1

MEASUREMENT PROBLEMS MEASUREMENT PROBLEMS OF TFP PERFORMANCE IN AN OF TFP PERFORMANCE IN AN

UNSTABLE ECONOMY: UNSTABLE ECONOMY: ARGENTINA 1990-2004ARGENTINA 1990-2004

““A Case of the Tyranny of Methodology”A Case of the Tyranny of Methodology”

ARIEL COREMBERGARIEL COREMBERGECLAC Buenos AiresECLAC Buenos Aires

CONICET/IDESCONICET/IDESacorem@fibertel.com.aracorem@fibertel.com.aracoremb@yahoo.com.ar

The 2008 World Congress on National Accounts and Economic Performance Measures for Nations

May 12–17, 2008Key Bridge Marriott (Arlington, VA—minutes from Washington D.C.)

2

• (T. J. Koopmans 1947) This research applies the famous Koopman`s motto: “It is necessary to avoid Measurement without theory”

• (J. Hicks 1981)“ The measurement of capital is one of the nastiest jobs that economists have set to statisticians”.

• Z.Grilliches (1990): “procyclical fluctuations in ‘productivity’ do not make sense if we want to interpret them as a measure of the growth in the level of technology or the state of economically valuable knowledge of an economy. The US. Economy did not forget 4% of its technology between 1974 and 1975.”

• D. Jorgenson (1995): the productivity term A reflects all those effects on output growth that are not investment, where investment is understood as the commitment of current resources in the expectation of future returns, implying that these returns can be internalised by the investor.

• Heymann, Galiani y Tomassi (2004): Agents make economic decisions taking the cycle as a trend: a correct identification of TFP is important in the context of sustainability of macroeconomic configuration, wealth perceptions and competitiveness of an economy

3

OBJETIVEOBJETIVE• IDENTIFICATION OF TREND AND IDENTIFICATION OF TREND AND

GROWTH PROFILEGROWTH PROFILE

• SUSTAINABILITY OF GROWTH AND SUSTAINABILITY OF GROWTH AND COMPETITIVENESS IN THE LONG RUN: COMPETITIVENESS IN THE LONG RUN: STRICT TFPSTRICT TFP

• EMERGING UNSTABLE ECONOMIES EMERGING UNSTABLE ECONOMIES WITH RELATIVE PRICE VOLATILITY WITH RELATIVE PRICE VOLATILITY AND INCONCISTENT MACROPOLICIESAND INCONCISTENT MACROPOLICIES

• ARGENTINA 1990-2004: WHAT ABOUT ARGENTINA 1990-2004: WHAT ABOUT STRICT TFP? STRICT TFP?

4

ARGENTINA-STYLIZED FACTS: 90´SARGENTINA-STYLIZED FACTS: 90´S

• 90`s: deregulation & privatization, trade & financial

openness

• apparent productivity gains compensate real

appreciation due to convertibility plan

• Lower relative capital costs: K intensity

• Previous Underutilization of K: Ku (60-75)

• After Brazilian devaluation, inconsistent macro

policy could not generate S and TFP in l.r.

5

ARGENTINA-STYLIZED FACTS: 02-08ARGENTINA-STYLIZED FACTS: 02-08

• 2002 Devaluation and “Corralito”: double real

exchange rate: import substitution and increase in

exports + growth in aggregate demand:

improvement in terms of trade as permanent

• High decreases in labor costs and Previous

Underutilization of K & L : Ld(ul: 20-10%) and Ku

(55-73)

6

SOURCE OF GROWTHSOURCE OF GROWTH

1. Increase in capital intensity

2. Improvements in productive organization independently of factor endowments (TFP)

3.3. Changes in the composition of GDP, K, LChanges in the composition of GDP, K, L

4.4. Quality changes in primary inputsQuality changes in primary inputs

5.5. Cyclical Changes in the Utilisation of Cyclical Changes in the Utilisation of Productive Factors: Labour intensity L and Productive Factors: Labour intensity L and K UtilizationK Utilization

7

y

k

Eb

E*

E**

yb

y*

0

B

A

kbka

BASIC SOURCE OF GROWTH: BASIC SOURCE OF GROWTH: TFP OR kTFP OR k

ya

Ea

8

y

k

E0

E1

E1’

y0

y1

0

A

B

k0k1

TFP OR KTFP OR KUU

9

TFP TYPOLOGIESTFP TYPOLOGIES

• APARENT TFPAPARENT TFP: INCLUDE QUALITY, COMPOSITION & UTILIZATION EFFECTS

• STRICT TFPSTRICT TFP : SHIFTS IN PRODUCTION FUNCTION DUE TO ORGANIZATION IMPROVEMENT (excludes Q, C, U effects)

