Impact of services liberalization on rm performance: evidence from the Ukrainian rm-level data

Transcription

1 Impact of services liberalization on rm performance: evidence from the Ukrainian rm-level data Oleksandr Shepotylo and Volodymyr Vakhitov August 12, 2010 Abstract This paper investigates the impact of services sector liberalization in - on performance of manufacturing and services sector rms in Ukraine. Using a unique database on Ukrainian rms in -, we look at external push for liberalization in services sectors as a source of exogenous variation that allows investigating the impact of services sector liberalization on total factor productivity (TFP) of a manufacturing rm. During the investigated period, Ukraine has achieved substantial progress in opening its services sector for competition of private and foreign owned rms. The services liberalization was accompanied by increase in productivity of services providers. The productivity of services sector has increased by almost 40 percent. As a result of services sector liberalization, total factor productivity of rms that used liberalized services more extensively has increased. Results indicate that one standard deviation increase in the services sector liberalization leads to 3.6 percent increase in total factor productivity. Interestingly, exporting rms have lower eect of services sector liberalizartion on total factor productivity. Key words: services liberalization, productivity, trade JEL codes: F14, G28, L80 1 Introduction The impact of services sector liberalization on performance of manufacturing rms has recently got considerable attention. The positive links between service sector liberalization and rm performance are documented for the Czech Republic, Sub-Saharan Africa, and India (Arnold et al., b,a, ) However, the ongoing debates about the benets of services sector liberalization for economic growth require to collect additional evidence from rm-level data in other Kyiv School of Economics and Kyiv Economic Institute, oshepotylo@eerc.kiev.ua Kyiv School of Economics and Kyiv Economic Institute, vakhitov@eerc.kiev.ua 1

2 countries in order to get more clear and nuanced picture. In this regard, Ukraine is an obvious country to study the impact of the services sector liberalization on economic performance of manufacturing rms. Ukraine has substantially liberalized the services sector over the past eight years. However, the degree of liberalization diered substantially across sub-sectors. It can be argued that the liberalization of services was not an internal response of the policymakers to changing economic situation, but rather an external push by the trading partners to liberalize trade in services as a prerequisite to the WTO accession. As a result, the government liberalized the sub-sectors that were crucial to reach an agreement on the WTO accession banking, nancial sectors, telecommunications, and business services and did not reform other sub-sectors utilities, retail and wholesale trade, and transportation. Using a unique database on Ukrainian rms in -, we look at external push for liberalization in services sectors as a source of exogenous variation that allows investigating the impact of services sector liberalization on total factor productivity (TFP) of a manufacturing rm. Our methodology is mainly based on Arnold et al. (b), as we rst estimate the production function using Olley and Pakes (1996) methodology to estimate production function for each industry and to compute total factor prodcutivity for each rm. At the second stage, we regress TFP on rm-specic index of services liberalization, controlling for rm-specic heterogeneity and market structure of manufacturing industries. The main ndings of the paper are as follows. Services sector liberalization episode coincided with the rapid growth of labor productivity in services sector, increase in the share of private rms, and even higher increase in the share of foreign-owned rms. Moreover, the labor productivity of private-owned services rms is higher than state-owned services rms, while foreign-owned rms have the highest labor productivity. The services sector liberalization, in turn, has a positive impact on TFP of manufacturing rms. An increase of the services liberalization index by a standard deviation increases TFP in manufacturing sector by 3.6 percent. Interestingly, exporting rms benet less from the services liberalization, even though, on average, they are more productive. The structure of the paper is as follows. Section 2 gives the background information on services sector performance in Ukraine. Section 3 discusses the theoretical and empirical literature on the impact of services liberalization on economic performance. Section 4 describes the data and present results of estimaing the impact of liberalization on performance in manufactung sector. Section 5 concludes and discusses directions for the future research. 2

