Business Environment and Entrepreneurial Activity

Description
Business Environment and Entrepreneurial Activity

BUSINLSS LNVIRONMLN1 AND
LN1RLPRLNLURIAL AC1IVI1\
by
Dashke·ich Volha
A thesis submitted in partial íulíillment oí
the requirements íor the degree oí
Master oí Arts in Lconomics
National Uni·ersity Kyi·-Mohyla Academy`
Lconomics Lducation and Research Consortium
Master`s Program in Lconomics
2005
Appro·ed by ___________________________________________________
Ms.S·itlana Budago·ska ,lead oí the State Lxamination Committee,
__________________________________________________
__________________________________________________
__________________________________________________
Program Authorized
to Oííer Degree Master`s Program in Lconomics, NaUKMA
Date __________________________________________________________
National Uni·ersity Kyi·-Mohyla Academy`
Abstract
BUSINLSS LNVIRONMLN1 AND
LN1RLPRLNLURIAL AC1IVI1\
by Dashke·ich Volha
lead oí the State Lxamination Committee: Ms.S·itlana Budago·ska,
Lconomist, \orld Bank oí Ukraine
Lntrepreneurial acti·ity is one oí the main sources íor economic growth,
inno·ation and job creation. 1his thesis pro·ides the analysis oí the iníluence oí
go·ernment regulations on the entrepreneurial acti·ity across countries around
the world. 1he entrepreneurial acti·ity is deíined as the number oí small and
medium-sized enterprises across 31 countries and as the share oí pri·ate sector in
GDP across 24 transition countries. 1he impact oí go·ernment regulations on
the business acti·ity is e·aluated using indicators concerning the business
registration process and the contract eníorcement procedure. 1his work also
in·estigates the entrepreneurial acti·ity in the iníormal sector oí the economy.
1he results oí the study indicate that go·ernment regulations negati·ely iníluence
on the de·elopment oí small and medium-sized business. Stricter regulation oí
entry is led to higher le·el oí corruption and a greater size oí the iníormal
economy.

1ABLL Ol CON1LN1S
List oí 1ables.................................................................................................................. ii
List oí Appendixes.............................................................................. iii
Acknowledgements.......................................................................................................i·
ClAP1LR 1. Introduction......................1
ClAP1LR 2. Literature re·iew..................3
ClAP1LR 3.Methodology......................9
ClAP1LR 4. Data description....................16
ClAP1LR 5.Lmpirical analysis and results.............21
ClAP1LR 6. Conclusions.....................28
Bibliography.........................31
Appendix.......................................................................................................................33

ii
LIS1 Ol 1ABLLS
^vvber Page

1. Speciíication 1ests íor the Model with the Dependent Variable as the number
oí SMLs.............................22
2. Speciíication 1ests íor the Model with the Logarithmic Dependent Variable as
the number oí SMLs.......................23
3. Speciíication 1ests íor the Model with the Dependent Variable as the number
oí SMLs per capita.......................24
4. Speciíication 1ests íor the Model with the Dependent Variable as the share oí
pri·ate sector in GDP ,°,.....................25

iii
LIS1 Ol APPLNDIXLS
^vvber Page

1. Assumptions about the Business and Contract Lníorcement......33
2. Procedure oí business registration in Belarus in 2003..........35
3 Descripti·e Statistics oí Starting Business Indicators............3¯
4. Descripti·e Statistics oí Lníorcing Contracts.............39
5. Descripti·e Statistics oí SMLs, the Pri·ate Sector and the Iníormal
Lconomy............................41
6. Lmpirical Lstimation oí the Model with the Dependent Variable as the
Number oí Lnterprises......................42
¯. Lmpirical Lstimation oí the Model with the Logarithmic Dependent Variable
as the Number oí Lnterprises......................44
8. Lmpirical Lstimation oí the Model with the Dependent Variable as the
Number oí Lnterprises Per Capita.................45
9. Lmpirical Lstimation oí the Model with the Dependent Variable as the Share
oí Pri·ate Sector in GDP ,°,...................4¯
10. Lmpirical Lstimation oí the Model with the Dependent Variable as the Share
oí the Iníormal Lconomy in GNI ,°,..................48

i·
ACKNO\LLDGMLN1S
I would like to express the sincere gratitude and appreciation to my ad·isor Dr.
Iryna Lykyanenko íor her continuous encouragement and insightíul super·ision
through the whole process oí the thesis writing. I am also wants to thank the
research workshop proíessors 1om Coupé, Polina Vlasenko and Volodymyr
Bilotkach íor their timely help and ·aluable comments.

C b a ¡ t e r 1
IN1RODUC1ION
1he condition oí the business en·ironment plays an important role in shaping the
nature oí entrepreneurial acti·ity and the dynamics oí new enterprises. 1he
burden oí business regulation ·aries across countries in the world. lea·y
business regulation promotes high le·els oí corruption and decreases the
incenti·es to be an entrepreneur.
1he main idea oí this work is the analysis oí the business en·ironment in the
world, in particular, the analysis oí íactors which determine the business
en·ironment, and how business en·ironment ía·ours the creation oí new íirms
and the de·elopment oí business.
I deíine the business en·ironment as the number oí íactors, which are conduci·e
to business de·elopment. 1he main íactors that determine the business
en·ironment are go·ernment and bureaucratic regulation, capital constraints,
in·estments, corruption, law, taxation and property rights.
In this paper I will use the íollowing methodology. I would like to concentrate
my attention on the impact oí go·ernment regulations on the business
de·elopment across countries around the world. I will deíine the entrepreneurial
acti·ity as the number oí small and medium-sized enterprises ,SMLs, across 31
countries, as the number oí small and medium-sized enterprises per capita across
31 countries and as the share oí pri·ate sector in GDP across 24 transition
countries. Also I would like to analyze the interaction between business
de·elopment and the iníormal sector oí the economy across 106 countries

2
around the world. In the empirical part oí my research I will e·aluate the impact
oí go·ernment regulation on the business acti·ity using the íollowing
determinants oí business en·ironment: costs oí obtaining legal status to operate a
íirm as a share oí per capita GNI, time that it takes to obtain legal status to
operate a íirm, in business days, number oí procedures that a íirm has to
complete with in order to obtain a legal status, the minimum capital required as a
percentage oí income per capita, indicators oí eníorcing contracts ,a number oí
procedures, time in calendar days and oííicial costs as a percentage oí the debt
·alue,, GDP per capita.
I am going to use the panel data íor 2002 and 2003. Data sources are Doing
Business Databases 2004,2005 oí the \orld Bank, 1ransition report 2004 oí
Luropean Bank íor Reconstruction and De·elopment and Statistical Agencies oí
diííerent countries.
Lntrepreneurial acti·ity and regulation oí entrepreneurial de·elopment ha·e
recei·ed considerable attention in the economic literature. 1he entrepreneurial
acti·ity is one oí the main sources íor economic growth, inno·ation and
creation oí jobs. Go·ernment regulations aííect the de·elopment oí small and
medium-sized business. Stricter regulation oí entry is led to higher le·els oí
corruption and a greater relati·e size oí the unoííicial economy.
My contribution to existing research in this íield will be the analysis oí business
en·ironment and íirm creation using the latest data and using the diííerent
deíinitions oí business en·ironment and entrepreneurial acti·ity.
1he structure oí the paper is as íollows. In Chapter 2, I analyze rele·ant literature
íor my research. Chapter 3 describes the methodology that I will use in íurther.
1he data, main results and their interpretation are presented in Chapter 4 and
Chapter 5, and Chapter 6 includes conclusions and discussion.

3
C b a ¡ t e r 2
LI1LRA1URL RLVIL\
1here is a broad literature concerning entrepreneurship, entrepreneurial acti·ity
and business regulations. I would like to structure all rele·ant papers to my topic
in the íollowing way: entrepreneurial de·elopment and economic growth,
deíinition and determinants oí entrepreneurial acti·ity in the literature, costs and
beneíits oí regulation, entrepreneurial acti·ity and go·ernment regulations.
Analyzing the interaction oí entrepreneurial acti·ity and economic de·elopment,
I would like to concentrate on transition countries. 1he entrepreneurial acti·ity is
·ery important íor transition because it is a good source íor economic growth
and íor creation oí new products and jobs. McMillan and \oodruíí ,2002,,
analyzing such transition countries as Russia, China, Poland and Vietnam,
coníirm that enhancement oí entrepreneurial de·elopment is responsible íor the
stable economic growth in Poland and China, while in Russia the slow
de·elopment oí entrepreneurial acti·ity leads to stagnation. In related paper,
Berkowitz and DeJong ,2001, ·eriíy relationship between entrepreneurial acti·ity
and economic growth in post-so·iet Russia. 1hey measure entrepreneurial acti·ity
as the number oí legally registered small pri·ate íirm and they take the a·erage
annual growth in real per capita income obser·ed between 1993-199¯ and
between 1993-2000. 1heir empirical results show that an additional 1.¯ legally
registered enterprises per 1000 inhabitants are associated with an increase in real
economic growth oí 2.5° annually o·er the period 1993-199¯, and 1.5° annually
between 1993-2000 ,Berkowitz and DeJong, 2001:3,.