10

Identification of Basic Source of Identification of Basic Source of GrowthGrowth

• Intensive Form

• Extensive Form

• Abramovitz: Solow´s Residual is the measurement of our ignorance

• Grilliches: non-measure factors and error measurement are in the Residual

dt

Ad

dt

kds

dt

ydK

lnlnln

dt

Ad

dt

Lds

dt

Kds

dt

YdKK

lnlnlnln

11

Index NumbersIndex Numbers

• Tornquist Index

n

i

tititi

T

dt

Xdvv

dt

Xd

1

,1,,

ln5.0

ln

12

dt

Xd

dt

Xd

dt

Xd tiB

tiOC

,, lnlnln

dt

Xd

dt

Xd

dt

Xd tiU

tiBQ

,, lnlnln

Composition Effects

Quality Effects

Utilization Effects

dt

Xd

dt

Xd

dt

Xd tipot

tiaju

,, lnlnln

13

EXHAUSTIVE GROWTH EXHAUSTIVE GROWTH ACCOUNTINGACCOUNTING

dt

Kd

dt

Kd

dt

Kd

dt

Kds

dt

Qd

dt

Qd

dt

Ad ukQCB

K

BCSW lnlnlnlnlnlnln

dt

Ld

dt

Ld

dt

Lds

dt

Kd

dt

Kd

dt

Kds

dt

Qd

dt

Ad

dt

Ad ulCQ

L

ukCQ

K

Csa lnlnlnlnlnlnlnlnln

dt

Ld

dt

Ld

dt

Ld

dt

Lds

CQulU

L

lnlnlnln

14

MAIN RESULTSMAIN RESULTS

15

GDPGDP

16

GROSS DOMESTIC PRODUCT IN ARGENTINA1993=100

70

80

90

100

110

120

130

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

LASPEYRES INDEX

OPTIMAL INDEX

Source: Author´s estimation based on National Accounts- INDEC

17

SHARE OF TRADABLES INDUSTRIES IN THE ARGENTINE GDP-current prices-

0,0%

5,0%

10,0%

15,0%

20,0%

25,0%

30,0%

35,0%

40,0%

45,0%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003Source: Author´s estimation based on National Accounts- INDEC

18

TABLE 1 GROSS DOMESTIC PRODUCT IN ARGENTINA

Average Annual Growth Rates Type of Index/ Period

1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

Laspeyres93 7,78% -2,62% 5,94% -2,65% 3,38% -10,02% 7,77% 2,97% Optimal 7,96% -2,36% 5,86% -2,46% 3,50% -8,96% 8,15% 3,20% Substitution 0,17% 0,26% -0,08% 0,19% 0,12% 1,06% 0,38% 0,23%

19

LABOR INPUTLABOR INPUT

20

• Hs. Worked vs Jobs

• Inclusion Self Employed

Another Input

Imputation of Labor Compensation of Self Employed with the Wages of Employees by Sector

Imputation of Internal Rate of Return of the Industry

Jobs

Adjustment by Part Time jobs And Double Occupation-

Employees Self Employees and others

Nr. Of Workers

Hourly Labor

By worker

Hours WorkedBy worker

Nr. Hours Worked

Labor Differentation

Explicit

Implicit

Hours Worked Differentiated by Worker

characteristics

Hours Worked By industry

MEDICION DEL INSUMO TRABAJO ( OECD)

21

dt

Ld

dt

Ld

dt

Ld

dt

Ld

dt

Ld UulCQO lnlnlnlnln

Weights: wages by industry

dt

Ld

dt

Ld

dt

Ld UBQ lnlnln

22

LABOUR INPUT IN ARGENTINA BY INDICATOR1993=100

70

80

90

100

110

120

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Hours Worked

Jobs

Source: Author´s estimation based on National Accounts- INDEC

23

TABLE 1 LABOUR INPUT IN ARGENTINA*

Annual Average Growth Rates

Labour Input Indicator / Period

1990-1994

1994-1995

1995-1998

1998-2001

1991-2001

2001-2002

2002-2004

1990-2004

Jobs 2,44% -2,87% 3,76% -0,34% 1,54% -5,68% 6,33% 1,67%

Hours Worked 4,09% -3,64% 3,27% -1,60% 1,57% -10,60% 10,46% 1,86% Labour Intensity 1,64% -0,77% -0,49% -1,25% 0,03% -4,92% 4,13% 0,19%