3 2 Evidence on services sector liberalization in Ukraine in Structure of services sector in Ukraine The share of services in the Ukrainian total value added has been constantly increasing from 44 percent in to 47.7 percent in. 1 Still, the share is well below the average for middle income countries, which was equal to 60 percent in (Francois and Hoekman, 2010) and much lower than the average for the EU countries, which reached 65.5 percent in. 2 Looking at disaggregated services data reveals a large variation of performance of services sub-sectors in Ukraine. The nancial services grew substantially faster than other sectors of economy and increased their share from 3.5 percent in to 7.3 percent in. The combined category of real estate, rental, informational technologies (IT), research and development (R&D), and legal services increased its share from 6.6 percent in to 9.3 percent in. At the same time the share of transport and telecommunications reduced from 14.4 percent in to 10 percent in. The utilities followed a similar declining pattern, falling from 5 percent in to 3.7 percent in. Figure 1 presents more detailed decomposition of services sector, 3 available for the year. The wholesale and retail trade was the largest Ukrainian services subsector in, accounting for 13% percent of the total value added, supporting and auxiliary transport activities accounted for 4.3 percent, while land transport and transport via pipelines came at the third place with 4 percent of the total value added. At the same time, water transport and air transport accounted for only 0.3 percent. Despite growing importance of nancial services, the share of nancial intermediation, insurance, and auxiliary nancial activities remained relatively low at 4.5 percent. 2.2 Liberalization of the services sector in Ukraine Ukraine has substantially liberalized its economy in general, and the services sector in particular in -. According to the EBRD transition indicators, measured by the European Bank of Reconstruction and Development (EBRD) and reported in the annual EBRD transition report, the liberalization has been achieved in most dimensions of economic policy reforms. The dimensions reported by the EBRD indicators include large and small scale privatization, enterprise restructuring, price liberalization, trade and exchange rate liberalization, and competitiveness. In addition, a separate set of indices measures the progress of reforms in services sub-sectors banking, nancial, telecommunications, electricity, trans- 1 The source of data is the State Statistics Commettee of Ukraine. To calclulate the share of value added by services sector, we use Input-Output tables, - 2 The WTO database on services reports the prole of the EU services sector in 3 More disaggregated input-output table is available from the Ukrstat website, 3

4 Figure 1: Share of service sectors in total value added. Note: Input-Output table in, Source: State Statistical Commettee of Ukraine. 4

5 Progress of economic reforms Ukraine, - Trade liberalization Competition Enterprise restruct Price liberalization 3 5 Larege scale priv Small scale priv Source: EBRD Transition reports, - Figure 2: Progress of the overall economic reforms in Ukraine in - portation, and public infrastructure. Indices range from 1 to 4.3, higher values representing a better progress. The level of 4.3 means that the best practice frontier has been reached. According to Figure 2 that shows the progress of reforms in -, Ukraine has made a substantial progress on its way towards the WTO membership, liberalizing its trade and exchange rate policies. It has also improved the regulatory environment for small scale privatization. At the same time, there were no changes in price liberalization and competitiveness. While price liberalization was really not an issue, because by the index of price liberalization has already reached value 4 which is just one step away from the highest possible level, the index of competitiveness has stayed at the level 2, which is far away form the best practice frontier. According to Figure 3, reporting the progress of Ukraine in reforms of the services sub-sectors, Ukraine has substantially liberalized services in a number of the sub-sectors. The market access has been improved and the barriers to entry considerably reduced in banking, insurance, telecommunications, and public infrastructure sub-sectors. The banking sector has experienced a steady liberalization. For example, in the amendments to the law on banking allowed to open branches of foreign banks after the WTO accession, simplied procedure for opening of banks and subsidiaries, and clearly dened under which circumstances the National Bank of Ukraine can turn down the application by the foreign bank to operate in Ukraine. The law also dened limiting terms for accreditation of the foreign 5

6 Progress of reforms in services sub-sectors Ukraine, - Transportation 1 3 Public infrastructure Telecommunications Electricity Banking Financial Source: EBRD Transition reports, - Figure 3: Services sub-sectors liberalization in Ukraine in - banks (up to 3 months). The law on telecommunications in declared principles of equal access and fair competition; introduced the policy towards standardization and harmonization with the world standards; specied detailed procedures for frequency auctions and rules for licensing. The law on insurance in substantially liberalized the insurance sub-sector. 2.3 Performance of services sector and liberalization This section documents performance of services sector during -. We rst describe performance of services sector by reporting services sector productivity across industries and over time in Table 1. Productivity is measured as the value added per worker in thousands of hryvnas. To make the numbers comparable across time periods, the value added was deated by sector-specic deators available from the Ukranian Statistical oce and measured in constant prices of. 4 Looking at the cross-section of services sub-sectors, the water distribution has the lowest productivity across all services sub-sectors. Coincidently, it is the sector where no reforms are reported during -. Other sectors with low productivity are hotels&restaurants and land transport (including pipeline transport). Insurance sector, on the other hand, has the highest productivity in the sample. Sales of motor vehicles and air transport are examples of sectors with high value added. Panel B of Table 1 reports 4 Over the investigated peiod, the exchange rate between hryvna and US dollar was remarkably stable at the level of 5 hryvnas per dollar that makes conversion to the internationally comparable units easy 6

