the composition of foreign direct investment pptx

Trang 1

Publi

wp 5257 TL

POLICY RESEARCH WORKING PAPER 2786

The Composition of Foreign Direct Investment and Protection of Intellectual

Property Rights

Evidence from Transition Economies Beata K Smarzynska

“The World Bank

Development Research Group Trade

February 2002

Trang 2

Pouicy Riseance Wonnine Paren 2786 Abstract

‘Wile exiting erate has exained the impact of inlets property protection on the voline of foci ‘ict nsx EDD, es known abouts fect, ‘on the composition of FD nows Smarayosa

“drones he gueton empl, sing aun ent ` Union She fins cae weuk protection deters foreign

‘avestorsin technology intensive sectors dt ch heal ‘on tlle property ight The esas alo iniate ‘hata weak intellect property seine encourages Investors 0 under roe fotsing on dso ‘athe than los prodution The lave eet esent inal sectors, not ut thon lying bevy on intellect Proper proteton “Thspaperaproctof file propery protein Tad, Deslopment Research Grop—ispartfslnsrffortntherooptn amine theese on exonomic act Copisof the paper are val ce rome Worl Bak, 1618 Steet NW, Washington, DC 20453 Pez conte Paina Flew room MC3-333, telephone 102-473.2724, fn 202 512-1158, email addres plewtt@woldbalc org Do Reseach Working Paper aes posted onthe Web aps Jeson.wonldhank or The author may be contacted at annrayska@ worldbankorg February 2002 26 pages)

Ti amin ig of km pra ce of ar sh cp es cre sissy ete ei ea ay pes Te pr yeh ae te ha oc ole cach Te Fer ny arf heh Ds a mony np soft ok sce tn ‘sen hr

Trang 3

COMPOSITION OF FORFIGN DIRECT INVESTMENT AND PROTECTION OF INTELLECTUAL PROPERTY RIGHTS: EVIDENCE FROM TRANSITION

ECONOMIES Beata K Smarzynskea *

The World Bank

‘Keywords Foreign dts savestment, flies property sights, Transition economies, Fimlove data JEL classifieation:F23 034,

Trang 5

L INTRODUCTION

Protection of iatllectual property rights (IPRS) has Been prominent itm on interations policy agenda Despite the intraducton ofthe Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), many developing economies are not eager to strengthen thee [PR lepsition and its ‘enforcement faring thatthe loses resulting ffom this action would ourwegh its benefits This paper contributes to a beter understanding of potential gain from stronger IPR protection by providing ‘empirical evidence indicating that the extent of IPR protection ina host country affects the composition oF

foreign diet investment (FDI it receives More specially, this study fads that @ weak IPR regime eters foreign investment in high technology sectors where intellectual propery sights play an important role, Moreover, it tits the focus of FDI projects from manufacturing to distribution

“The relationship between IPR protection and FD is ute comples On te one hand, weak IPR regime increases the probability of imitation, which makes host county’ fess tractive location for foceign investors On the other hand, tong protection may shift the preference of multinational corporations ftom FI towards licensing As surveys of multinationals have shown, the importance of 1 protection varies between industries The concer aout the IPR regime also depends on the purpose of an investment project, being the highest in de case of RAD files and the lowest for projets fecusing exclusively on sales and disrbution (ee Mansfield, 194 and 1995)

“This pper investigates two hypotheses that emerge rom the above studies, Ft it tests whether forvian investors in IPR sersitve sectors (at indented by Mansfield, 1995) ae more affected by the extent of inllectual property proteton ina host country than investors in general Secon, it examines if the IPR regime influences foreign investor's choice between setting up production faites and

engaging in atvties focused solely on dstsbution,

‘A unique firm-level dataset used inthis study allows for a more in depth examination of this phenomenon than was possible in the ere Harti which concentrated mostly on aeeregate inflows and casestudies The dataset was compiled fom a worldwide survey of companies conducted by the European Bank for Reconstruction and Development (EBRD) in 1995 The survey recipients were asked Whether they had underaken FOI in tweny-our economies ia Essten Europe and the former Soviet, Union and it's, what rype of projets they were engaged in These responses were supplemented with, ‘information on frm characteristics and host county specifi variables

Trang 6

‘study, more crude in nature but taking ito account all IPR laws onthe books a well s her enforcement is also used

‘The empirical analysis confirms the Aypotheses thus indicating that weak protection of Iiellestel property rights ha a significant impact on the composition of FDL inflows, Fis, deter foreign investors in four technology-intensve sectors: drugs, cosmetics and healthcare products, chemicals; machinery and equipment, and electrical equipment These are the sectors in which, according to survey studies, IPRS play partculatly prominent role Second, weak protection encourages foreign {nvestors to setup distribution facies rather than to engage in local production Interestingly, this effet fs significant inthe case ofall investors, not jus those in sensitive industries, Final the results suggest that investors respond to both lawson the Books and ther enforcement The above findings are robust 19 contols fo privatization, transition progress, couption level and effectiveness of the legal sytem

{in addition to an inns interest in wanton, a focus on Easter Europe andthe former Soviet Union can offer insights nt the broader question of the role of FI in economic development throughout the world, While investment in other developing regions has been studied extensively, one Finding ofthat research has been the importance of previous investment experience asa determinant of curent FDI flows (se Hallward-Dremeier, 1996) Thus, the impact of cuent policy variables may be obscured and ‘overcome by’a long history of past policies, for which iis difficult contol Transition economies offer almost a natural contol since FDI in the region was negligible prior to 1989, Therefore, the results ofthis paper sugges thatthe importance of IPR protection in developing countries may have been understate in

past research,

This study is structured as follows The net section briefly reviews the related literature and formulates the hypotheses tobe tested, Section Il describes the econometric speifctions and the data fet In Section 1V, empirical results are presented Section V concludes the study

