Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
21
Productivity Cycles in Public and Private Manufacturing Sectors:
Evidence from Turkey
Hakan Yilmazkuday*
Temple University
Abstract This paper compares the productivity cycles of public and private manufacturing
sectors in Turkey by using a regime shifting model applied through the multimove Gibbssampling approach over the quarterly period 1988Q1:2006:Q4. By considering timing of the
business cycles for the sample period, it is shown that: 1) the public sector has higher
productivity growth rates compared to the growth rates of private sector and total productivities,
in both low and high productivity growth regimes; 2) the productivity in public sector is
procyclical in periods of real shocks, such as stagnation or earthquakes (i.e., 1989 and 1999
crises); 3) the productivity in private sector is procyclical in periods of financial shocks (i.e.,
1994 currency crisis, 1998 Russian crisis and 2001 financial crisis); 4) the productivity in the
public sector has a smoothing effect in terms of reducing the effects of private sector
productivity cycles.
Keywords: Regime-Shifts; Productivity Cycles; Private Sector; Public Sector; Turkey
JEL Classification: E32, E44
1. Introduction
Productivity is a measure relating a quantity or quality of output to the inputs required to produce
it. Since business cycles have been one of the most attractive sources of study for economists, for
both developed and developing countries (see Kim and Nelson, 1999, for an extensive number of
business cycle analyses), the relation between the business cycles and the productivity growth
has an important role in explaining the specifics of the cycles (see Hall, 1990; Caballero and
Lyons, 1992; Basu and Fernald, 1995, 2000; Burnside, 1996; Burnside et al., 1996; Vecchi, 2000;
Hayashi and Prescott, 2002; Basu et al., 2004; Miyagawa et al., 2005). Most of the related
studies, such as Hulten and Schwab (1991); Morrison and Schartz (1996); Boisso et al. (2000),
focus on the effects of business cycles on the productivity for developed countries. Nevertheless,
there is less evidence on the developing countries.
This study bridges the gap by considering the experience of the Turkish economy over the
quarterly period 1988Q1:2006Q4. In order to find the productivity cycles of Turkey, we apply a
univariate Markov regime switching model originally introduced by Albert and Chib (1993). We
also consider the distinction between the productivities of public and private manufacturing
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
22
sectors in our analysis. In the related literature, studies such as Davies (1971, 1977), Boardman
and Vining (1989), Galal et al. (1994), Dewenter and Malatesta (1998), Megginson et al. (1994),
Majumdar (1996) have found private sector to be more productive, while studies such as Tyler
(1979), Caves and Christensen (1980), Millward (1988), Nelson and Primeaux (1988), Bruggink
(1982), Parker and Wu (1998) have found public sector to be more productive or no statistically
significant differences between two sectors.1
For the Turkish economy, there have been also studies to compare public and private sector
productivity levels: Cakmak and Zaim (1992) compare the productivity of the two sectors for the
Turkish cement industry, and they find that there is no significance difference between them;
Bagdadioglu et al. (1996) compare the productivity of the two sectors for the Turkish electricity
distribution, and they find that private sector is more productive; Ozmucur (2003) compares the
wage and productivity differentials in Turkish public and private manufacturing over the annual
period 1950-1998, and he finds that labor productivity is higher in the private sector; Zaim and
Taskin (1997) extensively compare the rate of change in productivity of Turkish public and
private sectors over the annual period 1974-1991, and they find that public sector has performed
worse compared to the private sector; Turut and Becker (2000) compare the productivity levels
of public and private sector, and they again find that productivity is higher in the private sector;
Karadag et al. (2005) compare the total factor productivity (TFP) changes between Turkish
public and private sectors over the annual period 1990-1998, and they find that while there is no
evidence for TFP growth in the private sector, there is evidence for it in the public sector.2
Instead of the studies mentioned above, which mostly compare the annual productivity levels of
public and private sectors (except for Zaim and Taskin, 1997, and Karadag et al., 2005, who
compare the annual rate of change in productivity), this paper compares the productivity growth
rates and cycles of these sectors by using quarterly data. This comparison is important to figure
out the trend in productivity series and to depict the relation between productivity and business
cycles in sectoral terms. In technical terms, the aim of our related model is to generate regime
probabilities for different average (mean) growth measures of the productivity in public and
private manufacturing sectors, by applying a Gibbs-sampling approach. As we talk in more
details in the text, our methodology has important advantages over the maximum likelihood
fitting approach of Hamilton (1988). First, the messy calculations entailed in the direct
calculation of the likelihood function are avoided, thus, the simulation algorithm is relatively
easy to implement. Second, posterior distribution of all unknown parameters and functions
thereof are obtained by simulating standard distributions, such as the multivariate normal and
inverted gamma. These posterior distributions convey much more information than the mode and
curvature summaries that arise from the maximum likelihood framework.
