Climate Change and Inflation in Eastern and Southern Africa

Abstract

This study analyzes the dynamics of key climate change indicators and their implications on food prices in Eastern and Southern African Countries. The study uses descriptive and quantitative analysis of monthly data covering ten countries over the period 2001 to 2020. The descriptive analysis reveals that the sampled countries have experienced various climate change events with increasing intensity in the last two decades. Additionally, three of the countries in the sample ranked in the list of countries most affected by extreme weather events in 2019 are at risk of either frequent events or rare but extraordinary catastrophes. The quantitative analysis showed that supply shocks measured using rainfall amounts and imported food price inflation are the main determinants of food inflation, whereas oil prices, subsidies, and imported inflation are the key determinants of overall inflation. At a macro level, the analysis shows that all countries have various climate change policy initiatives in place but are still vulnerable to climate change risks. This implies a need for sector-specific climate change policy options that are most effective. In addition, the adoption of renewable sources of power such as wind and solar and appropriate irrigation practices is important.

1. Introduction

Climate change continues to pose a huge threat to human development [1]. Developing countries are most vulnerable to climate change effects, frequently experiencing extreme weather patterns such as drought, floods, heatwaves, storms, precipitation variations, and changes in sea level with devastating effects on agriculture, food security, nutrition, housing, health, infrastructure, and incomes [1,2,3]. These developments threaten efforts to reduce extreme poverty, especially in low-income countries, and have led to reversals of gains for certain groups in terms of income, health, and education outcomes besides increasing global inequalities.

Rainfed agricultural systems in Africa are particularly vulnerable to climatic change. It is estimated that more than 70 percent of the population in Africa lives in rural areas and is largely dependent on agriculture [4,5,6]. Moreover, more than 25 percent of the continent’s Gross Domestic Product (GDP) is derived from agriculture which is more effective in reducing poverty than non-agricultural growth [4,7]. Climate change not only increases uncertainty in agricultural production but also reduces the nutrients in the soil that ensure healthy crop production hampering productivity. Thus, agricultural produce lacks proper nutrients, is less, and is subsequently sold at relatively higher prices. Higher and rising food prices signal the existence of imbalances in supply and demand and growing resource scarcity, driven either by demand factors such as growing population and income or by supply factors such as reduced productivity due to climate change and reduced agricultural land due to soil degradation and conversion to other uses [8].

Earlier theoretical literature hypothesized that increases in prices arising from excessive money growth were inherently inflationary and the effects of money supply on price level were long-term and significant [9,10,11]. The monetarist argument hinged on unanticipated changes in the growth rate of the money supply, which often led to a surge in commodity prices [11,12,13,14]. Other theorists ascribed inflationary pressures to climate change and weather variability [15,16,17], which often affect food production, food processing, food availability, and food access [17,18,19]. The main driving factors include the delayed provision of government social safety nets and subsidies, underdeveloped infrastructure, and inefficient foreign trade networks [18,20,21,22], which hinder internal distribution and international trade [23,24,25,26].

The empirical literature on climate change’s impact on food prices, particularly in developing countries, remains largely unexplored [27,28,29,30,31]. Existing empirical studies have largely modeled inflation dynamics with monetarist features and directly augmented the approach with structural supply-side factors, [32,33,34,35,36,37] for cross-country studies and [38,39,40,41,42,43,44] for country-specific studies. However, recent studies have shown that supply-side constraints pose a major problem in managing prices in Africa and these supply constraints are largely related to the outcomes of climate change, calling for explicit modeling of climate change risk indicators [45,46,47].

Empirical studies on the determinants of food inflation, proxied climate change using average temperature changes [33,44], observed rainfall amounts [40], and rainfall deviation from the long-run trend [42,44]. Other studies used agricultural produce of the most consumed cereal to be an indirect proxy for climate change. The authors in [41] used the price of maize to capture the effects of government controls on maize prices up to 1990, after which, maize prices captured supply constraints that fed directly into the consumer price index. The study [42] also used cereal prices to capture the effects of climate change in the short term. The foregoing reviewed studies reveal a dearth of knowledge on climate change–price relationships based on eastern and southern African data. Yet these countries have continuously experienced unpredictable weather patterns and variable food production with attendant impacts on food prices and inflation. Additionally, countries in the eastern and southern regions depend on agriculture relative to countries in western, central, and northern Africa. The weight of food in the sampled countries is significant, accounting for about 40 percent of households’ consumption basket, compared to 30 percent for other developing countries [34].

