Chapter

Chapter 3 Effectiveness of the Programs: What Was Achieved and How

Author(s):
International Monetary Fund. Independent Evaluation Office
Published Date:
April 2006
Share
  • ShareShare
Show Summary Details

This chapter examines the extent to which the objectives of programs were achieved, and what factors contributed to success or lack thereof. The objectives focused on are the same ones highlighted in Chapter 2: external viability, fiscal sustainability, inflation, and growth. The roles of conditionality and “country ownership” are explored in the context of program implementation.

Substantial Progress Toward External Viability

As discussed in Chapter 2, programs initially defined external viability in terms of exiting from reliance on exceptional financing. They measured progress toward this goal with indicators of external current account adjustment and the trajectory of debt and debt-service obligations. Jordan will continue to benefit from debt relief under the 2002 Paris Club agreement until 2007, but the 2002 agreement was regarded as an “exit” one. As discussed below, other indicators suggest that Jordan has made significant progress toward achieving external viability, including substantial improvement in the balance of payments, a greatly improved debt profile, and an increased resilience to external shocks. The following types of comparisons are used to measure this progress (although not all are relevant or feasible in each case): (1) progress over time (i.e., since the beginning of IMF program involvement in the late 1980s); (2) progress compared to the objectives set out in the various arrangements; and (3) comparisons vis-à-vis other countries.

Marked improvement in the external current account balance

Compared to the early 1990s, there has been a marked turnaround in the external current account balance, measured with or without official transfers (Figure 3.1 and Figure 3.2). The main factors that contributed to this outcome were increased exports and a strong recovery in remittances and, since 2003, a marked pickup in official transfers. Since 2000, a major source of export growth has been output from “Qualifying Industrial Zones” (QIZs), which enjoy quota- and duty-free access to the United States. After growing at less than 2 percent a year during 1996–2000, the dollar value of exports grew at over 15 percent a year during 2001–03, with most of the increase going to the United States. The QIZs have boosted nontraditional exports and converted the United States from being a minor market for Jordanian exports to the largest single country export market.

Figure 3.1.Current Account Balance, Including Official Transfers

(In percent of GDP)

Source: IMF.

Figure 3.2.Current Account Balance, Excluding Official Transfers

(In percent of GDP)

Source: IMF.

Developments in Jordan’s external current account balance have also been significantly influenced by foreign grant receipts. These have sometimes fluctuated widely on account of political factors. For example, they were cut by governments in some Arab countries and the United States in the early 1990s because of Jordan’s perceived support of Iraq in the 1991 war in the Middle East. More recently, they were boosted substantially in 2003 by the United States partly to compensate Jordan for the disruption to its economy caused by the U.S.-led war in Iraq. Abstracting from these bumps, official transfers recorded in the balance of payments averaged about $400 million or 5½ percent of GDP a year between 1992 and 2002.

How did the adjustment that occurred in the external current account compare with program projections? Figure 3.1 and 3.2 show that since the mid-1990s (i.e., beginning with the first EFF) the current account balance has tended to improve by even more than projected. One common factor was that programs tended to project declines in official transfers, while in fact these transfers fluctuated around a roughly constant trend for most of the period under review. We also found that under the 1994 and 1996 EFFs, an important factor was that import levels (in relation to GDP), which were projected to rise, actually fell. Under the more recent arrangements (i.e., the 1999 EFF and the 2002 SBA), stronger-than-projected growth in exports and remittances were also important factors.

The way in which this external adjustment was brought about—in particular, the composition between the “burden” borne by fiscal adjustment and changes in an approximate measure of the private sector savings-investment balance—often differed markedly from that which underlay the original program design. The most notable findings from this analysis were as follows (see Table 2.2 for a summary and Annex 13, Table A13.3 and Table A13.4 for details):

  • Over the period of the two early SBAs combined, the fiscal adjustment eventually exceeded its target (albeit with all of the adjustment concentrated in one year, 1992, when growth rebounded). The private sector savings-investment balance declined significantly, but by less than programmed.
  • Over the period of the three EFFs, none of the targeted fiscal adjustment was achieved while the external current account deficit continued to narrow. Contrary to the program strategy, all of this adjustment came from a large increase in the private sector savings-investment balance.
  • Under the 2002 SBA, both the external current account and fiscal deficits widened by more than targeted. The private sector savings-investment balance was higher than programmed.

Program documents were largely silent on the divergences from the implicit structure of adjustment that had underlain program design and said nothing about what the implications might be for future design. (The 2002 SBA was a partial exception.) Private savings and investment are extremely difficult to predict in the short term, especially at times of considerable economic uncertainty, so errors in projections are not especially surprising. However, it is striking that there was no analysis of the issue and the implications for design of subsequent programs.1

Debt burden indicators are on declining paths

Key indicators suggest that Jordan has made substantial progress toward debt sustainability. The external debt/GDP ratio and the debt-service ratios have both fallen significantly (Figure 3.3), and recent debt sustainability analyses carried out by IMF staff indicated that the debt/GDP ratio was likely to continue on a downward path in line with the official medium-term debt targets that have now been incorporated into Jordanian law (see Box 2.1).2 Various stress tests undertaken by the staff also suggest that the debt/GDP path would be fairly resilient to all but very large and compound shocks, of which a large exchange rate shock was expected to be the most severe.

