Topic 2 - Exploring the Short-Run Impact of Economic Growth and Inflation on Unemployment
in Vietnam: An ARDL Approach
1. Research Question
This study aims to investigate the short-run relationship between key macroeconomic indicators
—specifically, GDP growth and inflation—and the unemployment rate in Vietnam. While
traditional economic theory, such as the Phillips Curve and Okun’s Law, suggests inverse
relationships between these variables, the empirical evidence for emerging economies remains
mixed and context-dependent.
Accordingly, the central research question addressed in this paper is: “Do GDP growth and
inflation significantly affect the unemployment rate in Vietnam in the short run?”
To answer this, we employ the Autoregressive Distributed Lag (ARDL) modeling framework,
which allows for the analysis of variables with mixed orders of integration and accommodates
both short-run and long-run dynamics. Given the lack of long-run cointegration among the
variables, the analysis will primarily focus on short-run dynamics and forecasting implications.
2. Empirical Model
To empirically examine the relationship between economic growth, inflation, and unemployment
in Vietnam, this study employs the Autoregressive Distributed Lag (ARDL) model as the
primary methodological framework. The ARDL model is particularly suitable when the variables
are a mixture of I(0) and I(1), and it provides reliable estimates for both short-run and long-run
relationships. However, since the bounds test results in this study do not support the presence of
cointegration, the analysis focuses solely on the short-run dynamics.
The general ARDL(p, q₁, q₂) model specification used in this paper is expressed as follows:
The selection of optimal lag length (p, q₁, q₂) was determined using the Akaike Information
Criterion (AIC) through the auto_ardl() function in R. After estimating the model, diagnostic
tests including the Augmented Dickey-Fuller (ADF) test for stationarity and the Bounds test for
cointegration were conducted. Due to the absence of long-run cointegration, an Error Correction
Model (ECM) was not applicable. Forecasting was subsequently carried out based on the short-
run ARDL specification using hypothetical macroeconomic scenarios.
3. Data Collection and Variable Description
a. Data Source and Sample Structure
This study utilizes annual time series data for Vietnam covering the period from 2000 to 2022.
All macroeconomic indicators were collected from the World Bank’s World Development
Indicators (WDI) database, ensuring consistency, reliability, and international comparability.
The sample includes 23 observations, which is deemed sufficient for ARDL estimation given the
model’s flexibility with small sample sizes. The data was extracted using the WDI package in R
and cross-checked manually through the World Bank DataBank portal to ensure data integrity.
b. Variable Description and Measurement
This study investigates three key macroeconomic variables—unemployment, economic growth,
and inflation—all measured on an annual basis and expressed in percentage terms to ensure
comparability.
- Unemployment Rate (UNEMP):
The dependent variable represents the percentage of the labor force that is actively
seeking but unable to find employment. It reflects labor market conditions and is a
critical indicator of economic slack. A higher unemployment rate signals underutilization
of human capital and can negatively impact productivity and social welfare.
- GDP Growth Rate (GDPG):
This variable measures the annual percentage change in real gross domestic product,
capturing the pace of economic expansion or contraction. According to Okun’s Law, an
increase in GDP growth is typically associated with a decrease in unemployment, as
stronger economic activity creates more jobs. Therefore, a negative relationship is
expected between GDPG and UNEMP in the short run.
- Inflation Rate (INFL):
Defined as the annual percentage change in consumer prices, inflation reflects the overall
price level dynamics in the economy. According to the Phillips Curve hypothesis, there
may be a short-run trade-off between inflation and unemployment, implying that rising
inflation could coincide with falling unemployment. This variable is included to assess
whether this inverse relationship holds in the Vietnamese context.
Overall, the inclusion of both GDP growth and inflation allows for a comprehensive analysis of
how macroeconomic conditions shape labor market outcomes, particularly within a developing
economy setting.
4. Descriptive Statistics
Before proceeding to model estimation, a descriptive analysis is conducted to summarize the key
features of the dataset and provide preliminary insights into the behavior of the variables over
time. Table 1 presents the summary statistics for the unemployment rate (UNEMP), GDP growth
(GDPG), and inflation rate (INFL) for Vietnam from 2000 to 2022.
Table 1. Summary Statistics
Variable Obs Mean Std. Dev. Min Max
UNEMP 23 2.23 0.38 1.77 3.18
GDPG 23 6.27 1.47 2.91 8.48
INFL 23 5.90 4.23 0.63 23.11
The average unemployment rate over the period was approximately 2.23%, with relatively low
volatility, suggesting a generally stable labor market. The GDP growth rate had a mean of
6.27%, reflecting robust economic performance with moderate fluctuations. Inflation averaged
5.90%, but with a higher standard deviation, indicating more pronounced variability in price
levels, consistent with occasional inflationary episodes during the sample period.
