A. Introduction

1. Sierra Leone has experienced a sharp rise in inflation over the past three years, driven by a succession of external shocks and loose fiscal and monetary policies. Inflation increased from 18 percent in 2021 to 37 percent in 2022, before peaking at 56 percent end October 2023. It has since decreased to 36 percent as of May 2024, but remains elevated compared to peers. It also remains above the Bank of Sierra Leone’s (BSL’s) medium-term objective of single-digit inflation.

2. In this chapter, we ask what role fiscal and monetary policies played in the recent inflation surge. Our main finding is that macroeconomic policies played a decisive role in driving the sharp increase in inflation. The key policy implication is that the macroeconomic tightening supported by the new ECF arrangement will be critical in continuing to bring inflation down towards the BSL’s medium-term objective of single digit inflation.

3. The remainder of this chapter is organized as follows. In Section B, we document that the recent inflation surge was characterized by external shocks and loose fiscal and monetary policy. In Section C, we use a structural vector autoregression (SVAR) to quantify the role of fiscal and monetary policies in the recent increase in inflation relative to other shocks. We find evidence that monetary and fiscal policies explain more than half of the inflation surge in 2022 and 2023.

B. The Consumer Price Index and the Recent Surge in Inflation

4. The CPI basket contains 440 items with weights derived from the Sierra Leone Integrated Household Survey of 2018. Food makes up 40 percent of the basket, and includes staples such as rice, palm oil, cassava, Kinni-fish, and chicken. The weights on housing and utilities (8.9 percent), and transport (8.6 percent) are also significant (Text Table 1).

^{Text Table 1.}

Sierra Leone: COICOP of the CPI Basket

(Percent)

Sources: Statistics Sierra Leone.

^{Text Table 1.}

Sierra Leone: COICOP of the CPI Basket

(Percent)

	Vintage
COICOP	2008	2018
Food and non-alcoholic beverages	41.9	40.3
Alcoholic beverages, tobacco and narcotics	1.7	1.0
Clothing and footwear	7.3	7.7
Housing, water, electricity, gas and other fuels	13.7	8.9
Furnishings, household equipment and routine household maintenance	5.9	5.6
Health	11.4	7.6
Transport	7.8	8.6
Communication	2.0	4.7
Recreation and culture	1.5	2.6
Education	2.9	3.1
Restaurants and hotels	0.9	6.1
Miscellaneous goods and services	3.1	3.9
Total	100.0	100.0

Sources: Statistics Sierra Leone.

View Table

^{Text Table 1.}

Sierra Leone: COICOP of the CPI Basket

(Percent)

	Vintage
COICOP	2008	2018
Food and non-alcoholic beverages	41.9	40.3
Alcoholic beverages, tobacco and narcotics	1.7	1.0
Clothing and footwear	7.3	7.7
Housing, water, electricity, gas and other fuels	13.7	8.9
Furnishings, household equipment and routine household maintenance	5.9	5.6
Health	11.4	7.6
Transport	7.8	8.6
Communication	2.0	4.7
Recreation and culture	1.5	2.6
Education	2.9	3.1
Restaurants and hotels	0.9	6.1
Miscellaneous goods and services	3.1	3.9
Total	100.0	100.0

Sources: Statistics Sierra Leone.

5. Inflation surged dramatically after 2021 (Text Figure 1a). Inflation peaked at 56 percent in October 2023 before decreasing to 36 percent in May 2024.

6. Global food-price shocks and regulatory price increases contributed to the recent surge in inflation. Following shocks to global food prices, depreciation of the Leone, as well as domestic crop failures, food inflation peaked at 65 percent in October 2023, with inflation rates for some staples above 100 percent (Text Figure 1b). This sharp increase in food inflation resulted in a cost-of-living crisis, with the World Food Program estimating that 78 percent of households faced severe food insecurity in 2023. Inflation in the transport sector surged due to the government’s efforts to rationalize fuel and energy subsidies while curbing budget overruns. This involved a substantial hike in fuel pump and electricity prices, which rose by 40 percent and 91 percent in August and September 2023, respectively. Inflation in the hospitality industry reflects higher utility costs, including electricity and water.

7. Nevertheless, external shocks cannot fully account for the recent surge in inflation. First, despite facing similar external shocks, inflation in Sierra Leone is substantially higher than in peers (Text Figure 1c and 1d). Second, the surge is broad-based and persistent (Text Figure 1a). Third, the timing of the external shocks documented in the previous paragraph do not match the start and peak of the inflation surge.

View Full Size

Sources: Statistics Sierra Leone, IMF staff calculations, and WEO data.

• Download Figure
• Download figure as PowerPoint slide

8. Over the same period, fiscal and monetary policy were loose. The overall primary deficit in percent of GDP increased from 1.2 percentage points of GDP in 2019 to around 3.6 percentage points in 2022. Government spending increased during and in the wake of the COVID-19 pandemic as a result of emergency spending, the quick-action economic recovery program aimed to keep Sierra Leone’s youth employed in public work programs, the bail out of several state-owned agencies, increased cash transfers, and extensive road projects. A security incident in July 2022 and deadly floods in August led to additional spending to resettle affected persons. Furthermore, spending pressures picked up ahead of the general elections in June 2023. Meanwhile, the resulting surge in the government’s domestic financing need exceeded the domestic financial sector’s lending capacity. Consequently, the BSL began to purchase large volumes of government securities to finance the growing fiscal deficit. Year-on-year (yoy) reserve-money growth increased from 8.7 percent in December 2021 to 32 percent in December 2022, before accelerating to 63 percent in June 2023. This acceleration of reserve money growth is highly correlated with the rise in inflation (Text Figure 2a). While the BSL increased its policy rate by 8 percentage points between January 2022 and March 2024, ex-post real interest rates continue to be negative (Text Figure 4b).

