Inflation targeting: considerations for central banks in Africa

1: Introduction

As elsewhere in the world, countries in Africa have faced a sequence of large inflationary shocks in recent years including those caused by the Covid pandemic, the subsequent war in Ukraine and the ensuing commodity price shocks. These inflationary impulses were often accompanied by a sharp fall in output growth causing income per head to stall or even decline in a number of countries on the African continent. Government debts, which were already on an upward trend, generally increased further, adding to the monetary policy challenge.^{footnote [2]} And according to the International Monetary Fund (IMF)/World Bank, over one half of low-income countries (LICs) on the continent currently have external debt either in distress or have high probability of it becoming so.

Against this challenging background some countries in Africa have been considering – sometimes following IMF prompting – the adoption of a formal inflation targeting regime either to help get the inflation rate back down, or to maintain inflation at low levels.

Formal inflation rate targeting frameworks have not previously been widely adopted in Africa, where the most common monetary policy regimes are based either on some degree of managing the exchange rate, or a monetary aggregate target. Table A shows the most recent IMF classification of African countries by monetary regime (IMF (2024b)): whether an exchange rate anchor, a monetary target, inflation target or some other arrangement. Almost one half of countries were recorded as having an exchange rate anchor while a further quarter had a monetary aggregate (shown in the columns). Furthermore, the vast majority of countries were judged to have, in practice, some form of managed exchange rate even for those without an explicit exchange rate anchor (shown in the rows). Of the 54 countries, only five were recorded as having formal inflation targeting regimes: Ghana, Kenya, Mauritius, South Africa and Uganda. Only the latter three of those were classified as also having a floating exchange rate.

This raises a number of questions about existing policy frameworks and the desirability and pressures for change. In this paper, we review how the benefits of inflation targeting arise, some of the consequent implementation and operational issues that need to be addressed to maximise those benefits, and the particular challenges for African countries.

2: Theory of inflation targeting

2.1: Models of inflation targetting

The basic macroeconomic model used to illustrate the advantages of inflation targeting (IT) is based on the expectations-augmented Phillips Curve (EAPC). A mathematical expression of this model is presented in Annex 1 and careful analysis of that model yields a number of conclusions.

The model of inflation assumes that if the inflation target regime is credible then economic agents, especially wage and price setters, have constant inflation expectations. On the other hand, if the regime is not credible then inflation expectations would vary, for example, in line with the actual inflation experience.

Nominal interest rates can be used to control inflation, but the policy rate has to be changed sufficiently to stabilise inflation. If inflation expectations remain fixed, interest rates must change to offset the inflationary shock. The policy response required is greater:

1. the larger the shock;
2. the smaller the impact of interest rates on excess demand; and
3. the smaller the impact of demand on inflation.

These sensitivities may be intuitively obvious but are usefully confirmed by the model.

If inflation expectations do not remain constant, but change in the same direction as actual inflation, then a minimum stabilising condition in the model is that the policy rate must change in the same direction and by more than the change in expected inflation. This is the canonical ‘Taylor principle’ – namely that real interest rates rise when inflation is above target (Woodford (2001), Davig and Leeper (2007)). Furthermore, the size of that change in real interest rates – nominal rates minus expected inflation – needs to be large enough to offset the shock, with the same sensitivities (1)–(3) as above, determining the scale. This rule applies whatever the inflation expectations mechanism (eg backward or forward looking).

In practice, we see many policy reactions where there is a positive shock to inflation, but interest rates are raised less quickly than actual inflation. If inflation expectations then rise by more than the policy rate, real interest rates would not increase, and the model predicts that there would be no disinflationary pressure – indeed any fall in real interest rates could end up stoking inflation. Under such conditions, the model predicts that inflation will keep increasing, and that is often observed to be the case when the interest rate reaction is too slow and/or too small. Early, proportionate responses in interest rates can be much less costly overall than a sequence of small, delayed reactions.

If the policymaker lacks credibility, then, by definition, inflation expectations will move more, and more quickly, and in this model, the policy reaction then needs to be greater and faster, to stabilise inflation.

In summary, interest rates need to rise by at least as much as the change in inflation expectations plus an extra term which is inversely related to the strength of the transmission mechanism, and sufficient to overcome the shock (Annex 1, equation 6b).

Like all models, this formulation is only a rough approximation to reality. The model’s results nevertheless capture much of what inflation targeting is meant to achieve. The model demonstrates that the main benefit of inflation targeting regimes arises if they can add credibility over time, keep inflation expectations grounded and thus reduce the extent to which interest rates need to respond to shocks. Ideally, all the design parameters of an inflation target regime should be decided to enhance credibility.

The results also highlight that it is easier and less costly to control inflation if the transmission mechanism of monetary policy is stronger. That may seem obvious, but it explains why many central banks, as well as focussing on managing credibility and influencing inflation expectations, are keen on improving the transmission mechanism where they can. Much of that work is focussed on the financial sector where the central bank can directly promote change.

Finally, the model shows how policy can be more complex and challenging when there is any element of a supply-side shock (defined as a shock which drives inflation and demand in opposite directions, such as an increase in commodity prices). In the short global history of inflation targeting up to 2020, shocks to the oil price (rather than to demand) were consistently the most challenging.^{footnote [3]} The most complex shocks – like those caused by the Covid pandemic – combine both demand and supply impacts, making it difficult to assess what is happening to each of demand, inflation and expectations.

To anticipate some of the analysis below, we also note that demand shocks can come from persistently loose fiscal policy. The challenge then is not just economic – tight monetary policy in response to a loose fiscal stance can lead to an unbalanced economy – but can also have political consequences.

2.2: Demand and supply shocks

As noted, supply shocks can be the most challenging aspect of inflation targeting and the most common form is a rise in world commodity prices.

In a successful IT regime, the conventional response to a supply-side shock is to leave policy largely unchanged and to ‘look through’ the rise in inflation with the expectation that the associated fall in output should prevent any rise in inflation becoming embedded in expectations. Although a (one-off) negative supply shock will increase the price level, that should only temporarily increase the rate of inflation. The annual inflation rate should revert to its starting point 12 months or so later providing that the initial rise in inflation does not have ‘second round’ effects through increasing other prices and wages.

In the case of a commodity price shock, if the commodity price returns to its previous level, that reversal should have an equal and opposite negative effect on the measured inflation rate and a positive effect on output. However, policy lags mean that any active intervention to smooth out this saw tooth effect, would risk adding to instability.

Volatility in inflation and output is not good for medium-term growth and stability as there may be some scarring effects eg, from a temporary reduction in investment. Correctly handled the medium-term effects should be small.

The correction to output and inflation following a supply-side shock will depend on the credibility of the policy framework. The more credible the monetary authority and its target, then the less likely that inflation expectations will become de-anchored in the face of a supply shock and the smaller the disruption. Hence a credible IT regime can help reduce the impact of such shocks. We address some ways to reinforce credibility in the rest of this article.

2.3: The role of the exchange rate

The EAPC model in Annex 1 is shown for a closed economy. Similar economic arguments can be applied to an open economy. In a floating rate regime, if export and import prices are set in a world currency then the exchange rate should change so that import prices in domestic currency are in line with domestic prices in the medium term. But exchange rate dynamics can be important as we discuss below.

In a credible fixed peg (or currency board) exchange rate regime, a country essentially outsources inflationary control to the country of the currency to which it is fixed – typically the dollar or the euro. Credibility then hinges on the ability to maintain that nominal fixed exchange rate. This has been a successful regime in many countries globally but most notably where they are large exporters of valuable commodities (eg oil-producing Gulf states) or large entrepot economies (eg Hong Kong). In such cases, countries have been able to enhance exchange rate credibility by amassing a large pool of foreign currency reserves which help convince stakeholders that the peg can and will be maintained. African countries have not made a similar accumulation of foreign currency reserves, even in those countries such as Nigeria and Angola, which have significant oil exports.^{footnote [4]}

For emerging market and developing economies (EMDEs), which have an exchange rate regime which is not seen to be reliably pegged, the exchange rate can play an important role in the transmission of monetary policy and of other domestic and external shocks affecting inflation. If one starts from a stable nominal exchange rate and it depreciates exogenously, assuming that prices are fixed in world markets, import prices would rise and so likely would overall prices in the domestic economy. Measured annual inflation rates therefore would also rise temporarily, even if it is a one-off step depreciation in the exchange rate, and a consequent step change in the price level. If the exchange rate continuously depreciates then this scenario could generate continuing higher inflation domestically. That apparent pass-through of import prices to overall domestic prices is why a depreciating exchange rate is often blamed for ‘causing’ inflation.

