241Acta všfs, 3/2011, vol. 5
Applicability of Bankruptcy Models
at Agricultural Companies
Použitelnost bankrotních modelů
na zemědělské podniky
JAN SUŠICKÝ
Abstract
This part will be dedicated to the use of bankruptcy models in the evaluation of agricultural
companies. Bankruptcy models are tools used for the evaluation and prediction of
future development of companies. Their use is clearly linked to the problems associated
with the use of bankruptcy models developed abroad. These models, resulting from different
economic circumstances, often had difficulty in actually describing the real situation
of companies in the Czech Republic. In this part were used 6 bankruptcy models for the
analysis of agricultural companies (Z’’-Score, Z-Score, ZETA, IN99, IN01, IN05 and a model
created by Taffler and Tisshaw). The most successful models were Z-Score, ZETA and the
Czech model IN99. The Czech model IN01 was identified as the least suitable model.
Keywords
model, bankruptcy, ratio indicator, agricultural companies
Abstrakt
Tento příspěvek bude věnován bankrotním modelům, které jsou velmi využívaným nástrojem
hodnocení a predikce budoucího vývoje podniků. S jejich využitím jsou bezpochyby
spojeny také problémy, související s možností využití modelů vytvořených v zahraničí.
Tyto modely, vzniklé za odlišných ekonomických podmínek, mnohdy jen s obtížemi reálně
popisují skutečnou situaci podniků v České republice. V tomto příspěvku bylo pro analýzu
zemědělských podniků použito 6 bankrotních modelů (Z“-Score, Z-Score, ZETA, IN99, IN01,
IN05 a model vytvořený Tafflerem a Tisshawem). Jako nejúspěšnější modely se ukázaly
modely Z-Score, ZETA a český model IN99. Jako nejméně vhodný model se ukázal český
model IN01.
Klíčová slova
model, bankrot, poměrový ukazatel, zemědělské podniky
Introduction
Financial analysis has recently progressed from one-dimensional models to so-called
bankruptcy models (also known as the prediction of financial distress). These methods of
financial analysis are the result of efforts to use a single indicator to determine whether
a company is financially healthy or not.
One-dimensional models, which are represented by relative indicators, do not affect the
company as a whole, but only a certain area. This disadvantage removes multivariate mod-
Acta všfs, 3/2011, vol. 5242
els that assess the overall situation by the individual ratios which are combined to produce
a single value. The resulting figure would be a likely clue to the future development of the
company. These models can be divided into two groups.
The first group is defined as credibility models. The result of these models is a rating that
reflects the current creditworthiness of a company. This information is used to assess
whether the company will be viable in the future or not. The result of these models is the
inclusion of the company into one of the groups that characterizes the expected rating
of the company. Examples of credibility models are the Grünwald model or the Kralicek
Quick Test [1]. The second group consists of bankruptcy models, which aim to predict the
impending bankruptcy of the company. An example of bankruptcy model is the Altman
Index, written by E.I. Altman, which has been issued in the U.S. in a few modifications. In
the Czech Republic, bankruptcy models were also developed based on economic conditions
in the Czech Republic. The most popular are the IN Indexes created by Inka and Ivan
Neumaier [6].
Bankruptcy models, which were generated abroad, and based on other business conditions,
often do not include specific economic and business environment factors in the
Czech Republic. In particular, insolvency, which affects many companies and subsequent
secondary insolvency, in which companies receive as a result of unpaid debts incurred in
insolvency business partners.
Important factors that may reduce the successful application of Altman models for Czech
companies are different conditions under which the index was created and subsequently
tested. This is the difference in the external economic environment of business (growth
rate of inflation, etc.), legal environment (impact on payment discipline, the possibility of
recovery of outstanding debts, enforcement of tax obligations, investment options) and
different methodological conditions (different way to create input data, option pricing
etc.).This part will deal with evaluating the success of selected Czech and foreign models
applied to agricultural companies. It particularly aims at evaluating which models are
most successful in predicting the possible bankruptcy of agricultural companies, which
are the least successful.
On the other hand, from the available literature, studies and research conducted in the
Czech Republic and abroad cannot be perceived that the bankruptcy model developed
in one country cannot be successfully applied in another country. Mostly it is a “universal
model” of its successful use depends on economic conditions in which businesses are
located and the reliability of accounting data on which the model is based.
