can imports discipline collusive firms?: the case of the philippine ... · can imports discipline...

Philippine Journal of DevelopmentNumber 64, First Semester 2008

Volume XXXV, Number 1

Can Imports Discipline Collusive Firms?:The Case of the Philippine Cement Industry

RAFAELITA M. ALDABA*

ABSTRACTApplying a conjectural variations (CV) model introduced by Haskeland Scaramozzino (H&S model 1997), the paper examines the impact oftrade liberalization on the Philippine cement industry where allegedcartel activities have taken place after the entry of the world's BigThree cement firms: Holcim, Cemex, and Lafarge. In the H&S model,the relationship between firm behavior and competition is estimatedwith price cost margin (price minus marginal costs over price) as indi-cator of competition and profitability. The model is extended to assessthe impact of imports on competition using import penetration ratio asproxy for trade policy.

The paper focuses on the following questions: did the removal ofimport restriction and reduction of tariffs affect competition in thecement industry? Are imports effective in disciplining domestic firmsand reducing their market power? The results imply that imports donot seem to affect profitability and competition in the industry. Giventhe ability of firms to engage in anticompetitive behavior and the ab-sence of an effective competition policy in the Philippines, the gainsfrom trade liberalization are nullified. The country's experience in thecement industry illustrates that trade liberalization is not a substitutefor competition policy. For imports to effectively discipline the market,trade liberalization must be accompanied by strict competition policy.

* Senior Research Fellow, Philippine Institute for Development Studies, NEDA sa Makati Building, 106Amorsolo St., Legaspi Vil lage, 1229 Makati City, Phil ippines. Email for correspondence:[email protected].

PHILIPPINE JOURNAL OF DEVELOPMENT 20082

INTRODUCTIONThere are three main channels through which trade liberalization affects a country’seconomic performance. These are: (i) static gains due to resource allocation ef-fects, (ii) dynamic gains through technical change leading to productivity growth,and (iii) competitive effects arising from domestic competition which Levinsohn(1993) called the imports-as-market discipline hypothesis. In the presence of im-perfect competitive markets, trade liberalization can result in additional gains byreducing the deadweight losses created by domestic firms’ market power. Withmore competition, domestic firms (whether foreign or locally owned) which reapoligopoly profits in a protected environment are forced to become more competi-tive. Under imperfect competition, firms do not take prices as given and as a result,prices do not equal marginal cost.

Assuming that perfect competition exists, trade policy analysis shows thatfor a small country without domestic distortions and where income distribution isnot an issue, free trade is optimum even if other countries impose trade restric-tions. Under imperfect competition, trade policy focuses on an old, theoreticallyrobust insight that international competition curtails market power (Levinsohn1993). As Bhagwati (1968) indicated, protection of domestic enterprises allowsdomestic firms to increase their market power, hence, trade liberalization is a pow-erful and administratively simple way to enhance competition.

In the Philippines, there are a lot of empirical work that evaluate the impact oftrade liberalization on resource allocation (the first of the three channels men-tioned earlier) pioneered by Medalla (Medalla, Bautista, Tecson and Associates1995; Medalla 1998; see also Cororaton and Cuenca 2000; Cororaton 2003; andHabito, Cororaton, Intal et al. 1999; among others). However, empirical studies ontrade and competition (the third channel), as well as studies on the impact of tradereforms on productivity (the second channel), are still few and limited to single-year cross-section analysis based on the traditional structure-conduct-perfor-mance (SCP) model. In the Philippine manufacturing industry, trade and competi-tion studies by Imbat and Tanlapco (1993), L. de Dios (1994), and E. de Dios (1986)are based on SCP profitability equations which regress price cost margin on importpenetration ratio as a proxy for trade policy and other factors such as concentra-tion ratios and capital output ratios. The results of these studies provide generalsupport for the import discipline hypothesis in the manufacturing industry.

In this paper, the competitive effects of trade reforms will be analyzed byfocusing on the cement industry. Alleged cartel activities have taken place in theindustry after the entry of the world’s big three cement firms, namely, Holcim,Cemex, and Lafarge. Applying a conjectural variation (CV) model introduced byHaskel and Scaramozzino (1997), the paper attempts to determine a firm’s behav-ior and by extending the model to include a trade policy variable, it examines the

ALDABA 3

impact of trade liberalization on competition in the industry. In the Haskel andScaramozzino model (H&S), the relationship between firm behavior and compe-tition is estimated with price cost margin (price minus marginal costs over price)as indicator of competition and profitability. The model is extended to assess theimpact of imports on competition using import penetration ratio as proxy fortrade policy.

GOVERNMENT POLICIES: PROTECTION, REGULATIONAND LIBERALIZATIONCement is a homogeneous product characterized by geographic markets. Its manu-facturing is basically resource-based with cement plants located in or near lime-stone quarry areas. It has a limited shelf life and is characterized by relatively hightransport and handling costs. In Luzon, cement is distributed by land and by bothland and water in Visayas and Mindanao.

Historically, the Philippine cement industry thrived under a government-sanctioned cartel. Due to the economic slump in the early 1970s, cement firmspushed for government regulation to prevent cut-throat competition. The govern-ment allocated supply, controlled prices, and regulated entry in the industry. How-ever, in the absence of the necessary firm-level information to efficiently performthese tasks, the government delegated the setting of production quotas to theindustry association.