TL RELATED LITERATURE AND HYPOTHESES TO BE TESTED

“The connection benween technological capabilities ofa Frm and its decision to undertake FDI is highlighted in Dunning’ (1993) OL! paradigm, which explains activities of multinational corporations ia temns of owaersip (0), localization (L) and interalizaton advantages (0)! When sling its products broad, a fim is atlas intlly disadvantaged relative to local producers Thus, in order to compete fectively with indigenous Firms, a foreign producer must possess some ownership advantages They can systems, innovatorycapicity, trademarks, eputation, or ther assets, Ownership advantages aurea firm's ability {ake the form of 2 superior production technology or improved organizational and market

Trang 7

to enter the host country’s market, but do not explain why the forign presence should be established ‘through prodetion rather than expos This issue is, in tum, addressed by localization adhantages that srise due to differences in facor quality, costs and endowments, intematonsl transport and communication coss, overcoming trade restriction, and host government policies The lst advantage, Interalization, explains why 2 foreign fim prefers to retain fll control over the production process instead of licensing its intangible assets to local fms This decision may be atributble t9 high transaction costs involved in regulating and enforcing licensing contrasts,

Weak IPR protection inreases the probability of initaton, which erodes a fim’ ownership advantages and decreases localization advantages ofa hos county At the same time, @ weak IPR system increases the benefits of intemalizatin, since it is assosisted with a greater risk of the licensee's breaching the contact and acting in divest competition withthe see An inadequate FPR cegime, ‘therefore, deters FDI and encourages exporting A strong IPR system may also have a negative impact on

FDI by making licensing a viable alternative tdiret investment? Thus, the overall lationship between ‘he level of PR protection and FDI is ambiguous

The results of empirical studies exploring the impoct of IPR protection on FDI lead to mixed conclusions, Ferantino (1993) finds no statistically signteant relationship between the extent of US, tite sales in a freign country and that country's membership in an international patent or copytght ‘convention Similarly, Maskus and Konan (1984) who employ the Rapp and Rezck (1990) index of IPR inate and Pack (1987) index donot obi statistically significant results Lee and Mansfield (1996), onthe cher hand, show that he strength of a

protection as well as Primo Braga and Fink (2000) who use

county's IPR protection, as perceived by 100 U.S fms surveyed, is positively correlted withthe volume of US FDI inflows into that country Smith (2001) aso finds a pose correlation between sales of US, affiliates and the strength of IPR protection in host entry, None of these studies, however, looks atthe impact ofthe IPR regime onthe composition of FD inflows

Intellectual property rights do not play an equally important role in all sectors of even inal technology-intensive industries For instance, Mansfield (1995) mentions that IPR protection may be less veil in sectors such as automobile production, in which firms frequently cannot make ase of 3 competitor's technology without many complex and expensive inputs On the other hand, the IPR regime is likely to be important for sectors such as drugs, cosmetics and heath cae products chemicals machinery and equipment; and electrical equipment”

Trang 8

‘Additionally, a survey of US manufacturing firms conducted by Mansfield (1994 revealed that the importance of IPR regimes for investment decision depends onthe purpose of the investment projec For instance, inthe case of investment in sales and distribution outlets, only about 20 percent of survey ‘espondents were concerned with IPR protection In the ease of investment in rudimentary production ie, involving basic technologies) and assembly files, 30 percent of respondents viewed IPR rolection as important This percentage increased to $060 for investments in manufictring components and complete products and to 80 when RAD facilities were involved

Case studies from transition economies echo the concerns of foreign investors about weak IPR protection and are consistent with the survey findings For instance, Sharp and Bar-(1997, p 110) ‘mention that ICL (company producing synthetic organic chemicals) and Zeneca, a pharmaceutical company, “are wary of pracy and doubtful about transfering ether product or process know-how to these countries {Le ansition economies Both companies, however, recogize that eventually Con and Easter Europe andthe FSU (Former Soviet Union] willbe important markets That is why Zeneca is investing in developing its distribution links in high value-added areas such as medica

‘equipment and healthcare systems." Similar pcture emerges from the case study of Shel

“Shell provides know-how tis Russian parmers where necessary, But des not pass on anything it regards as commercially sensitive A relevant exampie is Shell's contract ‘with the Russian R&D testitute for Elemen-Organie Compounds (INEOS) to produce & ‘ew construction psi, called Nori Shell wll supply the chemical intermediates for rwducion, while the technology will be Russian There is no question ofthe Russians {ther supplying the intermediates or obtaining acess othe more up-to-date technology ‘used by General Electie forthe manufacture of Nori nthe United States” (Sharp and Barz, 1997, p 107-108

supplies and

‘Such examples are obviously nt resete to Eastern Burope andthe former Soviet Union Lan snd Young, (1996, p 73, footnote 9) present a case fom China: “Local staff wovk

of to foreign ailiates manufacturing detergents discovered the contents of production by repeatedly ‘ning the combination They then moved out to Set up thet own fms In only afew years, more than ten small local fms were manufacturing detergent™*

In the light ofthe theoretical prediction presened above a well asthe conclusions emerging from interviews with foreign investors, the followin testable hypotheses emerge (1) FDI in sectors relying heavily on protection of intellectual propery i likely tobe deterred bya weak IPR regime, Is not clea, however, that this should be te for FDI inflows in general (2) In counties with weak protection of nthe laboratories

Trang 9

intellectual property investors may be mor inclined to engage solely in distribution activities rather han in local production, These tw hypotheses ae tested inthis study

IIL ECONOMETRIC SPECIFICATION AND DATA Econometric specification

“To test the first hypothesis, a probit model of the determinants of investment decision is timate, The model is ofthe following form FDlig = if FOI, > 0 FDIig = Votherwise where por, XP din Kes +e