24

TABLE 1 LABOUR INPUT IN ARGENTINA*

Annual Average Growth Rates

Labour Input Indicator / Period

1990-1994

1994-1995

1995-1998

1998-2001

1991-2001

2001-2002

2002-2004

1990-2004

Undifferentiated 4,09% -3,64% 3,27% -1,60% 1,57% -10,60% 10,46% 1,86% Quality Effect 0,11% 0,06% -0,04% 0,10% 0,06% 0,10% -0,19% 0,03% Laspeyres 4,19% -3,58% 3,23% -1,50% 1,63% -10,50% 10,28% 1,89% Composition Effect 0,07% 0,14% 0,03% 0,17% 0,09% 0,46% -0,64% 0,03% Optimal 4,26% -3,43% 3,26% -1,33% 1,72% -10,03% 9,64% 1,92%

Total Effect 0,18% 0,21% -0,01% 0,27% 0,15% 0,57% -0,83% 0,06%

25

CAPITAL SERVICESCAPITAL SERVICES

26

dt

Kd

dt

Kd

dt

Kd

dt

Kd ukP

CP

QP

OP lnlnlnln

dt

Kd

dt

Kd

dt

Kd WPQ lnlnln

i

tititPt

Ist

Pt

Kt rKPKPwLPQ ,,,

Measurement Problems of K Measurement Problems of K services in Unstable Economiesservices in Unstable Economies

•Cohorts Aggregation: Efficiency Growth Profile•User cost•Utilization Indicators

27

CAPITAL SERVICES IN ARGENTINA1993=100

80

90

100

110

120

130

140

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

WEALTH STOCK LASPEYRES INDEX Asset Prices

WEALTH STOCK T. INDEX Asset Prices

CAPITAL SERVICES T. INDEX User Cost

Source: Author´s estimation based on National Accounts- INDEC

28

TABLE 1 CAPITAL STOCK SERVICES IN ARGENTINA 1990-2004

Annual Average Growth Rates Total Durable

Equipment Construction*

Laspeyres Wealth Capital 2,03% 1,84% 2,09% Composition Effect 0,93% 0,41% 1,03% Optimal Wealth Capital 2,96% 2,25% 3,11% Quality Effect -0,50% 0,76% -1,05% Optimal Capital Services 2,46% 3,01% 2,06% Total Effect 0,43% 1,16% -0,03%

29

TABLE 1 PROXY INDICATORS OF INSTALLED CAPACITY USE

Assumption Statistical Coverage in Argentina

Output Gap Usual interpretation of Okun: Natural Unemployment Rate

According to the analyst subjectivity, econometric estimations

Employment Rate Complementarity

Hours Worked Complementarity

By Household Surveys in Urban Conglomerates

Energy Consumption Complementarity Energy Demand Surveys Representativity Indicators limited to Manufacturing Indust.

30

CAPITAL SERVICES: ADJUSTMENT BY UTILISATION INDICATORS1993=100

70

80

90

100

110

120

130

140

150

160

170

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

POTENCIAL CAPITAL SERVICES

GDP

CAPITAL SERVICES adj. by hours worked

CAPITAL SERVICES Adj. by Manufacturing Indicator

CAPITAL SERVICES Adj. by Energy

Source: Author´s estimation based on National Accounts- INDEC

31

TABLE 1 CAPITAL SERVICES UTILISATION IN ARGENTINA

Annual Average Growth Rates

Indicator / Period 1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

Potential Capital Services 3,55% 3,01% 3,75% 1,83% 3,08% -1,59% 1,14% 2,46% Capital Scies. Adj. By Manuf. Indicator 9,84% 1,89% 3,24% -2,02% 3,97% -7,12% 7,40% 3,62%

