Chapter 2 Gross Domestic Product

François Lequiller

What is gross domestic product (GDP)?

How do we calculate it?

How can we compare GDP through time and across countries?

What is left out of the calculation?

How reliable are GDP statistics?

How do they guide policymakers?

In this chapter, our aim is to give a definition of gross domestic product—often referred to as GDP. GDP is the most frequently used indicator in the national accounts.

The results provided by the national accounts are now such a familiar part of everyday economic information that there is a tendency to forget how extremely ambitious the original project was—and still is. It is no accident that the two major creators of modern national accounts (Simon Kuznets of the United States and Richard Stone of the United Kingdom) were both awarded Nobel prizes for economics.

Other Nobel Prize winners—Hicks, Meade, Samuelson, Tinbergen and Leontief—also played a major role in national accounting development.

Read the Nobel prize lectures written by Kuznets and Stone, the fathers of national accounting:

Kuznets was awarded his prize in 1971 “for his empirically founded interpretation of economic growth which has led to new and deepened insight into the economic and social structure and process of development”.1

Stone followed in 1984 “for having made fundamental contributions to the development of systems of national accounts and hence greatly improved the basis for empirical economic analysis”.2

Despite its importance, national accounting is a fairly recent invention, with the first quarterly measurement of the economy being done only in the 1920s, by an economist called Colin Clark. Angus Maddison, who led a project that has estimated GDP for many countries 1,000 years back in history, pointed out that economic growth was much slower in the past “and therefore seemed irrelevant or uninteresting”.3 As Diane Coyle points out in her book, GDP: A Brief but Affectionate History, serious attempts to measure the economy consistently through time, and in a consistent way across countries, started with the rapid economic growth of the nineteenth century.4

But it wasn’t until the Great Depression and the Second World War made state planning of how to allocate resources essential, that the concept of GDP—with its quarterly reports and detailed breakdowns—and its systems of accounting for inflation, were formalised and, eventually, standardised.

ONS resource

Investigate the ONS GDP estimates.

Today, British policymakers, firms, investors, banks and journalists wait impatiently every month for the ‘GDP monthly estimate, UK’ and every three months for the ONS publication called ‘UK GDP first quarterly estimate’. They need to wait for 40 days after the end of the quarter while the estimate is calculated. A quarterly GDP estimate is one of the most keenly awaited statistics in every country, not just the UK, because of the information it gives about the state of the economy.

But the GDP estimate tells more than one story. On 11 February 2022, the first estimate of Quarter 4, 2021, from the beginning of October to December, came out. “UK fastest growing G7 member in 2021, but ‘middle of the pack’ for pandemic recovery”, said The Guardian. “UK economy rebounds with fastest growth since WW2”, reported BBC News. “Brexit Britain roars as GDP bounces back from pandemic”, The Express decided. “Record GDP growth does not reflect the reality on the ground”, according to the British Chambers of Commerce.

Figure 2.1 Quarterly GDP growth

Real GDP growth, UK, Quarter 1 (Jan to Mar) 2008 to Quarter 4 (Oct to Dec) 2021

Office for National Statistics – ‘GDP quarterly national accounts’

ONS commented that, “UK gross domestic product (GDP) is estimated to have increased by 1.0% in Quarter 4 (Oct to Dec) 2021. Compared with the same quarter a year ago, GDP increased by 6.5%. Following the large 9.4% fall in 2020 because of the initial impact of the coronavirus (COVID-19) pandemic and public health restrictions, UK GDP saw an annual rise of 7.5% in 2021”.5

All in all, this quarterly publication epitomises the concepts that will be discussed in this chapter:

While the growth in 2021 Q4 was 1.0%, and this may be seen as small, it must never be forgotten that national accounts variables are measured in hundreds of billions of pounds.

2.1 What is gross domestic product (GDP)?

gross domestic product (GDP)
GDP combines in a single number, and with no double counting, all the output (or production) carried out by all the firms, non-profit institutions, government bodies and households in the UK during a given period, regardless of the type of goods and services produced, provided that the production takes place within the country’s economic territory. It is calculated quarterly or annually, but it can also be calculated monthly.

In the case of the UK, the annual gross domestic product (GDP) at current prices for 2021 was estimated at £2,317 billion. Quarterly GDP, which is one fourth of the annual one (see Section 2.5.1), was around £580 billion, so that 1% of quarterly GDP (the growth rate in 2021 Q4) amounts roughly to £5.8 billion. To put this in context, this corresponds to roughly the quarterly earnings of over one million people, a substantial number.

2.1.1 The standards for GDP

The standards governing national accounts are enshrined in two international reference manuals: the System of National Accounts 2008, which is recognised globally, and its European version, called the European System of Accounts 2010. The global manual is co-signed by the five major international economic organisations: the United Nations, the International Monetary Fund, the OECD, the World Bank and the European Commission. The European manual is largely compatible with the global manual and includes additional useful details. These manuals have contributed substantially to improving the international comparability of data.

As well as the standards, both the UN and Eurostat provide guidance on how the statistic should be compiled and interpreted:

2.2 Measuring output

Measuring a country’s total output is not a simple matter; therefore, national accountants have had to devise innovative methods of calculation.

2.2.1 The problem of double counting

The output of a single firm can be measured fairly easily. In the case of a firm making pasta, for example, it can be measured as tonnes of pasta made during the year, or, if we multiply the number of tonnes by the price of the pasta, by the amount of output valued in pounds. But we shall see that it makes little sense to add together the output measured in pounds from all firms to arrive at a macroeconomic figure. That is because the result of this calculation depends heavily on the way the firms are organised.

Take again the example of the pasta manufacturer and compare two different production scenarios in a given region. Suppose that, in the first year, there is only one firm—let us call it National Food Ltd—that makes both the pasta and the flour used to make the pasta. Its output amounts to £100,000 corresponding to 100 tonnes of pasta, with each tonne valued at £1,000 pounds. Now suppose that, the following year, National Food is split into two, with new firm, National Flour Ltd, specialising in making flour and selling £30,000 worth of it to a firm called National Pasta Ltd, which uses it to carry out the final production of (not surprisingly) pasta. As the factories and workforces of the firms are unchanged, National Pasta makes the same quantity of pasta as National Food did the year before (100 tonnes) and at the same price (£1,000 per tonne).