For example, from a study dealing with the application of bankruptcy models at 16 Czech
agricultural companies implies that the best models for predictions of bankruptcy of agricultural
companies is the S-Index (Springate), Tamari index and the Czech model IN95
[6]. From another similar study implies that the most successful models were the ZETA
model and the Z-Score model. The model created by Taffler and Tisshaw was identified as
the least suitable model [5].
243Acta všfs, 3/2011, vol. 5
The main reasons for selecting the agricultural sector for the analysis of explanatory power
of selected models is the large number of bankruptcies in this sector and also the availability
of financial data.
1	 Bankruptcy Models Used and Chosen Companies
For the evaluation of bankruptcy models, only companies were used, which are legal entities.
The reason was that these companies are registered in the commercial register and
thus have an obligation to disclose their financial statements in the collection of documents
in the Register of Documents.
1.1	 Analyzed Companies
Due to the fact that when analyzing the success of bankruptcy models, the verification is
performed on a group of prosperous and bankrupt companies, it was necessary to define
these two groups.
Businesses that have been included in the sample companies which have gone to bankruptcy
were selected from clusters defined in the database MAGNUS [9] follows:
company is in liquidation,•	
company is in bankruptcy,•	
company is in preliminary reports,•	
business is in alignment,•	
company is in receivership.•	
The basic assumption for assessing the success of bankruptcy models is to obtain the
necessary financial statements. Although Czech companies have incorporated the obligation
to publish their financial statements in a collection of documents, this requirement
is not met by all companies. This obligation is not fulfilled in particular companies that
are in some form of insolvency (bankruptcy, liquidation, etc.). For the analysis, based on
a larger number of financial statements, it is also desirable to have access to some paid
databases.
Despite the above complications, I managed to get sufficient financial statements for
presentation of the analysis of bankruptcy models. The main source used for subsequent
analysis of the financial statements was:
Database MAGNUS [9],•	
Database Internet Securities [10]•	
Information Server of the Czech judiciary [11].•	
The original set of financial statements consisted of approximately 1200 historical financial
statements. These are companies that are currently in decline and companies which
have been assessed as prosperous. For the analysis of companies in bankruptcy financial
statements were used for the period up to five years before the crisis. This choice, however,
I had to adjust and reduce due to incomplete data. It was mostly financial statements de-
Acta všfs, 3/2011, vol. 5244
rived from the database MAGNUS [9], where there was incomplete information for some
companies which prevented their use for the calculation of bankruptcy models. Where
missing data could not be completed from the database Internet Securities [10], these
statements were not used for analysis.
Furthermore, the original choice had to be excluded from the financial statements, which
did not provide the necessary data for analysis. Reasons needed for exclusion were as
follows:
shortened form of financial statements - the company publishes financial data only•	
in abbreviated form, which is insufficient for the use of models based on much more
detailed data (for example, the requirement for the amount of interest expense, shortterm
liabilities, etc.),
the absence of some important items of financial statements - for example, sales of•	
own products and services. This lack is probably due to an error in data processing,
obviously incorrectly classified reports - some reports have been for two years in•	
a row exactly the same, all the items of financial statements have been completed
with the same number, etc.
Financial statements of problematic companies, which were used for subsequent analysis,
the complete financial statements (full version balance sheet and profit and loss
statement for years 2004 - 2008) over the five years prior to the time when it slipped
into a crisis situation that is listed above. Among the analyzed financial statements were
not included financial statements from 2008, which represents the deep period of the
financial crisis. In this period there was a dramatic change in the health of companies,
and so their comparability with data from previous years is limited. During the analysis it
was found that the success of the bankruptcy models, using the group’s financial statements
for the years 2008 - 2009, was significantly higher, which reduces their overall
comparability.
To select a group of prosperous companies, it was decided to use companies that have
been identified as prosperous by the ranking of“The EVA Ranking Czech Republic”and by
the sector ranking CZECH Sector Award [12]. The reason for choosing this form of prosperous
companies is that the source of data is the paid database Magnus [9] that provides
information about individual companies in the form of financial statements (balance sheet
and profit and loss statement). From this database, it is possible to obtain information only
with certainty about which company is insolvent (using specifications such as bankruptcy,
receivership etc.). For companies that do not have these characteristics, it is impossible,
without a deeper analysis of financial data, to state the smooth development of a prosperous
enterprise and therefore there is no possibility of insolvency. Further selection
was made based on criteria specified in the database Securities [10] (companies with the
highest turnover, regular positive economic results, etc.).
