Turkey - Technical Assistance for Treasury Data Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No: 22468 IMPLEMENTATION COMPLETION REPORT (CPL-34770; SCL-3477A; SCPD-3477S) ONA LOAN IN THE AMOUNT OF US$ 9.20 MILLION EQUIVALENT TO THE REPUBLIC OF TURKEY FOR A TA FOR TREASURY DATA 06/28/2001 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS (Exchange Rate Effective February 23, 2001) Currency Unit = Turkish Lira TL 1 = US$ 0.000001 US$ 1 = 960.189 GOVERNMENT FISCAL YEAR January 1 December 31 ABBREVIATIONS AND ACRONYMS CAS Country Assistance Strategy CFA Certified Financial Accountant CPA Certified Public Accountant DMS Data Management System DPC Data Processing Center EIC Economic Information Center ERR Economic Rate of Return GOT Government of Turkey GTZ Govenunent of Gernany IFO German Research Institute IT Information Technology LAN Local Area Network N/A Not Applicable OEF Oxford Economic Forecasting PC Personal Computer PFMP Public Financial Management Project PSBR Public Sector Borrowing Requirement SOE State-Owned Enterprises TOR Terms of Reference Vice President: Johannes F. Linn Country Director Ajay Chhibber Sector Director: Pradeep Mitra Sector Manager: Samuel Otoo Task Team Leader Ismail Arslan  TURKEY TA FOR TREASURY DATA CONTENTS Page No. 1. Project Data 1 2. Principal Performance Ratings 1 3. Assessment of Development Objective and Design, and of Quality at Entry 2 4. Achievement of Objective and Outputs 3 5. Major Factors Affecting Implementation and Outcome 8 6. Sustainability 8 7. Bank and Borrower Performance 9 8. Lessons Learned 10 9. Partner Comments 11 10. Additional Information 18 Annex 1. Key Performance Indicators/Log Frame Matrix 19 Annex 2. Project Costs and Financing 21 Annex 3. Economic Costs and Benefits 23 Annex 4. Bank Inputs 24 Annex 5. Ratings for Achievement of Objectives/Outputs of Components 26 Annex 6. Ratings of Bank and Borrower Performance 27 Annex 7. List of Supporting Documents 28   Project ID: P009097 Project Name: TA FOR TREASURY DATA Team Leader: Ismail Arslan TL Unit: ECSPE ICR Type: Core ICR Report Date: June 28, 2001 1. Project Data Name: TA FOR TREASURY DATA L/C/TF Number: CPL-34770; SCL-3477A; SCPD-3477S Country/Department: TURKEY Region: Europe and Central Asia Region Sector/subsector: BF - Public Financial Management KEY DATES Original Revised/Actual PCD: 10/08/91 Effective: Appraisal: 06/10/92 MTR: Approv'al: 05/28/92 Closing: Borrower/lmplementing Agency: GOT/TREASURY Other Partners: GTZ STAFF Current At Appraisal Vice Presidenit: Johannes Linn Wilfred Thalwitz Country Manager: Ajay Chhibber Sector Manager: Pradeep Mitra Team Leader at ICR: Ismail Arslan Barbara Lewis ICR Primary Auithor: Elliott Hurwitz 2. Principal Performance Ratings (HS=Highly Satisfactory, S=Satisfactory, U=Unsatisfactory, HL=Highly Likely, L=Likely, UN=Unlikely, HUN=Highly Unlikely, HU=Highly Unsatisfactory, H=High, SU=Substantial, M=Modest, N=Negligible) Outcome: S Sustainabilitv: L Institutional Development Impact: SU Bank Performance: S Borrowt-er Performiance: S QAG (if available) ICR Quality at Entry: S S Project at Risk at Any Time: No Introduction The Treasury in Turkey is one of the key agencies charged with formulation of macroeconomic policy. (Others are the Ministry of Finance, State Planning Organization, and the Central Bank) In addition to responsibility for supporting broad macroeconomic policy, Treasury (at project inception) was responsible for foreign economic relations and trade, public sector borrowing, financial sector regulation and supervision, international capital flows, management of the portfolio of state economic enterprises, and supervision of the insurance sector. (In 1995 foreign trade became the responsibility of the  Undersecretariat of Foreign Trade, which was split off from Treasury. Financial sector regulation and supervision units transferred to the newly established Bank Regulation and Supervisory Agency in 2000). At the start of the 1990s, Treasury's information systems and data management capabilities were obsolete and inadequate to support its responsibilities. Numbers reported by different departments were inconsistent. Many data series were marked by late reporting, inconsistent definitions, and other problems. Most staff did not have computers, and there was no infrastructure for sharing data. In addition, there was a need to bolster the agency's capabilities in database design, systems management, and analytic and modeling techniques. This project was prepared with the request of the Turkish Government. Total project cost was US$30.05 million, with US$9.20 million contributed by the Bank, US$1.95 by the Government of Germany (GTZ), and US$18.90 by the government. 3. Assessment of Development Objective and Design, and of Quality at Entry 3.1 Original Objective: The Treasury Data Systems Project (TDSP) objectives were to upgrade the technology, information, and analytic resources essential to Treasury's role in macroeconomic management. Specifically, the project aimed to (a) improve the agency's monitoring and policy design functions; (b) provide timely and efficient access for Treasury staff to a consistent, centralized data source; (c) enable Treasury to draw upon and contribute to databases of multilateral institutions and other agencies; (d) ensure that organizational procedures support efficient data processing and analysis; and (e) support investment in human capital. Evaluation of Objectives The objectives were clear and realistic, consistent with the Bank's goals as stated in the CAS, and consistent with the country's goals. They were highly relevant to Treasury's mission and important for economic development. Underlying the objectives was the assumption that better economic data would facilitate government efforts to improve macroeconomic policy setting and, implicitly, macroeconomic performance by reducing the Public Sector Borrowing Requirement (PSBR), improving the performance of SOEs, etc. Given the weaknesses in economic data and lack of analytical capability, it was realistic to assume that better data would assist in formulating a more balanced policy. 