The minimum GDP growth rate occurred during years of external economic shocks (e.g., the
global financial crisis or the COVID-19 pandemic), while the maximum inflation rate exceeded
23%, underscoring the susceptibility of the Vietnamese economy to price instability in certain
periods.
Overall, the descriptive results suggest that while the labor market remained relatively steady,
macroeconomic variables like inflation exhibited greater volatility, warranting further
investigation into their short-run effects on unemployment through the ARDL framework.
5. Discussion of Quantitative Results
a. Stationarity Testing
To ensure the validity of the ARDL model, it is essential to examine the order of integration of
each variable. The Augmented Dickey-Fuller (ADF) test was applied to assess whether the series
are stationary at level or require differencing.
The ADF test results reveal that all three variables—unemployment (UNEMP), GDP growth
(GDPG), and inflation (INFL)—are non-stationary at level but become stationary after first
differencing. Specifically, the test statistics at level do not reject the null hypothesis of a unit
root at conventional significance levels, while the first-differenced series exhibit statistically
significant ADF statistics, indicating integration of order one, or I(1).
These findings support the application of the ARDL model, which allows for the inclusion of
variables with mixed orders of integration (I(0) and I(1)), provided none is I(2).
b. Cointegration Test (Bounds Test)
To examine the existence of a long-run equilibrium relationship among the variables, the ARDL
Bounds Test was conducted. The F-statistic from the test was 3.173, and the corresponding p-
value was 0.2313, exceeding the 10% critical upper bound. As a result, the null hypothesis of no
cointegration could not be rejected.
This implies that there is no statistically significant long-run relationship among
unemployment, GDP growth, and inflation over the sample period. Consequently, the focus of
the empirical analysis shifts toward understanding the short-run dynamics of the variables.
c. Estimation Results
Following the rejection of long-run cointegration, the short-run ARDL model was estimated
using the optimal lag structure selected by the Akaike Information Criterion (AIC). The model
includes two lags of the dependent variable (UNEMP), one lag of GDPG, and contemporaneous
values of INFL and GDPG.
The estimation results indicate several notable relationships:
- GDP growth (GDPG) has a statistically significant and negative short-run effect on
unemployment (β = –0.14575, p = 0.0138). This finding is consistent with Okun’s Law,
which suggests that higher economic growth reduces unemployment.
- Interestingly, the first lag of GDPG exhibits a positive coefficient (β = 0.11097, p =
0.0440), indicating a short-term reversal or delayed effect possibly linked to labor market
adjustment frictions.
- Inflation (INFL) also shows a significant negative relationship with unemployment (β
= –0.02739, p = 0.0456), supporting the short-run Phillips Curve hypothesis that
inflationary pressures can temporarily reduce unemployment.
- The lagged unemployment terms were found to be statistically insignificant, suggesting
that the labor market does not exhibit strong autoregressive behavior within this short-run
framework.
The model explains approximately 68% of the short-run variation in unemployment (R² =
0.6801) and is statistically significant overall (F = 6.379, p = 0.0023), indicating a good fit.
d. Forecasting Analysis
To evaluate the model’s predictive capacity, a hypothetical forecast was generated based on
plausible macroeconomic assumptions. Under the scenario where GDP growth is projected at
6.0%, inflation is held at 4.5%, and the lagged unemployment rates are 2.2% and 2.25%
respectively, the model forecasts the unemployment rate to decline to approximately 2.07%.
This result highlights the model’s potential utility in short-run macroeconomic policy
simulations. However, caution is warranted for long-term forecasting due to the absence of
cointegration, meaning deviations from equilibrium may not self-correct over time.
e. Overall Implications
The empirical evidence suggests that Vietnam’s labor market responds significantly to short-
term changes in GDP growth and inflation, but lacks a stable long-run equilibrium path.
Policymakers aiming to reduce unemployment should consider the short-term trade-offs revealed
in the model, particularly in managing economic cycles and inflation dynamics. Given that both
GDP growth and inflation exhibit significant short-run effects, coordinated fiscal and monetary
policies can be crucial in sustaining employment levels, especially during periods of economic
uncertainty.
Source
The macroeconomic data used in this study—unemployment rate, GDP growth, and inflation—
were retrieved from the World Bank’s World Development Indicators (WDI) database. All
data are annual and country-specific for Vietnam, covering the period from 2000 to 2022.
World Bank. (2023). World Development Indicators. Retrieved from
[Link]