International Monetary Fund Washington, D.C.

Transcription

1 2004 Internatonal Monetary Fund June 2004 IMF Country Report No. 04/64 Aprl 3, 2004 January 29, 200 January 29, 200 February 24, 2004 January 29, 200 Morocco: Selected Issues Ths Selected Issues paper for Morocco was prepared by a staff team of the Internatonal Monetary Fund as background documentaton for the perodc consultaton wth the member country. It s based on the nformaton avalable at the tme t was completed on Aprl 2, The vews expressed n ths document are those of the staff team and do not necessarly reflect the vews of the government of Morocco or the Executve Board of the IMF. The polcy of publcaton of staff reports and other documents by the IMF allows for the deleton of market-senstve nformaton. To assst the IMF n evaluatng the publcaton polcy, reader comments are nvted and may be sent by e-mal to publcatonpolcy@mf.org. Copes of ths report are avalable to the publc from Internatonal Monetary Fund Publcaton Servces th Street, N.W. Washngton, D.C Telephone: (202) Telefax: (202) E-mal: publcatons@mf.org Internet: Prce: $5.00 a copy Internatonal Monetary Fund Washngton, D.C.

2

3 INTERNATIONAL MONETARY FUND MOROCCO Selected Issues Prepared by Abdourahmane Sarr, Ludvg Söderlng, and Jacques Bouhga-Hagbe (all MCD) Approved by Mddle East and Central Asa Department Aprl 2, 2004 Contents Page Overvew...3 I. Assessng the Sustanablty of Publc Debt n Morocco...4 A. Summary...4 B. Introducton...4 C. Publc Debt n Morocco...6 D. The Need for Fscal Adustment n the Medum Term...5 E. Concluson...8 Boxes. Publc Debt Management... Tables. Publc Sector Debt Sustanablty Framework, Fscal Polcy Reacton Functon, Publc Sector Debt Sustanablty Framework, Fgures. Prmary Balances and Central Government Debt Characterstcs of External Debt Characterstcs of Domestc Debt Bankng System Balance Sheet and Internatonal Comparson Debt Vulnerablty Factors n Emergng Market Countres Comparson of Overborrowng Debt to GDP rato Under Alternatve Fscal Scenaros...8 Appendx. Publc Debt Dynamcs Equaton...9 Reference...20

4 - 2 - II. Impact of the Barcelona Process on Morocco s Trade...2 A. Introducton...2 B. Development n Trade Patterns Snce C. Concluson...27 Appendx II Tables. Estmated Results from the Gravty Model, Excludng Regonal Trade Dummes Gravty Model Estmatons wth Regonal Trade Dummes...30 Fgures. Trade Creaton Effects wth the EU Estmated Regonal Trade Effects...3 III. On the Long-Term Determnants of Workers Remttances n Morocco...32 A. Introducton...32 B. Workers Remttances n Morocco...33 C. Potental Long-Run Determnants of Workers Remttances...35 D. Concluson and Polcy Implcatons...46 Tables. Balance of Payments, ADF (4) Statstcs for Unt Root Tests Resdual Msspecfcaton Test Statstcs: Frst Order Test Statstcs for Co-ntegratng Rank Unrestrcted Estmates of the Co-Integratng Vectors and Adustments Coeffcents Error Correcton or Short Run Model...64 Fgures. Workers Remttances Housng Prces and Effectve Exchange Rate Potental Long-Run Determnants of Workers Remttances Devatons of Remttances from the Long-Run Value as Defned Evoluton of Transfers and Some of Ther Potental Long-Run Determnants Transfers and Total Deposts and Ther Ftted Values Remttances Growth Rates and Ther Ftted Values (Short Run) Error Correcton Model Resduals and One-up Chow Test...65 Appendces 3. The Mcroeconomcs of Workers Remttances A Summary of the Econometrc Detals...58 References...66

5 - 3 - Overvew. Ths selected ssues paper ams to contrbute to the analyses of recent developments n the Moroccan economy and ts polcy challenges over the medum term. 2. Chapter I assesses the sustanablty of publc debt n Morocco. At 68 percent of GDP, government debt s a source of vulnerablty and could constran growth by breedng expectatons of hgher taxes and nterest rates. The paper uses the analytcal framework for assessng debt sustanablty n emergng market countres endorsed by the Executve Board and compares Morocco s vulnerabltes to those of the average emergng market country. The analyss s complemented by a revew of the characterstcs of government debt and the nsttutonal framework for publc debt management. The paper concludes that the characterstcs of government debt compare favorably wth those of an average emergng market country and that the current level of debt would appear sustanable udgng by the average fscal performance of the last 0 years. The fscal poston has however deterorated n recent years and, unless mproved, would put the debt-to-gdp rato on an unsustanable path wth no reslence to shocks. A front loaded fscal adustment amed at reducng the debt-to-gdp rato below 60 percent would enhance the countres credblty vs-à-vs domestc and foregn nvestors and would be an mportant component of a hgh growth strategy. 3. Chapter II studes the effects on Morocco s trade pattern of the ongong ntegraton wth the European Unon (EU) wthn the Barcelona process a key component of Morocco s strategy to ncrease growth and reduce poverty over the medum term. Morocco s one of the frst Medterranean countres to have sgned and mplemented a blateral Assocaton Agreement wth the European Unon (AAEU), callng for a gradual elmnaton of tarffs on non-agrcultural goods. Usng a gravty model of blateral trade, ths chapter addresses three man questons: () has the Barcelona process resulted n trade creaton vs-à-vs the EU? () s there any evdence of trade dverson? and () s there any mpact on ntra-regonal trade? The results ndcate that Morocco has ndeed benefted from trade creaton wth the EU, especally on the mport sde. Moreover, trade dverson does not appear to be a maor concern, whle the desred strengthenng of ntra-regonal trade has yet to take hold. These results would encourage the Moroccan authortes to contnue, or even accelerate, ther ntegraton efforts wth the EU. Whle accompanyng broader multlateral lberalzaton would be desrable, the apparent lack of trade dverson suggests that ths s not a precondton to deepen ntegraton wth the EU. 4. Chapter III examnes emprcally the long-run relaton between workers remttances and ther varous potental determnants. Transfers from Moroccans lvng abroad play an mportant role n the current strength of Morocco s external poston and have a sgnfcant mpact on the conduct of monetary polcy. Thus, any analyss of the long-term sustanablty of Morocco s external poston requres a good understandng of the behavor of workers remttances and ther long-run determnants. The fndngs of the emprcal analyss suggest that altrusm or soldarty, attachment to homeland, and economc growth n the countres where remttances orgnate could ndeed be the man long-run determnants of workers remttances n Morocco. There s no evdence that portfolo dversfcaton motves are sgnfcant among the long-run explanatory factors of provdng workers remttances to the home country. These fndngs suggest that the rsks of a sudden end or reversal of transfers from Moroccans lvng abroad are lmted n a foreseeable future. The evdence also ndcates that Morocco could further take advantage of ts potental as an nvestment destnaton n general and for ts ctzens lvng abroad n partcular. The acceleraton of the authortes broad-based structural reform agenda would support the achevement of ths goal.

