48 
UDC: 339.743(437.1/.2:497.4) 
In this paper we test the theory of 
purchasing power parity for the Czech 
Republic and Slovenia in comparison to 
Austria, Germany, France and Italy by 
employing data from January 1992 to 
December 2001. Results of unit root tests 
indicate that the eight time series of the 
real exchange rates of the koruna and 
the tolar are integrated of order one. 
Though some cointegration was found 
among the nominal exchange rates and 
selected consumer price indices, the 
presented results do not support the 
theory of purchasing power parity in any 
of the two observed economies. 
Key words: purchasing power parity, 
exchange rate, cointegration 
A b s t r a c t 
UDK: 339.743(437.1/.2:497.4) 
V tem prispevku preverjamo teorijo 
paritete kupne moči na Češkem in v 
Sloveniji v primerjavi z Avstrijo, Nemčijo, 
Francijo in Italijo na osnovi podatkov od 
januarja 1992 do decembra 2001. 
Rezultati testov enotnega korena so 
pokazali, da je vseh osem časovnih vrst 
realnega deviznega tečaja krone in 
tolarja integriranih prvega reda. Čeprav 
smo dokazali kointegracijo med 
nominalnimi deviznimi tečaji in indeksi 
cen življenjskih potrebščin, dobljeni 
rezultati ne potrjujejo teorije paritete 
kupne moči. 
Ključne besede: pariteta kupne moči, 
devizni tečaj, kointegracija 
I z v l e č e k 
JEL: E31, F31, F41 
Darja Boršič, Ph.D. 
Jani Beko, Ph.D. 
University of Maribor 
Faculty of Economics and Business 
Pariteta kupne moči na Češkem in v Sloveniji: 
Empirično preverjanje 
PURCHASING POWER PARITY IN THE CZECH 
REPUBLIC AND SLOVENIA: AN EMPIRICAL TEST 
NG, ŠT. 1–2/2007 IZVIRNI ZNANSTVENI ČLANKI/ORIGINAL SCIENTIFIC PAPERS 
5 
1 Introduction 
In the last few decades, the validity of the theory of purchasing power parity 
(PPP) has been scrutinized in numerous empirical papers. Froot and Rogoff (1995), 
Sarno and Taylor (2002) and Taylor and Taylor (2004) present reviews of relevant 
literature. Explicit research on PPP theory has yielded varying results, partly as a 
result of the different estimation techniques, observation periods and data sets 
that have been employed; and partly because of factors that complicate the law of 
one price, such as obstacles to international trade, the inclusion of transaction 
costs, pricing-to-market strategy, discretionary exchange rate management and 
changes in the structure of price indices. Researchers, however, agree on two 
issues related to this exchange rate theory (Rogoff 1996): first, real exchange 
rates tend to converge on levels predicted by PPP in the long run; and second, 
short-run deviations from the PPP relationship are substantial and variable. While 
there is a great deal of empirical work on PPP theory for developed market 
economies, similar studies for transition countries are rather rare. Varamini and 
Lisachuk (1998) analyze the case of Ukraine for the period 1992–1996 and gain 
evidence in favor of PPP, despite some short-run deviations. Christev and 
Noorbakhsh (2000) deal with six Central European Countries (Bulgaria, the Czech 
Republic, Hungary, Poland, Romania, and Slovakia) in the period from 1991 to 
1998. They find moderate proof of long-run equilibrium of prices and exchange 
rates, but conditions for the law of one price are violated. Pufnik (2002) and 
Payne et al. (2005) examine the Croatian economy, finding no support for PPP 
theory. Barlow (2004) also tests the theory for the Czech Republic, Poland and 
Romania using Johansen cointegration tests, but the conclusions for the time 
period 1994–2000 are mixed regarding different combinations of the exchange 
rates of selected countries. 
The present paper aims to expand the investigation of PPP for two advanced 
transition countries: the Czech Republic and Slovenia. Considering different views 
on how the process of economic transformation since the beginning of the nineties 
and its effects on reforming countries’ price mechanisms are compatible with 
rigorous assumptions of the theory of PPP (see Brada 1998), there is an obvious 
need for further empirical evaluation to supply clear-cut evidence on 
macroeconomic forces that govern the exchange rate behavior in the aforementioned 
economies. Because the majority of transition countries have undergone 
several phases of economic restructuring, these most likely also triggered shifts 
in their equilibrium real exchange rates. This suggests that, when comparing 
developed market economies with those still under economic reforms, the degree 
of a country’s similarity, especially in terms of trade pattern, level of development 
and the structure of relative prices, could importantly affect the assessment of 
PPP. In order to provide detailed estimates, this study is based on separate testing 
of PPP in the Czech Republic and Slovenia with reference to their main trading 
partners from the EU–15, i.e. Austria, Germany, France and Italy. From 1992 to 
2001, these four countries accounted for 56 percent of Slovenia’s exports and 
imports. In the same period their share in Czech exports amounted to 48 percent 
and they also covered 51 percent of Czech imports on average. 
