An econometric forecast of China’s pet industry using ARIMAX, cross-country comparison and multi-year scenario projections.
35p
APMCM paper
03
macro regressors
2027
forecast horizon
01 / Overview
What this package actually contains.
The APMCM 2024 paper forecasts China’s pet population and industry development using ARIMAX and cross-country comparisons. Exogenous variables include disposable income, consumer confidence and CPI, with scenario projections through 2027.
Objectives
01Test stationarity and prepare the pet-population series for time-series modelling.
02Quantify relationships with income, confidence and inflation variables.
03Compare China with the United States, France and Germany and project future demand.
Competition narrative
Problem, model, result, paper.
The page foregrounds the modelling decision and result before implementation details.
Stage 01
Macro data
Pet counts, disposable income, confidence and CPI by year.
01
Stage 02
ADF testing
Check stationarity and choose differencing.
02
Stage 03
ARIMAX
Forecast cat and dog populations with exogenous regressors.
03
Stage 04
Country comparison
Compare growth patterns across four markets.
04
Stage 05
Scenario output
Projections to 2027 and industry implications.
05
02 / Package highlights
What this work demonstrates.
Combines disposable income, consumer confidence and related economic indicators.
Projects pet populations and industry development through 2027.
Compares China with the United States, France and Germany.
03 / Model
The modelling choices behind the result.
01
Stationarity and exogenous variables
ADF tests were used to inspect whether the series needed differencing. The model then combined an ARIMA structure with external regressors, allowing income, confidence and inflation to influence the forecast rather than treating pet counts as an isolated series.
Cat and dog populations were modelled separately.
Exogenous variables were scaled before fitting.
Fitted values were plotted against actual observations to inspect model behaviour.
02
Projection and comparison
Future values for the exogenous variables were supplied as a scenario. The fitted models produced multi-year projections, and a second workflow compared China with the US, France and Germany using exponential smoothing on the country-level series.
04 / Key code
Selected modelling code.
Only verified source excerpts and pseudocode taken from the project papers are shown.
Verified project codePython
ADF test and ARIMAX fit
Original Python excerpt for stationarity testing and exogenous-regressor modelling.
problem1.py
from statsmodels.tsa.stattools import adfuller
from statsmodels.tsa.statespace.sarimax import SARIMAX
def adf_test(series):
result = adfuller(series)
print('ADF Statistic:', result[0])
print('p-value:', result[1])
def arimax_model(target, exog_vars, df):
model = SARIMAX(
df[target],
exog=df[exog_vars],
order=(1, 1, 1),
seasonal_order=(0, 0, 0, 0),
)
return model.fit(disp=False)
Verified project codePython
Scenario forecast
Original Python excerpt projecting cat and dog populations.
