ARIMA and GARCH

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/329315962
MODELING AND FORECASTING INFLATION IN KENYA: RECENT INSIGHTS

FROM ARIMA AND GARCH ANALYSIS
Article · November 2018
CITATIONS READS
149 2,730
1 author:
Thabani Nyoni
University of Zimbabwe
357 PUBLICATIONS 853 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Neuromarketing Mix in the 21st Century View project
THE IMPACT OF MACROECONOMIC VARIABLES ON STOCK MARKET DEVELOPMENT IN ZIMBABWE (1990-2018) View project
All content following this page was uploaded by Thabani Nyoni on 30 November 2018.
The user has requested enhancement of the downloaded file.

Vol-5, Issue-6 November-December 2018 eISSN: 2394-9163
MODELING AND FORECASTING INFLATION IN KENYA: RECENT

INSIGHTS FROM ARIMA AND GARCH ANALYSIS
Thabani Nyoni
Department of Economics
University of Zimbabwe, Harare, Zimbabwe
Abstract
Using annual time series data, we study inflation in Kenya over the
period 1960 – 2017 using both ARIMA and GARCH approaches. Based on
both minimum AIC and Theil’s U, the study presents the ARIMA (2, 2, 1)
model, the ARIMA (1, 2, 0) model and the AR (1) – GARCH (1, 1) model.
Our diagnostic tests indicate that the CPI series is I (2), the residuals of
the ARIMA (2, 2, 1) and ARIMA (1, 2, 0) models are stationary and that
both the ARIMA (2, 2, 1) and the ARIMA (1, 2, 0) models are stable. The
results of this study indicate that annual inflation in Kenya is likely to
continue rising. The policy implications emanating from this study are 3 –
fold and are envisioned to assist policy makers in restoring and
maintaining price stability in Kenya.
Key words: ARIMA, Consumer Price Index (CPI), Forecasting, GARCH,

Inflation, Kenya
0.0 Introduction
The practice of forecasting inflation has generally been considered an important

input in monetary policy making (Fisher et al, 2002). Nothing is more important
to the conduct of monetary policy than understanding and predicting inflation
(Kohn, 2005).It is now well understood that expected (future) inflation is
important for the design and implementation of monetary policy by central
banks (Huang, 2012). In fact, inflation forecasting can be considered a
comparative advantage of a central bank as it maintains information advantage
about the state of the economy over the public (Huseynov et al, 2014).In Kenya,
the second half of 2017 was characterized by general macroeconomic stability,
decline in food prices, and uncertainties with regard to the prolonged election
period. The Central Bank of Kenya (CBK) conducted monetary policy with the
aim of keeping overall inflation within the government target range of 2.5 and
7.5% (CBK, 2017). The priority of price stability over the other policy goals
seems to be politically accepted in most countries, if not appropriately
mentioned in the laws governing the central bank (Gallego, 2002). Price stability
remains the primary objective of monetary policy formulation and
implementation (CBK, 2017). Kenya’s Vision 20301 cannot be achieved without
reliable inflation forecasts.Motivated by the need to provide more insights into
1
Kenya Vision 2030 is a long – term socio – economic blueprint whose main
objective is to transform Kenya into a newly industrializing, middle – income
country. If this main objective is to be achieved, then we cannot rule out the
need for policy makers to be well equipped with reliable inflation forecasts.Policy
makers, as noted by Nyoni (2018k); can get prior indication about possible
future inflation through inflation forecasting.
DIMORIANREVIEWwww.dimorianreview.comA PEER REVIEWED MULTIDISCIPLINARY INTERNATIONAL INDEXED E-JOURNAL
16
the dynamics of inflation in Kenya, in this research; we forecast Kenya’s inflation

over the next decade.
A few studies have modeled and forecasted inflation in Kenya, for example;
Uwilingiyimana et al (2015), Lidiema (2017) and Fwaga et al (2017). These
studies have either employed ARIMAor GARCH approaches, separately; no study
has attempted to rely on more than one modeling technique. This paper, whose
main objective is to model and forecast inflation in Kenya; is quite different from
these existing studies in the sense that I rely on both ARIMA and GARCH
analysis in trying to model and forecast inflation in Kenya. The results of the
study are envisioned to assist policy makers in maintaining general
macroeconomic stability in Kenya which will in turn play a pivotal role in the
achievement of the much touted Kenya’s Vision 2030. The rest of the paper is
organized as follows: literature review, materials & methods, results and
conclusion & policy prescriptions; in chronological order.
1.1 Literature Review
The major goals of economic policies include high unemployment; and stable
and increased growth (Mutwiri, 2017). Although there is no common concession
that all these goals are compatible, there is agreement on the roles various
instruments can and should play in aiding the realization of these goals
(Friedman, 1968).It is almost needless to mention that in most economies (both
developing and developed),bothmonetary and fiscal policies compliment each
other in trying to achieve the desired economic goals. While various theories of
inflation, basically do not agree on their views on inflation, it is imperative to
note that both fiscal and monetary policies play a pivotal role in the economy.
Below is a brief overview ofthe twomost celebrated and widely used theories of
inflation nowadays:
1.1 Theoretical Literature Review
1.1.1. Keynesian Theory of Inflation
Having been developed by the great economist, John Maynard Keynes; the
Keynesian theory argues that increase in savings will not lead to lower interest
rates, as long as the economy suffers under – employment. The Keynesian
theory avers that an increase in general price level (inflation) is caused by an
increase in aggregate demand (well, above aggregate supply). Keynesians
believe that if the economy is at full employment output level, an increase in
government expenditure, a rise in private consumption and a rise in private
investment would result in a rise aggregate demand; thereby causing inflation.
Keynes himself suggested that governments should play a more active role in
the economy – he advocated for government intervention in order to stabilize
the economy.
1.1.2 Monetarist Theory of Inflation
The father of monetarism is the great economist, Milton Friedman, a Classical

economist by origin; who challenged the Keynesian theory by arguing that
government intervention would destabilize the economy. Monetarists strongly
challenge the Keynesian view that government spending stimulates national
output. Monetarists assume a crowding – out effect of government spending on
private investment, especially if the latter is deficit – financed. The monetarist
school of thought believes that the major cause of inflation is monetary
17
mismanagement. In fact, Friedman (1967) is well known for his popular

argument that inflation is and everywhere a monetary phenomenon. Monetarists
advocate for the use of fixed money growth rate rules in order to ensure
monetary stability in the economy.
1.2 Empirical Literature Review
The empirical literature on inflation forecasting points to different techniques in