National Food National Pasta
Year 1 output £ 100,000
Year 2 output £ 30,000 £ 100,000

Figure 2.2 Two pasta firms

Output of two hypothetical firms in years 1 and 2

In the first year, the output in this region will be worth £100,000; in the second year, the value of total output could be the sum produced by National Flour (£30,000) and that of National Pasta (£100,000), resulting in a total of £130,000. But it would clearly be absurd to use this total as our macroeconomic indicator of activity in the region. It shows an increase of 30% (130,000/100,000 = 1.30, often written as +30%, or more simply 30%), when in fact no change at all took place at the macroeconomic level. The same quantity of pasta was produced at the same price. All that changed was the legal and commercial organisation of the firms.

value added
The value of output minus the value of all inputs (called intermediate goods).

This problem generated the national accountants’ innovative idea of calculating the contribution of each firm not as its output, but as its value added. This expression is profound since it consists of measuring the value that the firm adds to that of the firms that supply its inputs.

Let us consider the pasta example again. Compared with the situation in the first year, when there was only National Food, the value added by firm National Pasta is not equal to £100,000. That is because National Pasta buys £30,000 of flour, whereas previously it had made this flour itself and did not count this as output. Therefore, the national accounts system proposes calculating the value added of National Pasta as £70,000 (100,000 – 30,000). In other words, the value of the firm’s output minus the value of the products used to carry out its production during the period.

intermediate goods
Goods, created by other firms, consumed by firms in the production of their output. The value of intermediate goods is subtracted from GDP to avoid double counting.

The products consumed in the production process during the period are known as intermediate goods. By deducting their value from that of output, one eliminates the double counting that occurred earlier when summing of the output of National Flour and National Pasta. In the second year, the output of flour was in fact counted twice: once in the value of the output of firm National Flour (£30,000) and a second time in the value of the output of National Pasta (whose £100,000 of output in fact includes the value of the flour bought and used in the production process).

2.2.2 Summing the value added

If one applies this same reasoning to all firms, calculating for each its value added, it is then possible to add together the value added of each firm, without double counting. The result will be an indicator that is independent of the way firms are organised. This is illustrated in the following table, which includes the farm that produced the wheat from which the flour was made. For the sake of simplicity, let us assume the farmer uses no intermediate consumption; he obtains his wheat solely from his labour and machinery, without buying seeds or fertilisers. As can be seen from Figure 2.3, the sum of the output of each unit changes, but the sum of the value added of each unit remains equal to £100,000, regardless of the pattern of organisation.

Year 1 Farmer National Food
Input Labour + machinery + wheat Labour + machinery + wheat
Output £ 10,000 £ 100,000
Intermediate consumption 0 £ 10,000
Value added £ 10,000 £ 90,000
Year 2 Farmer National Flour National Pasta
Input Labour + machinery Labour + machinery + wheat Labour + machinery + flour
Output £ 10,000 £ 30,000 £ 100,000
Intermediate consumption 0 £ 10,000 £ 30,000
Value added £ 10,000 £ 20,000 £ 70,000

Figure 2.3 The sum of the value added

Value added by farmer and two hypothetical firms in years 1 and 2

The way in which GDP estimates are compiled is a good illustration of the three essential rules followed by national accountants when they use statistics in the microeconomy to compile macroeconomic indicators:

  • Avoid double counting.
  • Devise aggregates that are economically significant (aggregates whose value is independent of non-economic factors).
  • Create indicators that are measurable in practice.

How it’s done Summing value added

This is why GDP is defined as being equal to the sum of the value added of each firm, government institution and producing household in a given country. By convention, and as explained later, these value-added terms are referred to as gross value added or GVA. Written formally, this is:

\[\text{GDP} = \text{GVA}_{i}\]

for each firm i. We can then rewrite this:

\[\text{GDP} = {(\text{Output} - \text{Intermediate consumption})}_{i} = \text{Output}_{\text{i}} - \text{Intermediate consumption}_{i}\]

This means that GDP is the sum of every firm’s output less the sum of every firm’s intermediate consumption.

The composite formula for GDP (known as an aggregate) constitutes a macroeconomic indicator of output that is independent of the pattern of the organisation and avoids double counting.

2.2.3 Calculating real GDP and real GDP growth

real GDP
Nominal GDP, adjusted for the effects of changes in prices. Also known as GDP in volume terms.
nominal GDP
A measure of the market value of the output of the economy in a given period. Also known as GDP in current prices.
real GDP
Nominal GDP, adjusted for the effects of changes in prices. Also known as GDP in volume terms.
The value of the set of goods and services that a household actually enjoys over a given period, for example, spending on food, recreational activities and clothes.

Refer to Figure 2.1. The comments of the ONS refer not to GDP growth as such, but to the growth of real GDP. What does this expression mean?

A key distinction discussed in this chapter is between GDP and other aggregates measured in the prices at the time of measurement, and the same aggregates adjusted to remove the effect of price changes over time. The latter are intended to identify changes in quantities having adjusted for the effects of changes in prices.

In this book, we will use the following terms:

Economics often attempts to distinguish between growth in the volume of output and consumption and its price. In macroeconomics, a key concern is differentiating changes in national accounts aggregates at nominal prices that stem from a change in the quantities from those that are the result of a change in their prices.

Back in the factory of National Food and its subsidiaries, let us suppose that the output of pasta is worth £100,000 in the first year and £110,000 in the second. An economist will immediately want to know if this 10% growth (which may be described as “nominal” or “in value” or “at current prices”) is due to an increase in the quantity of pasta, or to an increase in its price. An increase in quantity is good news, while an increase in prices, inflation, (see Chapter 1) tends to be bad news.

Keeping in mind the aim of separating the good growth (the quantities) from the bad growth (inflation), national accountants have developed sophisticated methods for separating out movements in nominal GDP into two components:

Appendix 5 (“Deflators and the national accounts”) discusses the topic of deflation in more detail. A brief account of deflation in regard to GDP is given in the next section.

The GDP deflator is essentially a price index with weights that are the proportionate quantities of all of the goods and services included in GDP.