Overall, the empirical literature has evolved alongside the theoretical literature. Study findings for developing economies generally reveal that the inflationary process transcends beyond monetarist explanations and have thus incorporated a structuralist approach to model inflation dynamics. In this study, climate variables are considered because of their effect on agriculture and energy generation, which are important in explaining food and electricity prices. Additionally, a combination of food and electricity items constitute a significant weight in the consumption basket of households in developing countries.

This study, therefore, seeks to: (1) analyze the climate disaster events and climate risk indicators in the eastern and southern Africa region and (2) empirically establish the impact of climate change risk indicators on food prices and overall inflation in the eastern and southern African region. The study hypothesizes that (1) climate change measured by temperature and rainfall variability will directly and negatively affect food and overall inflation and (2) erratic weather patterns will indirectly and negatively impact on food and overall inflation. The study uses qualitative assessment to address the first objective and quantitative-panel estimation techniques to assess the second objective.

2. Data and Methods

The study used monthly time-series data for a sample of eight countries from the eastern and southern African region, namely: Zambia, Malawi, Rwanda, Burundi, Uganda, Tanzania, Kenya and Mozambique. The study covers 20 years from 2001 to 2020. The selected estimation sample was conditioned on the availability of data on the main variables of interest and also the availability of high frequency monthly data. Zimbabwe and Ethiopia were dropped from the estimation sample due to a lack of monthly time series. The trend analysis retains the ten countries since it utilizes annual data. Overall consumer price and food price indices, lending interest rates, exchange rates, and GDP were obtained from respective country databases while the cereal price index was obtained from the FAOSTAT database. The data for rainfall and temperature was obtained from the World Bank Climate Change Knowledge Portal. Data on subsidies was sourced from the World Bank, World Development Indicators database. The global oil price index was obtained from the IMF commodity price database. Data on foreign prices were obtained from the Federal Reserve Bank of St. Louis.

Consistent with previous studies, this study specifies an inflation model incorporating proxies for climate change in the form of monthly mean rainfall amounts, variability in rainfall, and variability in temperature [33,34,35,41,42,44,48]. The general equation for estimation is specified in Equation (1):

$${\pi }_{it}={\alpha }_i+\beta X_{it}+{\pi }_{it-1}+z_t+{\epsilon }_{it}$$

(1)

where ${\pi }_{it}$ is the food/overall consumer price index for country i in time t, $X_{it}$ represents exogenous variables driving inflation, ${\pi }_{it-1}$is lagged dependent variable, ${\alpha }_i$is country-specific effects, $z_t$ is a vector of time-specific effects and ${\epsilon }_{it}$ is the error term. Specifically, the model to be estimated is specified in Equation (2):

$$\begin{gathered} \mathit{CP}{I}_{it}={\alpha }_i+{\beta }_1\mathit{REE}{R}_{it}+{\beta }_2\mathit{In}{t}_{it}+{\beta }_3\mathit{GD}{P}_{it}+{\beta }_4\mathit{Climat}{e}_{it}+{\beta }_5\mathit{Foreign}{P}_t \\ +{\beta }_6\mathit{cerealprice}{s}_t+{\beta }_7\mathit{oilp}{r}_t+{\beta }_8\mathit{Subsi}{d}_t+{\beta }_{9}CP{I}_{it-1}+z_t+{\epsilon }_{it} \end{gathered}$$

(2)

where the consumer price index (CPI) captures food CPI and overall CPI. Overall CPI and food CPI are the dependent variables in this study. REER is the Real Effective Exchange Rate, Int is the lending interest rate, and climate refers to the three proxies of climate change risk indicators, namely: monthly mean rainfall amounts, variability in rainfall, and variability in temperature; global prices are represented by global oil price index–oilpr, US overall consumer prices–foreignp and cerealprices–global cereal prices. Government intervention is captured by subsidy. The study uses dynamic panel data model consistent with previous studies. Details on the estimation method are in Appendix A.