forecasting inflation (Zivko & Bosnjak, 2017).Many studies, for example; Libert
(1983), Hill & Fildes (1984), Paulos et al (1987),Texter & Ord (1989), Bokhari &
Feridun (2006), Etuk et al (2012) as well as Kelikume & Salami (2014), amongst
others; support the strength of ARIMA as a modeling technique in forecasting
inflation. On the other hand, there is also a good number of researchers who
argue that the GARCH approach to modeling and forecasting inflation is the best,
for example; Bollerslev (1986), Caporale et al (2003), Verbeek (2004), Nor et al
(2007), Zivot (2008), Brooks (2008), Wang (2009), Kozhan (2010),
Osarumwense & Waziri (2013) and Moroke & Luthuli (2015) amongst others. In
this study, I employ both ARIMA and GARCH approaches in order to obtain a
rigorous analysis of inflation in Kenya.
Employing the simplest form of the GARCH model, the GARCH (1, 1) model;
Banerjee (2017) analyzed inflation in 41 countries over the period January 1958
to February 2016 and found out that in the long – run, conditional volatility of
inflation is 3.5 times greater in developing countries compared to advanced
countries. Relying on a simple GARCH (1, 1) model, Sek & Har (2012) analyzed
inflation dynamics in three Asian countries, namely Korea, Philippines and
Thailand over the periods, September 1985 to June 2010, January 1985 to June
2010 and January 1987 to June 2010 respectively; and found out that there is
lower volatility of inflation and also persistence of inflation declines in the
countries under study.In Croatia, Zivko & Bosnjak (2017), analyzed inflation
using monthly inflation data from January 1997 to November 2015; employing
ARIMA and ARCH family models and they found out that the ARCH (1) model is
the best model for explaining CPI volatility in Croatia while the ARIMA (0, 1,
1)(0, 1, 1)12 is reliable in forecasting CPI in Croatia.In Pakistan, Saleem (2008);
looked at inflation dynamics using data over the period January 1990 to May
2007, and employed VAR, ARCH, GARCH and EGARCH methodologies and found
out that inflation is indeed volatile in nature and the time effect model is
significant.
Closer to home, in Africa; Habimana et al (2016) modeled and forecasted

Consumer Price Index in Rwanda using monthly data from February 1995 to
December 2015, and employed the ARIMA approach; and found out that the
ARIMA (4, 1, 6) model was the best – fit model. Benedict (2013) studied
inflation in Ghana using data over the period January 1965 to December 2012
and employed various ARCH family type models and found out that the EGARCH
(2, 1) model is superior in performance.In Nigeria, Kuhe & Egemba (2016);
studied inflation using annual time series data from 1950 to 2014; and made use
of the ARIMA technique, and found out that ARIMA (3, 1, 0) was the optimal
model. Molebatsi & Raboloko (2016) investigated inflation in Botswana using
monthly CPI data from January 2005 to December 2014, employing ARIMA and
ARCH type models and found out that volatility for Botswana’s CPI is low. In
Zimbabwe, Nyoni (2018k) modeled and forecasted inflation using monthly data
from July 2009 to July 2018 and employed a GARCH approach, and found out

18
that the behaviour of inflation over the study period follows an AR (1) – IGARCH
(1, 1) process.
Back to Kenya, Uwilingiyimana et al (2015) studied inflation using ARIMA –

GARCH models, and employed monthly data from January 2000 to December
2014 and found out that the ARIMA (1, 1, 12) – GARCH (1, 2) model provide the
optimal results. More recently, Lidiema (2017), modeled and forecasted inflation
rate in Kenya and employed, SARIMA and Holt – Winters Triple Exponential
Smoothing models; using monthly data from November 2011 to October 2016
and found out that the SARIMA model was a better model than the Holt –
Winters Triple Exponential Smoothing model. Both studies, that is;
Uwilingiyimana et al (2015) and Lidiema (2017) point to the fact that monthly
inflation rate in Kenya is likely to continue (into the near future, the short – run)
on an upward trend. In another Kenyan recent study, Fwaga et al (2017),
studied inflation using data over the period January 1990 to December 2015 and
employed EGARCH and GARCH approaches, and found out that the EGARCH (1,
1) model is the best model for forecasting Kenyan inflation data.
From the empirical literature review, it is quite clear that most authors either use
ARIMA or ARCH family type models to model and forecast inflation. While my
basic methodological approach may be very similar to the one used by
Uwilingiyimana et al (2015), it is quite unique and still different in the sense that
I employ both ARIMA and GARCH approaches separately and then compare both
their parsimony (using the AIC values) and forecast accuracy (using the Theil’s
U).
2 Objective
Through this research article, an attempt is made to study the rend of inflation
in Kenya over the period 1960 – 2017.
3 Methodology
3.1 Model building & estimation procedures
3.1.1 The Moving Average (MA) model
Given:
= + +…+ ………………………………………………….…….……………………. [1]
where μt is a purely random process with mean zero and varience σ2. We say
that equation [1] is a Moving Average (MA) process of order q, commonly
denoted as MA (q). CPI is the Consumer Price Index in Kenya at time t, ɑ0 … ɑq
are estimation parameters, μt is the current error term while μt-1 … μt-q are
previous error terms. Hence:
= + ……………………………………………………….………..……………………………. [2]
is an MA process of order one, commonly denoted as MA (1). Owing to the fact

that previous error terms are unobserved variables, we then scale them such
that ɑ0=1. Since:
E( )=0 ……………………………………..……………………………….…………………………………….. [3]
for all t, it therefore; implies that:

19
E( )=0 …………………………………………………………………………………………………………… [4]
and:
Var( ) ( )σ2 ………………..…………………………………….…………………………………. [5]
where μt is independent with a common varience σ2. Hence, we can now re –

specify equation [1] as follows:
= + +…+ ……………………………………………….…………………………………. [6]
Equation [6] can be re – written as:
CPIt= + ……………………………………………………………….……….………………….. [7]
We can also write equation [7] as follows:
= + ……………………...…………………………………………..……………………….. [8]
where L is the lag operator.
or as:
=ɑ(L) …………………...……………………………………………………………..…………………… [9]
where:
ɑ(L)=θ(L)2 ………………………………………………………………………….………….……………….. [10]
3.1.2 The Autoregressive (AR) model
Given:
CPIt=β1CPIt-1+…+βpCPIt-p+μt …………………………………..…………….………………………. [11]
Where β1 … βp are estimation parameters, CPIt-1 … CPIt-p are previous period

values of the CPI series and μt is as previously defined. We say that equation
[11] is an Autoregressive (AR)process of order p, and is commonly denoted as
AR (p); and can also be written as:
CPIt= +μt ………………………………………...………………………..…………………… [12]
Equation [12] can be re – written as:
CPIt= +μt ……………………………...………………………………………………………… [13]
or as:
β(L)CPIt=μt …………………………………………………………………………………………………….. [14]
where:
β(L)=ɸ(L)3 ………………………………………………………………………………………………………… [15]
2
Defined in equation [22].
3
Defined in equation [23].
20
or as:
CPIt=(β1L+…+βpLp)CPIt+μt …………………………………………………………….……………… [16]
Thus:
CPIt=(β1L)CPIt+μt ……………………………………….………………………………….……………… [17]
is an AR process of order one, commonly denoted as AR (1).
3.1.3 The Autoregressive Moving Average (ARMA) model
As initially propounded by Box & Jenkins (1970), an ARMA (p, q) process is