Foreign exchange rates reflect the relative demand and supply for the currencies concerned, by diverse agents, including importers and exporters, foreign exchange traders, cross-border investors and any market interventions by the authorities. The latter can intervene either to transact or to change market structure and access (such as exchange controls) as well as influencing the rate via monetary and fiscal policy announcements.

The actions of foreign exchange market participants are not exogenous to domestic policy actions and therefore, it is seldom useful to consider the exchange rate as the true source of an inflationary shock – one needs to look at what has caused the exchange rate to change.

The market for foreign exchange facilitates trading and price discovery, like any other market. One important difference is that, as it ties in with expected returns on financial market investments, the exchange rate is necessarily forward looking and the price today will reflect not just what has happened already, but any expectations of what will happen in future. The bilateral exchange rate can therefore ‘jump’ on news that has a significant implication for the future value of either currency.

To determine the appropriate policy response to an exchange rate depreciation (or appreciation), policymakers need to identify and understand the reason why the exchange rate has changed, ie what is the true source of the shock? Domestic or foreign? Policy changes or other economic developments? Perhaps the most frequent cause of downward pressure on the exchange rate is that domestic monetary conditions are too loose (are themselves inflationary), and the exchange rate should then depreciate in line. The obvious appropriate policy response would be to tighten monetary policy and correct the problem at source.

If excessive demand and inflation is caused by overly-expansionary fiscal policy, then again, the exchange rate is likely to depreciate. The most appropriate policy response should be fiscal tightening. Monetary tightening to offset fiscal loosening can offset the short-run consequences and temporarily stabilise both the exchange rate and inflation rate but would leave the economy with an unbalanced policy mix that is not sustainable in the medium-term. We return to this issue in Section 3.2 when discussing fiscal dominance.

In the face of a negative real terms of trade shock, such as a rise in global commodity prices impacting commodity importers or a sudden loss of export markets, then the external trade account will likely go into deficit and the exchange rate should then be allowed to depreciate to facilitate external balance. In this case monetary policy should not be tightened in response to the first-round price effects. Only if expectations threaten to become de-anchored should policy adjust.

Higher import prices, in domestic currency terms, are inevitable if world prices rise, and part of the desired adjustment is for higher prices of imports to discourage import volumes.^{footnote [5]} Of course, higher world commodity prices mean that the domestic economy is worse off. But that loss is unavoidable and cannot be offset by intervening in the nominal exchange rate.

Persistent external imbalances can only be fixed by a change in the real exchange rate, not by monetary policy. For a shock that requires a fall in the real exchange rate, the easiest way to achieve that is through a fall in the nominal exchange rate. If instead the nominal rate is held fixed, then the alternative route to adjustment is through a fall in domestic prices. That can be a prolonged and very costly mechanism. Prices and wages do not adjust downwards easily, unless induced by a significant fall in output. This is one of the main arguments for floating nominal exchange rates – it enables changes in the real exchange rate to occur more easily.

Global shocks are often in the form of step changes, which have a one-off but possibly prolonged pass-through to domestic output and prices. Persistent downward pressure on the exchange rate over time is most often due to higher domestic inflation rates compared to key trading partner countries. In that case, changes in the exchange rate would be the consequence rather than the cause of the inflation. If domestic inflationary pressures are responsible, then domestic monetary policy tightening would be appropriate and by returning inflation to the desirable level, that would also stabilise the exchange rate.

On the other hand, if it is the foreign inflation rate that is falling then the exchange rate should be allowed to depreciate unhindered since that would be consistent with maintaining domestic inflation at the previous level (exchange rate depreciation would offset the lower foreign inflation rate such that import price inflation remained stable).

Only very occasionally, the foreign exchange market might itself be a source of the shock. For example, disorderly trading conditions, ‘lumpy’ cross-border capital flows or inefficient market structures, any of which may cause some volatility and temporary pressures. An IT regime with a floating exchange rate should seek to make its foreign exchange market as liquid and efficient as possible to reduce that risk. Other options such as ‘smoothing interventions’ in the exchange rate or capital controls may be employed but are usually expensive to implement and maintain effectively.

In summary, in an IT regime the monetary authority should not seek to control the nominal exchange rate directly as a means of inflation control as that can only cause unwanted real exchange rate distortions while not addressing the underlying inflationary factors. Rather the monetary authority should try to understand the sources of pressure causing the exchange rate to change and then work on those sources. The exchange rate, real and nominal, will find its own equilibrium if and when domestic inflation is kept under control.

3: IT policy in practice

IT regimes (combined with a free-floating exchange rate) are a relatively recent development in central banking. Historically most currencies were fixed against the price of gold and then in the aftermath of the Second World War against the dollar in the Bretton Woods System (with the dollar pegged to gold). This system broke down in the early 1970s. Most major countries moved towards having more exchange rate flexibility. But there was a growing recognition that following the high and volatile inflation in the 1970s and early 1980s there was a need for some form of domestic monetary anchor. Targets for the growth in monetary aggregates were tried but were found to be unreliable indicators of future inflation.

There was also a growing consensus that the job of controlling inflation should be the responsibility of the central bank rather than the government and, in fact, should be the central bank’s main macroeconomic responsibility. Keeping inflation low and stable was increasingly viewed as the key way in which a central bank could support long-term economic growth.

The first central bank to introduce an IT regime was the Reserve Bank of New Zealand in 1990 since when they have been adopted widely first in developed countries and subsequently in a number of EMDEs. In general, IT regimes have been relatively successful in keeping inflation low and close to target. Inflation did rise substantially above target in many major countries in 2021–22. However, because inflation expectations remained largely contained, many central banks were able to reduce inflation back towards target relatively smoothly through raising their policy rates, without causing recession.

3.1: Challenges of implementing an IT framework

There are though, in practice, a number of challenges to implementing an IT regime. The model described in Annex 1 is useful for understanding the forces at work, but not so useful as to guide policy in real time. Many key variables cannot be directly observed or measured, including potential output (ϒ*), the derived output gap, the equilibrium real interest rate (𝒓*) and even inflation expectations. Other key data, such as price inflation and GDP growth are only available after a time lag and sometimes subject to revision. The quantitative relationships between the key variables, in particular the impact of monetary policy on the economy (the transmission mechanism), are also uncertain.

It can also be difficult in practice to distinguish between a supply and demand shock, especially in real time, when many shocks have some element of both. For example, before the Covid epidemic it was difficult to know the extent to which it would reduce domestic supply as well as demand and hence the gap between the two was highly uncertain and changing over time. Many countries ran higher domestic fiscal deficits during the pandemic, supported by a loose monetary stance. Towards the end of the pandemic, it seems with hindsight that demand bounced back before supply constraints had all been alleviated and that may have started the global inflationary surge.

More recently, it has been a challenge for EMDE policymakers to understand the extent that the depreciation of many of their currencies against the dollar since Spring 2022 has been caused by domestic or global factors. Globally, there was a relative price shock from higher commodity and food prices following the war in Ukraine. This directly represented a negative terms of trade shock for any net importer of those commodities and a potential indirect capital account shock for many, from the rise in US policy rates as the Federal Reserve acted to bring US inflation back to target.

