A mercantilist misreading of GDP figures

Gross Domestic Product pops up everywhere in the news. In the summer of 2016, for example, the news that Ireland’s GDP had increased by 26.3 per cent in 2015 had people scratching their heads. Last September’s issue of The Economist raised the perennial question of whether GDP figures released by the Chinese government are reliable. Then Japan began revamping its GDP calculations after some contradictions appeared in official statistics. This is not counting the routine articles that follow the periodic releases of official estimates.

There is much to be debated in relation to the use and misuse of GDP. But among the many forms of GDP misuse, one is obvious, frequent, and dazzling.

One of the main accounting identities used in national income accounting and basic economics courses states that GDP is equal to the sum of consumption, investment, government expenditures (excluding transfers) and exports. In other words, it is the sum of domestic production flows to domestic consumers, domestic purchasers of investment goods, domestic governments, and foreign importers. In yet other words, the production side of GDP is equal to its expenditure side: everything that is produced is purchased.

This is an accounting identity, which means that it is true by definition and cannot be false. It is necessarily true because anything produced that is not purchased by domestic consumers, businesses, governments and foreign importers will pile up in inventories. Inventories are a form of unintentional business investment. Investment, in GDP numbers, is defined as including such inventories as well as fixed capital investment. This is how accounting identities are necessarily true in the real world: a residual adjusts as a matter of definition.

We could write our accounting identity as:

GDP = consumer expenditures + business investment + government expenditures + exports

However, this is true provided that we took consumer expenditures, business investment, and government expenditures as including only goods and services produced domestically. As its name indicates, gross domestic product is made of domestic production only.

In the statistics that are actually collected, however, consumer expenditures (normally represented by C), business investment (I), and government expenditures (G) include some imported goods and services. The Chinese-made football you bought at Sports Direct was captured in C; the printing press a newspaper company bought from Germany was part of I; and the salary of the foreign consultant hired by the government was included in G. Consequently, it would not be correct to write our accounting identity as GDP = C + I + G + X (where X represent exports), because spending on imports is captured in the right hand side of the equation (in C, I and G).

To solve this statistical problem, the accounting identity is written as:

GDP = C + I + G + X – M

The term –M cancels the imports that are hidden in C, I, and G. It does not mean that imports are a subtraction from national income.

It is easy to be misled. The problem is compounded by the fact that X – M is often grouped inside parentheses so that the accounting identity is remembered as:

GDP = C + I + G + (X – M)

For the non-expert, the last equation can easily suggest that (X – M) is the balance of trade. This interpretation error is further encouraged by experts who call (X – M) “net exports”. To repeat, it is only “net exports” if you forget that –M is used only to cancel the imports that, in the process of data collection, were included in C, I and G. In other words, the term –M is a statistical trick.

Imports are not deducted from GDP. They cannot reduce GDP because, by definition, they are not part of it. Buying a Chinese football from Sports Direct does not reduce national income. It is deducted in the (X-M) term to prevent the consumption of such footballs from increasing national income as a result of their inclusion in C.

Or, to put it another way, if Donald Trump were to ban the import of Chinese footballs, it might reduce M. (This would make US consumers poorer as they would have to buy more expensive footballs, but I am considering only accounting matters here). The reduction of M would certainly not increase GDP because, arithmetically, the M component of C would also be reduced, leaving GDP unchanged.


Pierre Lemieux is an economist affiliated with the Department of Management Sciences at the Université du Québec en Outaouais. His latest book is ‘Who Needs Jobs? Spreading Poverty or Increasing Welfare’ (Palgrave Macmillan, 2014). A longer version of this article was published in the Cato Institute’s ‘Regulation’ magazine.

Leave a Reply

Your email address will not be published. Required fields are marked *