Прогнозування ввп України на основі даних різної частоти
Date
2017-03-28
Journal Title
Journal ISSN
Volume Title
Publisher
Видавництво Львівської політехніки
Abstract
Досліджено залежність ВВП України від курсової динаміки гривні, спрогнозовано
квартальне та річне значення ВВП України за допомогою модифікацій моделі MIDAS,
яка дає змогу працювати із вхідною інформацією різної частоти. Для розрахунку
параметрів моделі та прогнозування значень ВВП використано щомісячні статистичні
дані офіційного курсу гривні до долара США та квартальні і річні показники
фактичного ВВП за попередні періоди від 2001 р. Оцінена якість запропонованих
моделей та отримано висновок про їхню достовірність і доцільність практичного
застосування. Запропоновані методи реалізовано у програмному середовищі R.
GDP statistics is usually quarterly and with a significant delay, and the data of many other economic indicators (average wages, unemployment, exchange rates, etc.) are monthly or have an even higher frequency. Such indicators often carry important information about the current state of the economy and it is important to use this data with a high frequency to obtain qualitative short-term forecasts. That is why methods that use mixed frequency data are becoming increasingly popular in predicting current systemstates and in short-term forecasting. The article examines the dependence of Ukraine’s GDP on the exchange rate dynamics of the hryvnia, the quarterly and annual GDP of Ukraine is predicted with the help of modifications of the MIDAS model, which allows to work with input information of mixed frequencies. Monthly statistics of the official exchange rate of the hryvnia against the US dollar and quarterly and annual GDP values for the previous periods since 2001 were used to calculate model parameters and forecast GDP values. To build the model, data on the exchange rate dynamics of the hryvnia against the US dollar were used for the period from January 2001 to December 2016, gross domestic product in actual prices from the 1st quarter of 2001 to the 4th quarter of 2016 and gross domestic product in actual prices from 2001 to 2016 year. Data source – statistical information from the site of the National Bank of Ukraine and the State Statistics Service of Ukraine. The model is implemented in the R environment using the midasr package. The stationarity test was carried out using one of the most common criteria – the augmented Dickey-Fuller test, the hypothesis of their stationarity was confirmed by the results of testing the logged time series. When constructing a MIDAS model with constraints on parameters, the normalized exponential Almon lag coefficients are used. The hAh restriction test on coefficients of MIDAS regression was performed, the quality of the constructed models was assessed. We get the predicted value of GDP for the 1st quarter of 2017 in actual prices – 468654.2 million UAH and 2579342 million UAH – the annual GDP for 2017. According to the operative assessment of the State Statistics Service of Ukraine GDP in the I quarter of 2017 compared to the previous quarter (taking into account the seasonal factor) decreased by 0.3%, and compared to the first quarter of 2016 – increased by 2.4% %. That is, the forecast adequately reflects the actual data and was received quarterly faster. In general, MIDAS models allow to practically combine the annual and quarterly GDP statistics and the monthly official dollar exchange rate to the Ukrainian hryvnia, as a result of which it is possible to analyze the dynamics and interrelation of the economic indicators.
GDP statistics is usually quarterly and with a significant delay, and the data of many other economic indicators (average wages, unemployment, exchange rates, etc.) are monthly or have an even higher frequency. Such indicators often carry important information about the current state of the economy and it is important to use this data with a high frequency to obtain qualitative short-term forecasts. That is why methods that use mixed frequency data are becoming increasingly popular in predicting current systemstates and in short-term forecasting. The article examines the dependence of Ukraine’s GDP on the exchange rate dynamics of the hryvnia, the quarterly and annual GDP of Ukraine is predicted with the help of modifications of the MIDAS model, which allows to work with input information of mixed frequencies. Monthly statistics of the official exchange rate of the hryvnia against the US dollar and quarterly and annual GDP values for the previous periods since 2001 were used to calculate model parameters and forecast GDP values. To build the model, data on the exchange rate dynamics of the hryvnia against the US dollar were used for the period from January 2001 to December 2016, gross domestic product in actual prices from the 1st quarter of 2001 to the 4th quarter of 2016 and gross domestic product in actual prices from 2001 to 2016 year. Data source – statistical information from the site of the National Bank of Ukraine and the State Statistics Service of Ukraine. The model is implemented in the R environment using the midasr package. The stationarity test was carried out using one of the most common criteria – the augmented Dickey-Fuller test, the hypothesis of their stationarity was confirmed by the results of testing the logged time series. When constructing a MIDAS model with constraints on parameters, the normalized exponential Almon lag coefficients are used. The hAh restriction test on coefficients of MIDAS regression was performed, the quality of the constructed models was assessed. We get the predicted value of GDP for the 1st quarter of 2017 in actual prices – 468654.2 million UAH and 2579342 million UAH – the annual GDP for 2017. According to the operative assessment of the State Statistics Service of Ukraine GDP in the I quarter of 2017 compared to the previous quarter (taking into account the seasonal factor) decreased by 0.3%, and compared to the first quarter of 2016 – increased by 2.4% %. That is, the forecast adequately reflects the actual data and was received quarterly faster. In general, MIDAS models allow to practically combine the annual and quarterly GDP statistics and the monthly official dollar exchange rate to the Ukrainian hryvnia, as a result of which it is possible to analyze the dynamics and interrelation of the economic indicators.
Description
Keywords
ВВП, модель, MIDAS, дані різної частоти, прогноз, метод Альмона, GDP, model, MIDAS, mixed frequency data, forecast, Almon’s method
Citation
Зомчак Л. М. Прогнозування ввп України на основі даних різної частоти / Л. М. Зомчак, В. Т. Марусіна // Вісник Національного університету “Львівська політехніка”. Серія: Менеджмент та підприємництво в Україні: етапи становлення і проблеми розвитку. — Львів : Видавництво Львівської політехніки, 2017. — № 875. — С. 278–292. — (Управління національним господарством).