The commodity shocks, if persistent, would suggest the real and hence nominal exchange rate should adjust downwards for many commodity importing countries. And the US interest rate policy changes might need to be at least partially replicated if higher domestic interest rates are needed to limit capital outflows. Rising domestic sovereign debt would suggest the need to tighten fiscal policy.

Whatever the shocks, if the consequence is that domestic inflation rises and inflationary expectations rise significantly, then clearly domestic policy interest rates need to be increased – at least enough to offset that rise in inflation expectations.^{footnote [6]}

Domestic shocks, especially from monetary and fiscal policy, may be the hardest to identify and respond to. Initially a rise in demand may be welcomed, rather than correctly seen as unsustainable. And it may be difficult politically to identify and then admit that the government or central bank is the source of a destabilising shock.

Multiple large shocks in recent years have not been the normal experience of IT regimes over the previous three decades. The period since the onset of the pandemic has arguably been the biggest challenge to IT regimes to date, greater even than the Global Financial Crisis of 2007–09. Inflation increased so much in developed countries in 2021/22 partly because the inflationary pressure was not anticipated. The extent of the rise has increased short-term inflation expectations above central bank targets in both advanced countries and EMDEs.^{footnote [7]} This appears to have made it more costly to get inflation back under control. In our view and with the benefit of hindsight, had monetary policy been tightened sooner that could have limited both the subsequent rise in inflation, and the degree of interest rate tightening. Once inflation was rising, the direction of the policy response was clear, even if the scale was not because of the mixture of supply and demand-side factors. The experience of developed countries at least, is consistent with the results from our EAPC model.

Several simple approaches can be used to help identify future inflationary pressures and to separate out demand and supply shocks. One example commonly followed by developed economies, is to calculate and monitor a ‘core’ inflation rate which excludes the prices of primary products such as energy and food. These components of inflation are both more volatile – under any conditions – and most likely to reflect global supply-side shocks. Although targeting core inflation explicitly is not advised, as it could lead to a loss of credibility with the public, it can nonetheless help to calibrate the extent of supply-side shocks. US Federal Open Market Committee (FOMC) members regularly refer to core inflation measures in their commentary (Powell (2024)). A rise in core inflation may be more worrying as an indicator of inflationary expectations becoming embodied. Wage rises based on ‘cost of living’ claims can also indicate rising inflation expectations. In some countries, labour strikes become more frequent at higher inflation rates.

3.2: Policy challenges in developing countries

Policy trade-offs are likely to be more challenging in developing countries, such as most African states, than in developed countries. First, and probably foremost, there is a risk of fiscal dominance – the subordination of the central bank’s objectives to that of the government – in the setting of monetary policy. Two particular challenges bring political issues to the fore: the short-run correlation between output and inflation and the impact of fiscal deficits on inflation.

The first issue stems from the fact that changes in policy interest rates (and money growth) usually affect output (albeit temporarily) before inflation. That timing differential matters. It could, for example, be exploited by politicians ahead of election years, to reduce interest rates to stimulate growth in the short run, with the cost in inflation and reversal coming later. If done repeatedly, this can result in a bias towards ever higher inflation over the medium term. That explains why many central banks have been designated as an independent monetary authority with delegated powers to set interest rates so to meet an inflation target in the medium term. The degree of central bank independence is often thought to be a key part of establishing credibility in inflation control.

The second issue is that lots of countries currently have high government debts and deficits. Developed countries, with a broad tax base and with more advanced financial markets, generally find it easier to finance a large debt stock. Many developing countries have a relatively narrow tax base and hence low tax revenues (relative to GDP) combined with huge pressures to spend to alleviate poverty and poor health. Many also have sizeable interest payments on past accumulated debts, especially if issued in foreign currency to attract international investors. High fiscal deficits may directly add to pressure on the central bank to limit or slow the raising of policy rates since that would increase the cost of government financing and, through the transmission mechanism, may in the short run reduce output growth and employment to a greater degree than if fiscal deficits were smaller.

Historically, over hundreds – in fact thousands – of years, governments (and other rulers) worldwide have very often wanted to spend more money than was available to them. History records many examples of punitive taxes and unpaid sovereign debts.^{footnote [8]} In the past a common reason was to finance wars, sometimes for other forms of personal indulgence, but perhaps more recently has included actions taken for perceived benefit of the population – but even the latter do not always grow the overall size of the economy and may have squeezed out more productive investment.

Sovereign defaults can be traumatic but historically they did not lead to sustained increases in prices. Inflation is a product of the modern world in which there is a greater ability to create fiat money without backing, ie money that is not supported by a corresponding increase in real national wealth or production.

Governments in developing countries may have the best of intentions to use government spending to improve the welfare of the population, but fiat currency regimes provide little natural constraint on the ability of governments to spend unproductively in excess of available resources.

Without some legal constraints being built into the monetary regime, the central bank may come under pressure to directly lend to the government and/or buy government securities in the primary market at subsidised interest rates. Such direct financing of the government by the central bank should be avoided, if possible, especially if it results in a permanent increase in the non-interest-bearing liabilities of the central bank.^{footnote [9]} ‘Printing money’ has invariably been a cause or a significant contributor to high inflation.^{footnote [10]}

In countries where the ability to collect taxes is low, and sovereign debt markets are small, an argument could be made that it is legitimate for the government to claim resources through an inflation tax. But this is seldom equitable or sustainable and it is not advocated here.

The risk of ‘fiscal dominance’ can be addressed by a strong governance framework to ensure that the central bank (and its key officials) have the required degree of operational independence. To ensure a successful IT regime, the central bank should have independence both in law and in practice, with limits on its direct financing of the government and a clear mandate.^{footnote [11]} Monetary policy decisions should not be subordinated to party politics.

The inflation target itself can be set by the government, the central bank or jointly. But once the target has been set it should be seldom changed (to preserve credibility), and the central bank (or its Monetary Policy Committee (MPC)) needs to have operational independence to change the policy rate and other parameters of the monetary framework as it sees fit to meet the inflation target. Such a framework should be designed to give interest rate and liquidity control making it less costly for the central bank to control inflation.

Regardless of deficit financing, a necessary condition for a successful inflation targeting framework over the medium term is that the government’s fiscal position should be sustainable. If a country has a high and rising government debt then, to control inflation, monetary policy is likely to be set tighter than it otherwise would have been. That internal balance may mean that inflation stays under control for a long period. But loose fiscal and tight monetary policy requires high real interest rates that will likely lead to instability through external imbalances.

One mechanism for external imbalances to exert pressure is that the higher nominal interest rate needed to keep inflation down also attracts short-term capital inflows which can lead to an appreciation of the nominal and real exchange rate. These capital inflows are matched by a rising current account deficit. Such trends cannot be sustained indefinitely. Eventually as the stock of net external assets deteriorates, the capital inflows will lessen and financial crisis is likely to ensue, involving a sudden and large exchange rate depreciation.

The sources of pressures on fiscal expenditure may be genuinely pressing for welfare reasons. Commodities such as energy and food tend to make up a much larger part of individuals’ consumption baskets in developing countries than in high-income ones. Therefore, a rise in global food or energy prices can have a very large negative impact on real incomes especially for the poorest in society. In turn, this may encourage governments to provide subsidies that increase deficits further.

The large weight of food and other commodities in the consumption basket – often over 40% in countries in sub-Saharan Africa^{footnote [12]} compared with under 10% in the UK – means that inflation in (open) developing countries is particularly sensitive to external shocks. Exchange rate depreciation also tends to be reflected more into higher general consumer prices in developing than in high-income countries.^{footnote [13]} This higher exchange rate pass through partly reflects the fact that EMDEs usually have a higher dependency on imports and a greater share of these imports have prices determined in world markets. Hence, they are invoiced in foreign currency, usually in dollars.^{footnote [14]} For example, in sub-Saharan Africa two thirds of imports are at prices fixed in dollars for the average (median) country.^{footnote [15]} In these cases, exchange rate depreciation will automatically result in higher import prices measured in the domestic currency.