Consistent with [35], climate variables are proxied for by rainfall and temperature. We use high-frequency rainfall and temperature data to capture the variability in the food supply that leads to high food prices. Specifically, the variables are defined as average monthly rainfall, variations in monthly rainfall and monthly temperatures from the long-term mean. Improved rainfall is expected to lead to a reduction in food inflation as supply increases, attributed to good agricultural production. However, high fluctuations in rainfall and temperature are expected to be inflationary. The effect on inflation from temperature variability is contrary to the average monthly rainfall variable since high temperatures are associated with drought, which is inflationary. Temperature changes lead to low rainfall and drought, which also affect hydropower generation capacity with direct effects on the price of electricity, a component of the consumer basket, and indirect effects through electricity price linkages with other components in the consumer basket such as food and non-food non-fuel items. The expected a priori sign between temperature and inflation is positive.

Macro variables are expected to have differential effects on food inflation. Real income growth increases the demand for real money balances leading to a reduction in the price of non-tradables if it is assumed that their price is domestically set and moves in tandem with overall demand in the economy. Increases in GDP imply increased availability of goods and services domestically and a reduction in the general price level. At the same time, higher GDP growth also implies high imports and hence affects the degree of pass-through to domestic prices. Thus, depending on the proportion of imported goods in GDP and the price of those imported goods, an increase in GDP can also be inflationary. The expected a priori sign is ambiguous [49,50,51]. Exchange rate movements impact inflation through the import channel from the international market. Depreciation of the currency is inflationary while appreciation is deflationary. Thus, the expected sign could be positive or negative depending on the degree of dependence on imported goods [32,33,34,35,38,41]. Import price is included as a measure of the cost of goods and services bought by residents from a foreign country including intermediate goods as well as final products. Higher import prices are expected to increase domestic prices [52].

Foreign food inflation and foreign inflation are expected to transmit to domestic inflation. For countries with a significant proportion of import content of food and other items in the consumption basket of households or with strong links with the international food supply chain, an increase in foreign food prices would increase domestic prices. In Kenya, the transmission of international food prices to domestic food inflation is weak. However, the transmission of foreign inflation through imported goods is found to be strong. The expected a priori sign is therefore positive [41,42,43,44]. Price movements in the international oil market transmit through to the domestic fuel prices in all the countries under consideration, as they are mainly importers of fuel. An increase in the international oil price would therefore lead to an increase in domestic fuel prices, which then leads to an increase in food prices due to increased transportation costs. Therefore, the expected a priori signs for both international oil prices and domestic oil prices to food inflation would be positive [41,42,44].

3. Results

Trend Analysis on Climate Change Risk Indicators and Disaster Events

Most countries across the world including Africa have experienced rising temperatures and variability in rainfall patterns (Figure 1 and Appendix D). The rainfall patterns have been erratic in Kenya, Tanzania, Uganda, Burundi, and Rwanda while extreme mean surface temperature changes have been witnessed in Zambia and Tanzania [53]. Most of these countries have experienced productivity decline and high food prices during periods of extreme weather as was the case during El Nino and La Nina seasons [54].

Scientists are confident that many of the observed changes can be linked to the levels of carbon dioxide and other greenhouse gases in the atmosphere, which have increased due to human activities. The sampled countries are experiencing high levels of greenhouse gas emissions with rising carbon dioxide (CO₂) emissions recorded in Kenya, Tanzania, Ethiopia, Zimbabwe, Mozambique, Uganda, and Zambia [55] (Figure 2). Moreover, WMO [1], the climate risk index of 2019, showed that several countries under review were at risk of either frequent events or rare but extraordinary catastrophes with implications on food prices and overall inflation.

