ECONOMICS FORECAST REPORT FORECAST VIETNAM MONTHLY CONSUMER PRICE INDEX FROM JANUARY 2019 TO JUNE 2019

INTRODUCTION

The Consumer Price Index (CPI) is a crucial economic indicator that significantly influences society by reflecting the inflation rate of a country over a specific period Fluctuations in the CPI can greatly affect the domestic market, as surpassing certain thresholds may lead to inflation or deflation, potentially destabilizing the economy in the short term Analyzing and calculating the CPI not only enables approximate predictions about future economic conditions but also serves as a vital alert for the government and citizens, allowing them to implement necessary measures to mitigate or adapt to adverse economic effects.

Understanding the significance of the Consumer Price Index (CPI) in the economy, our group utilizes insights from Economics 1, Economics 2, and Economic Forecasting We conduct a quantitative analysis of the monthly CPI data from January 2014 to December 2018 to support our project titled "Forecast Vietnam Monthly Consumer."

PRICE INDEX FROM JANUARY 2019 TO JUNE 2019" Paper structure consists of 3 sections:

METHODOLOGY AND DATA DESCRIPTION

Exponential Smoothing Method

Single exponential smoothing method formula:

TIEU LUAN MOI download : skknchat@gmail.com

- Ft - Forecast demand level for period t

- Ft-1 - Forecast demand level for period t-1

- At-i - Actual demand level for period t-i

The method of double exponential is the double smoothing, which reflects the changing trend of demand.

- Formula of double exponential smoothing method:

2.1.3 The Holt exponential smoothing method:

When the data set is biased, the single and double exponential smoothing method will produce huge errors The Holt method has a trend adjustment.

The Holt equation is written as follow:

Estimated current average value: F t = α A t + (A1- α)(AF t -1 + T t-1 )

Forecast for the future period: F t+h = F t + h.T t In which:

- β = The exponential constant estimates the trend (0 Exponential smoothing => Simple exponential smoothing => Single

The result are as follow:

Table 3-1 Result of the Single Exponential Smoothing Method

Source: Author’s calculation using Eviews

Forecasted values of CPI in Vietnam from January 2019 to June 2019 are as in the following table:

TIEU LUAN MOI download : skknchat@gmail.com

Table 3-2 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Single Exponential Smoothing Method

Graph the differences between observed CPI and predicted CPI using Eview

We run the following command: line cpi cpism

Figure 3-3 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Single Exponential Smoothing Method

Source: Author’s calculation using Eviews 3.1.2 Double Exponential Smoothing

In Stata, we use the following steps:

Proc=>Exponential smoothing => Simple exponential smoothing => Double

Table 3-3 The result of Double Exponential Smoothing

Table 3-4 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Double Exponential Smoothing Method

Figure 3-4 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Double Exponential Smoothing Method

In Stata, we use the following steps:

Proc => Exponential smoothing => Simple exponential smoothing => Holt – winter: No seasonal The results are as follow:

Table 3-5 The result of Holt Exponential Smoothing

Table 3-6 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Holt Exponential Smoothing Method

Figure 3-5 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Holt Exponential Smoothing Method

First, we have to determine whether the data is Multiplicative or Additive We run the following command: line cpi

The result is as follow:

The line graph indicates that the data is multiplicative After identifying the data type, we will proceed with the subsequent steps in EViews.

Proc => Exponential smoothing => Simple exponential smoothing => Holt – winter- Multiplicative

The result is as follow:

Table 3-7 The result of Winter Exponential Smoothing

Table 3-8 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Winter Exponential Smoothing Method

Graph the differences between observed CPI and forecasted CPI using Eview

Figure 3-7 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Winter Exponential Smoothing Method

Source: Author’s calculation using Eviews 3.1.5 Analytics Method

From Winter Exponential Smoothing, we have already determined the model as Multiplicative

- Step 2: Separate the seasonal elements from the series

Proc → Seasonal Adjustment → Moving Average → Adjustment Method → Ratio to moving average – Multiplicative → Factors: sr

- Step 3: Estimate the trend function and forecast

Type the command: Genr t = @trend(2014M01) → ls cpi c t cpi(-1) We have the following result:

Table 3-9 Result of Analytics method

We have the following pair of hypothesis:

H 1 : The model is not insignificant

From the above table, we can see that variable CPI(-1) have p-value = 0.0051 < α 0.05 => Reject H0 => The model is significant

In Stata, choose View -> Residual Diagnostics -> Serial Correlation LM test

From the above table, we can see that p-value = 0.1947 > α = 0.05 => Does not reject H0 => Collinearity doesn’t exist

Histogram – Normality test: In Stata, choose View -> Residual Diagnostics ->

Histogram – Normality LM test We have the following graph:

Since number of observations is 59 -> The model pass Normality test Heteroskedasticity Test:

In Stata, choose View -> Residual Diagnostics -> Heteroskedasticity Test We have the following result:

From the above table, we can see that p-value = 0.1316 > α = 0.05 => Does not reject H0 => Heteroskedasticity doesn’t exist

{ H 0 : p=0 (Omitted variable does not exist)

In Stata, choose View -> Stability Diagnostics -> Ramsey RESET Test We have the following result:

Table 3-12 Omitted variable test result

From the above table, we can see that p-value = 0.000 < α = 0.05 => Reject H0 =>

We will proceed to forecast CPI from 2014M01 2015M01 and obtain the following result:

Figure 3-9 Forecasted values of CPI in Vietnam from January 2019 to June 2019 using Analytics Method

- Step 4: Combine the trend and seasonal factors to give the final forecast results

In Stata we use the following command Genr cpip = cpiaf1*sr and obtain the following result:

Table 3-13 Predicted values of CPI in Vietnam from January 2019 to June 2019 using Analytics Method

To draw a graph we use the following command line cpip cpi

Figure 3-10 Predicted values of CPI in Vietnam from January 2019 to June 2019 using Analytics Method

Source: Author’s calculation using Eviews 3.1.6 Forecast using ARIMA model

- Step 1: Check the stationary of series:

In Eview, choose view → Unit Root Test → Level & Intercept and obtain the following result:

Table 3-14 Unit Root Test result

We have P-value = 0.0000 < 0.05, => Reject H 0 => CPI is stationary at level

- Step 2: Estimate the parameters and select the model of the parameters

In Eview, choose view → Correlogram Specification → 1st different & lags to 24

Table 3-15 Correlogram Specification of CPI

From the above result, we choose p = 4 and q = 3 Type the following command ls d(cpi) c ar(4) ar(3)

Table 3-16 Seasonal ARIMA model result

- Look at above table, authors see that both Inverted AR Roots and Inverted MA Roots are smaller than 1 => The model is stable

- White noise test: View → Residual diagnostics → Correlogram Q-statistics

- Result: p-value > 0.1 => No White noise

Table 3-17 Stable and invertible test result

Figure 3-11 Forecasted result of CPI in Vietnam from January 2015 to January

- Step 4: Forecast outside the sample

Table 3-18 Predicted result of CPI in Vietnam from January 2019 to June 2019 using ARIMA Method

Graph using the command line cpip1 cpi

From Data description, we can see that the data has seasonal factor

- Step 1: Extract Seasonal factor from the data

Proc → Seasonal Adjustment → Moving Average → Adjustment Method → Ratio to moving average – Multiplicative → Factors: sr

- Step 2: Check the stationary of serie

Next we check to see if the non-seasonal series is stationary or not In Eview, choose view → Unit Root Test → Level & Intercept and obtain the following result:

Table 3-19 Unit Root Test result

We have P-value = 0.0000 < 0.05, => Reject H 0 => CPI is stationary at level

- Step 2: Estimate the parameters and select the model of the parameters

Next we choose p and q In Eview, choose view → Correlogram Specification → 1st different & lags to 24

Table 3-20 Correlogram Specification of CPI

From the above result, we choose p = 1 and q = 2 Type the following command ls d(cpi) c ar(1) ar(2)

Table 3-21 Non-seasonal ARIMA model result

- Look at above table, authors see that both Inverted AR Roots and Inverted MA Roots are smaller than 1 => The model is stable