Figure 3.3.External Debt and Debt Servicing

Source: IMF.

In terms of cross-country comparisons, Jordan has made substantial progress in reducing its debt and debt-service burdens, but they still remain significantly higher than the average for other lower-middle-income countries or other countries in the region (Table 3.1).

Table 3.1.Cross-Country Comparison of Debt Burden Indicators1
In Percent of Gross National IncomeIn Percent of Exports of

Goods and Services
199020002003199020002003
Total external debt
Jordan219.087.085.0335.0208.1189.7
Low-income developing countries56.045.040.0396.0214.0180.0
Lower-middle-income countries27.038.034.0139.0133.8105.0
Middle-income countries30.037.236.0132.0116.0101.0
Middle East and North Africa33.022.021.0115.071.064.0
Total debt service
Jordan16.58.811.820.412.616.4
Low-income developing countries4.03.03.020.013.013.0
Lower-middle-income countries4.03.03.018.010.08.0
Middle-income countries3.03.03.015.010.08.0
Middle East and North Africa4.03.012.08.0
Sources: World Bank (2005).

There are some differences (magnitudes, but not trends) between data presented here and the time series data in Figure 3.3—owing to differences in sources.

Sources: World Bank (2005).

There are some differences (magnitudes, but not trends) between data presented here and the time series data in Figure 3.3—owing to differences in sources.

Economic growth has been the main factor driving the decline in the debt/GDP ratio. The value of the stock of public and publicly guaranteed external debt at end-2003 was approximately the same as the level at end-1990—about $7.5 billion, while its ratio to GDP fell to 76 percent from 180 percent over the same period (Annex 13, Table A13.5). This suggests that new borrowings offset the impact of various debt cancellations and debt reduction schemes that Jordan benefited from during the 1990s. During most of the period under review, the authorities sought to build up international reserves to support the exchange rate peg and as a matter of deliberate policy favored external financing over domestic financing of the government budget deficit.

External balance has become less vulnerable to the impact of shocks but adjustment to any future decline in grants would still be crucial

Jordan’s geopolitical situation makes it susceptible to a wide range of potential shocks. Several developments in the structure of the economy suggest that it is now less vulnerable to these shocks than it was in the late 1980s: (1) significantly greater diversification in the composition and geographic destination of its exports; (2) a markedly lower debt-service burden; and (3) a large cushion of external reserves (equivalent to over nine months of current merchandise imports in 2004). However, the continuing importance of foreign aid and remittances in the balance of payments invites comparisons with the late 1980s and early 1990s when these inflows fell sharply on account of external economic and political shocks.

The IMF staff’s medium-term projections underlying the post-program monitoring framework project a decline of about $1 billion in grants between 2005 and 2007 (equivalent to almost 10 percent of GDP) while growth increases from 5 percent to 6 percent a year. On the balance of payments side that decline in grants is projected to be compensated by a strong growth in exports, portfolio flows, and remittances that allows an increase in imports. However, on the fiscal side, projections assume that public expenditures decline by the same percentage points of GDP as grants—from 38 percent to a historically low level of 28 percent of GDP. Whether such massive fiscal adjustment is feasible and the type of measures necessary to achieve are not discussed—other than references to “fiscal measures to be identified” in the fiscal tables. A simple scenario that assumes that the baseline framework does not adjust to the projected decline in grants and that external reserves thereby bear the full burden of this decline (which is admittedly a crude assumption and may well represent the outer range of the risks) suggests that the stock of reserves would be halved in about three years (i.e., by about 2008). This exercise suggests that the substantial reserve cushion would indeed give Jordan some time to adjust to even quite large shocks. However, that time would still be limited.

Limited Progress Toward Fiscal Sustainability

Fiscal outturns through the long period of IMF program involvement can be summarized as follows (see Annex 9 for a more detailed discussion):

  • A massive reduction in the fiscal deficit occurred during the period of the first two SBAs, far exceeding program targets (Figure 3.4 and Table 3.2).3 This adjustment occurred despite program slippages in the first SBA, and all took place in 1992 as growth rebounded from the crisis-induced recession.
  • Throughout the period of the three EFFs, there was essentially no further reduction in fiscal deficits (before or after grants) and deficit outcomes consistently fell short of program targets.
  • The period of the 2002 SBA was characterized by a widening fiscal deficit (before grants), reflecting increased expenditures. The deficit was more than covered by increases in grants to levels that exceeded substantially the average levels of the preceding decade.

Figure 3.4.Projections and Outturns of Fiscal Indicators

(In percent of GDP)

Source: IMF.