3.2 Revised Objective: N/A 3.3 Original Components: The original components were: 1. Develop a Data Management System (DMS) that (a) provides an appropriate technology environment for macroeconomic data analysis; (b) contains centralized databases to support this work. 2. Undertake institution-building initiatives that include (a) training in selected economic, financial, and technical subjects; (b) consultant assistance for special studies and economic model-building; (c) management and training consultancy support; and (d) establishment of a Treasury library and Economic Information Center open to the public. The DMS was originally intended to be built on the existing mainframe computer and included design and development of economic databases, personal workstations, LAN and communications systems, and -2 -  technical training for Data Processing Center (DPC) staff. The institutional development component was intended to introduce modeling tools and undertake analytic studies to enhance staff capabilities in relevant areas. It was anticipated that organizational processes would be introduced to facilitate efficient data-sharing and management. A professional training program would also be undertaken to upgrade the skills of Treasury staff. Component 1: DMS development was intended to be achieved via the following tasks: (1) Update mainframe and install LAN services; (2) Convert existing data; (3) Design new databases by means of intensive end-user consultation; (4) Create new databases; (5) Train Treasury staff in use of new system. Component 2, institutional development, consisted of (1) Provision of short-term training and preparation of plans for long-term training; (2) Implementation of analytic studies and tools to permit advanced macroeconomic and foreign trade analyses, as well as specialized studies; (3) Accomplishment of organizational changes needed to take advantage of the new system. While the original objectives and components were not revised, two substantive changes were made after project initiation. First the Bank proposed and the government agreed that, because of the shift in the prevailing IT paradigm, the TDSP would shift to a client-server architecture instead of using the existing mainframe. Second, it was later decided that instead of establishing a physical Economic Information Center, dissemination of Treasury economic information over the Internet would instead serve the same purpose more efficiently. 3.4 Revised Components: N/A 3.5 Quality at Entry: Quality at entry was satisfactory. The design was straightforward, aimed squarely at the objectives as stated. The project approach fit the prevailing computing paradigm, and the institutional development components effectively complemented development of the DMS. The assumption that better data would facilitate better macroeconomic policy formulation was reasonable, and the project had the strong backing and enthusiastic participation of the government. Assessment of Design As noted, the project design was simple, and in accordance with the prevailing computing paradigm. This design was not risky. The first component was well within the implementation capabilities of the Treasury Data Processing Center (DPC) supplemented by relatively modest consultant support. The second component fit well within the capabilities of the General Directorate of Research (which was responsible for project implementation), augmented by other Treasury components as required. 4. Achievement of Objective and Outputs 4.1 Outcome/achievement of objective: Outcome overall was satisfactory, with highly-leveraged results from a relatively small Bank investment. Interviews with a range of Treasury officials and those of other economic agencies as well as private economists indicate that the quality and timeliness of Treasury economic data improved considerably over the period of the TDSP. And there is ample evidence that TDSP has made Treasury operations more efficient. The project fully succeeded in an ambitious improvement of Treasury's automation infrastructure, comprising servers, personal computers, LAN, database system, e-mail, and installed software. Use of computers by technical staff has become routine, and the ratio of computers to staff is nearly 1 to 1 - 3 -  (compared with a project goal of 1 to 2). The Treasury web site is now widely used for data dissemination. Economic studies and attendant staff development was successful. Extensive model development and training was successfully achieved in macroeconomics; leading economic indicators; and sources and uses of funds, with somewhat less progress in the two remaining areas. Short-term training of Treasury staff in utilization of the new systems was successful, with