4
In the economics literature entrepreneurial acti·ity has a range oí deíinitions.
Some papers examine entry and exit oí íirms, others concentrate on the dynamics
oí íirm growth. In such empirical works as Klapper, Lae·en, and Rajan ,2004,
and Djanko· et al. ,2001,, entrepreneurial acti·ity is measured as íirm`s entry.
Desai, Gompers, and Lerner ,2003, determine it as the entry and exit rate, the
a·erage íirm size, the industrial ·intage ,a weighed-a·erage measure oí íirm age,,
and the skewness oí the íirm size distribution. O·aska and Sobel ,2004, deíine
entrepreneurial acti·ity as the number oí new íirms and the number oí new
patent and trademark applications. 1he last ·ariable is associated with economic
growth, wealth and with the high-tech entrepreneurial inno·ation that is
generated by large íirms. Kaya and Uçdogruk ,2002, concentrate on rates oí
íirm`s entry and exit.
1he substantial part oí researches has íocused on íirm entry. 1he work done by
Geroski ,1995, is a summery oí all present iníormation about entry. 1he author
points out some stylized íacts about entry using data on entry and case studies.
Lntry is common. Large numbers oí íirms enter most markets in most years,
but entry rates are íar higher than market penetration rates. Although there is a
·ery large cross-section ·ariation in entry, diííerences in entry between industries
do not persist íor ·ery long. In íact, most oí the total ·ariation in entry across
industries and o·er time is within` industry ·ariation rather than between`
industry ·ariation. Lntry and exit rates are highly positi·ely correlated. 1he
sur·i·al rate oí most entrants is low, and e·en successíul entrants may take more
than a decade to achie·e a size comparable to the a·erage incumbent. De no·o
entry is more common but less successíul than entry by di·ersiíication. Lntry
rates ·ary o·er time, coming in wa·es which oíten peak early in the liíe oí many
markets. Diííerent wa·es tend to contain diííerent types oí entrant. Costs oí
adjustment seem to penalize large-scale initial entry and ·ery rapid post-entry
penetration rates` ,Geroski, 1995,.

5
Using empirical literature on entry Geroski ,1995, recei·es eight stylized empirical
results about entry. Lntry seems to be slow to react to high proíits. Lconometric
estimates oí the height oí entry barriers suggest that they are high. On the other
hand, entry rates are hard to explain using con·entional measures oí proíitability
and entry barriers. Lntry seems to ha·e only modest eííects on a·erage industry
price-cost margins. ligh rates oí entry are oíten associated with high rates oí
inno·ation and increases in eííiciency. 1he response by incumbents to entry is
selecti·e. Prices are not usually used by incumbents to block entry. Both íirm size
and age are correlated with the sur·i·al and growth oí entrants ,Geroski, 1995,.
\hat all oí this adds up to is a presumption that entry is generally a poor
substitute íor acti·e ri·alry among incumbent íirms in a market. Lntry can be an
important iníluence on the e·olution oí industry structure and períormance, also,
it is so only selecti·ely. lurther, entry seems to play an important role in
stimulating industry e·olution at precisely those times when the current acti·ities
oí incumbent íirms are most out oí line with exogenous changes in costs and
demand. In short, not only is entry an imperíect mechanism íor getting prices
right in markets, it is a mechanism íor getting product and process speciíications
right ,Geroski, 1995:43¯,.
In analyzing go·ernment regulation oí business, it is ·ery important to
understand who is better oíí and worse oíí, consumers ,the Public Interest
theory, or go·ernment and incumbents ,the Public Choice theory,. Djanko· et al.
,2001, in their research íind out that stricter regulation oí entry is not associated
with higher quality products, better regulation oí en·ironment`s pollution, or
competition. Stricter regulation oí entry is associated with sharply higher le·els oí
corruption, and a greater relati·e size oí the unoííicial economy. 1hereíore, the
regulation oí entry ser·es the Public Choice than the Public Interest ,Djanko· et

6
al., 2001 and lisman and Sarria-Allende, 2004,. Lntry is regulated, because it is
proíitable íor the regulators.
Beíore examination oí the impact oí go·ernment regulations on entrepreneurial
de·elopment, it is important to ·iew the classiíication oí barriers to market entry.
Robinson and lairchild ,2002, classiíy barriers as institutional and social. 1hey
deíine institutional barriers as íormal, cultural and legitimacy. Go·ernment, laws,
íinancial markets and lending institutions characterize íormal institutional
barriers. Such íormal barriers can impede entry to a market ií the market does not
ha·e the appropriate institutions to enhance entrepreneurial acti·ity. Cultural
entry barriers are the íollowing: language, slang, dress and etiquette. Legitimacy is
a generalized perception or assumption that the actions oí an entity are desirable,
proper, or appropriate within some socially constructed system oí norms, ·alues,
belieís and deíinitions ,Robinson and lairchild, 2002:11,.
Djanko· et al. ,2001,, using a data set oí 85 countries, create the íollowing
indicators íor their analysis oí the impact oí go·ernment regulation oí entry: the
number oí procedures íor starting business, the oííicial time íor completing the
process, and its oííicial cost. Klapper, Lae·en, and Rajan ,2004, and Desai,
Gompers, and Lerner ,2003, in their empirical papers use these ·ariables íor
measuring oí start-up business procedures. 1he \orld Bank uses the same
methodology íor constructing its Doing Business Databases.
Go·ernment regulations aííect entrepreneurial acti·ity. In related empirical
works, Desai, Gompers, and Lerner ,2003,, lisman and Sarria-Allende ,2004, and
Klapper, Lae·en, and Rajan ,2004,, íind out that regulations ha·e a negati·e
impact on íirm entry. A study done by Klapper, Lae·en, and Rajan ,2004,
in·estigates the interaction oí business en·ironment and íirm entry on industry-
and country- le·el. 1hey use a database oí íirms across a number oí de·eloped
and transition countries in Lurope. 1hey íind that entry regulations hamper entry,

¯
especially in industries that naturally should ha·e high entry ,Klapper, Lae·en,
and Rajan, 2004,. In addition, they íind out that those regulations, which are
ía·orable íor de·eloping intellectual property rights and the íinancial sector
enhance íirm creation. As emphasized by lisman and Sarria-Allende ,2004, low
entry regulation oí industries has little impact on the quantity and a·erage size oí
íirms in this industry comparing with high entry regulation. Countries with low
entry regulation ha·e íirm`s growth by creation oí new ones, while countries with
high regulation oí entry are associated with íirm`s growth by the expansion oí
existing íirms.
Desai, Gompers, and Lerner ,2003, analyze the impact oí institutional
en·ironment ,index oí íairness, protection oí property right, íormalism index,
index oí the interíerence oí courts, and a measure oí start-up procedures, on
entrepreneurial acti·ity. 1hey use a cross-country ,33 Luropean countries,
approach. Greater íairness and stronger property rights leads to higher rates oí
entry and lower rates oí exit, greater judicial interíerence and íormalism lead to
lower entry.
It is necessary to mention the work oí O·aska and Sobel ,2004, that inspects the
rates oí entrepreneurial acti·ity in post-socialist countries. Such íactors as low
go·ernment corruption, credit a·ailability, sound monetary policy, high íoreign
direct in·estment, contract eníorcement, low regulations and taxes are associated
with the higher rates oí entrepreneurial acti·ity. 1hey íind out that diííerent
íactors ha·e less or more iníluence on such measures oí entrepreneurial acti·ity
as creation oí new íirm and the patent and trademark de·elopment. Credit
a·ailability and go·ernment corruption aííect íirm`s creation. Go·ernment
corruption is more harmíul íor the creation oí small íirms than large ones. At the
same time sound monetary policy, go·ernment corruption and credit a·ailability
ha·e less iníluence on the patent and trademark acti·ity. ligh íoreign direct

8
in·estment is not so important íor creation oí íirms, but has a positi·e impact on
the patent and trademark acti·ity.
It also worth to mention paper done by Kaya and Uçdogruk ,2002,, they analyze
the determinants oí entry and exit and their ·ariation among 1urkish
manuíacturing industries. 1hey deíine entry and exit rates as proportions oí
entrants and exitors to total number oí íirms respecti·ely. 1he determinants oí
entry and exit rates are proíit margin, concentration rate, growth rate, labor
producti·ity, wage and producti·ity diííerentials, a·erage wage rate, ad·ertisement
intensity and capital intensity. Lmpirical results oí this work show that proíit
margin, concentration ratio, growth rate and capital intensity are the main
determinants oí íirm entry, while concentration ratio, growth rate and capital
intensity iníluence on exit rate.
I would like to summaries the theoretical background oí the relations between the
business en·ironment and the entrepreneurial acti·ity. Go·ernment regulations
aííect the creation and the de·elopment oí small and medium-sized enterprises.
My contribution to existing research will be the íollowing. I will use the latest data
about entrepreneurial acti·ity and business en·ironment indicators. Djanko· et al.
,2001,, Klapper, Lae·en, and Rajan ,2004, and Desai, Gompers, and Lerner
,2003, deíine go·ernment regulations as indicators íor measuring oí start-up
business procedures. In my research I will expand the deíinition oí go·ernment
regulations by adding the set oí ·ariables about the contract eníorcement.
Lntrepreneurial acti·ity I will deíine as the number oí small and medium-sized
enterprises and as the share oí pri·ate sector in GDP. Also I will analyze the
entrepreneurial acti·ity in the iníormal economy.

9
C b a ¡ t e r ²
ML1lODOLOG\
I will explore the impact oí business en·ironment on an entrepreneurial
acti·ity across countries around the world.
1he dependent ·ariable is an entrepreneurial acti·ity. I determine it as the
number oí small and medium-sized enterprises across 31 countries, as the
number oí small and medium-sized enterprises per capita across 31 countries and
as a share oí pri·ate sector in GDP around 24 transition countries. 1hese datasets
are panel. 1he number oí SMLs is the total number oí pri·ate enterprises
registered in each country during the period oí time 2002 and 2003. 1he scale oí
the country has a great iníluence on the number oí SMLs as the dependent
·ariable. 1o account this aspect I will use the number oí SMLs per capita in each
country. I recei·e this indicator di·iding the total number oí SMLs by the
population.
Lxplanatory ·ariables are íactors oí business en·ironment that iníluence the
entrepreneurial acti·ity. I will deíine the business en·ironment using indicators on
starting a business and eníorcing contracts. Starting a business includes a vvvber of
¡roceavre. required to register a íirm, a·erage tive spent during each procedure,
oííicial co.t oí each procedure as a percentage oí GNI and the vivivvv ca¡itat
required as a percentage oí income per capita.