Table 3.2.Envisaged Versus Actual Fiscal Adjustment(Change over horizon of arrangement, in percentage points of GDP)
Arrangement/Horizon/VariablesLevel at (t−1)1Envisaged

Adjustment
Actual

Adjustment
Shortfall (−)/

Excess (+)
1989 SBA (t=1989)
Horizon: (t−1) to (t+1)
Fiscal balance excluding grants–23.77.75.2–2.5
Revenues31.80.23.73.5
Expenditures55.5–7.5–1.5–6.0
Grants7.3–0.32.42.7
Fiscal balance including grants–16.47.47.60.2
1992 SBA (t=1992)
Horizon:(t−1) to (t+1)
Fiscal balance excluding grants–18.07.011.84.8
Revenues26.40.61.81.2
Expenditures44.4–6.4–10.03.6
Grants7.3–5.2–4.40.8
Fiscal balance including grants–10.71.87.45.6
1994 EFF (t=1994)
Horizon:(t−1) to (t+1)
Fiscal balance excluding grants–6.42.10.5–1.6
Revenues32.4–0.5–0.20.3
Expenditures38.8–2.6–0.7–1.9
Grants4.6–1.2–0.60.6
Fiscal balance including grants–1.80.9–0.1–1.0
1996 EFF (t=1996)2
Horizon:(t−1) to (t+2)
Fiscal balance excluding grants–4.92.3–4.2–6.5
Revenues30.9–0.6–5.0–4.4
Expenditures35.8–2.9–0.8–2.1
Grants3.8–1.20.01.2
Fiscal balance including grants–1.11.1–4.2–5.3
1999 EFF (t=1999)
Horizon:(t−1) to (t+2)
Fiscal balance excluding grants–10.36.31.5–4.8
Revenues28.32.30.0–2.3
Expenditures38.6–4.0–1.5–2.5
Grants3.9–1.30.72.0
Fiscal balance including grants–6.45.02.3–2.7
2002 SBA (t=2002)
Horizon:(t−1) to (t+2)
Fiscal balance excluding grants–8.1–1.4–3.4–2.0
Revenues26.4–1.10.81.9
Expenditures34.50.34.2–3.9
Grants4.41.25.94.7
Fiscal balance including grants–3.7–0.22.52.7
Sources: IMF and IEO calculations.

Level at (t−1) as estimated at time of program formulation, in percent of GDP.

The comparison between envisaged and actual adjustment during the 1996 EFF is especially difficult because of the large data revisions that took place after the program was finalized.

Sources: IMF and IEO calculations.

Level at (t−1) as estimated at time of program formulation, in percent of GDP.

The comparison between envisaged and actual adjustment during the 1996 EFF is especially difficult because of the large data revisions that took place after the program was finalized.

The composition of revenues changed broadly in line with the aims of the fiscal structural reforms, but attempts to synchronize the tax and import tariff reforms in order to prevent an erosion of revenues in relation to GDP failed to achieve their objective. The earlier consumption tax was replaced by a more broadly based GST whose coverage expanded over time, but at a slower pace than anticipated. Income tax reforms lowered marginal tax rates and reduced the number of tax bands, but efforts to limit widespread tax exemptions and deductions were only partially successful.

After the initial sharp reduction in the early 1990s, there has been limited progress in expenditure consolidation. Total government expenditures have hovered around 35 percent of GDP until the recent grant-induced pickup. Many of the earlier expenditure rigidities remain, although there has been progress on some:

  • The system of open-ended food subsidies that developed in the aftermath of the 1988–89 crisis was rationalized and eventually replaced by a system of targeted direct support to the poor.
  • Significant changes were eventually made to the two public pension systems—for the civil service (beginning with the 1996 EFF), and for the military (mainly under the 2002 SBA). However, the most important changes came relatively late in the period of IMF program involvement and their impact will only be felt gradually over the medium term.
  • Little progress was made in addressing two other major rigidities—the government wage bill and military expenditures.

A key remaining challenge for fiscal management in Jordan is to define a “normal” level of grants around which medium-term expenditure plans can be built, in order to avoid a situation in which expenditures increase with aid flows but are difficult to cut when aid flows fall significantly. A 2004 Public Expenditure Review conducted jointly by the World Bank and the authorities also highlighted the need to develop a framework for expenditure policy that goes beyond short-term perspectives, and which incorporates principles to guide fiscal policy over the long term, including parameters for defining the optimal level of government spending and debt.

Over the years, visits by IMF teams to conduct program negotiations or undertake Article IV consultations provided the main occasions for comprehensive reviews of fiscal developments and policies going forward. Based on interviews in Jordan and discussions with IMF staff in Washington, the eval uation team formed the view that the authorities—especially Ministers of Finance—came to rely on such visits to integrate information from various fiscal agencies to form a coherent picture of recent developments, that served as a basis for policy discussions. A macrofiscal unit established in the Ministry of Finance in 2004 is designed to strengthen the ministry’s macrofiscal analytical capacity in order to undertake on a routine basis some of the activities that the authorities had come to rely on the IMF to provide. During the evaluation team’s visit to Jordan it learned that the 2004 budget cycle was the first time the authorities employed a medium-term expenditure framework to establish initial allocations contained in budget guidelines issued by the Ministry of Finance to spending ministries/units.

Success in Achieving and Maintaining Low Inflation

Under the IMF-supported programs, inflation outcomes have tended to be lower than targeted (Figure 3.5).4 Prior to the explicit pegging of the dinar to the U.S. dollar, this reflected tighter monetary conditions than projected from 1992 onward. With the exchange rate serving as the nominal anchor from the mid-1990s, inflation continued to be lower than projected, even in periods when broad money grew faster than projected.

Figure 3.5.Inflation 1

(In percent of GDP)

Sources: IMF.

