WealthPast logo WealthPast
People
Blog
Calculators
Proverbs & SayingsOn This DayHolidaysAboutContact

Trygve Haavelmo’s Econometrics: Probability, Identification and the Correlation Trap

Trygve Haavelmo’s econometrics explains why probability and identification matter when prices and quantities are determined together.

Written by Ibraham

Trygve Haavelmo helped transform econometrics from curve fitting into a framework for testing economic relationships under uncertainty. His central warning remains financially important: observed variables can be determined together. A correlation between price and quantity, investment and growth, or credit and default does not by itself identify which mechanism caused which outcome.

Trygve Haavelmo before 1950 — historical photographic portrait, existing source crop
Trygve Haavelmo before 1950 — historical photographic portrait, existing source crop. Credit: Unknown author Unknown author. Wikimedia Commons source · Public domain. Existing source crop or thumbnail where applicable; no AI alteration.

From Frisch’s institute to a probabilistic foundation

Haavelmo was born in Norway in 1911, entered the University of Oslo in 1930 and completed his economics degree in 1933. He worked with Ragnar Frisch, studied statistics in London and taught at Aarhus before wartime work in the United States. His 1944 monograph The Probability Approach in Econometrics provided the foundation for his later doctorate.

He joined the Cowles Commission in 1946, returned to Norway in 1947 and became professor of economics and statistics at Oslo in 1948. The University of Oslo records later books on economic evolution (1954) and investment theory (1960). In 1989 he received the economics prize for clarifying the probability foundations of econometrics and analyzing simultaneous economic structures.

Why a fitted line may answer the wrong question

Economic data are often generated by systems. A market price affects quantity demanded, but supply conditions also affect both the observed price and quantity. If an analyst simply fits quantity on price without identifying which shifts came from supply and which from demand, the resulting slope can mix different relationships.

Haavelmo’s probability approach treats an economic model as a claim about how data are generated under uncertainty. Identification asks whether the available variation and assumptions can distinguish the parameter of interest. Estimation comes afterward. This sharpens the model-building tradition seen in Jan Tinbergen’s econometric policy work and complements Richard Stone’s accounting discipline.

A fictional price-and-quantity trap

Suppose a shop sells 100 units at $10 in one month and 120 units at $12 in the next. It would be wrong to conclude that a higher price caused customers to buy more. A festival may have shifted demand outward, raising both price and quantity. Alternatively, a new premium product may have changed the measured mix. Two observations do not identify a demand curve.

The figures are invented. To estimate price response, an analyst needs a credible source of variation, a model of the system and uncertainty around the result. Even then, the conclusion applies under stated assumptions; it is not a timeless constant.

Three safeguards for financial analysis

  • State the data-generating story: explain why variables moved, not just that they moved together.
  • Separate identification from precision: a narrow standard error cannot rescue the wrong causal parameter.
  • Keep structural change visible: regulation, technology and behavior can alter relationships after estimation.

Haavelmo’s lesson is not that useful inference is impossible. It is that probability, theory and institutional knowledge must work together. A forecast can be numerically precise and still answer the wrong economic question.

Sources and Method

This article separates documented history from the clearly labeled teaching example. It does not estimate personal wealth, promise a policy outcome or offer individualized financial advice.

Share with