simply a combination of AR (p) and MA (q) processes. Therefore, combining
equations [1] and [11]; an ARMA (p, q) can be specified as follows:
CPIt=β1CPIt-1+…+βpCPIt-p+μt+ɑ1μt-1+…+ɑqμt-q …...………………………………………… [18]
or as:
CPIt= + + ………………………………………………………..…………….. [19]
by combining equations [7] and [12].Equation [18] can also be written as:
ɸ(L)CPIt=θ(L)μt ……………………………..……………………………………………………………….. [20]
whereɸ(L) and θ(L) are polynomials of orders p and q respectively, simply

defined as:
ɸ(L)=1-β1L … βpLp ……………………..……………………………………………………………………. [21]
θ(L)=1+ɑ1L+…+ɑqLq ……………………………………………………………………..………………… [22]
A this juncture, it is quite essential to note that the ARMA (p, q) model, just like
the AR (p) and the MA (q) models; can only be employed for stationary time
series data; and yet in real life, many time series are non – stationary. For this
simple reason, ARMA models are not for describing non – stationary time series.
3.1.4 The Autoregressive Integrated Moving Average (ARIMA) model
ARIMA models are a set of models that describe the process (for example, CPIt)
as a function of its own lags and white noise process (Box & Jenkins, 1974).
Making predicting in time series using univariate approach is best done by
employing the ARIMA models (Alnaa & Ahiakpor, 2011). A stochastic process
CPIt is referred to as an Autoregressive Integrated Moving Average (ARIMA) [p,
d, q] process if it is integrated of order “d” [I (d)] and the “d” times differenced
process has an ARMA (p, q) representation. If the sequence ∆dCPIt satisfies and
ARMA (p, q) process; then the sequence of CPIt also satisfies the ARIMA (p, d, q)
process such that:
∆dCPIt= + +μt ……………………….…………………….………………. [23]
which we can also re – write as:
∆dCPIt= + +μt ………………..……………………………………………… [24]
where ∆ is the difference operator, vector β ϵⱤp and ɑϵⱤq.

21
3.1.5The Autoregressive Conditionally Heteroskedastic (ARCH) model
In financial time series, it usually makes sense to take into account a model that
describes how the varience of the errors evolves and such a model is non –
other – than the ARCH model. The intuition behind ARCH family type models is
that in reality, especially in financial time series; it is rare that the varience of
the errors will be constant over time and for that reason, it is quite noble to
consider models that do not assume that the varience is constant. To expagorate
the simple intuition behind the ARCH model, I begin by defining the conditional
varience of a random variable, :
=var(μt│μt-1, μt-2, …)=E[μt-E(μt)2│μt-1, μt-2, …] …………………….….……………….. [25]
assuming thatequation [3] also holds water in this case, such that:
=var(μt│μt-1, μt-2, …)=E[ │ , …] …………………………...…………………………….. [26]
Equation [26] implies that the conditional varience of a zero mean normally
distributed random variable μt is equal to the conditional expected value of the
square of μt.
= + ………………………………………………………………………………………………….. [27]
Equation [27] is referred to as an ARCH (1) model since the conditional varience
depends only on one lagged squared error. This equation cannot be seen as a
complete model just because we haven’t taken into account the conditional
mean. The conditional mean, in this case; describes how the dependent variable,
CPIt; varies over time. As noted by Nyoni (2018k); there is no rule of thumb on
how to specify the conditional mean equation; actually it takes any form deemed
adequate by the researcher. Therefore, the complete model consists of both the
conditional mean equation and the ARCH specification as illustrated by Nyoni
(2018k). Equation [27] can be generalized to case where the error variance
depends on p lags of squared errors as follows:
= + +…+ …………................................................…………….. [28]
Hence, equation [28] is an ARCH (p) model.
3.1.6 The Generalized ARCH (GARCH) model
The equation below:
= + +
…………...................................................................................…………….. [29]
is the simplest and yet most important case of a GARCH process, the GARCH (1,
1) model; where is the conditional varience, is the constant, is the
information about the previous period volatility, and is the fitted varience
from the model during the previous period. From equation [29], we know that:
Et-1[ ]= ………………………………………………………………………………..……………………… [30]
such that:

22
= +( + ) + - …………………………..…………………………………………….. [31]
which is clearly an ARMA (1, 1) model; this clearly implies that indeed, a GARCH
model can be expressed as an ARMA process of squared residuals. In this case:
= - [ ] …………………………………………………..……………………………………………… [32]
is the stochastic term. Given equation [31], we can deduce that the stationarity
of the GARCH (1, 1) model requires:
+ ˂1 …………………………………………………………………………………………………………… [33]
Taking the unconditional expectation of equation [29], we get:
= + + …………………………………………………………….…..………………………… [34]
so that:
= ………………………………………………………………………………………………………… [35]
For this unconditional varience to exist, equation [33] must be correct and for it
to be positive, then:
˃0 ……………………………………………….……………………………………….………………………. [36]
Equation [29] can be generalized into a GARCH (p, q) model where the current
conditional varience is parameterized to depend upon p lags of the squared error
and q lags of the conditional varience as follows:
= + +…+ + +…+ ……………………………………….…………… [37]
Equation [37] can be written as follows:
= +φ(L) +λ(L) …………………………………...………………………………………………. [38]
where φ(L) and λ(L) denote the AR and MA polynomials respectively, with:
φ(L)= L+…+ ……………………………………...……………………….……………………….. [39]
and:
λ(L)= +…+ …………………………………………………………………………………..………….[40]
or as:
= + + ………………………………...……………………………………….. [41]
where condition [33] is now generalized as:
+ ˂1 ………………………………………………...……………………………………………. [42]
If all the roots of the polynomial:
│1-λ(L)│-1 =0 ………………………………………………………………..………………………………. [43]

23
lie outside of the unit circle, we get:
= │1-λ(L)│-1+φ(L)│1-λ(L)│-1 ………………………………….………………….………… [44]
which is actually an ARCH ( ) process since the conditional varience linearly

depends on all previous squared residuals. Thus the unconditional varience is
expressed as:
E( )= ………………………………………………………………………..………… [45]
In the event that:
+…+ + +…+ =1 ………………………………………………………….………………………. [46]
then the unconditional varience will be .
Conditions [33] and [42] intuitively imply the same thing. In many financial time
series, these conditions are close to unity; implying persistant volatility. If:
+ =1 …………………………………………..…………………………………………………………….. [47]
or more generally:
+ =1 ……………………………………………………………….…..……………………….. [48]
or simply:
φ(L)+λ(L)=1 …………………………………………………………………………….……..……………… [49]
it implies that the resulting process is not covariance stationary. Hence the term
Integrated GARCH or IGARCH model, pointing to the fact that current
information remains vital when forecasting the volatility for all horizons.
3.2 Model Specification
Guided by our diagnostic tests (see tables 1 – 10), we specify the following
models:
3.2.1 ARIMA (2, 2, 1) model:
……………………………………………………. [50]
3.2.2 ARIMA (1, 2, 0)
……………………………………….………………………………. [51]
3.2.3 AR (1) – GARCH (1, 1) model:
The appropriate equations for the mean and varience were specified as follows:
= + + ; N(0; ) ………………………………………………….………………… [52]
= + + …………………………………………….………………………………………… [53]

