FORMAL AND INFORMAL INSTITUTIONS IN BOSNIA AND HERZEGOVINA
- DETERMINANTS AND RELATIONSHIPS -
Dr Adnan Efendic
1
Marie Curie IAPP Conference Zagreb, 01 September 2015
INTRODUCTION
A little bit about theory Some initial inputs about institutional
environment in BiH Empirical research on the link between formal
and informal institutions Remaining challenges for empirical research Implications
2
THEORY
Institutions - defined as FORMAL and INFORMAL rules of the game (North, 1990)
A good institutional arrangement is one that provides effective interaction between a set of formal and informal institutions (De Soto, 2000)
Informal institutions has been relatively neglected dimension, partly due to data limitations
3
THEORY
Formal and informal rules of the game - the meaning of “and” not very clear
The whole range of potential links, from unrelated to mutually endogenous, related either as complements or substitutes.
Possibilities: Complements (desirable) Substitutes (non-efficient interaction) Gaps (transition open up the gap) Informal institutions explain the formal once
4
EMPIRICAL INVESTIGATION
Quantitative investigation based on survey data and probability modelling
Determinants which affect confidence in formal institutions and reliance on informal institutions in the focus
The link between formal and informal institutions particularly investigated
5
FORMAL INSTITUTIONS IN BIH
Complex institutional structure, which was established through the DPA in 1995
14 different government levels and types Overlapping and indeterminate jurisdictions
between the levels of government A top-down approach in creating institutional
set-up in the country Experience: not widely recognized or viewed as
credible, lack legitimacy, difficult to enforce and more reliance on informal institutions is likely
6
FORMAL INSTITUTIONS IN BIH
7
FORMAL INSTITUTIONS (HFI - EF)
8
0
0.2
0.4
0.6
0.8
1
Aggregated2007
Regulation Trade Fiscal Gov't Monetary Investment Financial P ropertyRights
Labour
BIH AVERAGE 27
INFORMAL INSTITUTIONS IN BIH
“If you want to suceed, you should have your own network??”
Why?: different reasons, but primarily:To deal with (formal) institutional
inefficiencyPost-conflict nature – formal institutions
were overburdened if not absent so that people often had to rely on informal institutions to cope.
9
SURVEY DATA
Survey data from 2008, UNDPBiH (Early Warning System) and 2012 (RRPP project – 1 survey)
Sample: representative of the different entities, regions, municipalities, ethnic groups, genders and urban/rural areas
Pooled data from 3 surveys 2008 – (different intercepts); 7,617 observations
Construct varaibles which code respondents’ perceptions on “institutional questions”
10
DESCRIPTIVE STATISTICS
11
QUESTIONSDummy Valid
%Mean value
Q1. To what degree do you agree with this statement? "I am confident that the legal system will uphold my contract and property rights in business/household disputes".
FORMAL
0 – Agree 58.30.39
1 - Disagree 41.7
Q3a. Can you estimate how much you use informal links (like links with your relatives, friends, colleagues, etc.) in your everyday life activities?
INFORMAL
0 – Little/not 73.70.25
1 –Moderately/lot26.3
Q3b. Can you estimate how much you use informal rules (like unwritten rules, codes, norms, customs, conventions, etc.) in your everyday business (life) activities?
0 – Use little/not 76.2
0.241 –Moderately/lot 23.8
VARIABLES BASED ON THEORY
“Institutional variables” that control for: Formal institutions – confidence – property rigthts,
contracts Informal institutions – reliance – informal rules
and networks Direct costs of institutions (taxes, fees, admin
costs, …) Indirect costs of institutions (estimated costs –
time, lack of enforcements, lack of efficiency,…) Change in institutions – national/entity over the
last 5 years Istitutional structures – entities FBiH and RS Ethnic status – minority vv majority Residence – urban vv rural area Gender – male vv female Timing - surveys
12
LINK FORMAL-INFORMAL INSTITUTIONS
Institutional environment – formal and informal institutions - two (separate?) parts of one story
Theory – the link is very “complex” - complements/substitutes/gap
Possibilites in empirical context:1. “Common” observed variables influence
the both types of institutions in BiH2. Institutions “specific” influences3. Complexity hardly “(un)observed” in
the model13
SUPM MODELLING - FEATURES
SUPM - allows for a more complex (seemingly unrelated) pattern of joint determinations than simple simultaneity
Formal and informal institutions are related as joint outcomes of a wider system of influences
The relationship (FOR-INF) is modelled: implicitly through the unobserved correlations in the
error terms, explicitly by controlling for common observable
variables in the system,explicitly by controlling for equation specific
variables in the system.14
CONFIDENCE IN FORM ALINSTITUTIONS
(Confidence in contractand property rightsprotection in BiH)
RELIANCE ON INFORM ALINSTITUTIONS
(The estim ate of how m uchinform al rules/links have
been used in BiH)
SPECIFIC VARIABLES
- d irect costs paid(taxes, fees, adm in costs)
SPECIFIC VARIABLES
- Urban/Rural area- Gender of respondents
INDIRECT COSTS OF INSTIT.
(lack enforcem ent, lack effic iency,tim e-spending procedures, ...)
THE DIRECTION OF INST.CHANGE LAST FIVE YEARS
(Better, Sam e, W orse)
ENTITIES W ITH DIFFERENTINST. ENVIRONM ENT
(FB iH and RS)
ETHNIC STATUS
(M inority or Majority)
OM M ITED INFLUENTIALFACTORS
(e.g. the war and post-war effects,specific religious environm ent, ...)