Collusion took place through the firms’ informal agreement to set produc-tion quotas and to assign geographic markets among themselves (Lamberte et al.1992). This territorial arrangement restricted Luzon-based plants to sell only in theLuzon area while the Visayas/Mindanao-based plants were to confine their saleswithin the area. This practice divided the country into regional markets served bya dominant player which eliminated competition from taking place. The behavior offirms was described by a top executive of one of the major foreign-owned cementcompanies in the country as follows:

“People take positions that demonstrate to other people that theyare going to behave in a rational manner. By closing Calaca (AlsonCement’s bulk terminal in Batangas Province, south of Manila), Iam indicating that I’m prepared not to compete in Luzon. And whatdo I expect in return? I expect people to allow me to reasonablyclear market in Mindanao and the Visayas. And if they come andwreck my market (in those areas), I’ll open up Calaca again. It’s assimple as that.” 1

1 Interview with Tomas Clough, published in the Asian Cement Magazine as presented by the Confedera-tion of Homeowners Association for Reforms in Governance and Environment (CHARGE) during ahearing at the House of Representatives on cement.


As the government pursued market-oriented reforms in the 1980s, the indus-try was deregulated and liberalized. Government regulation was abandoned astariffs were significantly reduced and non-tariff barriers were eliminated. Effective31 March 1989, Central Bank Circular 1195 lifted the import restrictions on cement.In 1988, the tariff on cement was reduced to 40 percent from 50 percent in 1979.This was further cut down to 20 percent in 1989, to 5 percent in 1993–1994, and to3 percent in 1995–1997. As of 2008, its most favored nation (MFN) rate stands at5 percent while its ASEAN Free Trade Area (AFTA) rate is 3 percent.

Table 1 presents the effective protection rates2 arising from these trade re-forms. It is evident from the table that the tariff reduction has substantially reducedeffective protection in the industry from a very high level of 145 percent in 1987 tothe current rate of 4.3 percent. Under competitive conditions, a negative relationshipis expected between imports and protection with trade liberalization leading to in-creased imports. The table shows small import penetration rates in the industry.Except for the three years 1990, 2000, and 2001, these have not gone beyond 10percent. Imports were high in 1990, 1992, 1995, and 1996 due to demand shortages inthe domestic market. In 2000 and 2001, imports surged because of the high domesticprices in the country relative to world prices. During these years, there was a globalexcess supply of cement and depressed demand due to the generally unfavorableworld economic conditions after the 1997 Asian financial crisis.

Table 1. Effective protection rates and import penetration rates 1987–2003 (in percent)

Year MPR EPR Year MPR EPR

1987 3.95 144.8 1996 5.25 -2.311988 0.00 132.1 1997 2.36 -2.491989 2.05 -47.7 1998 1.40 8.061990 12.57 -16.85 1999 3.84 6.371991 0.14 -18.67 2000 12.40 3.511992 9.36 -19.09 2001 19.70 3.541993 0.00 -7.44 2002 5.10 4.371994 0.10 23.41 2003 0.78 4.311995 4.33 -8.02

Sources: Import data were from the Philippine Cement Manufacturers Corporations (PHILCEMCOR)while EPR estimates were from various research papers of the Philippine Institute for Develop-ment Studies (PIDS).

EPR: Effective Protection RateMPR: Import Penetration Rate

2 Effective protection rates (EPRs) measure the proportion by which value added, measured in domesticprices, exceeds the same value added measured at world prices. EPRs are measures of the netprotection received by domestic producers from the protection of their outputs and the penalty from theprotection of their inputs.

ALDABA 5

In the early 1990s, the cement companies invested in capacity expansion,however, they encountered serious financial difficulties due to the 1997–1998Asian financial crisis. Foreign companies came in and bought into the industrythrough mergers and acquisitions (M&As). The industry which used to be domi-nated by several family-owned firms is currently divided into three major groups:Holcim, Lafarge, and Cemex. The Holcim group controls Alsons, Bacnotan, Hi-Cement, and Davao Union while the Lafarge group consists of Continental, For-tune, FR, Iligan, Limay, Lloyds, Mindanao, and Republic. The Cemex group iscomposed of Apo, Rizal, and Solid. Grand has been acquired by the Japanesecompany, Taiheiyo Cement Philippines, Inc.

Table 2 compares firm market shares before and after the change in owner-ship and the wave of mergers and acquisitions in the late 1990s. The LaFargegroup’s average share increased from 31 percent (during the years 1994–1999) to34 percent (during the period from 2000–2004). The Cemex group’s average marketshare also rose from 19 percent to 21 percent during the same periods under study.The Holcim group’s average share fell from 36 percent to 31 percent. After thecompletion of the M&As, independent firm Northern had an average market shareof 7 percent while Grand Cement and Pacific had 4 and 2 percent, respectively.

In terms of geographic markets, Table 3 indicates that the major playersin Luzon are Solid and Hi-Cement. In the Visayas, the market is controlled byApo while in Mindanao, Davao Union leads. The same firms dominated theindustry since the early 1990s (see Aldaba 2002 for geographic market sharesfrom 1990 to 1999).