24
where:
0 ……………………………………………………………….……………………………………..…….[54]
3.3 The Box – Jenkins Methodology4
The first step towards model selection is to difference the series in order to
achieve stationarity. Once this process is over, the researcher will then examine
the correlogram in order to decide on the appropriate orders of the AR and MA
components. It is important to highlight the fact that this procedure (of choosing
the AR and MA components) is biased towards the use of personal judgement
because there are no clear – cut rules on how to decide on the appropriate AR
and MA components. Therefore, experience plays a pivotal role in this regard.
The next step is the estimation of the tentative model, after which diagnostic
testing shall follow. Diagnostic checking is usually done by generating the set of
residuals and testing whether they satisfy the characteristics of awhite noise
process. If not, there would be need for model re – specification and repetition
of the same process; this time from the second stage. The process may go on
and on until an appropriate model is identified (Nyoni, 2018i).
3.4Definition & Measurement of Inflation
There is no clear consensus on the definition and measurement of inflation

although there is a general agreement on the need to control inflation. Haberler
(1960) defines inflation as a condition of rising prices. According to Samuelson &
Nordhaus (1995), inflation refers to a rise in general level of prices. Inflation, as
defined by Stanford (2008); is a process whereby the average price level in an
economy increases over time.Jalil (2011) defines inflation as a progressive rise
in price level, usually over a period of time. Many previous studies, for example;
Wynne & Sigalla (1994), Saleem (2008), Stanford (2008), Bullard (2011),
Benedict (2013), Jacobs et al (2015), Uwilingiyimana et al (2015), Kuhe &
Egemba (2016), Molebatsi & Raboloko (2016), Habimana et al (2016), Lidiema
(2017), Zivko &Bosnjak (2017) and Fwaga et al (2017) amongst others;
generally agree on the measurement of inflation using the Consumer Price Index
(CPI). The CPI5, as defined by Stanford (2008); is a measure of the overall price
4
This technique was developed by Box & Jenkins (1970; 1974). For a more
simplified (and yet detailed) explanation of this technique, see Maddala (1992).
5
Retail price information is gathered on each type of product, and then weighted
according to its importance in overall consumer spending, to construct the CPI.
Monthly or annual changes in the CPI provide a good measure of the rate of
consumer price inflation (Stanford, 2008).
However, measuring CPI inflation is a tricky and at times controversial task.
Some economists believe that CPI measures actually overstate the true value of
inflation, because they do not take into account improvements in the quality of
goods and services which may offset, in some cases, some of the rise in their
prices. Other economists believe that CPI inflation may understate true inflation,
due to inadequate allowances for changes in the cost of housing and other hard
– to – estimate prices (Stanford, 2008).
25
level paid by consumers for the various goods and services they
purchase.According to Jacobs et al (2014), CPI is a measure of the price of a
fixed basket of goods and services, with unchanged composition and quality
overtime.Habimana et al (2016) defines CPI as a measure of the average change
over time in the price of consumer items, goods and services that households
buy for day to day living. This study relies on inflation that was measured in
terms of CPI.
3.5 Data Collection
An extensive times series data, as noted by Wabomba et al (2016); is required

for univariate time series forecasting. In fact, authors such as Chatfield (1996)
and Meyler et al (1998) argue that more than 50 observations are recommended
in order to build a reliable ARIMA model. This study is based on 57 observations
of annual Consumer Price Index (CPI) in Kenya.All the data used in this study
was gathered from the World Bank.
3.6 Diagnostic Tests & Model Evaluation
Stationarity Tests: Graphical Analysis
We can examine whether a series is stationary or not by analyzing the time plot
of that particular series. In this regard, a time plot of the CPI series is shown
below:
Figure 1
The above graph shows that CPI is not stationary since it is trending upwards
over the period under study. The implication is that the mean of CPI is changing
over time and hence we can safely conclude that the variance of CPI is not
constant over time.

26
The Correlogram (in Levels)
Figure 2
The correlogram above confirms our analysis from the observation of the time
series plot of CPI. The autocorrelation coefficients are very high particulary for
the first 6 lags and is very typical in non – stationary time series data.
The ADF Test
The Augmented Dickey Fuller (ADF test) was used to check the stationarity of
the CPI series. The general ADF test is done by running the following regression
equation:
CPIt= ct γCPIt-1+ CPIt-i+μt ……………………………………………………….…...……….. [55]
Where ct is a deterministic function of the time index t and ∆CPIj=CPIj-CPIj-1 is

the differenced series of CPIt. The null hypothesis H0: γ=1 is tested against the
alternative hypothesis Ha: γ≤1 using equation [] above. If the null hypothesis is
rejected, then the time series is stationary. The results of the ADF tests done in
this study are shown below:
Table 1 : Levels- intercept

Variable ADF Statistic Probability Critical Values Conclusion
CPI 11.09334 1.0000 -3.550396 @1% Not stationary
-2.913549 @5% Not stationary
Table 2: Levels- trend & intercept

27
Table 3: Levels- without intercept and trend & intercept

The Correlogram (at First Differences)
For a purely white noise process, the autocorrelations at various lags hover
around zero (Gujarati, 2004).
Figure 3
Table 4: 1st Difference: Intercept

Table 5 : 1st Difference- trend & intercept

CPI -1.921025 0.6299 -4.137279 @1% Not stationary
Table 6: 1st Difference- without trend and trend & intercept

28
Tables 1 – 6 as well as figures 1 – 3, indicate that the CPI series is not

stationary at levels. Hence the need to check stationarity at second differences.
The Correlogram (in 2nd Differences)
Figure 4
Table 7: 2nd Difference- Intercept

CPI -11.47370 0.0000 -3.557472 @1% Stationary
-2.916566 @5% Stationary
Table 8: 2nd Difference- trend & intercept

CPI -11.56959 0.0000 -4.137279 @1% Stationary
Table 9: 2nd Difference- without trend and trend & intercept

CPI -11.28174 0.0000 -2.608490 @1% Stationary
While figure 4 doesn’t make it clear whether our series is now stationary or not,
tables 7 – 9 confirm that the CPI series became stationary after taking second
differences.