6 - 4 - I. ASSESSING THE SUSTAINABILITY OF PUBLIC DEBT IN MOROCCO A. Summary 5. Morocco s publc debt has consderably decreased as a percentage of GDP n recent years. Although good fscal performance contrbuted to the declne n total debt durng most of the 990s, prvatzaton recepts accounted for most of the reducton as the fscal poston deterorated startng n The level and characterstcs of Morocco s publc debt compare favorably wth those of emergng market countres. Actve debt management through debt equty swaps and substtuton of domestc debt for external debt reduced the share of external debt by half over the last ten years. The share of external debt n total debt s low, external debt s mostly at low fxed nterest rates wth long maturtes, and external reserves cover the totalty of external publc and publcly guaranteed debt. Debt management has been modernzed, and domestc debt s reasonably dversfed across maturtes and nvestor bases. In the current envronment of low nflaton and sgnfcant fnancal deepenng, the government should not experence dffcultes meetng ts fnancng needs at reasonable nterest costs n the foreseeable future. 7. The recent developments n the fscal poston, f not reversed, would however put publc debt on an unsustanable path The authortes baselne scenaro assumes fscal adustment efforts n the medum term that would reduce the debt rato to 64 percent of GDP by Stress tests along ths adustment path suggest that ths target could be mssed by a large margn should adverse shocks, such as severe droughts experenced n the past, reoccur. 8. Although Morocco s debt profle s favorable, a front loaded fscal adustment would be desrable to further reduce the debt rato. Such an adustment amed at reducng the debt rato below 60 percent of GDP would ncrease the countres reslency to shocks and further enhance ts credblty vs-à-vs domestc and foregn nvestors. The resultng ncrease n confdence would be an essental component of a hgh growth strategy to reduce poverty and unemployment. B. Introducton 9. A sustanable publc debt s a key component of good macroeconomc management. A publc debt that s too hgh negatvely affects growth because t dscourages prvate nvestment by breedng expectatons of hgher real nterest rates and the assocated debt servce burden could lead to hgher taxaton and reducton of productve fscal expendtures. Moreover, a hgh publc debt reduces macroeconomc reslency by lmtng fscal flexblty to respond to shocks. Whle t can be argued that hgh publc debt ncurred to fnance expendtures that support prvate sector actvty and growth s desrable, emprcal Prepared by Abdourahmane Sarr

7 - 5 - evdence from emergng market countres suggests a negatve correlaton between government debt and growth. Economc theory also offers lttle practcal gudance on the optmalty of a gven debt level because conclusons drawn are n most cases model dependent (IMF 2003a). For these reasons, an analyss of the mplcatons of fscal polces for debt sustanablty s mportant for macroeconomc management. 0. To facltate the comparablty of publc debt sustanablty analyses and to emphasze ther mportance, the IMF Executve Board endorsed an analytcal framework that standardzed these exercses. 2 The analytcal framework allows to spell out assumptons on key macroeconomc varables that affect debt dynamcs (nterest rate, growth, exchange rate, and prmary fscal balance) n a baselne medum-term fscal scenaro. The relatve contrbuton of these varables for debt dynamcs s then hghlghted. The analytcal framework calculates the prmary balances that stablze publc debt ratos under alternatve scenaros and allows to conduct stress tests on the key macroeconomc varables. Stress tests are calbrated on the hstorcal volatlty of the macroeconomc varables to hghlght vulnerabltes. 3. The framework should however be supplemented by an analyss of a country s specfc characterstcs that may have a bearng on the sustanablty of ts debt. Ths s because the sustanablty of a gven debt rato vares from one country to another, and may depend on factors not captured by the key macroeconomc varables. These factors nclude the country s debt management capacty, the characterstcs of ts debt stock and debt market, and other relevant macroeconomc strengths or vulnerabltes. 2. Revewng the characterstcs of emergng market countres, the IMF concluded n a recent study that ther publc debt rato, at an average of 70 percent of GDP, was unsustanable. 4 In assessng the sustanablty of that debt rato, the study answered four basc questons: () Do prmary fscal balances under current fscal polces allow to stablze the countres publc debt ratos?; () Is there a postve correlaton between the countres prmary fscal balances and publc debt that suggests a track record of fscal adustment to keep publc debt on a sustanable path?; () Are the countres debt ratos excessve udgng by the benchmark debt ratos mpled by the real nterest rates on publc debt and by the prmary fscal balances and GDP growth rates the countres have been able to generate?; and (v) Would the countres debt ratos exceed ther sustanable levels n the event of plausble shocks based on experenced volatltes? In answerng these questons, the study recognzed that ts results could not be generalzed to all emergng market countres snce ther ndvdual characterstcs can vary sgnfcantly and the analyss s mostly backward lookng. 2 IMF (2002). 3 Appendx I brefly explans the debt dynamcs equaton used n the framework. 4 (IMF, 2003a).

