Project – Application of Time Series Analysis in Forecasting Inflation Rates in Nigeria (2005 to 2025)

CHAPTER ONE

GENERAL INTRODUCTION

Research Topic: Application of Time Series Analysis in Forecasting Inflation Rates in Nigeria (2005–2025)

1.1 Background to the Study

Inflation is one of the most important macroeconomic indicators used to assess the stability and performance of an economy. It refers to a sustained increase in the general price level of goods and services over time, resulting in a decline in the purchasing power of money. When inflation rises, consumers may find it increasingly difficult to afford essential commodities, while businesses may experience higher production costs, uncertainty in investment decisions, and difficulties in planning their operations. Inflation is commonly measured through changes in the Consumer Price Index (CPI), which tracks price movements across a representative basket of goods and services consumed by households. According to the International Monetary Fund (IMF, 2022), inflation in Nigeria has historically been persistent, with monetary conditions, exchange-rate depreciation, and structural weaknesses contributing to price instability. Similarly, Mankiw (2022) explains that sustained inflation affects the real value of money and can create economic uncertainty when price increases become difficult to predict. Consequently, reliable inflation forecasting is essential for economic planning, monetary policy formulation, business decision-making, and the protection of household purchasing power.

Time series analysis is a statistical approach used to examine observations collected sequentially over time and identify patterns that may help explain or forecast future values. It involves analysing historical data to identify trends, seasonal movements, cyclical fluctuations, and random variations. In economic applications, time series methods are particularly useful because variables such as inflation, interest rates, exchange rates, and output frequently exhibit dependence on their past values. Box, Jenkins, Reinsel, and Ljung (2015) explain that autoregressive integrated moving average (ARIMA) models provide a systematic framework for identifying, estimating, diagnosing, and forecasting time-dependent observations. Hyndman and Athanasopoulos (2021) similarly demonstrate that effective forecasting requires the identification of appropriate patterns, careful model selection, and evaluation of forecast accuracy. In Nigeria, where inflation has been affected by changes in food prices, energy costs, exchange rates, and monetary policy, time series analysis can help identify historical inflation patterns and generate forecasts that support economic decision-making. Therefore, applying suitable time series models to Nigerian inflation data may provide a more systematic basis for anticipating future price movements than relying exclusively on informal expectations or simple historical averages.

The application of time series analysis to inflation forecasting has gained importance because inflation does not necessarily follow a constant or predictable path. It may increase gradually, decline temporarily, or rise sharply following economic disturbances. Forecasting models must therefore account for the possibility that inflation exhibits persistence, structural changes, and different responses to economic shocks. Stock and Watson (2007), in their study of inflation forecasting, examined the performance of different forecasting methods and demonstrated the importance of evaluating models against suitable benchmarks. Box et al. (2015) also established that ARIMA modelling can capture temporal dependence through autoregressive and moving-average components, provided the model is appropriately specified. For Nigeria, these methods may help determine whether past inflation rates contain useful information for forecasting subsequent observations. However, the suitability of a particular model cannot be assumed in advance; it must be established through statistical estimation, diagnostic checking, and forecast evaluation. This makes the application of time series analysis relevant to a study that seeks to examine historical inflation behaviour and determine an appropriate forecasting model for Nigeria.

Nigeria provides an important setting for inflation forecasting because its economy has experienced persistent price pressures arising from both domestic and external factors. The country’s dependence on imported inputs and refined petroleum products, fluctuations in the exchange rate, infrastructure constraints, food supply disruptions, and changes in monetary and fiscal conditions can influence the cost of goods and services. The IMF (2022) identified monetary and exchange-rate factors as important drivers of Nigeria’s long-run inflation performance, while also noting the influence of structural impediments on domestic prices. These factors can affect inflation at different times and may cause historical relationships to change. For example, exchange-rate depreciation can raise the domestic cost of imported goods, machinery, fertilisers, and production inputs, while disruptions to agricultural production and distribution can increase food prices. The IMF (2026), in its analysis of Nigeria’s inflation developments, also described the effects of exchange-rate adjustments, fuel subsidy removal, food prices, and other supply-side pressures on recent inflation. These conditions demonstrate why forecasting Nigerian inflation requires a method capable of identifying patterns in historical observations while recognising that the underlying economic environment may change.

The period from 2005 to 2025 is significant because it covers different phases of Nigeria’s economic and monetary experience. During this period, inflation was influenced by changes in food supply, international commodity prices, global financial conditions, domestic monetary policy, and fluctuations in the exchange rate. The global financial crisis and subsequent economic adjustments created uncertainty, while the 2016 recession and the COVID-19 pandemic affected production, distribution, and business activity. More recently, the foreign-exchange reforms and removal of the petrol subsidy in 2023 contributed to major adjustments in domestic costs and prices. The IMF (2022) documented the persistence of inflation in Nigeria and the role of exchange-rate developments and monetary conditions, while the IMF (2026) described further inflationary pressures associated with the reforms and high food and energy prices during 2023–2025. Examining the period as a whole can help establish how inflation behaved across different economic circumstances and whether a model fitted to historical observations can produce useful forecasts. Nevertheless, the presence of major economic shocks means that model estimates and forecasts must be interpreted with caution, especially when historical patterns may not continue unchanged.