1 The targets under the 1999 EFF and the 2002 SBA were in relation to the CPI. Earlier targets were in relation to the GDP deflator.

Was monetary policy too tight?

A common criticism the evaluation team heard in Jordan was that the CBJ’s monetary policy—supported by the IMF—had focused excessively on accumulating international reserves, especially in the mid- to late 1990s, and that the high interest rates that had resulted harmed growth. With the exchange rate pegged to the dollar, the principal monetary target under programs related to international reserves. During periods of reserve loss, the CBJ used interest rate policy to support the peg (through increasing the yields of certificates of deposits (CDs)). But from about 1999, the CBJ allowed a steady decline in the CD yields by deliberately sterilizing only a small portion of excess liquidity associated with foreign exchange inflows. The aim of this strategy was to encourage commercial banks to do more lending to the private sector in order to stimulate economic activity. To a large degree, this policy was facilitated by a steady decline in U.S. interest rates from mid-2000 (Figure 3.6). It is noteworthy that international reserves continued to increase even as the CBJ eased monetary policy.

Figure 3.6.Jordanian Dinar and U.S. Dollar Interest Rates and International Reserves

Sources: Annual Reports of the Central Bank of Jordan; and IMF.

The authorities too raised concerns about the potential adverse impact of tight monetary policy on growth during the 1996 Article IV consultations. The staff response at the time included an argument that monetary policy was being tightened in part to compensate for weak fiscal performance—suggesting that if fiscal policy had been more restrained, it would not have been necessary to raise interest rates as high and to hold them at those high levels for as long as the monetary authorities did in order to support the exchange rate peg.

Beyond the issue of monetary-fiscal policy mix, is there a sense in which monetary policy could be said to have promoted “excessive” accumulation of international reserves? For emerging market economies with managed floating or fixed exchange rate regimes, Wijnholds and Kapteyn (2001) suggest a yardstick for gauging the adequacy of international reserves, based on two potential sources of capital flight that reserves should aim to cover: short-term external debt and internally generated capital flight which the authors proxy by a portion of broad money.5 Based on this approach, a time series of “adequate” levels of international reserves for Jordan was derived and compared with the actual levels of reserves (Figure 3.7).

Figure 3.7.“Adequate” Versus Actual Level of Reserves

(In millions of U.S. dollars)

Sources: IMF program documents and IEO calculations.

The results suggest that up to 1998, the level of reserves fluctuated around adequate levels—as suggested by the yardstick—but did not consistently exceed them. Since 1999, reserves have remained consistently above the calculated “adequate” levels. However, this period coincides with the period of monetary easing by the CBJ. Moreover, from 2003 onward, building some additional cushion of reserves in response to the pickup in grants—which was not expected to be permanent—seems appropriate. Therefore, given the constraints on independent monetary policy imposed by the exchange rate peg, it is difficult to argue that the CBJ could have done significantly more to ease monetary policy.

Growth Performance Was Strongly Influenced by Factors External to the Programs

Studies of Jordan have highlighted the sharp fluctuations of external flows in the form of aid from Arab countries and remittances as a major determinant of the volatility in its growth rates (Figure 3.8).6 For example, adverse shocks to these flows and the disruption caused by events leading up to the 1991 war in the Middle East were key factors in the poor growth performance from the mid-1980s through 1991. A spike in growth in 1992 took most observers by surprise, since the repatriation of some 300,000 Jordanians from neighboring countries was originally perceived as a negative shock to the economy. However, the returnees came back with substantial savings, and used some of these to spark a construction-led boom that lasted until the mid-1990s. Similarly, the most recent spurt in growth (7.5 percent in 2004) has also been the result of several geopolitical developments associated with the conflict in Iraq. Foreign aid increased sharply as did activities related to the reconstruction of Iraq.

Figure 3.8.Real GDP Growth

(In percent a year)

Sources: IMF program documents.

This volatility arising from external events makes it particularly difficult to assess the extent to which the reform programs supported by the IMF (in conjunction with the World Bank) may have contributed to long-term growth. There are at least two mechanisms through which the reforms may have affected long-term growth. First, they could have improved the resilience of the Jordanian economy to adverse shocks, thereby reducing the likelihood of the sharp output collapses seen in the 1980s. Second, they may have contributed to long-run capacity for growth through improving the incentives for investment, productivity growth, and improved resource reallocation.

As Figure 3.8 shows, output has become more stable during the last decade—averaging about 4 percent a year. The absence of major negative shocks comparable to those of the late 1980s may partly explain this outcome. However, an improved policy regime also appears to have been important; for example, the uncertainty created by the long illness of the late King Hussein seems not to have been translated into instability in financial markets.

In terms of factors influencing long-term growth, three elements are worth noting. First, exports have become an important source of growth. They almost tripled between 1990 and 2003. A large part of that expansion was fueled by exports to the United States, especially from the QIZs, for which the United States provided preferential access. It can be argued that the expansion of exports is more the result of such status than the long-term improvement in the overall trade regime associated with the reforms undertaken by Jordan during the 1990s. However, exports to the rest of the world—emanating from outside the QIZs—also performed well, doubling during the period. This performance suggests a positive impact of the overall improvement in the country’s trade regime.