How it’s done Converting nominal to real GDP

Recall that the £100,000 0f pasta production mentioned earlier equals 100 tonnes of pasta (the quantity) multiplied by £1,000 (the price per tonne). Similarly, in the national accounts, GDP also equals the quantity multiplied by the price of that quantity, which is:

GDP in nominal prices = real GDP × GDP deflator

Putting matters another way:

\[\text{real GDP} = \text{nominal GDP} \over \text{GDP deflator}\]

The same equation can also be expressed as a growth rate. This is usually regarded as a fundamental equation in national accounting. Over any period:

\[\begin{align*} 1 + \text{percentage growth rate}\\ \text{of GDP in nominal terms} &= \\ &(1 + \text{percentage growth rate of GDP in real terms})\\ \times &(1 + \text{percentage growth rate of the GDP deflator}) \end{align*}\]

Since we can observe what is happening to prices and quantities, this equation is an ideal tool to determine how much of an observed change in nominal GDP is due to a change in real GDP, and how much to a change in prices.

We can construct, by appropriate weighting of those prices, a GDP deflator as the price corresponding to GDP, and so calculate the changes in prices over time. Inserting this into the above formula will give us the change in real GDP.

Spending on fixed assets, also known as gross capital formation (GCF)

Note that the equations showing the breakdown into volume and price movements apply not only to GDP, but also to some of the other key variables in the national accounts, notably investment and consumption. So, for example, we can determine how much of a change in consumers’ expenditure in a particular period was due to a change in the quantities of goods and services they purchased and how much to changing prices for those goods and services.

ONS Resource

Real GDP and the GDP deflator feature prominently in the tables published by ONS, one of them being used to measure growth and the other to measure inflation. For example, the quarterly national accounts from February 2022 explain that:

The implied GDP deflator rose by 0.6% in Quarter 4 2021. Compared with the same quarter a year ago, the implied GDP deflator rose by 0.8%. This was mainly driven by the 3.6% increase in the implied price of household consumption, its highest rate since Quarter 3 (July to Sept) 2011.6

2.2.4 Comparing nominal and real GDP

Figure 2.4 illustrates the relationship between UK GDP in nominal prices, real GDP and the GDP deflator. This figure contains amounts expressed as indices base 100 in the year 2000. In other words, the two aggregates—nominal GDP and real GDP—are expressed in billions of pounds, but relative to the year 2000.

The index of nominal GDP reaches 211 in 2021, while real GDP is around 135 for the same year. The difference between the two lines is due to inflation. While the GDP deflator (inflation) does not appear as a separate line on the figure, it can be inferred as the gap between GDP at current prices and GDP in volume. The widening of this gap after the year 2000 indicates the existence of inflation. This is indeed the case, as can be seen from the fact that after 2000, GDP at current prices (the blue line) increases much faster than GDP in volume (the green line).

Figure 2.4 Gross domestic product, in nominal and real terms

Real and nominal GDP, UK, expressed as an index. Index: 2000 = 100.

Office for National Statistics – UK Economic Accounts - Table 1.1.2

Figure 2.5 shows the variations in UK’s GDP deflator. It can be seen that 2020 was characterised by fairly low inflation of 0.9% before accelerating to 2.6% the following year. For comparison, the table also shows the annual variation in the consumer price index (CPI). This index is another indicator of inflation that is better known and more frequently used than the GDP deflator, partly because it is available monthly and relates to the aggregate that is of most interest to people, namely consumption. The GDP deflator is, on the one hand, more general in scope than the CPI, since it also covers capital goods. But on the other hand, it measures only domestic inflation, with increases in import prices not directly taken into account.

Annual percentage growth rates 2017 2018 2019 2020 2021
GDP deflator 1.8 2.0 2.0 5.1 0.3
CPI 2.7 2.5 1.8 0.9 2.6

Figure 2.5 GDP deflator and consumer price index

GDP deflator and CPI, UK, 2017 to 2021

Macroeconomists tend to use real GDP most of the time. But GDP at nominal prices is used as the denominator to standardise many important aggregates, such as:

Ratios calculated as percentages of GDP, with both numerator and denominator usually expressed at nominal prices, are also used to make international comparisons of variables that would otherwise depend on the size of the country.

2.2.5 International comparisons of GDP and growth

Often, there is interest in comparing absolute levels of real GDP per head among different countries. Usually, the comparison is made for a particular year or series of years. How can this comparison among countries, or regions or zones—a so-called spatial analysis—be done?

Why do international comparisons use real rather than nominal GDP? Because the aim is to compare the quantities of goods and services produced in each country, and not the monetary value of this output, which would be affected by the differences among price levels.

Recall that, when analysing growth over time for a given country, real GDP is calculated by dividing GDP at nominal prices by a price index that is equal to 100 for a set base period. Exactly the same approach is used for spatial comparisons.

purchasing power parity (PPP)
The rate at which the currency of one country would have to be converted into that of another country so that one unit of currency can buy the same amount of goods and services in each country.

The comparisons of GDP across countries requires purchasing power parities (PPPs). It can be regarded as the hypothetical exchange rate at which countries’ currencies need to be converted into a common currency, so that a given amount of the first country’s currency will buy the same volume of goods and services in the second country as it does in the first.

Calculation of PPPs raises a number of methodological issues and problems which are discussed in more detail in the appendix to this chapter. But once calculated, their application to international comparisons is straightforward. The PPP can be regarded as the hypothetical exchange rate such that a given amount of the first country’s currency will buy the same volume of goods and services in the second country as it does in the first one.

Why not simply use the actual exchange rate seen on the currency markets? Because exchange rates are affected by a wide variety of factors, so they do not always reflect the difference in price levels between two countries. This means that they do not provide a true comparison of the volume of goods and services produced per head.

A moment’s reflection will show that this is exactly what we need to make a meaningful comparison between GDP in different countries. If we apply the PPP exchange rate to the second country’s GDP as expressed in its own currency, then we will have a quantity which can be directly compared with GDP in the first country expressed in its own currency terms.

2.2.6 Per capita comparisons

If we are comparing countries, it is often useful to focus on the comparative level of GDP per capita (GDP divided by population).

The simplest version of this ratio would be GDP at nominal prices (and expressed in a common currency, for example, the US dollar) divided by total population. However, this does not take into account the difference in the level of prices between countries. So most reports you will see use PPP comparisons, for the reasons discussed in the previous section.