Similarly, disaster events arising from climate risk indicators have increased. All the countries in eastern and southern Africa have had floods in the last decade as the main natural disaster with the greatest impact on human lives [56]. Storms have also been destructive in Zimbabwe, Mozambique, and Tanzania while landslides have been prevalent in Rwanda, Burundi, and Uganda. The intensity of drought has increased in the recent past. The East African countries experienced severe drought seasons in the years 2011, 2008 and 2009, 2005 and 2006, 2000 and 2001, of varying degrees of severity (Appendix B). More recently, the El Nino experienced in 2017 disrupted weather patterns in 2018 and 2019 [1]. This led to below-average rainfall in the short rainy season between October and December 2018, a delayed onset, and below-average rainfall during the March to May long rain season of 2019.

The climate disaster events in these countries have led to high economic losses including damages to crops and livestock, loss of lives and livelihoods through death and displacements, diminished food security, famine, low capacity to generate electricity, damage to infrastructure, and low agricultural supply leading to high consumer food prices across the selected countries (Figure 3). In some countries, rising food prices necessitated short-term measures such as a ban on exports of food crops and suspension of import tariffs on food imports to cushion against high food prices. Other countries such as Rwanda intervened through the targeted distribution of agricultural inputs. However, given the import dependence of the countries in the sample, some trade measures implemented to cushion consumers in respective countries led to an escalation of food prices in the neighboring countries as was the case in Rwanda following the ban of exports of food from Burundi in 2016.

Figure 3 panel 1, shows food inflation rates for selected countries in East Africa and corroborates the observed weather patterns, during which significantly high food inflation was experienced in 2008 and 2009, especially in Ethiopia. Food price increases averaged 59.1 percent, 34.1 percent, and 24.4 percent in 2008 for Ethiopia, Burundi, and Kenya, respectively, while they remained below 20 percent, in Uganda, Rwanda, and Tanzania. Food inflation rates for selected countries in the South African region show significant variability (Figure 3 panel 2). Similarly, the region has also experienced a fair share of weather vagaries. Malawi experienced acute food shortages from July to September 2018 and subsequently below-average rainfall. Mozambique also experienced cyclone Idai in March and cyclone Kenneth in April 2018 leading to a food crisis in the subsequent period to December 2018. These periods were accompanied by high food inflation as shown in Figure 3 panel 2. Maize-producing areas in Zambia and Zimbabwe were negatively impacted by severe drought in 2019, resulting in reduced domestic supplies, which triggered higher prices of maize. However, it is noteworthy that not all high food inflation is caused by the negative effects of climate change particularly in Zimbabwe where other domestic factors contributed to the high food inflation.

Table 1. Climate Risk Index 2019.

Country	CRI, 2019 (Rank)	CRI, 2000–2019 (Rank)	CRI Score	Fatalities in 2019 (Rank)	Fatalities per 100000 Inhabitants, (Rank)	Losses in Million US$ PPP (Rank)	Losses per Unit GDP in % (Rank)
Kenya	25	34	33	15	16	49	51
Uganda	31	66	42.1	21	19	68	63
Tanzania	67	122	66.5	27	47	88	95
Rwanda	42	117	53.3	45	21	99	67
Burundi	57	74	61.8	26	10	121	102
Ethiopia	72	60	69.3	39	74	67	81
Mozambique	1	5	2.6	2	3	4	2
Malawi	5	62	15.1	20	13	35	5
Zambia	59	123	63.3	80	90	56	32
Zimbabwe	2	15	6.1	6	2	21	3

Source: Eckstein, Kunzel and Schafer, Global Climate Risk Index, 2021.

provides the Climate Risk Index (CRI) rank for 2019 and 2000–2019 of countries under review. The table also provides the socio-economic impact of climate change in 2019 for the reviewed countries. The CRI indicates the level of exposure and vulnerability to extreme weather events, which countries should understand as warnings to be prepared for more frequent and/or more severe events in the future. The CRI analyses rank to what extent countries and regions have been affected by impacts of climate-related extreme weather events, mainly, storms, floods, drought, landslides, heat and cold waves [56]. Generally, the CRI shows that the intensity of extreme weather has been increasing in all the countries in our sample in the last 20 years. Specifically, in 2019, the CRI indicates that three of the countries in our sample were most affected by the impacts of extreme weather events, with Mozambique, Zimbabwe and Malawi ranked first, second, and fifth, respectively, in 2019.

