In this lesson · 8 sections
Gross domestic product is the market value of all final goods and services produced within a country's borders in a period. Every word in that sentence is doing work, and most misuse of the number comes from ignoring one of them. "Market value" means unpriced activity is invisible. "Final" means intermediate transactions are stripped out to avoid double counting. "Produced within a country's borders" means it is a production measure, not an income-of-residents measure. "In a period" means it is a flow, not a stock — GDP is not the wealth of a country any more than your salary is your net worth.
GDP is also a construct, assembled from surveys, tax records, customs data and a great deal of estimation, then revised for years afterwards. Treating a first estimate as a fact is the most reliable way to be wrong about the economy in public.
Three ways to measure the same thing
National accounts compute GDP three ways, which must agree in principle and never quite agree in practice.
The production (value-added) approach sums what each producer adds. Suppose a farmer sells wheat to a miller for CHF 100. The miller sells flour to a baker for CHF 180. The baker sells bread to households for CHF 300. Adding sales gives 100 + 180 + 300 = 580, which double-counts the wheat three times and the flour twice. Value added is output minus intermediate inputs: 100 for the farmer, 80 for the miller, 120 for the baker. That sums to 300 — exactly the value of final sales. Both routes agree because that is how the accounts are built.
The expenditure approach sums final purchases: Y = C + I + G + (X − M). Household consumption, gross fixed capital formation plus inventory change, government consumption and investment, and net exports. Imports are subtracted not because imports are bad but because C, I, G and X already contain imported content that was not produced domestically.
The income approach sums what production pays out: compensation of employees, gross operating surplus of firms, mixed income of the self-employed, plus taxes on production less subsidies.
The three estimates differ because the source data differ, and statistical offices reconcile them through supply-and-use tables. The size of the unreconciled gap is itself a rough indicator of how much confidence a given quarter's estimate deserves.
Nominal, real, and the deflator
Nominal GDP mixes changes in quantity with changes in price. Real GDP strips out prices, which is what you want if you are asking whether more was actually produced.
The GDP deflator is the ratio of nominal to real GDP. Suppose nominal GDP rises from CHF 800bn to CHF 832bn — an increase of 4.0% — while the deflator rises 2.5%. Real growth is 1.040 / 1.025 − 1 = 1.46%, call it 1.5%. Subtracting 2.5 from 4.0 to get 1.5 works here only because the numbers are small; the division is the correct operation and the approximation degrades quickly at higher inflation.
The GDP deflator is not the consumer price index. The deflator covers everything produced domestically, including investment goods and exports, and excludes imports. The CPI covers a basket bought by households, including imports. For an economy like Switzerland's, with a large export sector and heavy import content in the consumption basket, the two can diverge meaningfully — for instance when the franc appreciates, lowering import prices in the CPI while export prices in the deflator move differently.
Modern accounts also use chain-linking rather than a single fixed base year. Weights are updated each period and linked together, because using decade-old weights badly distorts growth in sectors where relative prices move fast — computing equipment is the standard case. A consequence non-specialists find surprising is that chain-linked components no longer add up exactly to the total.
Per capita, per hour, and the growth illusion
Headline GDP growth answers "did the economy get bigger". It does not answer "did people get better off". For that you need at least per-capita, and ideally per-hour-worked.
Take real GDP growth of 1.5% with population growth of 1.2%. GDP per capita grows 1.015 / 1.012 − 1 = 0.30%. Almost the entire headline was more people, not more output each.
This is not a hypothetical for Switzerland. Population growth driven by net immigration has been a persistent contributor to Swiss GDP growth, so Swiss headline growth systematically flatters the per-capita picture relative to economies with flat populations. When you compare Swiss growth to Japanese growth, comparing headline numbers is close to meaningless; comparing per-capita or per-hour figures is informative. The productivity lesson in this course takes the argument further.
Quarterly conventions, annualisation and the Swiss quirks
The United States reports quarterly GDP at a seasonally adjusted annual rate: the quarterly change compounded to a yearly pace. Switzerland and the euro area report the plain quarter-on-quarter change. A Swiss quarterly print of +0.3% corresponds to an annualised 1.003⁴ − 1 = 1.21%. A US print of +1.2% and a Swiss print of +0.3% describe the same speed. Comparing them directly is a rookie error and it appears in professional commentary constantly.
Switzerland has two idiosyncrasies worth knowing.
First, SECO publishes GDP both unadjusted and adjusted for major sporting events. Several large international sports federations are domiciled in Switzerland, and their licensing and broadcasting revenue in tournament years lands in Swiss output. The effect is large enough to distort quarterly growth without any change in underlying Swiss economic activity. Use the sporting-event-adjusted series when you are trying to read the cycle.
Second, Swiss trade statistics are published on more than one basis. The customs administration reports a headline series that excludes gold bullion, other precious metals, precious stones, art and antiques, alongside a broader series that includes them. Gold flows through Switzerland in enormous and volatile quantities relative to the size of the economy, and including them turns the trade balance into noise. Pharmaceuticals and chemicals concentration adds a second source of lumpiness: a single company's shipment schedule can visibly move a quarter.
Revisions: the number you traded on is not the number that survives
First estimates are built from incomplete source data. As tax returns, annual company accounts and structural surveys arrive, estimates are revised — sometimes for years, and occasionally by enough to reverse the sign of a quarter.
The mechanism matters for how you should behave. Revisions are not random error sprinkled on a true value; they are systematic in known ways. Early estimates lean on survey and monthly indicator data, which underrepresent small firms and new firms. Turning points are revised most, because the statistical models that fill gaps are estimated on the recent past and therefore extrapolate the old trend into the new regime. In practice this means official data tends to understate the severity of downturns in real time and to revise them worse later.
The investment implication is not to ignore GDP but to weight it correctly. A first-estimate quarter that comes in at +0.2% versus an expectation of +0.4% tells you far less than the market reaction implies. Benchmark revisions — periodic overhauls where methods and base years change — can move the entire level and growth history of a country's output.
What GDP misses, and why it was designed that way
Simon Kuznets, who did more than anyone to construct the US national accounts, warned Congress in 1934 that the welfare of a nation can scarcely be inferred from a measurement of national income. He was right, and the caveat has been ignored ever since.
Unpriced production. Household work, childcare, and volunteering produce real output and enter GDP at zero. If you pay a nanny, GDP rises; if you look after your own child, it does not. Imputed rent for owner-occupied housing is one of the few imputations the accounts do make, precisely because leaving it out would make homeownership look like a collapse in output.
Distribution. GDP is a total. It says nothing about who received it. An economy where all growth accrues to a narrow group and one where it is spread evenly are indistinguishable in the headline.
Depletion and damage. GDP is gross: it does not net off the consumption of fixed capital, and it certainly does not net off depletion of natural resources or environmental damage. Cleaning up a spill adds to GDP twice — once for the activity that caused it, once for the cleanup.
Quality and free services. Deflators try to adjust for quality improvement, and this is genuinely hard. Services delivered at zero price, funded by advertising, are captured only indirectly through the advertising expenditure that supports them.
Location versus ownership. GDP measures production inside borders; gross national income measures income accruing to residents. For countries hosting large multinational operations the two diverge sharply. Ireland is the notorious case: the relocation of intangible assets onto Irish balance sheets raised measured Irish GDP by roughly a quarter in a single year without a comparable change in Irish living standards. Switzerland, with its large multinational and commodity-trading presence, has its own version of the problem, which is why Swiss GNI and GDP are worth looking at together.
None of this makes GDP useless. It makes it a production statistic — an excellent one — that was never a welfare statistic and should not be pressed into service as one.
What to take away
- The production, expenditure and income approaches must give the same total by construction; the value-added chain of 100 + 80 + 120 = 300 shows why summing gross sales double-counts and why only final output belongs in GDP.
- Real growth is nominal growth divided by the deflator, not subtracted from it: 4.0% nominal with a 2.5% deflator is 1.46% real, and the shortcut fails as inflation rises.
- The US annualises quarterly GDP and Switzerland does not, so a Swiss +0.3% quarter and a US +1.2% quarter describe the same pace of growth.
- First estimates systematically miss turning points and are revised most exactly when the cycle turns, so a small surprise in an initial print deserves far less weight than markets usually give it.
- GDP counts marketed production inside a border and nothing else — not unpaid work, not distribution, not resource depletion, and not the difference between output produced here and income accruing to residents.
Sources
- United Nations et al., System of National Accounts 2008 — the international standard defining GDP and its components.
- Eurostat, European System of Accounts (ESA 2010) — the European implementation, which Switzerland follows closely.
- Simon Kuznets, National Income, 1929–1932, report to the US Senate, 1934 — the original warning against reading welfare from national income.
- Diane Coyle, GDP: A Brief but Affectionate History (Princeton University Press, 2014).
- Joseph Stiglitz, Amartya Sen and Jean-Paul Fitoussi, Report by the Commission on the Measurement of Economic Performance and Social Progress (2009).
- State Secretariat for Economic Affairs (SECO), quarterly GDP releases, including the series adjusted for major sporting events.
- Swiss Federal Statistical Office (FSO), national accounts methodology and gross national income statistics.
- Federal Office for Customs and Border Security, Swiss foreign trade statistics and their treatment of gold and precious metals.