, one of macroeconomic management andthe other of social cohesion. The problem ofmacroeconomic management results from thefact that as the country loses its domestic mar-ket, the propensity to import increases as aresult of growth in the gross domestic product(GDP); expansion produces a growing deficitin the country’s balance of trade that must besustained by an inflow of foreign capital. Thismakes the country highly vulnerable to thethreat of capital flight and periodic foreign-exchange crises of the kind that erupted mostrecently and dramatically in December .

These crises are managed by severe cutbacks indomestic demand and rising unemploymentthat, in turn, threaten social coherence.

The changing structure of industry also hasa direct effect through its impact on opportu-nities for social mobility. This is particularlytrue in clothing. The traditional garmentindustry is a cascade of operations, each ofwhich can be, and in practice is at one time oranother, separated off and subcontracted, cre-ating almost a continuum of firms arranged ina hierarchy of skill, power, and profitability.At the bottom of that hierarchy are firms thatdo simple sewing on the cheapest garments,often as home workers. Toward the top are arange of firms that actually design the gar-ments and cut the material into pieces that aresubcontracted for sewing, again arranged in ahierarchy of price and quality. At the peak arefirms that wholesale and retail the garments,often in combination with design. In the UnitedStates the latter tend to be large companies ofthe kind that are now entering into maquila-dora production in Mexico, but in Mexico, as

in France and Italy, many small producersown, or at least owned, a couple of retail out-lets. When all the elements of the structureexist in close geographic proximity, it is possi-ble to start at the bottom as an unskilled homeworker sewing cheap garments and work one’sway up the hierarchy, gradually acquiringmore skills and business sense and contactswith progressively higher levels on the chain.In our interviews in Mexico City we encoun-tered several family firms that were the prod-ucts of this process: The proprietors had be-gun their working lives helping their motherswith piecework at home.

The new kinds of subcontracting relation-ships between Mexican producers and foreignbuyers typically cut off the chain of subcon-tracting in Mexico at both ends: The span ofcontrol along the subcontracting chain is con-sidered too long for the quality and reliabilitythey are seeking, and they limit production tothe Mexican partners’ own facilities. At thesame time, they absorb the design and market-ing links of the chain. The result is to create asharp divide between workers and contractorsthat can be bridged only by people with accu-mulations of capital and industrial expertisethat a typical worker could never hope to ac-quire on the job. As Mexican firms lose designand marketing capability they also becomeincreasingly dependent on foreign partners,and in that sense mobility, even for those withcapital and expertise, is limited as well.

Methodology

The study is organized around the concept ofa commodity chain (Gereffi ), or, as it iscalled by other authors, a production chain orsupply chain (Fine and Whitney ). A com-modity chain consists of a series of linkagesstretching from raw-materials production atone end through manufacture and assembly to


wholesale and retail distribution at the otherend, and it generally encompasses importantsegments of a limited number of interdepen-dent industries. The process of industrial trans-formation can be understood in terms of therelationships along these chains. On any par-ticular chain certain points constitute leader-ship positions, and organizations that occupythese positions formulate strategy and drive thetransformation process. Leadership, however,varies over time and across industries. In auto-mobiles, manufacturing has historically driventransformation. In recent years retailers havedriven transformation in the traditional indus-tries that are the focus of this study (Gereffi). In this study we sought to map out thechains and the transformation process throughopen-ended interviews with key actors.

Because of their strategic importance in thegarment industry, we began interviewing retailmanagers, particularly managers in the dis-count retail chains that have proliferated inMexico at the turn of the twentieth century.We focused in these interviews on their expe-rience with local sourcing. The discount chainsare linked directly or indirectly to foreign com-panies that purchase in bulk throughout theworld. They thus constitute superhighways forthe entrance of foreign merchandise into theMexican economy, but they could as well serveas export channels for Mexican goods goingabroad. We then moved to interview Americancompanies buying from Mexican producers; arange of the Mexican producers, includingcompanies producing for exports as maquilado-ras and on their own account as well as com-panies focused exclusively on the domesticmarket; government agencies and nongovern-mental organizations (NGOs) concerned withthe promotion of the Mexican garment indus-try; and various other individuals and firmsoffering ancillary services to the industry.

The sample is in no sense random. Respon-dents were selected because of the strategic

importance of the place they occupied alongthe supply chain. Where possible, we used per-sonal contacts to obtain access. Although, as itturned out, that access was obtained in overhalf the cases through cold calls, respondentsclearly agreed to talk to us in many casesbecause of our credentials and the belief thatwe had useful contacts with government offi-cials or with potential customers. The closestour study came to a generally random selectionprocess was in Mexico City, where we selectedthe tallest building in the garment district,took the elevator to the top floor, and system-atically went from shop to shop seeking inter-views. The reception there was mixed, rang-ing from a three-hour interview in one shop toa three-minute exchange in another. We of-fered all our respondents confidentiality andanonymity; few, however, seemed to put muchcredence in that offer.

Overall, in the period from to ,

we interviewed managers in three discountretail chains that had recently opened in Mex-ico; three U.S. companies actively engaged inupgrading Mexican partners; nineteen Mexi-can clothing producers, five of which wereoperating as maquiladoras and two that de-signed for and sold directly to the internationalmarket; and one international consulting firmengaged in training personnel for “green-field” sites (i.e., investments in new plants,machinery, and equipment) in Mexico. Inaddition, we spent two days at a fair in Can-cún organized by Bancomext (Banco Nacionalde Comercio Exterior) to introduce U.S. buy-ers to Mexican garment producers. A total ofthirty U.S. companies and thirteen Mexicanproducers attended this event. We talked withmost of them informally and, in addition, ob-served one-on-one meetings between buyersand potential suppliers in which the formerevaluated the latter’s collections. We also metwith groups of local producers in Puebla andAguascalientes, which were essentially group

, , .


interviews. We met with leaders of the indus-try associations and state economic-develop-ment officials in both Puebla and Aguas-calientes, with federal officials in Mexico City,and with two NGOs working on upgradingsmall garment producers in different parts ofthe country. We also conducted, as explainedbelow, a separate study of empresas integra-doras. Material on garments is supplementedwith material collected separately from othertraditional industries.