Several countries in the sample have taken various long-term initiatives to mitigate against climate change. All countries have made commitments as Nationally Determined Contributions (NDCs) on aggregate 2030 emissions reduction targets. Most of the countries have developed Climate Smart Agriculture in the context of global obligations and local realities and as signatories to the Kyoto Protocol and Paris Agreements. Kenya, Ethiopia, Malawi, Zambia, and Zimbabwe also developed national climate change learning strategies in 2020 and 2021 aimed at raising awareness and strengthening climate change knowledge, including plans to mainstream climate change into the education curriculum. Other initiatives by other stakeholders include the most recent by the Central Bank of Kenya, which issued guidance on climate-related risk management aimed at enabling banks to integrate climate-related risks into their governance, strategy, risk management, and disclosure frameworks. Specifically, it requires CBK licensed banks to (a) embed the consideration of the financial risks from climate change in their governance arrangements; (b) incorporate the financial risks from climate change into their existing financial risk management practice and (c) develop an approach to the disclosure of the financial risks from climate change. However, the initiatives undertaken to mitigate against climate change are not yet adequate. The policies to promote environmentally friendly agriculture through climate-smart agriculture systems have not provided adequate mitigation against climate risks for these countries since over half of the countries have adopted climate-smart agriculture systems yet they are still vulnerable to risks of climate change (Table 1 and Appendix B). This analysis implies that it is not the lack of climate change policies or initiatives at a macro level that is an issue but rather the identification of effective climate change policies in each of the selected countries.

4. Empirical Results and Discussion

This section presents empirical results and discussions. The dependent variables are overall inflation and food inflation. Three different indicators of climate change are used as independent variables, namely: average monthly rainfall, variation in monthly rainfall, and variation in temperature and the results are shown in

Table 2. Empirical results with average monthly rainfall.

Independent Variables	The Dependent Variable Is Overall Inflation			The Dependent Variable Is Food Inflation
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Average Rainfall	−0.11(−2.34) ***	−0.06(−1.60) *	−0.05(−1.89) **	−0.028(−2.17) **	−0.02(−2.11) **	−0.03(−3.43) ***
GDP	−0.02(−1.36)	0.01(1.26)	0.03(0.58)	−2.10 × 10−10(−1.96) **	−1.96 × 10−10(−1.67) *	−5.53 × 10−10(−1.67) *
Oil prices	0.02(3.25) ***	0.01(2.46) ***	0.01(2.45) ***	0.27(0.57)	0.35(0.69)	0.68(0.32)
Foreign prices	0.42(2.56) ***	0.54(1.20)	0.88(2.76) ***
Inflation(−1)	0.13(11.3) ***	0.11(7.55) ***	0.08(8.97) ***	0.93(9.74) ***	0.91(9.75) ***	0.97(4.36) ***
Interest rate	−0.10(−1.24)		−0.45(−1.40)
Real Effective Exchange Rate		−0.53(−2.37) ***	−0.79(−1.72) *		−0.11(−0.56)	0.14(0.25)
Subsidy			−0.02(−0.57)			−0.05(−0.48)
Cereal price				0.006(2.31) ***	0.006(2.17) **	0.01(3.49) ***
No. Obs.	1216	1290	645	1621	1621	609
R2	0.74	0.58	0.84	0.44	0.45	0.47
J stats(p_value)	0.67(0.95)	7.29(0.12)	1.33(0.51)	1.89(0.92)	1.72(0.88)	9.04(0.10)

Note: For all the coefficients, the t–statistics are in parenthesis; *, **, *** denote 10%, 5% and 1%, significance levels, respectively.

,

Table 3. Empirical results with rainfall variability.