Forecasting inflation is also important for the Central Bank of Nigeria because price stability is a major objective of monetary policy. Monetary authorities use information about current and expected inflation when assessing policy interest rates, liquidity conditions, and other measures intended to influence aggregate demand and price developments. If inflation is expected to rise, policymakers may consider measures to limit persistent price pressures; if inflation is expected to moderate, the timing and intensity of policy adjustments may differ. Bernanke and Mishkin (1997) explain that inflation targeting places considerable importance on forward-looking policy decisions and the communication of monetary objectives. Similarly, the IMF (2022) emphasised the relevance of credible monetary policy frameworks in understanding Nigeria’s inflation performance. A reliable forecasting model cannot eliminate inflation or account for every unexpected shock, but it may help policymakers assess likely future movements using available information. Applying time series analysis to Nigerian inflation data can therefore contribute to the evidence base required for monetary policy planning, inflation monitoring, and the evaluation of alternative economic scenarios.

The need for accurate inflation forecasting is equally important to households, businesses, investors, and public-sector institutions. Households use expectations about future prices when making decisions concerning consumption, savings, and major purchases, while businesses rely on price expectations when preparing budgets, setting selling prices, negotiating contracts, and estimating production costs. Unexpected inflation may reduce the real value of savings, increase uncertainty in investment decisions, and complicate public expenditure planning. Hyndman and Athanasopoulos (2021) stress that forecast accuracy should be evaluated systematically rather than assumed from the apparent fit of a model to historical observations. In a related Nigerian study, Guobadia, Omoruyi, Izuogu, Omodio, and Chidera (2023) applied ARIMA-based methods to model and forecast inflation using historical data from 1981 to 2020, demonstrating the relevance of statistical forecasting techniques to Nigeria’s inflation experience. Ajobo et al. (2026) also examined seasonal ARIMA models for Nigerian headline, food, and core inflation, illustrating the potential value of considering different inflation measures and temporal patterns. These studies provide a foundation for further investigation into model selection and forecast performance using a defined historical period.

A further consideration in forecasting Nigerian inflation is the quality and comparability of the data used. The Consumer Price Index is periodically rebased to reflect changes in household consumption patterns and the composition of goods and services. In February 2025, Nigeria’s National Bureau of Statistics introduced a rebased CPI with December 2024 as the price reference base, replacing the previous 2009 base. The revised system updated the consumption basket and weighting structure to better reflect contemporary spending patterns (National Bureau of Statistics [NBS], 2025). Although rebasing improves the relevance of price measurement, changes in the base year, basket composition, and methodology can affect the interpretation of historical inflation rates. The NBS (2025) CPI documentation therefore provides an important source for identifying the definitions and reference periods underlying the available observations. For a study covering 2005–2025, the researcher must carefully select a consistent monthly or annual inflation series, document any methodological changes, and assess whether adjustments are necessary before model estimation. Accordingly, this study will apply appropriate time series techniques to historical Nigerian inflation data, compare suitable forecasting models, and evaluate their performance to identify a statistically defensible approach to forecasting inflation in Nigeria.

1.2 Statement of the Problem

Inflation has remained a major macroeconomic challenge in Nigeria, affecting the purchasing power of households, the cost of production, investment decisions, and the overall stability of the economy. Although inflation is monitored by the National Bureau of Statistics and the Central Bank of Nigeria, forecasting its future movement remains difficult because prices respond to several interacting economic conditions. Exchange-rate depreciation, food supply constraints, increases in energy costs, monetary developments, and changes in government policy may cause inflation to rise or fall unexpectedly. The International Monetary Fund (2022) identified monetary and exchange-rate factors as important explanations for the persistence of inflation in Nigeria, indicating that price movements cannot always be understood through a simple assumption of stability. The problem, therefore, is the difficulty of obtaining reliable forecasts that adequately reflect the historical behaviour of Nigerian inflation and provide useful information for future economic planning.