´´
49 
BORŠIČ, BEKO: PURCHASING POWER PARITY IN THE CZECH REPUBLIC AND SLOVENIA: AN EMPIRICAL TEST 
The paper consists of three additional sections. In Section 
2, after describing the general model of PPP and presenting 
the relevant data, the stationarity of real exchange rates is 
dissected. Section 3 proceeds with a search for cointegration 
among nominal exchange rates, domestic consumer prices 
and foreign consumer prices by relying on Johansen’s 
methodology (1991). Concluding remarks are given in the 
final section. 
2 The Model of PPP and Unit Root Tests of Real 
Exchange Rates 
The general model of testing for PPP can be specified in 
the following form (Cheung and Lai 1993): 
et = .0 + .1pt + .2pt* + .t (1), 
where et stands for nominal exchange rates, defined as 
the price of foreign currency in the units of domestic 
currency; pt denotes domestic price index and pt* foreign 
price index; while .t stands for the error term showing 
deviations from PPP. All the variables are given in 
logarithmic form. In the strictest version of PPP, there are 
the following assumptions: .0=0, .1=1, .2=-1. The symmetry 
restriction applies such that absolute values of .1 and .2 are 
equal, whereas the limitation of being equal to one is called 
the proportionality restriction (Froot and Rogoff 1995). 
Throughout this study we utilized monthly data series 
for Slovenia from January 1992 and for the Czech Republic 
from January 1993 to December 2001 (for both countries), 
when the euro was put into circulation. Primary data included 
monthly averages of nominal exchange rates and consumer 
price indices gathered from the central banks of individual 
countries. Each of the exchange rates has been defined as 
the koruna (CZK) or tolar (SIT) cost of a unit of foreign 
currency. Consumer price indices used in this study for 
Slovenia refer to January 1992, while for the Czech Republic 
they refer to January 1993. 
The empirical analysis starts off with the most restrictive 
version of Equation 1, .1=1, .2=-1, that is, with testing the 
properties of real exchange rates. In the context of relative 
PPP, the movements in nominal exchange rates are expected 
to compensate for price level shifts. Thus, real exchange 
rates should be constant over the long run and their time 
series should be stationary (Parikh and Wakerly 2000). The 
real exchange rates are a function of nominal exchange rates 
and relative price indices in two observed economies. They 
are calculated from the nominal exchange rates using the 
consumer price indices: 
REt = Et (Pt*/ Pt) (2), 
where REt stands for the real exchange rate, Et is the 
price of a foreign currency in units of domestic currency, 
and Pt* and Pt represent the foreign price index and the 
domestic price index, respectively. Taking the logarithms of 
Equation 2, the real exchange rates are defined as: 
ret = et + pt* – pt (3). 
The graph of a stationary time series is not supposed to 
reflect any kind of a time trend. Figure 1 presents the graphs 
of real exchange rates of the Czech koruna (CZK) and the 
Slovenian tolar (SIT) in comparison to the Austrian schilling 
(ATS), the German mark (DEM), the French franc (FRF) 
and the Italian lira (ITL). The graphs of exchange rates show 
that after an initial real depreciation, from 1996 onwards, 
the Slovenian tolar experienced a systematic real 
appreciation in comparison to the currencies of the selected 
market economies. As can be seen from Figure 1, the regular 
real appreciation of the Czech koruna against the currencies 
of the four developed market economies in the 1993–2001 
period was partly interrupted only in 1997 reflecting 
exchange rate instability due to a domestic currency crisis. 
Such a pattern in real exchange rate movements is explained 
in the literature by a range of factors, including inherited 
macroeconomic imbalances in transition countries, mixed 
performance of chosen exchange rate arrangements, 
monetary difficulties arising from huge capital inflows, the 
inflationary impact of wage and price adjustments, and real 
exchange rate appreciation due to the catching-up process 
(Halpern and Wyplosz 1997; Brada 1998). 
