August 10, 2006

Document of the World Bank

Report No. 33115-DJ

Republic of DjiboutiCountry Economic Memorandum

(In Three Volumes) Volume III: Technical Annexes

Social and Economic Development GroupMiddle East and North Africa Region

Unlocking Djibouti’s Growth Potential: The Road AheadPub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

REPUBLIC OF DJIBOUTI Country Economic Memorandum

Technical Annexes

Table o f Contents

Annex 1.1 : Growth Accounting Methodology ..................................................................... 1

Annex 1.3: Outstanding Statistical Challenges ................................................................... 6 Annex 2.1: Government Wage and Employment Model ................................................... 8 Annex 2.2: Social Impact o f A Reduction In Public Sector Nominal Wages ................. 10 Annex 3.1: Public-Private Wage Differential and Returns to Education ....................... 12 Annex 3.2: Djibouti’s Investment Climate . Costs and Regulations for Business

Annex 3.3: Field Survey on Wages and Prices in Djibouti. Ethiopia. and Yemen:

Annex 1.2: Panel Growth Regression ................................................................................... 4

Start.up, Operation and Ex i t .......................................................................... 20

March 2005 ....................................................................................................... 37

Annex 5.2: Road Condition o f the Paved Network ........................................................... 49 Annex 5.3: Port o f Djibouti’s Wage Bill and Accounts .................................................... 50 Annex 5.4: Estimation o f the Impact o f Traffic in Transit on Employment .................. 51 Annex 5.5: References ......................................................................................................... 53

Annex 5.1: Port o f Dj ibout i Data ........................................................................................ 47

Vice President: Christiaan Poortman Country Director: Emmanuel Mbi Sector Director: Sector Manager: M i r i a Pigato Task Team Leader:

Mustapha K . Nab l i

Paloma An6s Casero

ANNEX 1.1 : GROWTH ACCOUNTING METHODOLOGY

T o determine the sources o f economic growth, economists usually conduct a growth accounting exercise using the neoclassical growth model. In this model, the aggregate production function, which tells u s the maximum output that an economy could produce with a given amount o f inputs, i s typically specified in a Cobb-Douglas form as Y, =A, (LH); K , h , where Y i s output at factor cost, K i s the stock o f capital (machines and equipment), H i s human capital (proxied by the average number o f years o f schooling for the population over 25 years o f age),’ L i s labor (total number o f people worlung or total time worked), and t i s time. The parameter a represents the share o f labor in total compensation o f factors o f production. It i s generally assumed that Y i s positively related to K , H , and L at time t . A, i s defined as total factor productivity (TFP). A higher value o f A, means that people “work smarter” and learn how to obtain more output f rom a given supply o f capital and labor.

TFP growth i s measured as a residual. Using the previous equation and assuming that H i s embedded in labor (L), the growth rate o f output can be decomposed into the growth rate o f technology and the weighted growth rates o f the input factors:

AY AA, AL AA 1- -- + a 2 + (1 - a) T-1 4 - 1 Lr-1 Kr-1 4-1

Because the growth rates o f GDP, labor, and capital are available in the real sector data, TFP growth rates are obtained by subtracting f rom the GDP growth rate the sum o f the growth rates of labor and capital, with the appropriate weights a and (1-4.

, with TFP growth measured by 2 and derived as a residual.

The time series for the capital stock in constant prices were constructed by applying the procedure set forth in the next section. Gross fixed capital formation in constant prices was derived using the nominal series deflated by the GDP deflator.

Data on the employed labor force for the public sector are obtained from official sources. Where employment data for public enterprises and the private sector are not available (since the early 199Os), various assumptions were made to fill gaps (such as assuming a stable average wage and stable private-to-private sector employment ratios). The value used for the labor share a i s 0.4. 2

’ Benefits f r o m education are assumed to be embodied in workers.

for developing counties Bisat et al. (1997) use a share o f 0.3. Other cross-country regressions found a range between 0.3 and 0.5

1

Derivation of Capital Stock

In the absence o f actual data, the stock o f capital was estimated based on assumptions for capital accumulation. In empirical investigations, the stock o f real capital (K,) i s usually supposed to evolve over time as follows:

K, = Kt-, (1 - 6) + I, (1) where It stands for the level o f investment (or gross capital formation) and 6 i s the depreciation rate.

Given an estimate for a f irst year (1980 for example): equation (1) allows us to derive an estimated series for real capital stock. T o estimate the in i t ia l level o f real capital stock, in our case K1980, one can substitute the same capital accumulation pattern backward and assume that the economy was historically on a steady state path:

K,,,, = K1980-r (1-8Y + s qgs0 [l+E+ l + g ("):+ l + g ... + [ E ] ] l + g (2)

where T represents the number of years the calculation i s pushed backward, s i s the historical steady state investment-to-GDP ratio (ID"), Y/980 i s the 1980 level o f real GDP, and g i s the historical steady-state real GDP growth rate. For a number o f years sufficiently large, T+ 00, the f i r s t component o f equation (2) vanishes (meaning the whole stock o f capital can be assumed to be fully depreciated)? The relation between K/980 and Y1980 can be approximated by:

Assuming a depreciation rate o f 5 percent, a steady-state investment rate o f 15 percent, and a real GDP growth o f -0.13 percent (the average real GDP growth during 1980-2002), the capital stock-to-GDP ratio was about 3.076 in 1980.

An alternative way o f estimating the init ial capital stock, which yields similar estimates for the init ial capital stock-to-GDP ratio, i s t o assume that the stock o f real capital evolves over time according to:

The main difference between this equation and the earlier specification o f (1) i s the timing o f investment and the impact on the current period or next period's capital stock. Most growth models tend to use the specification o f (4) because current period output, Y,, can then be stated in terms o f current period capital stock, K,, rather than the previous period capital stock, Dividing both sides o f equation (1) by Y,] yields:

The in i t ia l year o f the sample i s 1980 because it i s the earliest for which GDP estimates are available.

l +g The sum in brackets reduces to - .

2

Simultaneously dividing and multiplying the left-hand side o f equation (5) by Yt leads to:

(Kt / r,-1) (r, / r, 1 = ( K - 1 / r,-1)(1- 6) + ( 4 - 1 / T-1)

On a steady-state path, the ratio o f capital stock to GDP i s constant: (K, / r , ) = (&, /<-I) = c

Equation (6) becomes: c ( l + g ) = c ( 1 - 6 ) + s (5) c ( g + 6 ) = s or c = s / ( g + 6 )

r 1

In this case, assuming a depreciation rate o f 5 percent, a steady-state investment rate o f 15 percent, and a real GDP growth of-0.13 percent, the ratio o f capital stock to GDP i s estimated at about 3.080 in 1980.

3

ANNEX 1.2: PANEL GROWTH REGRESSION List o f countries in the growth panel

MENA Algeria Egypt Iran, Islamic Rep. Jordan Morocco Tunisia Djibouti

AFRICA 1 ASIA Cote d’Ivoire I Bangladesh Cameroon Gabon Ghana Kenya Madagascar Mauritius Malawi Niger Nigeria Togo

Indonesia India Korea, Rep. Sri Lanka Malaysia Pakistan Philippines Thailand

LATIN AMERICA Argentina Bolivia Brazil Chile Colombia Costa Rica Ecuador Guatemala Mexico Paraguay Peru Uruguay

Principal Component Analysis

Reported below for the purpose o f reproduction, are the principal component analysis estimation results which include the principal factors, their associated eigen values, their relative contribution to the variance and their factor loadings.