‘The dependent variable takes onthe value of one if fim is invested in country c, and zor if fin has not undertaken FI in country c Thus foreach frm the numberof observations is equal othe number of possible destination countries in the sample To contol for unobserved firm characterises, firm specific dummy variable dae included, Additionally, country specific explanatory variables X- are included the model Since the impact of IPR protection and possibly other variables is expestd to Afr between seetors, dhe mode allows fora separate coeicient for high technology sectors in which IPRs play more prominent oe tis achieved by interacting X, witha dummy variable or these sectors Following the survey findings of Mansfeld (1995), the IPR sense sectors include: drugs, cosmetics snd health care products; chemicals; mschnery and equipment nd lecical equipment, The exors ae corrected fora correlation between observations forthe same destination country

Trang 10

‘As mentioned, the Fist pat ofthe mel desis the investor's dsson to eter apaicuar tos county Asin the einton stove, we allow fora fen impact of ost country characteris ‘nfs in PR-sensve sectors FDI jg “Lif FOL, > 0 FDI jg =0ghenise where DI}, = Xj8 + XØ, + dạy Xcf, + 8c

The second prt deserts the die Between song áp poluclon fies ín cuhy € Coos accompanied by ditrbuion networks) and engging in a project foewsing solely on Aisibion, conditional on FDI taking pace The dependent vaable eqns one in the case of nanufaeving EDI n the right hands, both fm (X) and Rost country CX specific varies ae

included, The model allows fora different impact ofthe intellectual property regime on IPR sensitive

aNUrACTERING, tif MANUFACTURING}, >0and DI, >0

MANUFACTURING, = Oi MANUFACTURING, sOand FOI, >0 vien MANUPACTURNG), « X,8 + XB +p IPR toc

Assuming that (6 9) aed normal variables with 2ro means and acorlationcoeficient of we estimate these equations (pobit with sample seleton) simultaneously by maximum likelbood The tre are corecte for a corelation between observation forthe same destination county The number of observation in the FDI decison equation is equal to the numberof Firms inthe sample, multiplied by the number of destination countries in the sample In the second decison equation, the number of observations i equal tthe total number of FDI projet inthe sample, The later number i smaller than the former because not al firms invest in al countries

Data

Trang 11

included Responses were obtained fom 1405 fims that answered questions regarding their undertaken and planned investments in Eastem Europe andthe former Soviet Union Additonal, information onthe faction of the projects (manufacturing, disibation, representative office) was collected” The data Set does not include any information on the time when each investment was undertake, Since the magnitude of FDI inflows was marginal before 1989, the information collected pris mosly tothe period 1989- ‘945 Since the objective of this study isto explore the impact of government policies onthe magnitude ‘and nature of FD inflows, firms in the cl, as and coal sector, which are likly 1 be acted wo natal resource endowments are excluded ffom the estimations

Measures of IPR protection

“The key variable inthe repression isa proxy forthe IPR regime The indices of patent rights protection developed by Rapp and Rozek (1990) and Ginarte and Park (1997) are the two most popular ‘measures employed inthe literature The former index, while widely used, is inadequate for the purpose ofthis pape since it covers only five counties from the dataset and pertains tothe preransiion period

The Ginarte-Park measure, on the other hand, cover ten wansition economies and includes information for 1995? tn order to test he hypotheses using the full dat se, the Ginarte-Park index was extended to ‘cover nine more counties *

‘The Ginate-Pak index takes imo account five categories of patent laws: (1) extent of coverage, (2) membership in international patent agreements, (3) provisions for los of protection, (4) enforcement ‘mechanisms, and (8) duration of protection Each ofthe categories isasignd a valve between 0 and {, and the unweighted sum ofthese values constiues the patent rights index (see Ginarteand Park, 1997 fr 1 detailed desriptién), Thus, the index ranges from zero to five withthe higher values indicating a strange level of protection, The index refers to 1995 othe closest year for which the information was 777 of he survey sspondents were chown for indepth interviews whose resus ae sessed in Lankes and enables (1996) EECs sa te FSU weee vitally closed to frig invenment bore 1989 (se Meyer, 1995; Durning and Rojec, 1995; Hanya, 197), The igus preseted inthe ube below teflet is stuaoe Cea i

[tora —— Souree Dang 37) sơ

Trang 12

available, Table | lit the index values Te highest seore inthe group of countries under consieraion ‘was obtained by Hungary (3.75), while the lowest seore of 2.52 belongs to Unbekstan and former YYogoslay Republic of Macedonia The average value ofthe index i 3.04 Fr comparison, the mean value ofthe index for 110 counties rated in 1995 was 2.67 A postive coefficient onthe Ginrte-Park index wil indicate that stronger patent laws are associated with a grester probability of FDI being undertaken, Th sdvantges of the Ginat-Park index are great level of detail and taking into account the treatment of foreigners Its main disadvantage is the fat that it focuses on the laws present on the ‘books but it doesnot eapure thelr enforcement

Since the issue of enforcement may havea crucial impact on forcign investors’ decisions, the paper also employs another index of IPR protection developed specifically fr this study This simple index captures both the legislative and the enforcement aspect of the IPR regime It is based on the escipions of IPR regimes provided by the International Intellectual Property Alliance in their recommendations for counties tobe placed on the U.S Special 301 Watch List These descriptions include the isue of enforcement and pay special attention to trademark and copyright laws Note thatthe actual placements onthe Special 301 Wath List have not been used in developing the inde, since they depend not only onthe extent of IPR violations ina specific country but also on the importance ofthe country tothe U.S interests, Again 1995 is used as a reference point inthe rating The table below resents the rating criteria, A higher value of the index eoresponds 0 stonger IPR protection, thus & positive coefficient on this variable i expected Table 1 lists the values ofthe index The correlation

‘beeen the Ginrte-Park measure and this index 57

Tadex of Ttelecaal Property Protection Dever

‘Coss adequate PR Tegilation present Oy Wie end of 1995, Some enforcement eflorts undertaken