Figure 1 shows the behavior of cement prices before and after the M&As.Prior to 1997, price movements in the industry were fairly stable with prices gener-ally rising during the dry season (January to May) and falling during the rainy

Table 2. Market shares before and after the M&A wave

Cement Firms Pre-M&A Post-M&A 1994–99 2000 2001 2003 2004

HOLCIM 0.36 0.31 0.29 0.31 0.31CEMEX 0.19 0.21 0.22 0.21 0.21LAFARGE 0.31 0.35 0.32 0.34 0.34

IndependentGrand 0.04 0.04 0.06 0.04 0.04Pacific 0.02 0.02 0.03 0.02 0.02Northern 0.06 0.07 0.09 0.07 0.07

Source: PHILCEMCOR data.


months (June to December). With the 1997–1998 crisis, prices dropped from P104per bag in March 1997 to P45 per bag in December 1998; but starting in January1999, prices began to increase.

Figure 1. Average ex-plant prices (at constant 1990 prices)

Sources: PHILCEMCOR data and Department of Trade and Industry-Price Monitoring Unit.

January 1993–December 2001

Table 3. Geographic market shares

L u z o n 1994–99 2000 2001 2003 2004

Limay 0.01 0.02 0.02Northern 0.06 0.07 0.09 0.07 0.07Bacnotan 0.08 0.05 0.07Solid 0.12 0.07 0.09 0.10 0.10Hi-Cement 0.11 0.10 0.08Republic 0.07 0.08 0.07 0.08 0.08F R 0.05 0.08 0.06 0.07 0.07Fortune 0.05 0.06 0.06 0.08 0.08Continental 0.03 0.04 0.05 0.06 0.06

Visayas 1994–99 2000 2001 2003 2004

Apo (Batangas) 0.01 0.05 0.03Apo/TTC 0.03Apo 0.04 0.08 0.07 0.11 0.11Lloyds 0.04 0.03 0.02Grand 0.04 0.04 0.06 0.04 0.04

Mindanao 1994–99 2000 2001 2003 2004

Davao 0.09 0.08 0.06Alsons 0.06 0.08 0.07Iligan 0.04 0.04 0.04 0.05 0.04Pacific 0.02 0.02 0.03 0.02 0.02Mindanao 0.02

Source: PHILCEMCOR data.

ALDABA 7

The price increases in 1999 coincided with the completion of most mergersand consolidations in the industry. Prices started to go up from P45 in December1998 to P70 in February 1999. Prices steadily rose to P97 by December 1999. In May2000, ex-plant price/bag was already P109 reaching P132 per bag in May 2001.Considering that the industry was facing an oversupply and low demand, pricecoordination was seen as the only explanation for the price increases. The threatof a price war and the fear that departure from such behavior may lead to costlyprice cutting, lower profits and market instability create incentives for firms tomaintain an implicit arrangement among themselves. In describing their behavior,the president of La Farge Philippines pointed out:3

In any industry you can fight for a bigger market share but some-one has to give up that market share… But obviously this wouldn’twork with our competitors here…they will fight and then we get into abloody battle where everyone ends up bruised. Given the opportu-nity, yes we can try and improve our market share. But I don’t want tolose out in the end and trigger another price war. … LaFarge has noplans of rocking the boat. We wouldn’t make a run on market share, wewant to keep it as stable as possible.

Note that during the period of rising prices, there was excess capacity in theworld market. The relatively high domestic prices in the country attracted importsleading to an increase in the import penetration ratios from 12.4 percent in 2000 to19.7 percent in 2001. However, the entry of imports did not last long as the cementindustry lobbied strongly for government protection through antidumping andsafeguard measures to prevent imports from coming in. The industry also threat-ened the government that they would pull-out their investments. In November2001, the Department of Trade and Industry imposed a temporary additional dutyof P20.60 per bag of imported cement. In turn, the industry promised that the firmswould not increase their prices during this period and committed to sell cementwithin the price range of P125 to P135 per bag.

The industry association, Philippine Cement Manufacturers Corporation orPHILCEMCOR, filed a dumping suit against Taiwan Cement Corporation andJapan’s Taiheiyo for allegedly dumping cement at US$36 and US$20 per metricton (MT) ex-terminal import price, respectively. Compared with prices in othercountries, the average spot price for cement in India was US$40 per MT in 1997and this dropped further to between $25–27 per MT in 1998. In Thailand, pricesranged from US$16 to US$18 per MT in late November 1998.4

3 Interview with Philip Roseberg, Philippine Star, 24 February 2001.4 Indian Express Newspapers (www.indianexpress.com/news).


Japan’s Taiheiyo finally decided to increase its prices in line with the prevail-ing domestic cement prices. Refusing to do the same, Taiwan Cement Corporationwas made the main respondent of PHILCEMCOR’s anti-dumping case. Aside fromthe dumping case, domestic firms in the Philippines retaliated by attacking themaverick Taiwanese firm by exporting cement to Taiwan at very low prices.5 Alsons(a company owned by Holcim) stated that it would sell its cement at significantlylower price than the prevailing domestic price in Taiwan to send a message thatPhilippine cement companies could play their game. Eventually, the Taiwanesefirm discontinued its operations in the Philippines. In 2005, the Taiwan Fair TradeCommission (FTC) punished 21 cement firms for engaging in international cartelactivities with total fines of NT$210 million (Taiwan FTC 2006).