Our main findings for the Turkish economy can be listed as follows: 1) the public sector has
higher productivity growth rates compared to the growth rates of private sector and total
productivities, in both low and high productivity growth regimes; 2) the productivity in public
sector is procyclical in periods of real shocks, such as stagnation or earthquakes (i.e., 1989 and
1999 crises); 3) the productivity in private sector is procyclical in periods of financial shocks (i.e.,
1994 currency crisis, 1998 Russian crisis and 2001 financial crisis); 4) the productivity in the
public sector has a smoothing effect in terms of reducing the effects of private sector
productivity cycles.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
23
The rest of the paper is organized as follows: Section 2 introduces the model, while Section 3
introduces data and our motivation. Section 4 depicts the Gibbs-sampling results for public and
private manufacturing sectors of Turkey. Section 5 concludes.
2. The Model
We use an approach which is originally introduced by Albert and Chib (1993), and then has been
modified by Kim and Nelson (1999, pp.218-219). This is a Bayesian Gibbs-sampling approach
applied to a Markov-switching model. In particular, we consider the following model with
Markov-switching mean:
yt = μ ( St ) + et
(1)
et ~ N (0, σ )
(2)
μ ( St ) = μ0 + μ1St
(2)
2
where yt the annual change in productivity; μ ( St ) is the state-dependent (mean) growth rate in
productivity; et is the error term with homoscedastic variance; and St ∈ {0,1} is the unobserved
two-state Markov-switching variable evolving according to the transition probabilities given
below:
⎛p
p = ⎜ 00
⎝ p10
p01 ⎞
⎟
p11 ⎠
(3)
where
p00 = Pr ⎡⎣ St = 0 St −1 = 0 ⎦⎤
p01 = Pr ⎡⎣ St = 0 St −1 = 1⎦⎤
p10 = Pr ⎡⎣ St = 1 St −1 = 0 ⎦⎤
(4)
p11 = Pr ⎡⎣ St = 1 St −1 = 1⎦⎤
and
p00 + p10 = 1
p01 + p11 = 1
(5)
According to our setting, we can determine the annual (mean) productivity growth trends for two
different regimes, namely high productivity growth and low productivity growth, by using the
Markov-switching nature of the model.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
24
3. Data and Motivation
In our analysis, we use the quarterly data of partial productivity of labor covering the period
1988:Q1-2006:Q4, for both public and private manufacturing sectors of Turkey together with the
total productivity data. From now on, we will use public, private and total for these productivity
series. For all the series, we use the data obtained from the Central Bank of the Republic of
Turkey (CBRT).3 These series are depicted in Figure 1.
As is evident by Figure 1, the public seems to have a higher productivity growth compared to
private and total. To see the specifics of the data, consider Figure 2 which shows the annual
growth of productivity for each of the series. The shaded areas of Figure 2 show the recessionary
periods determined by the OECD.4 These are: the 1988-89 stagnation that signaled the
shortcomings of the export-led growth strategy adopted during 1980s, the 1990-91 recession
caused by the Gulf War, the 1994 recession that followed the financial crisis in the same year,
the 1998 recession triggered by the Russian Crisis and the 1999 recession due to the earthquake,
and finally, the 2001 recession that followed the financial and currency crises that took place in
November 2000 and February 2001.5
Figure 2 suggests that the growth in the productivity of public and private manufacturing sectors
have different patterns. In particular, the correlation coefficient between public and private
productivity growth rates is -0.16; it is 0.10 between public and total productivity growth rates;
and it is 0.96 between private and total productivity growth rates. In other words, public sector
productivity growth has almost no relation with either private or total productivity growth rates.
As Vickers and Yarrow (1988); Martin and Parker (1997); and Willner and Parker (2007)
suggest, this may be due to the differences in the structures of competition in these sectors as
well as the type of owners (i.e., active or passive) that determine how these sectors are managed.