F [No enforcement efforts undertaken [Close adequate IPR legislation presen Oy Whe end of 1995, T [ack of adequate IPR legislation at th end of 1955,

‘urea consructed by te author bsed on IPA Special 30] Recommendations

Other contol variables

“The IPR regime may be correlated with ether host country characteristics such asthe overall progress in reform effectiveness of the legal system, corruption level, privatization policies and openness to trade, Therefore, addtional variables ar inched inthe regression o contol for these Factors

Trang 13

‘Multinational corporations are les likely to invest in vsky and unstable counties, and the peeeived riskiness of Easter Europe and the former Soviet Union has often been cited a a factor discouraging foreign capital inflows Lankes and Venables (1996) find negative association between [EBRD transition indicators and county risk as perceived by the interviewed firms, with the rank csoelaion coefficient equal 9-89 The tanson indicators rate the progres ofa county's reforms in the fol ving areas: price liberalization and competion, ade and exchange system, large-scale privatization, small-eae privatization, enterprise restructuring, and banking reform See EBRD (1994, p 11) for detailed description Thus, in the empirical analysis, the average ofthe EBRD indistors i sed

15a proxy for risks associated with undertaking FDI in a given host country Since the higher values of the vanstion index indicate @ greater progress in reform, one would expect to observe 8 postive soofeienton thisvaiable

urter, the effectiveness of the legal system is controled for using anther indiator produced by the EBRD (see EBRD, 1995, p 103) This indicator, ranging frm 1 104°, assesses the extent fo which legal rules affecting investment ae clear and accessible as well s adequately supported administratively nd juicilly, The value of on i assigned to counties where legal rules are usually very unclear and often conadictory, the availability of independent legal advice is limited and the administration of the Taw is substantially deficient The highest value (4*) is asigned to counties with clear and readily sscerainable laws, sophisticated legal advice available and well functioning courts Not the maxim score achieve by the counties in the sroup is 4

“Moreover, a measure ofthe extent of corrupt practice in the country is added tothe model The roasure is the 1999 Transparency Ineratonal Comupton Perception Index which pools information fiom ten diferent surveys of business executives, isk analysts and the general public The eriginal index ranges between 10 (highly clean) and 0 (highly corrupt) To facilitate interpretation of he results, were: scaled the index inthe following way: ce-scaledT index = 10 original index ‘Thus, a hiher index ‘alu corresponds higher evel of eonuption and a negative coefficient is expected”

Trang 14

Included in the model to conto for tariff jumping FDI The data refer to 1993 and come from the EBRD."

‘The exiting literature finds the host country’s market size tobe an important determinant of FDL inflows (see Dunning, 1993; Caves, 1996; Braunerjelm and Svensson, 1996) Most studies show that a large markt size encourages FDI inflows Therefore, the model includes GDP per capita which s a proxy for the purchasing power of local consumers and the population size whch reflects the potential size of the market Both variables come from EBRD (1994) and refer to 1993 They are entered in the logarithmic form nally, we contrl forthe corporate tx ate as higher taxation is likely to discourage investment Te figures (expressed as percentages) come fom PriceWateshousePaineWebbe, If several rates apply, the highest one was used.”

‘As explained above, i is necessary to include fim specie variables when testing the second ‘hypothesis Thus, standard variables found in most FDI studies are included inthe model These are frm sine (easred by the firm's sales in US dollars, entered a logarithm), R&D intensity (measured by RAD outlays as a percentage of net sales), advertising intensity (proxied by selling, general & ‘administative expenses asa percentage of net sles” and a proxy for produeton đnifiedion (the numberof four-digit SIC codes describing a firm's seit) Al information on frm characteristics was ‘obtained from the Warldsope database and pertains to Fiscal year 1993 rom 4/93 o 394) Aditonlly, ‘we control far investor's regional experience, proxied by dummy variable indicating whether a fm had

trading relationship with the region before 1989 The las variable comes from the EBRD survey TV EMPIRICAL RESULTS

Inpoct of IPR protection on probably of FDI aking place

Trang 15

postive sighs and are significant at least at the five percent level Inthe last egression bot coefficients ‘on the Ginarte-Prk index are significant

‘The ether variables also have the anticipated signs Population size is found t havea positive ‘impact on FD inflows in al industries IPR sensitive sectors donot appear tobe affested differently by this variable, GDP per capita is positively elated to FDI inflows and in majority of eases, it des not

affet the high technology sectors diferent As expected, progres in transition, greater effectiveness of the legal regulations governing investment and more advanced privatization process increase the probaly of FD in ll sectors Higher level of corruption and higher corporate tx rats, on the other and, deters foreign investors The coefficients on the interactions ofthese variables witha dummy for IPR sensitive sectors are not sigifcant The openness measure asa significant impact only onthe PR sensitive sectors A posible explanation is that fim in these sectors may be more reliant on imports as they tnd to transfer oly part of (rather than whol) production proces to the region for fear of losing thei intangible asset

‘Table 3 presents the estimation results with the stcond IPR measure, Unlike the Ginate-Park Index of patent rights protectin, this index captures both the legal and te enforcement aspect ofan IPR regime Iti also broader in scope as it persia to IPRS in genera ater than just patents As inthe previous table, we find that stronger IPR protection increases the probly that multinationals inthe four sensitive sectors will underake FDI In five out of sx cases, the coefficients are significant atthe five or one percent level, Additionally, in four regressions, the srenth ofthe IPR regime affects nt only te sensitive sectors bt also al investors The signs and significance levels of ether contro variables are ‘similar to those found in Table

‘The reason why all fms, not jus those in IPR sensitive sectors, may be affected by the extent of imelleewal propery protection is that an TPR regime may also play a signaling role As Lall (1997, p 24) point out“ the “signaling value” of the intellectual propery regime Its become extremely important in recent years In general, counties that seek to atract technologyintensive foeian investment aso offer strong protection to those investments.” As the eesuls in Tables 2 and 3 sugsest, sinalng takes place only ifthe legislative changes are accompanied by enforcement effort