Principal Findings

All the traditional industries, but especiallyclothing, are sensitive to fashion. This givesMexico the particular advantage, relative toother low-wage developing countries, of prox-imity to the U.S. market. The advantage iseven greater when producers are using U.S.materials that must be shipped into Mexicobefore the finished goods can be shipped out.(This advantage should be even greater still inthe domestic market.) The magnitude of thatadvantage is suggested by one brand-nameretailer who reported that shipment fromMexico to its Texas warehouse took four days,compared to thirty days from Korea. Anotherbrand-name retailer, a U.S. shoe company,estimated total time to market, from initialorder to receipt of the finished goods, at sevento eleven weeks in Mexico compared to four-teen to fifteen weeks in Hungary or Italy, eigh-teen in Portugal, and twenty-three to twenty-five weeks in Brazil, China, or Indonesia.

Against this advantage, the discount retailchains and American companies purchasing inMexico all identified a common set of obsta-cles to sourcing in Mexico. Mexican produc-ers were unable to meet quality standards; theycould not produce in sufficient volume; theirproduction cycle (or turnaround time) was toolong; and they failed to meet promised deliv-

ery schedules. These were all viewed as pro-duction problems, the legacy of the shelteredmarkets in which Mexican producers have tra-ditionally operated. They are distinct from theinexperience of Mexican companies with thecommercial practices and procedures involvedin selling internationally, which have been aproblem for Mexican companies seeking toexport for the first time. The American firmswe interviewed were all prepared to handle thecommercial problems for their Mexican sup-pliers, and commercialization was obviouslynot a problem in dealing with discount retailchains in Mexico itself. Therefore the centralquestion to emerge from the interviews is:Why haven’t Mexican producers been able tolearn how to meet international productionstandards? Or, since some Mexican producerscan meet these standards, how might Mexicanproducers be induced to learn faster or inlarger numbers?

The Nature of the Learning Process

An answer to this question is suggested by theexperience of American companies that havetried, with varying degrees of success, to de-velop Mexican sources. We interviewed sev-eral companies about this process. The com-panies were selected in an opportunistic fashionand not on the basis of a systematic survey.But we believe that companies actively en-gaged in upgrading their suppliers in Mexicoare relatively unusual. Mexican firms do nottypically engage in the practice of upgradingtheir suppliers themselves.

The impact of maquiladoras on the rest ofthe Mexican economy has been extremely lim-ited.2 The most extensively studied have beenthe automobile assembly plants (see, for exam-ple, Shaiken and Herzenberg ; Robinson). The plants of U.S. companies import


virtually all the parts that they use. Japanesecompanies have encouraged their home sup-pliers to locate around them in Mexico; thesesuppliers have not developed a second tier ofMexican contractors. Recently, as the cost ofproduction in Japan has increased, a specialeffort has been made to increase sourcing inMexico, but it has consisted almost exclusivelyof enhanced efforts to identify qualified Mex-ican producers, not to upgrade them. Infor-mation on other industries is more limited butis consistent with the automobile findings:Neither foreign firms operating in Mexico norMexican firms themselves have been particu-larly active in upgrading their supplier net-works. In this sense the discount retail chainsare typical.

The companies we interviewed, which didmake efforts to upgrade Mexican contractors,all managed brand names and all sourcedworldwide, purchasing in the United Statesand in a number of different developing coun-tries. They made it clear that in Mexico, as inmost countries in the world, few producers canmeet their standards initially. They thus madea substantial effort to develop new sources. InMexico this typically involves, first, compara-tive shopping in Mexico itself to find pro-ducers whose products meet some minimumstandards of quality at the outset. The U.S.company then visits the producer and inter-views the management to see whether thecompany has interest in and is capable ofupgrading its quality and producing in the vol-umes and time constraints that American pur-chasers typically require. This is a two-stageprocess that begins with an initial half-dayvisit. It is then followed by a whole-day eval-uation that serves as a diagnostic tool as wellas the basis for a business decision.

If the parties agree to go forward, the Amer-ican partner then undertakes to teach the Mex-ican company how to meet its standards. Thisinvolves a series of exchanges in which Mexi-

can personnel are virtually tutored by theirAmerican counterparts—sometimes in theMexican plants, sometimes at the facilities ofthe American customer in the United States,often in both places. One large shoe company,for example, when it began sourcing in Mex-ico, opened an office in Mexico City and hastwo engineers working out of that office per-manently assigned to each sourcing plant. Alarge clothing retailer reported that it takes atleast one and often one and a half years fromthe time it starts working with a potentialMexican partner to the time it receives its firstorder. To illustrate this, one retailer revieweda typical case: The process began with severalpreliminary visits of its personnel to Mexicoand of the potential partner to company head-quarters in the United States. Once upgradingwas begun in earnest, the process involved sixtrips of a three-person U.S. team to Mexicoand eight visits of a similar team of Mexicansto the United States, then heavy involvementof U.S. personnel in the initial productionruns in Mexico.

The learning process here involves what isknown in the literature variously as practical,implicit, or tacit knowledge. Its essential char-acteristic is that it is difficult to transmit ver-bally or in written instructions and instead itis taught by demonstration on the job as pro-duction is carried out. The U.S. garment firm,for example, in a process reminiscent of whatin England is called “sitting by Nellie,” has itsown people work side by side with the inspec-tors and watch what they are doing, picking upthe faults that the new inspectors miss andpointing out to them, case by case, what iswrong with the garment.

Historically, managerial theory and ad-vanced management practice have paid littleattention to knowledge of this kind. For Amer-ican manufacturing in particular, a sharp dis-tinction was made between formal and infor-mal engineering, and management looked only

, , .


to the former for improvement. But since themid-s, under the pressure of heightenedcompetition, particularly from Japan, thepriority accorded to formal knowledge hasbeen abandoned. A number of the techniquesborrowed from the Japanese or developed inresponse to the pressures of Japanese com-petition, such as total quality control and thekanban system of on-time delivery,3 are es-sentially ways of deliberately managing tacitknowledge, making it explicit, subjecting it todebate and discussion, and forcing progressiveimprovements in production processes (No-naka ). Part of what Mexican firms arerequired to do is thus not so much to learn astandard set of practices as to catch up with amanagerial revolution that has been occurringin industrialized countries only recently andeven there is far from complete.

In other ways this new emphasis on tacitknowledge is a competitive advantage for Mex-ico. It places an enormous premium on expe-rience in the industry. It values the knowledgethat comes out of growing up within an indus-try. As a result the existing skill within the tra-ditional industries of Mexico constitutes aconsiderable human capital. That skill is, how-ever, an asset specific to the industries in whichit resides; it will be lost if those industries failto make the transition and the resources aredispersed elsewhere in the economy. More-over, to make the transition and become com-petitive in world markets, this existing capitalneeds to be combined with modern managerialtechniques. Finally, the process of introducingthose techniques clearly involves a substantialcommitment on the part of both the Mexicansuppliers and their U.S. customers; it takesresources and its takes time to upgrade Mex-ican facilities. The latter seems to range froma year to a year and a half.