Utilisation Effect 6,29% -1,11% -0,52% -3,85% 0,89% -5,52% 6,26% 1,15%

32

SUMMARYSUMMARYEffects

ExcludedBias on Q Bias on

InputsBias on Q/L

& TFP

QC - -

LC&LQ (02-04) + -

L Intensity +Cyclical Phase - +

-Cyclical Phase + -

Ku +Cyclical Phase - +

-Cyclical Phase + -

KC&KQ - +

33

LABOUR PRODUCTIVITY IN ARGENTINA1993=100

80

85

90

95

100

105

110

115

120

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Jobs

Hours Worked

Source: Author´s estimation based on National Accounts- INDEC

34

TABLE 1 LABOUR PRODUCTIVITY IN ARGENTINA* BY TYPE OF LABOUR INPUT INDICATOR

Average Annual Growth Rates

Labour Input Indicator / Period

1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

Jobs 5,28% 0,25% 2,17% -2,31% 1,86% -4,34% 1,44% 1,34% Hours Worked 3,56% 1,02% 2,58% -1,06% 1,78% 0,58% -2,76% 1,04%

Labour Intensity Effect -1,72% 0,77% 0,41% 1,25% -0,07% 4,92% -4,20% -0,31%

35

TABLE 1 LABOUR PRODUCTIVITY IN ARGENTINA BY METHODOLOGY

Average Annual Growth Rates

Method/ Period 1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

QLaspeyres/L Undiff. 3,56% 1,02% 2,58% -1,06% 1,78% 0,58% -2,76% 1,04% QLaspeyres/L Laspeyres 3,51% 0,96% 2,61% -1,15% 1,74% 0,48% -2,57% 1,03% Q Optimal/LOptimal 3,62% 1,08% 2,51% -1,13% 1,77% 1,07% -1,52% 1,25% Total Effect 0,06% 0,06% -0,06% -0,07% -0,01% 0,49% 1,24% 0,21%

36

CAPITAL INTENSITY IN ARGENTINA BY INDEX AND INDICATOR1993=100

70

80

90

100

110

120

130

140

150

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

K/L Optimal Adj. by Utilisation

K/L Undifferentiated

K/L Laspeyres

K/L Optimal Index-Potencial Kl

K/L Jobs Undifferentiated

Source: Author´s estimation based on National Accounts- INDEC

37

CAPITAL INTENSITY IN ARGENTINA BY COMPONENTS1993=100

70

80

90

100

110

120

130

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

K/L Optimal Adj. by Utilisation

L Optimal Adj. by Labour Utilsation

Capital Services Adj. by Capital Utilisation

Source: Author´s estimation based on National Accounts- INDEC

38

TABLE 1 CAPITAL INTENSITY IN ARGENTINA BY METHODOLOGY

Average Annual Growth Rates

Method/ Period 1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

KW Laspeyres/L Undiff. Jobs -0,08% 5,58% -0,56% 1,84% 0,81% 3,66% -5,31% 0,11% KW Laspeyres/Undiff. Hours -1,84% 6,35% -0,17% 3,09% 0,67% 8,58% -9,51% -0,31% KW Laspeyres/Laspeyres hours -1,86% 6,29% -0,13% 3,00% 0,64% 8,48% -9,32% -0,31% KP Optimal/ Optimal Hours -0,89% 6,44% 0,39% 3,16% 1,20% 8,44% -8,56% 0,24% KP utilised/ Optimal Hours 5,48% 5,33% -0,12% -0,69% 2,21% 2,92% -2,27% 1,61% Total Effect 5,55% -0,25% 0,44% -2,53% 1,40% -0,75% 3,04% 1,50%

TABLE 1 COMPONENTS OF THE CAPITAL INTENSITY GROWTH*

Average Annual Growth Rates

Factor/ Period 1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

K 9,84% 1,89% 3,24% -2,02% 3,97% -7,12% 7,40% 3,62% L 4,26% -3,43% 3,26% -1,33% 1,72% -10,03% 9,64% 1,92% K/L 5,48% 5,33% -0,12% -0,69% 2,21% 2,92% -2,27% 1,61%