The exchange rate pass-through is also found to be disproportionally bigger for very large depreciations (20% or more).^{footnote [16]} Such a large rise in prices is thus more prone to dislodge inflationary expectations, making it more likely that EMDEs have to tighten policy in the face of such a shock. Most EMDEs are also likely to have less credibility built up from controlling inflation in the past.

All that said, the pass through in EMDEs is lower now than in the 1980s–90s and that seems to be associated with better policy frameworks – including the shift to inflation targeting in some cases, more exchange rate flexibility and increased central bank independence.^{footnote [17]} In other words, shifting to a credible inflation targeting framework may help reduce the policy trade-offs often associated with external price shocks. But no monetary framework can offset the fact that higher commodity prices make the economy worse off – unless the country itself is a producer of commodities in enough scale.

A further challenge of currency depreciation in EMDEs is that they tend to have a much higher share of their net external debt denominated in foreign currency than in richer countries, in part because of smaller domestic financial markets, but also because of the perceived added risk of default via inflation and depreciation. This foreign debt burden, and the cost of funding it, increases in domestic currency terms following nominal exchange rate depreciation. Fiscal consolidation can help contain the financial risks from currency depreciation. These challenges of exchange rate depreciation in developing countries has resulted in a ‘fear of floating’.^{footnote [18]} This has meant that the share of countries currently in sub-Saharan Africa with a floating exchange rate in practice (ex post) is much lower than stated in their official policy positions (de jure).^{footnote [19]}

In exceptional circumstances, if there is a substantial threat to financial stability or to the central bank’s ability to maintain price stability, there may be a role for intervening on a temporary basis to prevent currency depreciation.^{footnote [20]} But that generally requires a corresponding policy tightening that would justify the higher exchange rate over time. Currency intervention usually only works if it is supportive of, and supported by, a much wider package of policy measures.

4: Inflation and inflation targetters

4.1: Recent inflation

The IMF estimate that global annual inflation was 4.9% at end 2024 with it expected to drop to 4.0% by the end of 2025.^{footnote [21]} Richer countries tend to have inflation (and target) rates lower than this average, most frequently in the 2%–4% range. Inflation in poorer countries, including in Africa, tends to be higher (aqua parts of the bars in Chart 1). The IMF’s end-2025 inflation projection for sub-Saharan Africa is 12.6%; compared with a forecast for all emerging and developing countries of just 5.1%.

Nonetheless, in most African countries inflation has fallen from the post-pandemic peaks (Chart 2 – most dots are below the 45%-degree line) as the impacts of the pandemic have waned and policy has tightened. Also, inflation tends to be lower for countries with hard fixed exchange rates (aqua and orange dots in Chart 2).

The current differences in inflation rates among African countries with floating or managed exchanged rates seems to be reflected in the extent of currency depreciation against the strengthening dollar over the past two years (Chart 3). As discussed earlier, the causality is likely to run from inflation to the exchange rate rather than the reverse. The US took rapid action to contain its own inflation rate rise, effectively guiding it back towards target. Countries that did not react as rapidly or effectively – in both the developed and developing world – would see both their domestic inflationary pressures continuing to rise plus apparent pressures arising from their depreciation against the dollar.

4.2: Performance of inflation targets

In high-income countries, annual inflation targets have typically been set around 2%–2.5% (Chart 4). The reasons why ‘price stability’ has been set above zero in practice include:

(a) The official consumer price indices that are targeted are likely to overestimate true price inflation because they will not fully capture quality improvements, nor can they fully reflect switching from more expensive to cheaper goods and services as relative prices change.

(b) In order to boost demand when inflation is low, the central bank may want to lower its nominal policy rate to deliver a negative real interest rate. Given that the nominal policy rate has an effective zero lower bound, the central bank will not be able to generate a negative real rate if inflation (expectations) is below zero.

(c) Many nominal prices and wages are sticky downwards. In order to avoid potentially costly quantity adjustments in the economy (eg, declines in employment), average inflation above zero allows for some reduction in real wages to occur without cuts in nominal wages.

(d) Manufacturing price inflation will generally be much lower than service sector inflation reflecting sectoral productivity growth rates. At a small positive inflation rate, manufacturing prices on average will have an inflation rate close to zero.

(e) It is desirable to have inflation that has low volatility otherwise it creates uncertainty which can have a negative impact on the economy. Ultra-low or negative inflation can be more volatile than when it is low and positive. This in part reflects the price stickiness point (c) – some prices may be more difficult to adjust downwards than others (Lane (2020)).

For the first two decades of the 21^st century, central banks – mostly but not all from developed countries – were very successful in hitting their respective inflation targets. But as the Covid pandemic started to ease in 2021/22, inflation rates in many countries increased well above target, including in the United States, the euro area and the United Kingdom. Most of those countries which successfully brought inflation back close to target only achieved that during 2024.

Inflation targets can be set for various monetary regimes, not just for those with a formal IT framework. Sometimes these targets are forecasts or governmental ambitions rather than being actively pursued as a target by the central bank. Looking worldwide, we have identified 69 countries where there is evidence of some kind of formal numerical inflation target.^{footnote [22]} Whereas inflation was above target in most of these countries two years ago it is now fairly evenly dispersed around the respective central targets (Chart 5).

Of the 69 countries worldwide in our data set, we observe some countries (22) tending to focus on just on a point target (eg 2%), others (16) emphasising a range (such as 3%–5%) and a majority (31) with both a point and a range (eg 4% +/- 1%).^{footnote [23]} These differences in target specification are not always clear and consistent within a jurisdiction, especially where there is not a formal IT framework.

Countries adopting inflation targets should think carefully about the precise formulation and be consistent in its use to maximise the benefits from credibility. In particular, a range should not be used as a ‘get-out’ clause and the central bank should always be clear that they are targeting the centre of the range.

The point target, or centre of the range, tends to be higher for EMDEs such as many African countries, than for high-income countries. And countries with higher inflation targets tend to have wider ranges and more variable inflation as well as outturns less close to targets (Chart 6).^{footnote [24]} EMDEs usually have some form of target range reflecting their greater inflation volatility which likely arises from their susceptibility to the impact of supply shocks. A target range (or a point with a range) can, in any case, be a useful way of conveying the uncertainty of shocks to inflation and of policy responses especially for countries introducing or at an early stage of using an inflation target.

As shown earlier in Table A, the majority of African countries, whether or not they have a numerical inflation target, informal or formal, do not have formal inflation target regimes. Almost half the countries on the continent instead are said by the IMF to have hard fixed nominal exchange rates mostly against the euro with a few against the dollar or a hard currency mix.^{footnote [25]} Some countries in southern Africa (Eswatini, Lesotho and Namibia) have their currencies fixed against the South African rand in the Common Monetary Area. Around one third of countries have some form of managed floating or adjustable exchange rate with the monetary anchor usually provided by the growth in domestic narrow or broad money. That said, a small but increasing number of countries in Africa have adopted a formal inflation target coupled with a (more) flexible exchange rate. Most of these targets are set over the medium term although in South Africa it is set continuously. Some other countries in Africa have adopted an informal inflation target.

Although inflation in Africa tends to be higher than in other regions, there does not seem to be any systematic tendency for inflation to be above target currently among the inflation targetters. Many African countries with formal or informal targets in the 3.5%–5.0% range have reduced inflation back to target post-pandemic, eg South Africa, Kenya (Chart 6 orange dots). In some other countries inflation is currently a long way above formal (eg Ghana) or informal (eg Egypt) targets. The overall evidence suggests that inflation targeting can be made to work well in Africa, if policy is genuinely geared to meeting the target. But if monetary policy is not appropriately set, or is overridden by fiscal dominance, then the mere existence of a target will not help much.

