Too much, too soon: IMF conditionality and inflation targeting

by Gerald Epstein, University of Massachusetts

11 September 2006 | At Issue

A shorter, fully-formatted PDF version of this briefing is also available.

Gerald Epstein, co-director of the Political Economy Research Institute at the University of Massachusetts, finds that despite little evidence of the success of inflation targeting in promoting economic growth, employment creation or poverty reduction, the IMF is increasingly using loan conditions and technical assistance to push its use. There is an urgent need for viable alternatives that focus on employment generation, poverty reduction, export promotion and investment enhancement to be given more attention.

Ironically, employment creation has dropped off the direct agenda of most central banks just as the problems of global unemployment, underemployment and poverty are taking center stage as critical world issues1. The International Labour Organization estimates that in 2003, approximately 186 million people were jobless, the highest level ever recorded2. In addition, IMF economists estimate that economic growth needs to be sustained at seven per cent per year or more to reach the millennium development goal of reducing poverty by half by 20153.

the IMF-sponsored approach has narrowed both the goals and the tools of monetary policy

Yet, for the past decade or more, so-called ‘global best practice’ approach to central banking has not focused on economic growth or employment generation; instead, the best practice promoted by the IMF and other organisations promotes formal or informal ‘inflation-targeting’, in which keeping a low rate of inflation – in the low single digits – has been proposed as the dominant and often exclusive target of monetary policy. In this inflation-focused monetary policy, other important goals, such as rapid economic growth and employment creation, are seen as inappropriate direct targets of central bank policy; rather they are viewed as hoped for – even presumed – by-products of an inflation focused approach to monetary policy4. Thus, according to this orthodox approach to monetary policy, the focus of policy is on stabilisation, rather than growth or development, with an implicit assumption that once stabilisation is achieved, economic growth, employment creation, and poverty reduction will follow.

This orthodox view not only specifies the appropriate target of monetary policy, but also the appropriate tools or instruments. The orthodox approach has emphasised indirect, market-based instruments of policy, such as short term interest rates, as the primary and often exclusive tool of monetary policy5. This is in contrast to the ‘direct’, quantitative tools often used by central banks which have involved credit allocation methods, interest rate ceilings, and other ways to direct credit to priority economic sectors and goals. In short, the IMF-sponsored orthodox approach has narrowed both the goals and the tools of monetary policy.

After several decades of experience with this inflation-focused market-based approach, the policy record has been rather disappointing for many countries. In a number of countries, inflation has come down, to be sure, but it is questionable to what extent the drop in inflation is due to changes in domestic monetary policy, rather than the overall global fall in inflation6. But even if domestic monetary policy has reduced inflation, the hoped-for gains in employment have generally not materialized; and, for many countries following this orthodox approach, economic growth has not significantly increased. The key point, then, is this: despite what the orthodox approach maintains, employment generation and economic growth, are not automatic by-products of stabilisation-focused central bank policy.

Surprisingly, despite a disappointing record, this almost single-minded focus on inflation is gaining a more secure foothold in monetary policy circles and the circles are widening to include an increasing number of developing countries. According to a recent report by the IMF, an increasing number of central banks in emerging markets are planning to adopt inflation targeting as their operating framework. An IMF staff survey of 88 non-industrial countries found that more than half expressed a desire to move to explicit or implicit quantitative inflation targets. Nearly three-quarters of these countries expressed an interest in moving to “fully-fledged” inflation targeting by 2010 (See Table 1).

Importantly, to support and encourage this movement, the IMF is providing technical assistance to many of these countries and is willing to provide more. In addition, the IMF is considering altering its conditionality and monitoring structures to include inflation targets. In short, despite little evidence concerning the success of inflation targeting in its promotion of economic growth, employment creation and poverty reduction, and mixed evidence at best that it actually reduces inflation itself, a substantial momentum is building up for fully-fledged inflation targeting in developing countries. Promotion efforts by the IMF and western trained economists are at least partly responsible for this increasing popularity.