10
1he number oí procedures reílects the business registration process. It is the
number oí steps that entrepreneur has to íulíil to start a business. Bribes can
speed up this registration stage.
1he number oí days is the time that it takes to obtain legal status to operate a
íirm. 1he higher the oííicial costs and the number oí days associated with each
procedure the less probability that many entrepreneurs will register.
1he minimum capital requirement is the amount oí capital that the entrepreneur
needs to put into a bank account beíore registration process starts.
Lníorcing contracts characterize the eííiciency, íormalism and corruptibility oí
court system in the country. Contract eníorcement is ·ery important íor íirms in
their commercial transactions and íor access to íinance. Countries with higher
costs oí dispute resolution ha·e larger iníormal sectors. Ineííicient judicial system
is an impediment to the con·ersion oí iníormal enterprises into íormal ones
,Ayygari et al, 2003:14,. Lníorcing contracts include a vvvber of ¡roceavre., tive in
calendar days and officiat co.t. as a percentage oí the debt ·alue.
1he number oí procedures is warranted by law or court regulations that claim
interaction between the parties and the judge or a court oííicer. 1he time is the
number oí calendar days needed íor dispute resolution, counted írom the
moment the plaintiíí íiles the lawsuit in court until the moment oí settlement or,
when appropriate, payment ,this measure includes the days when actions take
place and the waiting periods between actions,. 1he oííicial costs are the costs oí
going through court procedures, including court costs and attorney íees ,Doing
Business 2005, \orld Bank,.
I expect to recei·e negati·e iníluence oí business registration and contract
eníorcement indicators on the entrepreneurial acti·ity.

11
1he le·el oí economic de·elopment aííects the attracti·eness oí becoming an
entrepreneur. Greater gross domestic product ,GDP, per capita pro·ides a larger
market potential and greater inírastructure íor start-ups. In my analysis, GDP per
capita is expressed in constant 1995 U.S. dollars. Ií the le·el oí economic
de·elopment plays a role in supporting entrepreneurial acti·ity, then GDP per
capita will ha·e a positi·e eííect on the entrepreneurial acti·ity.
1he econometric models are the íollowing:
í^1íRP i,t ~ ?0 - ?1 PROCíD i,t - ?2 D.Y i,t - ?² CO´1 i,t -
?1 C.Pí1 i,t - ?: íPROCíD i,t - ?ó íD.Y i,t - ?¨ íCO´1 i,t -
?º CDP¡c i,t - v i,t
í^1íRP¡c i,t ~ ?0 - ?1 PROCíD i,t - ?2 D.Y i,t - ?² CO´1 i,t -
?1 C.Pí1 i,t - ?: íPROCíD i,t - ?ó íD.Y i,t - ?¨ íCO´1 i,t -
?º CDP¡c i,t - v i,t
PR´íC1OR i,t ~ ?0 - ?1 PROCíD i,t - ?2 D.Y i,t - ?² CO´1 i,t -
?1 C.Pí1 i,t - ?: íPROCíD i,t - ?ó íD.Y i,t - ?¨ íCO´1 i,t - v i,t
í^1íRP i,t - the number oí small and medium-sized enterprises in country i
during the period t,
í^1íRP¡c i,t · the number oí small and medium-sized enterprises per capita in
country i during the period t,
PR´íC1OR i,t - the share oí the pri·ate sector in GDP in country i during the
period t,

12
PROCíD i,t - the number oí procedures required to register a íirm in country i
during the period t,
D.Y i,t - a·erage time spent during each procedure in country i during the period
t,
CO´1 i,t - the oííicial cost oí each procedure as a percentage oí GNI in country
i during the period t,
C.Pí1 i,t - the minimum capital required as a percentage oí income per capita in
country i during the period t,
íPROCíD i,t - the number oí procedures is warranted by law or court
regulation, that claim interaction between the parties and the judge or a court
oííicer in country i during the period t,
íD.Y i,t - the number oí calendar days needed íor dispute resolution in country
i during the period t,
íCO´1 i,t - the cost is incurred during dispute resolution as a share oí income
per capita in country i during the period t,
CDP¡c i,t - GDP per capita in country i during the period t.
I would like to analyze the interaction oí entrepreneurial acti·ity and the iníormal
sector oí the economy. lard go·ernment regulations oí business acti·ity
stimulate the de·elopment oí the corruption and the iníormal economy. 1he high
le·el oí the iníormal economy means that it is the high le·el oí the iníormal
entrepreneurial acti·ity in the country and weak de·elopment oí the oííicial one.
1hat is why I would like to analyze the iníluence oí business registration and

13
contract eníorcement íactors on the con·ersion oí the oííicial business acti·ity to
the iníormal one.
1he economists measure the iníormal economy as a percentage oí gross national
income. Ldgcomb and 1hetíord ,2004, characterize the iníormal economy in
the íollowing way: it i. tegat, bvt vvregvtatea, enterprises, employers and selí-
employed indi·iduals do not engage in criminal acti·ities, but do not comply
with standard business practice and taxation regulations, ca.b i. tbe vo.t covvov
veaivv of e·cbavge, ror/ covaitiov. for ror/er. are ivferior tbav iv tbe forvat ecovov,,
en·ironment protection, security and earnings are less protected than íor
workers in the íormal economy, both ev¡to,ea ava .etf·ev¡to,ea ror/er. engage in
the iníormal economy ,Ldgcomb and 1hetíord, 2004: 12-13,.
1he data about the iníormal economy as a share oí GNI is a·ailable íor 106
countries around the world íor 2002. I will estimate this dependent ·ariable using
OLS. 1he econometric model is as íollows:
í^ííCO^ i ~ ?0 - ?1 PROCíD i - ?2 D.Y i - ?² CO´1 i - ?1 C.Pí1 i -
?: íPROCíD i - ?ó íD.Y i - ?¨ íCO´1 i - v i
í^ííCO^ i - the iníormal economy as a share oí gross national income in the
country i.
I expect to recei·e the positi·e dependence between the iníormal economy and
starting business and contract eníorcement indicators.
I will check this regression íor the heteroscedasticity using Breusch-Pagan ,
Cook-\eisberg test. 1he null hypothesis states that there is a constant ·ariance.
lor checking this econometric model on the omitted ·ariables I will apply

14
Ramsey RLSL1 omitted ·ariable test. Ií the null hypothesis can not be rejected,
than the model does not ha·e omitted ·ariables.
lor the panel data databases it is important to employ econometric methodology
íor íull in·estigation oí panel data properties. lor the beginning it is necessary to
discriminate among the ordinary least squares and panel data techniques. Ií there
is no diííerence across cross-sectional units, and that the indi·idual eííects can
be ignored, then the ordinary least squares technique will pro·ide consistent and
eííicient estimation oí the parameters oí the model ,Greene, 2000, p.560,. Ií the
indi·idual eííects can not be ignored than the panel data model will be more
appropriate.
l-test helps to test the signiíicance oí the indi·idual eííects. Ií the null
hypotheses oí common intercept can not be rejected, than the OLS estimators
are eííicient, otherwise techniques íor the panel data are chosen.
1o test the appropriateness oí the random eííects I will use Breusch and Pagan
Lagrange multiplier test. 1he null hypothesis states that the indi·idual speciíic
disturbance does not ·ary across cross-sections. Ií the null hypothesis can be
rejected, than the random eííects model will be employed. 1he random eííects
approach establishes that the indi·idual - speciíic eííects randomly ·ary across
cross-sections.
1he íixed eííects approach means that the diííerence across cross-units can be
íixed by diííerences in the constant term.
1he lausman test distinguishes between random and íixed eííects. Under the
null hypothesis there is no correlation and the random eííects gi·es consistent
and eííicient estimators. Ií estimators under both eííects are signiíicantly

15
diííerent, than the null hypothesis is rejected. I will deal with the íixed eííects,
which gi·es consistent estimators.
It is necessary to check econometric regressions on the heteroscedasticity.
Breusch-Pagan , Cook-\eisberg test with the null hypothesis about the constant
·ariance can be applied.
1heoretically there is no problem oí endogeneity oí regressors in econometric
models due to the data construction.
1he soítware that is used íor analysis is Stata 8.0.
la·ing discussed the basics oí the econometric methodology, I will proceed to
the empirical part oí my study.

16
C b a ¡ t e r 1
DA1A DLSCRIP1ION

1he initial dataset consists oí statistics íor 125 countries and co·ers the period
írom 2002 to 2003. Main databases íor the analysis oí business en·ironment and
entrepreneurial acti·ity are Doing Business in 2004 and Doing Business in 2005.
1he Doing Business Database oííers a comprehensi·e sur·ey oí pri·ate sector
regulations in 145 countries. It consists oí such projects as Starting a Business,
liring and liring \orkers, Lníorcing Contracts, Getting Credit, Closing a
Business. 1he data íor all sets oí indicators in these databases are íor January
2003 and January 2004 respecti·ely.
1o make the business comparable across countries, 1he Doing Business
Database employed 10 assumptions. 1hese assumptions are represented in
Appendix 1.
\ith these assumptions the íollowing indicators were constructed: a number oí
procedures that a íirm has to complete to obtain a legal status, costs oí obtaining
legal status as a share oí per capita GNI, time that it takes to obtain legal status to
operate a íirm, in business days, minimum capital required íor starting a business
as a percentage oí income per capita.
lor example, íor starting business in Belarus entrepreneur needs ¯9 days, he,she
has to complete 16 procedures, the oííicial costs oí these procedures are 25,3°
oí per capita GNI and minimum capital required is 44,3° oí per capita GNI or
t1600 ,Doing Business Database, 2005,. It seems like the go·ernment is not

1¯
interested in the de·elopment oí small and medium-sized business. Appendix 2
represents the precise procedure oí business registration in Belarus in 2003.
Analysis oí data on entry regulation around the world leads to the íollowing
conclusions. 1he number oí procedures required to start up a íirm ·aries írom
the low oí 2 in Canada and Australia to the high oí 19 in Chad with the world
a·erage oí around 10. 1he minimum oííicial time íor such a startup ·aries írom
the low oí 2 business days in Australia to the high oí 215 in Congo ,Dem. Rep.,
in 2002 and 203 business days in laiti in 2003, with the world a·erage oí 55
business days in 2002 and 49 in 2003. 1he oííicial cost oí íollowing these
procedures íor a simple íirm ranges írom 0° oí per capita GNI in Denmark to
1268,4° per capita GNI in Sierra Leone in 2003, with the world-wide a·erage oí
93° in 2002 and 81,06° oí annual per capita income in 2003. Minimum capital
requirements ·ary írom the 0° oí per capita GNI in Azerbaijan, Australia,
Bangladesh and Brazil to 5053,9° oí per capita GNI in Syrian Arab Republic in
2003 with world-wide a·erage oí 289,6° in 2002 and 1¯0,54° oí per capita GNI
in 2003.
On a·erage it takes 6 ,¯, procedures, 25 ,31, days, 8° ,10,1°, oí the income per
capita to start a business and needed 44,1° ,61,2°, oí per capita GNI as the
minimum capital requirements in OLCD countries in 2003 ,2002, and 11 ,11,
procedures, 64 ,¯4, days, 212° ,292,8°, oí the income per capita as the oííicial
costs and 213,¯° ,300,4°, oí per capita GNI as capital requirements to do so
in Sub-Saharan Aírica in 2003 ,2002,.
lor an entrepreneur, legal entry is extremely cumbersome, time-consuming, and
expensi·e in most countries in the world. Appendix 3 represents descripti·e
statistics oí starting business indicators.

