1. Firm Ownership, FDI Spillovers and Business Environment Constraints: Evidence from Transition Economies
World Bank Workshop on Productivity
Jan Svejnar
University of Michigan
October 2005

2. Presentation based on following research
Sabirianova, Svejnar and Terrell -- large and medium size firms in Russia and Czech Republic
Gorodnichenko, Svejnar and Terrell -- Business Environment and Economic Performance (BEEPs) data on 26 transition economies
Commander and Svejnar -- BEEPs data

3. Basic Ideas
Economic development: local firms converging in efficiency to the world standard
=> Worry about relative efficiency of domestic and foreign firms, the effect of foreign firms on domestic ones, and constraints in the business environment
Do foreign firms have positive or negative horizontal and vertical spillovers on local firms’ efficiency
Effects on efficiency of a firm’s involvement in export and import
Does business environment affect efficiency?
Examine efficiency with which firms generate revenues from inputs (use revenue function)

4. Czech Republic – CEE model
Work with K. Sabirianova and K. Terrell CEE and CIS Models – Czech Republic & Russia as Cases
Czech Republic and Russia – good labs (clear shift of regime & mass privatization)
Czech Republic – CEE model
Making economies and firms compatible with the European Union
Development of market-oriented institutions and legal system
Substantial FDI inflows and low trade barriers
Russia – CIS model
Not aiming to join EU
Limited institutional and legal development
Modest FDI inflows and a more closed economy

5. 1. Data Large firm-level panel data from the Czech Republic and Russia
1,537 to 2,970 firms in the Czech Republic
15,035 to 19,209 firms in Russia
Most medium and large firms (>100 employees) in the industrial sector
Period: (data for used for IVs)
; ;

6. Table 1: Percentage Share of Industrial Firms by Ownership Type for Selected Years
Czech Republic
Russia
Ownership:
1992
1996
2000
1993
Foreign
3.5
12.6
30.7
1.8
5.6
Mixed
0.7
21.0
12.9
32.6
42.7
28.2
Private (local)
18.4
57.4
54.1
16.7
38.3
51.3
State
77.4
9.0
2.4
48.9
15.6
15.0
Number of Firms
1537
2283
2084
17923
17138
15035

7. 1. Central Tendency: Basic Methodology
Yit = revenue of firm i in period t
x's = inputs (capital and labor)
Zit = ownership types
Iit = industry categories
Tt = annual dummy variables
Also estimate the revenue function separately for early, middle and late transition: , and

8. Endogeneity of Ownership
Ownership as a predetermined variable:
E(Zitis)  0 for  t > s, E(Zitvi)  0, and E(Zitit) = 0
Method 1: 2SLS random effect estimator with fitted ownership probabilities as instruments
M = ministries under central planning
Method 2: Ministries and lags of levels and differences in input and ownership as instruments in the Blundell-Bond system GMM estimation

9. Central Tendency: Principal (Robust) Findings
Foreign firms more efficient than domestic firms with state, private and mixed ownership in both countries
Gap much bigger in Russia than Czech Republic
Firms with private and mixed ownership similarly efficient -- (somewhat) more efficient than SOEs in Czech Rep. but less efficient than SOEs in Russia

10. Table 2a: Average Effects of Ownership on Efficiency, 1992-2000 Czech Republic
OLS
QREG RE FE
2SLS-RE BB
Foreign
0.435**
0.413**
0.319**
0.275**
0.349**
0.657**
(0.019)
(0.021)
(0.017)
(0.024)
(0.037)
Mixed
0.122**
0.086**
0.110**
0.094**
0.097**
0.074*
(0.022)
(0.014)
(0.015)
(0.020)
(0.031)
Private
0.145**
0.115**
0.117**
0.075**
0.053*
(0.016)
(0.013)
(0.027)
No. of obs.
19,971
15,142 R2
0.754
0.526
0.741
0.656 …

11. Table 2b: Average Effects of Ownership on Efficiency, 1992-2000 Russia
OLS
QREG RE FE
2SLS-RE BB
Foreign
0.994**
0.885**
0.398**
0.176**
0.629**
0.771**
(0.021)
(0.015)
(0.019)
(0.022)
(0.029)
(0.049)
Mixed
0.124**
0.159**
-0.020**
-0.050**
-0.110**
0.081**
(0.008)
(0.007)
(0.018)
(0.016)
Private
0.163**
0.174**
-0.019*
-0.060**
-0.114**
0.140**
(0.009)
(0.017)
No. of obs.
153,402
140,658 R2
0.680
0.482
0.670
0.594
0.688 …

12. The Best and the Worst: The Quantile Approach
Estimate quantile regressions
Q is the th quantile of ln yit conditional on X and Z
Estimated  give the relative efficiency of firms with different ownership at the th quantile of the distribution of log yit
The quantile approach constrains efficiency to be the same for all firms in a given percentile of the distribution but permits coefficients to vary across percentiles
The quantile approach does not control for firm-specific heterogeneity

