# Causes of Deviations from the CNB’s Inflation Targets: An Empirical Analysis

**ABSTRACT** This paper provides an empirical analysis of the factors that caused deviations from the CNB’s inflation targets during the first ten years of inflation targeting in the Czech Republic. While in the short term, shocks to agricultural producer prices represent the most important factor, exchange rate shocks are clearly the most important medium-term factor. At the same time, it could not be proved that monetary policy created any major shocks contributing to the non-fulfillment of inflation targets, although some role thereof cannot be excluded either.

**0**Bookmarks

**·**

**47**Views

- [Show abstract] [Hide abstract]

**ABSTRACT:**In this paper, we (1) examine the interactions of financial variables and the macroeconomy within the block-restriction vector autoregression model and (2) evaluate to what extent the financial variables improve the forecasts of GDP growth and inflation. For this reason, various financial variables are examined, including those unexplored in previous literature, such as the share of liquid assets in the banking industry and the loan loss provision rate. Our results suggest that financial variables have a systematic and statistically significant effect on macroeconomic fluctuations. In terms of forecast evaluation, financial variables in general seem to improve the forecast of macroeconomic variables, but the predictive performance of individual financial variables varies over time, even though it strengthens during the 2008–2009 crisis. The results give some support for the risk-taking channel of monetary policy, as the level of the monetary policy rate is positively associated with the loan loss provision rate of commercial banks. Finally, a more stable financial system is found to contribute to faster economic growth. KeywordsMacroeconomic and financial linkages–Vector autoregressions–ForecastingEconomic Change and Restructuring 01/2012; - SourceAvailable from: cnb.cz
##### Article: DO FINANCIAL VARIABLES HELP PREDICT MACROECONOMIC ENVIRONMENT? THE CASE OF THE CZECH REPUBLIC

[Show abstract] [Hide abstract]

**ABSTRACT:**In this paper, we examine the pseudo out-of-sample forecasting ability of various financial variables – credit to private sector, the share of liquid assets in the banking industry, the share of non-performing loans, the loan loss provisions rate and equity prices – for inflation and GDP growth in the Czech Republic. To do so, we estimate a block-restriction vector autoregression model consisting of Czech macroeconomic and financial variables and euro area macroeconomic variables in 1999:1-2009:9. Our results suggest that financial variables have a systematic and statistically significant effect on macroeconomic fluctuations. In terms of forecast evaluation, financial variables in general seem to improve the forecast of macroeconomic variables, but the predictive performance of individual financial variables varies over time, in particular during the 2008-2009 crisis. Among other things, the results give some support for the risk-taking channel of monetary policy, as the level of monetary policy rate is negatively associated with the loan loss provisions rate of commercial banks. Finally, more stable financial system is found to contribute to faster economic growth.Theoretical and Applied Economics. 01/2011; 5(558)(supplement)(5(558)(supplement)):778-797. - SourceAvailable from: Borek VasicekJaromir Baxa, Jiri Bohm, Roman Horvath, Lubos Komarek, Petr Kral, Magdalena Morgese Borys, Filip Rozsypal, Branislav Saxa, Borek Vasicek

Page 1

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

46

CHAPTER 4

47

CHAPTER 4

CAUSES OF DEVIATIONS OF INFLATION FROM CNB TARGETS –

AN EMPIRICAL ANALYSIS

TOMÁŠ HOLUB

Page 2

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

48

1. INTRODUCTION

This paper offers an empirical analysis of the causes of deviations of inflation from the CNB targets

during the first ten years of inflation targeting. Section 2 presents a review of such causes as

contained in the existing literature dealing with inflation targeting in the Czech Republic, and it

applies a simple cross-correlation analysis and Granger causality tests to identify which explanatory

variables – and with what time lag – seem to have a statistically significant relationship to such

deviations of inflation from the target. Section 3 then offers estimates of two VAR models

examining impulse responses of the deviations of inflation from the target to the individual shocks,

and a variance decomposition of these deviations.

Based on those analyses, it can be concluded that the most important short-term factor of deviations

of inflation from the target are the shocks relating to prices of agricultural producers. In the

medium-term horizon, however, the development of the real exchange rate gap unequivocally takes

over as the most important factor. The main common macro-economic feature of the two periods of

the most significant undershooting of inflation targets indeed was a noticeable and unexpected

strengthening of the exchange rate of the Czech crown. Difficulties caused by the exchange rate

were further accentuated by the coincidence with other factors, although these were less significant

and – as such – they would have probably only resulted in a less distinct and merely short-term

undershooting of inflation targets. The analysis, however, does not provide an answer to the

question of why the target fulfilment was asymmetrically skewed in the undershooting direction

and why the periods characterised by depreciation corrections of the exchange rate did not see any

overshooting of the CNB targets.