8 The goal of ths paper s to assess the sustanablty of Morocco s publc debt n lght of ts specfc characterstcs and the methodology used n the IMF (2003a) study. Central government debt n Morocco represented 68 percent of GDP at end 2003 and s close to the average debt level of emergng market countres that the study assessed as unsustanable. 4. The paper s organzed as follows. Secton C descrbes the trends n Morocco s publc debt and fscal performance over the last ten years. Usng the IMF s (2002) debt sustanablty framework, the contrbutons of the key macroeconomc varables that affect debt dynamcs are dscussed. Secton C also descrbes the man characterstcs of Morocco s publc debt and the nsttutonal framework for publc debt management. Answers to the frst three of the four questons dscussed above are analyzed to assess the sustanablty of the country s debt. In ths context, the characterstcs of the Moroccan economy to those of the emergng market countres n the IMF (2003a) study are also compared. Secton D dscusses the mplcatons for publc debt of a baselne scenaro that assumes fscal consoldaton efforts and those of an alternatve scenaro that assumes a contnuaton of the current fscal polcy stance. The mplcatons for publc debt of stress tests (the fourth queston above) are used to draw fnal conclusons on the desrablty to reduce Morocco s publc debt rato. Secton E concludes the paper. C. Publc Debt n Morocco 5. Publc debt has declned sgnfcantly n recent years despte a worsenng n the fscal poston. Between 994 and 2003, total central government debt fell from 88 percent of GDP to 68 percent (Fgure and Table ). The debt rato however remaned hgh for most of the perod despte sgnfcant prmary surpluses. Ths s because low agrcultural output n drought years resulted n sgnfcantly postve real nterest rate/growth dfferentals that offset the debt reducng contrbutons of prmary surpluses. Interest expendtures, although declnng, averaged 5 percent of GDP over the perod contrbutng to the fscal defct whch averaged 3.7 percent of GDP between Snce 2000, the prmary balance has been n defct because expendtures fnanced by prvatzaton recepts, partcularly the wage bll, ncreased whle revenues declned because of ongong trade lberalzaton and new tax exemptons. Whle the fscal poston deterorated, the publc debt rato contnued to declne because of prvatzaton recepts. Prvatzaton recepts amounted to only 3.5 percent of GDP durng compared to over 2 percent of GDP durng

9 - 7 - Fgure. Prmary Balances and Central Government Debt, Prmary Balance (n Percent of GDP) 05 Central Government Debt (n percent of GDP) 8 7 Excl. grants & prvatzaton Incl.grants Incl. grants & prvatzaton Foregn Total Debt Domestc Source: Moroccan authortes and staff estmates External debt vulnerablty also declned consderably n recent years (Fgure 2). The share of external debt n total debt fell by more than half to only 27. percent between 994 and 2003 reflectng the government s polcy of substtutng domestc debt for foregn debt to pay off expensve debt and debt equty swap operatons. 5 Most of the outstandng debt s owed to blateral credtors and nternatonal organzatons at low fxed nterest rates wth an average maturty of years and at a dversfed currency composton. The weghted average nomnal nterest rate on outstandng external debt was only 3.7 percent at end In addton, external reserves covered the totally of publc and publcly guaranteed external debt at end Ths makes Morocco vrtually mmune to the mpact of exchange rate movements on ts net external debt poston (ncludng foregn reserves assets). Furthermore, udgng by the 35 percent of GDP threshold above whch the rsk of a debt crss has been emprcally observed (Renhart, Rogoff, and Savastano, 2003), and the fact that external debt s mostly owed to offcal credtors, Morocco appears sgnfcantly less vulnerable on the external sde. 5 The substtuton of domestc debt for external debt partly explans the ncrease n gross fnancng needs observed n Table n recent years. 6 Contngent clams on reserves are also lmted. Foregn currency deposts and nonresdent deposts convertble nto foregn currency represent less than percent of GDP. Short term external debt s only 6 percent of external debt.