Because of the potential access to the two paid databases of financial data of companies,
I managed to get a sufficient number of financial statements (observation), which is listed
in Table 1.
245Acta všfs, 3/2011, vol. 5
Table 1: Number of observations.
  ZETA
Z” –
Score
Z –
Score
Taffler
a Tisshaw
IN99 IN01 IN05
Bankrupt
companies
502 502 501 499 499 96 96
Prosperous
companies
472 472 472 472 470 400 400
Source: own elaboration.
Whereas that the selection of prosperous companies is more likely that some companies
were wrongly included in the prosperous enterprise analysis group as they were selected
for analysis in large quantities than companies in bankruptcy.
This gives in total 974 observations (from approximately 300 companies). The difference
between the number of observations in the ZETA model and the IN05 model (bankrupt
companies) is given a zero value of certain financial statement items entering into the
financial ratios (e.g. interest income) and the IN05 model can not be compiled.
1.2	 Bankruptcy Models
For this analysis the following models were selected:
Model ZETA, created by E.I. Altman in 1977 [1],•	
Z = 0,717 * A + 0,847 * B + 3,107 * C + 0,420 * D + 0,998 * E
WHERE
A = Working Capital/Total Assets
B = Retained Earnings/Total Assets
C = Earnings before Interest and Taxes/Total Assets
D = Market Value of Equity/Book Value of Total Debt
E = Sales/Total Assets,
Model Z’’- Score, created by E.I. Altman in 1999 [4],•	
Z’’ = 6,56 * A + 3,26 * B + 6,72 * C + 1,05 * D,
WHERE
A = Working Capital/Total Assets
B = Retained Earnings/Total Assets
C = Earnings before Interest and Taxes/Total Assets
D = Market Value of Equity/Book Value of Total Debt
Model Z - Score, created by E.I. Altman in 1968 [1],•	
Z = 1,2 * A+ 1,4 * B + 3,3 * C + 0,6 * D + 1,0 * E
WHERE
A = Working Capital/Total Assets
B = Retained Earnings/Total Assets
C = Earnings before Interest and Taxes/Total Assets
D = Carrying Value of Equity/Book Value of Total Debt
E = Sales/Total Assets,
Acta všfs, 3/2011, vol. 5246
	Model, created by Taffler and Tisshaw in 1977 [8],•	
Z = 0,53 * EBT / KZ + 0,13 * OA / CK + 0,18 * KZ / A + 0,16 * T / A
WHERE
A = Earnings Before Taxes/ Short-Term Liabilities
B = Current Assets/ Amount Of Liability;
C = Short-Term Liability/ Assets;
D = Sales/ Assets.
	Index IN99, created by Inka and Ivan Neumaier in 1999 [6],•	
IN99 = - 0,017 * A + 4,573 *B + 0,481 *C + 0,015 *D
WHERE
A = Assets / Amount Of Liability / Short-Term Liabilities
B = Earnings Before Interest and Taxes / Assets;
C = Sales / Assets;
E = Current Assets /( Short-Term Liabilities + Short Term Loans),
	Index IN01, created by Inka and Ivan Neumaier in 2001 [6],•	
IN01 = 0,13 * A + 0,04 * B + 3,92 * C + 0,21 * D + 0,09 * E
WHERE
A = Assets / Amount Of Liability / Short-Term Liabilities
B = Earnings Before Interest and Taxes / Interest;
C = Earnings Before Interest and Taxes / Assets;
D = Sales/ Assets,
E = Current Assets /( Short-Term Liabilities + Short Term Loans),
	Index IN05, created by Inka and Ivan Neumaier in 2005 [6].•	
IN05 = 0,13 * A + 0,04 * B + 3,97 * C + 0,21 * D + 0,09 * E
WHERE
A = Assets / Amount Of Liability / Short-Term Liabilities
B = Earnings Before Interest and Taxes / Interest;
C = Earnings Before Interest and Taxes / Assets;
D = Sales/ Assets,
E = Current Assets /( Short-Term Liabilities + Short Term Loans),
The reason for choosing the above models for subsequent analysis was mainly because
the models developed by Altman and also the model, which was created by Taffler and
Tisshaw in England [8] are among the world’s most widely used models. The above-mentioned
foreign bankruptcy models were also the basis for the development of a number
of new bankruptcy models.
