Gross Domestic Product (gdp)

[pic 1]

[pic 2][pic 3][pic 4][pic 5]

Introduction – background analysis and methods

Gross domestic product (GDP) is a measure of the output produced by an economy in a period of time. Real GDP estimates are commonly used to determine the economic performance of a whole country or region, and to make international comparisons. Moreover, its relevance lays on making predictions on the future path of a country and provide a better insight into economic decisions.

Since our study focuses on the Portuguese economy, it is worth taking a few steps back to look at the evolution of the GDP real growth rate over the years. Data shows a considerable fluctuation as a result of economic expansion or contraction, which is driven not only by internal conditions, but also by foreign intervention.

 GDP can be measured by the expenditure approach, among others. GDP under this method is calculated by summing up private consumption (C), investment (I), government spending on goods and services (G), and net exports (NX), which is equal to exports (X) minus imports (M) of goods and services:

With the purpose of estimating the real GDP growth for the Portuguese economy in the 3rd quarter of 2017 we used economic proxies. A proxy variable serves in place of an unobservable or immeasurable variable. Thus, it is worth explaining the proxies we used to estimate each of the GDP´s components:^[1]

Private Consumption is divided into goods and services. The former, in turn, can be split into durable and non-durable goods. A durable good is one that does not quickly wear out, like furniture, jewellery or automobiles, for this reason it makes sense to use proxies as vehicles sales or loans conceded to households for consumption, since people usually need to ask for credit to buy a car. Furthermore, indices that measure sales of technological equipment, books, clothing and footwear are helpful to understand if households are spending more in durable goods. On the other hand, a non-durable good is one that has a short life cycle, such as fuel and food, as so the retail confidence indicator and the index for sales in retail show how families have changed their consumption in non-durable goods. More specific indices, such as measures of expenditure in food, drinks, tobacco and fuel, are important to predict how the larger components of non-durable goods consumption evolved.

In addition, other proxies that reflect a more general picture of the economic context can be used to improve the estimation of both private consumption components. For instance, the consumer confidence indicator, which reflects households’ plans to make bigger purchases and the assessment on the financial situation of households in the last 12 months and an expectation for the next ones, as well as on the own country economic situation and evolution of unemployment.

Investment is the amount that households and businesses spend on items used to increase the capacity of production of more goods and services. Within this item are accounted not only investment in capital equipment and inventories for businesses, but also households' purchases of new housing.

As a result, sales of heavy vehicles and indicators related to construction and production are relevant proxies as they reflex the assessment on production decisions that result in its investment counterpart. Examples of such indices are services, industry, construction and manufacturing industry confidence indicators and evaluations of stocks of finished products, global demand for manufacturing industry and the performance of the construction sector. Moreover, the perspective on production for the next 3 months and the climate indicator, which reflects the conjuncture of the industry, trade, private and public construction and services, results as mix of all previous confidence indicators giving a bigger picture of the economic context.

More related with new housing, there is the cement sales indicator, the number of issued building permits for new houses and loans conceded to households for habitation, which reflect the decisions of families on investing in a new home.

Government Spending results of Government’s consumption of goods and services and investment in capital. This item corresponds to the sum of purchases of goods and services, wages of public servants and investment made by government. It is worth mentioning that interest’s payments and transfers, such as welfare and social security are not counted as government purchases. For this item we used the data published by the Government for the 3rd quarter of 2017 regarding: personnel expenses, purchase of goods and services, and investments made by the public administration.

Exports correspond to goods or services that are produced domestically and sold to foreigners. Portuguese exports depend on the economic performance of the main Portuguese trading partners, especially Spain, France, Germany, the United Kingdom and the US, among others. For this item we resorted to the predictions for the GDP of the Portuguese trading partners, namely US, Germany, Spain, France and United Kingdom. As a complement the economic sentiment and consumer confidence indicators of the Euro zone and European Union were used.

Imports correspond to goods or services brought into one country from another one. It should be subtracted from GDP because it represents foreign rather than domestic production. In order to predict imports, we used the predictions made on Consumption, Investment and Government Spending, as these macroeconomic variables represent domestic demand.

After deciding on the proxies to use, the growth rate of consumption for durable and non-durable goods, investment, exports and imports was estimated using OLS regressions. It is worth mentioning that, for all the proxies for which the data for September was not available, we assumed that they grew at the same growth rate as in the previous period.

Afterwards, the growth rates obtain were used to predict the variables real value in the 3rd quarter of 2017, through applying the growth rate on the previous quarter value. Using the value set by the Government for public administration in the relevant quarter and applying the formula for GDP computation (1), an estimate for GDP in the 3rd quarter of 2017 was obtained. Finally, the achieved value was used to compute both the quarter and homologous growth rate.

Results and Discussion  

        When looking at each of the GDP´s components, it was important to run some regressions with different proxies to test how well these explained the respective dependent variable. By looking at the adjusted R² value for the different combinations of proxies, we know that the closest it is from 100%, the better is the fit of the model. Although the adjusted R² value is a good indicator of the fitness of a model, we noticed that some of the models with the highest R² presented coefficient values that contradict the economic theory. Given that, we decided to run regressions with data for a longer period of time, in order to have a larger number of observations. Thus, we ended up with two different regression types: model 1 – with a larger number of observations and less proxies, and models 2 and 3 – with a smaller number of observations and more proxies, with the addition of dummies, since the data showed signs of seasonality. Despite the fact that models 2 and 3 always resulted in a higher adjusted R² value, they contradicted in almost all of the coefficients the expected signs. Hence, model 1 was adopted to explain each of the components.