HASKEL AND SCARAMOZZINO MODEL: A FRAMEWORK FOR AS-SESSING THE IMPACT OF IMPORTS ON DOMESTIC COMPETITIONThis brings us to the following questions: did the removal of import restriction andreduction of tariffs affect competition in the cement industry? Are imports effec-tive in disciplining domestic firms and reducing their market power? In addressingthese questions, the conjectural variations (CV) model by Haskel and Scaramozzino(H&S) is applied. With price cost margin (PCM) as a measure of competition andprofitability, the H&S model focuses on the relationship between PCM and CV bylooking at how the profits of one firm are affected by firm characteristics such asphysical capacity and financial condition.

Conjectural variations describe how firms think others will react to changesin their quantities, i.e., the CV summarizes the response of firm j to changes inquantity set by firm i. The H&S approach assumes that the conjectures of a firmabout the reactions of its rivals depend on its own and rival firms’ physical andfinancial capacity. The CV of firm i depends on the capacity utilization of otherfirms because a rival with excess capacity can easily respond to changes in quan-tities of its competitors. A competitor in financial distress is expected to respondless aggressively than one whose financial condition is more robust.

Price cost margins are often used to measure the degree of competition inthe market. Assuming homogeneous goods industry, the first order condition forprofit maximization yields:

(1)

5 At the hearing in the House of Representative's Committee on Trade and Industry, Mr. Jesus Arranzaasked the big three cement firms why they exported cement at P40/bag instead of competing with importsin the domestic market by selling at P110/bag which would kill imports considering that the averagelanded price was already P110/bag.

ALDABA 9

wherePCM is a firm i’s price cost margin,si is its market share, e is elasticity of demand, andli is firm i’s CV or expectations about the reaction of other firms to a change

in its quantity.For a competitive firm, si is small, li is zero, and PCM converges to zero. For

an oligopolist, li and PCM are positive and vary between 0 and 1.In estimating li, the H&S approach allows CV to depend on the actual ability

of other firms to respond based on physical capacity and financial capability. TheH&S model uses capacity utilization as an indicator of physical capacity whileborrowing ratio, cash liabilities ratio, and interest payment ratio are used as indica-tors of financial position. Assuming that a firm’s conjectures are affected by thecharacteristics of the firm itself and of other firms in the industry, the l equation ismodeled as:

(2)

whereCU refers to capacity utilization,FS to a vector of measures of financial status, andthe subscript –i refers to rivals.To allow responses to differ between major and minor players in the indus-

try, equation (2) is rewritten as:

(3)

wherew= L (major player), andF (minor player).The PCM equations for major and minor players are obtained by substitut-

ing equation (3) in equation (1).

For major players, the equation is given by the following:

(4)


For minor players, the PCM equation is:

(5)

L and F denote major players and minor players, respectively, and hit are disturbances,

FEi and TDi are firm fixed-effects and time dummies,the subscript –i refers to rivals,adj MSi is firm i’s market share adjusted for imports,CU is capacity utilization, andFS is a vector of measures of financial status.

K/Y or capital-sales ratio is a standard regressor used as a control variablesince PCM is measured as operating profit margin rather than price and marginalcost. As such, the estimated price cost margin consists of both pure profit andreturn to capital.

The implied conjectural variations or ëijs can be computed based on theparameters obtained from equations (4) and (5). To assess the impact of tradeliberalization on the cement industry, import penetration ratio (MPR)6 which isused as a proxy for trade variable is incorporated in the equations. The H&S modelwith foreign competition measured by MPR is given by the following:

(6)

ANALYSIS OF RESULTSThe data used in the analysis are taken from the financial statements of 18 cementcompanies obtained from the Securities and Exchange Commission. These compa-nies are: Northern, Bacnotan, and Limay in Northern Luzon; Hi-Cement, Solid,Republic, FR, Continental, Rizal, and Titan in the National Capital Region; andFortune, Grand, Apo, Davao Union, Alsons, Iligan, and Pacific in Visayas andMindanao. All the firms in the industry are represented, except for Lloyds, a rela-tively small firm in the Visayas. The period covered the years 1994 to 2003.

As earlier discussed, the major players in the industry are Solid Cement, Hi-Cement, Davao Union, and Apo Cement. These firms had the highest individual

6 Import Penetration Ratio = [Imports/(Output - Exports + Imports)]

ALDABA 11

annual market shares that ranged from 10 to 15 percent during the period 1994 to2001. These firms also had the highest average market shares during the period 1999to 2003 and accounted for an average total share of almost 38 percent of the market.

Table 4 reports the descriptive statistics on the main variables used. ThePCM is calculated using the following formula:

(7)

Data on the Cost of Goods Sold are taken from the firms’ income statements.This account consists of the following cost items: power and fuel; raw, packing,and production materials; depreciation; repairs and maintenance; transportation;personnel; and others. As Table 4 shows, the mean PCM is about 14 percent withmaximum value of 64 percent.