For checking the stationarity of real exchange rates, the 
augmented Dickey-Fuller (1979) test was used, taking into 
account the following equation: 
t 
m 
i 
t t i t i 
Y ß ß t .Y .. Y . 
= 
- - 
. = + + + . + 
1 
1 2 1 (4), 
where ß1, ß2, . and .i are parameters of the test, t is linear 
time trend, Yt is the tested time series, .Yt-i=Yt-i-Yt-i-1 and m 
is selected so that the residuals (.t) are white noise. We test 
the null hypothesis H0: .=0, which implies that there is a 
unit root present and the time series is non-stationary. 
Following Barlow (2004), Equation 4 was estimated 
assuming ß2=0. In order not to unnecessarily lose too many 
observations in a relatively short time series, the orders of 
augmentation were set to m=6 for all tests of unit root by 
using critical values according to MacKinnon (1991). 
Campbell and Perron (1991) prefer determining the time lags 
according to a t-test. They argue that a VAR with a maximum 
number of lags should be carried out. If the last included lag 
is statistically significant, it is appropriate to use it in ADF 
regressions. The number of lags should be reduced as long 
as the last included lag is statistically significant. Also Ng 
and Perron (1995) argue that information-based rules (AIC, 
SIC) tend to select too low truncation lags, while the t-test 
is supposed to provide results with more robust size 
properties in models. In the present analysis, the estimates 
are obtained on the basis of time lags which correspond to 
the minimum value of the Akaike Information Criterion 
(AIC) and are in line with the t-test approach. 
Results of the augmented Dickey-Fuller test are shown 
in Table 1. Each calculation is stated twice, according to the 
time lag determined by the two approaches described above. 
Although AIC and the t-test select different time lags, the 
results of the ADF test using both selection criteria do not 
´´
50 
NG, ŠT. 1–2/2007 IZVIRNI ZNANSTVENI ČLANKI/ORIGINAL SCIENTIFIC PAPERS 
Source: The Czech National Bank and Bank of Slovenia. 
Notes: L stands for logarithm, R for real; the next three letters (ATS, DEM, FRF, ITL) represent the currencies of Austria, Germany, 
France and Italy, respectively, while the last two letters (CZK, SIT) denote the currencies of the Czech Republic and Slovenia, 
respectively. 1992:01=100 for Slovenia, 1993:01=100 for the Czech Republic. 
Figure 1: Real Exchange Rates of the Czech koruna and the Slovenian tolar
51 
BORŠIČ, BEKO: PURCHASING POWER PARITY IN THE CZECH REPUBLIC AND SLOVENIA: AN EMPIRICAL TEST 
contradict, but are rather similar. The figures show that the 
eight time series of the real exchange rates of the koruna 
and the tolar are integrated of order one, which means we 
cannot reject the hypothesis of the presence of the unit root. 
Thus, the ADF test confirms the graphical results of nonstationarity 
in the observed time series. 
3 Cointegration Analysis and Comments on Results 
When all restraints in Equation 1 are omitted (.1.1, 
.2.-1), it becomes the least restrictive version of PPP. The 
only requirement that remains is the signs of the coefficients. 
This implies that we are looking for any linear relationship 
among the observed variables that has stationary properties. 
Taking into account the unstable characteristics of nonstationary 
time series, the existence of a stationary 
relationship among them is more important than deviations 
of coefficients from the strict theory of PPP (Liu 1992). If a 
cointegration among nominal exchange rates, domestic 
consumer prices and foreign consumer prices is found and 
it is presented by the cointegrating vector of (1, .1, .2) 
(Equation 1), the validity of the theory of PPP is proven. 
Since we are looking for a stationary linear combination 
of three variables, the Johansen cointegration test is 
appropriate to use. This method is based on a VAR and can 
be briefly described as follows (Johansen 1991): 
Yt = A1Yt-1 + …. + AmYt-m + BXt + .t, (5), 
where A1, …, Am and B are matrices and the parameters 
of the model, t ranges from 1 to T, Yt is a vector of k variables, 
which are integrated of the first order, Xt is vector of 
deterministic variables and .t is a vector of innovations. VAR 
in Equation 5 can be also written as: 
.Yt = .Yt-1 + .
- 
= 
- 
. . 
1 
1 
m
i 
i t i 
Y + BXt + .t (6), 
where . =. - . 