Trang 16

the earlier findings are quite robust In ten out of twelve regressions, the IPR measure i significant and ‘bears the expected sign Al egresion have a high explanatory power

In summary the empirical analysis indicates thatthe tenth of patent aw as wel asthe overall love of IPR protection (both Iw onthe books an their enforcement affect FI inflows in several high technology sectors where, as surveys show, IPRS play an important role, Moreover, there is some evidence thatthe overall strength of the IPR regime and its enforcement influences the investment ecision of multinationals ative in other sectors a wel

[Impoe of IPR proection on the choice of project ftion

‘Table 4 presents the empiial resus ftom the test of the second hypothesis, As mentioned above, the hypothesis was tested by looking jointly at two decisions: (1) whether or not FDI is taking place and (2) conditional on an FDI project being undertake, whether i involves sting up production facilites or fooasss solely on bildng distribution networks The resus ofthe investment desision with respect to host country characteristics are consistent with those found inthe earlier section, One interesting change is that the new model suggests that firms in IPR sensitive sectors are more stonely etered by corruption in host country than fiems in other industries This may be associated with investor's fear that inthe case of legal disputes on, for instance, patent iaingement higher level of corruption will ower the chances that the dispute wil be adjudicated fly." The coeffeients o im characteristics also have the expected signs Namely, the data suggest that that larger fms and those with gfewer Ínangjble assets, regional experience and more diversified production are more likely to ‘undertake investment in the region,

‘The hypothesis of interes is supported by the data, AS Table 4 indicates, foreign investors are ‘more likely fo engage in leal production, as opposed to focusing solely on setting up distribution networks, in counties with suonger IPR regimes, The relevant coefficient (Gnare-Park index) is statistically significant inal regressions Interestingly, this effec is significant for al sectors and does not appear to be suonger in the case of IPR sensitive industries.” These findings are consistent withthe

survey evidence provided by Mansfield (1994) and the Zeneca case duy cu above,

Trang 17

explanation shat countries with higher GDP per capita tend to have higher bar costs which make les production less dưacite The probaly of manufacturing FDI is positively affected by the tansiion ogress Since setting up a production plants more costly than sting up a distribution network alone, i 's not surprising the foreign investors choose the Former option in countries tha appear tobe more stable đe to an advanced reform process Manufacturing peojects are also mote Hiely to tke place in economies where the privatization process is more advanced as it Wings opportunities for acquiing domestic production facilites Corporate tation and opeaness 10 ade do not appear to have a

scaly sigifcam impact

‘Turing to investor characters, the findings india that fms possessing more intangible assets, as measured by R&D and advertising intensity ar les likely to undertake manufscturing project This may be due tothe fact thatthe potential for knowledge dissipation is greater when the production lakes place ina hos country than when the inal products are imported, Fim size appears to be, albeit nt ‘very robustly positively corelated with the probability of a manufacturing projet To the extent that, larger fms have more resources fr investment, this nding Would indicate thatthe choice between the to types of activites is affected by financial consti, Finally, we find that manufctring projects are ‘more likely to be undertaken by Fins without previous epinal experience

“Table S presents the results obtained using the other IPR index While inthis case, the support forthe hypothesis much weaker, other results ae broadly comparable with those in Table 4,

{Asan aditona robustness check, we estimated a multinomial logit model wit the lft hand side ‘rable eflectng three aptons avilable to a pote investor: () no investment at

(i investment solely in disibuton networks; and (i) investment in production facilites possibly accompanied by istbution networks Appendix Table 2A preseas the results Since the choice between manufacturing and non-manufictutng project sof interest tous, the resus represented relative to option (i, that is Jnvestmen ia distribution networks As the Fst part ofthe table indicates, a seven out of eight cases, the coefiient on IPR protection is postive and statistically sigifcam, indicating that stronger IPR regimes increase the lkethood of FD in production faites telaie to distribution only projects IPR protection, however, doot not appear fo have a statistically significant impact on the choice between investment in Aistrbuton networks and no investment at al

‘Samming , the empirical ress indicate that weaker protection of intellectual property iscowrages foreign investors fom undertaking local production and tits their preferences towards

Trang 18

project focusing on distribution alone, This isthe cae forall investors, ot just those in IPR sensitive V CONCLUSIONS

Governments ll over the world compete fiercely to ateact forcign dtet investment hoping that ‘multinational corporations will bring new technologies, management sil nd marketing know-how In corde to create an investment fiendly environment, its inmportant to understand the factors that influence FDI inflows as wel asthe determinants ofthe composition of such flows This study sheds some light on this issu by examining the impact of PR protection onthe structure of FD inflows

Unlike the earlier literature, which focused on aggregate FDI flows, we employ’ @ unique fim= level dataset describing investment projet in Eastem Europe andthe former Soviet Union Since this region was virvally closed to FDI before 1989, its sudden opening to frcign investment can be compared toa natural experiment Therefore, the dats set wed inthis study presents & unique opportunity to estimate the eect of IPR protection on FD in the absence of investment history It is possible that in ‘eatir studies, the lack of controls for past policy variables and investment history has obscured the

Jimpat of IPR proteston on FDI

‘Both hypotheses tested inthe study find empirical support Fis, the data indicate that investors in sectors relying heavily on protection of inteletul property are deterred by a weak IPR regime in 4 potential host country There is also some evidence that wesk IPR protection may discourage all ‘investors, ot just hose inthe sestive sectors Second, the lack of TPR protection deters investors fom

underaking local product

Interestingly, this effect is presen in all sectors, not only those relying hevily on IPR protection,