Because it takes time and resources, theprocess of upgrading is clearly an investment.But the investment is basically one of skill

transfer of a particular kind. The transfer mustbe made directly from the foreign client to theMexican contractor. Once transferred, the skillsare embedded in the ongoing practices of theorganization; they reside in the contractor, andif the contractor walks away from the relation-ship, he or she takes the skills along. Unlikeplant and equipment expenditures, there is nophysical asset that can be used to secure theinvestment and reprocessed if the contractorreneges on any agreed-upon payments. To theextent that the skills are particular to a givenclient and of no use in other contracting rela-tionships, there is little reason for the contrac-tor to walk away. But most of the skills are quitegeneral; there is inevitably a specific compo-nent, but typically the skills increase the capac-ity of the contractor to produce quality goodsefficiently for any client or, for that matter, forsale directly on the market. Thus the Mexicanfirm, once upgraded by its foreign client, hasevery incentive to jump ship and sell its newlyacquired skills to the highest bidder.

We encountered two cases in our interviewswhere the Mexican partner had apparentlydone this. One blue jeans contractor had beentrained by an American company with whomhe initially had an exclusive agreement, butwhen we interviewed him, he had abandonedthat relationship to work for a number of dif-ferent U.S. companies and was about to launchhis own brand. In the second case a U.S. shoecompany reported that it had acquired one ofits Mexican contractors by persuading the firmto leave the company originally responsible forupgrading its facility. In several other plants wevisited the company was obviously thinkingabout taking off on its own. Why, then, wouldclients ever make investments in upgradingcontractors in Mexico?

One possible answer is that the contractorrepays the client-tutor by charging pricesbelow the market value for the goods that itprovides during the learning period. This is


not generally true. In the cases we studied, nomerchandise was exchanged until after thecontractor had learned how to produce to theclient’s standards. But it is possible that someupgrading arrangements are financed in thisway. The transactions here are so complex,however, that it is possible that they aresecured in other ways. The upgrading is notnecessarily limited to tutelage. The Mexicanpartner is sometimes required to make com-plementary investments in plant and equip-ment. In several cases the American customersrequired their contractors to set up physicallyseparate facilities for the export portion of thebusiness in order to segregate exports from theoverhead associated with commercialization ofmanufacturing production in Mexico itself.The partnership generally includes access tomaterial supplies at favorable credit terms andoften to credit itself, which is a considerableadvantage to Mexican producers given thehigh interest rates and general shortage of cap-ital that have accompanied the opening of theMexican economy. Indeed, at real interest ratesranging as high as percent, this backdooraccess to the U.S. short-term credit marketmay be the most valuable part of the relation-ship for the Mexican partner and the biggestdeterrent to jumping ship.

If arrangements of these kinds were ableto solve the investment problem, one wouldexpect tutelage to be widespread, whereas it appears, as noted above, to be extremelylimited. Whatever forms of security can beworked out in these ways, they are evidentlynot enough to diffuse the tutelage arrange-ments broadly. What is it about the companieswe encountered that enabled them to overcomethe problems that seem to deter other firms?

We offer several conjectures on this score.4

The most plausible is linked to the character-istic that appears to distinguish these compa-nies from others engaged in outsourcing inMexico. The American companies we inter-

viewed are all brand-name producers with aworldwide sourcing strategy. Brand identifica-tion enables them to sell their product at pre-mium prices and thus generates an economicrent. That rent can be shared with contractorsin the form of favorable fees, thereby bindingthe contractors long enough to enable the com-pany that provides training to earn a return onits investment. A global sourcing strategygenerates further returns. In these strategiesMexican sourcing serves to diversify risk. Inaddition, the short turnaround time relative toother foreign locations enables the U.S. com-pany to balance its product line by includinga high-fashion component that attracts cus-tomers who then purchase other parts of thecollection. Without a nearby supplier the turn-around time would be too long to keep up withthe market. These returns are also a kind ofeconomic rent that can be shared to bind thecontractors to the tutoring company.5

The second conjecture rests on the fact thatthe knowledge about how to upgrade produc-ers in low-wage economies is a relatively recentinnovation. The companies we interviewed inMexico were all pioneers in global sourcing.Their strategy in this regard is new, developedover the last ten to fifteen years to take advan-tage of the low wages prevailing in developingnations in order to service the markets of ad-vanced industrial countries without becominghostage to the political and commercial risks ofthe extended supply chains this entails. Othercompanies sourcing in Mexico, to say nothingof Mexican companies that buy from local con-tractors, simply may not have the skills re-quired to upgrade their supply networks, andthe skills may not be generally available on themarket. This, rather than the difficulties ofsecuring the investment, may explain why par-ticular companies and not others are engagedin upgrading contracting networks.

Still another possible explanation is thatwhat those companies offer to their suppliers

, , .


is not a single set of techniques but rather con-tinuous access to state-of-the-art manufactur-ing production as it evolves over time. Again,their global sourcing strategy should put themin a unique position to do this. It enables themsystematically to benchmark and compare prac-tices across a wide variety of producers, to or-chestrate a competition among them, to selectthe best practices, and to diffuse these rapidlyacross their contracting network. Such tech-niques for the management of supply networksare part of the repertoire of techniques for themanagement of tacit knowledge that have de-veloped toward the end of the twentieth cen-tury and that are now widely applied in ad-vanced industrial countries. But they are notuniversal even in the United States and West-ern Europe, let alone in relationships that spanborders and countries at very different levelsof economic development. Some of the prac-titioners of these new techniques encouragetheir contractors to work with several clients,even competitors, thereby stretching the con-tractors’ capacities and generating a wider rangeof approaches to feed into the fund of alter-natives that the mother company is able sys-tematically to compare in order to generatecontinual improvements over time. Thus theapproach does not necessarily require Mexicancontractors to work exclusively for the clientswho initially upgraded their facilities, and oneof those clients whom we interviewed confirmedthat it did not seek exclusive relationships.