29
Testing for ARCH / GARCH effects
In this study, ARCH / GARCH effects were tested using the Langrange Multiplier
(LM) test as briefly described here: run the mean equation given by equation
[52] and save the residuals. Square the residuals and regress then on “p” own
lags to test for ARCH effects of order “p”. From this procedure, obtain and
save it. The test statistic, T (number of observations multiplied by ) follows a
distribution and the null and alternative hypotheses are:
In this research paper, the ARCH / GARCH effects test was done and the results
are shown below:
Chi – square (2) = 100. 904 [0.00]
The p – value of [0.00] indicates a significance of this LM test result at 1% level

of significance. This implies that there are (G) ARCH effects in the chosen model
and therefore it is appropriate to estimate a GARCH model
3.7 Evaluation of Various ARIMA & GARCH Models
3.7.1 Evaluation of various ARIMA models
While there are a number of model evaluation criterion in time series modeling
and forecasting, for example; Mean Error (ME), Root Mean Square Error (RMSE),
Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE); this
study will only be restricted to the most commonly used and highly celebrated
criterion, that is; the Akaike ‘s Information Criteria (AIC) and the Theil’s U in
order to select the best models (in terms of parsimony [AIC] and forecast
accuracy [Theil’s U]) to be finally presented in this study. A model with a lower
AIC value is better than the one with a higher AIC value. Theil’s U, as noted by
Nyoni (2018l); must lie between 0 and 1, of which the closer it is to 0, the better
the forecast method.
Table 10: Evaluation of various ARIMA models

Model AIC Theil’s U ME RMSE MAE MAPE
ARIMA (1, 2, 1) 265.8063 0.5923 0.51156 2.4495 1.2297 5.1141
ARIMA (1, 2, 2) 264.1763 0.58476 0.51031 2.369 1.1526 4.9281
ARIMA (1, 2, 3) 260.1191 0.58837 0.33075 2.238 1.116 5.0169
ARIMA (1, 2, 4) 260.3259 0.60023 0.42781 2.1997 1.1164 5.1493
ARIMA (1, 2, 5) 261.9480 0.60165 0.42759 2.1912 1.0966 5.1691
ARIMA (2, 2, 1) 258.5388 0.5945 0.42946 2.248 1.1141 5.0197
ARIMA (3, 2, 1) 260.5279 0.59479 0.43135 2.2477 1.1125 5.0207
ARIMA (4, 2, 1) 261.6730 0.60099 0.44474 2.2291 1.1085 5.0739
ARIMA (5, 2, 1) 262.7219 0.59832 0.37751 2.2083 1.0868 5.0917
ARIMA (1, 2, 0) 275.2863 0.55421 0.29024 2.7233 1.3528 4.7257
ARIMA (0, 2, 1) 263.9039 0.58942 0.5175 2.4521 1.2209 5.0812
ARIMA (0, 2, 0) - 0.56554 0.22679 2.9499 1.4518 4.8306
ARIMA (2, 2, 0) 256.5819 0.59144 0.41781 2.2489 1.1189 4.9939
ARIMA (3, 2, 0) 258.5478 0.59382 0.42675 2.2482 1.1158 5.0146
ARIMA (4, 2, 0) 260.2873 0.59866 0.45022 2.2424 1.1067 5.0484
ARIMA (5, 2, 0) 260.7676 0.59793 0.3858 2.2094 1.0903 5.0881
ARIMA (0, 2, 2) 265.4161 0.60387 0.47397 2.439 1.2497 5.1993

30
ARIMA (0, 2, 3) 259.4948 0.58049 0.30278 2.2664 1.1407 5.005

ARIMA (0, 2, 4) 258.7471 0.60226 0.404 2.2079 1.1186 5.136
ARIMA (0, 2, 5) 260.1093 0.59959 0.43166 2.1951 1.1101 5.1612
ARIMA (2, 2, 2) 260.2640 0.59565 0.44203 2.2419 1.111 5.0427
ARIMA (3, 2, 3) 262.5562 0.59869 0.36077 2.2041 1.1034 5.1375
As shown in the table above, the ARIMA (2, 2, 1) model has the lowest AIC
value whilst the ARIMA (1, 2, 0) model has the lowest Theil’s U.
3.7.2 Evaluation of Various GARCH models
Table 11: Evaluation of Various GARCH models
Model AIC Theil’s U ME RMSE MAE MAPE

GARCH AR (1) 167.9380 0.69312 0.76517 2.452 1.1916 5.7734
(1, 1)
GARCH AR (1) 213.3063 1.2137 4.5739 7.6782 4.5838 12.016
(0, 1)
GARCH AR (1) 169.5461 0.73338 0.82368 2.5328 1.2561 6.1782
(1, 1) AR (2)
3.8 Residual & Stability Tests
3.8.1 ADF Test of the residuals of the ARIMA (2, 2, 1) Model
Table 12: Levels- Intercept

μ1 -7.259603 0.0000 -3.560019 @1% Stationary
Table 13: Levels- intercept and trend

μ1 -7.38592 0.0000 -4.140858 @1% Stationary

μ1 -7.028539 0.0000 -2.609324 @1% Stationary
Tables 12, 13 and 14 indicate that the residuals of the ARIMA (2, 2, 1) model
are stationary and bear the characteristics of a white – noise process.

31
Stability Test of the ARIMA (2, 2, 1) Model
Figure 5: Stability Test
The figure above shows that the ARIMA (2, 2, 1) model is stable since the
corresponding inverse roots of the characteristic polynomials are in the unit
circle.
ADF Test of the residuals of the ARIMA (1, 2, 0) Model
Table 15: Levels- intercept

μ2 -9.993238 0.0000 -3.560019 @1% Stationary
Table 16: Levels- intercept and trend

μ2 -10.14207 0.0000 -4.140858 @1% Stationary

μ2 -9.709689 0.0000 -2.609324 @1% Stationary
Tables 15, 16 and 17 indicate that the residuals of the ARIMA (1, 2, 0) model
are stationary and thus bear the features of a white – noise process.

32
Stability Test of the ARIMA (1, 2, 0) Model
Figure 6: Stability Test
The figure above indicates that the ARIMA (1, 2, 0) model is also stable since
the corresponding inverse root of the characteristic polynomial is in the unit
circle.
4 Results and Discussion
4.1 Descriptive Statistics
Table 18: Statistics

Description Statistic
Mean 34.16
Median 6.95
Minimum 0.7
Maximum 172.4
Standard Deviation 47.839
Skewness 1.5078
Excess Kurtosis 1.0887
As shown in table 18 above, the mean is positive. The large difference between
the maximum and the minimum confirms the existence of an upward trend in
the CPI time series. The skewness is 1.5078 and the most striking feature is that
it is positive, implying that the CPI series has a long right tail and is non –
symmetric. The rule of thumb for kurtosis is that it should be around 3 for
normally distributed variables as reiterated by Nyoni & Bonga (2017) and yet in
this analysis, kurtosis has been found to be 1.0887. Therefore, the CPI series
not normally distributed.