Another complementary explanation can be the ongoing process of privatization in Turkey that
has started extensively from 1985. In particular, there are 186 different public companies that
have been privatized with different percentage of shares sold.6 As it is put by the Privatization
Administration of Turkey, the philosophy of privatization is as follows:
“The main philosophy of privatization is to confine the role of the state in the economy in the
areas like health, basic education, social security, national defense, large scale infrastructure
investments; provide legal and structural environment for free enterprise to operate and thus to
increase the productivity and the value added to the economy by ensuring more efficient
organization and management in the enterprises that should be commercialized to be competitive
in the market.”
Thus, to increase the productivity and efficiency in the economy is one of the main objectives of
privatization in Turkey. Due to this objective, it may be the case that after the privatization of
186 public companies, the public sector has now companies with higher productivity growth
rates compared to private sector companies.
Figure 3 depicts the level of privatization in terms of the percentage of Turkish GDP. As is
evident, the privatization has increased dramatically starting from late 1990s, which is consistent
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
25
with our view that privatization may have led to higher productivity growth rates in the public
sector.
Although the reasons for the difference in productivity growth rates between public and private
sectors are very important in terms of privatization process and competition policies, they are
beyond the scope of this empirical paper. Instead of focusing on these reasons empirically, this
paper attempts to put together the stylized facts of the productivity growth rates and cycles in the
public and private sectors. Related to our main objective, according to Figure 2, the growth of
the productivity series behave in a different manner in distinctive recessionary periods. In
particular, the public sector productivity seems to be procyclical in 1988-1989 and 1999
recessions while the private sector productivity seems to be procyclical in 1994, 1998 and 20002001 recessions. In other words, the public sector productivity seems to be procyclical in periods
of real shocks and the private sector seems to be procyclical in periods of financial shocks. To
test our observation, we apply our Markov-switching model in the next section by using the same
data as in Figure 2. Notice that by using the fourth log difference of the quarterly productivity
series as our growth measure, we also control for seasonality.7
4. Gibbs-Sampling Results
We use the Bayesian Gibbs-sampling approach to find the marginal posterior distributions of the
parameters of the model given by Equations (1)-(5). In particular, Gibbs-sampling is a Markov
chain Monte Carlo simulation method for approximating joint and marginal distributions by
sampling from conditional distribution. Selected studies on the Gibbs sampling are Albert and
Chib (1993); Casella and George (1993); Geman and Geman (1984); Gelfand and Smith (1990);
and Gelfand et al. (1990). In this paper, we use the one originally introduced by Albert and Chib
(1993). The main idea behind the method of Albert and Chib (1993) is that the unobserved states,
one for each time point, can be treated as missing data and then analyzed, along with the other
unknown parameters, via the simulation tool of Gibbs sampling.
In this sense, this methodology has two important advantages over the maximum likelihood
fitting approach of Hamilton (1988). First, the messy calculations entailed in the direct
calculation of the likelihood function are avoided, thus, the simulation algorithm is relatively
easy to implement. Second, posterior distribution of all unknown parameters and functions
thereof are obtained by simulating standard distributions, such as the multivariate normal and
inverted gamma. These posterior distributions convey much more information than the mode and
curvature summaries that arise from the maximum likelihood framework.
Moreover, the Gibbs sampling provides the posterior distributions of the states, and of future
observations, marginalized over all of the unknown parameters. This improves on plug-in
approaches in which unknown parameters are replaced by sample estimates. Finally, residual
analysis proceeds in a straightforward fashion by using the distribution of generated states to
compute the posterior distribution of the model residual.
Instead of the single-move Gibbs sampling of Albert and Chib (1993), we apply the multimove
Gibbs-sampling, originally motivated by Carter and Kohn (1994) in the context of a state-space
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
26
model and implemented by Kim and Nelson (1998) in a Markov-switching model, to obtain the
marginal posterior distributions of the parameters. Gibbs-sampling is run such that the first 2,000
draws are discarded and the next 10,000 are recorded. Following Kim and Nelson (1999), we
employ almost non-informative priors for all the model’s parameters. Specifically, for all of the
Gibbs-samplings that we run, we employ the priors that given in Table 1.