4.3: The choice of an inflation target for an African country

A number of factors should affect the choice of a formal, numerical inflation target. It could depend on global inflationary pressures, and in particular the inflation rate in the country’s main trading partners, the structure of the economy, the level of development, inflation history and local political economy considerations. As discussed above, less developed countries tend to have higher targets with wider ranges which to some extent reflects faster GDP growth and higher domestic (non-tradable) sector inflation.^{footnote [26]} And higher levels of inflation tend to be more variable. Overriding all these issues though, is the need to be credible so as to pin down inflation expectations – otherwise the benefits of IT will not be realised.

As shown in Annex 2, several emerging and developing countries, including in Africa, currently have point targets (or mid-points of ranges) between 4% and 5% which seems appropriate for developing countries that have a floating or managed exchange rate. It might be less credible to have a target lower than this if initial inflation is much higher. And it would be increasingly difficult to control inflation – and hence the target would be less credible – if the target were much above 5%.

To the extent that countries in Africa with a fixed nominal exchange rate against a hard currency also want to have an informal inflation target it would need to be lower since they will need to keep inflation close in line with the country to which the currency is pegged (adjusted for any trend differences in productivity growth). This is the case of countries in the WAEMU whose currency (the West African CFA franc) is fixed against the euro and have an informal inflation target of 2.0%.

For those high-inflation countries which would like to transition to an inflation target there is a question of whether to have a target path, interim targets or just announce a long-run target. Credibility is strengthened by choosing one long-run anchor that is right for the country and then sticking to it, whatever the shocks. Repeatedly changing the target is unhelpful, especially if the changes are seen to be linked to the path of actual inflation. A large initial gap to the ultimate inflation target need not damage credibility: the direction of policy, and the need to reduce inflation, should be very clear and policy will need to match. Once such a target is set, the central bank must adopt a credible policy designed to achieve it.

On balance, it might be best for credibility if the central bank communicates that it plans to hit an inflation target (range) of x by a certain date in the future rather than having a number of intermediate inflation targets during the transition.

The transition period to the steady state inflation target should be specifically considered by the central bank. If too long, then wage and price-setters will likely not believe that the central bank is committed to reducing inflation to the stated target. On the other hand, if the period is too short, the public may think that the central bank will be unwilling or unable to tighten policy sufficiently because of the potentially larger short-term output cost of hitting the target. Since a tightening of policy is thought to have its maximum impact on inflation at around 2–3 years, one might choose a transition period of, say, between 3 and 5 years so that a couple of years of tighter policy should see the emergence of a clear path to lower inflation.

Ideally the long-run target should be consistent with a broadly stable nominal exchange rate so that it is clear that domestic policy is responsible for controlling inflation. However, establishing the rate of inflation that is consistent with exchange rate stability is not straightforward. For example, there are many bilateral exchange rates, and they will not all be stable at the same time. A starting point is to observe that commodity prices and many manufactured goods are traded in world prices, and with developed countries targeting 2%–2.5% consumer price inflation, then the international inflation rate for these goods will be close to zero.

The domestic inflation target depends on what is happening to the relative prices of domestic, non-traded goods and especially services. As an EMDE develops, its GDP growth and productivity rates should be faster than for developed countries. Typically, internal services prices rise over time at a faster rate than in developed countries: consumers in richer countries with higher productivity levels in manufacturing etc, or historically wealthy, are often prepared to pay relatively more for personal services.

In summary, an EMDE may naturally have a slightly higher target than a developed country owing to different potential rates of growth. Another consideration is that some EMDEs may have specific economic and social relationships with nearby countries with whom they may prefer to have a relatively stable exchange rate. But credibility remains key and a formal target of much above 5% risks losing the benefits from an IT regime. And if a higher target is chosen, there will likely be some point at which it would need to be tightened. Recognising that possibility of reducing the target over time, if credible, could help induce a downward bias in actual inflation.

5: Operating an inflation target

As part of an IT regime, frameworks are needed to assess data, make inflation forecasts, recognise and discuss the shocks and the risks; to make and then explain the policy decision; and to carry out an ex-post evaluation of the policy impact. The main consideration behind all these operational choices should be to adopt arrangements that maximise credibility (in the local context). Sometimes, credibility will have to be prioritised over what is comfortable or convenient. Historically, central banks have not always been very transparent or accountable – but for monetary policy, demonstrating these characteristics can considerably enhance credibility.

5.1: Central bank governance

The inflation targeting decision makers at the central bank often meet as a formal MPC. On a regular basis, the MPC will assess current monetary conditions and future inflation prospects and then set the policy rate (or other monetary policy instruments). The committee will usually include the central bank Governor as Chair, Deputy Governors, the director of monetary policy/chief economist and the senior executive responsible for market operations (if not a Deputy Governor).

Such committee structures are often set out in statute or other formal provisions. In some regimes, the Governor is deemed to be the sole legal decision maker. But the Governor can still set up their own committee, with voting if desired, as an advisory function. The reason for preferring the formal committee structure is to build credibility by making it clear that decisions are not made on a subjective basis by any one individual (however wise). Having a committee structure with one member, one vote can both increase individual accountability and reduce the risk of group think. In practice, such arrangements have not been seen to damage the authority of the Governor. Governors are seldom outvoted and can always arrange not to be, by confirming the majority view. Governors also tend to lead press conferences to convey and explain decisions and are the ones most often held accountable by the legislature.

One issue to consider when creating or developing an MPC is whether to include external (non central bank staff) members. Some policy committees in high-income country central banks have external members (eg Bank of England and Bank of Japan) others do not (eg US Federal Reserve, European Central Bank (ECB)).^{footnote [27]} In Africa, some countries with a formal inflation target have external MPC members (eg Ghana and Kenya) and some do not (eg South Africa and Uganda).

Having external members changes the dynamic. They can challenge the hierarchy and reduce the risk of group think. Their presence may also enable internal members of the MPC to more freely express their personal view. That can help improve transparency and the perception of openness.

Having members with a variety of backgrounds and experience can increase diversity of thought. At the Bank of England there has been a mix of appointments, usually of economists but from business, finance and academic backgrounds. While the different groups have never formed voting blocs, they each have different experiences to draw on. In order to maximise the pool of potential candidates some central banks – including the Bank of England – have occasionally appointed non-citizens as external MPC members.

Although there are many benefits, there can also be challenges from including external members. They need to be seen as enhancing credibility, so require some demonstrable status, expertise and experience. Ideally, there should be a transparent and open recruitment process. In particular, they need to be seen as truly independent, especially of government, and they should not have a conflict of interest with any outside roles (eg a position or other interest in a financial firm). The length of their appointment needs to be carefully considered. If too short, then there will be rapid turnover, and it might not take long before the pool of experienced and credible potential candidates is diminished.

Unlike internal members, external members will not (and should not) have internal management responsibilities that may compromise their independence. If they are full-time, this raises the question of what other activities they should be permitted to undertake. External activities need to consider potential conflicts of interest. As non-executives there is also a question of what access should they have to their own staff/budgetary resources to pursue independent research and to otherwise confidential internal information, eg on market operations or supervisory information.

A particular concern can arise if an external member is not accepting of any of the internal models and processes or works outside the accepted economic paradigm. This can lead to a situation in which they are not acting effectively as a committee member but as a ‘lone wolf’. The issue is whether this is seen to enhance or detract from credibility. Open debate and discussion of technicalities or finely balanced judgements can be a positive if it is seen to be rational and well argued. But in extreme cases – such as where one person rejects the whole policy framework or approach – could be damaging, as could any debate that is seen to be about personalities.

The MPC does not usually include a representative of the Ministry of Finance as a voting member because of the need to be independent from party politics and to avoid fiscal pressures. In the United Kingdom, a representative from the Treasury attends the MPC meetings to aid communication flows (such as providing private Finance Ministry information), but s/he is not allowed to vote and may not express a government view on any decisions.

5.2: Transparency and accountability

This article has argued that inflation targets need to be credible in order to ground inflation expectations in the economy. As discussed above, a requirement for this is that the central bank is (and is seen to be) independent from government in making its monetary policy decisions. It can also be assisted by the policymakers being individually and collectively transparent and accountable. That will increase their incentive to achieve the inflation target and, therefore, the public’s confidence that the inflation target will be hit.