33
4.2 Results Presentation6
Table 19
ARIMA (2, 2, 1) Model:
………………………………………….. [56]
P: (0.01) (0.00) (0.82)
S. E: (0.29) (0.13) (0.25)
Variable Coefficient Standard Error z p – value
AR (1) [ ] -0.564292 0.294451 -2.76 0.0058***
AR (2) [ ] -0.538323 0.133071 -4.045 0.0000522***
MA (1) [ ] -0.0551283 0.247085 -0.2231 0.8234
ARIMA (1, 2, 0) Model:
……………………………………………………………..………………….. [57]
P: (0.00)
S. E: (0.12)
AR (1) [ ] -0.382082 0.124856 -3.06 0.0022***
AR (1) – GARCH (1, 1) Model
………………………………………………………….……………..………. [58]
P: (0.64) (0.00)
S. E (0.05) (0.00)
………………………………………………………..……….. [59]
P: (0.20) (0.00) (0.00)
S. E (0.00) (0.10) (0.07)
0.0241714 0.0518256 0.4664 0.6409
AR ( ) 1.07005 0.00264701 404.2 0.0000***
0.00708134 0.00550736 1.286 0.1985
ARCH ( ) 0.621061 0.101981 6.09 0.0000***
6
The *, ** and *** means significant at 10%, 5% and 1% levels of
significance; respectively.
34
GARCH ( ) 0.378939 0.0742271 5.105 0.0000***
ɑ1+β1 1
4.3 Interpretation & Discussion of Results
ARIMA (2, 2, 1) model
Both and (the AR components) are negative and statistically significant at

1% level of significance. (the MA component) is also negative but
insignificant. The AR components are closer to 1 (since they are well above 0.5),
implying that the series returns to its mean relatively slowly. The significance of
the MA component indicates that previous period shocks to inflation in Kenya do
not adequately explain current inflation levels.
ARIMA (1, 2, 0) model
(the AR component) is negative and significant at 1% level of significance.

Since is further from 1, it implies that the series returns to its mean quickly.
The results indicate that previous period inflation rates are quite important in
determining current rates of inflation in Kenya.
AR (1) – GARCH (1, 1) model
Since the positivity constraints specified in equations [54] are met, we safely
conclude that the estimated AR (1) – GARCH (1, 1) model is adequate. As
theoretically anticipated, and are greater than zero and is positive to
ensure that the conditional varience, is non – negative; thus the positivity
constraint of the GARCH model is not violated. Both and are positive and
statistically significant at 1% level of significance. is positive and statistically
significant, implying that strong GARCH effects cannot be ruled out. Actually, a
1% increase in previous period volatility leads to nearly 0.62% increase in
current volatility of CPI in Kenya. Since is positive and significant and indeed
less than one, it implies that the effect of old news on volatility is quite
important. In fact, an increase of 1% in previous period varience will lead to
about 0.38% increase in current volatility of CPI in Kenya. Since our empirical
model indicates that equation [47] has been confirmed, we conclude that the
estimated GARCH model is not covariance stationary; it is indeed strictly
stationary. Therefore, the estimated GARCH model is actually an IGARCH
model7. Thus current information remains essential for the forecasts of the
conditional variances for all horizons. Our findings, in this regard; are not quite
unique simply because they are similar to the findings by Moroke & Luthuli
(2015) and Nyoni (2018k).
7
The implications of IGARCH models are “too strong”. Therefore, there is need
for further research to consider fractionally integrated models such as the
Fractionally Integrated GARCH (FIGARCH) model, Fractionally Integrated
Exponential GARCH (FIEGARCH) model as well as the Fractionally Integrated
APARCH (FIAPARCH) model amongst others; when modeling and forecasting
inflation in Kenya.
35
4.4 Forecast Graphs

Figure 7: ARIMA (2, 2, 1) model forecast
Figure 8: ARIMA (1, 2, 0) model forecast
Figure 9: AR (1) – GARCH (1, 1) model forecast

36
The graphs 7, 8 and 9 (with a forecast range of 10 years, that is; 2018 – 2028)
indicate that Consumer Price Index in Kenya is likely to progress in an upward
trend, implying that inflation is likely to continue rising sharply in Kenya. The
most striking characteristic of the figures 7, 8 and 9 is that they strongly concur
in their forecasts; that inflation in Kenya is indeed expected to go up, of course;
as long as no appropriate counter – cyclical policy measures are taken.
5.0 Policy Implications

The study prescribes the following policy measures-
i) The CBK, in line with monetarism; ought to control money supply through
the use of a fixed monetary growth rate rule, commensurate with GDP
growth; in order to address inflation.
ii) The CBK can also make use of contractionary fiscal and monetary policy in
order to reduce spending and inflationary pressures.
iii) Policy makers in Kenya ought to consider supply – side policies such as
privatization and deregulation in order to reduce cost of doing business
and improve long – term competitiveness, productivity and innovation;
that will in turn lower inflation.
6.0 Conclusion
Monetary policy is more effective when it is forward looking (Svensson, 2005;
Faust & Wright, 2013). Maintenance of price stability continues to be one of the
main objectives of monetary policy for most countries in the world today (Nyoni
& Bonga, 2018a) and Kenya is not an exception. Achieving and maintaining price
stability will be efficient and effective the better we understand the causes of
inflation and the dynamics of how it evolves (Kohn, 2005). Therefore, inflation
forecasting is not unimportant in light of both fiscal and monetary policy making.
In Kenya, just like in any other country; inflation forecasts are envisioned to
enable the CBK to possess some control over future inflation. In this study, we
employ both ARIMA and GARCH techniques in order to model and forecast
inflation in Kenya.