=
m 
i 
i 
A 
1 
and .
= + 
. = - 
m
j i 
i j 
A 
1 
(7). 
Matrix . contains information about long-run variation 
of the time series. According to the Granger representation 
theorem (Engle and Granger 1987; Johansen 1991), matrix 
. can be divided into k x r matrices . and . with rank of r 
(r.k-1), so that .=..’ if . also has reduced rank r<k. Matrix 
. contains r linear cointegrating vectors, while matrix . 
presents adjustment coefficients of the error correction model. 
The number of cointegrating vectors is assessed by two 
statistics. The trace statistic (LRtr) tests H0: the number of 
cointegrating vectors is less than or equal to r, against the 
H1: the number of cointegrating vectors is k, where k is the 
number of endogenous variables for r=0, 1, ..., k-1. The trace 
statistic is specified as: 
.
= + 
= - - 
k
i r 
tr i 
LR r k T 
1 
( | ) log(1 . ) (8), 
where .i is the maximum eigenvalue of Ai in Equation 7. 
The maximum eigenvalue statistic (LRmax) checks H0: the 
number of cointegrating vectors is equal to r, and H1: the 
number of cointegrating vectors is equal to r + 1. LRmax can 
be calculated as follows: 
(9), 
where the abbreviations are the same as in Equation 8 
and the text above. 
Critical values for the Johansen cointegration test are 
stated in Johansen (1988) and Johansen and Juselius (1990). 
Osterwald-Lenum (1992) recalculated and extended them 
by handling the whole test sequence. Therefore, this study 
applies improved critical values of Osterwald-Lenum (1992). 
To undertake the Johansen cointegration test, an appropriate 
lag structure had to be found in order to remove serial 
correlation in the residuals. Estimation on the basis of VAR’s 
Akaike Information Criterion (AIC) and Final Prediction 
Error (FPE) gave the same lag specification for all eight 
cases under consideration. Figures for time lags are quoted 
next to the individual countries’ names in Table 4. 
Prior to cointegration analysis, it is necessary to establish 
the compatible orders of integration of the employed 
variables. For this reason, ADF tests were conducted for 
individual nominal exchange rates, domestic consumer 
prices and foreign consumer prices following the procedure 
described in the previous section. Results of unit root tests 
for nominal exchange rates are presented in Table 2, while 
Table 3 summarizes the unit root tests for selected consumer 
price indices. Again, AIC and a t-test were used to determine 
the number of time lags in ADF regressions. 
Notes: L stands for logarithm, R for real; the next three letters (ATS, DEM, FRF, ITL) represent the currencies of Austria, Germany, 
France and Italy, respectively, while the last two letters (CZK, SIT) denote the currencies of the Czech Republic and Slovenia, 
respectively. Critical values: -3.4890 (1%), -2.8870 (5%) and -2.5802 (10%). The subscripts indicate the value of m in Equation 4. 
Table 1: Results of the ADF Test for Real Exchange Rates of the Czech koruna and the Slovenian tolar 
´´
52 
NG, ŠT. 1–2/2007 IZVIRNI ZNANSTVENI ČLANKI/ORIGINAL SCIENTIFIC PAPERS 
All the nominal exchange rates in level form are found 
to be non-stationary, except the exchange rate of the 
Slovenian tolar to the French franc, which is stationary 
according to lag specification by t-test. A glance at the figures 
in Table 3 reveals that stationarity of five consumer prices 
is achieved (at the 5% significance level at least) only after 
the series are transformed into first differences. The Czech 
consumer price index, however, is specified as I(2). 
MacDonald (1993) claims that also in the case of different 
orders of integration, it is possible that the volatility of 
variables still implies a stationary linear combination among 
them. This is clearly impossible in the case of three variables 
being integrated of three different orders (Granger 1986). To 
retain consistency of the lag estimation criterion by ADF and 
cointegration tests, the integration orders of variables that enter 
cointegrating relations in our study were based upon the AIC. 
Table 4 and Table 5 list the results of the Johansen test by 
applying the basic model for PPP testing, i.e. Equation 1. 
From Table 4 it can be seen that for the Czech Republic 
limited evidence on cointegration among the nominal 
exchange rates and consumer prices was found, but only in 
comparison to France and Italy. In both pairs of countries 
the estimated coefficients appear to be statistically 
significantly different from zero. According to Equation 1, 
the signs of the coefficients of domestic prices should be 
positive, while signs of the coefficients of foreign prices 
should be negative. Thus, the signs of all cointegrating 
coefficients invalidate the PPP theory on the Czech data. 