IN Indexes [6] are among models that are most used in the Czech Republic and therefore
these have been included in the selection to be able to analyze the comparability of the
explanatory power of models developed abroad, thus based on completely different economic
conditions, with models based on the domestic economy and therefore should
produce results that better reflect the actual state of the domestic economy.
247Acta všfs, 3/2011, vol. 5
As mentioned above, the evaluation of the use of bankruptcy models for domestic companies
is also based on the assumption that the Czech bankruptcy models have to be
applied on domestic companies with greater success and greater explicitness than the
foreign bankruptcy models. From this reason, I chose both foreign and domestic bankruptcy
models.
2	 Evaluation at Bankrupt Companies
As shown in the Table 2 the most successful bankruptcy model was ZETA with 95% successful
predictions of bankruptcy. As the second most successful foreign model was the
Z-Score model which correctly classified 84% of companies in bankruptcy. From Czech
models the IN99 model showed the best results. The bankruptcy model reached 90%
correctly classified companies. Two other Czech models IN01 and IN05 achieved similar
results (71% respectively 74%).
The least accurate model was created by Taffler and Tisshaw, which reached only 54% success
rate. This corresponds to the results which receive Kopta in his study [5].
Table 2: Evaluation of bankrupt companies.
Model Bankruptcy Grey Zone Prosperity
ZETA 94.6% 4.0% 1.4%
IN99 90.2% 6.2% 3.6%
Z - Score 83.8% 5.6% 10.6%
IN05 74.0% 8.3% 17.7%
IN01 70.8% 13.5% 15.6%
Z’’ - Score 60.4% 21.7% 17.9%
Taffler a Tisshaw 54.1% 14.4% 31.5%
Source: own elaboration.
From the Table 2 is clear that the ZETA model incorrectly ranked only 1.4% of companies
in bankruptcy as prosperous companies. The second most successful model was the
IN99 model which incorrectly ranked only 3.6% bankruptcy companies as prosperous
companies.
Other Czech models showed more inaccuracies and as prosperous companies ranked
17.7% (IN05) and 15.6% (IN01) companies in bankruptcy. The worst results were shown
by the Z”-Score model (17.9%) and the model created by Taffler and Tisshaw, which ranked
among the prosperous companies even 31.5% of companies in bankruptcy.
From the Table 2 it is clear that the ZETA model ranked poorly in the gray area with only
4% of the analyzed sample of companies. The second most successful model was the ZScore
model which ranked, in the gray zone, only 5.6% of the companies threatened by
bankruptcy. Czech models showed more inaccuracies when included in the gray zone with
6.2% companies (IN99), 8.3% (IN05) and 13.5% (IN01). The worst results were shown by the
Acta všfs, 3/2011, vol. 5248
Z’’- Score model and the model created by Taffler and Tisshaw which included in the gray
zone 14.4% (Z’’- Score) and 21.7% of companies in bankruptcy (Taffler and Tisshaw).
3	 Evaluation of Prosperous Companies
Table 3 shows the percentage success of selected bankruptcy models in prosperous agricultural
companies.
As shown in the table below, in the prediction of prosperity, model created by Taffler and
Tisshaw was the most successful model with 98% successfully classified companies. The
second model, which achieved above average results was the Z’’- Score model. This model
correctly classified 78% prosperous companies. From Czech bankruptcy models the highest
result was achieved by the IN05 model which correctly classified 37% of prosperous
companies.
Table 3: Evaluation of prosperous companies.
Agriculture Prosperity Grey Zone Bankruptcy
Taffler a Tisshaw 98.1% 1.1% 0.8%
Z’’ - Score 78.4% 19.7% 1.9%
Z - Score 40.7% 50.8% 8.5%
IN05 36.8% 56.0% 7.3%
IN01 30.3% 67.8% 2.0%
ZETA 4.4% 64.6% 30.9%
IN99 2.8% 31.9% 65.3%
Source: own elaboration.