Table A2.1: Institutional Quality Variables Component Eigenvalue Cumulative R‘

PI 2.54 1.05 P2 0.02 1.05 P3 -0.05 1.03 P4 -0.08 1 .oo

Loadings PI P2 gel 0.95 -0.01 rq 0.42 0.10 law 0.92 0.03 cor 0.78 -0.07

IQ=Pl

Table A2.2: Macroeconomic Stability Variables Component Eigenvalue Cumulative R‘

PI 1.35 0.34 P2 1.02 0.59 P3 0.89 0.82 P4 0.74 1 .oo

Loadings P I P2 P3 P4 P -0.15 0.89 -0.41 -0.17 curacc 0.64 0.14 -0.24 0.72

exdebt -0.60 0.22 0.44 0.64 pubdef 0.46 0.39 0.77 -0.22

MS=(.3368 *Pl +O.2554*P2)/O0.5922

Table A 2.3: Human Capital Variables ComDonent Eiaenvalue Cumulative R’

P I P2

1.64 0.82 0.36 1 .oo

4

Loadings P I P2 mort -0.71 0.71 h l -0.71 0.71

HUM=O. 821 1 *Pl +O. 1 789 *P2

Table A2.4: Structural Reform Variables ComDonent Einenvalue Cumulative RL

P2 0.45 1 .oo

Loadings P I P2 creditpvt 0.71 0.71 tradegdp 0.71 -0.71

SR=Pl

5

ANNEX 1.3 : OUTSTANDING STATISTICAL CHALLENGES

There is an increasing awareness of the importance of good quality statistics among government officials. In 2002 the authorities created the Directorate for Statistics and Demographic Studies (DISEP). In February 2004, the authorities indicated their intention to participate in the General Data Dissemination System (GDDS). The ministry o f finance’s public Web site (http://www.ministere-finances.dj) was also upgraded, posting quarterly information on the real sector and monthly information on fiscal and monetary accounts (available thorugh the f i rs t quarter o f 2004), monthly price data (available through June 2004)’ and annual data on external debt (available through end-2003).

National Accounts

The existing national accounts estimates follow the 1968 System of National Accounts. The 1998 revision l ed to some improvements in the methodology o f compilation. A supply-side approach, however, remains the mainstay. There continues to be a focus on main GDP aggregates, with demand-side GDP aggregates available only in nominal terns. The full revision conducted in 1998 s t i l l falls short o f correcting the serious defects in Djibouti’s national accounts arising f rom underlying weaknesses in the basic source data. Laclung f ie ld survey data, national accounts estimates re ly heavily o n administrative data sources. Merchandise trade data, compiled by the National Directorate o f Statistics using customs data, are deemed to significantly underestimate imports and exports and do not fully capture transactions recorded by customs. Unl ike the usual practice in other developing countries with deficient customs data, the DISEP does not consistently adjust these data for undercoverage and valuation problems before using them to compile national accounts. Furthermore, GDP may be significantly underestimated because n o efforts are made to account for the contribution o f the growing informal sector.

The accuracy of national accounts data is still hindered by inadequate collection procedures, a lack of basic data sources, and the lack of coordination between the DISEP and other agencies involved in collecting and compiling national accounts source statistics, with regard to sample frames, concepts and definitions, and processing methods. Currently, national accounts compilation relies heavily o n fixed ratios and crude assumptions with n o clear statistical underpinning. Significant delays in releasing the data and obvious flaws in the compilation o f these data have led the Banque Centrale de Dj ibout i (BCD) to compile two parallel series o f national accounts, which leads to duplication o f efforts and can be a potential source o f confusion for users. With the change o f staff at DISEP, most o f the work done in collaboration with the Afr ican Development Bank (AfDB) has been lost and, as a consequence, the existing estimates have drifted away f rom the AfDB-revised estimates and are inconsistent.

Prices

The consumer price index (CPI) is the only available measure of inflation in Djibouti. The CPI compiled by the French government for French expatriate residents in Dj ibout i was used init ially as a proxy. But this index was not representative o f the country’s consumption basket. In late 1998, the staff began to track Dj ibout i price developments by uti l izing subindices f rom the French CPI that were representative o f Dj ibout i consumption, and weights derived f rom a recent Dj ibout i household survey. The new national CPI, available monthly, i s now considered the official measure o f monthly price developments since April 1999.

6

Fiscal Data

Fiscal data compilation improved over the past few years-data are now available monthly. The coordination unit in the Ministry o f Economy and Finance, established in 1998 to improve data coverage and timeliness, has made progress in fulfilling i t s mandate. Among the encouraging initiatives was the recent computerization o f the treasury accounts, which will help improve the coverage, quality, timeliness, and consistency o f cash management data. Another positive step i s the improvement o f data on domestic budgetary arrears through the comprehensive audit o f these arrears, which was finalized in 2002 with assistance f rom the European Un ion and the Wor ld Bank. Foreign-financed capital expenditure and some foreign-financed current expenditures and their financing have been reported in the budget since 1999.

Monetary accounts

Monetary statistics are generally adequate. T o improve sectorization o f monetary data, commercial banks were instructed in July 1999 to initiate a program for accurately identifying the residency status o f customers and to report balance sheet data to the B C D beginning with the end- December 1999 reporting period. Monetary statistics cover the central bank, two operating commercial banks, and one bank in liquidation.

External sector

The improvement in the database on external debt, launched in 2002, was completed in 2004, leading to a downward revision o f the outstanding stock o f debt by about US$40 mi l l ion. New balance-of-payments data, in l ine with the 5& edition o f the Balance of Payments Manual, were recently completed. The new dataset incorporates improvements involv ing transit trade to neighboring countries, imports by foreign mi l i tary forces stationed in Djibouti, and the treatment o f rents paid by the United States for i t s mil itary base. There i s a renewed urgency to review the balance-of-payments data. The revised current account balance i s now showing surpluses for most years since 199 1, mainly because o f revised data on imports o f goods and services. The new presentation separates for the f i rs t time private capital f lows from errors and omissions, which are s t i l l fairly large. T o strengthen external debt management, the authorities launched in 2002 a project to improve the database on external debt, with assistance from the Wor ld Bank and the United Nations Conference o n Trade and Development. The new debt management system became operational in 2003. The revised debt statistics are based on information received from creditors and covers the period f rom 1999 onward. As a result, the outstanding stock o f public and publicly guaranteed external debt (including arrears) was revised downwards at the end of 2003 by about US$41 mill ion. T h i s revision can be decomposed into an increase o f US$21.5 mi l l ion o f public enterprise debt and a decrease o f US$62.7 mi l l ion o f central government debt.

7

ANNEX 2.1 : GOVERNMENT WAGE AND EMPLOYMENT MODEL

Pnrnary Teachen (BI)

Consultant Sub Total

Table

L 8

Consultants

9

ANNEX 2.2: SOCIAL IMPACT OF A REDUCTION I N PUBLIC SECTOR NOMINAL WAGES

Table A 2.2.1 Djibouti: Poverty Impact of 10% Cut in Public Sector Nominal Wages (Headcount Ratio)

Householdsa

Pre-Reduction Post-Reduction Poverty Headcount Poverty Head Count

(YO of population) Extreme National Extreme National (% of population)

All Djibouti

Main Wage Source Non-public sector public secto?

Location Urban Rural

Household Head Male (77.4%) Female (22.6%)

Overall Income group Poorest 2nd quartile 3rd quartile Richest

Male Head: Income group Poorest 2nd quartile 3rd quartile Richest

Female Head: Income group Poorest 2nd quartile 3rd quartile Richest

13.4

14.8 8.9

8.2 39.5

11.9 18.6

53.6 0.0 0.0 0.0

50.7 0.0 0.0 0.0

61.3 0.0 0.0 0.0

48.7 13.8

51.6 14.8 39.6 10.5

42.1 8.4 82.1 40.6

46.3 12.3 57.1 18.7

100.0 55.1 95.0 0.0 0.0 0.0 0.0 0.0

100.0 52.5 95.5 0.0 0.0 0.0 0.0 0.0

100.0 62.0 93.5 0.0 0.0 0.0 0.0 0.0

Gini Coefficient' 40.8 40.5

49.7

51.6 43.6

43.1 82.9

47.3 57.9

100.0 95.6 3.2 0.0

100.0 96.4 3.2 0.0

100.0 93.5 3.1 0.0

Percentage Point Change

Extreme National

0.4

0.0 1.6

0.3 1 .o

0.4 0.2

1.5 0.0 0.0 0.0

1.9 0.0 0.0 0.0

0.6 0.0 0.0 0.0

1 .o

0.0 4.0

1 .o 0.8

1 .o 0.7

0.0 0.7 3.2 0.0

0.0 0.9 3.2 0.0

0.0 0.0 3.1 0.0

-0.3

Extreme per capita poverty line is Fdj 100,229 (USD 1.55/day) while the National(abso1ute) per capita poverty Line is Fdj 216,450 (USD 3.35/day) in 1996. a/ Households were ranked by adult equivalent expenditure.

bl Wage source by public sector is defined as households where all wage income is derived from employment in the public sector.