18
1o make the data about contracts eníorcement comparable across countries,
se·eral assumptions were made. Appendix 1 represents these assumptions.
1he dataset illustrates the diííerences in the eííiciency oí contract eníorcement
across countries. At best the creditor must complete 11 procedures in Australia
and spend 1 day to be get paid in Poland, it will cost 4,2° oí the claim amount in
attorney and court íees in Norway. Otherwise, the creditor needs to complete 58
procedures in Sierra Leone and spent 1459 days to be get paid in Guatemala, it
will cost 256,8° oí debt ·alue in Congo ,Dem. Rep.,. Appendix 4 represents
descripti·e statistics oí eníorcing contracts.
I collected data about the number oí SMLs across countries írom the oííicial
statistical agencies:
• Institute oí Statistics oí Albania,
• National Statistical Ser·ice oí Armenia,
• State Statistical Committee oí Azerbaijan,
• Ministry oí Statistics and Analysis oí the Republic oí Belarus,
• National Institute oí Statistics oí Belgium,
• Agency íor Statistics oí Bosnia and lerzego·ina,
• Central Bureau oí Statistics oí Croatia,
• Czech Statistical Oííice,
• Statistics Denmark,
• Statistics linland,
• National Institute oí Statistics and Lconomic Studies oí lrance,
• State Department íor Statistics oí Georgia,

19
• lederal Statistical Oííice oí Germany,
• Census and Statistics Department oí long Kong, China,
• lungarian Central Statistical Oííice,
• Central Statistics Oííice oí Ireland,
• National Statistical Oííice oí Korea,
• Central Statistical Bureau oí Lat·ia,
• Statistics Lithuania,
• State Statistical Oííice oí Macedonia,
• Statistics Netherlands,
• Statistics New Zealand,
• Statistics Norway,
• National Institute oí Statistics oí Romania,
• Russian State Committee íor Statistics,
• Statistical Oííice oí the Republic oí Slo·enia,
• National Institute oí Statistics oí Spain,
• Statistics Sweden,
• State Statistics Committee oí Ukraine,
• National Statistics oí United Kingdom
• Bureau oí Statistics oí USA.
1he database consists oí the iníormation about the number oí SMLs across 31
countries íor the period oí 2002 and 2003.
1he number oí SMLs ·aries írom the minimum oí ¯622 in 2002 and ¯¯95 in
2003 in Albania to the maximum oí 8 million in 2002 and 8,441 thousand in 2003

20
in Russian lederation with the a·erage oí 1 093 804 in 2002 to 1 142 06¯ in 2003.
1he descripti·e statistics oí SMLs is represented in Appendix 5, 1able 1.
1he data concerning the iníormal economy as a share oí gross national income is
a·ailable íor 106 countries around the world íor the period 2002. 1his data was
taken írom Doing Business in 2004. Switzerland shows the lowest le·el oí the
shadow economy - 8, 8° oí GNI. Georgia has the highest indicator oí the
unoííicial economy. It is 6¯, 3° oí gross national income. 1he world-wide
a·erage is 32, ¯° oí GNI. Descripti·e statistics oí the iníormal economy in GNI
is represented in 1able 3, Appendix 5.
1ransition report 2004 oí Luropean Bank íor Reconstruction and De·elopment
is the source oí the data oí the share oí pri·ate sector in GDP ,°,. 1he data is
collected íor 24 transition countries and a·ailable íor 2002 and 2003. 1he a·erage
share oí pri·ate business in GDP is 64°. 1he minimum indicator oí the share oí
pri·ate business in GDP is in Belarus - 25°. Czech Republic and Slo·ak
Republic show the highest le·el oí business sector - 80° in GDP ,1able 2,
Appendix 5,.
In the process oí the data collection I had some problems. It was ·ery diííicult to
íind out the systematic dataset with the number oí enterprises across 125
countries. 1hat is why I used data írom statistical agencies oí diííerent countries.
In some sites oí statistical agencies Lnglish ·ersion was not a·ailable or data was
old, thereíore I íound iníormation about number oí SMLs only íor 31 countries
among 125. It also was diííicult to collect data across all dependent and
explanatory ·ariables among 125 countries. 1hat is why I ha·e iníormation about
pri·ate sector íor 24 transition countries íor 2002 -2003 and iníormation about
the iníormal economy as the share oí GNI only íor 106 countries íor 2002.

21
C b a ¡ t e r :
LMPIRICAL ANAL\SIS AND RLSUL1S
1his section presents the results oí the empirical estimation oí the iníluence oí
íactors oí business en·ironment on the entrepreneurial acti·ity. 1he
entrepreneurial acti·ity is measured as the number oí enterprises, as the number
oí enterprises per capita and as the share oí pri·ate sector in GDP. Also I would
like to analyze the entrepreneurial acti·ity in the iníormal sector oí the economy.
ívtre¡revevriat actirit, a. tbe vvvber of evter¡ri.e..
I estimate the described earlier model using OLS and panel data techniques. l-
test rejects the hypothesis oí common intercept that is why the techniques íor
panel data are more appropriate. 1he Breush-Pagan Lagrange Multiplier test
rejects the hypotheses oí zero ·ariance oí the indi·idual eííects disturbance term,
thus I choose the random eííects. P-·alue oí the lausman test is equal to 1 and
estimators oí both íixed and random eííects are not signiíicantly diííerent írom
each other, that is why I reject the null hypotheses and select the random eííects
model as more appropriate íor the data. 1he lausman test is represented in
Appendix 6, 1able 2. 1he estimators will be consistent and eííicient under the
random eííects. 1able 1 represents the speciíication tests with p-·alue and
decision rules.

22
1able 1: Speciíication 1ests íor the Model with the Dependent Variable as the
number oí SMLs
Speciíication tests p-·alue Decision
Common ·s Diííerent eííects: l test 0,0000 Diííerent eííects
lixed ·s Random eííects: Breush-Pagan test 0,0000 Random eííects
Random ·s lixed eííects: the lausman test 1,0000 Random effects

1able 3 in Appendix 6 represents the estimation oí the initial econometric model
with the number oí enterprises as the dependent ·ariable. Such ·ariables as the
a·erage time spent during each procedure, oííicial cost oí each procedure,
minimum capital requirements as the percentage oí GNI and GDP per capita are
insigniíicant, but ha·e expected signs. I can not neglect these indicators, because
they ser·e as control ·ariables. 1he insigniíicance oí these coeííicients can be
explained by the insuííicient size oí the sample.
Variable the oííicial costs oí going through court procedures is signiíicant, but
has the unexpected sign. I can not neglect this ·ariable, because its iníluence on
the entrepreneurial acti·ity is ambiguous. Ií the íirm does not ha·e money íor
speeding up the dispute resolution process by bribes, than increasing in the sum
oí costs will coníirm the ineííiciency oí the court system. As a result the contract
eníorcement depresses the business de·elopment. But ií the íirm has money or
can borrow írom some íinancial institutions and use them íor acceleration oí the
court process, than the contract eníorcement ía·ours the de·elopment oí
business and stimulate the corruption.
1he number oí procedures required to register a íirm is signiíicant and has the
expected sign ,Appendix 6, 1able 3,. Ií the number oí procedures increases by 1
procedure, than the number oí enterprises will decrease in 2895¯. 1he
interpretation oí this ·ariable is not quite correct, because in some countries írom

23
the database, íor example in Albania and Belarus, the total number oí enterprises
is ¯ ¯95 and 3098¯ respecti·ely. I íind the logarithm oí the dependent ·ariable the
number oí SMLs íor more correct interpretation oí the results. I can not íind
logarithm oí all explanatory ·ariables, because some oí them, íor example the
oííicial costs as a percentage oí GNI and minimum capital requirements as a
percentage oí income per capita, ha·e zero meanings.
1able 2 presents speciíication tests íor justiíication between OLS and panel data
techniques oí the regression with the logarithmic dependent ·ariable as the
number oí SMLs. 1he random eííects model is more appropriate íor this
regression.
1able 2: Speciíication 1ests íor the Model with the Logarithmic Dependent
Variable as the Number oí SMLs
Speciíication tests p-·alue Decision
Common ·s Diííerent eííects: l test 0,0000 Diííerent eííects
lixed ·s Random eííects: Breush-Pagan test 0,0000 Random eííects
Random ·s lixed eííects: the lausman test 0.9801 Random effects

1he estimation oí the initial regression is in the Appendix ¯, 1able 2. 1he number
oí procedures required íor business registration process and the number oí
calendar days needed íor dispute resolution are signiíicant, but ha·e unexpected
signs. As in the pre·ious case I can not neglect these ·ariables, because they can
enhance the entrepreneurial acti·ity ií íirms can borrow money íor speeding up
the registration and dispute resolution processes by bribes. 1he oííicial costs, the
a·erage time needed íor business registration process, minimum capital
requirements and the number oí procedures needed íor contract eníorcement are
insigniíicant, but ha·e expected signs. 1he insigniíicance oí these ·ariables can be
explained by the insuííicient size oí the sample.