10 - 8 - Table. Morocco: Publc Sector Debt Sustanablty Framework, (In percent of GDP, unless otherwse ndcated) Actual Publc sector debt / o/w foregn-currency denomnated Change n publc sector debt Identfed debt-creatng flows (472) Prmary defct Revenue and grants Prmary (nonnterest) expendture Automatc debt dynamcs 2/ Contrbuton from nterest rate/growth dfferental 3/ Of whch contrbuton from real nterest rate Of whch contrbuton from real GDP growth Contrbuton from exchange rate deprecaton 4/ Other dentfed debt-creatng flows Prvatzaton recepts (negatve) Recognton of mplct or contngent labltes Other (specfy, e.g. bank recaptalzaton) Resdual, ncludng asset changes (2-3) Publc sector debt-to-revenue rato / Gross fnancng need 5/ n bllons of U.S. dollars Year 0-Year Hstorcal Standard Key Macroeconomc and Fscal Varables Average Devaton Real GDP growth (n percent) Average nomnal nterest rate on publc debt (n percent) 6/ Average real nterest rate (nomnal rate mnus change n GDP deflator, Nomnal apprecaton (ncrease n US dollar value of local currency, n Inflaton rate (GDP deflator, n percent) Growth of real prmary spendng (deflated by GDP deflator, n percent) Prmary defct Overall Balance 7/ Interest expendture Source: Moroccan Authortes and Fund Staff estmates / Publc sector refers to central government total gross debt 2/ Derved as [(r - π(g) - g αε(r)]/(gπgπ)) tmes prevous perod debt rato, wth r = nterest rate; π = growth rate of GDP deflator; g = real GDP growth rate; α = share of foregn-currency denomnated debt; and ε = nomnal exchange rate deprecaton (measured by ncrease n local currency value of U.S. dollar). 3/ The real nterest rate contrbuton s derved from the numerator n footnote 2/ as r - π (g) and the real growth contrbuton as -g. 4/ The exchange rate contrbuton s derved from the numerator n footnote 2/ as αε(r). 5/ Defned as publc sector defct, plus amortzaton of medum and long-term publc sector debt, plus short-term debt at end of prevous perod. 6/ Derved as nomnal nterest expendture dvded by prevous perod debt stock. 7/ Commtment bass, excludng prvatzaton, ncludng Fonds Hassan II expendtures

11 - 9 - Fgure 2. Characterstcs of External Debt, 2003 Debtors Guaranteed Debt 37% Yen 8% Other % Currency Composton US $ 25% Treasury Debt 63% Euro 56% Interest Rate Structure Credtors Floatng 24% Fxed 64% Pars club 37% Arab countres 4% Other Countres % Internatonal Insttutons 40% Sem-fxed 2% Source: Moroccan authortes. Commercal Banks 8% 7. Characterstcs of domestc publc debt are favorable and reflect a modernzed publc debt management framework. Almost all the domestc debt conssts of auctoned treasury blls that are spread over dversfed maturtes and nvestor bases (Fgure 3 and Box ), ssued n local currency, and at fxed nterest rates. The share of short term debt (less than year) has on average been kept at 25 percent to mnmze lqudty rsk and the average maturty of outstandng debt was 4 years at end A sgnfcant share of domestc debt s held by a healthy commercal bankng sector that enoys a comfortable resource base wth broad money exceedng 80 percent of GDP and reserves at the central bank equvalent to 0 percent of GDP. In ths context of abundant lqudty and low nflaton (.2 percent n 2003), nterest rates on domestc debt have remaned moderate over the yeld curve. 8. Notwthstandng ts sgnfcant holdngs of government paper, the bankng system does not appear overly exposed to government debt (Fgure 4). Moroccan banks use 80 percent of ther deposts to grant loans to the prvate sector and publc debt nstruments account for about 20 percent of ther portfolos. Nevertheless, bank credt to the prvate sector s equvalent to about 60 percent of GDP, a level comparable to other emergng market countres, although below the levels attaned n the most developed countres.

12 - 0 - Fgure 3. Characterstcs of Domestc Debt 6-month certfcates % By Product Type Other 5% Long-term 27% Maturty Structure Short-term 29% Auctons 94% Medum-term 44% Holders of Debt Yeld Curve, 2003 Other 6% Banks 36% 7% 6% 5% 4% Mutual Funds 9% Insurance Companes 29% 3% 2% 3 WEEK 26 WEEK 52 WEEK 2 YEAR 5 YEAR 0 YEAR 5 YEAR Source: Moroccan authortes.

13 - - Box. Morocco: Publc Debt Management Publc debt management has been modernzed. The nsttutonal framework s clearly defned and debt management polces are transparent. Publc debt management s the responsblty of the treasury and external fnance department of the mnstry of fnance. Each year, the treasury announces ts annual fnancng requrements and every month t announces the amounts to be rased on the aucton market. Aucton results are publshed and the treasury publshes annual reports on debt management. The treasury department holds regular auctons and mantans presence n the market. Government securtes of less than year are auctoned every week, maturtes rangng from 2 to 5 years are auctoned every month, and longer maturtes every quarter. Investors are nformed of the ssuance schedule and cancellatons are announced n advance. To stmulate the aucton market, a prmary dealer system has been put n place. Prmary dealers may submt noncompettve bds at auctons and are requred to provde quotatons to support the secondary market. The secondary market remans thn, but efforts are beng deployed to stmulate t. The treasury department ntroduced the technque of ssuance by assmlaton whereby new ssues characterstcs are dentcal to exstng ssues n order to ncrease market lqudty. A rsk management framework has also been put n place and the treasury actvely manages ts debt portfolo to lower borrowng costs whle mnmzng rsks. A benchmark portfolo has been establshed for external debt ncludng a currency composton, nterest rate, and maturty structures. Efforts are deployed to repay expensve external debt and reduce ts level through debt equty swaps and debt conversons. To take advantage of abundant lqudty n the domestc market, the treasury has also been substtutng domestc debt for external debt. On domestc debt, t ams to keep short term debt below 25 percent to mnmze lqudty rsk and extend the duraton of the portfolo. The treasury also attempts to algn expendture executon wth revenue collecton to smooth ts fnancng needs. Informaton sharng wth the central bank also takes place to facltate the latter s lqudty management. Source: Gudelne for Publc Debt Management, Accompanyng Document and Selected Case Studes. Internatonal Monetary Fund and World Bank 2003.