Two financial indicators, borrowing ratio and cash liabilities ratio, are com-puted as follows:

(8)

(9)

The firm’s capital stock consists of fixed capital including buildings,machines, transportation equipment, and other fixed assets such as furniture,fixtures, and office equipment. The value of capital is measured by the replace-

Table 4. Descriptive statistics

Variable Obs Mean Std. Dev. Min imum Maximum

Price Cost Margin (PCM) 109 0.1443515 0.2312443 -1.120171 0.6443633Borrowing Ratio (BR) 109 0.3405189 1.166529 0 11.91405Cash Liabilities Ratio (CL) 101 0.3393691 1.195246 0.0014901 11.54943Capacity Utilization Rate (CU) 109 0.54456 0.2347538 0.0342494 1.25Capital-Sales Ratio (K/Y) 109 3.871203 3.771939 0.0145845 24.78601Import Penetration Ratio (MPR) 109 6.024269 5.985602 0.0806322 19.0635Market Share (MS) 109 0.0694324 0.049295 0.0014862 0.3045608Market share adjusted for imports 109 0.0651423 0.0465906 0.0014291 0.2971806 (MSadj)Effective Protection Rate(EPR) 103 1.033736 0.0596734 0.9198 1.2341Construction Growth Rate(CGR) 109 0.9047147 14.58636 -27.07994 23.31661


ment cost of capital, that is, what it would cost today to replace the existingcapital of the firm.

The other indicators include capital-sales ratio which is given by theproportion of the replacement cost of capital to net sales. Capacity utilizationis obtained by getting the ratio of cement production to total installed capac-ity.7 The mean is about 54 percent. Import penetration is the proportion ofimports to total domestic supply of cement (production less exports plus im-ports). The mean is 6 percent with minimum and maximum values ranging from8 percent to 19 percent. Market shares are adjusted by the industry share ofimports and the mean is around 7 percent. This is not different from the meanof the unadjusted market shares. The value of the maximum market share is 30percent while the minimum is 0.1 percent.

Equations (4) and (5) are estimated by least squares dummy variable(LSDV) regression technique. Panel data have unobserved, time-constantfactors that affect PCMit known as unobserved or fixed effect and idiosyn-cratic error or time-varying error representing unobserved factors thatchange over time and affect PCMit. A traditional view of the fixed effectsmodel is to assume that the unobserved effect is a parameter to be estimatedfor each firm. To estimate fixed or unobserved effects, a dummy variabletechnique is applied, hence, it is called LSDV regression. Under fixed effects,the intercept is allowed to vary across firms which may be due to specialfeatures of each company.

The results presented in Table 5 suggest that firm interaction, which in turndepends on the firms’ physical capacity and financial condition, is an importantdeterminant of price cost margin. For major players, the most important determi-nants are the major firms’ own physical condition (adjMSit*CUit) and own financialstatus (adjMSit*BRit) as well as rival major players’ physical (adjMSit*CU-iL,t) andfinancial (adjMSit*BR-iL,t and adjMSit*CL-iL,t) characteristics. As the results show,the coefficient on own market share is positive and highly significant.

Based on the minor players’ responses, the results tend to indicate that thephysical and financial conditions of minor players do not seem to matter to majorplayers. The coefficients on the interaction variables MSit*CU-iF,t , MSit*CL-iF,t andMSit*BR-iF,t all turned out to be insignificant. The coefficient on major players’ K/Y is positive and highly significant.

The results for minor players show that the most important determinants aretheir own physical and financial conditions. The coefficient on own market shareis negative but not significant. Own market shares interacted with own CU is

7 Based on PHILCEMCOR data and Cement Manufacturers Association of the Philippines's (CEMAP)2005 Report.

ALDABA 13Table 5. Firm conjectures as determinants of PCM

Dependent Variable: PCM Major Players Minor Players

adjMSit 7.230078*** -0.2231883(0.680476) (1.066197)

adjMSit*CUit -6.826383*** 2.322326**

(0.7300138) (1.11604)adjMSit*CU-iL,t 0.4797935* -0.1211166

(0.2940344) (0.3928477)adjMSit*CU-iF,t -0.1404564 0.0013463

(0.1370133) (0.2048648)adjMSit*CLit -0.215368 0.9518469***

(0.160818) (0.0584158)adjMSit*CL-iL,t 0.3693449*** -0.3797919

(0.1287209) (0.3216298)adjMSit*CL-iF,t 0.0077585 -0.0283554

(0.0226771) (0.0342034)adjMSit*BRit -0.7517635*** 0.3899996***

(0.2015318) (0.1487846)adjMSit*BR-iL,t 0.0691973*** -0.0264484

(0.0210149) (0.0590076)adjMSit*BR-iF,t 0.0043805 0.0243232

(0.0279925) (0.1370107)(K/Y) it 0.0194575*** -0.0228076***

(0.0040007) (0.0012047)constant 0.2528175*** 0.4425579***

(0.0287328) (0.0227305)R2 0.81 0.85Number of observations 582 1984

Note: Robust standard errors in parentheses.*** Indicates significance at the 1 percent level* * Indicates significance at the 5 percent level* Indicates significance at the10 percent level

positive and significant at the 5 percent level. Both the coefficients on own marketshare interacted with own CL and own BR are also positive and highly significant.Minor players’ K/Y is negative and highly significant.