2. LITERATURE REVIEW AND BASIC STATISTICAL TESTS

Fulfilment of the CNB targets in the course of the first decade of inflation targeting was affected to

a significant extent by two episodes characterised by noticeable undershooting of the targets,

namely in the years 1998–99 and 2002–03. The existing literature has consented on a list of the

relevant causes of such distinct target undershooting (see e.g., Kotlán and Navrátil, 2003; Geršl and

Holub, 2006). They involved declining food prices (in both periods), low oil prices (in both

periods), a pause in deregulations (in 2002–03), fiscal and monetary restrictions (in 1998–99), a

growth slow-down in the EU (in 2002–03), and strengthening of the exchange rate of the Czech

crown (in both periods). To compare, the CNB assigned in its Inflation Reports the 1998–99 target

undershooting in the first place to the food prices, and then to the weak local demand, strong

exchange rate, and low oil prices in 1998. In the years 2002–03, the CNB pointed out particularly to

the disinflationary effects of regulated prices, food prices, appreciation of the exchange rate, and the

international developments.

A disadvantage regarding the above-described papers lies – with the exception of the CNB Inflation

Reports – in that they would not quantify the relative importance of the individual factors. At the

same time, they do not take into consideration any endogenous links of those factors, such as, e.g.,

the effect of the exchange rate, the monetary and fiscal policy or international developments on the

local food prices and on regulated prices, as well as the feedback effects of the price developments

on the monetary policy decisions. Without considering such links, any quantification would be

difficult, something which throws a shade of doubt also on the analysis of both of the two periods

as presented in the Inflation Reports. The CNB forecasts were compiled for both of those periods

CHAPTER 4

49

using short-term forecasting methods, and they were also assessed with these tools afterwards

concerning their fulfilment. Those methods were not particularly suitable for capturing any

medium-term endogenous relations in the economy (see Coats, et al., 2003). The existing QPM

model was not used until mid-2002, i.e. until the second exchange rate appreciation episode was

already peaking. Therefore, it was possible to use it in the analysis of the causes of deviations of

inflation from the target (see Filáček, 2007; Antoničová, et al., 2008) only starting at the beginning

of 2004 when inflation already commenced to return to the target.

Holub and Hurník (2008) express the opinion that the exchange rate development was the common

key feature of both of the two target undershooting episodes. For the sake of brevity, however, no

sufficiently detailed evidence for such claims has been presented, and/or, such evidence is only

mentioned in a footnote. The present paper removes the above-described shortcomings and brings

about a detailed empirical analysis of the causes of the deviations of inflation from the target.

The cited papers – despite their shortcomings – may help us compile a list of variables to be further

examined as regards their statistical and economic significance in the explanation of deviations of

inflation from the target. More specifically, this paper focuses on the following variables: the real

exchange rate, agricultural producer prices, crude oil prices in USD, foreign and domestic economic

activity, and domestic real interest rates.

All variables have been used on a quarterly basis in the form of their deviations from estimated

equilibrium levels, because the variable in focus – i.e., deviations of inflation from the target1 – may

also be considered as deviations from the equilibrium. As far as prices of crude oil and agricultural

producer prices are concerned, they always represent deviations from the trend as estimated with

help of the Hodrick-Prescott (HP) filter.2 In the event of the real exchange rate gap, real interest

rates gap, and domestic and foreign output gaps, we work with two alternative estimates. The first

one is based the structural Kalman Filter, which has been applied by the CNB in its analyses and

forecasts (see Beneš and N’Diaye, 2003); the other one is based on the HP Filter. The advantage in

using the Kalman Filter is in that the estimated “gap” values correspond to the view of the central

bank concerning the development of the Czech economy. On the other hand, however, its

application may cause certain distortions in the analyses in the form of an implicit “a priory”

presumption about the course of monetary transmission as captured by the QPM model applied by

the CNB. The application of an alternative estimate with help of the HP Filter, which represents a

non-parametric filter, therefore, may be understood as robustness check of the results.3

All the applied data and estimates correspond to the CNB forecast as published in its Inflation

Report I/2008, and cover the first ten years of inflation targeting, i.e., the period of time from the

first quarter of 1998 until the fourth quarter of 2007.

