Ordered Probit Model
ECONOMETRICS 2
How do macroeconomic factors influence inflation rates?
BY ABHIRAM G 22FMUCH020025
TABLE OF CONTENTS
1. Introduction ............................................................ Page 3
2. Review of Literature .............................................. Page 4
3. Objectives ............................................................... Page 5
4. Methodology ........................................................... Page 6
5. Results and Interpretations .................................. Page 8
6. Conclusion ............................................................. Page 10
7. Suggestions and Recommendations .................... Page 11
8. References ............................................................. Page 12
2
INTRODUCTION
Inflation is a critical economic variable that directly impinges on the purchasing power of
money, affecting individuals and firms. Indeed, understanding the factors of influence on
inflation is what policymakers and economists must do when designing appropriate strategies
to maintain stability in the economy. Important factors considered to shape an inflationary
trend include the unemployment rate, exchange rate volatility, GDP growth, government debt,
and growth of the money supply.
This paper now conducts an econometric analysis in order to investigate how these economic
variables influence inflation rates. Using an ordered probit model-a statistical technique
designed for modeling ordinal outcomes-the analysis categorizes inflation rates into three
levels: low, medium, and high. This paper tries to investigate the relationship between these
pivotal economic factors and inflation, therefore, providing useful lessons that can be drawn
about the dynamics of inflation and an increased understanding of the inflationary pressures
within an economy.
3
REVIEW OF LITERATURE
Inflation Dynamics: A Structural Econometric Analysis
J. Galí, M. Gertler (1999)
This seminal study looks into inflation dynamics through structural econometric modeling,
thus showing how variables like the unemployment rate and GDP growth interact with
inflation. Its focus is on econometric specifications and hence falls within the methodology
that has been applied in this project.
Econometric Inflation Targeting
G. Bårdsen, E. S. Jansen, R. Nymoen (2003)
This paper emphasizes the importance of econometric methodologies in analyzing inflation
models and targeting inflation. Its method of incorporating macroeconomic variables
especially qualifies it for an understanding of the influence of variables such as exchange rate
volatility and money supply growth on inflation.
An Econometrics Analysis of the Determinants of Inflation in Namibia
M. Likukela (2007)
This paper makes use of econometric models in establishing the inflation determinants for an
emerging market. This provides a comparison which would be useful to support the
perspective taken in this study. Certain variables employed in the estimation, like money
supply growth and unemployment, are used directly in the project.
Determinants of Inflation in Pakistan: An Econometric Analysis Using Johansen Co-
Integration Approach
D. F. Bashir (2011)
4
The present paper tries to investigate the determinants of inflation, using advanced
econometric tools and emphasizing the role of macroeconomic factors. Its methodology and
findings are of great value to the analysis of growth rates to inflation and other determinants
too.
OBJECTIVES
The key variables that drive inflation need to be analyzed in detail, placing particular
emphasis on variables like unemployment rate, exchange rate volatility, GDP growth,
government debt-to-GDP ratio, and money supply growth.
It tries to understand how these variables relate to inflation rates and their role in shaping
inflationary trends.
For this purpose, an ordered probit model of studying inflation-as an ordered category, low-
medium-high-is provided in detail through an econometric analysis.
The paper indicates that using econometric modeling is an effective means to comprehend the
complex nature of the relationship between macroeconomic variables and inflation.
RESEARCH METHODOLOGY
This paper has adopted an econometric approach in analyzing the determinants of inflation
rates, particularly the ordered probit model. The suitability of the model in this analysis is
that it considers inflation to be an ordered categorical variable-that is, low, medium, and high-
rather than being continuous. It will thus capture how the macroeconomic variables impact
different levels of inflation.
5
Definitions
The analysis will be based on a dataset of 50 entries each of the following variables:
Unemployment Rate: This reflects labor market conditions and its implications for inflation
through nominal wage dynamics.
Exchange Rate Volatility: This represents the external shocks of inflation by import costs and
foreign investment flows.
GDP Growth Rate: This proxy variable is for economic activity, hence demand-pull inflation.
Government Debt-to-GDP Ratio: It reflects fiscal sustainability and, therefore, its consequent
impact on inflationary expectations.
Money Supply Growth: Reflects the monetary policy impact on the rate of inflation.
The choice of the variables is such that they are fundamentals in setting up the
macroeconomic conditions and significantly affect inflation. Unemployment and GDP growth
reflect domestic health, exchange rate volatility reflects external pressures, and government
debt reflects fiscal pressure. Money supply growth has a direct relation to monetary
mechanisms and their relation with inflation.
DATA SOURCES
Unemployment Rate
Data sourced from the World Bank's World Development Indicators (WDI).
World Bank - WDI Database
Exchange Rate Volatility
Extracted from the International Monetary Fund (IMF)'s International Financial Statistics
6
(IFS).
IMF - International Financial Statistics
GDP Growth Rate
Data obtained from the OECD Economic Outlook Database.
OECD Economic Outlook Database
Government Debt-to-GDP Ratio
Gathered from the Bank for International Settlements (BIS) statistics.
BIS - Debt Statistics
Money Supply Growth
Sourced from the Federal Reserve Economic Data (FRED).
FRED - Money Supply Data
Model Specification
The ordered probit model is specified as:
Inflation_Factor=β0+β1(Unemployment_Rate)+β2(Exchange_Rate_Volatility)+β3
(GDP_Growth_Rate)+β4(Government_Debt_to_GDP)+β5(Money_Supply_Growth)+ϵ
Where:
Inflation_Factor is the dependent variable, categorized as low, medium, or high.