24
1he correlation matrix shows that PROCLD and DA\ ha·e correlation
coeííicient 0,62¯6 ,Appendix ¯, 1able 1,. Increasing in the number oí procedures
can increase the number oí days needed íor íirm to start operate legally.
ívtre¡revevriat actirit, a. tbe vvvber of evter¡ri.e. ¡er ca¡ita.
1he insigniíicance oí some estimators in pre·ious regressions can be explained by
the iníluence oí country scale. 1he scale oí the country has a great iníluence on
the total number oí enterprises in the country. lor accounting this dependence I
represent the new dependent ·ariable as the number oí enterprises per capita.
1his regression is estimated using the same algorithm as in the pre·ious case. l-
test do not coníirm the hypothesis oí the signiíicance oí the indi·idual eííects,
thus I proceed with the panel data. 1he Breush-Pagan Lagrange Multiplier test
rejects the hypothesis that the indi·idual speciíic disturbance does not ·ary across
cross-sections, thereíore the random eííect is more plausible in this situation. 1he
lausman test establishes the eííiciency and consistency oí estimators. As a result
I deal with the random eííect model. Speciíication tests are represented in 1able
3.
1able 3: Speciíication 1ests íor the Model with the Dependent Variable as the
Number oí SMLs per Capita
Speciíication tests p-·alue Decision
Common ·s Diííerent eííects: l test 0,0000 Diííerent eííects
lixed ·s Random eííects: Breush-Pagan test 0,0000 Random eííects
Random ·s lixed eííects: the lausman test 0.3825 Random effects

Lstimation oí the initial econometric model with the number oí enterprises per
capita as the dependent ·ariable shows the insigniíicance oí all explanatory
·ariables ,1able 2, Appendix 8,. It can be explained by the correlation between

25
some explanatory ·ariables. 1he correlation matrix shows that such ·ariable as
the number oí procedures required to register a íirm ,PROCLD, is highly
correlated with the a·erage time spent during each procedures. 1his ·ariable is
highly insigniíicant with unexpected sign and small meaning. Aíter running the
initial regression without PROCLD the situation has impro·ed ,Appendix 8,
1able 3,. 1he number oí days needed íor registration is signiíicant and has the
expected sign. 1he number oí enterprises per capita will decrease by 89
enterprises per million oí the population, ií the a·erage time needed íor
registration proceed is increased by 1 day. Such ·ariables as the oííicial cost oí
each procedure, the minimum capital requirements as a percentage oí income per
capita and GDP per capita ha·e expected signs, but insigniíicant. 1he indicators
oí the contract eníorcement are insigniíicant with unexpected sighs. I can not
disregard these ·ariables because their ser·ed as control ones. 1he insigniíicance
can be explained by the insuííicient size oí the sample.
ívtre¡revevriat actirit, a. a .bare of ¡rirate .ector iv CDP ;º)
1able 4 represents speciíication tests íor detecting between OLS and panel
techniques. l-test rejects the null hypothesis oí common eííects, thus I deal with
íixed eííects. 1he Breush-Pagan Lagrange Multiplier test points in ía·or oí the
random eííects. 1he lausman test reíuses the hypotheses oí the ·alidity oí the
random eííects, thereíore the íixed eííects model is more appropriate.
1able 4: Speciíication 1ests íor the Model with the Dependent Variable as the
Share oí Pri·ate Sector in GDP ,°,

Speciíication tests p-·alue Decision
Common ·s Diííerent eííects: l test 0,0000 Diííerent eííects
lixed ·s Random eííects: Breush-Pagan test 0,0000 Random eííects
Random ·s lixed eííects, the lausman test 0,0141 Iixed effects

26
Breusch-Pagan , Cook-\eisberg test íor heteroscedasticity does not reject the
null hypothesis oí constant ·ariance. P-·alue is equal to 0.3684. Variances oí
regression disturbances are constant across obser·ations.
Running the initial regression with the share oí pri·ate business as the dependent
·ariable, I recei·ed the íollowing results ,Appendix 9, 1able 2,. 1he minimum
capital required as a percentage oí income per capita is signiíicant and correctly
signed. Ií minimum capital requirements increase by 1°, the share oí pri·ate
sector in GDP will decrease by 0,00¯°. 1he time needed íor dispute resolution is
insigniíicant, but this coeííicient has the expected sign. Such ·ariables as the
number oí days, procedures and oííicial costs concerning business registration
process and the number oí procedures and oííicial costs concerning dispute
resolution process are insigniíicant with unexpected signs. 1hese ·ariables are
control ones, that is why I can not neglect them. 1he insigniíicance and
unexpected signs can be explained by the insuííicient size oí the sample.
ívtre¡revevriat actirit, ava tbe ivforvat .ector of tbe ecovov,.
1he iníormal economy in GNI ,°, is a·ailable only íor 2002. I estimate this
model using OLS techniques. Results are presented in Appendix 10, 1able 2.
I checked this regression íor the heteroscedasticity. Breusch-Pagan , Cook-
\eisberg test íor heteroscedasticity does not reject the null hypothesis oí
constant ·ariance. P-·alue is equal to 0.9352
1he p-·alue oí the Ramsey RLSL1 omitted ·ariable test is 0.0344. 1he initial
econometric model is misspeciíied. Such ·ariables as the number oí days and
oííicial costs concerning the contract eníorcement process are highly
insigniíicant, but ha·e expected signs. 1he a·erage time spent during each
procedure is insigniíicant with unexpected sign. 1he coeííicient oí capital

2¯
requirements is signiíicant, but has unexpected sign. 1he interpretation oí the
coeííicient oí this estimator is not logic, because the increase in minimum capital
requirements needed íor business registration process will lead to decrease in the
iníormal economy. In íact, the dependence should be in·erse. Aíter running the
initial regression without all these ·ariables the p-·alue oí the Ramsey RLSL1
omitted ·ariable test is 0.1232. I do not reject the null hypotheses about model
misspesiíication.
1he number oí procedures required to register a íirm and the oííicial cost oí
each procedure are highly signiíicant and correctly signing. Ií the number oí
procedures increments in 1 procedure, it will lead to increase in the share oí the
iníormal economy in GNI by 1, 19°. Increase in the oííicial cost will expand the
shadow economy by 0, 04° in GNI. 1he number oí procedures is warranted by
law or court regulations that claim interaction between the parties and the judge is
signiíicance. Ií the number oí these procedures increases by 1, than the iníormal
economy as a share oí GNI will grow by 0, 2¯° ,Appendix 10, 1able 3,.
1he p-·alue oí Breusch-Pagan , Cook-\eisberg test íor heteroscedasticity is
0.9598. 1he null hypotheses about constant ·ariance can not be rejected.
As a result indicators oí business registration process and dispute resolution
procedure enhance the de·elopment oí the iníormal economy. Strict go·ernment
regulations oí the business de·elopment lead to increase oí the transíormation oí
legal business enterprises to iníormal ones. 1hat is why high le·els oí the
iníormal economy indicate about considerable le·els oí the iníormal business
acti·ity.

28
C b a ¡ t e r ó
CONCLUSIONS
In this work I in·estigate the impact oí go·ernment regulations on the
entrepreneurial acti·ity across countries around the world.
I characterize the entrepreneurial acti·ity as the number oí small and medium-
sized enterprises across 31 countries, as the number oí small and medium-sized
enterprises per capita across 31 countries and as the share oí pri·ate sector in
GDP across 24 transition countries. All this panel datasets are a·ailable íor 2002-
2003. Also I analyze the entrepreneurial acti·ity in the iníormal sector oí the
economy. 1he share oí the iníormal economy in GNI ,°, is a·ailable íor 2002
across 106 countries around the world. I choose this ·ariable íor my analysis
because harder go·ernment regulations oí business de·elopment lead to higher
le·els oí corruption and a greater relati·e size oí the iníormal economy.
Main data sources are Doing Business Databases 2004,2005 oí the \orld Bank,
1ransition report 2004 oí Luropean Bank íor Reconstruction and De·elopment
and Statistical Agencies oí diííerent countries.
In this analysis I use the íollowing determinants oí the business en·ironment:
costs oí obtaining legal status to operate a íirm as a share oí per capita GNI, time
that it takes to obtain legal status to operate a íirm, in business days, number oí
procedures that a íirm has to complete with in order to obtain a legal status, the
minimum capital required as a percentage oí income per capita, indicators oí
eníorcing contracts ,a number oí procedures, time in calendar days and oííicial
costs as a percentage oí the debt ·alue,, GDP per capita.

29
Lmpirical analysis coníirms the theoretical expectations about negati·e iníluence
oí go·ernment regulations on the business de·elopment. Ií the number oí days
needed íor business registration process increases by 1 day than the number oí
small and medium-sized enterprises will decrease by 89 enterprises per million oí
the population. Increase in minimum capital requirements by 1° leads to
decrease oí the share oí pri·ate sector in GDP by 0.00¯°. Ií the number oí
procedures that a íirm has to complete to start operate legally increments in 1
procedure, it will lead to increase in the share oí the iníormal economy in GNI by
1, 19°. Increase in oííicial costs concerning business registration process will
expand the shadow economy by 0, 04° in GNI. Ií the number oí procedures is
warranted by law or court regulations necessary íor dispute resolution process
expands by 1, than the iníormal economy as a share oí GNI will grow by 0, 2¯°.
Such positi·e iníluence oí indicators oí the business registration procedure and
the contract eníorcement process on the de·elopment oí the iníormal sector
means the expansion oí the iníormal business acti·ity and the reduction oí
íormal one.
During the empirical analysis was íound the unexpected positi·e iníluence oí
some determinants oí the business en·ironment on the entrepreneurial
de·elopment. 1hese ·ariables are the oííicial costs oí going through court
procedures, the number oí calendar days needed íor dispute resolution process
and the number oí procedures concerning the business registration process. I can
not neglect this iníluence, because ií íirms ha·e money íor speeding-up the
business registration process and the contract eníorcement procedure by bribes,
than said íactors stimulate the de·elopment oí business acti·ity and, oí course,
enhance the corruption.
1he main conclusion oí this research work is that go·ernment regulations ha·e
the negati·e impact on the entrepreneurial acti·ity. Strict go·ernment regulations

30
oí the business íorce enterprises work unoííicially, thus stimulate the
de·elopment oí the iníormal economy. 1he go·ernment should adjust the
regulations oí small and medium-sized business, because it is one oí the main
sources íor the economic growth, inno·ations and job creation.