c/ The Gini Coefficient can take the value ranging from 0 to 100. The higher the value the greater the degree of inequality. Source: Authors' calculations using EDAM 1996

10

Table A 2.2.2 Djibouti: Poverty Impact of 10% Cut in Public Sector Nominal Wages (PovertyGap)

Pre-Reduction Post-Reduction

povertyline) poverty1 i ne) Poverty Gap (% of Poverty Gap(% of

Householdsa Extreme National Extreme National

All Djibouti

Main Wage Source Non-public sector public sectorb

Location Urban Rural

Household Head Male (77.4%) Female (22.6%)

Overall Income group Poorest 2nd quartile 3rd quartile Richest

Male Head: Income group Poorest 2nd quartile 3rd quartile Richest

Female Head: Income group Poorest 2nd quartile 3rd quartile Richest

31.3

32.0 28.1

27.7 35.1

30.7 32.8

31.3 0.0 0.0 0.0

30.7 0.0 0.0 0.0

30.7 0.0 0.0 0.0

38.8

39.7 35.1

34.2 50.4

38.0 40.9

57.3 19.2 0.0 0.0

56.5 19.0 0.0 0.0

56.5 19.0 0.0 0.0

31 .O

32.0 26.5

27.2 35.0

30.3 32.5

31 .O 0.0 0.0 0.0

30.3 0.0 0.0 0.0

30.3 0.0 0.0 0.0

38.9

39.8 35.8

34.4 50.8

38.3 40.7

57.7 20.6 2.4 0.0

57.1 20.5 2.4 0.0

57.1 20.5 2.4 0.0

Percentage Point Change

Extreme National

-0.3

0.0 -1.5

-0.5 -0.2

-0.4 -0.2

-0.3 0.0 0.0 0.0

-0.4 0.0 0.0 0.0

-0.4 0.0 0.0 0.0

0.1

0.1 0.7

0.1 0.3

0.3 -0.1

0.4 1.3 2.4 0.0

0.5 1.5 2.4 0.0

0.5 1.5 2.4 0.0

Notes: Extreme poverty line is Fdj 100,229 (USD 1.55Iday) while the National Poverty Line is Fdj 216,450 (USD 3.35Iday) in 1996. a/ Households were ranked by adult equivalent expenditure. bl Wage source by public sector is defined as households where all wage income is derived from employment in the public sector. Source: Authors' calculations using EDAM 1996

11

ANNEX 3.1 : PUBLIC-PRIVATE WAGE DIFFERENTIAL AND RETURNS TO EDUCATION

Methodology

Studies of public-private wage differentials are based on models of earnings determination developing by Mincer (1974). Variations in earnings are due to differences in human capital, measured by formal schooling and work experience. Other typical controls include location and gender dummies.

Assuming that there are two distinct labor markets: the public sector and formal private sector, denoted as 1 and 2 respectively, the corresponding wage functions are:

where lnWi i s the natural l og o f wages in sector i , pi i s the vector o f coefficients associated with

wage-determining attributes, X and pi i s a normally distributed disturbance term. More specifically the wage equation takes the fol low form:

where Wij i s the monthly earnings (in USD) for individual j in sector i ; PRIM, MIDDLE, SEC, VOC and UNIV are dummy variables indicating the highest level o f education an individual has completed in either primary, middle schooling, secondary, vocational or university; AGE and AGE2 are age and age squared, FEMALE and MARRIED are dummy variables for female workers and married workers and LOC are a set o f location variables covering urban and rural areas5.

The concern with estimating equations (1) or (2) using ordinary least squares (OLS) is that if some unobserved worker's characteristic that determine the wage are correlated with unobserved characteristic that determine the sector of employment, the results will be inconsistent. This i s the selection bias problem that Heckman (1979) highlighted. As Stelcner et.Al (1989) notes, OLS estimation without correcting for this implies that workers are randomly distributed between the public and private sector which i s questionable if wage differential exists. In this case, workers may prefer one sector over the other and an endogenous selection process will determine the assignment o f workers a sector.

~~ ~

W e excluded dummies for nationality because i t was found to be insignificant in subsequent regressions.

12

The equations that determine sector participation -commonly referred to as the switching equations, can b e stated as fo l lows:

I,’ = e,z + E , , i = 1,2 (4) I, = 1 (public sector) if I,’ 2 0

I, = 1 (private sector) if I,* 2 o Ii = 0 , otherwise

(5)

Where I* i s a partially observed index that describes the selection process and the outcome i s observed depending on whether I* i s positive or negative. Vector 2 are variables that are affect the selection process, for example father’s choice o f j o b sector, marriage status or unearned household income from remittances, benefits and rent. OLS estimation o f equation (1) and (2) will provide unbiased estimates o f wage levels only i f E, and E , are uncorrelated with p, and p,, respectively. If unobserved preferences or traits influence the selection process together with the usual wage determinants, this assumption i s violated and wage comparisons based on OLS are inconsistent .

Equations (4) and (5) summarize a selection process that involves two steps - a worker deciding on whether to obtain a public sector job and the employer determining if the person is suitable for the job. The worker compares the expected benefits with the cost o f applying and the employer uses the characteristics o f the applicants to select employee from a queue. Workers seelung a public wage offer and jo in ing the public sector are observed, however we cannot differentiate between a worker who chose not to work in the public sector f rom a worker who did but was not accepted by the government employer. In the overall selection process, the difference in wages wil l affect the worker’s choice while the applicant’s personal characteristic will influence the employer’s choice - hence a combination o f variables that determine wages and other personal characteristics are used to form vector Z, which determines the selection process.

We adopt a two-stage estimation method where in thefirst stage, probit equations are estimated to determine variables that affect the probability of working in the public sector and private sector. A selection term, Ai (or the inverse Mills ratio) i s constructed which i s added to each wage equation. This allows the earnings regression to be estimated consistently using OLS, talung the fol low form,

For comparison purpose, we estimate wage regression for both sectors, with and without the selection term.

The next section explores factors that affect an individual’s decision to participate in public or private sector employment. The characteristics that affect participation are then used to adjust the outcome o f earnings regression for each sector (ie. the selectivity adjusted approach) in the subsequent section. The results o f the regression are used to determine if there i s a public sector wage premium after controlling for human capital endowment and differences in private returns to education across sectors.

13

Estimation results

Labor market participation equation

This sub-section determines the extent to which factors such as education attainment, experience, gender and parent's occupation influence sectoral choices of workers. Participation equations are estimated using the probit model with the outcomes shown in Table A 3.1.1. The results provide the marginal effects or the probability o f each variable o f jo in ing the public or formal private sector calculated at the mean values o f the variables. For comparison purposes, the participation choice for self-employed workers i s also included.

The relationship between age and labor force participation is non-linear across all three sectors. For workers in the public sector, the probability of participation increases until an individual is 42.3 years and then starts decreasing. The turning point i s similar for individuals in the private sector (at 40 years) and who are self-employed (at 43.8 years).

Education affects positively participation into public and private sectors and reduces the likelihood of participation in the in formal sector. Workers with university education have a higher probability ofjoining the public sector as opposed to the formal private sector. The older a worker which serves as a proxy for more work experience, the higher the probability o f jo in ing each sector. Females are less likely to participate in the formal sector and are more likely to be in the informal sector.

Having a parent in the public sector raised the probability of finding employment in the public sector and reduce the probability of being in the informal sector. Income effects o n participation, measured by unearned household income i s statistically significant and negative as expected for public and private sector workers since higher non-labor income would reduce the need to seek formal employment.

Variables Effect Dev Effect age 0.011 *** (0.00) age squared/100 female married obtained certificate primary schooling middle schooling secondary schooling vocation training university parents work in public sector parents work in private sector In-unearned income

-0.013 *** (0.00) -0.068 *** (0.01) 0.009 ** (0.00)

0.042 *** (0.01) 0.048 *** (0.02)

0.106 *** (0.02) 0.134 *** (0.03) 0.144 *** (0.03) 0.138 *** (0.04) 0.010 ** (0.00)

-0.013 (0.01) -0.005 *** (0.00)

-0.005 *** (0.00) -0.008 *** (0.00) -0.008 *** (0.00) 0.009 *** (0.00) 0.001 (0.00) -0.009 *** (0.00) 0.039 '*' (0.02) 0.014 *** (0.00) -0.009 ** (0.00) 0.020 *** (0.01) -0.014 '* (0.00) 0.028 *'* (0.01) -0.004 (0.01) 0.016 ** (0.01) -0.014 (0.01) 0.023 +*' (0.01) -0.022 ** (0.00)

-0.003 (0.00) -0.019 *** (0.00) -0.002 (0.00) -0.011 ** (0.00)

-0.002 '** (0.00) -0.001 (0.00)

Pseudo R square 0.33 0.19 0.15 Note: standard errors are in brackets. *** denotes statistical sianificance at 7% level, ** at 5% and * at - 70%. Source: Authors' estimates based on EDAM 7996.