Based on these results, the conjectural variations for major and minor play-ers are estimated. The simplest type of oligopolistic behaviour is the Cournot casewhere lij = 0. When lij is between 0 and -1 (where lij = -1 implies price-takingbehaviour), expectation of price taking behaviour is indicated. When lij is be-tween 0 and 1 (where lij = 1 indicates monopolistic or cartel behavior), the conjec-tures imply strategic behavior. Haskel and Scaramozzino (1997) indicated that apositive sign is in line with cartel/fringe models.


Table 6 presents the average CV estimates and shows highly significantpositive signs for major players’ conjectures on other rival major players, lLL andon minor players, lLF. The F tests suggest that the point estimates for majorplayers’ conjectures on the behavior of other rival major players, lLL and on otherrival minor players, lLF are significantly different from zero. These results indicatethat major players’ conjectures with respect to other rival major players and rivalminor players tend to show strategic behavior.

With respect to minor players, the results show negative signs for bothminor players’ conjectures on other rival minor players, lFF and major players,lFL. However, the F tests show that these are not statistically different fromzero, hence, the null hypothesis of Cournot or non-strategic behavior cannotbe rejected. This tends to validate our characterization of the industry withminor players having Cournot or non-strategic conjectures with respect to theresponses of rival major players and minor players.

Table 7 presents the impact of imports arising from trade liberalization. Theresults differentiate the impact of import penetration on the profitability of majorand minor players. The table indicates that for major players, the coefficient onMPRt has the unexpected positive sign and is highly significant. This result tendsto suggest that for major players, imports do not have a disciplining effect with thecoefficient on MPR indicating that high import penetration is associated with highPCM or low penetration of imports is associated with low PCM. For minor players,the same result is obtained. On the overall, these results tend to provide weaksupport for the import discipline hypothesis.

After the imposition of safeguard measures on cement imports in 2002, thevolume of imports dropped from a high of 19 percent (see Table 1) to less than 1percent in 2003. Table 8 shows that while these were already removed in 2004,

Table 6: Estimates of conjectural variations

CV Estimate F test: lllllij=0

lLL 0.0687*** F(1, 559)=19.97 Prob>F=0.0000lLF 0.0442*** F(1, 559)=12.06 Prob>F=0.0006lFF -0.01942 F(1, 1953)=0.99 Prob>F=0.3208lFL -0.02985 F(1, 1953)=1.69 Prob>F=0.1941lLL = lLF F (1, 559)= 5.22 Prob>F=0.0227lFF = lFL F (1, 1953)= 0.49 Prob>F=0.4821

lLL refers to how major players expect other major players to respondlLF refers to how major players expect minor players to respondlFF refers to how minor players expect other minor players to respondlFL refers to how minor players expect major players to respond

ALDABA 15Table 7. Impact of trade reform on major and minor players' PCM

Dependent Variable: PCM Major Players Minor Players

MPR 0.078*** 0.071***

(0.018) (0.025)MSit 9.869*** -0.015

(0.803) (0.943)MSit*CUit -9.426*** 0.670

(0.878) (0.862)MSit*CU-iL,t 0.268 -0.135

(0.257) (0.269)MSit*CU-iF,t -0.104 0.077

(0.131) (0.145)MSit*CLit -0.894*** 1.445***

(0.221) (0.042)MSit*CL-iL,t 0.0479*** -0.519**

(0.019) (0.220)MSit*CL-iF,t 0.003 -0.005

(0.021) (0.035)MSit*BRit -0.149 -0.053

(0.155) (0.109)MSit*BR-iL,t 0.320*** -0.052

(0.119) (0.030)MSit*BR-iF,t 0.007 0.046

(0.020) (0.105)(K/Y) it 0.0120*** -0.037***

(0.005) (0.001)Constant -0.189*** 0.181

(0.091) (0.120)R2 0.83 0.90Number of observations 576 1879

Note: Robust standard errors in parentheses.*** Indicates significance at the 1 percent level* * Indicates significance at the 5 percent level* Indicates significance at the10 percent level

import penetration has remained low while prices continued to rise from an aver-age of P120/bag in 2003, P147/bag in 2004, P158/bag in 2005, and P170/bag in 2006.Production and consumption were falling while construction growth was eitherlow or negative.

The results tend to indicate that in an industry characterized by strategicbehavior, trade policy does not affect competition particularly when coordinationis done at an international level. In a global economy where domestic and foreignfirms compete in each other’s markets, collusion in markets (e.g., cement) take the


form of market sharing where firms serve only the domestic market without interna-tional trade. With economic integration, it becomes easier for foreign firms toattack a domestic market and for domestic firms to retaliate by exporting in theopposite direction (Lommerud and Sorgard 1999).

There is a growing number of collusive agreements among cement compa-nies prosecuted in different countries. Mehta (2007) wrote that in 1998, the Euro-pean Commission imposed fines amounting to EUR248 million on 23 cement pro-ducers which included La Farge and Holcim. In 2003, the German Cartel Officesanctioned six companies including La Farge and Holcim for price fixing and im-posed fines of EUR660 million. Mehta (2007) further noted that in 2005, La Fargeand Holcim were among the three cement firms sanctioned by the Romanian com-petition authority for engaging in anti-competitive activities with fines reachingEUR26 million.The Taiwan FTC found its domestic cement firms including Cemexto have reached an agreement in June 2001 with companies from Japan, Indonesia,Malaysia, Philippines, Thailand, and South Korea not to export to each others’market without consent.