Another problem is that inflation in Nigeria has exhibited changing patterns across different periods, making it necessary to determine which statistical forecasting method is most appropriate. A model that performs satisfactorily during a period of relatively stable inflation may perform poorly when the economy experiences major shocks or substantial changes in prices. The 2016 recession, the COVID-19 pandemic, and the economic reforms introduced in 2023 created different conditions for production, distribution, exchange-rate movements, and consumer prices. The IMF (2022) documented the persistence of Nigerian inflation and the influence of monetary and exchange-rate developments, while the IMF (2026) described the subsequent effects of exchange-rate adjustments, fuel subsidy removal, and food and energy costs. These developments raise an important statistical question: which time series model can adequately capture the historical behaviour of Nigerian inflation and produce forecasts with acceptable levels of accuracy? Without systematic model comparison and forecast evaluation, selecting a forecasting technique based solely on familiarity or convenience may produce unreliable estimates.

A further problem concerns the proper application of time series techniques to Nigerian inflation data. Inflation observations may contain trends, autocorrelation, changing variance, structural breaks, or seasonal patterns, depending on the frequency and definition of the series. If these characteristics are not examined before estimation, a model may be incorrectly specified and its forecasts may be misleading. Box et al. (2015) emphasised the importance of model identification, parameter estimation, and diagnostic checking in ARIMA analysis, while Hyndman and Athanasopoulos (2021) stressed the need to evaluate forecasts using appropriate accuracy measures and validation procedures. In addition, the NBS rebased Nigeria’s CPI in 2025, updating its reference period and consumption weights. This methodological change creates a need to verify the consistency of the inflation series selected for a study covering 2005–2025 (NBS, 2025). The problem is therefore not simply to generate future inflation figures, but to establish whether the selected data and forecasting model are statistically appropriate for the Nigerian economy.

Furthermore, although studies have applied statistical methods to Nigerian inflation, there remains a practical need to evaluate forecasting performance for the specific period covered by this research. Guobadia et al. (2023) applied ARIMA-based techniques to Nigerian inflation data, while Ajobo et al. (2026) investigated seasonal ARIMA models for headline, food, and core inflation. These studies demonstrate the usefulness of time series methods but do not remove the need to assess model performance for the particular dataset, time span, and forecast horizon selected for the present study. The central problem addressed by this research is therefore the need to apply appropriate time series techniques to Nigerian inflation data from 2005 to 2025, identify a suitable forecasting model, and evaluate its predictive accuracy using a systematic statistical procedure. The findings are expected to provide evidence on the usefulness and limitations of the selected forecasting approach, rather than assume that one model will accurately predict all future economic shocks.

1.3 Objectives of the Study

The main objective of this study is to apply time series analysis in forecasting inflation rates in Nigeria from 2005 to 2025.

The specific objectives are to:

1. Examine the trend and behaviour of inflation rates in Nigeria from 2005 to 2025.

2. Determine the time series characteristics of Nigerian inflation data, including stationarity, autocorrelation, and relevant trend or seasonal patterns.

3. Develop and estimate appropriate time series models for forecasting inflation rates in Nigeria.

4. Evaluate and compare the forecasting accuracy of the selected models to identify the most suitable model for the study period.

1.4 Research Questions

The study will be guided by the following research questions:

1. What were the trend and behaviour of inflation rates in Nigeria from 2005 to 2025?

2. What time series characteristics, including stationarity, autocorrelation, and relevant trend or seasonal patterns, are present in Nigerian inflation data?

3. Which time series models are appropriate for forecasting inflation rates in Nigeria?

4. Which of the selected time series models provides the most accurate forecasts of inflation rates in Nigeria?

1.5 Research Hypothesis

The following null hypothesis will guide the comparative evaluation of the forecasting models at the 5% level of significance.

H₀: There is no statistically significant difference in the forecasting accuracy of the selected time series models for inflation rates in Nigeria.

The hypothesis will be tested using an appropriate forecast-comparison procedure, depending on the models estimated, the forecast horizon, and the availability of out-of-sample forecast errors. A suitable test may be the Diebold–Mariano test, subject to its assumptions and the adequacy of the forecast evaluation sample (Diebold & Mariano, 1995).

1.6 Significance of the Study

This study is expected to be significant to monetary authorities, government policymakers, businesses, investors, researchers, and other stakeholders who depend on information about future inflation.

Central Bank of Nigeria: The study may provide empirical evidence on the usefulness of time series models in forecasting Nigerian inflation. The findings may support inflation monitoring, monetary policy analysis, and the assessment of future price movements. Bernanke and Mishkin (1997) emphasised the importance of forward-looking information in monetary policy frameworks, making reliable forecasting an important consideration for price stability.

Government and economic policymakers: The findings may assist policymakers in understanding historical inflation patterns and evaluating the usefulness of statistical forecasts in economic planning. The IMF (2022) identified monetary and exchange-rate factors as important influences on Nigeria’s inflation performance, demonstrating the relevance of reliable economic information to policy formulation.

Businesses and investors: The study may help businesses and investors appreciate how historical inflation data can be used to develop forecasts for budgeting, pricing, procurement, and investment planning. Hyndman and Athanasopoulos (2021) explain that systematic forecasting and forecast evaluation can improve decision-making where future conditions are uncertain.