Independent Variables	The Dependent Variable Is Overall Inflation			The Dependent Variable Is Food Inflation
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Rainfall variability	0.23(5.69) ***	0.13(2.15) **	0.06(1.93) **	0.05(2.09) **	0.05(1.93) **	0.04(4.91) ***
GDP	−0.01(−0.99)	0.01(1.36)	0.02(0.47)	−2.12 × 10−10(1.97) **	−2.00 × 10−10(1.70) *	−0.07(−1.85) *
Oil prices	0.16(0.15)	0.94(1.73) *	0.008(2.81) ***	0.54(0.93)	0.59(1.01)	0.07(0.19)
Foreign prices	0.69(3.60) ***	0.02(0.05)	0.98(2.96) ***
Inflation(−1)	0.13(10.7) ***	0.13(5.60) ***	0.08(12.6) ***	0.92(9.24) ***	0.90(9.56) ***	0.99(10.9) ***
Interest rate	−0.19(−1.81) *		−0.02(−1.12)
Real effective exchange rate		−0.54(−2.30) **	−0.65(−2.06) **		−0.09(−0.45)	026(0.82)
Subsidy			−0.05(−1.78) *			−0.11(−1.33)
Cereal				0.007(2.52) ***	0.007(2.34) ***	0.17(1.64) *
No. Obs.	1216	1308	652	1639	1639	703
R2	0.86	0.53	0.88	0.44	0.44	0.57
J stats(p_value)	4.39(0.49)	8.75(0.11)	2.47(0.28)	2.29(0.89)	2.17(0.82)	9.98(0.12)

Note: For all the coefficients, the t–statistics are in parenthesis; *, **, *** denote 10%, 5% and 1%, significance levels, respectively.

and

Table 4. Results with temperature volatility.

Independent Variables	The Dependent Variable Is Overall Inflation			The Dependent Variable Is Food Inflation
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Temperature variability	0.04(2.39) ***	0.05(2.50) ***	0.04(1.80) *	0.16(1.27)	0.18(0.66)	−0.03(−0.61)
GDP	0.01(1.22)	0.004(0.22)	0.006(0.04)	−2.10 × 10−10(−1.91) **	−1.13 × 10−11(−1.93) ***	−0.07(−1.82) *
Oil prices	0.01(2.08) **	0.92(2.53) ***	0.009(5.49) ***	0.58(0.97)	0.69(063)	0.67(1.73) *
Foreign prices	0.91(1.95) **	0.63(3.58) ***	0.51(2.32) **
Inflation(−1)	0.10(5.06) ***	0.08(7.26) ***	0.08(4.54) ***	0.91(9.61) ***	0.66(6.19) ***	0.90(13.7) ***
Interest rate	0.11(1.16)		−0.01(−0.49)
Real effective exchange rate		−0.008(−2.19) **	−0.007(−0.90)		−0.69(−1.92) **	0.10(0.49)
Subsidy			−0.11(−3.29) ***			−0.05(−0.09)
Cereal				0.005(2.14) **	0.008(2.79) ***	0.10(1.68) *
No. Obs.	1232	1163	624	1639	1599	664
R2	0.66	0.64	0.8	0.44	0.46	0.56
J stats(p_value)	7.50(0.11)	8.79(0.26)	0.01(0.99)	3.41(0.75)	3.91(0.56)	9.34(0.15)

Note: For all the coefficients, the t–statistics are in parenthesis; *, **, *** denote 10%, 5% and 1%, significance levels, respectively.

. A robust assessment was conducted using the rainfall anomaly index (RAI) as an alternative indicator for climate change. The results are largely similar to those obtained when average rainfall and rainfall variability was used. The results using RAI are reported in Appendix C.

The variation in rainfall and temperature are included as independent variables to assess the impact of volatility in climate variables on overall inflation and food inflation. To disentangle the effect of interest rates and exchange rates, in Table 2, Table 3 and Table 4, we present the results of the models with interest rates but without exchange rates in the second and fifth columns, respectively, where overall and food inflation are dependent variables. In separate models for overall and food inflation in the third and sixth columns, respectively, we reported results with exchange rate but without interest rate. In the fourth and seventh columns, we used subsidies and other transfers indicator to capture government policy intervention.