Trang 19

Bibliography

‘Ades, Alberto and Rafael Di Tella, 1997, "National Champions and Corption: Some Unplesant InterventonistArthmetis." The Economie Journal OT: 1023-1082 Baldwin, John, 1996 “The use of intellectual property from the innovation survey,” mimeo Statistics Canada, rights by canadian manufictrng firms: Findings ater JW 0d, various years, World Patent Law and Practice New York: M Bender,

Braunshjelm, Pontus and Roger Svensson 1996, “Host country characteristics and agglomeration in foreign direct investment." Applied Economics 28: 833-840 ‘Caves, Richard E, 1996, Multinational (Cambridge University Pres Enterprise and Economic Analysis, Second Edition Cambridge: Dunaing, Joho H 1991 “The prospects for foreign direct investment in Eastern Europe,” Discusion "Papers in International Investment and Busines Stidies, No 155, Universi of Reading Dunning, John H 1993, Mulnaional Enterprises "Adsizon-Wesley Publishing Company and the Global Economy Wekingham, England Dunning, Joha H and Matia Rojec 1993 Foreign Privatization in Central & Eastern Europe Ljublana, ‘Slovenia: CEEPN, EBRD Transition Report London Various yous

Estrin, Saul, Kiely Hughes and Sarah Todd 1997 Foreign Diect Investment in Central and Eastern Europe: Multinationals in Transition London and Washington: Pinter! Royal Insttte for Tuterasienal Afi,

Fereanno, Michael J 1993, “The effect of intelectual property rights on intematioal wade and investment," Welewirschaflches Archiv 129: 300-31 Garisoa, David, o8 (various years) Iwellecual Property: Eastern Europe and Commonwealth of Tidepende Stat, New Work; Oceana Publications, Ine Ginare, Juan C, and Waker G, Park 1997, “Determinants of patent rights: A cros-national duậy” ‘Research Policy 26 pp 283-301 Grubaveh, Stephan G, 1987, Determinants of direct foreign investment, Review of Ecanomes and Statistics 6, 149-182 Hallward-Driemeiee, Mary 1996, “Understanding foreign direct investment by Fems: Market pull, cost push and knowledge accumulation." mimeo, MIT Hunya, Gabor 1997 “Large privatisation, restructuring and foreign deet investment” in Salvatore Zecchini, ed Lesions from the Economie Transition Cental and Eastern Europe in the 19908,

Dordrecht, Boston and London: Kluwer Academie Publishers, pp 275-300, Internationa neleewal Property Alliance Special 30) Recommendations, Vatious years

Tal, Sanjaya 1997 “Investment, technology and international competitiveness” in John H Dunning and ‘Khalil A Hamdan, eds, The New Globalism and Developing Countries Tokyo, New York and Pars United Nations University Press, pp 232-9

La, Ping and Stephen Young 1996, “Foreign direct investment and technology wansfer:a case-study © Foreign direct investment in north-east China." Transnational Corporations 51) 57-83, Lankes, Hans-Peter and A.J Venables 1996, “Foreign direct investment in economic transition: the changing pater of investment,” Economics of Transition 4(2): 331-347, Le, lang Yeon and Fain Mansfield, 1996 “otllectual property protection and U.S foreign đức! investment” Review of Economies and Statistics 78(2): 181-186 Mansfield, Edwin 1994, “Incllecual property protection, foreign direct investment and technology wane,” IFC Disession Paper No 19 Mansfield, Edwin, 1995, "Intellectual property protection, foreign divect iavestnent and technology rane,” IFC Discussion Paper No 27, “Markusen, James R 1995, "The Boundaries of Multinational Enterprises and the Theory of Internationa] “Trade,” Journal of Economie Perspectives (2): 169-189

Trang 20

Maskus, K and D Konan, 1994 “Tradevelated intellectual property righs: issues and exploratory resus," in A Deardorff and R Stern, ed Anabical and Negodating Issues inthe Global Trading Sytem, Ann Atbor: University of Michigan, pp 401-446

Meyer, Klaus 1995 "Direct foreign investment in Eastem Europe The Role of Labor Coss,” Comparative Boonomic Shades 313) 69-88 Oxley, Joanne E 1999, “‘Instatonal envionment and the mechanisms of governance: the impact of intellectual property protection on the sucure of ldecfim aliansee” Journal of Economic

Behavir and Organization 38: 283308

PAIZ (Polish State Investment Agency), 1995, Major nestor Lit Warsaw

Primo Brags, Carlos and Carsten Fink, 2000, “Intemational transactions in intelectual property and developing counties,” International Jonna of Technology Management 1X2) 35-56, Rapp, Richard T and Richard P, Rozck 1990 "Benefit and costs of intellectual property protection in developing counties,” Journal of World Trae, 245) 75-102 Sharp, Margaret technological epabiies in Central and Easter Europe David A, ed The Teclalogy of Transition Science and Technology Policies for Transition and Michael Barz 1997 “Multinational companies and the transfer and iffision of new and the former Soviet Unio” in Dyker,

Couns Budapest: Central European Unversity Pres, pp 95-125,

‘Smaraynska, Beata K, 2000, “Technology Transfer and Forega lnvestoes' Choice of Entry Mode” The ‘World Bank Policy Research Working Paper No, 2314 ‘Smarzynska, Bats and ShangJin Wei 2000, “Corrption and composition of foreign dct investment Fem level evidence” NBER Working Paper No 1969, October