31
BIBLIOGRAPl\
Ayygari, M., Beck, 1. and Demirgüç-
Kunt, A., ´vatt ava Meaivv
ívter¡ri.e. acro.. tbe Ctobe: . ^er
Databa.e, Policy Research \orking
Paper Series írom the \orld Bank,
No 312¯, August 2003
Berkowitz, D. and DeJong, N. D.,
ívtre¡revevr.bi¡ ava Po.t·´ociati.t
Crortb, \illiam Da·idson \orking
Paper Number 406, October 2001
Daneiko, P., Lyno·a, 1., Pelipasya, I.
and Rako·a, L., ßetarv.iav bv.ive..: tbe
covaitiov, tevaevcie. ava ¡ro.¡ect,
Analytical report, Belarusian
Uni·ersity oí Pri·atization and
Management, 2003.
Desai, M.., Gompers, P. and Lerner,
J., ív.titvtiov., Ca¡itat Cov.traivt. ava
ívtre¡revevriat íirv D,vavic.: íriaevce
frov ívro¡e, lar·ard NOM Research
Paper No. 03-59, lar·ard Uni·ersity
and National Bureau oí Lconomic
Research, 2003.
Djanko·, S., La Porta, R., Lopez de
Silanes, l. And Shleiíer, A., 1be
Regvtatiov of ívtr,, Quarterly Journal
oí Lconomics 11¯, 1-35,2001.
Djanko·, S., La Porta, R., Lopez de
Silanes, l. And Shleiíer, A., Covrt.,
NBLR \orking Paper 4 \8890.
Ldgcomb, L.L. and 1hetíord 1.,
1be ívforvat ícovov,. Ma/ivg ít iv
Rvrat .verica, lILLD, 1he Aspen
Institute, lebruary 2004.
lisman, R. and Sarria-Allende, V.,
Regvtatiov of ívtr, ava tbe Di.tortiov of
ívav.triat Orgavi¸atiov, \orking Paper
10929, National Bureau oí
Lconomic Research, Cambridge,
No·ember 2004.
Geroski, P., !bat Do !e Kvor abovt
ívtr,. International Journal oí
Industrial Organization, Vol.13, 421-
440, 1995.
Green \. l., ícovovetric .vat,.i., 4
th

edition, New \ork Uni·ersity, 2000.
Kaya, S. and Uçdogruk, \., 1be
D,vavic. of ívtr, ava í·it iv 1vr/i.b
Mavvfactvrivg ívav.tr,, \orking paper
oí Lconomic Research Center,
Middle Last 1echnical Uni·ersity,
September 2002.
Klapper, L., Lae·en, L. and Rajan,
R., ßv.ive.. ívrirovvevt ava íirv
ívtr,: íriaevce frov ívtervatiovat Data,
1his draít: June 2004.
McMillan, J. and \oodruíí, C., 1be
Cevtrat Rote of ívtre¡revevr. iv
1rav.itiov ícovovie., Journal oí
Lconomic Perspecti·es, Volume 16,
43, 153 -1¯0, 2002.

32
O·aska, 1. and S. Sobel R.S.,
ívtre¡revevr.bi¡ iv Po.t·´ociati.t
ícovovie., No 04-06 \orking Papers
írom the Department oí
Lconomics, \est Virginia
Uni·ersity, October 2004
Robinson, J. and lairchild, G.B.,
´ociat ava ív.titvtiovat ßarrier. to
Mar/et ívtr,, 2002
Doivg bv.ive.. iv 2001, Report oí the
\orld Bank, 2003.
Doivg bv.ive.. iv 200:, Report oí the
\orld Bank, 2004.
ív¡rorivg tbe ßv.ive.. ívrirovvevt,
Report oí the \orld Bank, Belarus,
January 2003.

33
Appendix J: Assumptions about the Business and Contract Lnforcement

1o make the business comparable across countries, the íollowing assumptions
were employed ,Doing Business 2004, \orld Bank,.
1he business
· Is a limited-liability company ,Ií there is more than one type oí limited-liability
company in the country, the type most popular among domestic íirms is
chosen.,,
· Operates in the country`s most populous city,
· Is 100 percent domestically owned and has íi·e íounders, none oí whom is a
legal entity,
· las start-up capital oí 10 times income per capita, paid in cash,
· Períorms general industrial or commercial acti·ities, such as the production and
sale oí products or ser·ices to the public,
· Leases the commercial plant and oííices,
· Does not qualiíy íor in·estment incenti·es or any special beneíits,
· las up to 50 employees one month aíter the start oí operations, all oí them
nationals,
· las turno·er oí at least 100 times income per capita,
· las a company deed 10 pages long.

1o make the data about eníorcing contracts comparable across the countries, the
next assumptions were made ,Doing Business 2004, \orld Bank,:
· 1he debt ·alue equals 50 percent oí the country`s income per capita.
· 1he plaintiíí has íully complied with the contract ,the plaintiíí is 100 percent
right,.

34
· 1he case presents a lawíul transaction between businesses residing in the
country`s most populous city.
· 1he bank reíuses payment íor lack oí íunds in the borrower`s account.
· 1he plaintiíí íiles a lawsuit to collect the debt.
· 1he debtor attempts to delay ser·ice oí process but it is íinally accomplished.
· 1he debtor opposes the complaint ,deíault judgment is not an option,.
· 1he judge decides e·ery motion íor the plaintiíí.
· 1he plaintiíí attempts to introduce documentary e·idence and to call one
witness. 1he debtor attempts to call one witness. Neither party presents
objections.
· 1he judgment is in ía·or oí the plaintiíí.
· No appeals or post-judgment motions are íiled by either party to the case.
· 1he debt is successíully collected.

35
Appendix 2: Procedure of business registration in Belarus in 2003
S1ANDARDIZLD COMPANY
Legal Iorm: Private Limited Company
Minimum capital requirements: tJ600.
City: Minsk
Procedure J. Obtain an approval of the company name with the Ministry of
Justice.
1ime to complete 1 day
Cost to complete 1 base rate` ,as oí Jan. 12, 2004, 1 base rate`~BR 1¯,500, ? >8,
Procedure 2. Open a temporary bank account.
1ime to complete 1 day
Cost to complete No charge
Procedure 3. Notarize documents and pay a registry fee.
1ime to complete 1 day
Cost to complete
BR 332,500 ,? >150,
Procedure 4. Business registration with the State Registry.
1ime to complete 40 days
Cost to complete Luro 60
Procedure S. Obtain the approval of the company seal draft from the Registry.
1ime to complete 3 days
Cost to complete No charge
Procedure 6. Obtain the approval of the company seal draft from the local police
department.
1ime to complete 8 days
Cost to complete 1-2 base rates
Procedure 7. Prepare a business seal.
1ime to complete 1 day
Cost to complete
BR 50,000 ,? >23,
Procedure 8. Receive a Management Certificate for the CLO.
1ime to complete ¯ days
Cost to complete 1 base rate
Procedure 9. Notarize the received company documents.
1ime to complete 1 day
Cost to complete
BR 180,000 ,? >82,
Procedure J0. Register the company in the 1ax Office.
1ime to complete 6 days ,íor getting the account number oí payer,
Cost to complete none
Procedure JJ. Apply for a Statistical Number.
1ime to complete 1 day
Cost to complete none

36
Procedure J2. Apply for duplicates of a Company's taxpayer ID.
1ime to complete 4 days
Cost to complete 3 base rates` per 1 duplicate.
Procedure J3. Apply for a certificate for opening a bank account from the Social
Security Office, register the company with the Social Security Office
1ime to complete 2 days
Cost to complete none
Procedure J4. Notarize a card with signatures of a Company's Director and
Chief Accountant and a sample of the Company seal.
1ime to complete 1 day
Cost to complete BR 140,000 ,íour base rates íor each signature,
Procedure JS. Register the company in the Belgosstrakh (state insurance
company).
1ime to complete 1 day
Cost to complete
depending on the bank, there is either no charge íor opening a
regular bank account, or it could be up to 10 base rates.
Procedure J6. Open a regular bank account.
1ime to complete 1 day
Cost to complete no charge

3¯
Appendix 3: Descriptive Statistics of Starting Business Indicators
Means
4 oí
procedures
days cost min capital
Countries and regions
2003 2004 2003 2004 2003 2004 2003 2004
All countries 10 10 55 49 93 81,06 289,6 1¯0,54
Last Asia & Paciíic 10 10 64 55 ¯2,8 ¯6,2 844,3 115,1
Lurope & Central
Asia
11 10 4¯ 40 22 15,9 100,4 53,3
Latin America &
Caribbean
12 12 ¯8 ¯3 ¯3,8 62,6 89,9 29,9
Middle Last & North
Aírica
11 10 54 38 63,2 65 1385 341
OLCD: ligh income ¯ 6 31 25 10,1 8 61,2 44,1
South Asia 9 9 45 44 ¯6,3 52,3 86,1 0
Sub-Saharan Aírica 11 11 ¯5 64 292,8 212 300,4 213,¯