From Table 3 it is quite clear that the model created by Taffler and Tisshaw poorly ranked
among companies threatened bankruptcy, only 0.8% of the analyzed sample of prosperous
companies. The second most successful model was the Z’’- Score model which
ranked only 1.9% companies threatened bankruptcy. Czech models showed inaccuracies,
since most companies classified in the so-called the gray zone (IN05 56.0% and
IN01 67.8%). Among the companies threatened by bankruptcy the N05 model classified
7.3% of the analyzed sample of prosperous companies and the IN05 model only 2%
of prosperous companies. The exception is the IN99 model which classified the group
of companies threatened by bankruptcy, 65.3% of prosperous companies. This model
was definitely the worst of the selected bankruptcy models. The most successful model
for the identification of prosperous companies was the model created by Taffler and
Tisshaw.
249Acta všfs, 3/2011, vol. 5
4	 Evaluation of the Success of Bankruptcy Models
The following chart shows the order bankruptcy models based on success in bankruptcy
prediction of bankrupt agricultural companies.
	ZETA	 94.6%1.	
	IN99	 90.2%2.	
	Z - Score	 83.8%3.	
	IN05	 74.0%4.	
	IN01	 70.8%5.	
	Z’’ - Score	 60.4%6.	
	Taffler a Tisshaw	 54.1%7.	
The above order shows only the ability of bankruptcy models to correctly identify the
bankrupt agricultural companies. To assess the appropriateness of the use of models for
prediction of impending bankruptcy of companies, it is also necessary to determine its
success in the identification of prosperous companies. Although models are mainly focused
on bankruptcy, it is appropriate to take into account their ability to recognize prosperous
companies and thus ensure that their success in identifying bankrupt companies
is not only due to their setting, when most companies are automatically placed between
companies threatened with bankruptcy. Given that some models are bankruptcy models,
whose primary function is the ability to identify a company threatened by bankruptcy,
take into account in evaluating the more the ability to identify bankrupt companies than
the ability to identify prosperous companies. This evaluation is shown in the Table 4.
Table 4: Evaluation of prosperous companies.
Used model Correctly
classified
bankrupt
companies
Points
(2.5 times
achieved)
Correctly
classified
prosperous
companies
Points
(achieved
%)
Total
Points
Z - Score 83.8% 209.6 40.7% 40.7 250.3
ZETA 94.6% 236.6 4.4% 4.4 241.0
Taffler a Tisshaw 54.1% 135.3 98.1% 98.1 233.4
Z’’ - Score 60.4% 150.9 78.4% 78.4 229.3
IN99 90.2% 225.5 2.8% 2.8 228.2
IN05 74.0% 184.9 36.8% 36.8 221.6
IN01 70.8% 177.1 30.3% 30.3 207.3
Source: own elaboration.
Bankruptcy models Z-Score, ZETA and IN99 achieved very good results in the identification
of bankrupt companies. All three models were able to correctly identify more than
80% of bankrupt companies. The Z-Score model could correctly classify 83.8% of bankrupt
companies and 40.7% of prosperous companies.
Acta všfs, 3/2011, vol. 5250
In the case of the ZETA and IN99 model the success was achieved even above 90% (ZETA
94.6% and IN99 90.2%). However, these models correctly included only 4.4% respectively
2.8% of prosperous companies. This result reflects the facts that the models are quite
strict and most companies (whether prosperous or bankrupt) are immediately included
in bankrupt companies. Thereby their informative value slightly declines.
The model developed by Taffler and Tisshaw correctly assessed 98.1% of prosperous
companies and achieves the best results in area of analyzing prosperous companies. In
contrast, analysis of bankrupt companies, it achieved the worst result when it correctly
classified only 54.1% of companies in bankruptcy.
The Czech bankruptcy models achieved very good results in identifying companies
threatened by bankruptcy, but on the other hand, did not great success in the analysis of
prosperous companies. This result is due to their strictness, where most companies are
included in companies threatened by bankruptcy.
For overall evaluation of the success of bankruptcy prediction on bankrupt agricultural
companies, it is necessary to take into account the erroneous inclusion of bankrupt companies
in prosperous companies.
The following Table 5 takes into account the erroneous inclusion of bankrupt companies
in prosperous companies and the final order of bankruptcy models.
Table 5: Evaluation of bankruptcy models.