1 In the period 1998–2001, this is a deviation of net inflation from the middle of the target, which had been

extrapolated in a linear manner into the individual quarters from the year-end values. As far as the later period is

concerned, talk goes here of deviations of the headline inflation from the middle of the target range, or the CNB point

target.

2 We also tried to use year-on-year changes instead of deviations from the H-P trends because unexpected changes of

those prices may cause a deviation of inflation from the target irrespective of whether they concern shifts in the long-

term equilibrium or temporary fluctuation. However, the results were similar in quality, so we do not present them in

the subsequent text.

3 As concerns the real interest rates gap, there also exists a difference in that – as regards the Kalman filter – nominal

rates have been deflated by partly forward-looking expectations, while inflation expectations in the event of the HP

filter have been deemed as purely backward-looking.

Page 3

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

50

Testing of statistical significance of the above-described variables was first made with help of a

simple cross-correlation analysis and Granger causality pair-wise tests. One of the variables was

always represented by deviations of inflation from the target, while the other side was always

occupied by one of the above-described explanatory variables. The results are reported in Table 1.

The results of both tests are always shown for the time lag which would maximise the value of the

correlation coefficient, or the probability level of the Granger causality test.

All of the above-described explanatory variables have a statistically significant and in the majority

of cases economically intuitive correlation with the deviations of inflation from the target with a

time lag ranging from “zero” (for agricultural producer prices) up to 10 quarters (for foreign output

gap). In the event of real exchange rate, the time lag amounts to three quarters; in the event of real

interest rate gap it reaches 0-2 quarters (depending on the method of calculation). The only

surprising outcome is the long lag and the sign attached to the foreign output gap, which runs

counter to the economic intuition, indicating that the correlation may be spurious rather than

reflecting a true causal relationship.

All of the explanatory variables, with the exception of agricultural producer prices4 and the real

interest rate gap estimated with help of the HP Filter, at the same time, Granger-cause the

deviations of inflation from the target at least at the 10-percent significance level, usually with a

time lag of 1-2 quarters (only with the foreign output gap the time lag would extend to 10 quarters,

in line with the correlation analysis, making this relationship hard to interpret). These results,

therefore, justify further research of relationships between those variables and the deviations of

inflation from the target.

Table 1: Factors of inflation deviations from target – cross-correlation analysis and Granger

causality tests

Explanatory variables Correlation coefficient 1, 2)

Granger causality test – probability2)

Foreign output gap – KF

Foreign output gap –HP

Oil prices

Real exchange rate gap - KF

Real exchange rate gap - HP

Agricultural producer prices

Output gap – KF

Output gap – HP

Real interest rate gap – KF

Real interest rate gap – HP

-0.40* (10)

-0.44* (10)

0.51* (3)

-0.59* (3)

-0.52* (3)

0.58* (0)

0.55* (0)

0.47* (1)

-0.51* (2)

-0.75* (0)

4.0 % (10)

5.1% (10)

0.8% (1)

0.0% (1)

0.0% (1)

14.1% (1)

0.5% (2)

5.4% (1)

0.0% (0)

14.1% (4)

Source: Own calculations.

Notes: HP denotes estimates produced with help of the HP Filter, KF denotes estimates produced with help of the

Kalman Filter; 1) * denotes statistically significant results at the 5% probability level. 2) The numbers in

brackets show the time lag between explanatory variables and the deviations of inflation from the target

maximising the statistical significance of the identified relationship.

4 The statistical insignificance of agricultural producer prices has probably resulted from the fact that the relationship

– according to the performed correlation analyses (and in line with empirical experience) – was contemporaneous,

which cannot be captured by definition by the Granger causality test.

CHAPTER 4

51

Also examined was the reverse correlation and causality directed from inflation deviations from the

target to the analysed variables, which – however – is not presented in Table 1 for the sake of

brevity. The reverse causality was statistically significant at least at the 10% probability level with

almost all variables, with the exception of the foreign output gap estimated with help of the Kalman

Filter. In the case of oil prices, which can be deemed as a purely exogenous factor, this conclusion

is not very intuitive. It can be presumed – in view of the small time lag – that it rather represents a

co-incidence of both variables than any causal relationship. In the case of domestic variables,

however, the option of a reverse causality is not surprising and it only confirms opinions expressed

in the introduction to this Section, namely that a number of factors used in the literature so far to

explain deviations of inflation from the target is in reality of an endogenous nature, which needs to

be considered in the analysis.