training provided to 12 departments of the Undersecretariats of Treasury and Foreign Trade. Short-term training was also provided in accounting, statistics, international trade, and other relevant subjects. For long-term training, a plan has been developed to integrate and carry out this training as part of regular Treasury activities. However, the effort to introduce organizational processes that would facilitate efficient data sharing and management has had mixed success. Many departments, such as Domestic and External Debt, make full use of systems and in some cases have gone beyond the initial design capability. In order to provide reliability and accuracy of data, some general directorates have responsibility to enter data, which are the reference information of the main data, into the data bases. Therefore, the installation of the DMS has produced one more step forward in Treasury data processing and analytic capabilities. However, because of the complexity of some data bases, new requirements are not implemented easily in the system, so, in some instances the DMS is not sufficiently flexible to meet user needs efficiently (details are provided below). 4.2 Outpuits by components: Component 1-DMS Development: As noted earlier, development and implementation of the DMS was satisfactory, with substantial achievements considerably outweighing shortcomings. The installed infrastructure is well documented and performs very well. The network utilizes ATM switches, a highly-modem technology. Six modern servers support about 875 PCs; the system has the capacity to support around 1500 PCs, providing growth potential. Security systems and firewalls are installed to ensure data confidentiality. It should also be noted that the changeover from the mainframe-based architecture to a client-server architecture was accomplished without the need for additional funding. Principal Achievements: User needs were determined by an extensive consultant study, which established in detail the system requirements of each general directorate. Based on recommendations from consultants, the DMS acquired and installed an Oracle Relational Database Management System, which has been extremely successful. DMS also uses Graphical Query Language and Power Builder software to create departmental forms and reports. An e-mail system was installed that introduced additional efficiencies into Treasury operations. TDSP use of off-the-shelf software has been very cost-effective, as the potentially higher costs and delays of custom-developed software were avoided. As envisioned, Treasury DPC staff supplemented by consultants converted existing data, created the new databases within the DMS, and trained Treasury staff in use of the new system. As noted above, eleven projects were undertaken under TDSP, in most cases corresponding to a unique database and supporting application software: 1. Domestic Debt 2. External Finance 3. Asset and Debt Management 4. State Owned Enterprises 5. Banking 6. Sworn Bank Auditors 7. Insurance -4 -  8. Foreign Trade 9. Macroeconomic Indicators 10. Financial Institutions 11. Automated Document Archival Systems In eight of these eleven instances, Treasury directorates are making effective use of the DMS. For example, the General Directorate for Foreign Economic Relations uses the DMS to perform analyses in support of negotiations for new loans from multilateral institutions. And the General Directorate for Public Finance uses the DMS to manage external debt and monitor actual payments against what had been projected. When DMS was first installed, Public Finance issued quarterly reports on indebtedness, but in April, 2000, the directorate began monthly reporting. Data from the Bank's databases were transferred into the External Finance database in a protracted process that involved travel of Treasury staff to Washington. Bank staff report that Treasury debt reporting is among the best from member nations, and they attribute this quality to TDSP. The project introduced formal data-sharing arrangements between government agencies. Protocols between Treasury and other agencies using or generating economic data were concluded, and in general data exchange is routine and efficient. Such agreements are in effect with the Central Bank, the State Institute of Statistics, Istanbul Stock Exchange, and the Undersecretariat of Foreign Trade. Other economic agencies and private users also utilize data generated by the DMS. The Central Bank uses Treasury data on extemal debt in assessing the balance of payments, and its banking and foreign reserves departments are also heavy users of DMS data. And academic researchers make significant use of data from the Treasury website (www.treasury.gov.tr). In general, academic specialists and other economic agencies stated that the quality of Treasury economic statistics has improved significantly over the period of the TDSP. Shortcomings: As noted, in two instances Treasury Directorates are not making full use of the DMS. The SOE and Insurance General Directorates made little or no use of the DMS, continued the use of their own systems, and participated little in training programs. This was due to a mixture of a general resistance to change, reluctance to depart from existing systems that were known to be reliable, uneven support by directorate management, some perceived deficiencies in the DMS, and a heavy workload of routine reporting. Part