37
References
 A. Meyler, G. Kenny and T. Quinn (1998). Forecasting Irish inflation using

ARIMA models, Central Bank of Ireland – Research and Publications
Department, Paper No. 3/RT/98.
 A.H.S.M. Nor, T.Y. Ling and F. Maarof (2007). On the relationship between
inflation and inflation uncertainty: an application of the GARCH family
models, Sains Malaysiana, 36 (2): 225 – 232.
 C. Brooks (2008). Introductory Economics for Finance, 2nd Edition,
Cambridge University Press, Cambridge.
 C. Chatfield. The analysis of time series, 5thed, New York: Chapman &
Hall, 1996
 C. Lidiema (2017). Modeling and forecasting inflation rate in Kenya using
SARIMA and Holt – Winters Triple Smoothing, Science Publishing Group –
American Journal of Theoretical and Applied Statistics, 6 (3): 161 – 169.
 C. Uwilingiyimana, J. Mungatu and J.D. Harerimana (2015). Forecasting
inflation in Kenya using ARIMA – GARCH models, International Journal of
Management and Commerce Innovations, 3 (2): 15 – 27.
 CBK (2017). Central Bank of Kenya – Monetary Policy Statement, CBK,
Nairobi.
 D. Jacobs, D. Perera and T. Williams (2014). Inflation and the cost of
living, Reserve Bank of Australia – Bulletin, 1 (1): 33 – 46.
 D. L. Kohn (2005). Modeling inflation: a policy maker’s perspective,
International Research Forum on Monetary Policy, Conference Paper,
Frankfurt.
 D.A. Kuhe and R.C. Egemba, (2016). Modeling and forecasting CPI
inflation in Nigeria: application of Autoregressive Integrated Moving
Average Homoscedastic Model, Journal of Scientific and Engineering
Research, 3 (2): 57 – 66.
 D.E. Box and G.M. Jenkins (1970). Time Series Analysis, Forecasting and
Control, Holden Day.
 D.E. Box and G.M. Jenkins (1974). Time Series Analysis, Forecasting and
Control, Revised Edition, Holden Day.
 E. Zivot (2008). Practical issues in the analysis of univariate GARCH
models, Department of Economics, Washington University.
 E.H. Etuk, B. Uchedu and U.A. Victoredema (2012). Forecasting Nigerian
inflation rates by a Seasonal ARIMA model, Canadian Journal of Pure and
Applied Sciences, 6 (3): 2179 – 2185.
 F. Gallego (2002). Monetary policy transmission mechanism, Central Bank
of Chile.
 G. Haberler (1960). Inflation, it’s causes and cures, American Enterprise
Association, Washington DC.
 G. Hill and R. Fildes (1984). The accuracy of extrapolation methods: an
automatic Box – Jenkins package (SIFT), International Journal of
Forecasting, 3: 319 – 323.
 G. Libert (1983). The M – competition with a Fully Automatic Box –
Jenkins approach, International Journal of Forecasting, 2: 325 – 328.
 G.M. Caporale, N. Pittis and N. Spagnolo (2003). IGARCH models and
structural breaks, Applied Economics Letters, 10: 765 – 768.
 G.S. Maddala (1992). Introduction to Econometrics, 2nd Edition, MacMillan
Publishing Company, New York.

38
 I. Kelikume and A. Salami (2014). Time series modeling and forecasting

inflation: evidence from Nigeria, The International Journal of Business and
Finance Research, 8 (2): 41 – 51.
 I. Zivko and M. Bosnjak (2017). Time series modeling of inflation and its
volatility in Croatia, Notitia, 3 (3): 1 – 10.
 J. Bullard (2011). Measuring inflation: the core is rotten, Federal Reserve
Bank of St. Louis Review, 93 (4): 223 – 233.
 J. Faust and J.H. Wright (2013). Forecasting inflation, Elsevier – Economic
Forecasting, 2A (2013).
 J. Stanford (2008). A “how – to” guide: understanding and measuring
inflation, Canadian Centre for Policy Alternatives (CCPA), Canada.
 J.D.M. Fisher, C.T. Liu and R. Zhou (2002). When can we forecast
inflation? Economic Perspectives, Q1: 30 – 42.
 K. Molebatsi and M. Raboloko (2016). Time series modeling of inflation in
Botswana using monthly Consumer Price Index, International Journal of
Economics and Finance, 8 (3): 15 – 22.
 L. Paulos, A. Kvanli and R. Pavur (1987). A comparison of the accuracy of
the Box – Jenkins automated forecasting model, International Journal of
 L.E.O. Svensson (2005). Monetary policy with judgement: forecasting
targeting, International Journal of Central Banking, 1 (1).
 M. Benedict (2013). Modeling rates of inflation in Ghana: an application of
ARCH type models, University of Ghana, MPhil Thesis.
 M. Friedman (1967). AEA Presidential Address – The Role of Monetary
Policy, American Economic Review, 58: 1 – 17.
 M. Friedman (1968). The Role of Monetary Policy, American Economic
Review, 17: 1 – 15.
 M. Goodfriend (2007). How the world achieved consensus on monetary
policy, The Journal of Economic Perspectives, 21 (4): 47 – 68.
 M. Verbeek (2004). A Guide to Modern Econometrics, 2nd Edition, John
Wiley & Sons Ltd.
 M.A. Wynne and F.D. Sigalla (1994). The Consumer Price Index, Federal
Reserve Bank of Dallas – Economic Review, 2 (2): 1 – 22.
 M.S. Wabomba, M.P. Mutwiri and F. Mungai (2016). Modeling and
forecasting Kenyan GDP using Autoregressive Integrated Moving Average
(ARIMA) models, Science Publishing Group – Science Journal of Applied
Mathematics and Statistics, 4 (2): 64 – 73.
 N. Habimana, A. Wanjoya and A. Waitu (2016). Modeling and forecasting
Consumer Price Index (case of Rwanda), Science Publishing Group –
American Journal of Theoretical and Applied Statistics, 5 (3): 101 – 107.
 N. Moroke and A. Luthuli (2015). An Optimal Generalized Autoregressive
Conditional Heteroskedasticity model for forecasting the South African
inflation volatility, Journal of Economics and Behavioural Studies, 7 (4):
134 – 149.
 N. Saleem (2008). Measuring volatility of inflation in Pakistan, The Lahore
Journal of Economics, 13 (2): 99 – 128.
 N.M. Mutwiri (2017). Monetary policy tools and inflation in Kenya,
International Journal in Academic Research in Accounting, Finance and
Management Sciences, 7 (1): 86 – 97.
 O. Osarumwense and E.I. Waziri (2013). Modeling monthly inflation rate,
using Generalized Autoregressive Conditionally Heteroskedastic (GARCH)
models: evidence from Nigeria, Australia Journal of Basic and Applied
Sciences, 7 (7): 991 – 998.
39
 P. Wang (2009). Financial Econometrics, 2nd Edition, Routledge – Taylor &