7 changes in productivity in -. During the period, the average productivity has increased from thousand hryvnas to thousand hryvnas, or by 39 percent. Real estate activities, air transport, and water transport are the fastest growing industries, while insurance, water, and nancial intermediation recorded the lowest growth rates. Looking separately at productivity of services providers by types of ownership in Tables 2-4 reveals that the foreign service providers are the most productive, while the state-owned providers are the least productive. The state owned rms have average productivity for the whole period 12.6 thosand hryvnas. Panel A of Table 5 reports the share of private rms in the sub-sector value added. During -, private ownership has increased in all sub-sectors with low initial private ownership, for example water distribution, electricity, gas and heating, and has remained high in sub-sectors with high initial private ownership, for example insurance and auxiliary nancial intermediation. Panel B of Table 5 reports the share of foreign owned rms in the sub-sector value added. The share of foreign owned rms has been increasing in almost all sub-sectors, including water transport, post and telecommunications, insurance, auxiliary nancial intermediations, and renting of machinery and equipment. The share of foreign owned companies has substantially declined in nancial intermediation. Finally, it remained low in water distribution, air transport, land transport, and R&D sub-sectors. In order to establish the association between liberalization and productivity per worker in -, we estimate a simple regression model ln(y it /L it ) = α + β foreign it + γ private it + δ index jt + ζ ln(l it ) + θd j + ϑd t + ɛ it (1) where Y it is value added of rm i in year t deated by a sector specic price deator, L it is the rm employment, foreign it is the dummy variable, taking the value of one if the rm is primarily foreign-owned and zero otherwise, private it is athe dummy variable taking the value of one if the rm is private and zero otherwise, index jt is the EBRD index of reforms in sector j, and D j and D t are sector, and time xed eects. Columns (1) and (2) of Table 6 present point estimates of the coecients of equation 1. Both private- and foreign-owned rms are more productive than state-owned rms. On average, private rms are 50 percent more productive than state rms, while foreign owned rms are 237 percent more productive than state owned rms. The liberalization of a services sub-sector is associated with higher productivity of rms in the subsector. Columns (3) and (4) demonstrate that the state-owned rms are larger. Column (5) indicates that the services liberalization is associated with increase in the number of rms in the services sub-sectors. 7

8 A Labor productivity in services in Ukraine, - Year NACE, Rev. 1 Services sub-sector Total 40 Electricity, gas, steam Water Sale of motor vehicles Retail trade Hotels and restaurants Land transport Water transport Air transport Auxiliary transp. activities Post&telecommunications Financial intermediation Insurance Auxiliary n. activities Real estate activities Renting of equipment IT R&D Other Total B Growth in labor productivity in services, - Year NACE, Rev. 1 Services sub-sector Average 40 Electricity, gas, steam Water Sale of motor vehicles Retail trade Hotels and restaurants Land transport Water transport Air transport Auxiliary transp. activities Post&telecommunications Financial intermediation Insurance Auxiliary n. activities Real estate activities Renting of equipment IT R&D Other Total Table 1: Average labor productivity of rms in services sub-sectors in - 8

9 Year NACE, Rev. 1 Total Total Table 2: Average productivty in state-owned rms 9

10 Year NACE, Rev. 1 Total Total Table 3: Average productivty in majority domestic private rms 10

11 Year NACE, Rev. 1 Total Total Table 4: Average productivty in majority foreign private rms 11

12 A Share of private rms Year NACE, Rev. 1 Services sub-sector 40 Electricity, gas, steam and hot water supply Collection, purication and distribution of water Sale, maintenance and repair of motor vehicles Retail trade Hotels and restaurants Land transport; transport via pipelines Water transport Air transport Supporting and auxiliary transport activities Post and telecommunications Financial intermediation Insurance and pension funding Activities auxiliary to nancial intermediation Real estate activities Renting of machinery and equipment Computer and related activities Research and development Other business activities B Share of foreign-owned rms Year NACE, Rev. 1 Services sub-sector 40 Electricity, gas, steam and hot water supply Collection, purication and distribution of water Sale, maintenance and repair of motor vehicles Retail trade Hotels and restaurants Land transport; transport via pipelines Water transport Air transport Supporting and auxiliary transport activities Post and telecommunications Financial intermediation Insurance and pension funding Activities auxiliary to nancial intermediation Real estate activities Renting of machinery and equipment Computer and related activities Research and development Other business activities Table 5: Share of private and foreign-owned rms in sub-sector value added in - 12