3. ESTIMATES OF VAR MODELS

Consideration of the endogenous links among the individual variables can be performed by way of

estimates of the VAR models, which are presented in the present Section. Those VAR models at the

same time allow for the identification of not only the statistical significance of the individual factors

with help of examining impulse responses to shocks, but also for an analysis of the economic

significance via a variance decomposition of the deviations of inflation from the target. More

specifically, two models have been estimated, which can be expressed in general as follows:

Y

ttt

aLY

ν+=

(1)

[]

tttttttt

gapirgap pigapgdp gapczv gap ergappoil gap eaY_,_,_,_,_,_,_

=

(2),

where ea_gap denotes the foreign output gap (approximated by the effective eurozone, i.e., with the

individual countries weighed according to their shares in the Czech foreign trade), poil_gap denotes

the deviation of the USD crude oil price from the equilibrium, er_gap denotes deviation of the real

exchange rate from the equilibrium, czv_gap denotes deviation of agricultural producer prices from

the equilibrium, gdp_gap denotes the domestic output gap, pi_gap denotes deviation of inflation

from the target, ir_gap denotes the real three-month interest rate gap, while

residuals and L stands for the lag operator. Both of the models differed only in one sense: whether

they included the gaps of foreign and domestic demand, the real exchange rate, and real interest rate

estimated with help of the Kalman Filter or the HP Filter.

tν means the vector of

Shocks to the individual variables were identified in a standard manner on the basis of the Cholesky

decomposition, with the variables ordered according to Equation (2). This is a fairly standard

ordering, which corresponds to the view of the transmission of shocks in a small open economy

applying inflation targeting. The ordering means an implicit presumption that shocks to exogenous

variables (ea_gap and poil_gap) may have an immediate impact on the exchange rate of the Czech

crown, rather than the other way round. The exchange rate, together with the agricultural producer

prices and the domestic output gap may then directly affect inflation deviations from the target

which, however, does not influence immediately the said variables. The monetary policy then

responds, under the inflation targeting regime and via the set-up of the interest rates, to the

Page 4

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

52

development of all available information. The lag length in the model was determined at 2 quarters,

in view of the relatively short time series and making use of the standard tests.5

A variance decomposition of the deviations of inflation from the target is presented in Figure 1 for

the VAR model making use of estimates made with help of the Kalman Filter, and in Figure 2 for

the model with the input variables estimated with help of the HP Filter.

It has turned out that in the short term the most important factors of the deviations of inflation from

the target (apart from the shocks to inflation itself) are represented by the agricultural producer

prices and – in the case of the model making use of estimates based on the HP Filter – also the

crude oil prices. Their influence, however, would get weaker in the longer run. Shocks to the real

interest rates, i.e. monetary policy shocks, hold roughly a 15-percent share in both models on the

variance of the deviations of inflation from the target on the horizon of approx. 2-4 quarters, and

that share is at the edge of statistical significance. However, on the horizon of one year and longer,

the outspokenly most distinct factor is represented by the real exchange rate gap, which explains a

substantially larger portion of the variance of the deviations of inflation from the target (over 35 %)

than the other macro-economic variables, and its influence is statistically significant. As far as the

model making use of the HP Filter is concerned, the shocks to the domestic output gap are also at

the edge of statistical significance on the horizon exceeding 6 quarters; the model based on the time

series derived from the Kalman Filter, however, would not support this conclusion.

Figure 1: Variance decomposition of the deviations of inflation from the target (Kalman

Filter)

0

10

20

30

40

50

60

70

80

123456789 10

EA_GAP

CZV_GAP

IR_GAP

POIL_GAP

GDP_GAP

ER_GAP

PI_GAP

5 The application of the Schwarz criterion would result in preference of only one quarter; using the Akaike

information criterion would require the application of time lags in excess of two quarters, which would be facing

limitations in the form of short time series. Under this situation, a lag of two quarters was selected as a compromise

solution. We also examined the robustness of our results with respect to the selection of a shorter time lag, which

appeared to be satisfactory.