Used model
Number
of points
obtained
Incorrect classification
of bankruptcy.
companies
Adjustment
points
Total
Points
Final
Ranking
Z - Score 250.3 10.6% -10.6 239.7 1
ZETA 241.0 1.4% -1.4 239.6 2
IN99 228.2 3.6% -3.6 224.6 3
Z´´ - Score 229.3 17.9% -17.9 211.4 4
IN05 221.6 17.7% -17.7 203.9 5
Taffler a Tisshaw 233.4 31.5% -31.5 201.9 6
IN01 207.3 15.6% -15.6 191.7 7
Source: own elaboration.
For bankruptcy models, it is necessary to eliminate the possibility of incorrect classification
of companies threatened by bankruptcy, in prosperous companies. As is evident from
the above table, most companies threatened by bankruptcy, were included in prosperous
companies by the model created Taffler and Tisshaw (31.5%). This percentage is very
high, so this model fell in the final ranking to sixth place and therefore is inappropriate for
identification bankruptcy in agriculture. The most successful models are Z-Score, ZETA and
the Czech model IN99. Before considering erroneously inclusion of bankrupt companies
251Acta všfs, 3/2011, vol. 5
in prosperous companies it was obvious that the least suitable model for bankruptcy
prediction is the Czech IN01 model.
Conclusions
In the previous section, the success of models in predicting bankruptcy of agricultural
companies was described. Below is the final order of the success of the bankruptcy models
which based on a correct identification of bankrupt companies and correct identification
of prosperous companies. The final evaluation of the usefulness of models for the bankruptcy
identification is as follows:
	Z - Score1.	
	ZETA2.	
	IN993.	
	Z’’ - Score4.	
IN055.	
	Taffler a Tisshaw6.	
	IN017.	
Although the model developed by Taffler and Tisshaw achieves the best results in prosperous
companies, in analysis of bankrupt companies achieved the worst result when it
correctly classified only 54.1% of bankrupt companies, so therefore it is inappropriate for
identification bankruptcy in agriculture.
The most successful models are Z-Score, ZETA and the Czech model IN99. The Czech model
IN01 was identified as the least suitable model.
References
ALTMAN, E.I.1.	 (2006). Corporate Financial Distress and Bankruptcy. Hobokem, John Wiley
& Sons, New Persey, 2006. ISBN 978-0-471-69189-1.
BLAHA, Z.; JINDŘICHOVSKÁ, I.2.	 (1996). Jak posoudit finanční zdraví firmy. Management
Press, 1996. ISBN 80-85603-80-2.
BEAWER, William H.3.	 (1966). Financial ratios as prediction of failure. Journal of Accounting
Research. Vol. 4, 1966, 71-III.
HEINE, M.; ALTMAN, E. I.4.	 (2000). Predicting financial mistress of companies: Revisiting
the Z-Score and Zeta models. NY : New York University, 2000.
KOPTA, D.5.	 (2006). Využití ukazatelů finančního zdraví při hodnocení zemědělských podniků.
Hradec Králové : Hradecké ekonomické dny, 2006.
NEUMAIEROVÁ, I.; NEUMAIER, I.6.	 (2002). Výkonnost a tržní hodnota firmy. Grada Publishing,
2002. ISBN 80-247-0125-1
ŘEZBOVÁ, H.7.	 (2001). Kritéria hodnocení podnikatelských subjektů v zemědělství. Praha:
ČZU, 2001. Disertační práce.
TAFFLER, R. J.8.	 (1984). The audit going-concern in practice. Accounting Magazine. No.
88/1984. ISBN 105-733-021.
Database Magnus.9.	 [on-line]. [Access: 2011-03-07]. Available at: http://www.magnus.cz.
Database Internet Securities Inc.10.	 [on-line]. [Access: 2011-03-07]. Available at: http://
www.securites.com.
Acta všfs, 3/2011, vol. 5252
Information Server of the Czech judiciary11.	 . [on-line]. [Access: 2011-03-07]. Available
at: http://www.justice.cz.
Press release ČEKIA, Inc.12.	 [on-line]. [Access: 2009-02-14]. Available at: http://www.
cekia.cz/tiskove-zpravy.php?rok=2009#zprava-70.
Contact address / Kontaktní adresa
Ing. Jan Sušický
Czech University of Life Sciences Prague, Faculty of Economics and Management / Česká
zemědělská univerzita, Provozně ekonomická fakulta
(jan.susicky@cnb.cz)