TESTING THE IMPACT OF IMPORTS BASED ON SCP MODELThe impact of trade reforms on price cost margins is further examined by applyinga structure-conduct-performance (SCP) model. The traditional SCP argues that aconcentrated industry (structure) will facilitate collusion (conduct) and results inmonopoly pricing (performance). This suggests that firms in a concentrated mar-ket, if protected from competition through barriers to entry, are expected to gener-ate supra-normal profits. Thus, firms operating in oligopolistic industries withlarge market shares are more likely to coordinate their pricing and output or to

Table 8. Recent cement economic data, 2000–2006

In million 40 kg bags Average Import ConstructionDomestic Price Ratio Growth Rate

Year Production Imports Exports Consumption (P/bags) (in %) (in %)

2000 299.0 37.6 33.6 303.0 126 12.41 26.32001 284.5 58.4 46.6 296.3 144 19.71 -5.02002 334.9 16.1 37.2 313.8 85 5.10 -23.72003 326.7 2.23 41.4 287.5 120 0.78 -2.32004 326.4 0.32 38.1 288.6 147 0.10 6.92005 309.2 2.9 63.4 248.7 158 1.17 0.92006 301.0 6.7 55.2 252.5 170 2.60 4.6

Sources: PHILCEMCOR data, Board of Investments Industry Profile, and Department of Trade andIndustry-Price Monitoring Unit.

ALDABA 17

unilaterally engage in anticompetitive behavior resulting in higher profit margins. The model is estimated using firm-level data and is extended to include a

trade policy variable. Two trade proxies are applied: effective protection rate andimport penetration ratio.

(10)

whereMSit: market share of firm i which is expected to be positively correlated with

the degree of profitability;EPRt: effective protection rate, which is expected to be positively correlated

with the price cost margin;Kit/NSit: capital-sales ratio, which is a standard regressor used as a control

variable in looking at PCM determinants;CGRt : real growth rate of the construction industry; andIit: firm and time dummies.

The same model is tested using import penetration ratio, or MPR.

(11)

Based on firm-level data from 1995 to 2003, the two equations are estimatedusing fixed effects method with time dummies. Random effects and feasible gener-alized least squares (FGLS) estimators are also applied. The estimation results areset out in Tables 7 and 8 with EPR and MPR as trade policy variables, respectively.

Tables 9 and 10 show that based on the three methods applied, the coeffi-cients on EPR and MPR are negative and positive, respectively, and are signifi-cantly different from zero. This tends to imply that cement imports do not have adisciplining effect on the domestic market. Based on the FGLS results, the coeffi-cient on market share is positive and significant at the 5 percent level. The coeffi-cients on capital intensity and construction growth rate are both negative andhighly significant.

CONCLUSIONSUsing the H&S model, the paper estimated cement firms’ CVs and examined theimpact of trade liberalization on the industry. The results show that conjecturesand firm interaction as measured by the firms’ financial condition and physicalcapacity are important determinants of PCM. PCM depends both on own andrivals’ physical capacity and financial condition. The CV results validated theearlier characterization of the industry that conjectures differ between major and

PHILIPPINE JOURNAL OF DEVELOPMENT 200818Table 9. Determinants of PCM (EPR as trade policy variable)

Dependent Fixed Effects Random Effects Feasible GeneralizedVariable: Method (robust Method (robust Least Squares

PCM standard errors) standard errors) (FGLS)

EPR -2.649672*** -2.357227*** -2.021508***

(0.7550611) (0.6391348) (0.7152694)MS 1.541314 0.9969583** 0.8978004**

(1.238792) (0.4809817) (0.3742343)K/Y -0.015091 -0.0139713 -0.0109947**

(0.0123802) (0.0142424) (0.0050102)CGR 0.0052202 0.0051436 0.0066544*

(0.0042423) (0.0036999) (0.0038011)Constant 2.672306*** 2.410122*** 2.059295***

(0.6929306) (0.5667444) (0.6972674)R2 within 0.5140 0.5100between 0.0867 0.0788overall 0.3323 0.3456

Autocorrelation F=13.454 F=13.454 no autocorrelationtest# Prob>F=0.0019 Prob>F=0.0019

No. of obs 109 109 109

Note:# For FE regression model, the modified Wald test for groupwise heteroskedasticity is used while

the Woolridge test for autocorrelation in panel data (Ho: no autocorrelation) is applied.The robust standard errors are White’s heteroskedasticity-corrected standard errors.

*** Indicates significance at the 1 percent level* * Indicates significance at the 5 percent level* Indicates significance at the 10 percent level

minor players: major players are characterized by the tendency to engage in strate-gic behavior while minor firms are characterized by Cournot behavior.

The results also indicate that under imperfect competitive conditions, par-ticularly when major firms have a tendency to engage in strategic behavior, im-ports do not seem to have a disciplining effect on domestic firms. Given the behav-ior of domestic market players, the inclusion of import penetration ratio in the H&Sregression model shows that for both major and minor players, the coefficient onthe import ratio is unexpectedly positive and highly significant. The same resultsare obtained in the SCP model which also indicates weak evidence to support theimport discipline hypothesis.