Standard deviations

4 oí
procedures
days cost min capital
2003 2004 2003 2004 2003 2004 2003 2004
All countries 4 3 40 3¯ 180,6 164,8 ¯1¯,2 516,¯
Last Asia & Paciíic 2 2 5¯ 40 169,3 146 1313 156,2
Lurope & Central
Asia
3 3 28 23 16,8 9,6 111,6 54,5
Latin America &
Caribbean
4 3 45 46 85 62,¯ 193,4 61,6
Middle Last & North
Aírica
3 3 25 1¯ 85 86,2 1¯6¯,4 348,¯
OLCD: ligh income 4 4 29 25 14,9 8,6 8¯,4 ¯1,8
South Asia 1 2 28 28 6¯,3 31,5 192,5 0
Sub-Saharan Aírica 3 3 46 41 301,¯ 196,5 394,2 228,1

38

Maximum

4 oí
procedures
days cost min capital
All countries
2003 2004 2003 2004 2003 2004 2003 2004
All countries 19 19 215 203 129¯,6
1268,
4
562¯,2 5053,9
Last Asia & Paciíic 12 12 168 151 553,8 480,1 3855,9 394
Lurope & Central
Asia
19 16 118 123 65 46,2 450,8 23¯,9
Latin America &
Caribbean
19 1¯ 203 203 33¯,8 1¯6,1 699 182,4
Middle Last & North
Aírica
14 13 98 63 264,1 269,2 562¯,2 815,6
OLCD: ligh income 16 15 115 108 69,6 35,2 402,5 332
South Asia 10 11 88 89 191 91 430,4 0
Sub-Saharan Aírica 19 19 215 155 129¯,6 884,6 1¯56,1 ¯44,¯

Minimum
4 oí
procedures
days cost min capital
All countries
2003 2004 2003 2004 2003 2004 2003 2004
All countries 2 2 2 2 0 0 0 0
Last Asia & Paciíic ¯ 8 31 20 ¯,3 6,¯ 0 0
Lurope & Central
Asia
5 5 11 9 6,3 3,¯ 0 0
Latin America &
Caribbean
¯ ¯ 19 19 8 10 0 0
Middle Last & North
Aírica
6 5 34 11 5,3 4,9 ¯,4 2,1
OLCD: ligh income 2 2 2 2 0 0 0 0
South Asia ¯ ¯ 22 21 18,3 10,¯ 0 0
Sub-Saharan Aírica 6 6 26 26 8,¯ 9,1 0 0

39
Appendix 4: Descriptive Statistics of Lnforcing Contracts
Mean

4 oí procedures days cost
All countries
2003 2004 2003 2004 2003 2004
All countries 26 31 309 383 36,4 25,2
Last Asia & Paciíic 24 29 195 33¯ ¯¯,8 52,6
Lurope & Central
Asia
26 30 31¯ 35¯ 28,9 18
Latin America &
Caribbean
32 35 363 4¯2 38,8 20,3
Middle Last & North
Aírica
28 35 2¯8 419 15,¯ 16,¯
OLCD: ligh income 18 20 233 230 ¯,1 10,¯
South Asia 22 32 358 395 48,2 29,3
Sub-Saharan Aírica 31 35 233 45¯ ¯,1 42,8

Standard deviations

4 oí procedures days cost
All countries
2003 2004 2003 2004 2003 2004
All countries 11 11 230 235 ¯3,6 30,¯
Last Asia & Paciíic 4 5 62 129 104,¯ 4¯,¯
Lurope & Central
Asia
9 8 25¯ 188 51,5 9,5
Latin America &
Caribbean
11 8 286 268 95,3 10,1
Middle Last & North
Aírica
12 14 229 214 18,8 8,¯
OLCD: ligh income 4 5 160 2¯¯ 6,1 5
South Asia 6 11 60 38 31,1 9,6
Sub-Saharan Aírica 13 13 239 204 103,1 52,5

40

Maximum

4 oí procedures days cost
All countries
2003 2004 2003 2004 2003 2004
All countries 55 58 1460 1459 520,6 256,8
Last Asia & Paciíic 29 3¯ 2¯0 5¯0 269 126,5
Lurope & Central
Asia
44 46 1028 1028 254,¯ 4¯,9
Latin America &
Caribbean
4¯ 4¯ 1460 1459 440,5 41,2
Middle Last & North
Aírica
54 55 ¯21 ¯21 54,3 34,3
OLCD: ligh income 23 29 645 1390 28 21,1
South Asia 30 46 440 440 95 43,1
Sub-Saharan Aírica 55 58 895 1011 520,6 256,8

Minimum
4 oí procedures days cost
All countries
2003 2004 2003 2004 2003 2004
All countries 11 11 ¯ 1 0,3 4,2
Last Asia & Paciíic 19 22 ¯5 ¯5 1,8 5,4
Lurope & Central
Asia
16 1¯ 65 1 2,1 8,1
Latin America &
Caribbean
14 18 ¯6 155 2,4 4,2
Middle Last & North
Aírica
14 14 ¯ 2¯ 0,3 8,8
OLCD: ligh income 11 11 39 48 0,4 4,2
South Asia 15 1¯ 2¯0 350 ¯,6 21,3
Sub-Saharan Aírica 13 16 90 200 3,8 9,2

41
Appendix S: Descriptive Statistics of SMLs, the Private Sector and the
Informal Lconomy
1able J: Descriptive Statistics of SMLs
Number oí SMLs
2002 2003
Mean 1 093 804 1 142 06¯
Standard de·iation 1 880 061 1 960 399
Maximum 8 000 000 8 441 000
Minimum ¯ 622 ¯ ¯95

1able 2: Descriptive Statistics of the Share of Private Sector in GDP (°)
Share oí pri·ate sector in GDP ,°,
2002 2003
Mean 63.5 64
Standard de·iation 13.4 13.3
Maximum 80 80
Minimum 25 25

1able 3: Descriptive Statistics of the Share of Informal Lconomy in GNI
(°)
Share oí Iníormal Lconomy in GNI ,°,
2002
Mean 32,¯
Standard de·iation 14
Maximum 6¯,3
Minimum 8,8

42
Appendix 6: Lmpirical Lstimation of the Model with the Dependent
Variable as the Number of Lnterprises

1able J: Correlation Matrix
ENTERP PROCED DAY COST CAPIT EPROCED EDAY ECOST
ENTERP 1.0000
PROCED -0.0236 1.0000
DAY -0.0776 0.6276 1.0000
COST -0.2110 0.4494 0.4316 1.0000
CAPIT -0.0639 0.3881 0.1071 0.4773 1.0000
EPROCED -0.1639 0.3537 0.2411 0.4494 0.1153 1.0000
EDAY -0.0696 0.2451 0.2696 0.2646 0.0984 0.2902 1.0000
ECOST -0.1520 0.3922 0.2344 0.4800 0.0644 0.3034 0.1406 1.0000
GDPpc 0.0894 -0.6781 -0.4243 -0.5403 -0.2549 -0.4367 -0.3008 -0.4650
GDPpc
GDPpc 1.0000

1able 2: Random versus Iixed Lffects: 1he Hausman 1est
---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fixed random Difference S.E.
-------------+----------------------------------------------------------------
PROCED | -28941.41 -28957.09 15.67346 5138.129
DAY | -523.4268 -538.7161 15.28927 436.8473
COST | -683.2311 -786.2578 103.0267 685.5631
CAPIT | -44.42186 -43.98051 -.4413515 86.37763
EPROCED | 4227.291 4344.794 -117.5031 2021.672
EDAY | 264.5148 263.553 .9618483 62.75202
ECOST | 2487.123 2519.588 -32.46561 540.4738
GDPpc | 22.10944 11.15881 10.95063 57.56707
------------------------------------------------------------------------------
b = consistent under Ho and Ha; obtained from xtreg
B = inconsistent under Ha, efficient under Ho; obtained from xtreg
Test: Ho: difference in coefficients not systematic
chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 0.09
Prob>chi2 = 1.0000

43
1able 3: Initial Regression
Dependent Variable: the Number of Lnterprises

Random-effects GLS regression Number of obs = 62
Group variable (i): ID Number of groups = 31

R-sq: within = 0.4762 Obs per group: min = 2
between = 0.0009 avg = 2.0
overall = 0.0011 max = 2

Random effects u_i ~ Gaussian Wald chi2(8) = 23.30
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0030

ENTERP | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | -28957.09 13248.8 -2.19 0.029 -54924.25 -2989.92
DAY | -538.7161 1218.351 -0.44 0.658 -2926.64 1849.208
COST | -786.2578 1932.283 -0.41 0.684 -4573.463 3000.947
CAPIT | -43.98051 238.07 -0.18 0.853 -510.5892 422.6281
EPROCED | 4344.794 4533.109 0.96 0.338 -4539.935 13229.52
EDAY | 263.553 167.2049 1.58 0.115 -64.16251 591.2685
ECOST | 2519.588 1367.804 1.84 0.065 -161.2587 5200.436
GDPpc | 11.15881 23.95486 0.47 0.641 -35.79185 58.10948
_cons | 1020245 562948 1.81 0.070 -83112.71 2123603
-------------+----------------------------------------------------------------
sigma_u | 2144850
sigma_e | 63116.768

44
Appendix 7: Lmpirical Lstimation of the Model with the Logarithmic
Dependent Variable as the Number of Lnterprises
1able J: Correlation Matrix
logENT~P PROCED DAY COST CAPIT EPROCED EDAY ECOST
logENTERP 1.0000
PROCED 0.9673 1.0000
DAY 0.5874 0.6276 1.0000
COST 0.4506 0.4494 0.4316 1.0000
CAPIT 0.3765 0.3881 0.1071 0.4773 1.0000
EPROCED 0.3498 0.3537 0.2411 0.4494 0.1153 1.0000
EDAY 0.2820 0.2451 0.2696 0.2646 0.0984 0.2902 1.0000
ECOST 0.3807 0.3922 0.2344 0.4800 0.0644 0.3034 0.1406 1.0000
GDPpc -0.6825 -0.6781 -0.4243 -0.5403 -0.2549 -0.4367 -0.3008 -0.4650
GDPpc
GDPpc 1.0000