- White noise test: View → Residual diagnostics → Correlogram Q-statistics

- Result: p-value > 0.1 => No White noise

Table 3-22 Stable and invertible test result

- Step 4: Forecast outside the sample

Type the following command genr cpip3 = cpip2*sr and obtain the following predicted values:

Table 3-23 Predicted result of CPI in Vietnam from January 2019 to June 2019 using SARIMA Method

Graph using the command line cpip3 cpi

Figure 3-14 Forecasted result of CPI in Vietnam from January 2019 to June 2019 using SARIMA Method

Table 4-24 Predicts monthly CPI from January 2019 to June 2019 using different method

Single exponential Double exponential Holt exponential Winter exponential Analytics ARIMA models ARIMA model

Time smoothing method smoothing method smoothing method method without seasonal with seasonal smoothing method

Source: Author’s calculation using Eview

The Analytics method is the most effective approach for predicting Vietnam's Consumer Price Index (CPI) from January to June 2019, as it yields the lowest RMSE/MAPE of 0.226791 The forecast indicates that while Vietnam's monthly CPI values may fluctuate, there is an overall upward trend anticipated over the next six months, as detailed in the accompanying table.

Table 4-25 Predicted CPI from January 2019 to June 2019 using Analytics method

Source: Author’s calculation using Eview

A high Consumer Price Index (CPI) signals a potential risk of elevated inflation levels Economists caution that without effective measures to manage the CPI, the threat of sustained high inflation persists To address this issue, several recommendations have been proposed for the Vietnamese government aimed at stabilizing the CPI.

To stabilize the economy, the government should manage the price level and its related factors, including gold prices and exchange rates For instance, the State Bank should work to minimize the disparity between domestic gold prices and global prices, which would discourage speculators from purchasing USD to import gold, thereby protecting the exchange rate from volatility.

To enhance its economic stability, Vietnam should focus on attracting diverse sources of foreign currency, particularly through foreign direct investment (FDI), official development assistance (ODA), remittances, and tourism This strategy will contribute to stabilizing the exchange rate, improving the balance of payments, and boosting foreign exchange reserves.

Vietnam should implement additional monetary measures to control rising prices For instance, increasing the compulsory reserve ratio can directly influence the credit market, effectively tightening monetary policy by decreasing the amount of cash in circulation.

To enhance the appeal of the Vietnamese Dong (VND) and encourage the conversion from US Dollars (USD) to VND, it is essential to maintain a significant gap between deposit interest rates for the two currencies Additionally, the State Bank should consider injecting USD liquidity into the interbank market as needed, while also urging commercial banks and export enterprises to boost the supply of USD in the market.

We would also want to suggest that Vietnamese government cut back on their public spending in order to lower inflation rate.

General statistics office of Vietnam (2020) Monthly consumer price index.

[online] Available at: https://www.gso.gov.vn/default.aspx?tabidr0 [Accessed 22 May 2020].

Nguyen, Q and Nguyen, T., (2013) Giáo trình Kinh tế lượng 2nd ed

Gujarati, D.N and Porter, D.C (2017) Basic econometrics Usa: Mcgraw-Hill/Irwin.

Greene, W.H (2018) Econometric Analysis Harlow: Pearson Education Limited.

Jackson, E.A (2018) Comparison between Static and Dynamic Forecast in

Autoregressive Integrated Moving Average for Seasonally Adjusted Headline Consumer Price Index SSRN Electronic Journal.

Sanson, N (2015) Forecasting the consumer price index of Ghana using exponential smoothing methods (2019) Mathematical Theory and Modeling.

Tiêu đề	Economics Forecast Report Forecast Vietnam Monthly Consumer Price Index From January 2019 To June 2019
Tác giả	Tạ Mai Linh, Vũ Thị Thu Nga, Hoàng Minh Quyên
Người hướng dẫn	PhD. Chu Thị Mai Phương
Trường học	Foreign Trade University
Chuyên ngành	International Economics
Thể loại	report
Năm xuất bản	2020
Thành phố	Hanoi

Định dạng
Số trang	38
Dung lượng	0,99 MB