Second, a recent growth accounting exercise by staff to update some earlier studies done at the World Bank in 1994 and 2001 suggested a modest rebound in productivity growth in the period 2001–02.7 This may also be explained by increased productivity of the export sector that lifted overall productivity.

Third, while private investment did rebound during the mid-1990s, this appears to have been almost entirely due to the surge in housing investment by repatriated workers (Figure 3.9). Private nonresidential fixed investment fell during the 1988 crisis and recovered gradually thereafter—although to levels that were still low in comparison with earlier decades. The positive dimension is that the recovery seems to have been sustained during the 1990s without major fluctuations. An improved macroeconomic situation and a more transparent pricing and open trade regime may have helped the recovery.

Figure 3.9.Trends in Private Investment

(In percent of GDP)

Comparing growth targets and outcomes, program projections for Jordan did not exhibit a marked trend toward excessive optimism that has been noted in IMF-wide reviews on the subject (Figure 3.10).

Figure 3.10.Real GDP Growth

(In percent a year)

Sources: IMF program documents and IEO calculations.

Mixed Record on Program Implementation and Effectiveness of Conditionality

Thus far, this review has concentrated on the economic policy content of programs. It now turns to issues related to program implementation, including the effectiveness of conditionality. It starts with a look at evidence on the incidence of program interruptions, proceeds to analyze compliance with conditionality, and ends with a discussion of tensions between conditionality and ownership and how these were resolved.

Program interruptions

A review of program documents indicated two clear-cut cases of program interruptions: (1) the 1989 SBA which became inactive (in 1990) after only one disbursement, in the wake of Iraq’s invasion of Kuwait and the ensuing war in the Middle East; and (2) the 1996 EFF, which was allowed to lapse in 1998 in the aftermath of data revisions that cast doubt on the credibility of program monitoring (Box 3.1).

Timeliness of completing program reviews was used as a yardstick for identifying other incidents of (“nonfatal”) program interruptions.8 Timely completion of reviews was interpreted as indicating that a program is being implemented broadly as envisaged or is judged to be “on track” to achieve its objectives. Conversely, a long delay in completing a review suggested the existence of a problem—and triggered further investigation into the cause(s) of the delay. In order to exclude minor delays, and noting that nearly all reviews under Jordan’s IMF-supported programs were scheduled to be completed at approximately six-month intervals, cases where it took 12 months or more to complete a review were classified as program interruptions.

The above yardstick yielded two cases of nontimely completion of reviews. The first was a 13-month gap between the second and third reviews under the 1999 EFF, reflecting delay in the implementation of measures (e.g., adjustment of petroleum product prices) that the authorities and IMF staff had agreed on to close an emerging fiscal gap in 2001. The second was a 12-month hiatus between approval of the 2002 SBA and the first review. This was due to an extra round of discussions to take account of the impact of the 2003 war in Iraq.

All told, two war-related interruptions (in 1990 and 2002–03), and two interruptions due to problems with program monitoring (1998) or implementation (2001) were identified. There was only one case in which delay in completing a program review was linked to the implementation of an agreed measure (i.e., conditionality).

Effectiveness of conditionality

Macroeconomic conditionality

Compliance with macroeconomic conditionality was very high overall with the exception of fiscal performance criteria. The targets on the contracting of nonconcessional external debt and on net international reserves were always met. Similarly, except in 1993, the target for monitoring liquidity in the banking system (i.e., net domestic assets of the system or of the CBJ) was routinely met, often by large margins; the breaches of the targets in 1993 reflected stronger than expected expansion in credit to the private sector as well as data revisions that took place after the targets were set.

The key fiscal performance criteria were on some measure of bank financing of the fiscal deficit until the 1999 EFF, after which a PC on the overall deficit was introduced.9 Discussions with staff and officials indicated that the shift to the broader measure of the deficit was beneficial in terms of facilitating improved monitoring—not in the sense of yielding any immediate results in terms of better final outcomes but by bringing greater focus to the overall magnitude of any fiscal slippages and thereby to negotiations on remedial measures.

Jordan received more than the average share of waivers compared to other countries but the number does not appear to have been excessive. This result is based on a comparison of Jordan data with averages reported in IEO (2002) for the number of waivers per program year, and the share of the total number of quantitative performance criteria waived. Focusing only on periods when programs were active, it is estimated that the average number of waivers per program year (0.7) is similar to the average for nonconcessional arrangements (0.6), and that the share of performance criteria waived for Jordan (about 4 percent) is about double the average for nonconcessional arrangements (2.1 percent), but still low.

Structural conditionality

Most structural measures were implemented—although in a number of cases with considerable delay (see Annex 10). The delays were most prevalent in the fiscal area, especially with respect to tax laws. For example, the introduction of the GST, which was a performance criterion under the 1992 SBA, was waived twice. It was converted to a prior action for the 1994 EFF, and was eventually implemented in June 1994. Subsequently, a waiver was again required when it came to converting the GST to a VAT in 2000.