Smith, Pamela J 2001 “How Do Foreign Patent Rights’ Affect US Expons, AMiiste Sales and Licenses? Journal of Pematonal Economies 58 411-439 “The World Bank, 1996, World Development Report 1996, From Plan to Market New York: Oxford University Press ‘Wei, Shang-Jin, 1999, "Does Corruption Relieve the Burden of Taxes and Capital Convols for Intemational Investors?" World Bank Pliey Reseach Working Paper No 208, ‘Yang, Guifang (Lynn) and Keith Masks, 2000 “Intellectual property rights and licensing: an econometic investguion,” Weloirschaflches archi, foteoning Zloch-Chrisy, liana 1995 “Economic tansformatian, external ibalances and political risk in post- ‘communist Eastern Europe” in Zloch Christy, ana, ed, Privaization and Foreign Investments

in Eastern Bwope Wesiport, CT and London: Praeger, pp 215-238,

Trang 21

Table Measures of IPR protection

] Ginarte-Park Patent | TPR Index based on Rights Index peer man 7 IAzetalin 1 lacus” — 3p 237 2 ' — 3i [Cech Republic 59 3 sons 286 2 IPYRMseedeist 23 keo hungary 300 315 3 ' Ikerbisene 30 ' lenaaan haw 2s z 1 itn 2s 1 Moldoeet | 300 2 Poland 32 3 Romania 27 1 Russia bo 2 —— 3m 3 Stover xe aiken 1 [rurkmenisan 2 (Ukraine jurtekisune aot 232 2 L

Trang 22

‘Table2 Probe esl with Ginarte-Park index of patent rights protection

(Gina Pak ade (040) 0300280 (034) (0) EO (0) (029 (023 Bấm PR sesiive* Ginane Pa indevÌ 040% (048 — (049 04m 0049 oss (02 088 G49 031 |Gprperepia (0 000 ⁄008 01) (049 648 om (87 oz (009 046 |PR sesdiwe*GDPpereapia | 012 (0 049 010 0 005 010 001 019 001 00489 038

|Popuaton (005) os2 00) oss (008) ose (008) 0610 (006) 043m: (005) 040% lip sensitive * Population (099 000 (00) 001 (006) 009 (0608 002 (00) 000 (088 042

[Progres in reform om (018 .- HERsenidve* Progestin tem | 016-010 6018 (049

(corporate tax rate oor (08) IPR sensitive * Corporate tax rate 001 (00) Legal etetveness (048) 02 IPR sensitive * Legal effectiveness (008) 03 lCampien ape (0.08) [PR sensitive * Corrption (005) 002 [Privatization (0.00) oa PR sensitive * Privatization (089 9.00 lopenness IPR sensitive * Openness fo, ob rs rr TẾT” lcni lbmb> ch? ae 000 HN 000 5A 000 09 000 LS 008 lpeesdeR Log Likelihood CR0 — -IBĐI — -1480 086 036 034 2188 — 21739 034 038 "heo signin a 5s Tove, "7 a * at 10% Sand ors arn pares

Trang 23

“Toble.Probi results with IPR index

IPR index (040) om 608) aa (046) Oe 6640 oa (60) OO (oT) UPR sensitive * IPR index (60) D2 oss 60) (00 033 0A) 038 = os 020.12) 003 {GDP per capita (049 02 ogo A0 649 nas (00) (609) 043 JPR senstive* GDP pereapita | 00 (049 00) 0ø (049 ot 008 09) 001 (610 (629 048 Population (004 eas (007) oar™ = 08) (08) oom (009 oat IPR sensitive * Population (09) (00) oo (00 001 (00) oor (069 00 (069 048 [Progress in reform 6048) os oem" 6049

` ) | 01) 016 (corporate mx rate oz (oot) iP sensitive * Corporate ax rte 02" (089 LLegn fletivenesr (012) oa IPR sensitive * Legal effectiveness 010 | (007) {corruption IPR sense * Conption Privatization (088) — 069) Bs IR sensitive * Privatization @on 000 eon 096 lopensess -006 om IR sensitive * Openness 629 080% Isao Tae ees SRT

len rab> Chi mol 09 $267,693 000 000 TRỤ 000 S60 000 800 000 Pseudo R Likelihood 1G99 15955 0 042 0 _-16502_16034 1612216067 04 044 044

dente sg wc eve, 7 5%, * 1H Standard evar re apres

Trang 24

“Tabled Bivariate probit with sample election Manufacturing ve detrbuton projets Ginarte-Park index INVESTMENT DECISION ‘Ginate-Park index aa 609 ©» 019) 039 mỊ {PR sensitive *Ginare-Pack indy (09) 033 “008 019 CDP pec capa | t0, 029 19) am IPR sense * GDP pr capita 0009) -00 010 r1 Population (00, _ 009) ‘or {PR sensitive * Population (099 0n 06) ái Cemepie R aecative* Corton 047 1 20+

(006,

Comte ae oon 000 oot 0000

IPR sense * Corporate tx (000, 002 (080, ‘000 | Proeninelem on ‘acres

IPR sensitive * Progress in refom| ` oy IPR sensitive *Pevatiation

Openness

IPR seneve* Openness

Firm size oon) arrears 00) an ars

Trang 25

[DECISTON REGARDING RORET ae de oF VP ANUFNCTURNG WS DSTRITTION) oe aar a tab ir senie* Gane a ima 90s Am AM A0, sat đổ ônh aya) ay de HC 0B 636 oy eh 005 củnh Aêồ 016 a8 — om cố đâu oo 80s aay ae pert oo oa Progen Piven oa eam — cam tia im size | “007 iam oy) 0.06 0.06 aaa đl6ttt 06899% aD ney ^^” ay ay ay yan Aves cages anaes antes ay” ay 0A 6 ay Am aaee "hs ' ẻẻẽ nhs EU ayaa ei expos came games agree oaaee 036 Am iy ain — adm dm irate ay ah pea me 38 nan: GR qwnstesortno-o) | Zoster cam Sn pose a rr 8m Noro Soo yas ta Ne etae Qe bo bef mon Ie sft at goa mm âm wo âm |

Trang 26

‘Table Bivariate probit wid sample selection Manufacturing vx dtibation projects IPR index {avast ESTO TPRinder ae am SR— 8 G) G09 độ 0i) 000 PR seine time (0 605 G0) 0Á (009 60c đán gie 000 00g cepa Goes “Gages SD Ry

609 609 (0 PR seeive* OD? pep "` or 0M 00 Pepin (ay ace “aes Gane @09 630 PR sei * Poplin 6A oy `" 6000 contin Sie Gite TT aon

IER serie * Compa straws States Go aay

como tx 0) So @on Sop 01 60) IPR * Copia @ay co, 600 oat aan, “oa Prope areorm a eu IPR seni * Popes ia eto Sor om mtn PR sete * ration | se PR env * Open Fem se am "- - aon oy hiep meay đốn TỶ < (án |Aereglmesty e000) fant 66+

rte dein Gay 400 08Ế% oy ay 680 Rein rine đáo G0n ¬- 9 come ‘ Sse Sanne “Gagne sáp “he abe cones onthe owing page

Trang 27

nn a Gay a a5 “ơn [mem inder " 2 oma a) ay GP rere aH aise estes ane đ yaa, comport te -a6 (8, Pros infor ais) oats — or ‘om Openness aa ow mg fants ame gen anes mm HH

Trang 29

‘Table 2A Multinomial logic results Manufacturing vs distribution project vs no FDI

(DLN MANUFACTURING FACILITIES RELATIVE TO FDIIN DISTRIBUTION NETWORKS] ‘ina Park index @3y (02 (39 (030 Oss ose agg 090"

TPR ides (08 (02) (019 019 048 031 si bAeee |apP «si -Ư@deeh arson 029 (025, na 0030 02) 0029 (03) 2H lẽ

Population (00 GO) OH IT) (00) 0009 (000 (009 asa 041921 naires oats 027092 dares 640As+ 047, Progen efor 02) 6001 030 (036 0839" 098%+ 039 039 629 ĐỀU - 080 agate Corton an 1st (08) G09, G016 (G1) 2993 Corporat x (000 oo oars (000, Privatization eon oe any om Openness an oy a0 om, Firm ie (0089 (089) (009 00: 061 60 (089 om | RAD ineasty (009 (003) (0A9 nisi 913 oases oases (09)

— testy tases ase gages won @oH (060, vos (oo Production dverifeation | 001 00 001-001 0089 (009 (009) on) oo Resins experince om OH OH) GI OH OI OI) OD 79" 270.068 79 -0609%%- nese!

Trang 30

[NO ROT RELATIVE TO FDIIN DISTRIBUTION NETWORKS [Ginarte-Pak nde (040) 6029 030 029 6E 0Á 08 ODT

PR má am 0) (05) 009 oats oP per ean 1) a9) on 1s 24g 96s 608608 sore tin

Popultion cm eon 050 0.46 08 6m -a4gee ‘oon Progress in reform -800%% nesree my a2 ose om 080, 003 (cemsior or as 68) 028

lcepsaee (080, oo Privtation

Openness

Fi size oases 99) ogress 202A nares) (009 (000 ò0 RAD intensity 009+ eon, on, -0899% -o.0sees nose) 0U 60D (00D | |Asxensing nemiy (008) oes oouers eo 204991 04ers 09s (000) (000 @00) Production diversitention (009) ose (ony) see (00H 0H 0009 | dows 000-089 | Regional experince oan 26 sane aasree nase cam ae oH Gy oH C1 (00 | consant isgores 136619 729995 1811995 1610999 149 309] L 9) 3H) Q4) OTH đ30 (A0 Q69 (669)

sưa TRE HẾO 780 55 0N 5 595 a0 Ip=seeet ‘7 center sient To Teel, 15," at 1, Sandan eos ae preted arene 06 ĐẸP 02T 027 030 029 035 0

Trang 31

Policy Researen Working Ps TH Anher

.MBSSTSS Inesusly Aegrsen, HoiP, neque, a Hoh Achweret “dan Magsalf `WES2785 Atono, Pedepalen ard Leaing Gunnar S.Esaland in Arganthe Seas: Figs ana Deon lever

Thor Inpcaons fr Dacaraizaton

\WPS2TE7 ChlaLaber The Role alincome Varsity sneEeersla Co len Robos Gah Raa H Deen Gros Serton of Counties

`WES2788 Trade, orion Exnange and xery Jesper Jensen Pale intelslame Repubice! Deve Tar an Rdoem AgeAos Econene

rpc nd ington he Poor

`WBS2788 Innutzalon n Dorekglng Gourtos: Varin Gaunt Tre Poee and Orpanastens) Pahsndesisghan `WESBT7D Dewmssng and Posictety Gxa Marin Rama Ine Pubic rd Paste Sectors Constance Newman

siCoomta

\W52771 Exchange Rate Aepracaton, Labor Nosbet Fast Naa Figo, and inomaity area Fugazen

—

weser72 Govorrance Mater I: Updied ari Keutmarn ——— An Kon Pablo Zao Laban MBS2T7A Hoosehod Erdepreex ‘Sova Grown se Ling n ram Win P.M Vive lorahan HauglSon

—

.MÊS274 ChdLabrn ranetonh Vatnam - EleEơnends `BS2T75 Peleneeltealh De zalo Pravin K Te Veta ais 7.98

Vana iin Standards Survey Data

Trang 32

Polley Research Working Paper Series TH

'NPS2778 Tedteleayan Em Bafertanae 'WPS27T8 Techrlby and Ski Derr ‘nono

Anher

Gladys Lopes Aewveto lays Lop Acoveds 'PS278) Dolemman inten of Tecmeloy Adoption Glas Lipas Raced WPS2761 Mart Trangpon ENeeey Costs and Por

Wesera2 Gioal Capt! Flows and Francng \Wes27=5 Onnarsho, Competition nd Comriptan: Bribe Takers vrsus

‘ene Payers

\WeS2704 Financal and Leg! Canes im Grow: Dass Sie Mater? 'WPS2785 ImgroirgAi Quay Maco Cây An Eeohene Vauthon ft eropolian