Table 1: Inflation Targeting Countries, Current and Potential

Country When adopted Current inflation target per cent per annum Technical assistance requested from IMF
Israel 1997 1-3
Czech Republic 1998 3 (+/- 1)
Poland 1998 2.5 (+/- 1)
Brazil 1999 4.5 (+/- 2)
Chile 1999 2-4
Colombia 1999 5 (+/- .5)
South Africa 2000 3-6
Thailand 2000 0 – 3.5
Korea 2001 2.5-3.5
Mexico 2001 3 (+/-1)
Hungary 2001 3.5 (+/-1)
Peru 2002 2.5 (+/-1)
Philippines 2002 5-6
Slovak Rep. 2005 3.5 (+/- 1)
Indonesia 2005 5.5 (+/- 1)
Romania 2005 8.8
Costa Rica, Egypt, Turkey, Ukraine 1-2 years Yes
Albania, Armenia, Botswana Dominican Republic, Guatemala, Mauritius, Uganda 3-5 years Yes
Angola, Azerbaijan, Georgia, Guinea Morocco, Pakistan, Paraguay 3-5 years No
Belarus, China, Kenya, Kyrgyz Republic, Moldova, Serbia, Sri Lanka, Vietnam, Zambia more than 5 years Yes
Bolivia, Honduras, Nigeria, Papua New Guinea, Sudan, Tunisia, Uruguay, Venezuela No

Source: IMF, 2006, “Inflation Targeting and the IMF”, Tables 1 and 2.

While it might seem obvious that stabilisation-focused central bank policy represents the only proper role for central banks, in fact, looking at history casts serious doubt on this claim. Far from being the historical norm, this focus by central banks on stabilisation to the exclusion of development represents a sharp break from historical practice, not just in the developing world but also in the now developed countries as well7. In many of the successful currently developed countries, as well as in many developing countries in the post-Second World War period, development was seen as a crucial part of the central bank’s tasks. Now, by contrast, development has dropped off the ‘to-do list’ of central banks in most developing countries.

Inflation targeting and IMF conditionality

The promotion of inflation targeting has implications for the type of conditionality that the IMF imposes. As the IMF states: “Conditionality in Fund-supported programmes is intended primarily to ensure that Fund resources are used to support adjustment toward sustained external viability, and thereby to safeguard the capacity to repay the Fund. Traditionally, monetary conditionality consists of limits on monetary aggregates – specifically, a floor is set for the level of net international reserves (NIR) and a ceiling is established on the net domestic assets (NDA) or on base money.” 8

The IMF is concerned, however, that this NDA-NIR approach could allow for higher inflation than they might like, if, for example, larger than necessary increases in net international reserves result from inflows of capital. As a result, inflation targeting might require a further tightening of monetary conditions for countries undergoing IMF programs in order to maintain inflation rates in the low single digits. To the extent that such tightening slows employment growth even further, inflation targeting as part of IMF conditionality could have even more severe impacts on employment and poverty impacts for developing countries than the current NDA-NIR conditionality. This is illustrated by work James Heintz and I did in Ghana with a UNDP sponsored study.

A Ghana case study

In a recent report with my PERI colleague James Heintz, we studied the evolution of this conditionality and its impact in Ghana9.

In Ghana, monetary policy has been strictly limited by macroeconomic stabilisation agreements made with the IMF and World Bank in conjunction with IMF loans, the HIPC (Heavily Indebted Poor Country) initiative and the PRSP (Poverty Reduction Strategy Paper) process. In addition to the typical IMF requirements to limit credit to the government and increase foreign reserves, the Bank of Ghana has also been encouraged to focus on reducing inflation into the “low single digits.” This focus on inflation fighting and the other limitations imposed by IMF conditionality has reduced the prospects for rapid economic growth and broad-based employment generation in Ghana.

Monetary policy has been strongly influenced by the IMF Poverty Reduction and Growth Facility (PRGF) agreements and commitments undertaken by Ghana as well as commitments made in conjunction with the development of the Ghana Poverty Reduction Strategy, as part of the HIPC debt-relief initiative, and the Multi Donor Budgetary Support initiative. According to documents associated with these initiatives: “The focus of macro-stability in the medium term is to ensure prudent fiscal and monetary policy management to achieve price stability, maintain interest rate levels that are conducive to both savings and investments, and to ensure stable yet competitive exchange rates and a fiscally sustainable debt burden. The Government of Ghana with its development partners have defined a number of macroeconomic targets and measures.”10

Financial programming has been used since the 1970’s as part of the IMF’s lending programmes to least developed countries. This programming has now been folded into the PRSP and HIPC processes without much alteration. The programming uses a set of “identities” and extremely simple models (at best, a set of assumptions about the structure of the economy) to establish a set of targets that the IMF will monitor and the government will have to meet in order to receive the next installments of IMF loans, or qualify for HIPC relief and other donor support11.

The typical programme connects balance of payments constraints, the government fiscal deficit, and central bank policy in order to attempt to reduce indebtedness to a sustainable level, primarily by keeping economic growth in line with likely available foreign resources from export receipts, aid and capital inflows. Increasingly, reducing inflation into the low single-digits has become a central focus. Therefore, two key central assumptions of these programmes are (1) that inflation rates between 10 and 20 per cent are bad for economic growth and reducing inflation below that level will not reduce economic growth; and (2) that reducing government spending is good for the economy, because more government spending crowds out private investment.