If this is what is going on in the companiesupgrading Mexican facilities, however, thecapacity of a Mexican firm to compete inter-nationally once it does jump ship must deteri-orate progressively over time unless it man-ages quickly to hook up with a new foreignpartner. The practices observed in the onecontractor that had become cut off from hisoriginal American partner suggested that thismight be the case. This is also suggested bythe fact that the American partners whom we

interviewed continue to station their own per-sonnel in the Mexican partners’ facilities evenafter the initial training period and regularlysend additional personnel for random qualityinspections at the production site.

Minimum Order Size

In principle the problem of quality and effi-ciency within productive establishments canbe separated from the issue of minimum ordersize, which many clients and particularly thelarge discount retail chains cited as reasonswhy they did not source locally. To solve thesecond problem, many of our respondentssuggested association arrangements in which anumber of producers pooled their resourcesto take on a large order.

The government has recently created a newinstitutional structure, empresa integradora,designed to house such arrangements and facil-itate their development. This seems a promis-ing approach to the problem of minimum ordersize, but this organizational form has been slowto take off, and few such integradoras actuallyexist in Mexican manufacturing. To find outwhy, we conducted an in-depth study of twelveempresas integradoras in Cuernavaca, Puebla,Jalisco, Mexico City, and Tijuana. Regionalcultural factors, the education and training ofthe entrepreneur, and the availability of finan-cial resources were all found to be critical tointegradora success. Established cooperativerelationships between manufacturers and thedecentralized division of labor among manu-facturers also appeared to facilitate success.One integradora grew out of an association thatfor years had worked together at trade fairs andbought fabric together. The division of laboramong manufacturers involving marketing, in-spection, and other tasks serves to decentralizeauthority and to enhance trust among mem-bers of the integradora.


Although some small producers have formedempresas integradoras, these efforts have facedmany obstacles and relatively few such associ-ations exist in Mexico. Among the most im-portant challenges faced by integradoras arebuilding a culture of trust, gaining access tocredit, and overcoming bureaucratic barriers.Many small producers are reluctant to enterinto such associations and generally share lit-tle information about their sources of fabricand other production issues. Accountability isalso an important issue. A particularly dra-matic example of this problem was one largeintegradora outside Puebla. As one home-basedmanufacturer explained, each producer paidthe salary of a coordinator who later ran offwith all the money.

Empresas integradoras also encounter diffi-culty in gaining access to credit and findbureaucratic obstacles when they seek toexport. Like other garment manufacturers,integradoras face extraordinary interest ratesand payment cycles that lag behind loan sched-ules. The integradoras often compound ratherthan simplify credit problems because of thereluctance of financial institutions to lend tosuch associations. As one manufacturer ex-plained, “There were complications in lendingto five long-standing businesses. Someonewould have to put up their house and becomethe leader, which we didn’t want.” Finally,integradoras have experienced delays in gettingexport authorization because of the lack ofcoordination of government programs.

In the garment industry, however, the focuson the limits of the government’s integradoraprogram may be misplaced. It is after all stan-dard practice, not only in Mexico but through-out the world, for a “jobber” to meet largeorders through a network of subcontractors.The jobber is, in other words, already func-tioning as a kind of integradora. Thus the job-ber could upgrade its suppliers in the same waythat some foreign retailers work with larger

suppliers. If it is possible to upgrade these job-bers’ networks and maintain standards withinthem, it may not be necessary to develop newcontracting institutions. In this sense any setof policies that manages to diffuse the tutelagearrangements that exist between foreign clientsand Mexican contractors would also resolve theproblem of minimum order size.

Consultants

This leads to the question of why Mexicangarment firms have been so reliant on theseforeign partnerships at all. Why can they nothire consultants to help upgrade themselves?Indeed, not all foreign firms rely on their ownpersonnel to develop production facilities orcontractors abroad; a number use consultingservices. We identified and interviewed onesuch firm in the garment industry. Among itsother services it offered training in both pro-duction and management for shops in thedeveloping world seeking to export. The firmwill staff and train the personnel of a new pro-duction facility from scratch. Its program fordoing so has strong parallels to the in-houseprograms we encountered. It first hires a cadreof managers. In Mexico, interestingly, it drawsfor this purpose primarily on people whostarted but, often for financial reasons, wereunable to finish a technical education. Theconsulting firm uses its own personnel to trainthe managers in production techniques andthen hires the production workers for the newfacilities. The managerial candidates under thesupervision of the consultant’s personnel thentrain the production workers. At the same timethe managers in the new facility receive spe-cial functional training, including a classroomcomponent. All production training is on-the-job, using a variant of the tutelage wedescribed earlier. Our respondent estimatedthe total time needed to launch a new factory

, , .


in Mexico at six months to one year, which issomewhat shorter than the in-house programsdiscussed above. Although the source of thediscrepancy is not clear, the standards of effi-ciency and quality may not be identical; thetype of product may also vary. Most of ourrespondent’s clients seem to be multinationalcompanies in the United States producing rel-atively standardized products with limitedfashion content, but they claimed to offer thesame services to Mexican producers for anytype of clothing. We did not, however, findMexican firms using this type of service.

We were more successful in the furnitureindustry, where we found an association offirms in Ciudad Hidalgo, a relatively remotecity in Michoacán that had hired a consultantto help upgrade the quality and efficiency oftheir operations. We visited the city some timeafter the consultant, who had provided exten-sive advice on how to upgrade the quality ofthe product line and the efficiency of the pro-duction facilities, and we interviewed in-shopthe proprietors of several of the enterprisesabout what they had learned. It was clear inthese interviews that the people in these shopshad changed their practices at the consultant’sbehest, but they had essentially learned thenew practices by rote. They had no idea of theunderlying principles from which the consult-ant was working. This, in turn, reflected thefact that they had never seen the kinds ofproducts with which they were competing inthe international marketplace, which the con-sultant was using as a template to improvetheir own. Nor had they seen the foreign shopswhose practices the consultant was trying toget them to adopt. Thus one shop has re-designed the work flow on the consultant’sadvice, but aisles were clogged with work inprogress that completely undermined therationale for the streamlined plant layout it hadintroduced. In another shop, the proprietorshowed us how the consultant had suggested

they turn the knots on the wood to the inte-rior of the cabinet to improve the outside fin-ish, but he then indifferently forced the lockand bent the key on his model piece when heopened it up so that we could feel the knots onthe inside pieces of wood.