14

Wage equation

A wage regression using the Mincer framework is ofien used to calculate the private returns to education and the degree to which other individual characteristics may influence earnings. The model i s estimated for public and private wage earners using OLS with and without conditioning o n selection choice. Inclusion o f the selection term which i s constructed fi-om the earlier participation equation controls for the fact that workers do not randomly work in either sector allows us to test whether factors influencing the sorting o f workers into these two sectors also affect wage earnings. The independent variables common to each wage regression are dummies variables for rural and urban locations, dummies for completing education at various levels, age, female dummy and a marriage dummy.

Experience exerts a positive influence on wage offers particularly in the private sector. The results are presented in Table A 3.1.2. The Coefficient magnitudes differ for both specifications and indicate the importance o f correcting for selectivity. Referring to the selectivity corrected earnings regressions, experience exerts a positive influence on wage offers while the quadratic term has the expected negative sign. The curvature o f the wage-experience profi le i s steeper in the private sector than in the public sector, which i s consistent with the general pattern in many countries.

Being female has a negative impact on wages in the private sector. The coefficient on female workers in the private sector i s negative and statistically significant implying that female workers face an earnings penalty in the private sector. No clear conclusions could be drawn about females in the public sector given the statistically insignificant o f the female Coefficient though from the data, we observed that for similar profession such as biologist, teachers and administrative agents, males earned more on average.

Marriage has a positive impact on wage offers in the private sector. The coefficients for schooling attainment will be discussed in the next section.

Private sector workers have higher productivity than the average worker in the labor market- by contrast, public workers have below average productivity. The coefficient estimates o f the selection t e r n for individuals were statistically significant, being negative in the public sector and positive in the private sector. The positive coefficient for the private sector indicates that individuals who select jobs in the private sector have higher productivity than the average worker while the negative Coefficient for public sector workers imply that they have lower productivity that the average worker.

15

0.670 *** (0.12) 0.954 *** (0.19) 0.856 *** (0.07) 0.751 *** (0.10 vocational training 0.659 ** (0.13) 0.815 *** (0.18) 0.870 *** (0.06) 0.667 *** (0.19

0.942 ** (0.12) 0.977 *** (0.18) 1.137 *** (0.09) 0.822 *** (0.14 5.603 *** (0.57) 2.747 *** (0.93) 4.376 *** (0.28) 4.164 *** (0.42

-0.274 ** (0.11) 0.318 (0.18) _ - -

Note: Standard errors are in brackets. *** denotes statistical significance at 7% level, ** at 5%, and * at 70%. Source: Authors' estimates based on EDAM 1996.

Public-private wage differential decomposition

Using a simple average of the wage difference between public and private Cformal) workers to determine if there is a wage premium is misleading as it does not account for differences in human capital endowment and other personal or household characteristic that influences earnings. I t i s possible to breakdown various components that contribute to the difference between predicted public and private sector wages using an Oxaca-Blinder decomposition. There are four possible components, including the differential caused by selectivity bias,

premium endowment returns

selection

where bars denotes the mean o f the variables. Subscripts pub refers to the public sector while pvt denotes the private (formal sector). The f i r s t component i s the difference in the base wage (constant term) which i s often understood as the premium or pure rent f rom being in the public sector. The second component i s due to difference in human capital endowment o f the workers. The third i s due to the difference in private returns to the endowments and the final component i s due to the difference in the selection terms.

Table A 3.1.3 shows the decomposition results for Dj ibout i with and without conditioning o n selection. Both methods clearly show that the most important factor determining a positive wage differential in favor of public sector wages is a wage premium or rent. The premium i s much larger when characteristics that determine participation in both sectors are controlled for which i s capture in the second column.

I n Djibouti, public sector workers earn a rent, afier controlling for other wage determinants. Terrell( 1993) found in Hai t i that public administration workers earned a sizeable rent. Similarly, Lindauer and Sabot (1983) also found that the wage premium was the most important determinant o f public-private wage differential in Tanzania. In Djibouti, the positive rent earned by the public sector i s more l ikely due to the legacy o f French colonial rule that paid high wages to the public sector and o f rewarding the patronage and loyalty o f the elite class.

16

Total log mean differential

Components attributable to: Wage premium Human capital endowments Market returns

- Selectivity Corrected

0.038

2.856 -0.065 -1.800 -0.953

Total unexplained differential 0.0635 1.056 Source: Authors' calculations based on wage regressions in Table A 3.1.2.

Private returns to education

Differences in returns to schooling for workers in the public and private sector that may also help explain the queuing for public employment aside from the existence of higher base wages for government work. The results f rom the earnings functions can be used to estimate private rate o f return to each level o f schooling. The private returns to schooling, r t o different levels o f schooling can be computed as follows for each sector, i :

r(prim), = Bil IS,,, r(midd1e) = (B, - B l ) l SMIDDLE r(sec) = (B, - B,)/SsEC r(voc) = (B, - B,)/Svoc r(univ) = (B, - B3) I S,,,

where the length in years o f each schooling level i s given by , SpN-6, SMDDLE = 4 , SsEc=3, SVOC

=2.7 and SuNp2.

Table A3.1.4 below presents the computed private wage returns per year of schooling for different education levels completed in the public and private dformal) sector using results from the previous wage regressions. While the focus i s on the f i r s t two columns which utilizes coefficients f rom the selectivity correct wage equations, the returns are also computed using OLS coefficients for comparison.

17

Notes: Estimates for primary category is the earnings premium over no-formal education. Similaw, middle school category shows earning differential over primary school. Secondary category shows differential over middle school, vocational shows differenti I Source: Authors’ calculations based on wage regressions in Table A 3.1.2

Workers in the public sector in Djibouti earn larger private rates of return to schooling than do private sector workers with secondary, vocational and university education. For wage earners in the private sector, the private return to education reaches it peak at 8.9 percent after completion o f middle schooling, compared to 5.8 percent for those in the public sector. The private return to secondary schooling i s 6.6 per year of schooling for wage earners in the private sector relative to 9.9 percent for those in the public sector. The returns for private sector wage earners with vocational training further drops to 2.2 percent relative to the higher 10.5 percent received by public sector workers with the same qualification. Returns to education for workers with university level education are much higher in the public sector than in the private sector. Their private returns moving from high school to university i s 1.2 percent compared to 13.6 percent received by graduates working in the public sector.

As illustrated earlier, higher private returns do not, however, signal that productivity is higher (and hence better rewarded) in the public sector. The higher returns to education in the public sector are explained by the distorting government pay policies and the legacy o f rewarding educated labor for their patronage and loyalty to the ruling elite and the absence o f viable j o b opportunities for educated labor in the private sector.

Average returns to education in Djibouti can also be examined for all workers and for males and females separately. Separate wage regressions corrected for selectivity i s shown in Table X. Table Y provides results using the working population engaged in various sectors correcting for selectivity bias.

Variables age age squared/100 female dummy married dummy primary school middle school high school vocational training university constant selection term years of education

All 0.095 *** (0.03)

-0.104 *** (0.03) -0.875 *** (0.09) 0.260 *** (0.05) 0.597 *** (0.06) 1.003 *** (0.09) 1.328 *** (0.13) 1.348 *** (0.13) 1.596 ’** (0.14) 2.923 *” (0.67) 0.345 ** (0.17)

- - -

All (years of edu.) 0.103 *** (0.03)

-0.109 *** i0.03j -0.783 ’** (0.09) 0.270 *** (0.05)

- - _ - - _ _ - _ _ _ _ _ - _

0.367 ** (0.16) 0.116 *** (0.01)

Male 0.133 *** 10.03) -0.144 *** (0.04j

0.273 ” (0.12) 0.492 *** (0.07) 0.787 *** (0.10) 1.184 *** (0.14) 1.166 *’* (0.15) 1.500 *** (0.14) 2.177 *** (0.79) 0.377 ** (0.18)

-0.131 (0.08j

0.067 (0.23) 0.908 *** (0.14) 1.533 **’ (0.20) 1.795 *** (0.31) 1.860 *** (0.27) 1.923 *** (0.45) 1.838 (1.55) 0.470 (0.41)

Chi square 2006.10 2027.49 1497.17 779.65 No. of uncensored obs. 201 1 1997 1293 718 Note: Standard errors are in brackets. *** denotes statistical significance at 7 % level, ** at 5%, and *a t 70%. Source: Authors’ estimates based on EDAM 7996.