Effective communications can have a great deal of influence and so should be thought of as part of the monetary policy toolbox. Communications from monetary decision makers should continually restate the objective of policy (to avoid misunderstandings) and give an assessment of the current state of the economy and a view on the outlook. Ideally, the policy setter should convey their ‘reaction function’ – how policy would react in different circumstances.

Communications can take various forms including – but not limited to – regular press conferences by the Governor and senior central bank staff, speeches, timely publishing of minutes of MPC meetings with voting record, meetings with businesses around the country, regular published monetary assessments/inflation reports including inflation forecasts.^{footnote [28]} Written reports can also function as the definitive reference point for all central bank staff involved in disseminating the message.

There are broadly five audiences for monetary policy communication – financial market participants, people running businesses in the real economy, journalists/media, politicians, and the general public. Given they have different levels of financial sophistication and different interests, they require different forms (‘layers’) of communication.^{footnote [29]}

Of the five groups, financial sector participants will want the most detail (eg minutes of the MPC meetings, regular reports (including the inflation forecast)) and some will scrutinise every word of every report. Aside from understanding the reasons for the policy decision, the forecast will help market participants better understand the authorities’ reaction function and hence embody the expected policy path in asset prices. In general, that is likely to be helpful in making policy more effective. For example, if interest rates are expected to be lowered in future, then the term structure of rates will fall immediately that becomes the assessment, and the lower yield curve will bring forward some of the policy impact.

For real economy businesses, the communications need perhaps to be less frequent and detailed, but high-level messages will be important for managers and owners making decisions on output, investment, employment and especially pay and pricing. Wage and price setting are arguably where credibility counts most.

Journalists and politicians may have similar, high-level interests – both groups will be interested in the political consequences of monetary policy decisions. Monetary authorities must be careful in communicating with these groups. Central bankers must not have or show any political biases, yet they must be aware of the politics so as not to be surprised by negative reactions and need to mitigate risks to credibility by expressing themselves clearly and appropriately. Not many journalists or politicians are monetary policy experts, more likely they will have such experts among their advisors or contacts. Communications need to be simple enough for generalists to understand, backed by suitable educative materials. It is quite normal for central banks to offer training or ‘information gathering’ sessions to both media and politicians.

Popular media are likely to be the main channels for communication with the general public. Short simple messaging with supporting graphics/visualisation may be most effective since television, radio and social media all work with sound bites or equivalent. Especially when communicating to the general public, there is a need to be personable and demonstrate empathy but to avoid becoming a ‘celebrity’ which would distract from the messaging. Central banks must always show integrity. It is probably best to avoid overly high-frequency messaging to the public and communicate only when there is something substantial to be said.

Transparency leads into the more problematic area of accountability. Since the responsibility of setting interest rates gives (unelected) central bank policymakers substantial influence over the economy and the lives of its citizens, they should be accountable for their actions. Arguably it is the government which ultimately must hold the central bank to account, through its appointments process and via legislating the applicable central bank law.

At the same time the central bank should be (seen to be) free of party-political interference. Public transparency is one way to meet both objectives.

Public accountability can take different forms. If the inflation target is not met, then the central bank Governor needs to explain why. This can be in the form of a press conference or a speech or a letter to government. In the United Kingdom the Governor must also write an open letter to the Minister of Finance every third month of consecutive miss by more than one percentage point (in either direction), explaining why the target has been missed and what is being done to bring inflation back to target and by when.

In many countries, MPC members have to give evidence and answer questions to parliamentary or similar political committees. In a broader sense, MPC members are accountable to the general public. So, it is useful if they give regular public speeches, ideally across the country, setting out their views on the economy and the prospects for inflation.

A central bank needs to be aware that conflicting messaging from multiple MPC members can be confusing. While a debate about current or future policy can be beneficial, unresolved arguments that spill over into the public arena might not be good for credibility if they are not constructive in tone.

5.3: Which measure of inflation should be targeted?

Most, but not all, central banks that target inflation use the ‘headline’ CPI.^{footnote [30]} This is usually a high-profile statistic which is accessible to the public. It is important that it is measured independently by a statistics agency and not manipulable by government. Most individuals face personal inflation rates that differ somewhat from the average, and it is natural for some to suspect that inflation is being deliberately under recorded.

A central bank with a consumer price inflation target should be able to discuss its measurement with the statistical agency provider, including to investigate and raise any concerns about its potential weaknesses. Staff in the monetary policy department should have, or acquire, detailed knowledge on how the index is constructed in order to be able to interpret it.

Producing a short-term forecast – monthly for six months ahead say, broken down by components – is a good way to ensure that some staff focus on a detailed construction of the CPI. More broadly, as part of gaining credibility, central bank staff need to be regarded as the best experts on all aspects of inflation including how it is measured as well as its economic determinants.

An important aspect in analysing the data is looking for measures that help assess whether a rise in the headline consumer price inflation is likely to be temporary or permanent, for example, after a supply-side induced rise in commodity prices (eg food prices following a drought) – that is to say evidence (or not) of the impact of second round effects on inflation.

These measures will include non-commodity (‘core’) inflation, services sector inflation, wage inflation and measures of price inflation expectations such as from surveys (eg of households, companies, financial markets). Money and credit aggregates may also provide information on the likely path of future inflation. Although these indicators have often been unreliable in the past, some recent studies have found that the marked build up in broad money growth was a precursor to the recent big increase of inflation worldwide.^{footnote [31]}

5.4: Inflation forecasting

Changes in monetary policy only affect the economy after a time lag. Waiting to react to actual price inflation usually means too little reaction, too late. So, policymakers need a forecast^{footnote [32]} of the future as a basis for choosing the policy today that will most likely, based on current information, help meet the target for inflation over the target period (usually 2–3 years).

Forecasts of inflation play several roles. They should:

1. cover all known economic data and theory and combine all available information into a simple set of statistics that can deliver a clear message;
2. help identify the shocks affecting the economy by tracking forecasts against the outturn;
3. help convey the degree of uncertainty over the inflation outlook – the forecast can be seen and used as much more than just a single central projection;
4. help explain to the various external stakeholders the reason for policy actions; and
5. help in ex-post evaluation of policies.

The best way to think of a good inflation forecast is that it summarises everything known today about inflationary pressure. Tomorrow, something will change – news about the future is always accruing – so the central projection of a forecast will always be different from the outturn to some degree and considerable weight should be attached to the uncertainty, either through confidence intervals or scenarios.

In order to make the best possible inflation forecast it is necessary to have the best available theory, data, empirically estimated models, and judgement. These are big challenges for all central banks but likely especially so in developing countries. Key data on the economy are often limited and not available in a timely fashion. The (domestic) interest rate and credit transmission channels of policy are likely weaker and more uncertain than in more developed economies given the more limited role played by banks and financial markets in the economy and credit often being directed to the government rather than the private sector. These channels will likely strengthen with financial sector deepening, which may also be a central bank objective. There may also be limited technical capacity at the central bank to produce empirical models.

For the most part, external stakeholders do not understand the nature of forecasting. They overly focus on a single central projection and tend to think of it as crystal ball gazing. Hence a popular conception that the only good forecast is one that matches the actual outturn. But even the best possible forecast will not be able to anticipate new shocks and, paradoxically, the central projection should always be wrong because of unforeseen events.

This is a difficult challenge because inflation forecasts also have a role in influencing inflation expectations. If the forecasts have credibility, then that makes them more powerful an influence. If they are perceived to be terrible, then that might damage credibility.

The only way to balance this challenge, assuming that one cannot be right/lucky all the time, is to be completely transparent about the models, the issues, the ex-post evaluations etc, so as to establish an expectation around forecast (in)accuracy. Furthermore, it is worth remembering that policy should not automatically be based on a single central forecast. No forecast is statistically precise enough for that degree of dependence. The forecast is just one input into a policy decision. But if policy judgement differs markedly from that implied by a published forecast, then one needs to understand and be able to explain why.