Looking at Slovenia, values of LRtr and LRmax show that 
there is cointegration among the nominal exchange rates and 
consumer price indices in comparison to Austria, Germany 
and Italy. In all three cases the coefficients of domestic prices 
are proven to be statistically significantly different from zero, 
while for coefficients of foreign prices the standard errors 
are too high to conclude the same. The signs of estimated 
cointegrating coefficients, reported in Table 4, are again 
wrong to confirm PPP. Only the coefficient of Austrian 
consumer prices tends to have the right sign. In reference to 
France, there is no proof of cointegration either. In addition, 
the estimated coefficients are statistically insignificant and 
only the coefficient of French consumer prices has a sign 
corresponding to PPP theory. 
The presented results do not support the theory of PPP 
in any of the two observed economies. Such an outcome is 
in line with the rather weak empirical evidence on PPP 
reported for transition countries in the introductory part of 
this paper. The invalidity of PPP found in our study is also 
consistent with the real appreciation of the national 
currencies of the Czech Republic and Slovenia stated by, 
inter alia, Desai (1998) and Bole (1999). One part of real 
exchange rate appreciation can be attributed to the faster 
growth of domestic tradable prices compared to tradable 
prices of developed European economies, although this sort 
of real appreciation was substantially mitigated in Slovenia 
by monetary policy interventions on foreign exchange 
markets in order to preserve external competitiveness (Bole 
1999). In the Czech Republic, on the other hand, the 
contribution of relative prices of tradables to the real 
exchange rate appreciation was preponderant (Kovács et al. 
2002) and the domestic monetary authorities were, until 
1997, obliged to sustain the exchange rate peg. The Czech 
example, therefore, corroborates the findings of Barlow 
(2004) that implementation of a more rigid exchange rate 
policy in conditions of still volatile inflation and price inertia 
is to blame for violating PPP. 
Notes: L stands for logarithm, N for nominal; the next three letters (ATS, DEM, FRF, ITL) represent the currencies of Austria, Germany, 
France and Italy, respectively, while the last two letters (CZK, SIT) denote the currencies of the Czech Republic and Slovenia, 
respectively. Critical values: -3.4890 (1%), -2.8870 (5%) and -2.5802 (10%). The subscripts indicate the value of m in Equation 4. 
Table 2: Results of the ADF Test for Nominal Exchange Rates of the Czech koruna and the Slovenian tolar 
Notes: L stands for logarithm, CPI for consumer price index; C, S, A, G, F and I denote the Czech Republic, Slovenia, Austria, 
Germany, France and Italy, respectively. Critical values: -3.4890 (1%), -2.8870 (5%) and -2.5802 (10%). The subscripts indicate 
the value of m in Equation 4. 1Second difference. 
Table 3: Results of the ADF Test for Individual Consumer Price Indices in the Observed Countries
53 
BORŠIČ, BEKO: PURCHASING POWER PARITY IN THE CZECH REPUBLIC AND SLOVENIA: AN EMPIRICAL TEST 
In Slovenia, a far more important source of real exchange 
rate appreciation comes from faster growth of nontradable 
to tradable prices in comparison to relative prices of 
developed market economies. As documented in Kovács 
(2004), changes in relative labor productivity explain a 
considerable portion of nontradable/tradable relative price 
behavior in Slovenia in the 1992–2001 period. Besides this 
productivity-based real appreciation, relative mark-ups 
account for real appreciation in the case of Slovenia as well 
(see also Bole 2003). In addition, studies by Kutan and 
Dibooglu (1998) and Kovács (2004) imply that the variety 
of real shocks encountered by transition economies and 
expansive macroeconomic policies can significantly 
strengthen real exchange rate appreciation, the former via 
improving efficiency and boosting productivity, while the 
latter by increasing inflation differentials with respect to 
levels in developed market economies. 