COM M ON OBSERVED FACTORS
COM M ON UNOBSERVEDFACTORS
BASE-LINE MODEL RESULTSThe dependent var. in the 1st equation is DPROPER (0=’agree’; 1=’disagree’)The dependent var. in the 2nd equation is D3INFRULE (0=’use little/not’; 1=’use moderately lot’)
DPROPER D3INFRULE
Variable Coef. Std. error z-stat.P> Coef. Std. error z-stat.
P> CONSTANT - 0.82 0.08 - 10.6 0.000 - 0.30 0.08 - 3.78 0.000
D4DIR10 - 0.15 0.07 - 2.10 0.036 - - - -
D4DIR30 - 0.25 0.06 - 4.19 0.000 - - - -
D4IND5 - 0.04 0.08 - 0.54 0.586 - 0.45 0.06 - 7.15 0.000
D4IND20 - 0.12 0.06 - 2.24 0.025 - 0.27 0.05 - 5.91 0.000
D5NATWOR 0.18 0.09 2.08 0.038 - 0.11 0.09 - 1.22 0.223
D5NATSAME - 0.09 0.07 - 1.38 0.168 - 0.06 0.07 - 0.90 0.367
D5ENTWOR 0.67 0.09 7.33 0.000 - 0.22 0.10 - 2.24 0.025
D5ENTSAME 0.59 0.07 8.30 0.000 - 0.10 0.07 - 1.36 0.174
DENTITY 0.28 0.04 6.25 0.000 0.15 0.05 3.08 0.002
DMAJOR 0.05 0.05 0.96 0.338 - 0.15 0.05 - 3.07 0.002
DURBAN - - - - 0.12 0.04 2.90 0.004
DGENDER - - - - 0.11 0.04 2.55 0.011
SEPT08 0.21 0.04 4.79 0.000 - 0.18 0.05 - 3.91 0.000
NOV08 0.10 0.05 1.88 0.060 - 0.40 0.06 - 6.76 0.000
z z
P> z
MARGINAL EFFECTS AFTER SUPM
The dependent variables are DPROPER=1 and D3INFRULE=1
Variable Explanation of the variables dy/dx Std. error z-stat.
D4DIR10 Direct costs, 1=range 0-10%. - 0.01 0.01 - 2.11 0.035
D4DIR30 Direct costs, 1= range > 10-30%. - 0.02 0.01 - 4.20 0.000
D4IND5 Indirect costs, 1= range 0-5%. - 0.06 0.01 - 6.58 0.000
D4IND20 Indirect costs, 1= range > 5-20%. - 0.05 0.01 - 5.70 0.000
D5NATWOR Change in state instit.; 1=” worse” 0.01 0.02 0.03 0.977
D5NATSAME Change in state institutions; 1=’same’ - 0.02 0.01 - 1.43 0.151
D5ENTWOR Change in entity instit.; 1= ‘worse’ 0.02 0.02 1.17 0.241
D5ENTSAME Change in entity institutions; 1=’same’ 0.04 0.01 3.55 0.000
DENTITY Entity in BiH; RS=0, FBiH=1 0.04 0.01 5.93 0.000
DMAJOR Ethnic status; minority=0, majority=1 - 0.02 0.01 - 1.82 0.069
DURBAN Type of area; rural=, urban=1 0.02 0.01 2.90 0.004
DGENDER Gender; female=0, male=1 0.01 0.01 2.54 0.011
SEPT08 Survey from September 2008=1 - 0.01 0.01 - 0.68 0.498
NOV08 Survey from November 2008=1 - 0.04 - 0.01 - 5.13 0.000
Note, dy/dx is for discrete change of dummy variable from 0 to 1. Source: Authors’ calculations using STATA 10.
CONCLUSION ON THE LINK
A mutually endogenous relationship between confidence in formal institutions and reliance on informal institutions
The success or failure of formal institutions is mirrored by the decreasing or increasing role of informal institutions
Hence, the result supports substituting relationship between formal and informal institutions
18
SPECIFIC FINDING - DETERMINANTS
Indirect costs matters the most – the highest marginal effect in the model – higher indirect costs associated with lower confidence in F.INST and more reliance on INF.INST
Institutional structure matters – FBiH entitiy respondents are less confident in formal and use informal institutions more than those from RS.
Ethnicity matters - Minority - have lower confidence in F.INST and more reliance on INF.INST than majority
19
SPECIFIC FINDING - DETERMINANTS
Lack of improvements in inst. – lower confidence in F.INST and more reliance on INF.INST (no strong evidence for INF.)
Higher direct costs – higher expectations from F.INST – less confident
Urban vv Rural - more reliance on INF.INST
Male vv Female – more reliance on INF.INST
20
ROBUSTNESS PROCEDURES
Logit, probit, multinomial Logit
Full sample with “Don’t knows”
Different proxies for informal institutions (i.e. Informal rules or Informal networks)
Inclussion or exclussion of some ommited variables – model diagnostics
New survevy data from 2012 confirms the main finding - substitution 21
INVESTIGATE FURTHER
Ommited infuences – no possiblity to control Our analysis cannot identify fully the
underlying causal relationships between formal and informal instititons
Dynamic panel context would be an advantage
Comparison of the results with other countries – extend the research?
Different types of formal/informal institutions ...
22
IMPLICATIONS
Puts into question assumptions either that formal and informal institutions are unrelated or that informal institutions are exogenous with respect to formal institutions
The quantitative effects of indirect costs of institutions are substantially larger than the effects of direct costs
Informal institutions are more persistent than formal ones – changes in formal institutions do not influence the degree of reliance on informal institutions 23
THANK YOU!
24