There are a variety of macroeconomic models that are used to help forecast inflation. Semi-structural models tend to be the most popular. These can be quite large and will cover the key sectors of the economy, often in some detail. Some central banks put more weight on dynamic stochastic general equilibrium (DSGE) models (eg Bank of England, Norges Bank and Bank of New Zealand).^{footnote [33]} These models are more theoretical with microeconomic foundations and tend to be smaller.

There is likely a trade-off when choosing a macro model, between detail/reality and simplicity/ease of use. Bigger more detailed models are more likely to capture more aspects of the real economy in practice. But they may be more difficult to understand and to change. Large models are only as reliable as their weakest parts. All models, but especially larger ones, suffer from mis or over-specification and other varieties of model risk. Generally, it is harder to demonstrate good documentation, data sources and governance around a very large model. Smaller models may be easier to use but less detailed. DSGE models are both small and complicated to use in forecasting (because the necessary theoretical elements do not often fit past data well) – but are intended to allow for a better discussion of the economics.

Subject to staff resources, it is beneficial to have a number of different models. Semi-structural and DSGE models may be supplemented by sector specific models and purely statistical non-economic-theory-based models (such as vector autoregression). The latter can be a useful cross check on the forecasts derived from the macro models.^{footnote [34]} However, the benefits from a greater number of different models will need to be weighed against the investment in controls on model risk and governance/ownership. Bernanke (2024) sets out some desired characteristics for any forecasting apparatus (eg Recommendation 4).

Human judgement also plays a crucial role in the forecast. This might be because of known deficiencies of the economic models or because of important factors affecting the economy not captured in the models, including qualitative ones. For example, it was very hard to reflect the impacts of the Covid pandemic in any economic model, as none of the data sets covered any similar experience.

Relatedly, one needs to decide whether the forecast is to be produced solely by staff (eg as at the ECB and Federal Reserve), or jointly with the MPC (as at the Bank of England). If the MPC are involved in the forecast decision making, then they will become the owners of the forecast and, by implication, at least joint owners of the models that produced it. If they are not involved, then the forecast and models are owned by the staff.

Involving the policy decision makers enables them to closely review all the factors that are likely to influence future inflation and make their assessment in a coherent framework they have at least endorsed, thus potentially improving the decision-making process. But some MPC members may not be sufficiently confident in the forecasting processes or otherwise may wish to distance themselves from any precise set of numbers.

Any published forecast should not be inconsistent with policy decisions, to preserve credibility. A purely staff-driven forecast may not be consistent with policy decisions hence such forecasts are seldom published.

Macro-model generated forecasts are usually based on certain conditioning assumptions. These may include for fiscal policy, the exchange rate, commodity prices and the central bank’s policy rate itself. These, in turn, could be set exogenously, eg from financial markets for asset and commodity prices or from the government’s stated plans for fiscal policy. Although choosing conditioning assumptions this way is transparent and can be justified, the downside is that they may not actually reflect the monetary policymakers’ views of future developments of these variables, and they may be endogenous to the out-turn.

One of the most difficult variables to model is the exchange rate which, if floating, will react to any policy changes. But forecasting the exchange rate at all accurately is largely thought to be impossible. The best that one can aim for realistically is to have an exchange rate path which is economically consistent with the rest of the forecast.

A key conditioning assumption is for the policy rate.^{footnote [35]} There are a number of different approaches that can be used to decide on the path for policy interest rates in the forecast –optimal paths using classical control techniques, simple model approaches (eg a Taylor Rule as described in Annex 1) or using a reference assumption such as from surveys of financial firms or, if financial markets are sufficiently deep, the market yield curve.

A useful assumption to make, either in the central forecast or in an alternative scenario (below), is that the (nominal) policy rate remains unchanged at its current level. The forecast will then show what is expected to happen to inflation if the authorities make no changes in policy. If the forecast for inflation in the future is above (below) target it will highlight the need for an increase (reduction) in policy rates.^{footnote [36]} In some economic models, however, the resulting projection would not be dynamically stable and some models (eg those with forward expectations) may not even solve for a fixed interest rate assumption.

Some central banks instead reverse engineer the process and use and sometimes publish (eg the Riksbank), their own agreed MPC members’ forecast path of the policy rate that is consistent with the projection of inflation being on target at the end of the forecast horizon. This is intended to help shape what the private sector should expect to happen to the policy rate going forward based, at least, on the policymakers’ current available information.^{footnote [37]} The staff of the South African Reserve Bank (SARB) also publish from time to time a (Taylor rule) policy path derived from their Quarterly Projection Model, but this does not necessarily reflect SARB’s MPC’s policy intentions.^{footnote [38]}

Irrespective of how the central forecast is produced, it is important in any communication to give a sense of the main risks and uncertainties around it. Most central banks do this through a verbal description in their post-meeting policy statements and regular reports. Some central banks also use quantifiable scenario analysis either internally to help inform the policy judgement or also externally to help explain the risks around the forecast and information on the policymakers’ reaction function.

Scenario analysis, as recommended by Bernanke (2024), could include assuming different paths for the policy rate (to help assess the optimal policy path), different assumptions on key exogenous variables including the conditioning assumptions, different parameters on key model relationships or different off-model judgements. Considering more than just the central scenario, could also be a useful way to communicate any big differences of view among individual MPC members both in terms of the key factors affecting the inflation outlook, the structure of the economy and/or the appropriate policy response.

Considering different scenarios is likely to be resource intensive, so judgements would need to be made over the most policy-relevant risk(s) or uncertainties to select at the current conjuncture for the scenario(s) and whether they should be published or rather used only internally by policymakers.

It is useful to assess the errors from past forecasts to see if there is any systematic pattern of under or over prediction of inflation. It should be possible to decompose the errors into various factors. These should include those due to unexpected movements in the conditioning assumptions, since these will be known ex post, those due to model misspecification (such as the relationship between the exchange rate and price inflation), data revisions and (off-model) judgments. If the inflation forecast is built up from its components it would be useful to assess the forecast errors from the sub-indices such as goods versus services or core versus non-core inflation.^{footnote [39]} This could help increase the policymakers’ understanding of the economy and hopefully improve their forecasting performance.

6: Conclusions and overview

Most of the issues discussed in this paper concerning the practicalities of inflation targeting, are similar in nature for developed countries and EMDEs alike. The most important considerations are that successful inflation targeting relies on building credibility in the regime and having an effective transmission mechanism for monetary policy. What are the particular issues that may affect African countries? We focus on three.

6.1: The level of the inflation target

Many fast-developing African countries may justifiably have higher inflation rates than developed countries. As mentioned in Section 4.3 and elsewhere, as a country develops, its productivity growth in manufacturing and other traded items should be faster than in more developed countries but (subject to a stable exchange rate), the inflation rates of such items are determined by world prices. Meanwhile, the prices of non-traded services tend to rise in line with real incomes rather than in line with productivity. A richer person can pay proportionately more for health, education, and other services, even if they would be objectively more valuable to a poorer person. Improvements in many personal services as an economy develops tend to be in quality rather than output volumes. Therefore, the prices of non-traded items, especially services, tend to start relatively lower in a developing economy, but increase faster, justifying a higher average domestic inflation rate overall – even with a stable exchange rate.

These considerations would justify inflation targets being somewhat higher in Africa than elsewhere and it is common to see the more successful IT regimes in Africa adopting targets of, say, 3%–6% per annum rather than 2%–2.5% per annum. Over time, as countries develop, and their experience with IT improves, one may see targets starting to be lowered. Continued inflation of 3%–6% a year may not feel like ‘price stability’.

A downside of a higher inflation target is that the exchange rate probably depreciates against currencies of countries with lower targets. It is easy to blame the exchange rate for causing inflation when the source lies elsewhere. Quite often a nominal depreciation is consistent with maintaining a constant real exchange rate and/or reflects domestic monetary and fiscal policy settings. Any policymaker needs to understand why the exchange rate is depreciating, before deciding on the appropriate response. If domestic inflation is stabilised at a low level, one main cause of depreciation should be minimised.