4 Concluding remarks 
Testing for stationarity of real exchange rates of the 
Czech koruna and the Slovenian tolar showed no firm 
evidence in favor of PPP. After examining the stationarity 
of real exchange rates, the proportionality and symmetry 
restrictions were omitted. The Johansen cointegration 
technique was applied to find a long run linear relationship 
among chosen nominal exchange rates and individual time 
series of consumer prices. Although some cointegration was 
proven, the theory of PPP could not be confirmed. Regarding 
the low national price levels in both countries in question 
(see, for example, IEDP 2003; 2004) compared to levels in 
the EU–15, even after a decade of reforms, such a result is 
not unexpected. Another reason for failing to substantiate 
PPP could be the relatively short period of observation for 
such a long relationship to be detected among the observed 
variables.1 Since the early nineties both countries had already 
pursued a strategy of more or less successful gradual 
disinflation. Managing low variations of nominal exchange 
rates during periods of excessive inflation could also imply 
deviations from PPP. However, the empirical work 
completed so far reveals that the underlying cause of real 
exchange rate appreciation in Slovenia stems from 
differences in relative productivity gains and from steady 
price increases due to inadequate competition in the 
nontradable sector. 
Notes: ** (*) denotes rejection of the null hypothesis at the 1% (5%) significance level, respectively; figures in parentheses are 
standard errors. 1Critical values for LRtr at the 5% level are 29.68 (r=0), 15.41 (rd”1), and 3.76 (rd”2); and at the 1% level are 
35.65 (r=0), 20.04 (rd”1), and 6.65 (rd”2). 2Critical values for LRmax at the 5% level are 20.97 (r=0), 14.07 (r=1), and 3.76 (r=2); 
and at the 1% level are 25.52 (r=0), 18.63 (r=1), and 6.65 (r=2). 
Table 4: Results of the Johansen Cointegration Test for the Czech Republic and Slovenia 
1 Rogoff (1996) stresses that it takes three to five years for one 
half of the exchange rate deviation from the PPP level to be 
completed. 
´´
54 
NG, ŠT. 1–2/2007 IZVIRNI ZNANSTVENI ČLANKI/ORIGINAL SCIENTIFIC PAPERS 
References 
1. Barlow, D. (2004). Purchasing Power Parity in Three 
Transition Economies. Economics of Planning 36 (3): 
201–221. 
2. Bole, V. (1999). Financial Flows to a Small Open 
Economy: The Case of Slovenia. In The Mixed Blessing 
of Financial Inflows, Transition Countries in 
Comparative Perspective, ed. J. G<cs, R. Holzmann and 
M.L. Wyzan, 195–238. Cheltenham: Edward Elgar. 
3. Bole, V. (2003). Denarna politika v času odštevanja 
(Monetary Policy in the Time of Countdown). 
Gospodarska gibanja 346: 23–43. 
4. Brada, J.C. (1998). Introduction: Exchange Rates, Capital 
Flows, and Commercial Policies in Transition Economies. 
Journal of Comparative Economics 26 (4): 613–620. 
5. Campbell, J.Y., and P. Perron (1991). Pitfalls and 
Opportunities: What Macroeconomists Should Know 
About Unit Roots. NBER Working Paper Series: 
Technical Working Paper, no. 100. 
6. Cheung, Y., and K.S. Lai (1993). Long-Run Purchasing 
Power Parity During the Recent Float. Journal of 
International Economics 34 (1–2): 181–192. 
7. Christev, A., and A. Noorbakhsh (2000). Long-Run 
Purchasing Power Parity, Prices and Exchange Rates in 
Transition. The Case of Six Central and East European 
Countries. Global Finance Journal 11 (1–2): 87–108. 
8. Desai, P. (1998). Macroeconomic Fragility and Exchange 
Rate Vulnerability: A Cautionary Record of Transition 
Economies. Journal of Comparative Economics 26 (4): 
621–641. 
9. Dickey, D.A., and W.A. Fuller (1979). Distribution of 
Estimators for Autoregressive Time Series With a Unit 
Root. Journal of the American Statistical Association 74 
(366): 427–431. 
10. Engle, R.F., and C.W.J. Granger (1987). Co-integration 
and Error Correction – Representation, Estimation and 
Testing. Econometrica 55 (2): 251–276. 
11. Froot K.A., and K. Rogoff (1995). Perspectives on PPP 
and Long run Real Exchange Rates. In Handbook of 
International Economics Vol. III, ed. G. Grossman and 
K. Rogoff, 1647–1688. Elsevier Science. 
12. Granger, C.W.J. (1986). Developments in the Study of 
Cointegrated Economic Variables. Oxford Bulletin of 
Economics and Statistics 48 (3): 213–228. 
13. Halpern, L., and C. Wyplosz (1997). Equilibrium 
Exchange Rates in Transition Economies. IMF Staff 
Papers 44 (4): 430–461. 