These experiences with the consultant in thefurniture industry led us to believe that a keyingredient in upgrading partnerships in thegarment industry is the visit of Mexican per-sonnel to the partners’ facilities in the UnitedStates that precedes the visits of the partners’personnel to the Mexican facilities. It seemedthat the consultant would have been muchmore successful if he had first put his clientson a plane and flown them to the United States,or even to Mexico City, to see and discuss theproducts with which they were competing andto visit production facilities on which his advicewas modeled. Indeed, he might then have beenable to teach his clients not only how to do whatthe foreigners were doing but how actually tothink through and critique their own practicesthemselves. This approach might be attractivenot simply for the traditional firms that havebeen left out of the export boom but even forMexican producers that have found foreignpartners. This would be especially true if, assome of our conjectures about the tutelage pro-cess suggest, what Mexican producers are get-ting from their foreign partners is simply themost up-to-date production practices, not theskills of their foreign partners that the economyreally needs to survive in international compe-tition on its own—that is, the capacity for con-tinuous improvement in practice over time andto assume the tutelary role vis-à-vis their ownsubcontractors.

The other factor that is involved in the de-pendence on foreign partners for learning, asopposed to hired consultants, is credit. The dif-ficulties of securing investment in tacit knowl-edge that limit the willingness of foreign part-ners to invest in upgrading Mexican facilities


also make it difficult for the Mexicans to obtaincapital to invest in themselves. This problemhas been greatly aggravated by the generalshortage of working capital and the extremelyhigh interest rates that have accompanied theopening to trade, even before the peso crisis inDecember and much more so afterward.At the same time the credit crisis increases theadvantages of a foreign partner enormously, ifyou can find one, because one then has accessto the partner’s suppliers in the United Stateson favorable credit terms. Indeed, several ofour interview respondents suggested that theycould obtain working capital through foreignpartnerships on relatively favorable terms attimes when such capital would not be availableon any terms in Mexico. The extreme exampleof what the capital shortage was doing was onesmall producer who was reduced to buying justenough material in the morning so that hecould produce a day’s output, sell it in theevening, and have enough money to buy mate-rial for another day’s production. Such prac-tices foreclose economies of scale in purchas-ing and production altogether.

In principle, these credit problems call for an“investment subsidy” or a specialized loan pro-gram. But such a program would not be easyto administer, especially in Mexico. Applicantswould have to be screened for eligibility andthen monitored afterward. The general scarcityof credit promotes a strong incentive to divertfunds to other purposes, and without collateralit is difficult to penalize such diversions. Thedifficulties here are compounded by the natureof small firms in the garment industry and theMexican banking system. The garment indus-try is populated by family firms in which thehousehold and business accounts are oftenintermingled and confused. It requires a stronglocal banking system with roots in the commu-nity to distinguish the viable firms and judgethe integrity of the enterprise. But the Mexi-can banking system has passed through a

process of nationalization and reprivatizationthat has left the industry centralized in Mex-ico City, without locally oriented branches.

Bootstrapping

Given what an investment subsidy designed todiffuse foreign practices appears to entail, it isworth considering a much more broadly con-ceived policy to actually develop the requisitecapacities within the Mexican economy, with-out foreign intermediaries, by what one analysthas termed “bootstrapping” (Sabel ).Could a developing country such as Mexicoactually discover or invent world-class man-agement practices for itself ? The reason tothink it might be able to do so is that develop-ment of the skills at stake here has not histor-ically taken place through tutelage arrange-ments. Rather, these skills emerged first in theefforts of the Japanese economy to catch upwith the West in the aftermath of the SecondWorld War. Japan entered the postwar periodwith a reputation for cheap, second-rate man-ufactured goods, not unlike that of Mexico’straditional industries today. It managed inthe s to set a course of development thatby the s had made it preeminent in theefficient production of high-quality mass-produced goods, rapidly gaining share in thehome markets of its erstwhile competitors inthe United States and Western Europe. In thes these Western competitors then soughtto meet the Japanese challenge by appropriat-ing the techniques the Japanese had inventedin order to catch up and use them to recapturetheir original lead. In both episodes of com-petitive transformation, foreign practices playedan important role, but in neither case was theprocess essentially one of direct transfer offoreign practice.

The latest round of transformation in theUnited States and Western Europe has had

, , .


three key ingredients. First, companies devel-oped a set of standards and benchmarks toidentify concretely where their performancewas deficient. Second, they sought to identifythe precise institutions and practices that dif-ferentiated the benchmark procedures andpractices from their own. In the attempt to doso, they occasionally went as far as to establishjoint ventures with Japanese partners in orderto get firsthand exposure to their ways of doingbusiness. But they did not ever slavishly imitatethe Japanese. Instead, and this is the third ofthe key ingredients, they initiated a series ofinternal debates and discussions about what thecritical elements of Japanese practices were,whether these could be adopted whole, and, ifnot, how they might be altered to fit into theirown organizational practices. When it was notpossible to identify precise procedures usedelsewhere, they nevertheless sought to inventapproaches that might produce the desiredresult. The new practices and procedures thatconstituted the revolution in Western manage-ment in the s were not those actually bor-rowed from Japan but the practices and pro-cedures invented to facilitate the borrowing,namely, discussion and debate structured by aset of benchmarks and standards on the onehand and a set of alternative institutions andpractices on the other. These are what consti-tute the new techniques for managing tacitknowledge. They are basically the techniquesforeign retailers are applying to develop andmaintain the global sourcing networks thattheir Mexican partners are being drawn into.

U.S. manufacturers in the s generallysought to catch up with their Japanese com-petitors as rapidly as possible, in a single spateof institutional reform. It was only relativelyrecently, after they had bridged the initial gap,that they began to think in terms of continu-ous improvement, using the same proceduresand benchmarks or, when they are already inthe vanguard, standards and targets to stay

ahead of the game. By contrast, the Japanesein the postwar period had recognized that theycould not catch up in one sudden transfor-mation and sought instead to raise their per-formance gradually over time. For these pur-poses it is important not simply to have notsimply a single standard or set of benchmarksbut rather to think in terms of a hierarchy ofstandards that the practice can ascend gradu-ally over time. This hierarchy of standardsneeds to be matched to a typology that dividesthe market into segments that the firm canmove across as its standards rise. Mexico’sposition in international competition is closerto Japan’s in the s than to that of theUnited States and Western Europe in thes, and this idea of a hierarchy of stan-dards and markets would seem an importantaddendum to the North American approach.A number of people with whom we talkedwere already thinking in these terms: A Japa-nese government official working to increasethe backward linkages of the Japanese auto-mobile assembly plants used a three-tier sys-tem to rate potential Mexican suppliers; U.S.companies looking for contractors in Mexicouse a similar system. But in such a system, itdoes not appear that a policy designed to stim-ulate a bootstrapping process would be muchmore difficult to initiate than one more nar-rowly focused on investment subsidies.