View Full Size

Sources: Statistics Sierra Leone and IMF staff calculations.

• Download Figure
• Download figure as PowerPoint slide

C. Estimating the Effect of Macroeconomic Policy on Inflation

9. This section analyzes the extent to which the recent surge in inflation was driven by macroeconomic policy as opposed to other shocks to inflation. Box 1 provides an overview of existing empirical studies of inflation dynamics in Sierra Leone.

Methodology

10. We consider a three-variable SVAR

where y_t — [π_t,m_t,b_t]', π_t is quarter-on-quarter (qoq) CPI inflation, m_t is qoq reserve-money growth, b_t is the primary balance relative to GDP, a₀ is a constant, A_j is a coefficient matrix, p = 4 is the number of lags, and u_t is the vector of reduced-form errors related to the vector of uncorrelated structural errors ε_t := [ε^π_t,ε^m_t,ε^b_t]' via the relation u_t = Bε_t, for some invertible matrix B.

11. To assess the role of monetary and fiscal policy in driving inflation, we decompose ε_t into the shocks contained in £_t (see, e.g. Killian and Lütkepohl (2017, pp.109-116)). This requires estimating A_j and B. We estimate A_j by OLS using data spanning 2009Q2-2023Q4. B is unidentified without imposing further restrictions. We identify B using the heteroskedasticity-based method proposed in Lewis (2021), combined with mild restrictions on B.

12. We identify B up to column ordering by exploiting that the variance of the structural errors £_f is time-varying as in Lewis (2021). This assumption is testable using the Cragg-Donald statistic as suggested in Lewis (2021), and we find evidence that it holds in the empirical application we consider (all rank tests have a p-value smaller than 0.001). Since the method proposed by Lewis (2021) does not identify the column order of B, it yields six potential estimators, ${\hat{B}}_{\iota }$, i = 1,..., 6. We only keep the estimators ${\hat{B}}_{\iota }$ that satisfy:²

View Full Size

• Download Figure
• Download figure as PowerPoint slide

where * indicates that no restriction is imposed. The two restrictions are mild: they rule out that shocks to reserve-money growth can be disinflationary, and they rule out that a worsening of the primary balance causes a deceleration of reserve-money growth. We note that this identification scheme is different from the sign restrictions commonly employed in SVAR analyses. Indeed, the above approach yields a finite number of candidate structural covariance matrices. By contrast, a sign restriction approach would leave us with infinitely many structural covariance matrices, which makes the historical decompositions hard to interpret.

Results

13. The above approach leaves us with two candidate structural covariance matrices, ${\hat{B}}_1$ and ${\hat{B}}_2$. The historical decomposition for each of these is shown in Figure 3. The charts show the contribution to inflation net of the constant and the initial conditions, ${\hat{\pi }}_t$, of shocks to reserve money growth (yellow), the primary balance (purple), as well as other inflation shocks (blue). Intuitively, figure shows the decomposition of inflation in excess of its average value.

14. Idiosyncratic shocks to inflation (such as energy-price shocks) account for most of the inflation prior to 2017 and the inflation between 2020 and 2022. For both ${\hat{B}}_1$ and ${\hat{B}}_2$ we find that movements in inflation in these periods of time are primarily due to innovations ${\epsilon }_t^{\pi }$.

15. Between 2017 and 2020, macroeconomic policy was successful in containing inflationary pressures. The period 2017-2020 was disinflationary, and part of this disinflation is sometimes attributed to tightened macroeconomic policy. For instance, the Fiscal Strategy Statement 2019-2023 suggests that disinflation between 2017 and 2018 was due to tightened monetary policy. While we are unable to disentangle the effects of monetary and fiscal policy without imposing further restrictions on the structural covariance matrix, we find evidence that innovations in ${\epsilon }_t^{\pi }$ or ${\epsilon }_t^b$ contributed substantially to decreasing inflation. This shows that policy tightening between 2017-2020 was successful in containing inflation.

16. From 2022Q3 onwards, we find that more than half of inflation is due to macroeconomic policy shocks. For both ${\hat{B}}_1$ and ${\hat{B}}_2$, we find that innovations in ${\epsilon }_t^{\pi }$ or ${\epsilon }_t^b$ contributed substantially to increasing inflation. While we are unable to disentangle the effects of monetary and fiscal policy without imposing further restrictions on the structural covariance matrix, we nevertheless conclude that macroeconomic policy was the main driver of inflation from 2022Q3 onwards.

View Full Size

• Download Figure
• Download figure as PowerPoint slide

D. Conclusion and Policy Recommendations

17. Our analysis indicates that loose monetary and fiscal policies have played a significant role in driving the recent surge in inflation in Sierra Leone. We are able to show that fiscal and monetary policy account for more than half of the variation in inflation outturns in the recent inflation surge. We also find that policy tightening contributed to disinflation between 2017 and 2020. These findings underscore the need for continued policy tightening in the coming years to bring inflation down to the BSL’s single digit medium-term objective.