14 - 2 - Fgure 4. Bankng System Balance Sheet and Internatonal Comparson Bank Assets (n percent of GDP) M orocco Algera Tunsa Egypt France Span Portugal Greece Poland Hungary Share of Publc Debt n Bankng Sector Assets (n percent) M orocco Algera Tunsa Egypt France Span Portugal Greece Poland Hungary Loans to the prvate sector (n percent of deposts) M orocco Algera Tunsa Egypt France Span Portugal Greece Poland Hungary Bank Credt to the Prvate Sector (n percent of GDP) M orocco Algera Tunsa Egypt France Span Portugal Greece Poland Hungary Source: Fnancal Sector Assessment Program (2002)

15 - 3 - Developed countres, however, rely more on non-depost funds. 7 These characterstcs and the more than comfortable excess lqudty enoyed by the bankng system suggest that the government should not have dffculty meetng ts fnancng needs n the foreseeable future at reasonable costs. 9. Morocco s publc debt appears more sustanable than that of the average emergng market country. Among the vulnerablty factors for publc debt hghlghted n the IMF (2003a) study for these countres, Morocco seems to share only a smlar debt rato (Fgure 5). These countres on average have hgher debt to revenue ratos, hgher shares of external debt n total debt, lower fnancal deepenng (M3/GDP), and sgnfcantly more volatle fscal revenues. The low coeffcent of varaton of fscal revenue n partcular mples a stable revenue base to servce debt 8, whle the hgh fnancal deepenng mples a larger domestc fnancng source. 20. Judgng by the frst three crtera dentfed n the ntroducton of the paper, however, the verdct on the sustanablty of Morocco s publc debt s mxed. Morocco s publc debt rato appears sustanable from the pont of vew of past fscal performance, but not on the bass of current fscal polces. Frst, the average prmary surplus, real GDP growth rate, and mplct real nterest rate on publc debt over the last ten years (see Table ) suggest a sustanable debt level of 78 percent of GDP. 9 Relatve to the actual debt rato of 68 percent of GDP, ths benchmark debt level does not suggest overborrowng and s comparable to the benchmark debt found n ndustral countres (Fgure 6). The overborrowng rato (actual over benchmark debt) s below one (0.87), suggestng underborrowng. An equvalent calculaton for emergng market countres suggests a publc debt rato equvalent to 2.5 tmes ther benchmark debt level. Second, econometrc evdence shows a postve and sgnfcant relatonshp between the prmary surplus and the publc debt rato suggestng that adustment efforts have been made n the past to keep publc debt on a sustanable path (Table 2). The coeffcent for Morocco s sgnfcantly hgher than that of the average emergng market country. 7 Ths fndng was hghlghted by the background papers of the 2002 Fnancal Sector Assessment Program undertaken by the IMF and the World Bank. 8 Tax revenues are less volatle than output n Morocco, because the agrcultural sector whch s the man source of output volatlty s not taxed. 9 Wth an average prmary surplus at 0.7 percent, real GDP growth rate at 3.7 percent, and mplct real nterest rate at 4.6 percent, and usng d = p / (r-g) to calculate the benchmark debt where p s prmary balance as a percentage of GDP, r s the real mplct nterest rate on publc debt, g s the real GDP growth rate, and d s the debt to GDP rato.

16 - 4 - Fgure 5. Debt Vulnerablty Factors In Emergng Market Countres Publc Debt (percent of GDP) Morocco Industral Countres Emergng Markets External Publc Debt, 2002 (percent of total debt) Publc Debt, 2002 (rato to revenue; n percent) Morocco Mddle East & Afrca Asa Emergng Markets Industral countres Morocco Mddle East & Afrca Asa Emergng Markets Industral countres Comparson of Key Indcators of Publc Debt (percent) / Defaulters Nondefaulters M orocco Debt/Revenue Debt/GDP Share of external debt n percent of total debt Broad Money n percent of GDP Morocco MENA Asa Emergng Industral Volatlty of Revenues, (Coeffcent of Varaton) Source: IMF (2003a) and staff estmates. / Data are an average for Defaulters refer to countres that have defaulted snce 998. For Morocco, data s for 2002 only.

17 - 5 - Fgure 6. Comparson of Overborrowng 90 "Benchmark" Publc Debt Rato (percent of GDP) 3.5 Overborrowng Rato (actual over benchmark debt) Industral countres Morocco Emergng market economes Other developng countres 0 Industral countres Morocco Emergng market economes Other developng countres Source: IMF (2003a) and staff estmates. Thrd, however, on the bass of the average GDP growth rate and mplct real nterest rates of the last 0 years (Table ), the prmary surplus that would stablze the current publc debt rato would be at 0.6 percent of GDP compared to a prmary defct of.3 percent of GDP n A medum term scenaro that assumes the contnuaton of current fscal polces would therefore put the publc debt on an unsustanable path. D. The Need for Fscal Adustment n the Medum Term 2. The precedng analyses can be ustfably crtczed on grounds that they are mostly backward lookng. Indeed, to the extent a country can credbly commt to fscal consoldaton lookng forward, past debt dynamcs may no longer be relevant. Moroccan authortes have ndcated ther commtment to fscal consoldaton and are takng steps n ths drecton wth the am of reducng the defct to 3 percent of GDP by Ther medum term fscal consoldaton scenaro s underpnned by: () tax reform to stablze the revenue rato as trade lberalzaton proceeds, and () cvl servce reform to lower the wage bll. Along wth wde rangng structural reforms, fscal consoldaton s expected to ncrease growth to 5 percent over the medum term. The authortes do not have a debt reducton obectve, but ths scenaro would reduce the debt rato to 64 percent of GDP by The mplct nterest rate on debt s assumed to ncrease to ts hstorcal average of 6.5 percent