Generally, the results in Table 2, Table 3 and Table 4 reveal minimum sensitivity to the climate change indicator used as they remain largely unaltered in terms of direction and significance. In Table 2, we used average monthly rainfall in the six models with overall and food inflation as dependent variables. The results show that mean monthly rainfall has a negative and significant effect on both overall and food inflation. The results imply that good weather reflected in high rainfall leads to lower food prices and lower inflation since higher rainfall supports good harvest, particularly of vegetables, cereals, and legumes. This result is consistent with previous results that also established that higher rainfall leads to the availability of food products and reduces food prices, [29,38,39,40,57,58]. The significance of rainfall amounts points to the need to prioritize investment in policies that lead to reliable water supply such as irrigation and or policies to enforce water storage facilities in all households and improve the accuracy of weather forecasting technology. However, it should be noted that irrigation programs not only require massive investments but have been tried in some African countries, yet the issue of food sufficiency has not been solved. It may, therefore, be necessary to evaluate the impact of irrigation on food productivity against costs at the country level and integrate experiences of countries where irrigation policies have worked such as Israel and Asia.

The coefficient of foreign international prices and international cereal prices is positive and significant in the overall inflation and food inflation models, respectively. The two variables reflect the impact of external cost-push factors on domestic prices. The results reflect a strong transmission of foreign prices through imported goods to overall inflation and a strong transmission of cereal prices to domestic food inflation. The results are in line with other empirical studies such as [41,50,52,59,60]. In terms of policy, this result can be interpreted to signal a need for the countries in the sample to invest in food self-sufficiency rather than relying on imported food.

The coefficient for GDP is negative and significant in the food inflation models. The results imply that an increase in GDP reflects the increased availability of goods and services including agricultural goods and thus reduced food prices. Consistent with the theory of linkages between GDP and inflation, the result can also be interpreted to mean that the source of the increase in GDP is dominated by non-tradables as opposed to tradables. These results are consistent with previous work that showed a negative relationship between GDP and inflation [50,51].

In Table 3, we replaced amounts of rainfall with rainfall volatility, which is the second indicator of climate change used in this study. The effect of rainfall variability reported in the second row has a positive and significant effect on both overall and food inflation. Most of the countries in the sample depend on rainfed agriculture for food supply and support of cash crops. Thus, changes in the amount and timing of rainfall within the season and an increase in weather changes reduce agricultural supply including food products leading to high food prices and overall inflation. These results corroborate previous work, [48,61,62].

Whereas the coefficient of exchange rate and oil prices bears the expected sign and is significant in the overall inflation models, they are not significant in the food inflation models. The significance of the exchange rate in overall inflation reflects the effect of imported inflation on the components in the consumer price index that are imported. Similarly, the significance of the coefficient of oil in overall inflation and not in food inflation models would be reflecting oil pass-through to other components in the consumer basket such as transport, electricity, and non-food non-fuel items. The coefficient of interest rate is not significant whether it is included separately or together with the exchange rate in the same model implying that the effect of imported inflation is more significant compared to the effect of interest rate.

The results further show a positive and significant coefficient of the lagged overall and food inflation variables. As noted in [63], the extent of inflation inertia is usually taken as measuring the consequences of indexation or inflation expectations. It thus largely reflects expectations, which in the presence of quantity and price controls, might create self-enforcing expectations of increasing inflation. However, the magnitude of inertia is much higher in food inflation models than in the overall inflation models. In line with previous studies, it is possible that the high and significant levels of the lag of inflation would be reflective of the impact of structural factors such as price regulation and market inefficiency on inflation, which makes price formation inflexible. Under inefficient markets, the price would be inflexible due to, for example, a lack of competition and information, [41,43,50]. This finding is consistent with previous work, especially for emerging markets and developing countries [63,64,65].

In Table 4, we used temperature volatility, which is the third indicator of climate change. Similar to the results in Table 2 and Table 3, the indicator of climate change used does not alter the results of the other independent variables in terms of direction and significance in both the overall and food inflation models. Apart from the climate change indicators, the other main determinants of overall inflation remain international foreign prices, oil prices, exchange rates, and government subsidies while the determinants of food inflation include cereal prices and GDP.

The coefficient of temperature volatility is positive in both the overall and food inflation models but is significant only in overall inflation models. The significance of temperature volatility in the overall inflation model would be explained by the fact that temperature changes do not only affect food products but also other components in the consumer basket, particularly electricity, whose price increases when the production costs increase. The weight of electricity in the consumer basket is considerable in the countries under consideration. Moreover, prices of electricity are also linked to other components in the consumer basket including non-food non-fuel items. Similar results were reported by [29,44,66,67,68]. Based on this result, it can be argued that a policy focusing on cheaper sources of energy such as solar and wind would be beneficial in reducing the reliability of hydroelectricity.