β₁ to β₅ represent the coefficients for the independent variables.
ϵ is the error term.
This model quantifies the relationship between inflation levels and the selected
macroeconomic factors, offering insights into their relative importance and directional
effects.
7
Null Hypothesis (H₀):
The macroeconomic factors (unemployment rate, exchange rate volatility, GDP growth rate,
government debt-to-GDP ratio, and money supply growth) have no significant impact on the
likelihood of inflation falling into different categories (low, medium, or high).
Mathematically:
H₀: β₁ = β₂ = β₃ = β₄ = β₅ = 0
Alternative Hypothesis (H₁):
At least one of the macroeconomic factors has a significant impact on the likelihood of
inflation falling into different categories.
Mathematically:
H₁: ∃ βᵢ ≠ 0, where i ∈ {1, 2, 3, 4, 5}
RESULTS AND INTERPRETATIONS
The ordered probit model was used to analyze the impact of macroeconomic factors on
inflation, categorized as low, medium, or high. Below is a summary of the results obtained
from the model:
Coefficient Standard t-
Variable Significance
Estimate Error value
Unemployment Significant
0.1897 0.0948 2.0015
Rate at 5%
Exchange Rate Significant
0.331 0.1154 2.8699
Volatility at 1%
GDP Growth Not
0.0428 0.1253 0.3415
Rate significant
8
Government - Not
-0.0068 0.0067
Debt to GDP 1.0037 significant
Money Supply Not
0.1971 0.1445 1.3644
Growth significant
Intercept Values:
Threshold Estimate Standard Error
Low to Medium 2.4237 1.4344
Medium to High 5.8744 1.7429
Residual Deviance:
The residual deviance of the model is 47.30, and the AIC value is 61.30, indicating a
reasonable model fit.
INTERPRETATION
Unemployment Rate:
A coefficient value of 0.1897 infers that with increased rates of unemployment, the chances
of inflation moving from a lower to a higher category are also higher.
Exchange Rate Volatility:
The coefficient value is 0.3310, so this makes the volatility of the exchange rate highly
influential in inflation. Higher fluctuations in the exchange rate raise probabilities for higher
levels of inflation.
GDP Growth Rate:
9
However the impact it has on inflation, as shown from the results by the low coefficient of
0.0428, is very small and thereby statistically insignificant given the high value of the p-
value.
Government Debt to GDP:
The coefficient is negative, -0.0068, indicating a reverse relationship to inflation; the effect is
not statistically significant.
Money Supply Growth:
Money supply growth records a positive coefficient of 0.1971, meaning that with money
supply growth comes higher inflation levels; this has also been statistically insignificant.
Intercepts:
From the t-values, the threshold values that split the categories of inflation, namely low,
medium, and high, are far away from each other.
CONCLUSION
This paper applied the ordered probit model to study the impact of some key macroeconomic
variables such as the unemployment rate, exchange rate volatility, GDP growth, government
debt, and money supply growth on inflation. Results from the analysis show that among those
variables, unemployment rate and exchange rate volatility are significant to explain inflation,
while the relationship of these variables with inflation is positive. In other words, when the
unemployment rate and exchange rate volatility increase, the probability of higher inflation
also increases.
10
However, in this analysis, GDP growth, government debt, and growth in the money supply
did not affect inflation significantly. It may suggest that these variables, within this study, are
less influential on inflation than perhaps hypothesized.
The ordered probit model is useful in inferring the threshold that separates various levels of
inflation, with key cut-offs between "Low", "Medium", and "High" rates of inflation being
identified. Such relationships will always go a long way in helping policymakers develop
better monetary and fiscal policies for managing inflation within an economy.
This study contributes to an understanding of inflation dynamics and pinpoints those factors
of a macroeconomic setting that significantly influence inflation, while also providing a
quantitative framework for the assessment of such effects. Further research could be done to
include more variables or try different modeling techniques to strengthen these findings.
SUGGESTIONS AND RECOMMENDATIONS
Policy Interest in Unemployment and the Exchange Rate: Policymakers need to show more
interest in the levels of unemployment and fluctuations in the exchange rate to combat
inflation effectively.
Future studies on other variables may consider incorporating different macroeconomic
variables or employing alternative models to more accurately depict the complex nature of
inflation dynamics.
Long-run data should be used where possible for a better demonstration of long-run impacts
over time that are captured in the variable effects of macroeconomics.
11
Macroprudential policies should be implemented to help in the stabilization of the exchange
rate and maintaining the government's debt at sustainable levels to have low inflationary
pressures.
Central banks could also refine inflation-targeting frameworks, focusing on the more
significant variables identified in this study.
REFERRENCES
Galí, J., & Gertler, M. (1999). Inflation dynamics: A structural econometric analysis. Journal
of Monetary Economics, 44(2), 195-222. [Link]
Bårdsen, G., Jansen, E. S., & Nymoen, R. (2003). Econometric inflation targeting. The
Econometrics Journal, 6(2), 234-259. [Link]
Likukela, M. (2007). An econometrics analysis of the determinants of inflation in Namibia.
Unpublished Manuscript. Retrieved from [Link]
12
Bashir, D. F. (2011). Determinants of inflation in Pakistan: An econometric analysis using
Johansen co-integration approach. Australian Journal of Business and Management
Research, 1(10), 1-11. Retrieved from
[Link]
13