39

TOTAL FACTOR PRODUCTIVITY IN ARGENTINA1993=100

80

85

90

95

100

105

110

115

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Laspeyres 1

Laspeyres 2

Lasperyes 3

Laspeyres 4

Optimal Index

Strict: Optimal Index Adj. by Factors Utilisation

Source: Author´s estimation based on National Accounts- INDEC

40

TABLE 1 TOTAL FACTOR PRODUCTIVITY IN ARGENTINA BY METHODOLOGY

Average Annual Growth Rates

Method/ Period 1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

Laspeyres 1 5,3% -2,1% 2,4% -3,1% 1,5% -5,9% 3,7% 1,2% Laspeyres 2 4,3% -1,7% 2,7% -2,4% 1,5% -3,0% 1,3% 1,1% Laspeyres 3 4,6% -2,0% 2,7% -2,5% 1,5% -4,2% 2,5% 1,2% Laspeyres 4 4,6% -2,0% 2,7% -2,6% 1,5% -4,3% 2,5% 1,2% Optimal Index 4,2% -2,0% 2,3% -2,6% 1,2% -3,7% 3,2% 1,1% Strict 1,1% -1,5% 2,6% -0,8% 0,7% -0,6% -0,3% 0,5%

41

INTERNATIONAL COMPARISONS

Q, Q/L, TFP

USA EU ARG

70-95 95-04 70-95 95-04 95-04

Q 2.8 3.2 2.2 2.2 1.5 (1.9)

Q/L 1.3 2.4 2.4 1.4 -0.2 (0.3)

80-95 95-04 80-95 95-04 95-04

TFP 0.7 1.6 0.7 0.3 0.3a (-0.1S)

Q K L TFP

Hong Kong 66-91 7.3 8 3.2 2.3

Singapore 66-90 8.5 11.5 5.7 -0.3

Corea 66-90 10.4 13.7 6.4 1.6

Taiwan 66-90 9.6 12.3 5.1 2.4

42

SOURCE OF GROWTH IN ARGENTINA 1990-2004

59%

-38%

28%40%

56%44% 50% 57%

14%

-44%

23%

-24%

31%9% 6%

27%

28%

76% 28%28%

25% 49%53% 30%

18%63%

37%

7%

26%

33% 47%

33%

13%

62%

45%32%

19%6%

-3%

13%

68%

80%

41%

116%

43%59%

47% 40%

-100%

-50%

0%

50%

100%

150%

200%

1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

1990-1994

1994-1995

1995-1998

1998-2001

1990-2001

2001-2002

2002-2004

1990-2004

Co

ntr

ibu

tio

n t

o A

vera

ge

An

ual

GD

P G

row

h(%

)

K L TFP

OPTIMAL METHODOLOGYTRADITIONAL METHODOLOGY

Source: Author´s estimation based on National Accounts- INDEC

43

EXTENSIVE GROWTH PROFILEEXTENSIVE GROWTH PROFILE

This Results are This Results are analogous to analogous to Young (1997) Young (1997) y Timmer and y Timmer and Van Ark Van Ark (1990) for Nic(1990) for Nic´s experience.´s experience.

%% 1990-1990-20042004

1990-1990-20012001

2002-2002-20042004

KK 5757 5656 5050

LL 3030 2525 5353

Strict TFPStrict TFP 1313 1919 -3-3

Apparent TFPApparent TFP 4040 4343 4747

44

CONCLUSIONSCONCLUSIONS

• K estimation was done following OECD and SNA: econometric estimation of depreciation, hedonic prices, and rental equivalent taking into account the reliability of Argentine statistics: K/Q & K/L lower & greater dynamism than developed countries.

• Adjustment by User Cost: Greater Performance than KW: excluding this effect could overestimate the TFP

45

CONCLUSIONSCONCLUSIONS

• Non-measure of Composition and Quality Effects in L & Substitution in Q underestimate Q/L & Strict TFP (2002-2004)

• The adjustment by Labour Intensity and K utilisation reduce the pro cyclical behaviour of TFP

46

• The importance of competitive gains of the Argentine economy through improvements in apparent cyclical TFP, generated both during the nineties and after 2002 devaluation, are unquestionable. However, doubts arise about the ability of the Argentine economy to generate the necessary productivity gains in the strict sense, independently of composition and quality effects and cyclical variations in factor utilisation, that allow to maintain a sustainable long run growth.

• The extensive growth profile diagnosed for the Argentine economy, especially during the nineties, contrasts with assessments of other authors and institutions based on the traditional methodology: without adjusting by relative price effects and factor utilisation. On the contrary, our results are analogous to the evidence found by Young (1995) and Timmer and Van Ark (2000) for Nic´s countries.

• This conclusion is based not only in what Young (1995) called the “tyranny of numbers”, by assessing strictly the consistency of the country statistical information, but also a consequence of the “tyranny of the economic cycle, macroeconomic and methodological consistency”.