International Matchmaking: An Illustration

The limits of government policy are illustratedby one particular program we were invited toexamine closely, a program managed by Mex-ico’s Foreign Trade Bank (Bancomext) in part-nership with the Ministry of Commerce (Sec-retaría de Comercio y Fomento Industrial, orSECOFI) to link Mexican producers with out-side clients. The program was conceived as a


matchmaking operation, in which buyers frommajor U.S. and European department storeswere invited to Mexico to meet with potentialsuppliers. This program was run for severaldifferent industries. In the garment industry,the first meeting was held in . The Mex-ican producers brought samples of their mer-chandise, and the buyers set up booths wherethey met with the producers individually toexamine and criticize their products. Enor-mous effort was put into the organization ofthe meetings; the then secretary of commerce,who was also the chief Mexican negotiator forNAFTA, actually called the chief executiveofficers in the United States to urge them tosend representatives. But virtually no effortwas put into evaluation and follow-up. No onereally knows whether Mexican companiesmanaged to obtain any business from this exer-cise and, if not, why they failed to do so. It iscompletely unclear whether the meetings werea successful policy initiative and, if not, whatprecisely could be modified to make themmore successful.

The program was nonetheless administeredagain, in October , in essentially the sameway in which it was administered in . Thistime, however, there was considerably morediscussion and evaluation of the results. Sev-eral of the conclusions that emerged are worthemphasizing, partly to illustrate what was lostby failing to reflect on the experience the firsttime around but also because they feed intothe specific policy recommendation we areabout to put forward. The first conclusion isthat the large U.S. chains that were the focusof the first two meetings are the wrong tar-gets. Their standards of quality and minimumorder sizes are too far out of reach of the bulkof Mexican producers. The Mexican industrycan do better by targeting buyers from otherLatin American countries, whose levels ofincome and taste are closer to its own, andsmaller (but somewhat obscure) retail chains in

the U.S. that order in lesser quantities. Thesecond conclusion is that the promotion ofMexican products should focus on areas witha distinct national style, such as Mexican hand-icraft styles or formal garments for children(baptismal and communion dresses, for exam-ple). The third conclusion is that the kind ofMexican producers most likely to benefit fromprograms of this kind are unable to meet thenew orders without access to working capitaland hence that, to be effective, these match-making operations need to be supplementedby programs providing short-term credit tosmall enterprises. Bancomext developed a pilotcredit program for a group of producers thatobtained orders at the meetings from aColombian department store.

The Bancomext example suggests that thefirst step toward an effective policy is a newapproach to thinking about policy itself. In asense, what is required is to introduce into themanagement of government programs thosetechniques for managing and systematicallyupgrading practical knowledge that haveemerged in manufacturing production. But aprior task is to create a wider space for a prin-cipled approach to industrial policy, to artic-ulate a philosophy of government that, whilemore active and interventionist than theframework that currently dominates govern-ment thinking, cannot be reduced to tradi-tional clientelistic actions.

Toward a New Philosophy ofIndustrial Policy

Our examination of the problems of the cloth-ing industry suggests that a principled ap-proach to industrial policy might be builtaround three basic suppositions. The first ofthese would preserve the basic insight of neo-liberal thought by recognizing that the marketis a powerful instrument both for motivating

, , .


economic activity and for coordinating and di-recting the allocation of scarce resources, andthat economic science provides a way for under-standing how the market works toward theseends. Second, it must be recognized that, what-ever the ideological attractions of a marketeconomy, the scientific case for its effectivenessin no way precludes the interventions of theMexican state. This is because the unregulatedoperation of the market leads to a particulardistribution of income and power in the soci-ety that is not inherently just or necessarilycompatible with long-term social and politicalstability. This important caveat to the neo-liberal argument for market-oriented economicpolicies must be distinguished from the thirdpoint: A separate and distinct rationale for stateaction lies in the considerable difficulty in fullyunderstanding how a market economy operates(in theory, let alone in practice).

What we do understand implies that aneffective market economy must be supportedby a set of supplementary institutions and thateven when those institutions are in place therecan be significant instances of market failure,as appears to be the case, for example, in thetransfer of practical knowledge that we havebeen examining. These principles suggest anapproach to policy that is guided by the mar-ket and instructed by developments in the pri-vate sector without being completely depen-dent on the market to produce desirable resultsor necessarily acquiescing to market develop-ments. They imply as well that, in public pol-icy no less than in the production and com-mercialization of goods and services, constantdiscussion and reevaluation of practice mustsupplement theoretical economic knowledge.

Toward an Alternative Policy

What might an alternative policy look like?First, it should be conceived as an effort to

extend the process of adjustment already tak-ing place in the private sector. Second, it mustbuild on mechanisms for evaluation and learn-ing as well as pressures and processes designedto produce continual improvement over time.Third, it should build on the experience of andborrow mechanisms developed for this pur-pose since the mid-s in the laggard sectorsof advanced industrial countries that have beentrying to catch up with their competitors inthe international marketplace.

These general principles, when applied toa policy designed to bootstrap traditional in-dustries in Mexico, suggest an approach thatfocuses less on specific sets of government pol-icy initiatives and more on the role of govern-ment in catalyzing discussion and debate. Thebasic goal, in other words, is to develop aheightened public awareness of the need toupgrade the productive apparatus and com-mercial practices throughout Mexican society.More than any particular policy measure, theidea is to orchestrate a national discourse; todraw as many people as possible from a broadspectrum of the society—from the worker onthe plant floor to the politician in the legisla-ture—into the enterprise of making Mexicomore competitive at home and in the interna-tional marketplace; to generate a critical per-spective on productive and commercial prac-tices in the business and political community.The aim should be to focus discussion anddebate as much as possible on practice andaway from ideology and abstraction. Models ofhow to do this include the case method usedin business and legal education, grand roundsin medical education, and the design studio inart and architecture, in which students areassigned a particular problem and their solu-tions are then criticized by a jury of faculty. Aparticular example of how this might be done,one that might serve to initiate the process, isto invite state development agencies to a seminarin which each agency presents for discussion


and debate two case studies, one of a majordevelopment success and one of a develop-ment failure. Industry chambers, particularlyin traditional industries, could be encouragedto sponsor similar seminars in which each localchamber is asked to work up and present onecase of a rapidly developing enterprise (or con-tracting network) and one case of a decliningenterprise or network.