Figure 2.6 demonstrates the evolution of the levels of GDP per head combined with the relative movements of quarterly real growth between countries. In this graph, three lines are shown: France, Germany, UK. But there is another implicit one, the US, which is not shown as, in this figure, it is by definition equal to 100. Therefore, the graph must be read as the evolution of GDP per head of France, Germany and the UK, relative to the evolution of the US GDP per head, all expressed at purchasing power parities of 2015.

Figure 2.6 Evolution of GDP per head relative to the US (US = 100)

GDP per head relative to US, France, Germany and the UK, Quarter 1 (Jan to Mar) 2009 to Quarter 4 (Oct to Dec) 2021, using purchasing power parities of 2015. Index: US = 100

US GDP per head (here represented implicitly as equal to 100) is one of the highest in the world. So all the lines for the other three countries, throughout the time period, are less than 100. We also see that the German GDP per head (around 80% of the US one) is higher than the French (circa 70%) or British (circa 72%) figures and saw a much lower fall in 2020. What we see in this graph, in addition, is that the three lines show an overall fall during this period. From this, we may conclude that, if it was the objective for these countries to close the gap with the US, then there has so far been no success. Note that Figure 2.6 does not mean that the three other countries have experienced a decrease of real GDP per head since the crisis. It just means that their growth per capita has been slower than that for the US. The graph also shows that starting in 2012 Q2 and up to the beginning of the pandemic (2020 Q2), the growth in the UK had been significantly greater than that of France.

2.2.7 Reconciling aggregate output and aggregate demand

fixed assets
Property, plant and equipment. Fixed assets are not expected to be converted back into cash in the next year.
gross fixed capital formation (GFCF)
GCF minus the change in inventories
household final consumption expenditure
The total expenditure of households on consumption goods, including imputed rent for owner-occupied housing

In the national accounts, investment—the purchase of machinery (including software) and buildings (offices, infrastructure, dwellings) and the building up of stocks (inventories)—is known as gross capital formation (sometimes abbreviated as GCF). When stockbuilding (or changes in inventories) is excluded, leaving only the purchases of buildings, machinery and other fixed assets, the result is known as gross fixed capital formation (GFCF). We will designate total investment as I, the change in stocks as S and gross fixed capital formation as GFCF.


\[I = \text{GFCF} + S\]

GFCF measures total expenditures on products intended to be used for facilitating future production. These types of products are collectively known as fixed capital. Why not simply call them investment, as economists in fact often do? Because the word “investment” in everyday use applies as much to financial investment (“I invest in shares of the stock market”) as it does to investment in machinery and buildings. So to make a clear distinction between the two applications, the national accountants use this somewhat peculiar terminology. Finally, the word “gross” indicates that the expenditure is measured without deducting the consumption of fixed capital, sometimes called depreciation, due to wear and tear or to obsolescence.

Household consumption, which we will designate as C, is the main part of domestic demand and is what the national accountants call household final consumption expenditure.

This variable covers all purchases made by consumers: food, clothing, housing services (rents), energy, durable goods (notably cars), spending on health, on leisure and on miscellaneous services. These are all items that are consumed (in the sense of used up or destroyed) during the period. Why final consumption? It is in contrast to intermediate consumption, referred to earlier.

Consumption expenditure does not, however, include households’ purchases of dwellings, which are counted as household GFCF. This is because dwellings are obviously not consumed only during the present period, but also provide future services to households.

After GDP, household final consumption probably receives the most attention in the national accounts, representing more than 60% of GDP. Indeed, the original economic model providing the underlying framework for the national accounts could be said to be aimed at maximising this consumption. Today, of course, there is increasing concern that consumption should be sustainable in the longer term (sustainable development). But, either way, consumption is a key concept for economics.

Final consumption and investment are two of the main components of final macroeconomic demand. The great attraction of the national accounts is that they constitute a reconciled model of the economy, balancing supply and demand. In fact, a second fundamental equation of the national accounts can be written as follows:

\[\text{GDP} = \text{sum of final demand aggregates}\]

In order to grasp the origin of this essential accounting equation, let us return to the example of the pasta industry.

Farmer National Flour National Pasta
Year 2
Input Labour + machinery Labour + machinery + wheat Labour + machinery + flour
Output £ 10,000 £ 30,000 £ 100,000
Intermediate consumption £ 10,000 £ 30,000
Value added £ 10,000 £ 20,000 £ 70,000

Figure 2.7 Summing the aggregates

Output, intermediate consumption and value added, by farmer and two hypothetical firms, year 2

Recall that GDP is equal to total value added or, equivalently, to total output minus total intermediate consumption.

Add up the output:

\[\text{£10,000 of wheat} + \text{£30,000 of flour} + \text{£100,000 of pasta} = \text{£140,000}\]

Deduct the intermediate consumption:

\[\text{£140,000} - \text{£10,000 of wheat} - \text{£30,000 of flour} = \text{£100,000 of pasta}\]

If one simplifies matters by ignoring possible inventory accumulation in the factory and in the distribution chain, the £100,000 corresponds exactly to the purchases by households, in other words to household final consumption expenditure.

This example shows that GDP—the sum of all values added—is equal, by definition, to final demand—which, in this simplified case, consists only of household demand for pasta.

How it’s done Reconcile output and final demand

Only a small amount of elaboration is needed to bring this example much closer to reality. If one introduces a firm that makes the machinery used to manufacture pasta, it can be verified that GDP equals exactly the consumption of pasta plus the purchase of the machinery used to make it, that is, household consumption plus GCF. This opens up the system to GCF in addition to household consumption. In addition, if we assume that the economy is open to imports—which we designate as M—and that there is external demand reflected in exports, designated as X, the equation:

\[\text{GDP} = \text{sum of final demand aggregates}\]


\[\text{GDP} + M = C + I + X\]

The left-hand side of the equation consists of supply at the macroeconomic level, made up of domestic production (GDP) and external supply (imports). The right-hand side consists of final demand, broken down into domestic demand (household consumption and GCF) and external demand (exports). Equivalently, of course:

\[\text{GDP} = C + I + X - M\]

The left-hand side now consists solely of GDP, essentially the value that the economy is producing available for final uses. The right-hand side consists of the final uses—the major components of domestic demand together with net exports, which is simply the difference between exports and imports. This accounting equation is fundamental to analysing the economic condition.

The equation \(\text{GDP} = C + I + X - M\) provides a perfect illustration of the impact of demand on supply, according to Keynesian reasoning. It is no accident, in fact, that national accounting—although its beginnings pre-dated Keynes’s major contributions—was developed quickly during the 1940s when it became clear that it could be a powerful tool in implementing what economists would call a Keynesian approach to policymaking. During the Great Depression, Keynes had advocated increased government expenditures and lower taxes to stimulate demand and pull the global economy out of the depression.

To be fully precise, the above equation has to be made slightly more complex, as shown in Figure 2.8. In reality, there are more final uses than discussed above. One small item relates to the final demand from so-called NPISHs (non-profit institutions serving households). These comprise a miscellany of bodies such as churches, trade unions, charities, universities and political parties. But they account for only a small proportion of GDP (2.3% in 2021).

A much more important item is the final demand represented by general government consumption (20.8% of GDP in 2021). As a convention of national accounts, the output of the general government—which consists essentially of public services, such as education, health, defence, police, etc.—is seen as consumed by the government itself. In reality, these services are consumed by citizens rather than by government itself. But it is difficult to break down some of these services between households (final consumption) and firms (intermediate consumption). In this context, it has been seen preferable to attribute fictitiously these flows to the government itself.

Figure 2.8 also shows stockbuilding (changes in inventories). Inventories are essentially a short-term shock absorber between variations in production and in final demand from households and firms. Although usually small in absolute terms, stockbuilding nevertheless plays an important role in the short term, since the changes can be significant in relation to the overall size of GDP.

SNA code Component of GDP Value (£ billion) Contribution to GDP (%)
P31S14 Final consumption expenditure of households (ABPF) 1,233,327 60.3
P31S15 Final consumption expenditure of non-profit institutions serving households (ABNU) 50,653 2.3
P3S13 Final consumption expenditure of general government (NMRU) 456,149 20.7
P3 Final consumption expenditure 1,832,347 83.3
P51 Gross fixed capital formation (NPQR) 382,900 17.4
P52 Changes in inventories (and valuables) (ABMQ and NPJP) 5836 0.3
P5 Gross capital formation (YBIK) 388,736 17.7
P6 Exports of goods and services (KTMZ) 600,792 27.3
P7 Imports of goods and services (KTNB)* −629,704 −28.6
B11 External balance of goods and services -28,912 −1.3
Statistical discrepancy (GIXS) 6302 6302 0.3 0.3
B1_GE Gross domestic product (by expenditure approach) (BKVT) 2,198,473 2,198,473 100.0 100.0

Figure 2.8 Contributions to GDP

Real GDP and final demands, UK, 2021. (Note: *negative because subtracted from GDP).
CDID codes are provided in brackets.

Office for National Statistics – UK Economic Accounts – Table 1.1.2

The penultimate line of Figure 2.8 is “statistical discrepancy”. In theory, GDP as sum of value added equals total final demand. In the real world, ONS methods and sources for deriving the output measure are necessarily different from those for estimating the expenditure measure. This creates a short-term statistical discrepancy between the two aggregates.

How it’s done Contributions to growth

The ONS’s report on GDP estimates for October 2021 to December 2021 reported that household consumption, government consumption and gross capital formation all contributed positively to growth, while net trade subtracted from growth.7 How do the statisticians create these reports?

Recall that \(\text{GDP} = C + I + X - M\)

The equation provides a mathematical explanation of GDP growth in terms of its various components.

The sign Δ will be used to express the difference between two years (or two quarters), so that \(\Delta\text{GDP}_{t}\) signifies \(\text{GDP}_{t} - \text{GDP}_{t - 1}\), ; in other words, the difference between GDP in year (quarter) \(t\) and GDP in year (quarter) \(t-1\).

Using this notation, \(\text{GDP}_{t}/\text{GDP}_{t - 1}\) will be equal to the GDP growth rate for year (or quarter) t compared with year (or quarter) \(t-1\).

For illustration, consider a simplified equation linking GDP with the final demands. For this simplified equation, we assume that there are no imports and no government consumption. The equation can then be expressed in real terms as:

\[\text{GDP}_{t} = C_{t} + I_{t} + X_{t}\]

where C stands for final consumption, I for GFCF and X for exports.

Dividing both sides by \(\text{GDP}_{t-1}\)

\[\frac{\text{GDP}_{t}}{\text{GDP}_{t - 1}}\ = \frac{C_{t}}{\text{GDP}_{t - 1}} + \frac{I_{t}}{\text{GDP}_{t - 1}} + \frac{X_{t}}{\text{GDP}_{t - 1}}\]

Dividing and multiplying each term on the right-hand side by its value in \(t-1\):

\[\frac{\text{GDP}_{t}}{\text{GDP}_{t - 1}}\ = \frac{C_{t - 1}}{\text{GDP}_{t - 1}}\frac{C_{t}}{C_{t - 1}} + \frac{I_{t - 1}}{\text{GDP}_{t - 1}}\frac{I_{t}}{I_{t - 1}} + \frac{X_{t - 1}}{\text{GDP}_{t - 1}}\frac{X_{t}}{X_{t - 1}}\]

The verbal translation of this equation is as follows: GDP growth breaks down exactly into the contribution of consumption plus the contribution of investment plus the contribution of exports. Each contribution is equal to the weight of the variable multiplied by the growth rate of the same variable in the current period. The weight of the variable is equal to its value in the previous period divided by the GDP of the previous period.

This breakdown of growth is widely used by macroeconomists. The calculation of contributions to growth relies on the accounting identity between GDP and final demand. In macroeconomic tables expressed in growth rates, changes in inventories and net exports are never shown in terms of percentage growth rates but solely as contributions to growth.

2.2.8 Reconciling aggregate output with aggregate income

There is a second reconciliation, this time between output and the aggregate income of economic agents.

Any production activity generates income that is shared between the three factors of production: labour, capital and intermediate consumption. But, by definition, intermediate consumption is used up in the production process, as explained in Section 2.2.1. Since value added is equal to output minus intermediate consumption, this second macroeconomic reconciliation can be written more simply by eliminating intermediate consumption and using value added as the global indicator of output.

This means that there are now just two factors creating value added, namely labour and capital, which are compensated respectively by salaries and by the profits generated through production. It is these types of income that subsequently enable economic agents—households and firms—to consume and invest.

For example, the £100,000 of GDP of our now-familiar pasta industry are divided between the profits of the farmer, the profits of the two firms National Flour Ltd and National Pasta Ltd, and the salaries of the staff at firms National Flour and National Pasta. This will be discussed in more detail in Chapter 3.

2.3 Controversies and improvement

It is important to realise that, in order to achieve the aim of summarising a country’s entire economic activity in a set of internally consistent tables, national accounts must accept significant approximations and adopt conventions that are sometimes arbitrary.

These approximations and conventions may be open to criticism, but they have been the subject of lengthy discussions by national accountants, and they were often chosen for sound practical reasons.

2.3.1 How accurate is GDP?

ONS Resource

The ONS publishes its revisions to UK GDP. This report shows the extent, and the average, of revisions in the 30 years to 2018. It distinguishes the two sources of revisions: data content increases over time, and methodological improvements.

National accounts could better be called national accounts statistics because, without this qualifier, users may think they are as reliable as the business accounts of a company. This is not true. In particular, while the level of GDP for technical reasons is often expressed in millions of units of the national currency, users should be aware that they are far from being accurate at the level of millions, although growth numbers may be more accurate than levels.

National accounts’ quality is highly dependent on the quality of the statistical system that exists in a given country. Further, in all countries, to varying degrees, the data and methodologies applied to it do not cover every bit of economic activity that is taking place.

This means a number of adjustments need to be made to counteract such omissions. Therefore, national accounts data are approximations. It is not even possible to give a summary of the accuracy of the level of GDP. Indeed, national accounts—and in particular GDP—are not the result of a single big survey and/or methodology for which we could compile a confidence interval. They are the result of combining a complex mix of data from many sources—many of which require adjustment to put them into a national accounts database and which are further adjusted to improve coherence—often using subjective, if hopefully well-based, methods.

GDP levels can typically be revised by 1 to 3 percentage points when new and more reliable data becomes available (even before taking into consideration changes arising because of conceptual or methodological improvements). It can even happen, although rarely, that some countries modify their estimate of GDP by more than 15% (Italy in 1987, China in 2005).

2.3.2 Quality change

The quality of national accounts is not the same in all countries. Overall, the OECD Statistics Directorate believes it may be misleading to establish a strict order of ranking countries based on GDP per capita at purchasing power parity in those cases when countries are clustered around a narrow range of outcomes of less than five percentage points.

In practice, it is not always easy to distinguish between changes in nominal GDP stemming from changes in prices and those from changes in quantities. In the case of commodity goods, such as the pasta example, it seems relatively straightforward to do so.

But consider services—which in practice make up the great majority of the value added that cumulate to GDP. Here the split between prices and quantities is harder to assess. Suppose that an insurance policy premium, for example, is observed to increase in nominal terms. That might be simply a reflection of the fact that the insurance company has increased its price. But it might also be the case that the insurance company has changed the nature of the policy so as to increase the value to a policy holder. Maybe, for instance, the size of the excesses has been reduced or the range of adverse events that the policy insures against has increased. In essence, the quality of the policy has improved. This has more of the nature of a quantity change since the real value conveyed by the new policy to its holder is greater than the value implicit in the earlier policy.

Read more about this type of adjustment in Chapter 1. Consumer price indices, for instance, ought also ideally to allow for quality change as well as price changes. If the price of a product or service rises only because it has increased in quality, it is hard to think of this as inflation.

Problems for national accounting caused by changes in quality are not confined to services. Many goods are also subject to quality change over time. As an illustration, the nature and quality of a mobile phone has changed rapidly in recent years. Just counting the number of mobile phones as a way of determining the price/quantity split is not itself good enough. The quality change over time also has to be assessed, and taken into account.

2.3.3 Non-market activity

non-market transaction
A transaction covering goods or services that a producer supplies to others free of charge, or at prices that are not economically significant.

GDP is predominantly a measure of market-based transactions, although not exclusively so; it includes non-market transactions, such as expenditure by government on non-market services such as the NHS and schools. But around 70% is market based.

Consistent with this spirit, households’ internal production (cooking, cleaning, running errands) is not covered in the national accounts. The principal reason has been that inclusion would involve making very bold estimates of its market value. This leads to the joke first made by Paul Samuelson, another Nobel-prize-winning economist, that if a man marries his maid the result is a reduction in GDP.

ONS resource

The ONS household satellite accounts estimate the value of unpaid household production in 2015 and 2016.

Traditionally, the problem has been regarded as small; clearly, it is not. Not attributing value to the labour of (predominantly) women is clearly unfair, and can give misleading information. The ONS Household Satellite Accounts estimate that “the value of the UK’s unpaid household service work was estimated at £1.24 trillion – larger in size than the UK’s non-financial corporation sector; overall unpaid household service work was equivalent to 63.1% of gross domestic product”.8

Many services, such as an internet search, are free, but clearly have a value.

On the other hand, the national accounts include an estimate of the production of services in the form of the accommodation house owners provide for themselves. This is called imputed rent and is fairly difficult to estimate, since there is no observable transaction involved. However, if one were not to make this estimate, the change in GDP could be affected by a change in the proportion of households owning their own dwelling.

Actual rental payments by households certainly appear in the national accounts. By making an estimate of the rents people essentially pay to themselves for their housing, and also including this in the national accounts, eliminates the possibility of a change in the proportion of owner-occupation distorting the path for GDP.

2.3.4 How big is the underground and illegal economy?

ONS Resource

This paper sets out the basis for revisions, how they are included, and the challenges from data limitations in recording illegal activity.

Abramsky, Joshua and Steve Drew (2014), ‘Changes to national accounts: Inclusion of illegal drugs and prostitution in the UK national accounts’, ONS

The underground and illegal economy is, in principle, included in the national accounts, though the estimates are often based on indirect and approximate sources. The UK has included illegal drugs and prostitution activity in the estimates of UK GDP and its components, since 2014. In the case of the UK, these adjustments increase GDP by around 2.5 to 3%.

Why include these transactions? The European System of National Accounts states that illegal transactions to which all agents involved consent are included within the production boundary, and therefore the relevant items should be included in the national accounts.

However, the treatment of these transactions in the national accounts of the EU member states varies. Different types of illegal activity are included in different states, so the illegal activity in national accounts cannot easily be compared.

2.3.5 Consumption or investment?

GFCF is intended to capture capital expenditure that increases the stock of capital, which offers a stream of future benefits. For example, income spent on computers or machines is considered as investment, not consumption. This principle creates some interesting and non-obvious distinctions. One example is that R&D was previously considered as expenditure, but since 2014 is now capitalised.

From the household’s perspective, expenditure for the purchase of a house is recorded as GFCF, but expenditure on durable goods, cars in particular, is classified as consumption. Yet the services rendered by a car generally last a fairly long time, although obviously not as long as those of a house. This shows how it is necessary to draw a line somewhere between consumption and investment.

2.3.6 Does GDP measure wellbeing?

One other key issue is the extent to which GDP can be regarded as an indicator of national welfare or social wellbeing. Media comment sometimes implies that it is—at worst portraying GDP announcements as if indicating national success or failure. But this is a dangerous oversimplification.

On 18 March 1968, Senator Robert F Kennedy made a famous speech on the shortcomings of this type of national accounting as a measure of wellbeing. It “counts air pollution and cigarette advertising, and ambulances to clear our highways of carnage,” he said. Read the speech.

In one sense, more national output and national income might be regarded as a clear improvement in wellbeing, since more material resources would be available for people to use. But it might seem strange that GDP rises if there are more road accidents. This is partly because of greater activity by emergency services.

Many people fail to understand why GDP does not fall following major natural catastrophes. One would intuitively like to see GDP diminishing in such circumstances. But this would be to confuse a measure of output (GDP) with a well-defined measure of welfare, which GDP is not. Undoubtedly, major calamities destroy part of the economic wealth (buildings, houses, roads and infrastructure), but they do not, per se, constitute negative production and so do not directly contribute to a decline in GDP.

Destruction can indirectly affect production in a negative or positive way. When a factory is destroyed it ceases production, but it also must be rebuilt and this constitutes production. For this reason, paradoxically, it is possible for a natural catastrophe to have a positive impact (in the purely mathematical sense of the word “positive”) on GDP. At most, GDP is a measure of the contribution of production to welfare.

There are a great number of other dimensions to welfare that GDP does not claim to measure:

Why the bizarre title “gross domestic product” or GDP? It should be clear by now that “product” describes what one is trying to measure—the result of production. “Domestic” indicates that the output measured is produced within the economic territory of the country, a region or a group of countries. “Gross” means the consumption of fixed capital is not deducted (see below). “Net” domestic product, which does make this deduction is discussed in Section 2.4.2 below.

2.4.1 Gross national income (GNI)

gross national income (GNI)
The sum of a nation’s gross domestic product and the net income it receives from overseas. It was previously known as gross national product, or GNP.

“Domestic” is also in opposition to “national”, as in GNI or gross national income, which is the current title of what used to be referred to as GNP, or gross national product, in previous systems of national accounts (“GNP” is still widely used by the media out of habit).

GDP measures the total production occurring within the territory, while GNI measures the total income (excluding capital gains and losses) of all economic agents residing within the territory (households, firms and government institutions).

To convert GDP into GNI, it is necessary to add the income received by resident units from abroad and deduct the income created by production in the country but transferred to units residing abroad:

In 2017, when Ireland’s Central Statistics Office created a national statistic called Modified GNI, designed to account for some of the ways in which the country’s economy is unlike most others, it discovered that “[t]he Irish economy is about a third smaller than expected. The country’s current account surplus is actually a deficit. And its debt level is at least a quarter higher than taxpayers have been led to believe” in the words of the Financial Times.9

These differences can be large. For example, the earnings of workers living in France but working in Luxembourg have to be deducted from the Luxembourg GDP to obtain its GNI. Conversely, the earnings of the seasonal workers living in Poland and working temporarily in the UK have to be deducted from the UK GDP to obtain the UK GNI. But it is larger for a small country like Luxembourg, which pays out a substantial percentage of its GDP as workers’ earnings and other so-called primary income to the “rest of the world” (the term used by national accounts to signify “all countries other than Luxembourg”, in this case). Primary income includes interest paid on money invested in Luxembourg. Luxembourg also receives substantial primary income from abroad, including interest. Taking into account all these factors, GNI for Luxembourg is around 30% lower than its GDP.

Ireland is in a comparable situation to Luxembourg, since it pays out substantial dividends to the parent companies of the multinational firms that have set up there, partly, but not entirely, for tax reasons. The result is that Ireland’s GNI is 25.2% lower than its GDP.

For large countries like the UK, the difference between GDP and GNI is small (1.0% in 2020, as seen in the following table). While for Luxembourg and Ireland GNI is lower than GDP, the opposite also happens—Germany and Switzerland are cases in point.

Year 2020 UK Luxembourg Ireland
Gross domestic product 2,043,373 59,592 377,344
Net primary income −21,182 −17,775 −95,241
Gross national income 2,022,191 41,817 282,103
Difference between GDP and GNI (%) −1.0 −29.8 −25.2

Figure 2.9 Reconciliation of GDP and GNI for UK, Luxembourg and Ireland

Reconciliation of GDP and GNI for UK, Luxembourg and Ireland

£ Million (UK)/Million euros (Luxembourg, Ireland)

Organisation for Economic Co-operation and Development – Annual National Accounts

2.4.2 Net domestic product (NDP)

net domestic product (NDP)
GDP minus depreciation of a country’s capital goods

Although less widely used than GDP, net domestic product (NDP) is, in theory, a better measure of the wealth produced, since it deducts the cost of wearing out the machinery and other capital assets used in production.

To produce goods and services (the output), different inputs are required: labour (the labour force), goods and services (intermediate consumption) and capital (such as machinery). These represent the inputs into the production process.

In order to arrive at a genuine measurement of the new value created during the period, a deduction has to be made for the cost of using up capital (such as the wear and tear on machinery). This is known as consumption of fixed capital. When this consumption is deducted, the result is net value added (NVA), and NDP is the sum of these net values added. Formally:

\[\text{NDP} = \text{NVA}_{i}\]

for each firm, i.

2.4.3 Net national income (NNI)

net national income (NNI)
Gross national income minus depreciations

For similar reasons, net national income (NNI) is, in theory, a better measure than GNI of the income created, because net national income deducts the cost of using up capital assets.