18 - 6 - over , 0.5 percentage pont hgher than n Growth s proected at an average of 4.5 percent over Table 2. Morocco: Fscal Polcy Reacton Functon, / (Dependent varable: prmary surplus, percent of GDP) Explanatory Varables Total Publc debt (TD)--lagged 2/ Interacton wth TD (Morocco Dummy) Drought control (Morocco) 3/ Adusted R Number of Observatons / Equatons have been estmated wth Generalzed Least Squares allowng for fxed effects and correctng for heteroskedastcty. ndcates sgnfcance at the percent level. The full emergng market countres sample n (IMF, 2003a) has been used. 2/ Emergng Market Economes n EMBI global ndex at begnnng of 2002 plus Costa Rca, Indoneda, Inda, Israel, and Jordan. 3/ Drought control s defned as years n whch the value added n agrculture and overall GDP growth were negatve. 22. Adoptng ths forward lookng approach shows that shocks could stll put publc debt on an unsustanable path (Table 3 and Fgure 7). Ths s manly because of the volatlty of output growth. Stress tests on real GDP growth suggest that adverse shocks such as two consecutve years of severe drought could put the debt rato on an unsustanable path above 00 percent of GDP. Although the assumpton of the most extreme stress test on output growth overstates the rsk (2 standard devatons of negatve shocks on growth two consecutve years, and because output was more volatle n the past), the results nevertheless renforce the vew that publc debt rato should be further reduced. 0 Under the second most extreme scenaro (one standard devaton tests on growth and nterest rates), the debt rato could ncrease to about 80 percent of GDP. Ths debt rato s the maxmum rato past fscal and growth performances suggest s sustanable, as ndcated above, but reachng t could undermne nvestor confdence. In addton, the analyss does not take nto account contngent labltes such as those related to the actuarally unbalanced penson funds and weak publc banks. 0 The volatlty of output growth s one of the man determnants of overborrowng n emergng market countres (IMF, 2003a). See SM/04/2 (4/4/04).

19 - 7 - Table 3. Morocco: Publc Sector Debt Sustanablty Framework, (In percent of GDP, unless otherwse ndcated) Actual Proectons Debt-stablzng prmary balance 0/ Publc sector debt / o/w foregn-currency denomnated Change n publc sector debt Identfed debt-creatng flows (472) Prmary defct Revenue and grants Prmary (nonnterest) expendture Automatc debt dynamcs 2/ Contrbuton from nterest rate/growth dfferental 3/ Of whch contrbuton from real nterest rate Of whch contrbuton from real GDP growth Contrbuton from exchange rate deprecaton 4/ Other dentfed debt-creatng flows Prvatzaton recepts (negatve) Recognton of mplct or contngent labltes Other (specfy, e.g. bank recaptalzaton) Resdual, ncludng asset changes (2-3) Publc sector debt-to-revenue rato / Gross fnancng need 5/ n bllons of U.S. dollars Year 0-Year Hstorcal Standard Proected Key Macroeconomc and Fscal Assumptons Average Devaton Average Real GDP growth (n percent) Average nomnal nterest rate on publc debt (n percent) 6/ Average real nterest rate (nomnal rate mnus change n GDP deflator, Nomnal apprecaton (ncrease n US dollar value of local currency, n I. Adustment Scenaro Inflaton rate (GDP deflator, n percent) Growth of real prmary spendng (deflated by GDP deflator, n percent) Prmary defct Debt-stablzng II. Stress Tests for Publc Debt Rato prmary A. Alternatve Scenaros balance 0/ A. Key varables are at ther hstorcal averages n / A2. Prmary balance under adustment scenaro n A3. No polcy change (relatve to baselne) 8/ B. Bound Tests B. Real nterest rate s at hstorcal average plus two standard devatons n 2005 and B2. Real GDP growth s at hstorcal average mnus two standard devatons n 2005 and B3. Prmary balance s at hstorcal average mnus two standard devatons n 2005 and B4. Combnaton of -3 usng one standard devaton shocks B5. One tme 30 percent real deprecaton n / B6. 0 percent of GDP ncrease n other debt-creatng flows n / Publc sector refers to central government total gross debt 2/ Derved as [(r - π(g) - g αε(r)]/(gπgπ)) tmes prevous perod debt rato, wth r = nterest rate; π = growth rate of GDP deflator; g = real GDP growth rate; a = share of foregn-currency denomnated debt; and ε = nomnal exchange rate deprecaton (measured by ncrease n local currency value of U.S. dollar). 3/ The real nterest rate contrbuton s derved from the numerator n footnote 2/ as r - π (g) and the real growth contrbuton as -g. 4/ The exchange rate contrbuton s derved from the numerator n footnote 2/ as αε(r). 5/ Defned as publc sector defct, plus amortzaton of medum and long-term publc sector debt, plus short-term debt at end of prevous perod. 6/ Derved as nomnal nterest expendture dvded by prevous perod debt stock. 7/ The key varables nclude real GDP growth; real nterest rate; and prmary balance n percent of GDP. 8/ In partcular, the revenue rato declnes as trade lberalzaton proceeds and cvl servce reform s not undertaken 9/ Real deprecaton s defned as nomnal deprecaton (measured by percentage fall n dollar value of local currency) mnus domestc nflaton (based on GDP deflator). 0/ Assumes that key varables (real GDP growth, real nterest rate, and prmary balance) reman at the level n percent of GDP/growth rate of the last proecton year.