18

Note 2. Estimates -for primary category is the earnings premium over no--formal education. Similarly, middle school category shows earning differential over primary school. Secondary category shows differential over middle school, vocational shows differential o

].Source: Authors'calculations based on wage regressions from Tables A 3.1.2 andA 3.1.4

Referring to column one, an additional year of schooling once corrected for worker's sector allocation in Djibouti raises earnings by 11.6percent which is higher than the world average of 10percent.6 This figure may be overstated as it does not account for years o f repetition which i s high in Djibouti, particularly for secondary education. When compared to other Af i ican countries such as Ghana, Cote D' Ivoire and Kenya, returns to secondary and university schooling are lower in Dj ibout i (Schultz, 2003). In addition, the foregoing analysis o f public and private wages in Dj ibout i revealed that private returns depend whether the worker i s employed in the public sector or in the private sector. Estimates o n private returns to education should be treated with caution and future work should provide more accurate estimates by sector as wel l as control for selection bias.

Female workers who have completed primary and middle schooling earn relatively higher returns than their male counterparts- by contrast; male workers who have completed secondary schooling and tertiary education have higher returns to education than women. Returns are almost double for females relative to males completing primary and middle school at an annual rate o f 15.1 percent and 15.6 percent respectively. However this reverses with secondary schooling and higher levels o f education. Males receive higher returns on having completed secondary schooling and tertiary education.

See Psacharopoulos (1994) for wor ld averages. 6

19

- e

e U V

e -

e,

f 5 N

$ P x U

e, fi 8 4 e,

M .C

9 a

U E d a2

a2

k

N N

I L

!

I I L

k I I L ! ! e I

I L

e I

111

3 P -0

m

"xB 4 - 9, 3

r -V I 2 2 w m b r n

? a

, ?

e

? e

l

z I a

! a I

z

t! I

! I I a

I ! Y 4

t!

a

3 t!

b

8 c .-

-0 s f -1 W

a . .. a . a

a a a

I I

a f

rn Y

s

rn M

I 2 I a 2 I 2 a 1 a

2 ?

* .!

i t

3 a ? 1 1 a z ? 3

t

a

3 .!

0 m

0 0 . a

e . . . .. . .

h cu R P)

E e ti rn

m m

U w Y

a c c h m L

n 2

a a a

ANNEX 3.3:’ FIELD SURVEY ON WAGES AND PRICES IN DJIBOUTI, ETHIOPIA, AND YEMEN: MARCH 2005

Djibout i i s a small and poor coastal country (population o f less than 1 million) situated between Eritrea, Ethiopia, and Somalia. The country i s resource-poor, with a weak agriculture and manufacturing sector, and imports nearly al l o f i t s consumer and commercial goods. I ts main export partner i s Ethiopia, largely in re-exports. As detailed in the main body o f the CEM, the high cost-price structure found in Dj ibout i affects the country’s competitiveness and the provision o f key goods and services, further compromising growth in a struggling economy.

Against this background, a qualitative survey was carried out to validate official data with supporting qualitative evidence o f high wages and prices in Djibouti; and to conduct informal interviews with local business managers in order to examine barriers facing the private sector. The surveys and interviews were also conducted in Ethiopia and in Yemen, a neighboring country as wel l as a competing country for port services.

Motivation and objectives o f field survey

A f ield survey conducted by the IMF in 1999 (Staff Report, Selected Issues, 2001) revealed that Djiboutian wages, whether public sector or private sector, were very high when measured as a multiple o f wages, compared with i t s major trading partners, Ethiopia and Yemen. In addition to collecting information on wages, prices for common goods among the three countries were also researched, as prices have trended higher in Dj ibout i than in i t s neighbors, due in large part to the fact that Djibouti imports nearly 100 percent o f i t s products.

This annex attempts to update the research performed by the IMF in 1999 and validate whether similar conditions hold. Additionally, detailed interviews were held with the private sector in the field in order to identify commonalities in the challenges faced in attracting and retaining staff, in the face o f potentially less remunerative, but more secure, employment in the public sector.

Methodology

The field survey was carried out between February 26 and March 14, 2005. The wage and price data research was performed as a qualitative field survey, matching sample criteria (age, gender, professional occupation, sector o f employment-formal or informal-and size o f company). A representative sample o f respondents was obtained-for example, hotels and markets-which f e l l into an “average” range for the ~ o u n t r y . ~ The field survey was carried out in collaboration with Wor ld Bank staff as wel l as national staff in the respective capital cities. Additionally, key informant interviews were held with the managers o f three privately held companies in Djibouti, and hotel managers in the three countries. The wages and prices were converted into U.S. dollars using period average exchange rates for 2004.

This annex was prepared by Ingrid Ivins (MNSED) who also conducted the f ield survey on wages and

One country exception is Ethiopia, where we have the results f rom both a “luxury” hotel and averages for p s in Djibouti, Ethiopia and Yemen.

a more typical mid-range hotel.

37

~~

Table A3.3.1: Survey Breakdown by Respondent Category

Djibouti Ethiopia Yemen Managers, large companies 3 2 1

Local businesses (restaurants, supermarkets, stores) 3 3 2 Local entrepreneurs (tailors, vendors, shoe shine, etc) 18 13 20

Nationals 1 6 4

Total 25 24 27 ource: World Bank Survey March 2005

Multiple of Djiboutian

Main findings

-- 10

- 8 ;; P

Wage in formation

Average Monthly

The most striking wage differential between the three countries can be seen in the public wages. Average wages in the public sector were 10 times lower in Ethiopia, and 5 times lower in Yemen. In the private sector, the multiple ranged from 1 to 7, depending on the j o b classification.

-- 4

-- 2

0

Figure A 33.1: Djibouti vs. Comparators Average Monthly Civil Servant Wages (US$)

Djibouti Yemen Ethiopia

Source: World Bank Survey March 2005

Due to survey limitations, i t was not possible to compare precisely the same employment conditions which were surveyed in 1999. However, preliminary calculations (for example, for hotel staff), do indicate that inequities persist, particularly at the higher-skilled level (professional and managerial staff). The private sector wage multiples compared with those in Ethiopia are particularly striking, because the hotel surveyed in Djibouti was classified as rating “three” stars, while the luxury hotel in Addis Ababa (one o f two hotel categories represented) rates as a “premier” hotel and therefore measures several stars above in ranking. Despite the difference in hotel classes, the hotel wages o n average are s t i l l higher in Dj ibout i (see Table A3.3.3 for hotel wage comparisons between the three countries).

In the piece-work rates section o f the survey, both the total cost o f service (cost to the consumer) and the share o f labor in the total cost o f service are presented in Table A3.3.1. T h e share o f labor costs for Djibouti in the total cost o f service are on average higher than the shares for Ethiopia

38

and Yemen, which average about 30 percent o f the total cost, compared with Djibouti which ranges between 30 and 50 percent o f the total cost.

0.5

0.0

Figure A 3.3.2 Djibouti vs Comparators:

~ 3.0

Djibouti Ethmpia Yemen

Hotel cleaning staff 0 carpenters (sW) E carpenters (unskilled)

Source: World Bank Survey March 2005

Price in formation

W h i l e the price differentials comparing common goods in Dj ibout i to Ethiopia and Yemen may not be as striking as the wage differentials, the prices o f basic consumer goods are consistently higher in Djibouti, with Dj ibout i prices as a multiple o f Ethiopian prices ranging between 1 and 6, and ranging between 1 and 5 for Yemen. As mentioned previously, Dj ibout i i s wholly an importing economy, therefore a price-taker o n the global market. However the higher prices charged for consumer goods in Dj ibout i are not justified by the higher average wages in Djibouti, especially as seen in the public and service sectors.