Whatever the target, it is important that policy is set credibly so as to deliver it at least cost. Credibility takes time to build but can be easily lost. It requires more transparency and accountability than many central banks are accustomed to historically, which may also be seen as the required basis for more monetary independence.

6.2: Fiscal dominance

Fiscal dominance can be a problem in any country, but especially for developing countries. Pressures on government to spend money promoting the general welfare of the population seem to be much greater in the modern world than they were historically, especially given the advent of democracies over the past century or so. Yet developing countries tend not to have a broad tax base, in part reflecting the smaller size of middle-income groups within the society and in part reflecting the size of the ‘informal’ economy.

While the government could be justified in borrowing to invest in the infrastructure required by a fast-growing economy, markets for sovereign debt in local currency are likely to be limited, especially if the country has a history of high inflation or default. Sovereign debt in foreign currency can be both expensive, reflecting low credit ratings, and risky to service. It is not surprising that, at least, historically in many African countries, governments resorted either implicitly or explicitly to an ‘inflation tax’ – printing money is an easy way to raise cash but is inevitably inflationary.

Governments with access to central bank balance sheets can create money at will. It is not clear that such inflationary-inducing spending improves welfare in the long run, while high inflation clearly causes significant harm.

A necessary condition for successful inflation targeting is that the central bank does not finance government spending. In some countries, where low inflation objectives are genuinely shared, this may be achieved by custom and practice. In others it more likely requires need some changes to central bank laws. To be clear, central banks are never completely independent from the rest of the public sector – central bank governors tend to be appointed by governments, who also control the relevant legislation and set the mandates for the central bank. Rather, the legal or other restrictions on monetary financing should be interpreted as a commitment by the government to low inflation that makes it harder to backtrack from. Whatever the precise arrangements, if they are broken or changed that should be required to be done publicly.

6.3: Improving the transmission mechanism

It is likely that the financial sector is less well-developed in most African countries than in developed economies. A much larger share of bank credit is also likely taken by the government crowding out the private sector. The monetary transmission mechanism via interest rates may therefore be weaker. The central bank, sitting at the heart of the financial system, does have the ability to encourage and help deliver structural improvements to the functioning of domestic money markets and foreign exchange markets.

While a handicap, a weak transmission mechanism need not be a roadblock in practice. As emphasised in this article, credibility is the real key to an IT regime, more so than the actual impact of interest rates, which can be weak even in many developed countries. Communication combined with decisive, independent policy action is still possible. There are several examples of African countries successfully employing IT regimes, albeit at slightly higher inflation rates than for developed countries.

References

Adrian, T, Erceg, CJ, Gray, ST and Sahay R (2021), Asset Purchases and Direct Financing: Guiding Principles for Emerging Markets and Developing Economies during COVID-19 and Beyond, IMF Departmental Paper No. 2021/023.

Bernanke, B (2024), Forecasting for monetary policy making and communication at the Bank of England: a review, Bank of England, April.

Borio, C, Hofmann, B and Zakrajšek, E (2023), Does money growth help explain the recent inflation surge?, BIS Bulletin No. 67, January.

Calvo, GA and Reinhardt, CM (2000), Fear of Floating, National Bureau of Economic Research Working Paper No. 7993, November.

Davig, T and Leeper, EM (2007), Generalizing the Taylor Principle, American Economic Review, Vol. 97, No. 3, pages 607–35.

Gopinath, G, Boz, E, Casas, C, Díez, FJ, Gourinchas, P-O and Plagborg-Møller, M (2020), Dominant Currency Paradigm, American Economic Review, Vol. 110, No. 3, pages 677–719.

Gopinath G, and Gourinchas, P-O (2022), How countries should respond to the strong dollar, IMF, October 2022.

Gourinchas, P-O (2025), As One Cycle Ends, Another Begins Amid Growing Divergence, IMF, January.

Ha, J, Ayhan Kose, M, Ohnsorge, F and Yilmazkuday, H (2023), Understanding the Global Drivers of Inflation: How Important Are Oil Prices?, World Bank Policy Research Working Paper 10283, January.

Hardy, B, Igan, D and Kharroubi, E (2024), Resilience in emerging markets: what makes it, what could shake it?, BIS Bulletin 88.

Honohan, P and Orphanides, A (2022), Monetary policy in South Africa, 2007-21, WIDER Working Paper Series WP-2022-29, World Institute for Development Economic Research (UNU-WIDER).

Hooley, J, Nguyen, L, Saito, M and Towfighian, SN (2024), Fiscal dominance and inflation: evidence from Sub-Saharan Africa, Public Sector Economics (pse-journal.hr), September.

IMF (2020), Toward an Integrated Policy Framework, IMF Policy Papers 2020/046, October.

IMF (2023), ‘Managing exchange rate pressures in Sub-Saharan Africa – adapting to new realities’. Analytical Note in Regional Economic Outlook for Sub-Saharan Africa, April.

IMF (2024a), ‘A Tepid and Pricey Recovery’, Analytical Note in Regional Economic Outlook for Sub-Saharan Africa, April.

IMF (2024b), Annual Report on Exchange Arrangements and Exchange Restrictions 2023, December.

IMF (2025a), ‘Macroeconomic effects of a potential change in South Africa’s Inflation Target’, South Africa: Selected Issues, pages 48–66.

IMF (2025b), World Economic Outlook Database, April.

King, MA (2024), Inflation Targets: Practice Ahead of Theory, NBER Working Paper 32594, June.

Lane, PR (2020), Low inflation: macroeconomic risks and the monetary policy stance, Keynote speech at the financial markets workshop of the Economic Council.

Loewald, C, Steinbach, R and Rakgalakaneet, J (2025), Less risk and more reward: revising South Africa’s inflation target, South African Reserve Bank Working Paper No. 25, May.

Lucas Jr, RE (1976), Econometric policy evaluation: A critique, in ‘The Phillips curve and the labor market’, ed. K. Brunner and A. Meltzer, Vol. 1 of Carnegie-Rochester conference on public policy, Amsterdam, North-Holland.

Pettinger, T (2022), The relationship between oil prices and inflation, June.

Pill, H (2024), Cross-checking, Speech given at Institute of Chartered Accountants for England and Wales, October.

Powell, J H (2024) The Economic Outlook, speech at the National Association for Business Economics Annual Meeting, Nashville, Tennessee.

Reinhart, CM and Rogoff, KS (2009), This Time Is Different: Eight Centuries of Financial Folly, Princeton, New Jersey: Princeton University Press.

Selassie, AA (2025), Evolution and Challenges for Monetary Policy in Africa, Keynote Address for the 50th Anniversary of the Banco de Moçambique, May.

South African Reserve Bank (2021), Monetary Policy Review October 2021.

South African Reserve Bank (2024), Monetary Policy Review April 2024.

Svensson, L (2010), Inflation Targeting, Handbook of Monetary Economics, Vol. 3, Chapter 22, pages 1,237–302.

Taylor, J (1993), Discretion versus policy rules in practice, Carnegie-Rochester Conference Series on Public Policy 39, pages 195–214.

Taylor, J (1999), A historical analysis of monetary policy rules, in J Taylor (ed.), ‘Monetary Policy Rules’, University of Chicago Press, pages 319–41.

Tiedemann, J, Bizimana, O, Dalmau, L and Ambassa, M (2024), Understanding Inflation Dynamics: The Role of Global Shocks in The Central African Economic and Monetary Community (CEMAC), IMF Working Paper, February.

Woodford, M (2001), The Taylor Rule and Optimal Monetary Policy, American Economic Review, Vol. 91 No. 2, pages 232–37.

World Bank (2019), Global Economic Prospects, Special Focus: Currency Depreciations, Inflation and Central Bank Independence, June.

Annexes