20 - 8 - Fgure 7. Debt to GDP Rato Under Alternatve Fscal Scenaros Debt-to-GDP Rato and Stress Test on No Polcy Change Scenaro / (In percent) Debt-to-GDP Rato and Stress Tests on Adustment Scenaro / (In percent) No polcy change Authortes Adustment Scenaro Bound Test B2 (Most Extreme) / Bound Test B4 (Second Most Extreme) / Bound Test B4 (Second Most Extreme) / Bound Test B2 (Most Extreme) / Source: IMF staff estmates. / See Table 3 on Publlc Sector Debt Sustanablty Framework for the defnton of bound tests. E. Concluson 23. The paper suggests that a front loaded fscal adustment to reduce the debt below 60 percent n the medum term s desrable to provde Morocco wth room to absorb shocks. Past fscal performance and the characterstcs of the publc debt stock suggest that Morocco may sustan a hgher debt level than the average emergng market country. Other characterstcs of the Moroccan economy also compare favorably wth other emergng market countres, notably, favorable external vulnerablty ndcators, low nflaton, and low fscal revenue volatlty. However, the volatlty of output growth has been hgh. In vew of ths, a lower level of debt would ncrease Morocco s credblty vs-à-vs domestc and nternatonal nvestors n comparson to other emergng market countres.

21 - 9 - APPENDIX I Publc Debt Dynamcs Equaton 2 Debt dynamcs are modeled by the followng process: ) ) dt dt = ( r π ( g) g εα( r )) d ( g π gπ ) t where d s the debt-to-gdp rato, pb s the prmary balance, r ) s a weghted average of domestc and foregn nterest rates, α s the share of foregn currency-denomnated publc debt, π s the change n the domestc GDP deflator, and g the real GDP growth rate. Changes n the exchange rate (local currency per U.S. dollar) are denoted by ε, wth ε >0 ndcatng a deprecaton of the local currency. Based on the above equaton, the framework separates between dfferent channels that contrbute to the evoluton of the debt-to-gdp rato (Tables and 3 n the text)): the prmary defct (lne 4) and the endogenous/automatc factors (lne 7), whch nclude the real nterest rate, real GDP growth, and exchange rate movements. The contrbuton of the real nterest ) r π ( g) rate to the evoluton of the debt rato s gven by dt (lne 9), the contrbuton ( g π gπ ) of the real growth rate as g dt ( g π gπ ) (lne 0), and that of an exchange rate εα( r ) ) deprecaton as dt (lne ). The separaton of the dfferent factors allows an ( g π gπ ) assessment of ther relatve mportance for the evoluton of the debt rato. It also serve as the bass for stress tests, whch nclude a baselne check, n whch the key varables (such as real GDP growth) are assumed to be at ther hstorcal averages, and a number of temporary shocks to these varables (solated and combned) based on hstorcal averages and standard devatons. The results of these stress tests are summarzed along wth the baselne assumptons and proectons n Table 3. pb t 2 Based on Debt Sustanablty Template, Gudance Note November 8, Polcy Development and Revew Department, Internatonal Monetary Fund.

22 REFERENCES Chalk. N. and Hemmng. R Assessng Fscal Sustanablty n Theory and n Practce. Internatonal Monetary Fund Workng Paper WP/00/8. Internatonal Monetary Fund (2002) Assessng sustanablty. SM/02/ a Publc Debt n World Economc Outlook, September. 2003b Sustanablty assessment Revew of Applcaton and Methodologcal Refnements c Debt Sustanablty n Low Income Countres Toward a Forward Lookng Strategy SM/03/85. And World Bank, Gudelnes for Publc Debt Management: Accompanyng Document and Selected Case Studes. Renhart, C.M., Rogoff, K.S, Savastano, et al Debt ntolerance. Brookngs Papers on Economc Actvty. Treasury department, Morocco Mnstry of Fnance, «Rapport sur l évoluton de la dette ntéreure en 2003». Treasury department, Morocco Mnstry of Fnance, «Rapport sur la stuaton de la dette extéreure en 2003».

23 - 2 - II. IMPACT OF THE BARCELONA PROCESS ON MOROCCO S TRADE 3 A. Introducton 24. A closer ntegraton wth the global economy s a key component of Morocco s growth strategy. Increased openness to trade and captal flows can contrbute to mprove economc performance through a more effcent allocaton of resources, technology transfers, scale effects, compettve pressure, and mproved access to foregn savngs. Morocco s man trade lberalzaton efforts have been undertaken wth the European Unon (EU), frst under a cooperaton agreement, and more recently and more sgnfcantly wthn the Barcelona process. 4 Ths process, ntated n Barcelona n 995, ams at closer poltcal, economc, and cultural cooperaton between the EU and Southern and Eastern Medterranean countres (henceforth referred to as Medterranean countres). The man nstruments used to buld the economc partnershp are the blateral Assocaton Agreements wth the European Unon (AAEU), supported by grants under the MEDA program, and complemented by loans by the European Investment Bank (EIB). The prncpal obectves are to create a Euro- Medterranean free trade zone by 200, allevate the challenges from economc lberalzaton through capacty buldng and measures to enhance compettveness, and ncrease foregn drect nvestment nto Medterranean countres. Morocco s AAEU entered nto force on March, 2000, makng the country the thrd partcpant after the Palestnan Authorty (997) and Tunsa (998) The Barcelona process presents both opportuntes and rsks. Whle quantfcaton of ts mpact on welfare s beyond the scope of ths study, the effect of the process on Morocco s trade pattern wll be analyzed, usng a gravty model of blateral trade. In partcular, the followng questons wll be addressed: Has the Barcelona process resulted n trade creaton? Several factors could potentally mpede the generaton of trade between the EU and ts Medterranean partners. Exports creaton toward the EU could suffer f Morocco and other Medterranean countres are not compettve enough to penetrate the European 3 Prepared by Ludvg Söderlng. Research assstance was contrbuted by Tea Trumbc. 4 Morocco s also undertakng other trade lberalzaton efforts, but ther mpact wll largely be felt n the future. A free trade agreement was recently sgned wth the Unted States, another wth Tunsa, Egypt and Jordan, and yet another wth Turkey. Morocco has also recently reduced and smplfed the tarff protecton vsà-vs Most Favored Naton (MFN) countres. 5 An AAEU came nto force wth Israel n 2000, and wth Jordan n Egypt sgned an agreement n 200, and Lebanon and Algera n 2002 (all awat ratfcaton). Negotatons were concluded wth Syra n All aforementoned countres sgned cooperaton agreements n the md-970s. Cyprus, Malta (to become EU members n 2004) and Turkey (an EU customs unon partner) concluded frst generaton assocaton agreements wth the European Communty n the 960s and early 970s.