1able 2: Initial Regression
Logarithmic Dependent Variable: the Number of Lnterprises

Random-effects GLS regression Number of obs = 62
Group variable (i): ID Number of groups = 31

R-sq: within = 0.8522 Obs per group: min = 2
between = 0.9424 avg = 2.0
overall = 0.9387 max = 2

Random effects u_i ~ Gaussian Wald chi2(8) = 578.51
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

logENTERP | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | .1284903 .0077787 16.52 0.000 .1132443 .1437363
DAY | -.0008979 .0008202 -1.09 0.274 -.0025055 .0007096
COST | -.0000311 .0015634 -0.02 0.984 -.0030953 .0030332
CAPIT | -.0000351 .0001765 -0.20 0.843 -.000381 .0003108
EPROCED | -.0017768 .0028584 -0.62 0.534 -.0073791 .0038254
EDAY | .0002048 .0001068 1.92 0.055 -4.65e-06 .0004142
ECOST | .0006097 .0011824 0.52 0.606 -.0017079 .0029272
GDPpc | -1.33e-06 2.36e-06 -0.56 0.573 -5.96e-06 3.30e-06
_cons | .9662773 .1202702 8.03 0.000 .730552 1.202003
-------------+----------------------------------------------------------------
sigma_u | .13316436
sigma_e | .06379726
rho | .81332291 (fraction of variance due to u_i)

45
Appendix 8: Lmpirical Lstimation of the Model with the Dependent
Variable as the Number of Lnterprises Per Capita
1able J: Correlation Matrix
ENTERPpc PROCED DAY COST CAPIT EPROCED EDAY ECOST
ENTERPpc 1.0000
PROCED -0.1558 1.0000
DAY 0.0786 0.6276 1.0000
COST -0.1155 0.4494 0.4316 1.0000
CAPIT 0.0249 0.3881 0.1071 0.4773 1.0000
EPROCED -0.2459 0.3537 0.2411 0.4494 0.1153 1.0000
EDAY 0.1327 0.2451 0.2696 0.2646 0.0984 0.2902 1.0000
ECOST -0.1961 0.3922 0.2344 0.4800 0.0644 0.3034 0.1406 1.0000
GDPpc 0.0511 -0.6781 -0.4243 -0.5403 -0.2549 -0.4367 -0.3008 -0.4650
GDPpc
GDPpc 1.0000

1able 2: Initial Regression
Dependent Variable: the Number of Lnterprises Per Capita

Random-effects GLS regression Number of obs = 62
Group variable (i): ID Number of groups = 31

R-sq: within = 0.2386 Obs per group: min = 2
between = 0.0002 avg = 2.0
overall = 0.0001 max = 2

Random effects u_i ~ Gaussian Wald chi2(8) = 6.69
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.5707

ENTERPpc | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | 5.12e-06 .0005843 0.01 0.993 -.0011402 .0011504
DAY | -.0000887 .0000542 -1.64 0.102 -.0001949 .0000176
COST | -.0000221 .0000863 -0.26 0.798 -.0001913 .0001471
CAPIT | -2.16e-07 .0000106 -0.02 0.984 -.000021 .0000205
EPROCED | .0000802 .0001996 0.40 0.688 -.0003109 .0004713
EDAY | 6.65e-06 7.42e-06 0.90 0.370 -7.89e-06 .0000212
ECOST | .000026 .0000611 0.43 0.670 -.0000937 .0001458
GDPpc | 2.21e-07 5.79e-07 0.38 0.702 -9.14e-07 1.36e-06
_cons | .0390585 .0154204 2.53 0.011 .0088352 .0692819
-------------+----------------------------------------------------------------
sigma_u | .04381293
sigma_e | .00261036
rho | .99646281 (fraction of variance due to u_i)

46

1able 3: Initial Regression without Variable PROCLD
Dependent Variable: the Number of Lnterprises Per Capita

Random-effects GLS regression Number of obs = 62
Group variable (i): ID Number of groups = 31

R-sq: within = 0.2390 Obs per group: min = 2
between = 0.0002 avg = 2.0
overall = 0.0001 max = 2

Random effects u_i ~ Gaussian Wald chi2(7) = 7.42
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.3866

ENTERPpc | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
DAY | -.000089 .0000485 -1.84 0.066 -.0001839 6.02e-06
COST | -.0000222 .0000817 -0.27 0.786 -.0001824 .000138
CAPIT | -1.17e-07 9.67e-06 -0.01 0.990 -.0000191 .0000188
EPROCED | .0000827 .0001706 0.48 0.628 -.0002516 .0004171
EDAY | 6.55e-06 7.04e-06 0.93 0.352 -7.25e-06 .0000204
ECOST | .0000267 .0000582 0.46 0.647 -.0000874 .0001407
GDPpc | 2.33e-07 5.74e-07 0.41 0.685 -8.92e-07 1.36e-06
_cons | .0388854 .0131222 2.96 0.003 .0131664 .0646045
-------------+----------------------------------------------------------------
sigma_u | .04618406
sigma_e | .00256568
rho | .99692332 (fraction of variance due to u_i)

4¯
Appendix 9: Lmpirical Lstimation of the Model with the Dependent
Variable as the Share of Private Sector in GDP (°)
1able J: Correlation Matrix
PRSECTOR PROCED DAY COST CAPIT EPROCED EDAY ECOST
PRSECTOR | 1.0000
PROCED | -0.5326 1.0000
DAY | -0.2793 0.5449 1.0000
OST | -0.1348 0.1128 0.2232 1.0000
CAPIT | -0.0091 0.1644 0.0674 0.4268 1.0000
EPROCED | -0.2448 -0.1230 -0.2120 0.0458 -0.1448 1.0000
EDAY | -0.1108 -0.0125 0.0777 -0.0422 -0.0257 0.1203 1.0000
ECOST | -0.0442 0.0117 -0.0798 0.0776 -0.0119 0.2894 -0.0098 1.0000

1able 2: Initial Regression
Dependent Variable: the Share of Private Sector in GDP

R-sq: within = 0.2964 Obs per group: min = 2
between = 0.1486 avg = 2.0
overall = 0.1019 max = 2

F(7,17) = 1.02
corr(u_i, Xb) = -0.3726 Prob > F = 0.4504

------------------------------------------------------------------------------
PRSECTOR | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | .063616 .4122386 0.15 0.879 -.8061314 .9333633
DAY | .0110234 .0200569 0.55 0.590 -.031293 .0533398
COST | .0188905 .030205 0.63 0.540 -.0448364 .0826174
CAPIT | -.006645 .0031649 -2.10 0.051 -.0133224 .0000324
EPROCED | .0408503 .0716434 0.57 0.576 -.110304 .1920046
EDAY | -.0011913 .0016502 -0.72 0.480 -.004673 .0022904
ECOST | .0011962 .0070374 0.17 0.867 -.0136514 .0160439
_cons | 62.0396 6.377985 9.73 0.000 48.58323 75.49598
-------------+----------------------------------------------------------------
sigma_u | 13.59706
sigma_e | 1.0171835
rho | .99443475 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(23, 17) = 205.37 Prob > F = 0.0000

48
Appendix J0: Lmpirical Lstimation of the Model with the Dependent
Variable as the Share of the Informal Lconomy in GNI (°)
1able J: Correlation Matrix

INFECON PROCED DAY COST CAPIT EPROCED EDAY ECOST
INFECON 1.0000
PROCED 0.4129 1.0000
DAY 0.2256 0.5715 1.0000
COST 0.3521 0.2464 0.2333 1.0000
CAPIT -0.1504 0.1017 0.0390 0.1530 1.0000
EPROCED 0.2831 0.2667 0.1270 0.1191 0.1427 1.0000
EDAY 0.1171 0.1091 0.0676 0.1453 0.0771 0.1898 1.0000
ECOST 0.1073 0.1065 0.1968 0.1539 -0.0286 0.0506 -0.0002 1.0000

1able 2: Initial Regression

Dependent Variable: the Share of the Informal Lconomy in GNI

Source | SS df MS Number of obs = 106
-------------+------------------------------ F( 7, 98) = 6.89
Model | 6638.10429 7 948.300613 Prob > F = 0.0000
Residual | 13479.6811 98 137.547766 R-squared = 0.3300
-------------+------------------------------ Adj R-squared = 0.2821
Total | 20117.7853 105 191.597956 Root MSE = 11.728

INFECON | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | 1.354094 .4117764 3.29 0.001 .536937 2.171251
DAY | -.023929 .0413736 -0.58 0.564 -.1060336 .0581757
COST | .046042 .0139454 3.30 0.001 .0183677 .0737162
CAPIT | -.0047158 .0015553 -3.03 0.003 -.0078022 -.0016294
EPROCED | .3128295 .1405463 2.23 0.028 .0339199 .5917391
EDAY | .0014208 .0050281 0.28 0.778 -.0085574 .0113989
ECOST | .0037359 .0156219 0.24 0.811 -.0272652 .034737
_cons | 10.97582 4.259043 2.58 0.011 2.523883 19.42775

49
1able 3: Regression of the Model with the Dependent Variable as the
Share of the Informal Lconomy in GNI
Source | SS df MS Number of obs = 106
-------------+------------------------------ F( 3, 102) = 12.22
Model | 5319.90306 3 1773.30102 Prob > F = 0.0000
Residual | 14797.8823 102 145.077277 R-squared = 0.2644
-------------+------------------------------ Adj R-squared = 0.2428
Total | 20117.7853 105 191.597956 Root MSE = 12.045

INFECON | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
PROCED | 1.194281 .3553767 3.36 0.001 .4893931 1.899169
COST | .0407081 .0139093 2.93 0.004 .013119 .0682972
EPROCED | .2747924 .1414469 1.94 0.055 -.005767 .5553517
_cons | 11.81826 4.288197 2.76 0.007 3.312637 20.32387

4

doc_640936619.pdf