5. Conclusions

This study sought to establish the impact of climate change on food and overall inflation in selected eastern and southern African countries based on both descriptive and quantitative approaches. The descriptive analysis shows that all the sampled countries have experienced various climate change events with increasing intensity in the last two decades. Erratic weather changes manifested in floods had the greatest negative impact on human life in all the countries under review while in the recent past, storms have been prevalent in Zimbabwe and Mozambique. Analysis of the CRI in 2019 shows that three countries (Mozambique, Zimbabwe, and Malawi) under review appear in the list of the ‘bottom 10’, referring to countries that were most affected by extreme weather events in 2019 and should consider the CRI as a warning sign that they are at risk of either frequent events or rare but extraordinary catastrophes. High food prices resulting from the impact of erratic weather patterns on agricultural supply necessitated the implementation of short-term policy interventions which, however, was often counterproductive for countries that depend on one another to close food deficits as is the case for some countries in our sample.

Results from the quantitative analysis show that climate change, measured by amounts of rainfall, rainfall volatility, and temperature volatility, has significant effects on food and overall inflation. The results show that while total mean monthly rainfall has a negative and significant effect on overall and food inflation, rainfall variability has a positive and significant effect on both food and overall inflation. Similar results are found in the analysis between temperature variability and overall inflation. Temperature variability affects not only food products but also the capacity to generate electricity with implications on food prices as well as prices in the other components of the consumer basket including non-food non-fuel prices. The results show a strong transmission of foreign inflation through imported prices to domestic prices. However, the interest rate variable is not significant in explaining both overall and food inflation while subsidy is not important in explaining food inflation.

The study showed that short-term trade measures and domestic agricultural price policy measures do not provide lasting solutions to high food prices arising from extreme weather patterns. Long-term and effective climate mitigation measures are thus critical. Climate change action measures towards the attainment of net-zero emissions will require massive financial resources, multi-pronged approaches as well as coordinated actions across countries. Advanced economies that are progressive in the trajectory of climate change mitigation have allocated investment towards a green economy and carbon pricing. Conversely, most African countries including the ones covered in our sample, have limited fiscal space and high public debts, especially following the COVID-19 pandemic. Consequently, macro-level policy options for climate change mitigation requiring massive fiscal easing may not be feasible in the near to medium term. Based on the results of this study, therefore, it would be realistic to recommend that such countries consider focusing on the enhancement of sector-specific policies such as investment in other relatively cheaper sources of clean energy (solar and wind). This policy option will not only reduce carbon emissions but will also reduce reliance on hydroelectricity, therefore reducing electricity prices and inflation. Additionally, considering the dependencies of food imports across sampled countries and Africa in general, African governments can consider joint and coordinated climate action strategies at the continental level. For instance, this would include the development of a continental green growth investment package with clear responsibilities across African countries.

Based on the results of this study, it can also be argued that the significance of rainfall amounts in reducing prices implies a need to support the prioritization of investment in policies that lead to reliable water supply such as irrigation and food self-sufficiency strategies. However, some of these policies have been tried in some countries in our sample yet they have not attained food self-sufficiency as is the case in countries such as Israel where irrigation worked. Even as this policy is recommended, it may be necessary to evaluate the impact of irrigation on food productivity against costs at the country level and integrate experiences of countries where irrigation policies have worked. It is also worth noting that most countries in our sample launched climate change learning strategies in 2020 and 2021 implying that the understanding of green growth has not been achieved. Entrenched awareness would be one of the policy options, which has worked in advanced economies. At a macro level, the analysis supports the need for investment in further research to establish effective climate change policies and effective good practices in other countries that can be customized in Africa.

This study’s main limitation is the availability of time series data for climate change and weather-related indicators. In this study, it was not possible to empirically examine the impact of greenhouse emissions and weather indicators such as drought and flooding due to a lack of high-frequency data. The construction of such data would improve future research and enhance quantitative analysis on diversified measures of climate change.