Barriers facing the private sector: country comparisons

Djibouti

As Dj ibout i recovered from c iv i l war and internal instability in the 1990s, the private sector was l e f t aside and undeveloped. This contributed to a development lag in the private sector o f about 10 years as the government was preoccupied with attending to basic services. Interviews with local managers revealed very high local operating costs, double or triple those o f Djibouti’s neighbors, with one manager estimating that his secretary’s salary in Dj ibout i would be the equivalent o f a minister’s salary in neighboring Ethiopia. Adding to the burden o f business owners are the lack o f financial depth and tools, weak tax and investment incentives, and variable tax rates, a l l o f which add up to large disincentives.

A widespread challenge concerns the diff iculty in finding qualified local staff to f i l l administrative, supervisory, or managerial positions. Because Dj ibout i does not have a university, France offers a l imited amount o f funding for scholarships at French universities. However when the newly minted graduates return to Djibouti, many expect to jump straight to a managerial position, despite a lack o f practical experience, or turn to government employment or international companies (these jobs are also seen as more prestigious). For nonprofessional posts, the dearth o f vocational training programs was also cited as a problem. O n e way to address this

39

constraint would be to attract qualified expatriates to return to Djibouti, if only to train local staff on the ground.

Within the private sector, the few international industries include three o i l companies and two banks. There are local branches o f Coca-Cola and Courbeche bottling corporation. Salaries at the international industries are quite competitive and may pay up to twice as much as in the local private sector. Discussions showed that within the o i l companies, salaries are among the highest on the African continent. One manager gave an average salary for h is company o f US$1,688/month (however this figure covers 7 managers, out o f a total staff o f 70).

A looming problem exists within the o i l sector in that the anticipated arrival o f the management o f the port by a Dubai agency i s expected to lead to the closing or downsizing o f the local o i l companies. One o i l company has already relocated and the rest are scheduled to do so by the end o f 2005. Because the new port,has the right to hire international staff (not an option for the local companies), i t may absorb only about 50 newly unemployed staff out o f the estimated 1,500 to 2,000 that will lose their jobs. It remains to be seen whether the local international affiliates will find it financially viable to remain in Djibouti; firing staff i s very expensive, with legal codes favoring the staff, not the employers. One option for the o i l companies i s t o remain in Dj ibout i on a more l imited distribution basis in order to supply the French and American military bases.

Concerning the hospitality and tourism sector, there currently are n o government-sponsored programs to provide assistance or training to upgrade services and quality. The formation o f a type o f hospitality union, including the tourism sector, may be helpful to local entrepreneurs who would have a venue to discuss pressing issues as wel l as raise awareness o f the lack o f high- standard hotels and tourism services in Djibouti. The few attempts that have been made to help this sector were generated by nongovernmental organizations or foreign partners; however, the efforts were neither coordinated nor sustainable. Feedback from interviewees revealed that in recent years, the Ministry o f Tourism has been concerned with transportation issues only.

Dj ibout i i s also in a unique position o f being surrounded by countries that do not speak one o f i t s official languages. Some business owners are starting to develop in-house training programs to teach English to their staff, as the lack o f English-language s lu l ls i s seen as a potential drawback in the expansion of trade relations.

Interviews were held at several local establishments as wel l as at a 3-star hotel in Djibouti, in order to compare the salary structures. The estimated GNI per capita for Dj ibout i reached US$909 in 2003,9, a l o w figure compared to average salaries reported even for non-managerial staff (table A3.3.2). The salaries at the Djiboutian hotel seem generous, compared with the salaries at the Ethiopian hotels (Tables A3.3.3 and A3.3.4), especially the luxury Ethiopian hotel which i s rated up to 3 stars above the Dj ibout i hotel.

A population census for Dj ibout i has not been done since 1983, leading to uncertain population figures. Estimates are based o n extrapolations o f population growth, leading to calculations ranging between 450,000 and 700,000 (the UN estimate for 2003 was 702,000).

40

Source: World Bank Survey March 2005

Ethiopia

Ethiopia faces some o f the same challenges as Djibouti; however, the private sector i s growing, at a modest, though less volatile, rate. The Ethiopian economy, which i s s t i l l in the process o f emerging from a Soviet-style command system, i s s t i l l largely dominated by the government. A privatization program was launched in 1994, but it has been limited to small businesses such as shops and restaurants, and has not turned into an engine o f growth for the economy.

Although many barriers were cited as disincentives for investment in Ethiopia, a factor found to be more o f a constraint than in Dj ibout i i s the issue o f land rights (also a substantial constraint in Yemen). Repatriating land back to original owners i s a slow and highly disputed process. Although a land auction does exist for private investors, i t i s very expensive and there i s not enough supply to meet demand. Access to capital i s also a challenge, as banks suffer f rom a high percentage o f NPLs (about 35 percent) and require high collateral or guarantees in order to lend.

In contrast to Djibouti, many higher education institutions do exist in Ethiopia, mostly in Addis. However the universities are often seen as old-fashioned and outdated. Most university graduates enter the public sector workforce or international organizations after graduation. Salaries in the private sector are often higher than in the public sector, but as elsewhere, the public sector i s seen as a safe place to gain employment, with ensured tenure and benefits.

The salary structures for two hotels in Ethiopia were examined: the f irst i s a 5-star hotel and the second concerns averages for a typical mid-range hotel. Although the hotel salaries are high for Ethiopia, especially the luxury hotel (GNI per capita was about US$lOO in 2003), as mentioned previously, total compensation for the luxury hotel s t i l l compares favorably with that at the Djiboutian hotel. It should be noted; however, that the salary structure at the Ethiopian luxury hotel depends highly on the service charge o f 10 percent added to every bill, here called variable salary, which i s distributed evenly to al l employees. The average variable salary in the table represents an annual average; however the variable salary can fluctuate seasonally. For most staff, excluding managerial staff, the variable salary i s quite a bit higher than the fixed salary, though it does provide an incentive for a l l hotel staff to provide service at the highest standards. The fixed salary starts at a level slightly above the official minimum wage in Ethiopia. The salaries o f a

41

mid-range Ethiopian hotel, as seen in table A3.3.4, are considerably lower than in the luxury hotel, although they start to converge slightly at the professional level.

Table A3.3.3: Salary Survey, “Luxury” Ethiopian hotel Ave. fixed Ave. variable salary/month, salary/month,

Position US$ US$ Total Housekeeper 40 196 236 Reception staff 92 196 288 Back office staff 115 196 311 Mgr, Food Services 138 196 334 Technical Staff 161 196 357 Guest Services Mgr 167 196 363 Sales Mgr 184 196 380 Department Head 403 196 599 I Top Management 778 196 974

Source: World Bank Survey March 2005

Table A3.3.4: Salary Survey, “Mid-range” Ethiopian hotel Monthly salary

Position (US$) Housekeeping Staff 35 WaiterMlaitress Staff 46 Reservation/Reception Desk 69 ProfessionaVBack-office staff 248 Manager, Food Services 230 Manager (general) 46 1

Source: World Bank Survey March 2005

Although Ethiopia shares some o f the same problems in Dj ibout i in stimulating growth in the private sector, i t has shown that with government investment programs, such as the new, modem airport, plus educational and microcredit opportunities, many o f these obstacles can be overcome.

Yemen

Yemen i s facing i t s own growth crisis, however for reasons dissimilar to Dj ibout i and Ethiopia: the country i s running out o f oil, traditionally a major export. Private investment, which jumped after unification, has basically stagnated in recent years. Yemen faces many o f the same barriers as Dj ibout i and Ethiopia; however, i t also has advantages, such as a well-functioning port in Aden and an open trade regime. Still, land issues persist, the judicial system i s regarded suspiciously, and high collateral i s needed to borrow, even from Islamic banks. A Private Sector workshop was held in Yemen in February, where investors raised many issues as constraints, including those mentioned above, and added further concerns such as minimal government support for small investors and high operating costs-for example, cit ing the planned removal o f fuel subsidies. Due to these barriers, most businesses are small and family-run, with a very small minority o f large companies being owned by the same family or group o f investors.