The estimation results are given in Table 2.1 According to Table 2, when the annual growth
trends are taken into account, the low (mean) annual growth rate of the productivity in the public
sector is around 2.40%, and the high (mean) annual growth rate is around 6.63%. Similarly, the
low (mean) annual growth rate of the productivity in the private sector is around 2.05%, and the
high (mean) annual growth rate is around 5.06%. Note that the public sector has higher
productivity growth rates compared to both private sector and total in both low and high
productivity regimes. This result is consistent with Karadag et al. (2005) and our observation
through Figure 1 in which we depict all of our series. As we have discussed in details above, this
results may be due to the differences in competition structures of the two sectors as well as due
to the privatization process in Turkey that have privatized 186 different companies since 1985.
We also provide additional explanations for this result below.
The results in the first two rows of Table 2 are also interesting for us. Note that, given that the
economy is in state 0 in the previous period, p00 is the probability of being in state 0 in the
current period. Similarly, given that the economy is in state 1 in the previous period, p11 is the
probability of being in state 1 in the current period. By using these Gibbs-sampling results, we
can calculate the expected duration of a high-productivity-growth state and the expected duration
of a low-productivity-growth state. As shown by Kim and Nelson (1999, pp.71-72), the expected
duration formula is given by:
E(Dj ) =
1
1 − p jj
(6)
where E ( D j ) represents the expected duration of state j . For different manufacturing sectors of
the Turkish economy, the expected durations of state 0 and state 1 are given in Table 3.
Note that, since we work with quarterly data, the durations represented in Table 3 are in quarterly
terms, regardless of the growth measure that is used. According to Table 3, the expected duration
of the low (mean) productivity growth rate of 2.40% in the public sector is around 5 quarters;
while the expected duration of the high (mean) productivity growth rate of 6.63%, is around 26
quarters. Similarly, the expected duration of the low (mean) productivity growth rate of 2.05% in
the public sector is around 4 quarters; while the expected duration of the high (mean)
productivity growth rate of 5.06%, is around 17 quarters. Finally, the expected duration of the
low (mean) productivity growth rate of 2.33% in the total is around 4 quarters; while the
expected duration of the high (mean) productivity growth rate of 5.74%, is around 18 quarters.
These figures suggest that the public sector productivity has longer durations compared to other
two series. Depending on the turning points of the productivity cycles, this difference in the
durations may be a signal for the smoothing effect of the public sector productivity on the
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
27
productivity of other series. To investigate this possibility, we consider the probabilities of lowproductivity for each sector in Figure 4.
Figure 4 supports us for the smoothing feature of the public sector. In particular, the public sector
behaves countercyclical to the private sector and the total in almost all recessionary periods.
Figure 4 also supports our motivation that the public sector productivity drops seem to be related
to the real shocks (i.e., 1989 and 1999 crises), and the private sector productivity drops seem to
be related to financial shocks (i.e., 1994 financial crisis, 1998 Russian financial crisis and 2001
financial crisis).
In addition to the explanations that we have given in Section III for the difference between public
and private productivity cycles (i.e., the type of owners or the process of privatization), another
possible explanation for our empirical results may be the capital and finance structures of the two
sectors. In particular, while the private sector mainly finances its capital from financial
institutions, which are directly affected through financial crises, the public sector is mainly
financed by the government, which is directly affected through real shocks. Another
complementary possible explanation may be the export oriented private sector. More specifically,
while the public sector mostly serves to the domestic market, the private sector serves to the
foreign markets as well. This fact may give rise to productivity drops in private sector due to the
terms of trade effects of the financial crises. In particular, a private company that cannot sell its
product adds its production to its inventory instead of selling it, which, in turn, leads to
underproduction of the firm with the very same capacity after some while. Similarly, the public
sector is mostly affected through the demand shocks in the domestic economy. By the same
token, the public sector may also end up with underproduction due to the underused capacity.
These results may also be connected to the traditional business cycle literature where labor
hoarding is one of the key components. In the traditional literature, a small role is attached to
monopoly power and increasing returns to explain productivity fluctuations. Instead of it, labor
hoarding, which means the quasi-fixity of the labor input, is seen as the primary source of
cyclical changes in productivity. This approach states that productivity falls in recessions
because the firms retain their workers. This is consistent with the observation in the data that, as
the economy enters a recession, the percentage fall in output exceeds the percentage fall in labor
input; thus, the productivity is procyclical. In particular, Rotemberg and Summers (1990) find
that productivity is more procyclical in industries and in nations where labor hoarding appears
more important. In this context, both the procyclicality of the public sector productivity during
the real shocks (i.e., 1989 and 1999 crises) and the procyclicality of the private sector during
financial shocks (i.e., 1994 financial crisis, 1998 Russian financial crisis and 2001 financial crisis)
may be related to the relevant labor hoardings in such sectors. This may also give more insight
about the business cycle specificities of these sectors.