Francis Group, London.
 P.A. Samuelson and W.D. Nordhaus (1995). Economics, 15th International
Edition.
 P.A. Texter and J.K. Ord (1989). Forecasting using Automatic
Identification Procedures: A comparative analysis, International Journal of
 R. Kozhan (2010). Financial Econometrics with E – Views, Ventus
Publishing.
 S. Banerjee (2017). Empirical regularities of inflation volatility: evidence
from advanced and developing countries, South Asian Journal of
Macroeconomics and Public Finance, 6 (1): 133 – 156.
 S. Huseynov, V. Ahmadov and S. Adigozalov (2014). Beating a random
walk: hard times for forecasting inflation in post – oil years? Munich
Personal RePEc Archive (MPRA), Paper No. 63515.
 S.E. Alnaa and F. Ahiakpor (2011). ARIMA (Autoregressive Integrated
Moving Average) approach to predicting inflation Ghana, Journal of
Economics and International Finance, 3 (5): 328 – 336.
 S.K. Sek and W.M. Har (2012). Does inflation targeting work in emerging
East – Asian Economies? Panoeconomicus, 5 (2012): 599 – 608.
 S.M. Bokhari and M. Feridun (2006). Forecasting inflation through
econometric models: an empirical study on Pakistani data, Dogus
Universitesi Dergisi, 7 (1): 39 – 47.
 S.O. Fwaga, G. Orwa and H. Athiany (2017). Modeling rates of inflation in
Kenya: an application of GARCH and EGARCH models, Mathematical
Theory and Modeling, 7 (5): 75 – 83.
 T. Bollerslev (1986). Generalized Autoregressive Conditional
Heteroskedasticity, Journal of Econometrics, 31: 307 – 327.
 T. Jalil (2011). Macroeconomic theories of inflation, International
Conference on Economics and Finance Research, IACSIT Press, 4 (2011):
459 – 462.
 T. Nyoni (2018i). Box – Jenkins ARIMA Approach to Predicting net FDI
inflows in Zimbabwe, University of Munich Library – Munich Personal
RePEc Archive (MPRA), Paper No. 87737.
 T. Nyoni (2018k). Modeling and forecasting inflation in Zimbabwe: a
Generalized Autoregressive Conditionally Heteroskedastic (GARCH)
approach, University of Munich Library – Munich Personal RePEc Archive
(MPRA), Paper No. 88132.
 T. Nyoni (2018l). Modeling and Forecasting Naira / USD Exchange Rate in
Nigeria: a Box – Jenkins ARIMA approach, University of Munich Library –
Munich Personal RePEc Archive (MPRA), Paper No. 88622.
 T. Nyoni and W.G. Bonga (2017). Population growth in Zimbabwe: A
Threat to Economic Development? DRJ – Journal of Economics and
Finance, 2 (6): 29 – 39.
 T. Nyoni and W.G. Bonga (2018a). What determines economic growth in
Nigeria? DRJ – Journal of Business and Management, 1 (1): 37 – 47.
 Y.F. Huang (2012). Forecasting Chinese inflation and output: A Bayesian
vector autoregressive approach, Munich Personal RePEc Archive (MPRA),
Paper No. 41933.

40
View publication stats

ARIMA and GARCH

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ARIMA and GARCH

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

MODELING AND FORECASTING INFLATION IN KENYA: RECENT INSIGHTS

Article · November 2018

Neuromarketing Mix in the 21st Century View project

The user has requested enhancement of the downloaded file.

MODELING AND FORECASTING INFLATION IN KENYA: RECENT

Key words: ARIMA, Consumer Price Index (CPI), Forecasting, GARCH,

The practice of forecasting inflation has generally been considered an important

the dynamics of inflation in Kenya, in this research; we forecast Kenya’s inflation

1.1 Literature Review

1.1 Theoretical Literature Review

1.1.1. Keynesian Theory of Inflation

1.1.2 Monetarist Theory of Inflation

The father of monetarism is the great economist, Milton Friedman, a Classical

mismanagement. In fact, Friedman (1967) is well known for his popular

1.2 Empirical Literature Review

The empirical literature on inflation forecasting points to different techniques in

Closer to home, in Africa; Habimana et al (2016) modeled and forecasted

DIMORIANREVIEWwww.dimorianreview.comA PEER REVIEWED MULTIDISCIPLINARY INTERNATIONAL INDEXED E-JOURNAL

Back to Kenya, Uwilingiyimana et al (2015) studied inflation using ARIMA –

3.1.1 The Moving Average (MA) model

= + +…+ ………………………………………………….…….……………………. [1]

is an MA process of order one, commonly denoted as MA (1). Owing to the fact

E( )=0 ……………………………………..……………………………….…………………………………….. [3]

for all t, it therefore; implies that:

DIMORIANREVIEWwww.dimorianreview.comA PEER REVIEWED MULTIDISCIPLINARY INTERNATIONAL INDEXED E-JOURNAL

E( )=0 …………………………………………………………………………………………………………… [4]

Var( ) ( )σ2 ………………..…………………………………….…………………………………. [5]

where μt is independent with a common varience σ2. Hence, we can now re –

= + +…+ ……………………………………………….…………………………………. [6]

Equation [6] can be re – written as:

CPIt= + ……………………………………………………………….……….………………….. [7]

We can also write equation [7] as follows:

where L is the lag operator.

=ɑ(L) …………………...……………………………………………………………..…………………… [9]

ɑ(L)=θ(L)2 ………………………………………………………………………….………….……………….. [10]

3.1.2 The Autoregressive (AR) model

CPIt=β1CPIt-1+…+βpCPIt-p+μt …………………………………..…………….………………………. [11]

Where β1 … βp are estimation parameters, CPIt-1 … CPIt-p are previous period

CPIt= +μt ………………………………………...………………………..…………………… [12]

Equation [12] can be re – written as:

CPIt= +μt ……………………………...………………………………………………………… [13]

β(L)CPIt=μt …………………………………………………………………………………………………….. [14]

β(L)=ɸ(L)3 ………………………………………………………………………………………………………… [15]

CPIt=(β1L+…+βpLp)CPIt+μt …………………………………………………………….……………… [16]

CPIt=(β1L)CPIt+μt ……………………………………….………………………………….……………… [17]

is an AR process of order one, commonly denoted as AR (1).

3.1.3 The Autoregressive Moving Average (ARMA) model

As initially propounded by Box & Jenkins (1970), an ARMA (p, q) process is

CPIt=β1CPIt-1+…+βpCPIt-p+μt+ɑ1μt-1+…+ɑqμt-q …...………………………………………… [18]

CPIt= + + ………………………………………………………..…………….. [19]

ɸ(L)CPIt=θ(L)μt ……………………………..……………………………………………………………….. [20]

whereɸ(L) and θ(L) are polynomials of orders p and q respectively, simply

ɸ(L)=1-β1L … βpLp ……………………..……………………………………………………………………. [21]

θ(L)=1+ɑ1L+…+ɑqLq ……………………………………………………………………..………………… [22]

3.1.4 The Autoregressive Integrated Moving Average (ARIMA) model

∆dCPIt= + +μt ……………………….…………………….………………. [23]

which we can also re – write as:

∆dCPIt= + +μt ………………..……………………………………………… [24]

where ∆ is the difference operator, vector β ϵⱤp and ɑϵⱤq.

DIMORIANREVIEWwww.dimorianreview.comA PEER REVIEWED MULTIDISCIPLINARY INTERNATIONAL INDEXED E-JOURNAL

3.1.5The Autoregressive Conditionally Heteroskedastic (ARCH) model

=var(μt│μt-1, μt-2, …)=E[μt-E(μt)2│μt-1, μt-2, …] …………………….….……………….. [25]