Conditionality related to the reform of direct taxes was not as successful in achieving its goals. At the urging of IMF staff, the authorities submitted a draft amendment of the income tax law to parliament in 2000 that sought to broaden the tax base by reducing exemptions. By the time parliament passed the amendment, the list of exemptions had increased rather than decreased. Several officials involved in the process said that the amendment had been rushed to parliament to meet a deadline in the IMF-supported program. They argued that if they had had more time to have substantive consultations with members of parliament before the formal submission of the draft to parliament, the outcome might have been more favorable. All in all, at about 3 percent of GDP, the contribution of direct taxes remains low by international standards. Staff estimates that if appropriate reforms are implemented, direct taxes could increase overall revenues by about 1½–2 percent of GDP over the medium term.

Box 3.1.The 1998 Data Misreporting Episode

During May and June 1998, IMF staff learned of major revisions to official estimates of real GDP growth and fiscal accounts that painted a very different picture of economic performance in 1996 and 1997 than that previously discussed by the Executive Board. The Executive Board had recently discussed Jordan (April 1998) in the context of the 1998 Article IV consultation and completion of the fourth review under the prevailing EFF.

The data revisions sharply reduced growth estimates from 5 percent a year to 0.3 percent and 3 percent in 1996 and 1997, respectively, and substantially raised the fiscal deficit (excluding grants) from 4–5 percent of GDP to 7–8 percent. Apart from the impact of the downward revision in GDP growth, the revised fiscal positions also reflected significant changes to revenue and financing data.

The revisions cast serious doubt on the effectiveness of the EFF-supported program and raised fundamental questions about the integrity of the system in place for monitoring performance under the program. IMF management called for a series of internal reviews and subsequently commissioned an independent inquiry into the episode. Meanwhile IMF staff and the authorities were unable to agree on measures to bring the program back on track; no further reviews (beyond the fourth) were completed, and the EFF was allowed to lapse. However, the IMF scaled up its technical assistance to Jordan in the areas of fiscal management and real sector statistics to help the authorities begin to address institutional weaknesses that may have contributed to the problem. Subsequently, a new EFF arrangement was approved by the Executive Board in April 1999, before the internal reviews were completed (and before the independent inquiry had got off the ground).

The reports on the episode were presented to the Executive Board in October 1999 at the time of the completion of the first review under the new EFF. The reports found that available updated information was not provided to IMF staff in a timely manner by the authorities, but also that IMF staff had gone along with noncompliance with reporting requirements (for example, a requirement for monthly reporting of fiscal data to the IMF had not been implemented). The internal reviews called for more vigilance by staff in monitoring the quality of data received from countries, and for correcting deficiencies in the design of program monitoring.

Following Executive Board consideration of the reports, the IMF issued a news brief, part of which read: “Directors…discussed the problems which had come to light concerning misreporting of fiscal and national accounts data in 1996–98 under the previous extended arrangement. They were very concerned both by the misreporting itself and because it had contributed to a belated policy response by the authorities and a weakening of economic performance in Jordan. However, Directors were reassured by the recent improvements in the compilation of statistics and in the monitoring of fiscal accounts; the timeliness with which they are now reported to the Fund and released to the public; and the authorities’ commitment to maintaining the current open and cooperative relationship with the Fund.”

The authorities voluntarily repaid the IMF the last disbursement made under the 1996 EFF.

In discussions with the evaluation team, a number of former Jordanian officials claimed that the reports presented to the Board about the episode were unfair in laying the bulk of the blame on the authorities. They emphasized their reliance on IMF staff for updating the macroeconomic framework for the programs and for the government’s budget. Moreover, they pointed out (as did one of the reports to the Executive Board) that some of the updated information on fiscal performance in 1996 and 1997 had been presented to parliament in December 1997 and was part of the Budget Law that was passed in early 1998, well before the Board discussion of the 1998 Article IV consultation and the fourth review under the EFF. It is not clear why information that was in the public domain had not been offered to IMF staff, nor why IMF staff had not been able to pick it up on its own from the Budget Law.

What was the impact of the misreporting episode on IMF policy advice? Given the extent of staff concerns about a debt overhang at the time, it is likely the IMF would have prescribed greater fiscal adjustment if the available data had indicated a larger fiscal imbalance. It is unlikely that evidence of much lower growth performance would have exerted much influence on IMF policy advice. As noted in Chapter 2, although the 1999 EFF discussed a countercyclical role for fiscal policy, it still proceeded to call for a large fiscal adjustment despite a recognition that the economy had been in a recession for a couple of years. In any case, IMF staff was probably aware of widespread skepticism about official growth estimates in Jordan. It is noteworthy that a group of Executive Directors who visited Jordan in 1996 reported that they encountered a widespread view that the economy was growing at much less than the IMF-endorsed official rate of about 6 percent a year at the time. Also, country reports on Jordan prepared by the Economist Intelligence Unit in the second half of the 1990s routinely characterized official growth estimates as being too rosy.

Tariff reforms featured prominently in structural conditionality and implementation was successful, although it is difficult to attribute this to IMF conditionality. Regional as well as international impulses, including accession to the World Trade Organization and the free trade agreements with the European Union and the United States, seem to have been important drivers of the reforms. Trade liberalization also figured prominently as conditionality in many World Bank lending operations during the period.