14. Institute for Economic Diagnosis and Prognosis (IEDP) 
(2003). Slovenija in Avstrija – Ravni cen in plač (Slovenia 
and Austria – Levels of Prices and Wages. Bilten EDP 
26 (2–3). Maribor. 
15. Institute for Economic Diagnosis and Prognosis (IEDP) 
(2004). Slovenija in Češka – Ravni cen in plač (Slovenia 
and the Czech Republic – Levels of Prices and Wages). 
Bilten EDP 27 (1). Maribor. 
16. Johansen, S. (1988). Statistical Analysis of Cointegration 
Vectors. Journal of Economic Dynamics and Control 12 
(2–3): 231–254. 
17. Johansen, S. (1991). Estimation and Hypothesis Testing 
of Cointegration Vectors in Gaussian Vector 
Autoregressive Models. Econometrica 59 (6): 1551–1580. 
18. Johansen, S, and K. Juselius (1990). Maximum 
Likelihood Estimation and Inference on Cointegration – 
with Applications to the Demand for Money. Oxford 
Bulletin of Economics and Statistics 52 (2): 169–210. 
19. Kovács, M.A. (ed.); J. Benes; M. Klima; J. Borowski; 
M.K. Dudek; P. Sotomska-Krzysztofik; F. Hajnovic; and 
T. Žumer (2002). On the Estimated Size of the Balassa- 
Samuelson Effect in Five Central and Eastern European 
Countries. NBH Working Paper 5, July. 
20. Kovács, M.A. (2004). Disentangling the Balassa- 
Samuelson Effect in CEC5 Countries in the Prospect of 
EMU Enlargement. In Monetary Strategies for Joining 
the Euro, ed. Gy. Szap<ry and J. von Hagen, 79–105. 
Cheltenham: Edward Elgar. 
21. Kutan, A.M., and S. Dibooglu (1998). Sources of Real 
and Nominal Exchange Rate Fluctuations in Transition 
Economies. The Federal Reserve Bank of St. Louis, 
Working Paper 1998-022A. 
22. Liu, P. C. (1992). Purchasing Power Parity in Latin 
America: A Co-Integration Analysis. Weltwirtschaftliches 
Archiv 128 (4): 662–679. 
23. MacKinnon, J. (1991). Critical Values for Cointegration 
Tests. In Long-Run Economic Relationships: Readings 
in Cointegration, ed. R.F. Engle and C.W.J. Granger. 
Oxford: Oxford University Press. 
24. MacDonald, R. (1993). Long-Run Purchasing Power 
Parity: Is It For Real? The Review of Economics and 
Statistics 75 (4): 690–695. 
25. Ng, S., and P. Perron (1995). Unit Root Tests in ARIMA 
Models with Data-Dependent Methods for the Selection 
of the Truncation Lag.” Journal of American Statistical 
Association 90 (429): 268–281. 
26. Osterwald-Lenum, M. (1992). A Note with Quantiles of 
Asymptotic Distribution of the Maximum Likelihood 
Cointegration Rank Test Statistics. Oxford Bulletin of 
Economics and Statististics 54 (3): 461–472. 
27. Parikh, A., and E. Wakerly (2000). Real Exchange Rates 
and Unit Root Tests. Weltwirtschaftliches Archiv 136 (3): 
478–490. 
28. Payne, J.; J. Lee; and R. Hofler (2005). Purchasing Power 
Parity: Evidence from a Transition Economy. Journal of 
Policy Modeling 27 (6): 665–672. 
29. Pufnik, A. (2002). Purchasing Power Parity as a Long-Run 
Equilibrium: Co-Integration Test in the Case of Croatia (1991– 
1996). Croatian Economic Survey 1996–1999: 29–54. 
30. Rogoff, K. (1996). The Purchasing Power Parity Puzzle. 
Journal of Economic Literature 34 (2): 647–668. 
31. Sarno, L., and M.P. Taylor (2002). Purchasing Power 
Parity and the Real Exchange Rate. IMF Staff Papers 49 
(1): 65–105. 
32. Taylor, A.M., and M.P. Taylor (2004). The Purchasing 
Power Parity Debate. NBER Working Paper 10607. 
33.Varamini, H., and H. G. Lisachuk (1998). The Application 
of Purchasing Power Parity to Ukraine by Using the 
Cointegration Approach. Russian and East European 
Finance and Trade 34 (3): 60–69.