5. Conclusions
We have analyzed the productivity cycles of public and private manufacturing sectors in Turkey
by using a regime shifting model applied through the multimove Gibbs-sampling approach. By
considering timing of the business cycles, we find that the public sector has higher productivity
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
28
growth rates compared to the growth rates of private sector and total productivities, in both low
and high productivity growth regimes; the productivity in public sector is procyclical in periods
of real shocks, such as stagnation or earthquakes (i.e., 1989 and 1999 crises); the productivity in
private sector is procyclical in periods of financial shocks (i.e., 1994 currency crisis, 1998
Russian crisis and 2001 financial crisis); and finally, the productivity in the public sector has a
smoothing effect in terms of reducing the effects of private sector cycles.
A possible explanation for our results is the capital and finance structures of the two sectors. In
particular, while the private sector mainly finances its capital from financial institutions, which
are directly affected through financial crisis, the public sector is mainly financed by the
government, which is directly affected through real shocks. Another complementary possible
explanation may be the export oriented private sector. More specifically, while the public sector
mostly serves to the domestic market, the private sector serves to the foreign markets in addition.
This fact may give rise to productivity drops in private sector due to the terms of trade effects of
the financial crises. In particular, a private company that cannot sell its product in financial crisis
periods adds its production to its inventory instead of selling it, which, in turn, leads to
underproduction of the firm with the very same capacity. Similarly, the public sector is mostly
affected through the demand shocks in the domestic economy. By the same token, the public
sector may also end up with underproduction due to the underused capacity.
Analyzing and testing empirically the reasons for the difference in productivity growth rates
between public and private sectors are also very important in terms of privatization process and
competition policies. Considering capital and finance formation of the two sectors together with
their market structure would also be interesting empirically. Another interesting study would
include the type of owners into the analysis that determine how these sectors are managed. All of
these empirical issues are beyond the scope of this paper, but they may be subject of future
research.
Endnotes
*
Department of Economics, Temple University, Philadelphia, PA 19122, USA. E-mail:
[email protected].
1. For the comparison of the efficiency levels in public and private manufacturing sectors, see
also the overviews in Megginson and Netter (2001); Willner and Parker (2007).
2. For other analyses on the productivity of the Turkish manufacturing sector, see Krueger and
Tuncer (1982); Yildirim (1989); Uygur (1990); Aydogus (1993); Gokcekus (1997); Onder and
Lenger (2003); Taymaz and Saatci (1997). See Denizer et al. (2007) for the comparison of the
banking efficiency in Turkey before and after the financial liberalization. Also see Willner (2001)
for studies comparing the efficiency of public and private sectors in US, UK, Sweden, Brazil,
Canada, Australia, Finland, Germany and Switzerland.
3. http://www.tcmb.gov.tr/ is the web page of CBRT.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
29
4. http://www.oecd.org/document/29/0,2340,en_2649_34349_35725597_1_1_1_1,00.html is the
webpage of the OECD that publishes the OECD Composite Leading Indicators: Reference
Turning Points and Component Series.
5. See Akay and Yilmazkuday (2008) for a complete business cycle analysis of Turkey.
6. See http://www.oib.gov.tr/program/uygulamalar/completely_privatized.htm for the list of all
companies that have been privatized in Turkey.
7. See Yilmazkuday (2009) about the relation between seasonality and growth measures.
8. For the estimation, we modify the GAUSS codes written by Kim and Nelson (1999).
References
Akay, K., Yilmazkuday, H. 2008. “An Analysis of Regime Shifts in the Turkish Economy,”
Economic Modelling, 25, 885-898.
Albert J.H. and S. Chib. 1993. “Bayes Inference via Gibbs Sampling of Autoregressive Time
Series Subject to Markov Mean and Variance Shifts,” Journal of Business and Economic
Statistics, 11, 1-15.