The primary targets and measures have been structured by IMF financial programming. Under the standard financial programming methods implemented by the IMF, target ceilings are set for central bank monetary and credit expansion and floors are established on net foreign reserves12. The original motivation for these restrictions was to ensure the ability of countries to reduce their foreign debt and remain solvent, including protecting the ability of the IMF to get repaid. Recently, other goals, such as reducing inflation, increasing foreign exchange reserves and “creating room for private investment,” have been emphasised.

Table 2 illustrates how the programming works. As mentioned above, the main targets are net domestic assets ceilings (NDA) – sometimes called domestic credit ceilings – which directly limit the amount of credit that the Bank of Ghana can create, and net international reserve floors (NIR), which require monetary and fiscal policy to operate so to preserve a minimum level of international reserves. If either target is threatened – that is, if international reserves are too low or if net domestic assets are too high – then the policy calls for tightening monetary policy, usually raising interest rates or reducing monetary expansion.

A key and troubling implication of this programming approach is that there is no clear set of conditions under which expansionary monetary policies are called for, even in a situation of slow growth and high unemployment. Even if both targets are met, programming does not call for expansionary monetary policy. This is largely because there is no explicit operational target for economic growth, employment creation, or poverty reduction. The bias of financial programming is therefore highly contractionary.

Table 2: IMF financial programming with Net Domestic Asset and Net International Reserve targets

NDA higher than programmed (threatened) NDA lower than programmed (not threatened)
NIR higher than programmed (not Threatened) Only the NIR target has been met. Reflects overexpansion of money. Policy: tighten Both targets have been met. Policy: No need for tightening
NIR Lower than programmed (threatened) Neither target has been met. Policy: tighten. NIR has not been met. Policy: keep tight to stem further reserve losses

Source: Adapted from Blejer, et. al., (2002), Table 1.

In addition, inflation targets have recently been stressed as additional goals of IMF programmes. Inflation targeting (IT), or more generally, anti-inflation focused central banking, reinforces this contractionary approach to central banking by adding an additional contractionary constraint.

Traditional financial programming did not incorporate explicit inflation targets. Now, however, there are new commitments made by central banks to the IMF and associated organisations to reduce inflation. This essentially adds an additional restriction on central bank policy along with the traditional commitments with respect to domestic credit ceilings and reserve floors13.

As Table 3 demonstrates, even if the NDA ceiling is met, if the inflation target is not met, then the central bank will still be expected to tighten credit. There are many problems associated with financial programming, and even more when inflation targeting is added to it.

Table 3: Financial Programming with Net Domestic Asset and Inflation Targets

Inflation target (IT) higher than programmed (threatened) NDA lower than programmed (not threatened)
Actual Net Domestic Assets (NDA) relative to programme assumptions: Higher than programmed (Not threatened) NDA and IT give the same signal. Policy: tighten NDA and IT give different signals: NDA – tighten; IT – no tightening needed. Policy: tighten
Actual Net Domestic Assets (NDA) relative to programme assumptions: Lower than Programmed (Threatened) NDA and IT give different signals: IT – tighten; NDA – no tightening needed.Policy: tighten NDA and IT give the same signal. Policy: no tightening needed.

Source: Adapted from Blejer, et. al., (2002), Table 2.

There are many problems with financial programming that economists have highlighted:

  1. It is based on identities that, in practice, often have large and variable measurement errors (“errors and omissions”), thereby rendering precise targets problematic14;
  2. Their policy prescriptions are based on the assumption of constant or even one-for-one economic relationships – for example a stable velocity of money or a constant relationship between domestic credit and the money supply – relationships which turn out to be highly unstable and often not one-for-one;
  3. They often leave out other important channels of monetary policy besides changes in the money supply, channels such as credit and asset prices; and
  4. There are important variables that are assumed to be exogenous to monetary variables, which are, to the contrary, often affected by monetary policy, for example, the level of unemployment.