As part of the effort to focus and direct thedebate, the government should encourage thedevelopment of standards and benchmarks.These provide both a target for policy and thecriteria for judging its success. The federalgovernment might do this by requiring thatany project it funds build in a set of standardsto serve as a threshold for admission to theproject, as well as a second set of standardsthat serve both as a program target and a setof criteria for evaluating the outcome. Thestandards might in principle focus on out-comes—for example, delivery time, quality,efficiency, and the like. But standards shouldalso focus on processes, such as inspection,inventory control, quality control, quality cir-cles, and so on.

The process of generating these standards,the debate about what appropriate standardsand benchmarks are, is at least as important asthe standards and benchmarks themselves.An example of the kind of standard-settingprocess that needs to be encouraged is Gua-najuato , which the footwear chamber inthat state created as a threshold that firms hadto achieve to gain access to a set of state-rundevelopment programs. There is now a debateat the national level as to whether this stan-dard should be extended to the shoe industryas a whole or whether other states should beencouraged to develop their own standards. Athird alternative would be to use as a nationalstandard the International Organization forStandardization (ISO) of the EuropeanEconomic Community.6

The kind of debate that is emerging aroundGuanajuato is the key to the policy we areproposing. The debate is actually more impor-tant than the particular way in which the issueis resolved. If properly orchestrated, it willforce the participants to reflect on practice.Nevertheless, the outcome of the debate maynot be irrelevant; there is a lesson here too. Wetend to think, as suggested earlier, that it isimportant to avoid a single set of absolutestandards. The relevant standard depends verymuch on which segment of the market theindustry is targeting at any moment. The stan-dard should shift upward over time as thecountry develops or with technological ad-vances. Standards should thus be a movingtarget. And a variety of standards at anymoment will help to pick out benchmarks andcall attention to alternative practices. The factthat Guanajuato has set a standard differentfrom ISO means that ISO firms canserve as a source of ideas for where the indus-try might move next. Were Guadalajara todevelop a higher standard than León, practicein Guadalajara could become a benchmark forfurther upgrading. The León standard, theGuanajuato standard, and the ISO stan-dard would then constitute a hierarchy acrosswhich firms or contracting networks mightthink of moving over time.

The development of standards needs to beaccompanied by a parallel effort to develop atypology of market segments that can then beset alongside the hierarchy of standards toguide industrial strategy. This is the broaderlesson embodied in the Bancomext insight thatMexican producers are more likely to findmarkets at this time in Latin America oramong smaller retail chains in the UnitedStates than in the prestigious New York de-partment stores at which their developmentprogram was originally directed. Divorcedfrom this broader lesson, the Bancomext pol-icy is likely to trap the industry in a low level

, , .


of development. But linked to a typology ofmarkets and a hierarchy of standards, it be-comes a way station in a strategy for the grad-ual upgrading of the productive system overtime. The development of market typologiescan be fostered, like the development of stan-dards, by requiring that a market analysis bebuilt into any development project that thefederal government funds. Such an analysisshould identify the segment of the market towhich the targeted enterprises are currentlycatering and the segment toward which theproject is designed to help them move.

Conclusions

It is useful to return in conclusion to the cen-tral theme of the paper: There is a growingdivision within the Mexican economy between,on the one hand, a relatively small group ofproducers that have managed to adjust to theopening of the economy to trade and are pros-pering in the newly created North Americanmarket and, on the other, a large group ofsmaller producers that have been unable tomeet international standards of quality andreliability and are floundering even in theirown national marketplace. The garment in-dustry is thus in many ways symptomatic ofthe Mexican manufacturing sector: The rap-idly expanding subcontracting industry dom-inates the aggregate statistics and makes theindustry the outstanding success, at least fromthe Mexican point of view, of the NAFTAstrategy, but it masks the stagnation anddecline of the smaller, traditional producersand the progressive loss of the domestic mar-ket to imports. In an economy with significantexcess labor reserves, there seems no reasonwhy the second development pattern shouldfollow from the first, especially in a traditionalsector such as garments, which is extremelylabor-intensive and has a fund of tacit knowl-

edge embodied in a skilled labor force and acadre of managerial and technical experience.In garments at least Mexico should be able toexpand its exports through subcontractingrelationships and retain its domestic market. Itbecame apparent early in this study that itsinability to do so is associated with problemsof quality and reliability within the traditionalsector, and we looked for clues among firmsthat had successfully overcome these prob-lems—largely with the help of an Americanpartner—as to how the lagging firms in theindustry might do so.

Ultimately we arrive at two rather differentsolutions. One is to take the upgrading processin the successful firms as a model and to try totransfer or extend it to the lagging sector. Themodel seems to have two salient characteristics.One is how the foreign partner works as atutor to its Mexican contractors. An extensionof this model would presumably look for con-sultants to play this role. The other character-istic is the investment in intangible assets andthe difficulties of securing such investmentswhen they are made by parties other than thosein which the newly transferred knowledge re-sides. The importance of the credit implicitin these arrangements has been augmentedby the general shortage and high interest costof working capital in Mexico and by howmaquiladora-type arrangements facilitateaccess to working capital in the United States.A direct attempt to extend this model through,for example, a government development pro-gram would thus concentrate on the provisionof consulting services as a substitute for therole of the foreign partner and on special loanprograms to overcome the capital constraintsthat small producers appear to face. The dif-ficulties with implementing such a programand the limitations of upgrading through con-sultants lead us to consider a second strategyof bootstrapping, in which the laggard firmsare encouraged to upgrade themselves through


a process of self-criticism and self-examina-tion in light of visits to best-practice facilitiesand benchmarks that measure the gap betweenbest and prevailing practices.

The bootstrapping strategy might actuallybe better suited to the traditional garment sec-tor than the tutorial approach in maquiladorafirms. The traditional sector, as noted, consistsof a long chain of subcontracting relationshipsthat stretch from the design and cutting roomsbackward to progressively smaller shops and,ultimately, to home workers. Historically therehas been considerable mobility along thischain, with pools of people at each stagethinking strategically about how to gather theknowledge and contacts required to move upto the next level. People are, in other words,already involved in a process that looks verymuch like bootstrapping, and in this sensethe strategy we are proposing in many wayssimply formalizes, codifies, and, hopefully,improves on a process already in progress.