CHAPTER 4

53

Figure 2: Variance decomposition of the deviations of inflation from the target (HP Filter)

0

10

20

30

40

50

60

70

80

12345678910

EA_GAP

CZV_GAP

IR_GAP

POIL_GAP

GDP_GAP

ER_GAP

PI_GAP

The impulse responses of the deviations of inflation from the target to shocks affecting the

individual variables are shown in Figure 3 for the model using variables derived from the Kalman

Filter and in Figure 4 for the model using variables derived from the HP Filter. Some conclusions

would be qualitatively identical for both models. Shocks hitting the agricultural producer prices are

statistically significant for short time lags; however, their impact would gradually decline and

would quickly become statistically insignificant. The real exchange rate shocks are most effective

with a lag of 4-5 quarters, and an overvaluation (undervaluation) of the real exchange rate results in

a statistically significant decline (increase) of inflation below (above) the target. The maximum real

exchange rate pass-through into inflation equals roughly 28–38%, which approximately matches the

conclusions from the previous studies applicable to the Czech Republic (see Babestkaia, 2007).6

Both of the models also consent that international demand shocks would not have any statistically

significant impact on the deviations of inflation from the target. The said conclusions, therefore,

may be considered as reasonably robust.

6 Usually, the pass-through of nominal exchange rate shocks to inflation would be analysed, with results slightly

lower than the figures quoted here. However, under the presumption that a portion of the exchange rate shock would

be reflected in inflation with a delay shorter than one quarter, it is required to have more than a proportionate shock to

the nominal exchange rate to change the real exchange rate.

Page 5

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

54

Figure 3: Impulse responses of deviations of inflation from the target to shocks (Kalman

Filter)

-1.2

-0.8

-0.4

0.0

0.4

0.8

12345678910

Response of PI_GAP to EA_GAP

-1.2

-0.8

-0.4

0.0

0.4

0.8

123456789 10

Response of PI_GAP to POIL_GAP

-1.2

-0.8

-0.4

0.0

0.4

0.8

123456789 10

Response of PI_GAP to ER_GAP

-1.2

-0.8

-0.4

0.0

0.4

0.8

123456789 10

Response of PI_GAP to CZV_GAP

-1.2

-0.8

-0.4

0.0

0.4

0.8

123456789 10

Response of PI_GAP to GDP_GAP

-1.2

-0.8

-0.4

0.0

0.4

0.8

12345678910

Response of PI_GAP to IR_GAP

Response to Cholesky One S.D. Innovations ± 2 S.E.

Figure 4: Impulse responses of deviations of inflation from the target to shocks (HP Filter)

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to EA_GAP

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to POIL_GAP

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to ER_GAP

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to CZV_GAP

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to GDP_GAP

-1.5

-1.0

-0.5

0.0

0.5

1.0

123456789 10

Response of PI_GAP to IR_GAP

Response to Cholesky One S.D. Innovations ± 2 S.E.

CHAPTER 4

55

However, some partial differences exist between the two estimates, which urge for cautiousness in

the interpretation of their results. The model applying inputs from the HP Filter identified shocks

hitting the global crude oil prices as statistically significant for short-time lags, while the model

with the Kalman Filter did not. The same applies in the medium-term horizon also for the impact of

domestic output gap shocks, which is surprising – among other things – also in view of the fact that

the output gap estimated with help of the Kalman Filter, contrary to the HP Filter, explicitly

considers the relationship between this unobservable variable and inflation. The last noteworthy

difference relates to the influence of real interest rates, which are statistically significant in both

cases but – contrary to expectations – only in the short horizons of 2-3 quarters and, moreover, in

both of the models with the opposite sign (an intuitive one in the model with inputs from the

Kalman Filter, and an counter-intuitive one in the model with inputs from the HP Filter). Such

results can most probably be assigned to the existence of certain general difficulties concerning the

identification of the monetary transmission with help of the VAR models (see for example, the

discussion in Arnoštová and Hurník (2005) and the references contained in this paper), and to

uncertainties in the measuring of real interest rates (forward-looking vs. backward looking inflation

expectations).

4. CONCLUSION

The results presented here thus show on the whole that the most important role in relation to the

deviations of inflation from the target was probably played by the exchange rate. This reflects the

fact that the main common feature of the two most important periods of undershooting of the

inflation targets in the years 1998–99 and 2002–03 was represented by a significant and unexpected

strengthening of the exchange rate of the Czech crown (roughly by 8 % towards DEM, and by 15 %

towards euro, respectively). The difficulties caused by the exchange rate were further accentuated

by certain concurrence with other short-term factors, including in particular the development of

agricultural producer prices and probably also the development of the crude oil prices. Those

factors, though, were of lesser significance and they would probably in themselves have resulted

only in less distinct and shorter episodes of inflation targets undershooting.7

Based on the econometric methods applied, it is not possible to prove or exclude that the monetary

policy in itself would create any significant shocks contributing to the non-fulfilment of inflation

targets. In any case, it seem obvious at least that the monetary policy did not respond to exchange

rate developments sufficiently quickly and strongly to prevent distinct episodes of non-fulfilment of

inflation targets. In the first place, it attempted to restrain the scope of appreciation by foreign

exchange interventions, the effect of which – however – proved to be fairly small and could not

prevent the occurrence of relatively long periods of the exchange rate overvaluation (see Geršl and