Alternatives do exist

Alternatives to this destructive approach to central bank policy do exist. Erinc Yeldan from Bilkent University in Ankara, Turkey and I directed a series of country studies undertaken by a team of researchers working on a project on alternatives to inflation targeting. The countries covered in this project are Argentina, Brazil, Mexico, India, The Philippines, South Africa, Turkey, and Vietnam. I also worked with a UNDP-sponsored study of economic policy which targets employment for South Africa.15

A range of alternatives were developed in these papers, all the way from modest changes in the inflation targeting framework to allow for more focus on exchange rates and a change in the index of inflation used, to a much broader change in the overall mandate of the central bank to a focus on employment targeting, rather than inflation targeting. Some of the alternative policies focus exclusively on changes in central bank policy, while for other countries, changes in the broad policy framework and in the interactions of monetary, financial and fiscal policy are proposed. Some incorporate explicit goals and targets, while others prefer more flexibility and somewhat less transparency. But all of the studies agreed that the responsibilities of central banks, particularly in developing countries, while including maintaining a moderate rate of inflation, must be broader than that, and should include other crucial ‘real’ variables that have a direct impact on employment, poverty and economic growth, such as the real exchange rate, employment, or investment. They also agree that in many cases, central banks must broaden their available policy tools to allow them to reach multiple goals, including, if necessary, the implementation of capital management techniques.16

The major lesson of these case studies and auxiliary materials is that there are well thought out and plausibly viable alternatives to inflation targeting that can focus more on important social, real sector outcomes such as employment generation, poverty reduction, export promotion and investment enhancement. If inflation targeting is resilient as the ‘big idea’ of modern central bank policy because many perceive that there is no alternative, these case studies can provide an antidote by showing that viable alternatives are plausible and can be further developed and put into practice. In fact, doing so would be consistent with long-standing historical practice17.

Gerald Epstein is professor of economics at the University of Massachusetts
A shorter, fully-formatted PDF version of this briefing is available.

I thank my colleagues and co-authors who have contributed to the ideas here, including Ilene Grabel, James Heintz, Jomo, K.S., Terry McKinley, Leonce Ndikumana, Robert Pollin, Erinc Yeldan and the other members of the Alternatives to Inflation Targeting Project, and Lynda Pickbourn for excellent research assistance. All errors are mine.


1. Heintz, James. 2006b. Globalization, economic policy and employment: Poverty and Gender Implications. Employment Strategy Department. Employment Strategy Papers. 2006/3. Geneva: ILO.

2. International Labour Organization, 2004a. Global Employment Trends 2004. Geneva: International Labour Organization.

3. International Monetary Fund, 2005. Monetary and Fiscal Policy Design Issues in Low-Income Countries. Policy Development and Review Department and The Fiscal Affairs Department. Washington, D.C.: IMF, August 8, 2005. p.8

4. Allen, Mark, Ulrich Baumgartner and Raghuram Rajan. 2006. Inflation Targeting and the IMF, Washington, D.C.: IMF, March 16.

5. Masson, Paul R. Miguel A. Savastano and Sunil Sharma. 1997. The Scope for Inflation Targeting in Developing Countries, Washington, DC: IMF WP/97/130.

6. Ball, L. and Sheridan, N. (2003) Does Inflation Targeting Matter? IMF Working Paper No. 03/129. Roger, Scott and Mark Stone. 2005. On Target? The International Experience with Achieving Inflation Targets, Washington, D.C.: IMF, Working Paper, WP/05/163.

7. Epstein, Gerald. 2006. Central Banks as Agents of Economic Development, WIDER Working Paper, No. 2006-54.

8. IMF, 2006

9. Epstein, Gerald and James Heintz. 2006. Monetary Policy and Financial Sector Reform For Employment Creation and Poverty Reduction in Ghana. PERI.

10. IMF, 2004a, p. 38

11. Easterly, William. 2002. An Identity Crisis? Testing IMF Financial Programming. Center for Global Development. Working Paper No. 2. August.

12. Barth, Richard and William Hemphill. 2000. Financial Programming and Policy: The Case of Turkey. Washington, D.C.: IMF Institute, IMF.

13. Blejer, Mario I., Alfredo M. Leone, Pau Rabanal and Gerd Schwartz, 2002. Inflation Targeting in the Context of IMF-Supported Adjustment Programs, IMF Staff Papers, vol. 49, no. 3, pp. 313-338.

14. Easterly, 2002

15. For the PERI/Bilkent project see Epstein and Yeldan, eds. forthcoming; and Pollin, et. al., 2006. The PERI/Bilkent Project was made possible with the generous support of UN-DESA, the Ford Foundation and Rockefeller Brothers Fund.

16. Ocampo, Jose Antonio. 2002. Capital-Account and Counter-Cyclical Prudential Regulations in Developing Countries. WIDER, Discussion Paper, August. Epstein, Gerald, Ilene Grabel and Jomo, K.S., 2005. Capital Management Techniques in Developing Countries: An Assessment of Experiences from the 1990's and Lessons for the Future. in Gerald Epstein, ed. 2005, Capital Flight and Capital Controls in Developing Countries. Northampton: Edward Elgar Press.

17. Epstein, 2006