13. The Best and the Worst: The Panel Data Approach
Estimate random effects model
For each firm i calculate efficiency as
RE allow efficiency to vary across firms but constrain input coefficients to be identical for all firms

14. Principal Findings Across the Distributions
Domestic firms less efficient than foreign firms at virtually all levels of the efficiency distributions – much more pronounced in Russia
Russia - difference between foreign and domestic firms is always greater among the more efficient firms
Czech Republic - same phenomenon only until
Later, relative efficiency of the worst Czech SOEs drops -- worst firms privatized last and soft budget constraints (SBCs)
Russia - large efficiency differentials between foreign and domestic firms -- SBCs and limited competition

16. An Alternative Measure of the Frontier
Rather than using the entire distribution of foreign-owned firms, take only the most efficient firms in an industry – say 33% -- as the frontier
Estimate the (inverse) distance from the frontier as the ratio of each firm’s efficiency to the mean efficiency of the frontier foreign firms within a two-digit industry in each period
Use random effect (RE) estimates to obtain our measure of efficiency

17. Distance of Domestic Firms to the Frontier by Period
Notes: The frontier is defined as the mean efficiency of the top third of foreign owned firms in a 2-digit ISIC industry. The efficiency estimates are obtained from the translog function (specified in equation 1) estimated with a random effect estimator for each period separately. The (inverse) measure of the firm’s distance to the frontier is calculated as the ratio of the firm’s efficiency to the frontier in its industry. Percentiles are constructed from the distribution of the firm-specific distance to the frontier for each ownership type.

18. Start-up Firms
Hypothesis -- new foreign firms have higher initial efficiency than domestic firms
Finding – existing foreign firms are more efficient than foreign start-ups; which are more efficient than domestic start-ups, which are in turn more efficient than existing domestic firms
Possible exception -- Czech start-ups with mixed ownership (some have foreign investors)
Implication -- emerging market economies are increasing efficiency as they create new, especially foreign, firms
Overall, foreign green-field investments have a positive effect on efficiency of the host economy

19. Creaming? Hypothesis -- foreigners acquire more efficient firms
Foreign firms move instantly ahead by taking over better firms rather than because of restructuring capabilities
Remaining local firms hence display a lower average efficiency on account of negative duration dependence
Findings support this hypothesis: part of the superior performance of foreign firms is due to selective acquisitions of local firms
Effect is larger in the Czech Republic than Russia
Hypothesis that foreign investors select less efficient firms and turn them around is not supported

20. Table 9: The Effect of Local Firm Characteristics on the Probability of Acquisition by Foreign Investors
Czech Republic
Russia
dF/dX
Mean(X)
(Efficiency
0.750**
0.006
0.047**
0.036
)t-1
(0.087)
(0.010)
Mixedt-1
1.634
0.114
-0.193**
0.359
(1.872)
(0.047)
Privatet-1
2.030**
0.582
-0.114*
0.314
(0.509)
(0.052)
Mixedt-1* Time
-0.297
0.678
0.080**
1.705
(0.177)
(0.013)
Privatet-1* Time
-0.351**
2.960
0.058**
1.703
(0.113)
Time
0.606**
4.475
-0.004
4.359
(0.097)
lnKt-1
0.548**
11.464
0.085**
0.596
(0.060)
(0.006)
Unconditional probability (%)
2.121
0.407

21. Differential Rates of Learning
Hypothesis – Domestic firms learning more slowly
Findings – Local firms are improving efficiency more slowly than foreign firms – consistent with the hypothesis that domestic firms acquiring knowledge slower than foreign firms
Foreign firms are more likely to move up in the overall efficiency distribution over time and stay in the top group
Mobility more pronounced in Russia than Czech Republic

22. Table10a: Average Effects of Ownership on Efficiency, 1992-2000, Czech Republic Time–Varying Effects
OLS
QREG RE FE
2SLS-RE BB
Foreign
0.303**
0.280**
0.149**
0.140**
0.208**
0.337**
(0.031)
(0.025)
(0.029)
(0.046)
(0.043)
Mixed
0.023
0.002
0.009
0.022
0.003
-0.002
(0.035)
(0.037)
(0.023)
(0.063)
Private
0.144**
0.142**
0.089**
0.103**
0.105**
(0.022)
(0.018)
(0.019)
(0.033)
 * Foreign
0.006
0.018**
0.033**
(0.007)
(0.005)
(0.006)
(0.008)
 * Mixed
-0.013
-0.003
0.020*
-0.004
(0.009)
(0.010)
(0.011)
 * Private
-0.038**
-0.031**
-0.018**
-0.012*
(0.004)
 * State
-0.025**
-0.017**
-0.016**
-0.010*
-0.001
-0.017*