Aydogus, O. 1993. “Turkiye Imalat Sanayiinde Ithal Ikamesi, Ihracat Artisi, ve Toplam Faktor
Verimliligi Iliskileri, 1971-88,” METU Studies in Development, 20, 453-473.
Bagdadioglu, N., C. M. Waddams Price, T. G. Weyman-Jones. 1996. “Efficiency and
ownership in electricity distribution: a non-parametric model of the Turkish experience,” Energy
Economics, 18, 1–23.
Basu, S. and J. Fernald. 1995. “Are Apparent Productive Spillovers a Figment of Specification
Error?” Journal of Monetary Economics, 36, 165–188.
Basu, S. and J. Fernald. 2000. “Why is Productivity Procyclical? Why do we Care?” NBER
Working Paper No: 7940.
Basu, S., J. Fernald, and M. Kimball. 2004. “Are Technology Improvements Contractionary?”
NBER Working Paper Series, No.10592.
Boardman, A. E., and A. R. Vining. 1989. “Ownership and Performance in Competitive
Environments: A Comparison of the Performance of Public, Mixed and State-owned
Enterprises,” Journal of Law and Economics, 32, 1–33.
Boisso, D., S. Grosskopf, and K. Hayes. 2000. “Productivity and Efficiency in the U.S.: Effects
of Business Cycles and Public Capital,” Regional Science and Urban Economics, 30, 663-681.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
30
Bruggink, T. H. 1982. “Public versus Regulated Private Enterprise in the Municipal Water
Industry: a Comparison of Operating Costs,” Quarterly Review of Economics and Business, 22,
111–125.
Burnside, C. 1996. “Production Function Regressions, Returns to Scale, and Externalities,”
Journal of Monetary Economics, 37, 177–201.
Burnside, C., M. Eichenbaum, and S. Rebelo. 1996. “Sectoral Solow Residuals,” European
Economic Review, 40, 861–869.
Caballero, R. J., and R. K. Lyons. 1992. “External Effects in U.S. Procyclical Productivity,”
Journal of Monetary Economics, 29, 209–225.
Cakmak, E. H. and O. Zaim. 1992. “Privatization and comparative efficiency of public and
private enterprise in Turkey: The cement industry,” Annals of Public and Cooperative Economy,
63, 271–284.
Carter, C. K. and P. Kohn. 1994. “On Gibbs Sampling for State Space Models,” Biometrica,
81, 541-553.
Casella G. and E. I. George. 1993. “Explaining the Gibbs Sampler,” The American Statistician,
46, 167-174.
Caves, D. W. and L. R. Christensen. 1980. “The Relative Efficiency of Public and Private
Firms in a Competitive Environment: the Case of Canadian railroads,” Journal of Political
Economy, 8, 958–976.
Davies, D. G. 1971. “The Efficiency of Public versus Private Firms: the Case of Australia’s Two
Airlines,” Journal of Law and Economics, 14, 149–165.
Davies, D. G. 1977. “Property Rights and Economic Efficiency – the Australian Airlines
Revisited,” Journal of Law and Economics, 20, 223–226.
Denizer, C., M. Dinc, and M. Tarimcilar 2007. “Financial liberalization and banking
efficiency: evidence from Turkey,” Journal of Productivity Analysis, 27, 177-195.
Dewenter, K. L. and P. H. Malatesta. 1998. “State-owned and Privately-owned Firms: an
Empirical Analysis of Profitability, Leverage and Labour Intensity,” mimeo, School of Business,
University of Washington.
Galal, A., L. Jones, P. Tandon, and I. Vogelsang. 1994. Welfare Consequences of Selling
Public Enterprises. New York: Oxford University Press.
Gelfand, A. E. and A. F. M. Smith. 1990. “Sampling-Based Approaches to Calculating
Marginal Densities,” Journal of the American Statistical Association, 85, Theory and Methods,
398-409.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
31
Gelfand, A. E., S. E. Hills, A. Racine-Poon, and A. F. M. Smith. 1990. “Illustration of
Bayesian Inference in Normal Data Models Using Gibbs Sampling,” Journal of the American
Statistical Association, 85, Application and Case Studies, 972-985.
Geman, S. and D. Geman. 1984. “Stochastic Relaxation, Gibbs Distribution and the Bayesian
Restoration of Images,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 6,
721-741.