Although the financial sector has undergone substantial reforms, very few of these were covered by conditionality in IMF-supported programs. Current and former officials of the CBJ interviewed suggested that the goal of reaching international standards provided enough motivation for undertaking many of the reforms. Interestingly, some of these reforms—for example, in the areas of monetary operations, banking supervision, and improvements to the payments system—were undertaken with the help of IMF technical assistance. A few officials suggested that tying technical assistance recommendations “too closely” to conditionality might have been counterproductive; instead of the reforms being driven by the authorities’ own perceived need to meet international standards, they would have been seen as having been imposed by the IMF.

For a long time IMF staff has urged the authorities to adopt a strategy for domestic fuel pricing that would link automatically domestic energy prices to world prices—and has provided technical assistance on the subject. It became a structural benchmark in the 1999 EFF, but it was not met. It was not reinstated as a condition in the 2002 SBA. This is an area where little progress has been made.

In contrast, solid progress was made in reforming the pension system, albeit with significant delays. The first step was to replace the civil service system by the private pension scheme administered by the SSC for all new hired civil servants. A critical step was implemented in 2003, where the previous military pension system was closed to all new entrants, hence completing the unification of the public and private sector pension systems.

During the evaluation team’s visit to Jordan, several current and former senior officials highlighted the importance of IMF conditionality for implementing reforms in a number of sensitive areas. They cited as examples, the reorientation of the tax system toward a broad-based domestic consumption tax (GST/VAT), rationalization of food subsidies, and pension reform. In each case, the recommendations of IMF technical assistance reports had provided important input which had been drawn on by the Minister of Finance in internal debates within the government, and sometimes in the government’s interaction with interest groups outside government. The GST experience was referred to often; the general view was that although it had been a long drawn-out process, it had eventually succeeded in achieving what it set out to do. Similarly, pension reform, which had also been a very slow process, was also characterized as an eventual success, although it was acknowledged that most of the benefits of the reform will come in the future.

Ownership

Drawing on the literature on “ownership” of economic reforms and programs, the evaluation endeavored to gauge the authorities’ ownership of Jordan’s IMF-supported programs, and whether this made a difference to program implementation.10 Four aspects of ownership were defined: (1) contribution of the authorities to program formulation, including the choice of measures to be subject to conditionality; (2) alignment of the policy content of programs with national development plans/programs and with annual budgets; (3) involvement of other stakeholders; and (4) support of the political leadership of the country.

We found that the process of formulation of Jordan’s IMF-supported programs was highly centralized around the Minister of Finance, but with significant input from the CBJ. We also learned that IMF staff always prepared the first draft of the Letter of Intent (LOI), which then served as the basis for negotiations.11 Based on interviews with Jordanian officials and IMF staff, the evaluation team formed the view that the authorities did not have much control over the formulation of the early programs. However, by about the time of the 1996 EFF, the CBJ had developed enough confidence and expertise to exert significant influence over the monetary and exchange rate policy components of the programs. It would take another few years (until the time of the 1999 EFF) before the authorities would have the same influence over the fiscal component.

A number of former and current officials we interviewed in Jordan expressed the view that often disagreements with IMF staff on structural reforms revolved around the preparation required and the pace of implementation, rather than on the desirability of the reforms per se. This view is given credence by the staff report on the request for the 1999 EFF which noted that IMF staff would have liked a faster pace of structural reforms, but that the authorities were concerned to commit to only measures and a pace they thought they would be able to deliver.

With respect to alignment of program content with national development plans, program objectives have been consistent with key objectives spelled out in the government’s Social and Economic Development Plans.12 In discussing the programs, IMF staff reports have sought to link macroeconomic objectives with the authorities’ broader development goals. This was especially the case for the 2002 SBA-supported program and the government’s PSET. Initially, IMF staff was concerned about the scale of the programs envisaged in the PSET and whether there would be enough “new” external grant financing for it. In the end, with the encouragement of IMF staff, the authorities worked with World Bank staff to incorporate the PSET into the government budget.

With respect to the third aspect of ownership—involvement of other stakeholders—several people interviewed by the evaluation team in Jordan made the point that where legislative action was needed to implement particular measures, it is important for the government to engage members of parliament in discussing proposals and options at an early stage, and that IMF staff could usefully contribute by explaining the rationale for its views to members of parliament and other interest groups, and in this way contribute to an informed consideration of the measures. Of course, engagement with the legislative branch would be expected to be at the invitation of the government.

Reflecting on economic reforms in Jordan, Hunaidi (2004) underscored the important role of parliament.13 She reported that implementation of many reforms involved extensive negotiations with parliament, and that this often led to modifications of understandings about draft laws and policies reached between the government and the IMF or World Bank. In her view, this engagement was critical to sustaining the momentum for reforms. She compares the fallout from two social unrests related to the removal of subsidies—the “fuel riots” of 1989 and the “bread riots” of 1996—and attributes the more muted impact of the latter to the government’s active engagement of parliament in the formulation of the measure. She argues that by contrast, the government was seen as acting alone in 1989, when the price of petroleum products (highly subsidized at the time) was raised substantially, and triggered riots in major cities. She notes that part of the problem with the 1989 measure was that it had been taken in haste, without a thorough assessment of its implications on other sectors.14

An analysis of the political economy of the introduction of the GST in Jordan comes to a similar conclusion (see Annex 11). Introduction of the GST was controversial, with a number of stakeholder groups seeing it as an imposition from abroad. The initial failure to communicate adequately the economic rationale for the GST to broader groups contributed to some of the difficulties encountered in the initial introduction of the tax, but greater efforts to broaden the debate, both inside and outside parliament, on why the tax was needed facilitated subsequent amendments.