23. Table 10b: Average Effects of Ownership on Efficiency, 1992-2000, Russia Time–Varying Effects
OLS
QREG RE FE
2SLS-RE BB
Foreign
0.693**
0.616**
0.296**
0.107**
0.465**
1.155**
(0.040)
(0.029)
(0.025)
(0.028)
(0.132)
(0.051)
Mixed
0.299**
0.373**
0.134**
0.093**
-0.012
0.496**
(0.017)
(0.014)
(0.012)
(0.013)
(0.144)
(0.026)
Private
0.332**
0.383**
0.124**
0.071**
0.006
0.548**
(0.016)
(0.015)
(0.122)
(0.027)
 * Foreign
0.131**
0.152**
0.080**
0.068**
0.060**
0.077**
(0.010)
(0.007)
(0.005)
(0.006)
 * Mixed
-0.024**
-0.016**
-0.023**
-0.021**
-0.014**
-0.027**
(0.003)
(0.002)
(0.004)
 * Private
-0.013**
-0.022**
-0.020**
-0.019**
-0.034**
 * State
0.014**
0.021**
0.013**
0.002
0.037**
(0.001)
(0.011)

24. Conditional (β) Convergence?
As a complement, estimate a conditional convergence equation
ip = efficiency of firm i in period p
κ = steady state efficiency level of firms with ownership Zip
η = (-log of) the speed of convergence of firms with ownership Zip to their steady state efficiency level
I = industry dummy variables controlling for factors that affect steady state efficiency levels of firms in different industries
P = period dummy variables

25. β Convergence Findings
Foreign owned firms converge to a higher steady state level of efficiency
In Czech Republic all types of firms converge to their respective steady states at a similar speed
In Russia foreign owned firms are converging faster

27. Positive or Negative Spillovers
Positive or Negative Spillovers? Sabirianova, Svejnar, Terrell (JEEA 2005)
If domestic firms are not catching up to the world efficiency standard in general, do they converge to it in industries with a greater foreign presence or vice versa?
Does greater foreign presence have a positive effect on foreign firms and a less positive or negative effect on domestic firms?

28. Findings on Spillovers
In both economies, greater presence of foreign firms in a given industry has a negative average effect on the efficiency of domestic firms and a positive effect on the efficiency of foreign owned firms
In the Czech Republic the negative spillover effect on domestic firms is alleviated and reversed over time
In Russia the negative spillover effect becomes stronger over time

29. The Effect of Foreign Presence on the Efficiency Gap, 1993-2000

30. Development, Institutions and Culture: Czech - Moscow - St Petersburg Comparisons
Compare basic results in the Czech Republic to those in the Moscow and St. Petersburg regions (similar size)
Moscow more advanced and St. Petersburg more western business culture than the rest of Russia – could bring about similar results to those in the Czech Republic
But Moscow and St. Petersburg embedded in the Russian legal, institutional and less open economic system
Q: Does Moscow’s and St. Petersburg’s corporate behavior (efficiency) resemble more that of the Czech Republic or Russia?

31. Development, Institutions and Culture (2)
Answer: Basic results for both Moscow and St. Petersburg resemble those for Russia as a whole
Implication: A country’s institutions, legal system and economic openness matter more than a region’s level of economic advancement or proximity to western business culture

32. Summary and Conclusions (1)
Two settings for evaluating development policies
In neither policy setting have domestic firms started to catch up to the world standard set by foreign firms
CEE model has succeeded in maintaining, but not closing, the relative foreign-domestic gap
CIS model has led to an increase in the gap
Even best firms not catching up to foreign counterparts
Privatization to domestic owners -- a major pillar of these policies -- has not been a means for domestic firms to start closing the efficiency gap (“wrong wealth transfer”?)

33. Summary and Conclusions (2)
Why this pattern?
Foreign start-ups more efficient that domestic ones
Foreign firms creaming
Domestic firms not learning as rapidly as foreign firms

34. Summary and Conclusions (3)
Implications for economic development
Czech Republic (CEE) is meeting the development challenge by increasing foreign firms’ share of GDP and exports
Russia (CIS) faces a development challenge – with increasing globalization and WTO entry

35. 2. Work with Y. Gorodnichenko and K
2. Work with Y. Gorodnichenko and K. Terrell BEEPs data (2005 wave, including retrospective data)
Industrial and service firms from 26 transition economies
Responses of CEOs to questionnaires => economic and financial data plus CEO perceptions
Augmented by industry-level data from recent input-output tables
Focus on vertical and horizontal spillovers plus effects of exports and imports
Dependent variable is sales revenue
Specification: Augmented Cobb-Douglas revenue function

38. 3. Work with Simon Commander BEEPs data (2002 and 2005 waves; panel)
Foreign firms more efficient than domestic firms, but their rate of increase of efficiency is not higher
Domestic private and state-owned firms similarly efficient
Effect of perceived differences in business environment (constraints) on efficiency is very limited
Effects of greater export orientation and perceived competition also limited