13 (1) (2) (3) (4) (5) lnrva_prod lnrva_prod lnempl lnempl ln(n) private (54.66) (54.56) ( ) ( ) foreign (55.72) (55.69) (-39.40) (-39.41) ln(employment) (73.98) (73.97) liberalization index (2.22) (1.75) (10.70) Constant (42.70) (11.93) (150.89) (62.95) (7.27) N t statistics in parentheses p < 0.05, p < 0.01, p < Table 6: Labor productivity and employment depending on rm ownership and sub-sector liberalization 3 Impact of services sector liberalization on productivity in manufacturing This section discuss the links between services sector liberalization and productivity of manufacturing rms and estimates the size of the eect of services liberalization on total factor productivity of manufacturing rms in Ukraine in Evidence on the role of services sector in productivity increase The evidence on the impact of the services on productivity in the downstream manufacturing rms is scarce. Arnold et al. (b) analyzed the impact of liberalization on services on the performance of approximately ten thousand manufacturing rms in the Czech Republic in and established the positive correlation between the productivity of manufacturing rms and the liberalization of the services that was higher for the rms that use the service inputs more intensively. A one-standard-deviation in the foreign presence in services was associated with a 3.8 percent increase in the total factor productivity. Arnold et al. () found a similar association between rm's productivity and availability of services for the Sub-Saharan Africa. Finally, Arnold et al. (a) found a positive impact of liberalization of services sector in India in 13

14 1993- on productivity of manufacturing rms. Moreover, the impact was higher for the rms with a substantial foreign ownership. According to their estimates, a one-standard-deviation increase in the index of services liberalization resulted in a six percent productivity increase for domestic rms and 7.5 percent productivity increase for foreign enterprises. Fernandes () studied structure and performance of services sector in transition countries in -. The main ndings included the following. There was a clear divide that separated more productive services in the Czech Republic, Hungary, Poland, Slovakia, and Slovenia (EU-5) from services sector in the Southern East Europe and Ukraine. Ukraine had the lowest level of labor productivity in most sub-sectors. In terms of growth rates, however, the results are mixed. Over investigated period, Ukraine lagged behind most countries in the region in growth of labor productivity in wholesale, retail, and repair of motor vehicles and transport, storage, and communications sectors. At the same time, in - Ukraine has achieved the highest annual growth in labor productivity in nancial intermediation and real estate, renting, and business activities. Turning to the determinants of labor productivity, Fernandes () emphasized the importance of the usage of information and communication technologies (ICT), intensity of skilled labor, and progress of policy reforms for performance of services sub-sectors. Finally, the paper further linked liberalization in services sector with an increase in labor productivity in downstream manufacturing sub-sectors in EU-5 countries, Estonia, Lithuania, Bulgaria, and Romania. 3.2 Channels of impact There are several channels that can explain how a rm level productivity is inuenced by the liberalization of services. Aghion and Schankerman (1999) describe the following mechanism: liberalization of services leads to a wider choice of service providers and lower costs of services. Increased supply of services, in turn, heightens competition and induce changes in the distribution of productivity of manufacturing rms, which facing the increase in supply of services invest in the cost saving technologies. Rajan and Zingales (1998) stress the importance of nancial sector for growth of manufacturing rms. More competitive nancial sector lowers the wedge between internal and external nancing of projects that allows investing in riskier projects with higher expected returns. Therefore, liberalization of nancial sector has a positive impact on industry level productivity. Manova () shows that nancial sector development lowers the credit constraints for exporting rms, thus, helping them to extend production and export. Other channels are also mentioned in the literature. In the presence of xed costs and learning-by-doing externalities, expansion of the output leads to higher productivity. Investment in infrastructure electricity, gas, and water supply has a positive impact on productivity of rms by providing private rms with essential inputs and attracting foreign direct investments to areas with developed infrastructure. Anos-Casero and Udomsaph () estimate 14