In any case, it does not appear necessary tochoose between the two approaches to upgrad-ing, any more than it seems necessary to choosebetween exports and the domestic market. Thebenchmarking and broader debate aroundwhich the bootstrapping strategy is built shouldserve to facilitate the learning arrangementsassociated with either the foreign partnershipsor consultants. And in the case of foreign part-nerships, it might allow the maquiladoras tocreate or maintain the skills in design and mar-keting downstream and management in a sub-contracting chain upstream that they nowseem to give up when they enter into a rela-tionship with a foreign partner for upgrading.

Notes

Acknowledgments: This was a joint research under-taking of the Massachusetts Institute of Technol-ogy (MIT) and the Universidad Nacional Autó-

noma de México. It was supported by funds fromthe World Economy Laboratory at MIT and a grantfrom the MacArthur Foundation to the Center forInternational Studies at MIT.

. Own calculations based on an Expansión sur-vey of the largest firms in Mexico. The exactnumber of firms varies from to ; it goesfrom firms in that accounted for . per-cent of total Mexican exports, including maquilado-ras, to firms in that accounted for .

percent of total exports (Expansión ‒).. This finding is pervasive in the literature (see,

for example, Gonzales-Aréchiga and Ramírez a,b, c; Wilson ). On linkages withinMexican industry itself, see Rabellotti (). It isnot clear, however, whether the apparent weaknessof these interindustry linkages is a peculiarly Mex-ican phenomenon. Only the last of the studies citedin this note compares Mexico to other countries, andthis is a comparison with Italy, where the interfirmlinkages are believed to be unusually strong.

. The kanban system is a complex administra-tive and production organization that includes ajust-in-time supplier-client system to manage toolchanges, product changes, material purchasing, andplanning. It thereby reduces stocks and work inprogress.

. Formal models that capture elements of thisprocess have been developed by Caballero andHammour () and by Hansen (; ). Theproblem of inducing investments in upgrading hereis a specific instance of what Caballero and Ham-mour call the “appropriability” problem. Theseconjectures are thus basically about how the appro-priability problem is resolved by particular firms.

. In the current depressed conditions of theMexican economy, the investments that the new dis-count retail chains made initially may also act as arent and provide an inducement for them to take onthe task of upgrading local contractors. The invest-ments are a sunk cost. To earn a return upon them,the companies must try to minimize their losses, holdwhat customers they can, and survive until domes-tic demand revives. One strategy for doing this wouldbe to substitute lower-cost Mexican goods for theproducts they were importing from abroad, but to

, , .


do so without losing the reputation that differenti-ates them from other retail outlets. The contributionof this strategy to survival and the long-term profitthat survival will generate is thus a kind of rent thatcould be used to bind the producers that it trains.

. ISO is a set of evolving international stan-dards for businesses or organizations that initiallydeveloped in the United Kingdom in the s.These guidelines and requirements apply to suchtasks as inquiries and orders, doing the job or work,checking the work, and delivering the product. Theintended effect of the systematic evaluation andimplementation of these procedures is to improvethe quality and productivity of economic units.

References

Alvarez Galván, José Luis, and Enrique DusselPeters. . “Causas y efectos de los programasde promoción sectorial en la economía mexi-cana.” Comercio exterior , : ‒.

Bancomext (Banco Nacional de Comercio Exterior).. Sistema de comercio exterior-México (SIC-M). Mexico City: Bancomext.

Caballero, Ricardo, and Mohamad L. Hammour.. “On the Ills of Adjustment.” Journal ofDevelopment Economics , (October): ‒.

Expansión. ‒. “Las empresas más impor-tantes de México.” Expansión (Mexico City).

Fine, Charles, and Daniel Whitney. . “Is theMake-Buy Decision Process a Core Compe-tence?” MIT Working Paper, Massachusetts In-stitute of Technology, Cambridge, Mass., January.

Gereffi, Gary. . “The Organization of Buyer-Driven Global Commodity Chains: How U.S.Retailers Shape Overseas Production Networks.”In Commodity Chains and Global Capitalism, ed.Gary Gereffi and Miguel Korzeniewicz, pp. ‒

. Westport, Conn.: Praeger.Gonzales-Aréchiga, Bernardo, and Jose Carlos Ra-

mírez. a. “Estructura contra estrategia:Abasto de insumos nacionales a empresas expor-tadoras.” In Subcontratción y empresas transna-cionales: Apertura y restructuración en la maquila-

dora, ed. Bernardo Gonzales-Aréchiga and JoseCarlos Ramírez, pp. ‒. Tijuana: El Colegiode la Frontera Norte.

———. b. “Perspectivas estructurales de laindustria maquiladora.” In Subcontratción y em-presas transnacionales: Apertura y restructura-ción en la maquiladora, ed. Bernardo Gonzales-Aréchiga and Jose Carlos Ramírez, pp. ‒.

Tijuana: El Colegio de la Frontera Norte.———, eds. c. Subcontratción y empresas trans-

nacionales: Apertura y restructuración en la maquila-dora. Tijuana: El Colegio de la Frontera Norte.

Hansen, Gordon. . “Industry Agglomerationand Trade in Mexico.” Unpublished Ph.D. diss.,Department of Economics, Massachusetts Insti-tute of Technology, Cambridge, Mass.

———. . “Incomplete Contracts, Risk, andOwnership.” International Economic Review ,

(May): ‒

Nonaka, Inkujiro. . The Knowledge-CreatingCompany: How Japanese Companies Create the Dy-namics of Innovation. New York: Oxford Univer-sity Press.

Rabellotti, Roberta. . “External Economies andCooperation in Industrial Districts: A Compar-ison of Italy and Mexico.” Unpublished Ph.D.diss., University of Sussex, Brighton, U.K.

Robinson, Elizabeth. . “A Comparative Studyof the Economic Effects of External and Inter-nal Linkages Achieved through Compensatory-Type Investments: The Mexican AutomobileIndustry.” Unpublished Ph.D. diss., Departmentof International Business, George WashingtonUniversity, Washington, D.C.

Sabel, Charles. . “Bootstrapping Reform: Build-ing Firms, the Welfare State, and Unions.” Pol-itics and Society , : ‒.

Shaiken, Harley, and Stephen Herzenberg. .

Automation and Global Production: Automobile En-gine Production in Mexico, the United States, andCanada. La Jolla: Center for U.S.-Mexican Stud-ies, University of California at San Diego.

Wilson, Patricia A. . Exports and Local Devel-opment: Mexico’s New Maquiladoras. Austin: Uni-versity of Texas Press.


dusselpeters.com · 2020. 7. 17. · Created Date: 20200620052100Z

Documents