Gokcekus, O. 1997. “Trade Liberalization and Productivity Growth: New Evidence from the
Turkish Rubber Industry,” Applied Economics, 29, 639-645
Hall, R. E. 1990. “Invariance Properties of Solow’s Productivity Residual,” in P. Diamond (ed.)
Growth, Productivity, Unemployment, Cambridge, MA: MIT Press.
Hamilton, J. D. 1988. “Rational Expectations Econometric Analysis of Changes in Regime: An
Investigation of the Term Structure and Interest Rates,” Journal of Economic Dynamics and
Control, 12, 385-423.
Hayashi, F. and E. C. Prescott. 2002. “The 1990s in Japan: A Lost Decade,” Review of
Economic Dynamics, 5, 206–235.
Hulten, C. R. and R. M. Schwab. 1991. “Public capital formation and the growth of regional
manufacturing industries,” National Tax Journal, 44, 121–134.
Karadag, M., A. O. Onder, and E. Deliktas. 2005. “Growth of Factor Productivity in the
Turkish Manufacturing Industry at Provincial Level,” Regional Studies, 39,: 213-223.
Kim C. J. and C. R. Nelson. 1998. “Testing for Mean Reversion in Heteroskedastic Data Based
on Gibbs-Sampling-Augmented Randomization,” Journal of Empirical Finance, 5, 131-154.
Kim C. J. and C. R. Nelson. 1999. “State-Space Models with Regime Switching,” MIT Press:
Cambridge, Massachusetts.
Krueger, A. O. and B. Tuncer. 1982. “Growth of Factor Productivity in Turkish Manufacturing
Industries,” Journal of Development Economics, 11, 307-325.
Majumdar, S. K. 1996. “Assessing Comparative Efficiency of the State-owned, Mixed and
Private Sectors in Indian Industry,” Public Choice, 96, 1–24.
Martin, S., and D. Parker. 1997. The Impact of Privatisation. Ownership and Corporate
Performance in the UK. London and New York: Routledge.
Megginson, W. L., R. C. Nash, and M. van Randenborgh 1994. “The Financial and Operating
Performance of Newly Privatized Firms: An International Empirical Analysis,” Journal of
Finance, 49, 403–452.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
32
Megginson, W. L. and J. M. Netter. 2001. “From State to Market: A Survey of Empirical
Studies on Privatization,” Journal of Economic Literature, 39, 321–389.
Millward, R. 1988. “Measuring Sources of Inefficiency in the Performance of Private and
Public Enterprises in LDCs,” In Privatization in Less Developed Countries, edited by P. Cook
and C. Kirkpatrick. Brighton: Wheatsheaf.
Miyagawa T, Y. Sakuragawa, and M. Takizawa. 2005. “Productivity and the Business Cycle
in Japan: Evidence from Japanese Industry Data,” Empirical Analysis of Economic Institutions
Discussion Paper Series No.80.
Morrison, C. J. and A. E. Schwartz. 1996. “State infrastructure and productive performance,”
American Economic Review, 86, 1095–1111.
Nelson, R. A., and W. J. Primeaux, Jr. 1988. “The Effects of Competition on Transmission
and Distribution Costs in the Municipal Electric Industry,” Land Economics, 64, 338–346.
Onder, A. O. and A. Lenger. 2003. “Productivity in Turkish Manufacturing Industry: A
Comparative Analysis on the Basis of Selected Provinces,” Yapi Kredi Economic Review, 14, 1729.
Ozmucur, S. 2003. “Wage and Productivity Differentials in Private and Public,” EconWPA
Paper No: 0303002.
Parker, D. and H. L. Wu. 1998. “Privatisation and Performance: a Study of the British Steel
Industry under Public and Private Ownership,” Economic Issues 3, 31-50.
Rotemberg, J. J. and L. H. Summers. 1990. “Inflexible Prices and Procyclical Productivity,”
The Quarterly Journal of Economics, 105, 851-874.
Turut, S. A. and K. Becker. 2000. “Performance Differences between Public and Private Firms
in Turkey,” Atlantic Economic Journal, 28, 3.
Taymaz, E. and G. Saatci. 1997. “Technical Change and Efficiency in Turkish Manufacturing
Industries,” Journal of Productivity Analysis, 8, 461–475.