15 that a one-standard-deviation increase in provision of infrastructure correlates with 9.8 percent increase in rm-level productivity. More competitive and cost ecient services sector also induces entry of new rms by reducing xed costs of entry. Empirically, Eschenbach and Hoekman () showed that ECA countries that improved provision of inputs from the services sector attracted more inward FDI since the transition has started. On average, multinational companies (MNC) are more productive than domestic rms (Helpman et al., ). In addition, there evidence of positive spillovers form MNCs to domestic rms through backward and forward linkages, and knowledge spillovers (Matthias Arnold and Javorcik, ; Javorcik, ). Therefore, liberalization of services changes the distribution of rms' productivity and leads to the increase in the average productivity in other sectors. 4 Results and discussion 4.1 Data The data for the study come from several statistical forms annually submitted to National Statistics Oce (Derzhkomstat) by all commercial rms in the country. The data are restricted and not available for public use. The sample covers years from to with some gaps in the data for use of services in and. The total number of rms in the dataset exceeds 400,000 and covers all sectors except budgetary organizations (public schools, public hospitals, museums, etc.) and banks. We use the Financial Results statement to construct our output measure. The statement provides data on gross and net sales, as well as the material, wage, social security and administrative costs. We created our value added variable using the net sales less material and other operational costs. This variable was corrected for ination using the sector-specic deators provided by Derzhkomstat. Employment comes from the Enterprise Performance statement. The variable is measured as the year-averaged number of enlisted employees, which is a rough estimated of the total labor used in the production process. The Balance Sheet statement is the source of the capital measure for which we used the end-of-year value of the tangible assets. For every rm we created a comprehensive prole which includes all its industry and territory codes in every year. The dictionary comes from the les accompanying the Enterprise Performance statement. The territory code is available at the level of raions and cities of oblast subordination (which corresponds to the European NUTS3 category of statistical regions), whereas the industry code is at the level of four digits. At this level the Ukrainian industrial classication (KVED) is fully compatible with the European NACE (Rev. 1) classication. We chose the rms from the sections E (Electricity, Gas and Water Supply), G (Wholesale and Retail Trade; Repair Of Motor Vehicles, Motorcycles and Personal and Household Goods), H( Hotels and Restaurants), 15

16 I (Transport, Storage And Communication), J (Financial Intermediation), and K (Real Estate, Renting and Business Activities). Due to unclear nature of the material costs and, as a result, exaggerated output and productivity measures we dropped several sub-sectors and groups: KVED 7415 (Holding companies), KVED 6712 (Portfolio management), KVED 6603 (other types of insurance), and KVED 51 (Retail trade). We amended this dictionary with the database of the State property fund and the rms' FDI reports to construct a set of ownership variables. Firms indicating a state or municipal ownership and rms from the Banned from Privatization list were coded as state owned in a given year. The rest rms, as well as the ones with the private share above 50% were denoted as private. If the foreign share in the private rms exceeded the domestic private share, such rms were marked as primarily foreign owned. Other private rms were marked as primarily domestic owned. Brown and Earle () used the same classication and found it to be the most suitable for the transition economies. We further excluded observations with zero or negative output and employment in a year assuming that they indicated non-operational rms. After cleaning the data we have between 62,000 to 78,000 rms in the Services sector per year, totaling at 356,860 observations. 4.2 Methodology and results We follow the methodology developed by Arnold et al. (b) with some modications that are possible due to richer set of data, having services input at rm-level in particular. First stage. Production function estimation Assume that production technology of a rm i at time t is described by a production function Y it = F (L it, K it, M it, Ω it ) (2) where Y it is output, L it is labor, K it is capital, M it is material and services inputs, Ω it is the productivity, unobservable by an econometrician, but known by the rm. We further assume the Cobb-Douglas production function, and, after taking logs, derive the following equation y it = β 0 + β 1 l it + β 2 k it + β 3 m it + u it (3) u it = Ω it + η it where lower-case letters represent upper-case variables in the log form and η it is a shock that is not known by the rm when it chooses inputs at time t. At the rst stage, we estimate a production function by the Olley and Pakes (1996) methodology (OP) in order to extract the TFP for all manufacturing rms. Signicant entry and exit of rms, caused in part by liberalization of services sector and other regulatory reforms, creates a simultaneity and selection bias. Entry and exit decisions are functions of the productivity that, in turn, depends 16