of the resistance also appeared to be due to internal institutional resistance, as the Treasury DPC and Treasury management made a number of efforts to induce cooperation and meet the directorates' stated data processing requirements. In the insurance area, however, despite repeated efforts on the part of DPC, the TDSP-prepared module was not sufficiently responsive to the Directorate's requirements. The eleventh project, the automated archival system, has been deferred and is expected to be- accomplished in the future. Treasury DPC and management are continuing to work with the directorates noted to improve the capabilities and user-friendliness of the DMS, and it is expected that further system development will enhance system utilization by these directorates. With regard to SOEs, it should be noted that even though the SOE General Directorate makes little use of the DMS, individual enterprises have made considerable use of this data, and in many cases have been able to better analyze and improve their operations as a result. While the TDSP staff has recommended that it would be safer and more efficient to utilize the DMS, the SOE General Directorate believes that the continued use of their existing Excel system is adequate. As mentioned earlier, there are some issues in the manner in which Treasury utilizes the DMS which will -5-  have to be resolved through a collaborative process among the Treasury Departments. The functional duties of debt management officers within Treasury has led to different approaches in data storage and data processing of DMS. For example, the General Directorate for Public Finance is responsible to follow the actual transactions (drawings and payments) of extemal debt operations and to announce this information to the public. Therefore, in order to produce accurate and reliable extemal debt data, the responsibility for maintenance of reference tables in the DMS (for instance the information of foreign lenders, borrower, project name etc.) is given to one staff in the General Directorate of Public Finance. Other users can process the reference information on a "read only " basis. While entering, a new loan agreement into the system, the certain reference information should be provided by the staff from the General Directorate of Public Finance. This sometimes causes delays in entering the data into the system. As another example, the system is not able to accommodate multi-tranche loans, thus requiring staff to develop a "workaround" solution that requires entering a separate loan number for each tranche. However, modifications to the system are planned to be made before end-2001 to accommodate multi-tranche loans. Overall, these shortcomings are relatively minor when compared to the benefits derived from the implementation of DMS. Component 2-Institutional Development: Achievement was satisfactory overall: Short -Term Training: As noted earlier, short-term training needs were identified through a consultant study, which was revised with the Bank's help after the split-off of the Undersecretariat of Treasury and Foreign Trade. The purpose of the short-term training was to provide the initial capacity for Treasury staff to utilize the DMS. The recommended training was delivered to around 300 staff from 12 departments of Treasury and the Foreign Trade Undersecretariat. The success of this effort was demonstrated by the fact that staff have been productively using the system for the past several years as additional databases and analytic capabilities have been brought on line. Refresher training was provided throughout the duration of the project. It should be noted that due to resource constraints some general directorates have not been provided with the amount or frequency of training they believe necessary. Also due to resource constraints, some specialized software identified during project implementation (e.g., for portfolio risk management analysis) has not yet been acquired. Long-Term Training: The purpose of this sub-component was to institutionalize on-going training, so that (1) new Treasury staff would have the requisite skills to utilize the DMS; (2) new applications and knowledge could be introduced. A consultant report in this area was prepared, which provided the basis for Treasury to establish in-house programs. A Steering Committee for training, with the participation of each department, was also established to follow up long-term needs. This committee has met periodically and developed recommendations for Treasury management. In addition, a survey was performed (which is planned to be repeated annually) to identify these needs, and an action plan prepared. Thus the goal of institutionalizing on-going training has been achieved. Controller Training Proeram: With the goal of improving the audit of Bank projects, TDSP envisaged a two year CPA/CFA training program for four Treasury Auditors, Halfway through project implementation, Treasury requested a change to in-country shorter-term training which would benefit a larger number of auditors. The Bank agreed with this request, but for a number of reasons it was delayed, requiring the project to be extended. Eventually, a revised program for 20 auditors was deferred to the Public Finance Management Project, under which it is currently scheduled to be completed by December 31, 2001. Thus - 6 -  controller training intended for the TDSP was not completed by the close of