Analisis Konsumsi Energi di Amerika Serikat

Purpose – This paper examines the effect of various energy prices and population sizes on energy consumption in the United States. Methods – This paper begins the analysis by conducting integration tests on the existing variables. The test produces a mixture of variables I(0) and I(1), so the model chosen is the ARDL model. The bound test indicates cointegration, and the analysis is continued by estimating the conditional ECM model. Findings – Both the ARDL and conditional ECM model estimates find that the prices of all components of energy consumed, as well as population size, are good explanations of energy consumption behavior. Significant autoregressive factor up to lag 3 indicates that energy consumption behavior is strongly influenced by energy consumption in the past. Coal prices, crude oil prices, and natural gas prices only affect contemporaneously. Population size affects both contemporaries and three lags in the past. The estimation results with the conditional ECM in the long-term model estimation segment confirm the results from the ARDL model. The short-term ECM conditional estimation results suggest that all variables significantly affect the dependent variable. Implication – The lag effect of the dependent variable implies that energy consumption in the present is highly dependent on consumption in the lamp life. The historical effect of various energy source prices means that consumers only consider current prices in deciding how much energy to consume. Originality – This research contributes to modeling the influence of various variables on energy consumption in the United States. Abstrak Tujuan – Paper ini menguji pengaruh berbagai harga energi serta ukuran populasi terhadap konsumsi energi di Amerika Serikat. Metode – Paper ini mengawali analisis dengan melakukan uji integrasi atas variabel-variabel yang ada. Uji tersebut menghasilkan campuran antara variabel-variabel I(0) dan I(1), sehingga model yang terpilih adalah model ARDL. Uji kointegrasi dengan bound test menyarankan keberadaaan kointegrasi sehingga analisis dilanjutkan dengan mengestimasi model conditional ECM. Temuan – Estimasi model ARDL dan conditional ECM menemukan bahwa harga semua komponen energi yang dikonsumsi serta ukuran populasi menjadi penjelas yang baik terhadap perilaku konsumsi energi. Faktor autoregressive signifikan sampai lag 3 menandakan bahwa perilaku konsumsi energi sangat dipengaruhi oleh konsumsi energi di masa lampau. Harga batubara, harga minyak mentah, dan harga gas alam hanya berpengaruh secara sejaman. Ukuran populasi memiliki pengaruh baik sejaman maupun tiga lag di masa lampau. Hasil estimasi dengan conditional ECM, dalam segmen estimasi model jangka panjang, mengkonfirmasi hasil dari model ARDL. Hasil estimasi