17 ln(k) ln(l) ln(m) Firms Obs Food and Tobacco 0.057** 0.246*** 0.730*** (0.022) (0.013) (0.010) Textile and leather 0.074* 0.477*** 0.490*** (0.037) (0.022) (0.014) Wood and Paper *** 0.728*** (0.022) (0.019) (0.016) Printing 0.073* 0.419*** 0.508*** (0.028) (0.019) (0.016) Coke, chemistry, plastics *** 0.689*** (0.057) (0.018) (0.019) Non-metallic minerals *** 0.722*** (0.031) (0.020) (0.020) Metallurgy *** 0.644*** (0.026) (0.023) (0.026) Machinery and equipment 0.064** 0.459*** 0.492*** 5209 (0.021) (0.020) (0.015) High tech machinery 0.071* 0.368*** 0.510*** (0.035) (0.026) (0.014) Vehicles and transport *** 0.473*** (0.046) (0.047) (0.024) Furniture and others *** 0.628*** (0.027) (0.031) (0.024) Bootstrap standard errors in parentheses * p < 0.05, ** p < 0.01 Table 7: Estimation of production function by Olley and Pakes on liberalization and the decisions are taken simultaneously with the decision on the vector of inputs. The Olley and Pakes methodology accounts for those issues and correct the estimation of the parameters of the production function. Further, the TFP is estimated residually, after accounting for contribution of labor, capital, and material and services inputs to the rm output T F P it = y it ˆβ 0 + ˆβ 1 l it ˆβ 2 k it ˆβ 3 m it. The point estimates of the coecients of the production function estimated separately for 11 industries are presented in Table 7. The OP methodology is restrictive in a sense that it requires observed investment decisions in order to estimate the production function. Practically, it means that the observations with no investments are dropped from the regression. In our case, about two-thirds of rms report zero investments in a given year, considerably restricting the sample size. To avoid the limitation and check robustness of the estimation of the production function parameters, we also estimate the production function controlling for rm-level xed eects (FE). 17

18 ln(k) ln(l) ln(m) Firms Obs R2 overall Food and Tobacco 0.039*** 0.405*** 0.595*** (0.005) (0.008) (0.004) Textile and leather 0.036*** 0.512*** 0.400*** (0.006) (0.010) (0.005) Wood and Paper 0.038*** 0.433*** 0.520*** (0.007) (0.012) (0.006) Printing 0.079*** 0.472*** 0.389*** (0.005) (0.009) (0.004) Coke, chemistry, plastics 0.036*** 0.414*** 0.554*** (0.007) (0.012) (0.005) Non-metallic minerals 0.036*** 0.417*** 0.577*** (0.007) (0.013) (0.007) Metallurgy 0.043*** 0.387*** 0.554*** (0.006) (0.012) (0.006) Machinery and equipment 0.056*** 0.509*** 0.468*** (0.007) (0.011) (0.005) High tech machinery 0.049*** 0.505*** 0.415*** (0.006) (0.012) (0.005) Vehicles and transport 0.061*** 0.523*** 0.478*** (0.014) (0.028) (0.011) Furniture and others 0.051*** 0.548*** 0.465*** (0.007) (0.014) (0.006) Standard errors in parentheses * p < 0.05, ** p < 0.01 Table 8: Estimation of production function using rm-specic xed eects 18

19 Second stage. Impact of services liberalization on TFP of manufacturing rms The estimated T F P is further regressed on the index of service liberalization that is a rm-specic variable, reecting the rm-level importance of various service inputs, that is computed according to the formula service liberalization it = j a ijt index jt (4) where service liberalization it is the rm-specic index of service liberalization, a ijt is the share of input sourced from the service sector j in the total input of services for a rm i at time t, and libindex kt is the measure of liberalization in the service sector j at time t. 5 The constructed index of services liberalization takes into account liberalization of services sub-sectors and weights more heavily services sub-sectors used by the rm more extensively. The assumption here is that the liberalization of the sub-sector used by the rm more extensively is more important and has higher impact on rm's performance. One of the concerns that is important to keep in mind when interpreting the results is endogeneity of the constructed measure of services liberalization. However, as we argue earlier, the preassure to reform the services sector to a large extent was external push from the WTO members rather than internal response of the local policymakers to economic conditions within the country. The estimated regression takes the following parametric form ln(t F P it ) = α + service liberalization it β + exporter it γ (5) + exporter it service liberalization it δ + D i µ + D jt λ + ɛ it where T F P it is the rm j level of total factor productivity at time t, exporter it is the dummy variable that takes value of one if the rm i exported in year t and zero otherwise, D i are rm-specic xed eects unobserved by the researcher, and D jt are industry-time specic xed eects capturing market structure of industry j, and sector-specic macroeconomic shocks at time t. Results are presented in Table 8. In column (1) the dependent variable is log TFP estimated by Olley and Pakes method, while in column (2) the dependent variable is log TFP estimated by xed eect method. Comparison of point estimates of the coecients in columns (1) and (2) shows that the results are robust to the dierent methods of estimating the production function. A positive and signicant coecient of the services liberalization index indicates the positive eect of services sector liberalization on total factor productivity of Ukrainian manufacturing rms. Increase in rm specic index of services liberalization by one standard deviation results in a 3.6 percent increase in TFP according to the results in column (1). Exporters are 24 percent more productive than non-exporters, conrming higher productivity of exporting manufacturing rms (see, for example, Bernard et al., ). At the same time, the interaction term between services liberalization and exporting status is slightly negative, which is a puzzling result which needs further research. 5 At this point we have services use data at the rm level for - 19

20 ln(tfp) OP ln(tfp) FE services liberalization 0.096*** 0.088*** (0.015) (0.015) exporters 0.236*** 0.244*** (0.052) (0.051) exporters x services liberalization * * (0.022) (0.022) industry-time specic xed eects Yes Yes rm specic xed eects Yes Yes rms obs Overal R Standard errors in parentheses * p < 0.05, ** p < 0.01 Table 9: Services liberalization and total factor productivity in Ukraine in - 5 Conclusions Ukraine has substantially liberalized its services sector in -, which resulted in growing share of private- and foreign-owned rms. As the data suggests, the privatization and foreign ownership is strongly associated with higher productivity per worker productivity per worker in state-owned rms was lower in levels and growth relative to the productivity per worker in private- and foreign-owned rms. During the period, the productivity of services sector has increased by almost 40 percent. As results suggest, the increase in services sector liberalization had a positive spillover eect on the total factor productivity of manufacturing rms one standard deviation increase in services sector liberalization results in 3.6 percent increase in the total factor productivity. We also conrm the fact that exporters are more productive than non-exporters, but nd a puzzling result that services sector liberalization has a stronger positive eect on non-exporting manufacturing rms. References Aghion, P. and Schankerman, M. (1999). Competition, entry and the social returns to infrastructure in transition economies. Economics of Transition, 7(1):

21 Anos-Casero, P. and Udomsaph, C. (). What Drives Firm Productivity Growth? Research Working papers, 1(1):154. Arnold, J., Javorcik, B., Lipscomb, M., and Mattoo, A. (a). Services reform and manufacturing performance: Evidence from India. World Bank, mimeo. Arnold, J., Javorcik, B. S., and Mattoo, A. (b). Does services liberalization benet manufacturing rms? evidence from the czech republic. Arnold, J. M., Mattoo, A., and Narciso, G. (). Services inputs and rm productivity in sub-suharan africa: Evidence from rm-level data. Journal of African Economies, 17(4): Bernard, A., Eaton, J., Jensen, J., and Kortum, S. (). Plants and productivity in international trade. The American Economic Review, 93(4): Eschenbach, F. and Hoekman, B. (). Services policy reform and economic growth in transition economies. Review of World Economics/Weltwirtschaftliches Archiv, 142(4): Fernandes, A. M. (). Structure and performance of the service sector in transition economies. Economics of Transition, 17(3): Francois, J. and Hoekman, B. (2010). Trade and Policy in Services. Journal of Economic Literature. Helpman, E., Melitz, M. J., and Yeaple, S. R. (). Export versus fdi with heterogeneous rms. American Economic Review, 94(1): Javorcik, B. S. (). Does foreign direct investment increase the productivity of domestic rms? in search of spillovers through backward linkages. American Economic Review, 94(3): Manova, K. (). Credit constraints, equity market liberalizations and international trade. Journal of International Economics, 76(1):3347. Matthias Arnold, J. and Javorcik, B. S. (). Gifted kids or pushy parents? foreign direct investment and plant productivity in indonesia. Journal of International Economics, 79(1): Olley, G. S. and Pakes, A. (1996). The dynamics of productivity in the telecommunications equipment industry. Econometrica, 64(6): Rajan, R. G. and Zingales, L. (1998). Financial dependence and growth. American Economic Review, 88(3):