that project. Studies: As noted earlier, studies were undertaken in several areas with the goal of building up Treasury's analytic capabilities. While there were some shortfalls, the main goals were achieved, with the greatest progress in the areas of macroeconomics; leading economic indicators; and sources and uses of funds. In these studies, the consultants worked very closely with Treasury staff, with the goal of transferring knowledge by "learning by doing." When the studies were completed, Treasury staff took over operation and further development of the models and other analytic tools that had been developed. A sophisticated study in the area of foreign trade was completed, but was accompanied by less transfer of skills to Treasury staff than had been envisioned. And in the area of consolidation of fiscal accounts, although several analytical tools were installed and papers written, the anticipated work was not completed. In the area of Financial Early Warning Indicators, for instance, threshold values for certain variables that are believed to be reliable indicators of a prospective financial crisis are monitored. Similarly, a "Market Pressure Index" is calculated monthly that estimates the probability of a financial crisis within the next 12 months. These capabilities, as well as those described earlier, were enabled by the TDSP. Treasury also utilized its macroeconomic model to estimate the economic effects of the recent earthquakes, thereby permitting the government to modify the assumptions it was using to set spending and revenue targets. As discussed in greater detail below, in the performance of the studies Treasury made a strong effort to select consultants who were pre-eminent in their fields. However, this often did not result in the selection of individuals who were effective-or sufficiently available-to transfer knowledge to Treasury staff. As noted earlier, a physical Economic Information Center was not established, but Treasury economic infornation has been disseminated over the Internet, thereby accomplishing the same goal more efficiently. The library, training center, conference facility, and publication center were all developed as envisioned, and are currently used for in-house training as well as seminars open to the public. 4.3 Net Present Value/Economic rate of retutrn: It was not possible to compute an ERR, and a measure of cost-effectiveness was not practical. 4.4 Financial *ate of return: N/A 4.5 Institutional development impact: Institutional Development Impact was substantial: * Prior to the TDSP, Treasury data was rudimentary and inconsistent, reporting was not timely, and many Treasury departments (and staff) collected and analyzed their own data-with little coordination between staff. Treasury publications contained inconsistent information. TDSP dramatically improved this situation, making data series more consistent, timely, and available to a greater audience. This significantly improved the quality of information available to economic policy officials, to the research community, and to international organizations. * TDSP provided Treasury with a much-improved capability to formulate and monitor economic policy. DMS is a key instrument utilized for Treasury's meetings with multilateral organizations, and the General Directorate for Economic Research produces monthly, and in some cases weekly, reports on the current economic program (which are also used for the IMF Stand-by Agreement). -7-  1. Calculations of the public sector borrowing requirement -a critical parameter for Turkish economic policy officials-can now be generated quarterly for the central government, which was not possible before DMS. 2. The General Directorate for Economic Research regularly forecasts major economic aggregates and Leading Economic Indicators. 3. In the perforrnance of the 5 studies described earlier, Treasury staff "learned by doing" and thereby enhanced agency capabilities in each of the designated areas. In macroeconomic modeling, for instance, Treasury staff is now changing the parameters of the models without help from outside consultants. * Treasury staff received short-term training in accounting, statistics, international trade, and other topics that would upgrade their skills in relevant areas. * The functioning of the Undersecretariat of Treasury is much more efficient than it was before TDSP. This has been made possible by the use of a common data base, and possession of modem computing and communications tools-including the Internet and e-mail. In addition, because they have better information and can better plan borrowing needs, Treasury is more efficient at managing external debt, for example, by optimizing the selection of the currency in which loans are denominated. Regarding domestic debt, Treasury has much more current information on conditions in the credit markets, and can "time" its offerings to obtain better terms. 5. Major Factors Affecting Implementation and Outcome 5 1 Factors outside the control of government or implementing agency: e Soon after project inception, it became clear that the computing paradigm was shifting from a mainframe-based architecture to a client/server arrangement. As noted earlier, after extensive discussions, the Bank and the GOT agreed to shift the DMS architecture accordingly, which was accomplished at no additional cost to the project. 5.2 Factors generally subject to government control: