Most economics textbooks don’t devote a lot of space to explaining economic statistics, but they usually provide a short discussion of gross domestic product, or GDP. In limited space, they typically are able to explain some of the important features of GDP—that it’s a comprehensive measure of income and expenditure, that it measures transactions at market prices, that it excludes non-market production, and that it isn’t a measure of well-being. But I’ve also observed that the typical American textbook definition of GDP misses one of its most important features and often leads to misinterpretation.
GDP is fundamentally a measure of value added in production
Let’s consider a simple example of production. Let’s say Henry has a food truck and sells Korean tacos. His produces and sells his output (Korean tacos), using purchased inputs of various food items, as well as other materials such as food containers and napkins. He also pays for fuel to operate the truck, and if he leases the truck, he makes lease payments. His contribution to GDP, or value added, is the value of his output (that is, the food he sells) during a given period of time, less the value of the goods and services that are used up in producing that output. GDP is equal to the sum of value added of all producers that are residents of the country or other economic area that is being measured, valued at the prices paid by purchasers.
GDP is part of the national economic economic accounts and is based on that accounting system. The national accounts include a number of identities, so there are other, equivalent definitions of GDP. In fact, there are three common approaches for measuring GDP—the production approach (sum of value added), the expenditure approach (sum of final expenditures), and the income approach (sum of income from current production). But most manuals focusing on national accounts start with the production approach. For example, the System of National Accounts 2008 (SNA), which is the international standard for national accounts, says, “The underlying rationale behind the concept of gross domestic product… is that it should measure the total gross value added from all institutional units resident in the economy.”
But that’s not the definition of GDP taught in most textbooks
Most economics textbooks, on the other hand, define GDP using the expenditure approach. For example, the fifth edition of Greg Mankiw’s Principles of Macroeconomics provides the following definition of GDP: “Gross domestic product is the market value of all final goods and services produced within a country in a given period of time.” The online macroeconomics course from Tyler Cowen and Alex Tabarrok’s MRUniversity uses a similar definition, though they substitute “finished” for “final” goods and services. Both sources also explain that the expenditure and income approaches are conceptually identical and that there may be a statistical discrepancy between them. However, the production approach is typically given short shrift.
Why is that a problem?
While what the textbooks are teaching isn’t technically incorrect—these are, after all, conceptually equivalent ways of measuring GDP—I do think that focusing on the expenditure approach instead of the production approach leads economics students to certain misinterpretations of GDP.
For example, I’ve often heard people say that the production of intermediate goods and services “isn’t counted” in GDP. But if we treat the production approach as the principal definition of GDP, we understand that producers of intermediate goods and services are generating value added—their output is almost always worth more than the value of the intermediate goods and services that went into its production—so the production (value added) of intermediate products is part of GDP.
Another very common confusion comes about because imports are a subtraction in the calculation of GDP using the expenditure approach. The familiar expenditure formula says that GDP is the sum of household final consumption expenditure, government final consumption expenditure, gross capital formation (or “investment”), and exports, less imports. This identity sometimes leads even credentialed economists to the fallacy that GDP can be boosted simply by cutting imports. Indeed, during the campaign Trump’s trade policy advisers Wilbur Ross and Peter Navarro (now Secretary of Commerce and Director of the National Trade Council) wrote a white paper on trade policy that made exactly that mistake. They argued that the effects of policies to eliminate the trade deficit could be measured directly, without using “any complex computer model,” as simply an addition to GDP in the amount of the reduction in the trade deficit. This analysis implausibly argues that all of the other components of the GDP expenditure formula can be held constant while eliminating the trade deficit.
The production approach, on the other hand, makes clear that imports are neither a subtraction nor an addition to GDP. They simply aren’t part of value added, so they aren’t included. Imports certainly may be correlated with GDP (value added). In some rare cases, a reduction in imports could, in fact, lead to a temporary shift to domestic suppliers, which could produce a temporary boost to GDP. But the more common relationship is that imported goods or services are used by domestic producers as intermediate inputs, which leads imports to move procyclically—increasing (along with GDP) during expansions, and decreasing (again, along with GDP) during recessions. The relationships between imports and GDP can’t be read directly from the GDP final expenditure formula.
Another quibble with the textbook definition is the term “final” (or “finished”) goods and services. The SNA never uses these terms because “final expenditures” (or “final uses”) is not necessarily a characteristic of a good or a service. For example, if a household buys an airline ticket for vacation travel, that’s counted as part of “final consumption expenditure,” even though the ticket for the business traveler sitting next to them will be counted as intermediate consumption. The service of airline travel is neither fundamentally “final” or “intermediate”—it all depends on who’s buying it.
Finally, when we get into certain technicalities about explaining the expenditure and income approaches, I think we ultimately have to go back to the production approach to explain why the measures take the form that they do. For example, why are imports subtracted in the expenditure approach? For imports of consumption or investment goods or services, the textbooks typically say that they need to be subtracted because they are counted in consumer spending or investment spending even though they are not produced domestically. But many imports are not targeted for final consumption or investment, but are intended for use as intermediate inputs (think crude petroleum or memory chips). To explain why these imports also need to be subtracted, we more or less have to go back to the explanation of GDP as a concept of value added—if imported intermediate goods weren’t subtracted, the expenditure approach would overstate the nation’s total value added. Similarly, under the income approach, national accountants remove capital gains from their measures of income. Again, the reason for doing that goes back to the concept of value added—capital gains are income derived from changes in prices of existing assets, rather than from the output of goods and services.
So while the textbook definitions of GDP aren’t awful, I do think they would be better if they: (a) discussed the production approach along with the expenditure and income approaches, and (b) explained that GDP is principally designed to be a measure of production or economic activity, and that using the production approach is perhaps the clearest way to present that concept.
Some really good points here. However, we can’t be too critical about the textbooks emphasizing the final expenditures approach when that is indeed the U.S. featured measure of GDP. Even academics would be hard pressed to find any information or estimates related to the production approach in the U.S. statistical system. I wonder if European textbooks emphasize the production approach since that is the featured measure in most of those countries.
Nice point about how the production approach clearly reveals the rationale for why all imports–not just imports consumed in final uses–must be subtracted to obtain the correct measure of GDP. However, it should be noted that in practice it is difficult to determine exactly how certain imports are being used. That is why many countries rely on indirect procedures to determine intermediate use of imports. (Much easier to simply subtract all imports!)
Thanks, Bob. These are good points.
While you’re right that the U.S. has always emphasized the expenditure approach in presenting GDP estimates, one change recently made by BEA is that the definition of GDP appearing on p. 4 of the GDP release starts with the production approach. (I admit that I sort of pushed that change.
More fundamentally, I think it’s easy for statistical agencies to give too much emphasis to how they collect and compile the data relative to what the data really mean.
If you import a good and then you sell it for a higher price than the one you pay for the import, you are not creating value?
Interesting and useful (especially to me, as an intro teacher).
I like to tell students that GDP *is* production, but we measure it in practice by expenditure (firms’ sales) and by income. (Is that right?) But we have to be very careful how we define “expenditure” and “income”, otherwise it will not be true that income=output=expenditure.
“You produced some goods and didn’t sell them? Ah! You must have sold them to yourself!”
“You earned income from capital gains? Ah! That doesn’t count as income!”
I tell them that accountants “cheat”, by weaseling out of any counterexample they can come up with, in order to make those accounting identities true, so that “income” and “expenditure” both mean “production”.
Nick – These are good points. I certainly wouldn’t suggest that teachers omit the expenditure and income approaches, but if I were teaching I’d start with the production approach. And when statistical agencies compile benchmark estimates of GDP, they in practice are combining and balancing all three approaches.
Your comment about the special definitions that national accountants use for expenditure and income is exactly what I was trying to describe when I said that these definitions ultimately go back to the production approach. Are national accountants cheating? I don’t know, but “income” and “expenditure” don’t mean the same thing in the context of national accounts as in the common vernacular. Then again, every field finds it useful to have its own specialized jargon.
I cannot but agree with you ! When I teach GDP in my French university class, I present it first as the “sum of production without double counting” (i.e. the sume of value added). Only after do I explain the two other approaches. It is obviously also the presentation that I make in the OECD manual “Understanding National Accounts”.
Thanks, François. I’m adding a link to your OECD manual, Understanding National Accounts,” which is an excellent resource for analysts who are interested in learning about national accounts. It takes an international perspective, with examples drawn from a number of countries.
I always made up a simple numerical example that illustrated the identity of the value-added approach and the expenditure approach.
1. GDP is the value of final goods and services produced in the economy during a given period. This method is known as the expenditure approach. It gives us GNP (Gross National Product) at market prices.
2. GDP is the sum of value added in the economy during a given period. The value added by a firm is the value of its production minus the value of the intermediate goods used in production. This method gives us GDP (Gross Domestic Product) at factor cost
3. GDP is the sum of incomes in the economy during a given period. This method gives us NNI (Net National Income) at factor cost.
Relationship between the three approaches.
– GDP at factor cost + Net Factor Income from abroad = GNP at factor cost.
– GNP at factor cost + Indirect Taxes – Subsidies = GNP at market prices
– GNP at factor cost – Depreciation = NNI at factor cost
These definitions are pretty close, but I’ll suggest some modified wording to coincide with the most recent standards. The following definitions are from the current version of the international standard (SNA 2008):
1. GDP is equal to the sum of the final uses of goods and services measured at purchasers’ prices, less the value of imports of goods and services. (expenditure approach)
2. GDP is the sum of gross value added of all resident producer units plus that part of taxes on products, less subsidies on products, that is not included in the valuation of output. (production approach) SNA 2008 recommends, but doesn’t require, that output should be measured excluding taxes on products, less subsidies on products, which is called valuation at “basic prices.” If a country measures value added at basic prices, then taxes on products, less subsidies on products needs to be added to the sum of gross value added to obtain GDP at purchasers’ prices.
3. GDP is equal to the sum of primary incomes distributed by resident producer units. (income approach)
4. SNA 2008 no longer uses “GNP”; the concept is now designated as “gross national income (GNI). GNI is equal to GDP less primary incomes payable to non-resident units plus primary incomes receivable from non-resident units (that’s essentially “net factor income from abroad”).
5. Value added at factor cost is no longer emphasized by the SNA; the current version emphasizes value added at basic prices. But if a country does measure value added at factor cost, the relationship is GDP = sum of value added at factor cost, plus all taxes on products, less all subsidies on products, plus all other taxes on production, less all other subsidies on production.
6. GNI less consumption of fixed capital (i.e., depreciation) = net national income (NNI)
Great points! It is sometimes hard to integrate each vision into a single concept of GDP, and it seems that considering at first (at least) the value-added approach could be a good start. Beside the OECD manual, do you have any other reference more suited for students learning economics for the first time? I teach introduction to economics based mainly on Mankiw’s books but I’d love to have more references with more comprehensive views.
Eurostat has a manual, Building the System of National Systems, that they use for training national accounts statisticians. While the whole manual would be too much for an introductory economics class, chapter 3 on “Basic concepts” would be good. Or if you wanted to emphasize the U.S. national accounts, there are some materials on BEA’s website, like this primer.
Statistics Canada also has a manual that it uses for training purposes and for informing the general public about how Canada’s national accounts are structured.
I’ve long thought that our approach to measuring GDP was dependent on the data we have and much of that comes from tax data. Our most important taxes are income taxes and that influences what we measure. That is a major reason we measure final demand rather than production at each stage as in an input-output approach. Maybe if we used a value added tax (VAT) rather than an income tax we would have much better and timely data on what was produced at each stage of production and our approach to measuring GDP would be very different.
You’re right that there are three conceptually equivalent approaches to measuring GDP. Countries choose to estimate GDP using any or all of the approaches, depending on the available data. For the United States, the earliest approach was the income approach, reflecting the availability of data on national income, and it was followed by the expenditure approach. The production approach was the last one adopted in the US, and it still isn’t as well known here as the other approaches. Other countries, however, often emphasize the production approach. Most countries with advanced statistical systems tend to utilize more than one approach and attempt to reconcile them, at least for the benchmark estimates.
But the main point of my post is not about how GDP is estimated, but about how we teach and understand GDP. Teachers (at least in the US) have not emphasized the production approach. I think students are likely to have less confusion if we teach the production approach as the primary definition of GDP, regardless of whether that’s the primary approach used by the statistical office for compiling the estimates.
I agree with you that the production approach should be the basic approach: it emphasises what is produced in the country rather than what is produced by residents of the country. Nowadays, there is not much of a difference in the two numbers expect for a few countries.
Mr. Moulton, and others on this string. Is it any wonder that we Economics students are confused by a relatively simple, albeit critical, concept like GDP when the experts and Professors can’t agree on the definition? You have succeeded in making this subject so complicated that it is virtually impossible to unravel what you are trying to clarify. Speak English, please. I am one of the older people in my class at JWMI. I have been hearing about and watching GDP my whole life. I, along with millions of citizens, should have a reasonably cogent idea of what GDP is. However, this entire string looks to me as if it is academics who are trying to have an academic discussion instead of actually clarifying the concept for the people who are most affected by it.
I am loving the blog.
This post got me thinking. Both the SNA and, if I recall correctly, the System of Environmental Economic Accounts attribute value added to institutional units, which in the SEEA are defined as:
An institutional unit is an economic entity that is capable, in its own right, of owning assets, incurring liabilities, and engaging in transactions and other economic activities with other entities.
I think this is the same definition as in the SNA. In practice, this means Households, Governments, Corporations and NPISHs.
So what happens when an ecosystem is granted the same rights, obligations and liabilities as a person?
It seems to me that this opens the door to attributing value added to the environment. I would be interested to hear your thoughts.
Attributing value added to the environment is certainly an attractive idea, since we all understand that production requires not only traditional inputs of labor and capital, but also the inputs derived from environmental assets. I think the tricky aspect is valuation. In particular, because we may not pay for environmental inputs like clean air and water, there may not be a standard way of valuing these inputs the way we do for labor and capital. There may be a way to value them in terms of opportunity costs (the value of goods and services forgone in order to maintain the quality of the air and water), but it certainly would be trickier to measure than traditional value added or GDP.
Thanks Brent. I agree with you regarding valuation. It would indeed be tricky (and controversial), and perhaps the largest impediment to a comprehensive set of accounts.
It is also possible that these natural assets with personhood will engage in real transactions. If that happens, it will be interesting to see how these transactions are dealt with in the national accounts.
Great post. I have been emphasizing this in my undergraduate macro classes as it is important to distinguish between concepts and measurements to have a solid understanding of both.
It is actually one of my pet peeves about shortcuts taken undergraduate macro books. I wonder if anyone reading this thread has found an undergraduate textbook that defines GDP as production and starts its measurement with the approach more closely related to its definition.