Support of the programs by the political leadership was difficult to gauge. Some commentators have suggested that, for the most part, the political leadership remained “above the fray,” leaving the reforms to be carried out by technocrats.15 They note that the situation changed with the ascension of King Abdullah, who has taken an active interest and leadership role on economic issues. Moreover, in a survey of stakeholders in Jordan that we commissioned as part of this evaluation, a large proportion of respondents indicated that key political leaders publicly voiced support for Jordan’s IMF-supported programs (Annex 12)—a result that is at odds with the view that the political leadership was on the sidelines.

Poverty

Measuring trends in poverty is complicated by differences in views on what is the appropriate measuring rod and by difficulties in obtaining a consistent time series. Nevertheless, several broad conclusions are suggested by the various measures (Table 3.3). First, poverty rates rose markedly as a result of the 1988–89 crisis, with the number of people below the national poverty line rising fivefold. Inequality also seems to have increased. Second, the share of the population below the poverty line declined in the mid-1990s but this improvement subsequently tapered off (and may have been partly reversed) leaving overall poverty rates still well above precrisis levels. However, the share of the population living in extreme poverty (i.e., those living on less than $1 a day) did continue to decline in the 1990s, from 6.6 percent in 1992 to 4 percent in 2002—probably reflecting the better targeting of the social safety net support that replaced more generalized food subsidies.

Table 3.3.Poverty Indicators
198719921997199912001120042
Poverty line (current annual per capita income in Jordanian dinars)148.0261.0313.5313.5312.0436.0
Poverty measures
Head count index (percent of population)33.014.411.713.613.515–30
Poverty gap index (percent)40.33.62.53.033.01
Number of poor people (in thousands)87.4554.5538.5668709
Inequality (national)50.360.430.360.360.36
Sources: World Bank (2004) and USAID (2005).

World Bank projections.

USAID estimates.

The head count index is the percentage of the population living below the poverty line deemed appropriate for the country by the authorities.

The poverty gap index measures the depth of poverty in a country or locality. It increases with the distance of the poor from the poverty line.

Income inequality refers to the extent of disparity between high and low incomes. The higher the ratio (Gini coefficient), the greater the extent of inequality.

Sources: World Bank (2004) and USAID (2005).

World Bank projections.

USAID estimates.

The head count index is the percentage of the population living below the poverty line deemed appropriate for the country by the authorities.

The poverty gap index measures the depth of poverty in a country or locality. It increases with the distance of the poor from the poverty line.

Income inequality refers to the extent of disparity between high and low incomes. The higher the ratio (Gini coefficient), the greater the extent of inequality.

1

Other IEO evaluations have shown that the difficulty of predicting private sector response in program design, and the tendency not to step back and reconsider the hypotheses that underline that design, are frequent occurrences in IMF arrangements, so Jordan is not especially unusual in this respect. See, for example, Evaluation of Prolonged Use of IMF Resources (IEO, 2002), and Fiscal Adjustment in IMF-Supported Programs (IEO, 2003).

2

IMF, 2003a (updated in IMF, 2004b).

3

The outturns shown in the graphs are based on the most recently available data. The “projections” contain the outcome for the year preceding the first program year (i.e., year t–1) as reported at the time of program formulation. This is done in order to convey a sense of the order of magnitude of adjustment envisaged at the time of program formulation. Because of subsequent data revisions (especially to the national accounts), there are a few large discrepancies in the starting points between “projections” and “actuals.”

4

Except for the upward spike in 1989, which largely reflected an underestimation of the price increases that were already in the pipeline from the preceding sharp depreciation of the exchange rate. Inflation declined quickly as the exchange rate stabilized.

5

The indicator is calculated as the sum of short-term external debt and a risk-adjusted proportion of broad money. For the latter component, the authors suggest 10–20 percent of broad money for countries with a managed float or fixed exchange rate regime. The application to Jordan below, used 20 percent of broad money and applied a country risk factor of 1 (the highest).

7

The methodology used is a growth accounting framework with physical and human capital as well as labor force growth (see IMF, 2004a).

8

Program reviews are an important part of the mechanism for monitoring performance under IMF-supported programs, and for the recalibration of program targets/projections to reflect recent developments. For a discussion, see Mussa and Savastano (1999).

9

The 1999 EFF had both. The fiscal deficit PC was on a “beforegrants” basis under the 1999 EFF but switched to an “after-grants” basis for the 2002 SBA.

10

The literature drawn on included Johnson and Wasty (1993), and Boughton (2003).

11

Senior government officials interviewed in Jordan claimed that this did not detract from their “ownership” of the contents of the LOI.

12

Brief summaries of the objectives and macroeconomic policies contained in the three most recent plans are presented in Annex 7.

13

The author is a former Minister of Planning in Jordan.

14

For example, there was no adjustment in transport tariffs, which were also controlled at the time, even though the cost of transportation had gone up substantially. See also Ryan (1998).

15

See, for example, Sud (2005).

    Other Resources Citing This Publication