24 market. On the mport sde, Morocco s AAEU, lke those of other countres, provde for a sgnfcantly more rapd reducton n tarffs on non-competng captal goods and ntermedate nputs than on competng fnal goods. As a result, effectve protecton ncreases ntally, to be reversed only over tme. Is there evdence of trade dverson? Smlar to all preferental trade lberalzaton efforts, the Barcelona process n general and the AAEUs n partcular rsk dsplacng trade from more compettve non-eu producers to less compettve producers (tarffs excluded) wthn the EU. As a result, a porton of the welfare gans from trade lberalzaton could potentally accrue to neffcent European producers, whle the Medterranean countres pay the prce n terms of lost fscal revenue from tarffs. Has there been any mpact on ntraregonal trade,.e., on trade among Medterranean countres? One of the obectves of the Barcelona process s to ncrease foregn drect nvestment (FDI) n Medterranean countres. However, as long as trade barrers are hgher between Medterranean countres than those vs-à-vs the EU, there s a rsk of a hub and spoke effect, whereby FDI s drected toward the EU rather than toward Medterranean countres. By contrast, ncreased ntraregonal ntegraton could counter ths effect by expandng the market sze both for ntermedate and fnal goods. B. Developments n Trade Patterns Snce 990 Analytcal Framework: The Gravty Model of Blateral Trade 26. The gravty model s a tool often used to analyze blateral trade patterns. Its smplcty and hgh level of statstcal explanatory power have contrbuted to promotng ts wde use. The basc gravty model relates some measure of blateral trade (mports, exports, or both) to the economc sze of two countres, and the geographcal dstance between them. Populaton (or GDP per capta), s often also ncluded, along wth other varables that could nfluence blateral trade. The general specfcaton used here s: M = Y α Y β P λ P δ D φ X e ε where M s mports to country from country. Only south-south and north-south trade s consdered, gven that the focus of the analyss s on Morocco and that hence ntra-ndustral country trade (n partcular ntra-eu trade) s unlkely to provde an approprate reference for the trade flows studed here. Y and Y are GDP n country and respectvely, P and P are populaton n country and, D s the dstance between country and, X s a vector of varables descrbng ether country or, or both. Ths vector ncludes the share of agrculture n GDP of the exportng country and dummy varables for trade between partners sharng the

25 same language, for partners borderng each other, for prmary commodty exporters, and for landlocked countres. 6 ε s a normally dstrbuted error term, wth zero mean. 27. Blateral trade can be expected to depend postvely on the sze of the two economes, and negatvely on the dstance between the countres. A large populaton s generally consdered to relate negatvely to trade, snce ths would mply a larger domestc market and a hgher degree of auto-suffcency. 7 Trade s also lkely to be hgher between borderng countres and countres sharng a common language, and lower for landlocked countres. The share of agrculture n GDP can be expected to correlate negatvely to exports, snce trade protecton tends to be partcularly hgh aganst agrcultural products. Trade n prmary commodtes should deally be excluded from the gravty model, gven that terms of trade swngs can cause sgnfcant volatlty n trade. However, IMF s Drecton of Trade Statstcs (DOTS) provdes data for total trade only, and a dummy varable for commodty exporters s used nstead. 8 There s no strong a pror reason for ths dummy to be postve or negatve. Emprcal Results 28. The econometrc analyss s based on a dataset coverng blateral trade for 90 countres. Three-year averages are calculated for four separate regressons for the perods 99 93, , , and (latest avalable). Dummy varables for trade between Medterranean countres 9 and the EU, ntra-medterranean trade, and Medterranean trade wth the rest of the world are ncluded to analyze any changes n trade patterns over tme. Any changes n Medterranean trade patterns as seen by changes n the coeffcent for the relevant regonal dummes can reasonably be attrbuted manly to the Barcelona process, snce t s the prncpal trade lberalzaton ntatve undertaken n that regon durng the studed perod. Another set of regressons s carred out usng regonal trade dummes for the EU, the Mddle East and North Afrca (MENA), Asa, sub-saharan Afrca, European developng countres, Latn Amerca, and non-eu ndustral countres to control for 6 The dummy for landlocked countres takes the value f one of the countres s landlocked, 2 f both are, and zero otherwse. All other dummes take the value or zero. 7 Moreover, for a gven level of GDP, a large populaton ndcates a low level of economc development and hence generally a low export capacty. It has, however, occasonally been argued that that a large populaton allows for scale effects and a more effcent dvson of labor and would therefore affect trade postvely. 8 The UN s COMTRADE database does provde nformaton on trade by maor product group. Explotng ths could be useful for further research. 9 Medterranean countres are Algera, Egypt, Jordan, Morocco, Syra, and Tunsa. Lebanon and the Palestnan Authorty are excluded due to ncomplete data.