A hotel was also interviewed in Sana’a, the capital o f Yemen, to further compare salary differentials between Djibouti’s neighbors. T h e Yemeni hotel i s a 3-star hotel, roughly placing it somewhere between the hotel interviewed in Dj ibout i and the luxury hotel interviewed in Ethiopia. As seen in table A3.3.5, salaries are fairly l o w compared with the hotel surveyed in

42

Djibouti. However, they are s t i l l competitive when considering Yemen’s annual GNI per capita o f ~ ~ $ 5 3 0 .

Position Bellbov

Ave. saIary/month, US$ 75

Waiter Doorman Steward Head Waiter Cook Asst. M r., Restaurant 250

Source: World Bank Survey March 2005

Conclusion

The findings o f the qualitative surveys conducted in Djibouti, Ethiopia, and Yemen confirm that higher labor costs and the paucity o f slul led workers constitute a major constraint to private sector development in Djibouti. The field survey validates official statistics on relative wages and prices in the three countries. Finally, the interviews with the managers o f local f i r m s in Dj ibout i also revealed that rigid labor market regulations remain a significant barrier to private investment.

43

HousekeeDina S ta f f 2531 351 1501 7.31 1.7 Waiter (hotel) Reservation/Reception Desk ProfessionaVBack-office s ta f f

13.Piocc-workatos

321 46 100 7.0 3.2 467 69 6.8 731 248 3.0

I I

Manager, Food Services (hotel) Manager (general)

Unweiahted averaae

I a) Total Cost of Service (US$) I

230 250 1576 46 1 3.4 670 181 167 3.7 4.0

I I Making pair of pants Making shir t Makina simDle skir t

17 7 8 2.4 2.1 11 3 3 3.3 3.5 17 6 2 2.9 8.9

Making simple wooden bed 253 52 271 4.9 0.9 Making simple wooden chair 84 7 27 12.2 3.1 ReDair of one t i r e Duncture 0.58 1.62 Washing one medium-size car Haircut (men) Shoeshine [with Dolish)

0.58 1.08 11 0.81 1.62 14.0 6.9

0.28 0.12 0.27 2.4 1.0 Shoe repair (replace two heels) Unweighted average

1/ Mid-range hotel. 2/ Assuming 26 working days per average month

1.13 0.35 0.54 3.3 2.1 49 8 32 6.4 1.6

44

US$ per item

Cows milk I l i ter I 2.251 1.561 0.971

Djibouti prices as multiple of:

1.41 2.3

Product I Quantity Djibouti Ethiopia Yemen I 1 Ethiopia Yemen

Mangoes I 1k9 I 2.251 0.351 I 6.51

12. Fruit & veoetables I I I I I I I I

Bananas 1 1ka 1 1.131 0.291

I o n s

3.91

Unweighted average 5. Restaurant meals

One serving of tea One serving of coffee Plate of rice (no meat) Plate of Dasta

1.23 1.32 0.52 0.9 2.4

0.12 0.05 1.69 0.17 0.05 9.8 31.2

0.40 0.38 0.46 1.08

Unweighted average

45

0.29 0.39 6. Recreational & other services

Ticket for football match 1 ticket

Movie ticket 1 ticket Video rental per day

1.73 1.62 0.35 1.08 1.15 1.08

Annex Table 3: Comparison of Hotel Salaries by Position Hotel Average Monthly Salaries (US$) Diibouti Yemen Ethiopia Ethiopia

“luxury” hotel “mid-range’’ hotel

Housekeeping Staff 253 150 236 35 WaiterNVaitress Staff 32 1 100 46 Reservation/Reception Desk 467 288 69 ProfessionaVBack-office staff 73 1 31 1 248 Manager, Food Services 250 334 230 Manager (general) 1576 579 46 1 Source: World Bank Survey March 2005

46

ANNEX 5.1 : PORT OF DJIBOUTI DATA"

Note: Approximately 80 percent of the petroleum products throughput is transit to Ethiopia Source: Djibouti authorities

lo See World Bank Transport Sector Review, 2005, led by Jean-Jacques Crochet (MNSIF)

47

Container throughput Port o f Djibouti 1994 - 2003 (TEU) Year I1994 I1995 I1996 I1997 I1998 11999 I2000 I2001 I2002 12003

74708 GRAND TOTAL 83219 105757 148872 136217 129456 127126 147908 178005 243487 Source: Djibouti authorities

48

ANNEX 5.2: ROAD CONDITION OF THE PAVED NETWORK

Road Length Number (km) Roads Condition

International Road Corridor Djibouti Port - Balbala

I Balbala - Arta Turnoff l R N l 124 1 Good I

241 RN1 10 Very Good

Arta Turnoff - Doubalala Doubalala - Dikhil (North Corridor)

I Dikhil - Galafi (North Corridor) l R N l 196 1 VervGood I

RN1 51 Fair R N 1 38 Good

Doubalala - Ali Sabieh (South Corridor) Ali Sabieh - GuelilC ( South Corridor)

I Other National Paved Roads I I 189 I I

RN 5 13 Fair RN 19 9 Very Good

~

Djibouti - Douda R N 2 8 Fair Route de Dorale (RN 1 - Dorale) R N 3 7 Bad

I Route d 'Arta (RN 1 - Arta) I R N 4 1 8 I VervGood I Route de 1 'Unit6 (RN 1 - Assa Hougoub) Route de 1 'Unit6 (Assa Hougoub - Tadjoura) Route Lac Assal (RN 9 - Lake Assal)

RN 9 113 Good

R N 9 9 Fair

RN 10 16 Fair I Route de Randa (RN 9 - Randa) I RN11 I 2 5 I Bad I

Route de Mandela

TOTAL

RN 17 2 Very Good

430 (*) excluding the urban road network

Source: MPWH- BCEOM

49

ANNEX 5.3: PORT OF DJIBOUTI’S WAGE BILL AND ACCOUNTS

Total Government & Public Enterprise Wage Bill Total Public Enterprise Wage Bill

Percentage change Port of Djibouti

Percent of total-Public Enterprise Percent of total-Govt & Public Ent. Percent of GDP

1999 2000 2001 19,411 20,798 21,208 4,717 6,189 6,404

31.2 3.5 1,932 2,232 2,342 41.0 36.1 36.6 10.0 10.7 11.0 2.0 2.3 2.3

2002

Value added Wages and salaries Taxes

2 1,644 6,923 8.1 2,573 37.2 11.9 2.4

1993 1996 1997 1998 1999 2000 69.2 73.9 72.6 75.4 80.9 80.6 40.9 48.8 43.9 31.2 27.9 29.3 0.1 0.1 1.9 1.2 1.8 2

I GDP at current Drices (in millions of DE”) 195,273 98,267 101,870 105,210 1 Source: Djibouti authorities and company accounts

Source: Djibouti authorities and company accounts

50

ANNEX 5.4: ESTIMATION OF THE IMPACT OF TRAFFIC IN TRANSIT ON EMPLOYMENT

Institution Port Autonome International de Diibouti

1. Direct effect Direct employment o f activities in transit amounts to 4700- 4800 permanent workers” (see Table A 5.4.1) 12.

Employees 1100-1200

Dockers employee^'^ Employees working in transit-related and other ancillary a~tivities’~ Chemin de Fer Djibouto-Ethiopien Ministrv of TransDort and Eaui~ment”

1000

1500

500 400

Source: Le Brishoual & Khelif (2004)

2. Indirect effect Indirect effects on employment are diff icult t o assess owing to lack o f data on multiplier effects on employment. L e Brishoual and Khelif (2004) estimate indirect jobs in a l o w case scenario o f 1500 persons and a high case o f 5000 personsI6.