Households and consumers: Although the study will not directly determine future living costs, its findings may improve understanding of inflation trends and the usefulness of statistical forecasts. Better-informed inflation expectations may help households and consumers consider how changing prices affect purchasing power and expenditure planning.

Researchers and students: The study may serve as a reference for future research in statistics, economics, econometrics, and applied forecasting. It may also provide a basis for comparing ARIMA, exponential smoothing, and other appropriate forecasting methods using Nigerian economic data. Box et al. (2015) provide the methodological foundation for ARIMA modelling, while Diebold and Mariano (1995) provide a framework for comparing predictive accuracy.

National Bureau of Statistics and other statistical institutions: The study may demonstrate the importance of maintaining consistent historical data and documenting methodological changes in price measurement. The NBS’s 2025 CPI rebasing highlights the need to understand the reference periods and methods underlying inflation statistics when conducting long-term forecasting studies.

1.7 Scope of the Study

The study will focus on the application of time series analysis to forecasting inflation rates in Nigeria over the period 2005–2025. The geographical scope will be Nigeria, and the study will rely on secondary data obtained from credible statistical and economic institutions.

The principal variable will be the Nigerian inflation rate, measured through an appropriately selected Consumer Price Index-based inflation series. The researcher will determine whether monthly or annual observations are most suitable for the analysis based on data availability, consistency, and the forecasting objectives.

The methodological scope will include descriptive statistical analysis, graphical examination of inflation trends, stationarity testing, model identification, parameter estimation, diagnostic checking, and forecast evaluation. Potential methods may include ARIMA, seasonal ARIMA where appropriate, and exponential smoothing. The final models will be selected according to the characteristics of the data rather than predetermined assumptions.

The study will evaluate the predictive performance of the selected models using suitable measures such as mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE), where appropriate. Forecast comparison will be based on an out-of-sample evaluation procedure to reduce the risk of judging model performance solely by how well the models fit historical observations.

1.8 Limitations of the Study

One possible limitation is the availability of a consistent inflation series covering the entire period from 2005 to 2025. Nigeria’s CPI rebasing in 2025 updated the price reference base and consumption weights, creating a need to verify the comparability of historical observations before estimation (NBS, 2025).

Another limitation is that time series models rely substantially on historical patterns. Unexpected developments, such as major exchange-rate adjustments, fuel price changes, supply disruptions, or policy changes, may cause actual inflation to differ considerably from forecast values. Box et al. (2015) emphasise that appropriate model specification and diagnostic checking are essential, but these procedures cannot eliminate uncertainty associated with future economic shocks.

The study may also be limited by the forecasting horizon and the number of available observations. Annual data provide fewer observations than monthly data over the same period, which can limit the estimation of complex models. Monthly observations may provide more information about short-term movements, but they require careful consideration of seasonality and changes in measurement.

Finally, no single accuracy measure is sufficient in every situation. For example, MAPE may become problematic when actual values are zero or very close to zero, while RMSE gives greater weight to large forecast errors. Hyndman and Athanasopoulos (2021) recommend selecting accuracy measures that are suitable for the data and the purpose of forecasting. The study will therefore interpret the results in light of the selected evaluation methods and acknowledge the limitations of the preferred model.

1.9 Operational Definition of Terms

Inflation: A sustained increase in the general price level of goods and services over time, resulting in a decline in the purchasing power of money.

Inflation rate: The percentage change in a selected price index over a specified period. In this study, it refers to the Nigerian inflation series selected for statistical analysis.

Time series analysis: A statistical method used to examine observations recorded sequentially over time to identify patterns and support forecasting.

Forecasting: The process of estimating future values of a variable using historical observations and an appropriate statistical model.

Consumer Price Index (CPI): An index that measures changes in the prices of a representative basket of goods and services consumed by households.

ARIMA model: An autoregressive integrated moving average model that uses autoregressive, differencing, and moving-average components to represent time-dependent data.

Seasonal ARIMA (SARIMA): An extension of ARIMA that incorporates seasonal patterns when such patterns are present in the data.

Stationarity: A statistical property in which the relevant characteristics of a time series, such as its mean and variance, remain stable over time under the specified definition.

Autocorrelation: The correlation between observations of a time series and its own past values at different time lags.

Forecast accuracy: The degree to which predicted values correspond to actual observed values, assessed using appropriate error measures.

Mean Absolute Error (MAE): The average of the absolute differences between actual observations and forecast values.

Root Mean Squared Error (RMSE): The square root of the average squared differences between actual observations and forecast values.

Model selection: The process of identifying the most appropriate forecasting model based on statistical properties, diagnostic results, and forecast performance.

Project – Application of Time Series Analysis in Forecasting Inflation Rates in Nigeria (2005 to 2025)