Holub, 2006). At the same time, the CNB – like a number of other analysts – may have

underestimated the strength of the exchange rate pass-through into inflation.8 It is also possible to

express intuitively the hypothesis that in particular in the first of the above-described “troubled”

7 It remains an open question whether the given concurrence was only a coincidence or whether it was caused by

some hidden causal factors. For example, the development of the global economy could have affected the prices of

crude oil, the global prices of food as well as the development on the foreign exchange markets.

8 In the first of the said periods, the short-term forecasting methods were applied, which – however – could hardly

have a sufficient guidance in the data from the period of fixed exchange rate; in the other case, the exchange rate pass-

through into inflation in the newly introduced QPM model was expertly restrained due to a lack of confidence in

sufficient downward flexibility of prices.

Page 6

EVALUATION OF THE FULFILMENT OF THE CNB’S INFLATION TARGETS 1998–2007

56

periods, monetary policy in itself could contribute to the appreciation of the foreign exchange rate,

when it was keeping the nominal interest rates high and the real interest rates in a restrictive

position, even though the currency crisis was already fading out. It cannot be excluded that such

policy may have been based also on certain asymmetry of the preferences of the CNB concerning

the announced inflation targets, i.e., higher concerns regarding their overshooting than their

undershooting as a result of the understanding of inflation targeting as a means for achieving

disinflation and obtaining credibility, or any other monetary policy goals (e.g., efforts for the

overall stabilisation of the situation after the preceding currency crisis).

Similarly, it is not possible to find out with help of the selected approach why the non-fulfilment of

the target was on average significantly skewed towards its undershooting and, for example, no

overshooting of the targets occurred during the periods of exchange rate depreciation.

REFERENCES

ANTONIČOVÁ, Z., K. MUSIL, L. RŮŽIČKA, AND J. VLČEK (2008): “Vyhodnocení kvality a úspěšnosti

prognóz – historický přehled” (Evaluation of the Quality and Success Rate of Forecasts –

A Historic Overview), in this volume.

ARNOŠTOVÁ, K. AND J. HURNÍK (2005): “The Monetary Transmission Mechanism in the Czech

Republic (evidence from VAR analysis),” CNB Working Paper, 4/2005.

BABETSKAIA, O. (2007): “Transmission of Exchange Rate Shocks Into Domestic Inflation: The

Case of the Czech Republic,” CNB Working Paper, 12/2007.

BENEŠ, J. AND P. N’DIAYE (2003): “A Multivariate Filter for Measuring Output and the NAIRU,”

in The Czech National Bank’s Forecasting and Policy Analysis System, edited by W. Coats, D.

Laxton, and D. Rose. Prague, Czech National Bank, pp. 99–118.

COATS, W., D. LAXTON, AND D. ROSE, EDS. (2003): The Czech National Bank’s Forecasting and

Policy Analysis System,” Prague, Czech National Bank (http//:www.cnb.cz).

FILÁČEK, J. (2007): “Why and How to Assess Inflation Target Fulfilment,” CNB Working Paper,

11/2007.

GERŠL, A. AND T. HOLUB (2006): “Foreign Exchange Interventions Under Inflation Targeting: The

Czech Experience,” Contemporary Economic Policy, vol. 24, no. 4 (October), p.p. 475–491.

HOLUB, T. AND J. HURNÍK (2008): “Ten Years of Czech Inflation Targeting: Missed Targets and

Anchored Expectations,” Emerging Markets Finance and Trade, vol. 44, no. 6

(November/December), pp. 59–79 (forthcoming).

KOTLÁN, V. AND D. NAVRÁTIL (2003): “Inflation Targeting as a Stabilisation Tool: Its Design and

Performance in the Czech Republic,” Finance a úvěr – Czech Journal of Economics and

Finance, vol. 53, no. 5–6, p.p. 220–242.

#### View other sources

#### Hide other sources

- Available from Tomas Holub · Jul 3, 2014
- Available from cnb.cz