Table A.5.4.2: Indirect Employment Resulting from the Traffic in Transit I Services I Indirect EmDlovment

Port-related services Facilities building, shops, restaurants, hotels, banking and financial institutions Road building, shops, restaurants, banking and financial system Transport-related services

Source: Le Brishoual & Khelif (2004)

3. Induced effect

Some 6,000 to 9,000 people are employed due to the induced effects o f traffic in transit. Port activities have potentially large induced effects on local employment”. Given the lack of

l1 Estimates by K h e l i f and L e Brishoual. Certain activities are clearly l inked to the volume o f traffic in transit (handling, ship services, freight forwarding activities). I t remains more complex to identify public sector employment generated by the volume o f traffic in transit. l2 I t differs f rom the estimates o f the employment o f transport sector as a whole. Indeed, estimates o f the employment o f urban and interurban vary between 4,000-5,000 permanent workers. 10,000 people work in the transport sector (port o r land transport) in Djibouti. If we add temporary workers, the figure goes upt to 12,000 l3 Dock labor i s hired on a dai ly basis, depending o n the dai ly requirements (Dock Labor Pool). The Pool has some 3,000 registered dock workers o n record, but o n average only 1,000 i s required per day. Workers report for work in the morning at the port gate. If employed, they are paid, in cash, for that particular day. Most port workers are unskil led and untrained; they obtain skills through experience. l4 Such as ship and cargo-related services such as freight forwarding but also ship repairs and ship agent. l5 An important number o f employees in the Ministry manage road user charges and the road find. l6 A conservative figure would range between 7000 and 8000 workers for combined direct and indirect jobs, a result o f port and transit-related activities.

51

available data in Djibouti, we have calculated a low-case scenario using the lowest induced coefficients calculated for French ports. The induced coefficients are between 0.75 and 1''. The number o f people employed due to induced effects o f traffic in transit could represent 6000 to 8000 persons. This figure i s probably a minimal estimate because, in developing countries, salaries have usually higher induced effects than in developed countries, and the induced coefficient calculated in France was limited to port activities and did not take land transport into account.

l7 They are function o f the revenue-effect of the salaries paid to workers working directly or indirectly for the port (or land transport in the case o f Djibouti).

Grosdidier de Matons (1999: 613).

52

ANNEX 5.5: REFERENCES

Acemoglu, D. 2003. “Root Causes: A Historical Approach to Assessing the Role o f Institutions in Economic Development.” Finance and Development (June).

Alesina, A. E. Glaeser, and B. Sacerdote. 2001. “Why Doesn’t the U S have a European-Style Welfare System?”, National Bureau o f Economic Research (NBER) Working Paper No. 8524

Anos Casero, P., and A. Varoudakis. 2004. “Growth Private Investment and the Cost f Doing Business in Tunisia,” MENA Workmg Paper Series No. 34, World Bank, Washington, DC.

Balino, T., and A. Ubide. 1999. “The Korean Financial Crisis o f 1997 - A Survey o f Financial Sector Reform”, International Monetary Fund, IMF Working Paper WPl99128.

Batra, G., D. Kaufmann, and H. Andrew. 2003. “Investment Climate Around the World: Voices o f Firm from the World Business Enviroment Survey”. World Bank. Washington D.C.

Bisat, A., M. El-Erian, and T. Helbling. 1997. “Restoring the Boom-Growth, Investment, and Savings in the Arab Economies.” International Monetary Fund, IMF Working Paper ~ ~ 1 9 7 1 8 5 .

Borensztein, E., J. Gregorio, and J. Lee. 1998. “How Does Foreign Direct Investment Affect Economic Growth?” Journal of International Economics 45 (1): 1 15-35.

Djibouti. Direction Nationale de la Statistique. 199 1. “Annuaire Statistique de Djibouti.” RCsultats de 1991.

. 1999 “Comptes Nationaux de Djibouti, 1990-1998.”

. 2000. “Annuaire Statistique de Djibouti.” RCsultats de 2000.

, 2002. “Secteur informel: Contraintes fiscales et institutionnelles.”

Diwan, I. 1994. “Public Sector Retrenchment and Severance Pay: Nine Propositions.” In “Civil Service Reform in Latin America and the Caribbean”, ed. S. Chaudhry, G. Reid, and W. Malik. World Bank Technical Paper No. 259, World Bank, Washington, DC.

Dridi, J., and A. Toure, 2003, “Growth and Employment in Djibouti”, mimeo, International Monetary Fund.

Favaro, E. 2005, “A guide to Economic Growth Analysis”, mimeo, World Bank.

Grosdidier de Matons, J. 1999, “Droit, Economie et Finances Portuaires” , Paris : Presses de 1’Ecole Nationale des Ponts et ChaussCes.

Haltiwanger, J. and MSingh. 1999. “Cross-country evidence on public sector retrenchment”. The World Bank Economic Review, 13(1): 23-66

Heckman, J. 1979. “Sample selection bias as a specification error.” Econometrica 47: 53-161.

53

International Monetary Fund. 200 1 “Djibouti: Staff Report - Selected Issues”.

Khan, M. and M. Kumar. 1997. “Public and Private Investment and the Growth Process in Developing Countries”, Oxford Bulletin of Economics and Statistics 59( 1): 69-88

King, R., and R. Levine. 1993. “Finance and Growth: Schumpeter Might be Right.” The Quarterly Journal of Economics 108 (3): 717-37. L e Brishoual, M. and Khelif. .2004., “Rapport de Mission pour 1’Etude Diagnostique du Cadre IntCgrt.”

Levine, R. 1999. “Financial Development and Economic Growth: Views and Agenda”, World Bank, Policy Research Workmg Paper Series: 1678 1999

Lindauer, D., and R. Sabot. 1983. “The Public-Private Wage Differential in a Poor Urban Economy.” Journal of Development Economics 12 (3): 137-57.

Marshall, M. and K. Jaggers. 2000. “Polity IV Project: Data Users Manual P4v2000”.College Park: INSCWCICDM

Mincer, J., 1974, “Progress in Human Capital Analysis o f the Distribution o f Earnings”, National Bureau o f Economic Research, (NBER) Working Papers 53.

Nabli, M., and M. Veganzones-Varoudahs. 2004. “Reforms and Growth in MENA Countries: New Empirical Evidence.” MENA Working Paper Series No. 36, World Bank, Washington, DC.

Orac, Jane. 2001. “Fiscal Impacts o f Reform Strategies for Russia’s State Service: A Pay and Employment for the Executive Branch.” World Bank mimeo.

Pinto, B. and A. Joshua. 2004. “Managing Volatility and Crises: A Practioner’s Guide Overview”, National Bureau o f Economic Research (NBER) Workmg Paper No. 10602

Psacharopoulos, George. 1994. “Returns to Investment in Education: A Global Update.” World Development 22 (9): 132543.

Psacharopoulos, George. 1996. “RR Manual: A Program to Estimate the Rate o f Return to Investment in Education.” World Bank, Washington, DC.

Rama, M. 1997. “Efficient Public Sector Downsizing.” World Bank Policy Research Working Paper No. 1840, World Bank, Washington DC.

Rama, M. and D. MacIsaac, 1999. “Earnings and welfare after downsizing: Central Bank Employees in Ecuador”. The World Bank Economic Review, 13( 1): 89-1 16

Rodrik, D., R. Haussman, and A. Velasco. 2004. “Growth Diagnostics.” Kennedy School o f Government, Harvar d University .

Ruppert, E. (2000), “Rationalizing Public Sector Employment in the MENA Region”, Middle East and North Africa, Chief Economist Office, World Bank Working Paper Series, No. 19

54

Schultz, P. 2003. “Evidence o f Returns to Schooling in A h c a From Household Surveys: Monitoring and Restructuring the Market for Education.” Yale Economic Growth Center Discussion Paper No. 875, Yale University.

Serven, L. and A. Solimano. 1992, “Private Investment and Macroeconomic Adjustment: A Survey” Wor ld Bank Research Observer 7(1): 95-1 14

Stelcner, M., J. van der Gaag, and W. Vijverberg. 1989. “A Switching Regression Model o f Public-Private Sector Wage Differentials in Peru: 1985-1986.” Journal of Human Resources 24 (3): 545-59.

Tansil, A. 1999. “Public-Private Employment Choice, Wage Differentials and Gender in Turkey.” Yale Economic Growth Center Discussion Paper No. 797, Yale University.

Terrell, K. 1993. “Public-private wage differential in Haiti.” Journal of Development Economics 42: 293-314.

USAID. 2001. “Republic o f Djibouti: Investor’s Road Map”, United States Agency for International Development.

Wor ld Bank .2004. “Djibouti Public Expenditure Review”, Wor ld Bank, Washington D.C.

.2005. “Djibouti: Transports Sector Review”, Finance, h v a t e Sector, and Infrastructure Department, Middle East and North A h c a Region, Wor ld Bank, Washington DC.

,2005. “Doing business in 2005: removing obstacles to growth”, Wor ld Bank, Washington D.C.

55