However, economists tend to prefer GDP or GNI (over NDP and NNI) for two reasons:

The UK is an exception in this regard. Historically, the UK has experienced lower investment relative to its GDP than has been the case in most comparable countries. In itself, that has not been a favourable outcome, but it does mean that the UK’s capital stock is relatively low in relation to its GDP. In turn, that means that its capital consumption is also smaller. So when this amount is deducted from GDP to obtain NDP, the UK’s ranking in terms of the latter rises considerably.

For example, UK NDP is some 12% lower than its GDP but in the US NDP is 16% lower than its GDP. So the gap is some 4% smaller when compared in NDP terms rather than by using GDP.

2.5 GDP in policymaking

2.5.1 Quarterly and accounts and seasonal adjustments

In the United Kingdom, a recession is defined as a negative economic growth for two consecutive quarters. Other countries define the concept differently.

One of the crucial objectives of macroeconomic statistics is to help the authorities and macroeconomic agents make the right decisions at the right moment. It would not be appropriate to launch a policy boosting the economy when the upswing has already started, or conversely to “cool down” the economy when it is already entering recession.

Therefore, it is desirable to have the most refined possible information regarding the economic cycle and its turning points. In this context, the annual national accounts arrive far too late. Moreover, exclusive reliance on annual averages can in fact be misleading about the true state of the economy. Hence, it is important to compile accounts that are more timely than annual ones. In most developed countries, national accounts are published quarterly. A few countries, such as Canada and the UK, produce monthly estimates.

The information given by such estimates is much used by commentators and forecasters, whether in the Finance Ministry helping in the preparation of the government budget, or in private research institutions such as those connected to the large banks, or in international organisations such as OECD. When they are published, new national accounts typically receive a great deal of attention in the media and elsewhere, because of the information they give about the state of the economy.

In theory, the quarterly (or more frequent) accounts should be no different from the annual accounts as regards the basic principles and the definitions of the variables. The difference is merely that the size of the flows experienced. For example, in the quarterly accounts, flows will be roughly only a quarter as large as those in the annual accounts (as is logical, given that one calendar quarter accounts for only three months out of 12). Conversely, the annual flows are equal to the sum of the flows for the four quarters.

However, for practical reasons and because the data sources on which they rely are less extensive, the quarterly accounts are simplified compared with the annual accounts. Along the same lines, the detail of the transactions in the accounts for institutional sectors is not as great in the quarterly accounts as in the annual accounts. This enables the ONS to reduce the workload entailed in the more frequent calculation of quarterly accounts.

Another important feature of the quarterly accounts is the seasonal adjustment. This consists in eliminating, by means of complex statistical processes based on moving averages, the changes from one quarter to the next that are due simply to seasonal effects. For example, the output of transport services rises systematically and steeply before Christmas and the summer holidays. It is, therefore, better to eliminate the impact of this seasonal effect in order to know whether holidaymakers actually consumed more or less in the quarter or month in question than in the previous quarter or month. Things are so arranged that the sum of the quarterly seasonal adjustments for the year as a whole is zero. In other words, the sum of the quarterly seasonally adjusted figures remains equal to the unadjusted figure for the year.

2.5.2 Contributions to quarterly real GDP growth

Figure 2.10 represents the contributions of each significant element of the demand to the growth of the UK GDP in volume for the twelve quarters from Quarter 1 in 2019 to Quarter 4 in 2021. The line in black is the growth of the quarterly real GDP. The last histogram represents the cumulated contributions to the cumulated growth over these eight quarters. As can be seen, as regards the evolutions of some quarters, the element “change in inventories” can have an important positive (Quarter 4, 2020) or negative (Quarter 2, 2019) contributions. Changes in inventories are a “shock absorber” on a quarter to quarter basis, as explained in Section 2.2.5. The same is true for net trade, which had a very significant negative contribution in Quarter 3 2020, followed by some compensation in Quarter 1 2021.

Figure 2.10 Contributions to quarterly real GDP growth

Contributions to growth of GDP in volume, UK, Quarter 1 (Jan to Mar) 2019 to Quarter 4 (Apr to Jun) 2021

Office for National Statistics – UK Economic Accounts - Table 1.1.2

It is interesting to analyse the cumulated contributions represented in the last histogram. Here we see that, over eight quarters, it is changes in the growth of private consumption that has been the most sizeable contributor to changes in GDP.

2.5.3 Identifying recessions

Periods of time when output is falling are often termed recessions.

The UK annual national accounts date back to 1948 and have appeared in quarterly form since 1955. Figure 2.11 shows that GDP is presently around four and a half times its level in 1955, and correspondingly national income is higher by the same multiple.

But the growth has not been uniform and there have been periods where output has fallen. (These are shown by the shaded bars in the chart.) This was the case in the early to mid-1970s, the early 1980s, in the early 1990s, in 2009 at the time of the global financial crisis and most markedly in 2020 through the Coronavirus (COVID-19) pandemic. In the case of the global financial crisis, the subsequent recovery was also conspicuously slow, particularly in comparison to 2020.

In the United Kingdom, a recession is defined as a negative economic growth for two consecutive quarters. Other countries define the concept differently.

These episodes of falling GDP are sometimes referred to as recessions. But this term needs careful use.

Figure 2.11 Rise and fall in UK GDP since 1955

Level of real GDP, UK, Quarter 1 (Jan to Mar) 1955 to Quarter 4 (Oct to Dec) 2021 (1955 = 100)

Office for National Statistics – ‘GDP quarterly national accounts’

Commentators have sometimes been guided by a rule of thumb which evidently first derived from American presidential speechwriters. The rule of thumb is that one period of falling output does not constitute a recession, but that two consecutive quarters do.

This usage is illogical and sometimes misleading. Suppose in one episode, GDP falls sharply by 1% in the first quarter, then recovers by 0.1%, only to fall by a further 1% in the third quarter. The cumulative fall in output would be 1.9%. But the two quarters rule of thumb would say this was not a recession.

On the other hand, if output fell by 0.1% in quarters 1 and 2 and then rose by 0.1% in the third, the cumulative output loss would be a much less worrying 0.1%. But the rule of thumb would judge that there had been a recession.

2.6 Summary

2.7 Further reading