Tyler, W. G. 1979. “Technical Efficiency in Production in a Developing Country: an Empirical
Examination of the Brazilian Plastics and Steel Industries,” Oxford Economic Papers, 31, 477–
495.
Uygur, E. 1990. “Policy, Productivity, Growth and Employment in Turkey, 1960-1989 and
Prospects for the 1990s,” MIES Special Topic Study, ILO, Geneva.
Vecchi, M. 2000. “Increasing Returns, Labour Utilization and Externalities: Procyclical
Productivity in the United States and Japan,” Economica, 67, 229–244.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
33
Vickers, J. Yarrow G. 1988. Privatization: an Economic Analysis. Cambridge, MA: MIT Press.
Willner, J. 2001. “Ownership, Efficiency, and Political Interference”, European Journal of
Political Economy, 17, 723–748.
Willner, J. and Parker D. 2007. “The Performance of Public and Private Enterprise under
Conditions of Active and Passive Ownership and Competition and Monopoly”, Journal of
Economics, 90, 221–253
Yildirim, E. 1989. “Total Factor Productivity Growth in Turkish Manufacturing Industry
Between 1963-1983: An Analysis ”, METU Studies in Development, 16, 65-69
Yilmazkuday, H. 2009. “The Effects of Currency Crises in Emerging Markets on the Industrial
Sector: An Alternative Regime-Shifting Approach”, Emerging Markets Finance and Trade, 45,
31-48.
Zaim, O. and Taskin, F. 1997 “The Comparative Performance of the Public Enterprise Sector
in Turkey: A Malmquist Productivity Index Approach”, Journal of Comparative Economics, 25,
129-157.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
Table 1 - Priors for the Parameters
Parameters
Mean
Standard Deviation
p00
0.8
0.16
p11
0.8
0.16
μ0
0
1
μ1
0.5
1
σ2
–
–
Note: The prior distribution of σ is improper.
2
34
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
35
Table 2 - Gibbs-Sampling Results
Public
Private
Total
M1
SD1
MD1
M2
SD2
MD2
M3
SD3
MD3
p00
0.77
0.10
0.79
0.77
0.11
0.79
0.76
0.12
0.77
p11
0.96
0.03
0.96
0.94
0.05
0.95
0.94
0.04
0.95
μ0
2.40
0.77
2.39
2.05
0.82
2.07
2.33
0.81
2.33
μ1
4.23
1.01
4.26
3.01
0.91
3.02
3.42
0.90
3.41
σ
42.34
9.78
41.00
68.47
12.62
67.10
33.53
6.80
32.84
2
Notes: Mi , SDi and MDi stand for the mean, the standard deviation and the median, respectively. Results are
obtained by Gibbs-sampling. The first 2,000 draws are discarded and the next 10,000 are recorded. The productivity
growth rates have been multiplied by 100 in order to have the results in percentage terms. The sample size is 72 in
all equations.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
36
Table 3 - Expected Durations
Sector
p00
p11
E ( D0 )
E ( D1 )
Public
0.77824
0.96201
5
26
Private
0.77690
0.94034
4
17
Total
0.76540
0.94529
3
18
Notes: p00 and p11 are the same as in Table 2. E ( D0 ) and E ( D1 ) stand for the durations of state 0 and
state 1 of the productivities, respectively.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
37
Figure 1 – Productivity Series
250
200
150
100
50
0
1987
1990
1993
Public
1996
Private
1999
2002
2005
Total
Notes: All series are seasonally adjusted. The base year is 1997 (=100) for all series. The correlation
coefficient between public and private is 0.95; it is 0.97 between public and total; and it is 0.99 between
private and total.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
38
Figure 2 – Annual Growth of Productivity
Notes: The annual growth rates are calculated by the fourth log difference of the quarterly productivity
series. The shaded areas show the recessionary periods determined by the OECD. The correlation
coefficient between public and private is -0.16; it is 0.10 between public and total; and it is 0.96 between
private and total.
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
Figure 3 – Privatization in Turkey
% of GDP
3
2
1
0
1987
1990
1993
1996
1999
2002
2005
Source: Privatization Administration of Turkey (http://www.oib.gov.tr/index_eng.htm).
39
Yilmazkuday, International Journal of Applied Economics, 6(2), September 2009, 21-40
Figure 4 – Probabilities of Low-Productivity
Notes: The shaded areas show the recessionary periods determined by the OECD.
40