The above findings indicate that in assessing the impact of trade liberaliza-tion, market structure and firm behavior matter. The results imply that given strate-gic behavior by major players in the industry, imports do not seem to affect profit-ability and domestic competition. The ability of firms to engage in anti-competitive

ALDABA 19Table 10. Determinants of PCM (MPR as trade policy variable)

Dependent Fixed Effects Random Effects Feasible GeneralizedVariable: Method (robust Method (robust Least Squares

PCM standard errors) standard errors) (FGLS)

MPR 0.2829674*** 0.2517363*** 0.2158837***(0.0806355) (0.0682554) (0.076386)

MS 1.541314 0.9969583** 0.8978004**(1.238792) (0.4809817) (0.3742343)

K/Y -0.015091 -0.0139713 -0.0109947**(0.0123802) (0.0142424) (0.0050102)

CGR -0.1585028*** -0.1405092*** -0.1182544***(0.0438284) (0.0365126) (0.0429358)

Constant 0.0413451 0.069541 0.052062(0.1301338) (0.0877807) (0.0696948)

R2 within 0.5140 0.5100between 0.0867 0.0788overall 0.3323 0.3456

Autocorrelation F=13.454 F=13.454 no autocorrelationtest# Prob>F=0.0019 Prob>F=0.0019

No. of obs 109 109 109

Note:# For FE regression model, the modified Wald test for groupwise heteroskedasticity is used while

the Woolridge test for autocorrelation in panel data (Ho: no autocorrelation) is applied.The robust standard errors are White's heteroskedasticity-corrected standard errors

*** Indicates significance at the 1 percent level* * Indicates significance at the 5 percent level* Indicates significance at the10 percent level

behavior weakens the competitive effects arising from trade liberalization. In aglobal economy where domestic and foreign firms compete in each other’s mar-kets, collusion in markets, in this case the cement industry, takes the form ofmarket sharing where firms serve only the domestic market without internationaltrade. Hence, trade policy does not affect competition especially when coordina-tion is done at an international level.


REFERENCESAldaba, R.M. 2002. Of cartels and collusion: an analysis of the Philippine

cement industry. Toward a national competition policy for the Phil-ippines, edited by E. Medalla. Makati City: Philippine Institute forDevelopment Studies.

Bhagwati, J.N. 1968. The theory and practice of commercial policy. New Jersey:Princeton University Press.

Cororaton, C. and J. Cuenca. 2000. An analysis of Philippine trade reforms in 1995–2000 using the 1994 APEX model. PIDS Discussion Paper Series No. 2000-36. Makati City: Philippine Institute for Development Studies.

Cororaton, C. 2003. Analyzing the impact of trade reforms on welfare andincome distribution using a CGE framework. PIDS Discussion PaperSeries No. 2003-01. Makati City: Philippine Institute for Develop-ment Studies.

De Dios, E. 1986. Protection, concentration and the direction of foreign directinvestments. The Philippine Review of Economics & BusinessXXIII(1&2).

De Dios, L. 1994. A review of the remaining import restrictions. ResearchPaper Series No. 94-08. Makati City: Philippine Institute for Devel-opment Studies.

Fair Trade Commission. Competition Policy Information and Research Center(Taiwan) Newsletter Issue 1, Vol. 10. February 28, 2006.

Habito, C., P. Intal, Jr., C. Cororaton, et al. 1999. WTO and the Philippineeconomy: an empirical and analytical assessment of post-WTO tradereforms in the Philippines. An unpublished report submitted to theDepartment of Agriculture.

Haskel, J. and P. Scaramozzino. 1997. Do other firms matter in oligopolies? TheJournal of Industrial Economics XLV(1).

Imbat, M.M.C. and E.M.F Tanlapco. 1993. Influence of import competitionand seller concentration on the profitability of the Philippine manu-facturing sector. Philippine Review of Economics and Business30(2):234-55.

Indian Express Newspapers [online] (www.indianexpress.com/news).Lamberte, E. D., E. de Dios, D. A. Flores, J.F. Tabbada and E. Ramiro. 1992. Barriers

to entry study. Vols. I & II. United States Agency for International Devel-opment. Unpublished paper.

Levinsohn, J. 1993. Testing the imports-as-market discipline hypothesis. Journalof International Economics 35(1-2): 1–22.

Lommerud, K.E. and L. Sorgard. 1999. Trade liberalization and cartel stability. Re-view of International Economics 9(2): 34–55.

ALDABA 21

Medalla, E.M. 1998. Trade and industrial policy beyond 2000: an assessment of thePhilippine economy. PIDS Discussion Paper Series No. 98-05. MakatiCity. Philippine Institute for Development Studies.

Medalla, E.M., G. Tecson, R. Bautista, J. Power and Associates. 1995. Catching upwith Asia’s tigers. Vols. I & II. Makati City: Philippine Institute for Devel-opment Studies.

Mehta, P. 2007. Cement cartels: flavour of the day. The Financial Express. June 10.Roseberg, P. 2001. Interview published in the Philippine Star, 24 February 2001.

can imports discipline collusive firms?: the case of the philippine ... · can imports discipline...

Documents