Gerhard Tintner, Scope and Method of Econometrics: Illustrated by Applications to American Agriculture (1949)

Tintner’s journal article explains econometrics as a connection between theoretical economic relationships, numerical estimation, and empirical testing. It establishes the method’s place within economics, examines its statistical foundations, and illustrates its possibilities and limitations through American agricultural demand and supply. Its governing concern is how quantitative inquiry can inform policy without deciding society’s objectives.

Econometrics has to be distinguished from mathematical economics and from statistical economics.

Mathematical economics formulates models and derives their implications; econometrics estimates parameters and tests theoretical claims against observations. Tintner nevertheless treats mathematical and literary theory as continuous: symbolic formulation clarifies assumptions rather than producing an entirely separate kind of knowledge. Statistical description likewise cannot dispense with theory, since selecting and organizing observations already presupposes conceptual commitments.

This defence of theory-guided measurement does not establish an exclusive method of economic inquiry. Historical research remains appropriate where numerical evidence is scarce, while institutional analysis explains the legal framework of economic activity. Introspection can suggest hypotheses about consumption, but these require checking against evidence beyond the investigator’s experience. Econometrics must establish its value through results.

Tintner also distinguishes empirical consequences from political judgements. Lower tariffs might increase national product while compromising self-sufficiency; an estimate of these effects cannot determine their relative desirability.

But the choice of policy is a matter of politics, ethics and similar considerations.

Economic laws remain conditional, and welfare propositions depend on specified objectives. This distinction also informs Tintner’s reservations about statistical decision theory: a risk function requires judgements about the consequences of errors, whereas agreement on economic policy goals cannot be assumed.

The statistical discussion concerns the adaptation of inference to non-experimental economic evidence.

Statistics is necessary if we want to proceed from the abstract formulations of mathematical economics to numerical results.

Tintner discusses maximum likelihood, least squares, confidence limits, and hypothesis testing alongside identification. Following Carnap, he distinguishes probability as a hypothesis’s degree of confirmation from probability as limiting relative frequency. The former concerns theoretical credibility but remains insufficiently developed for practical econometrics; the latter supplies working procedures despite philosophical difficulties.

The agricultural application demonstrates why identification precedes estimation. Observed prices and quantities alone cannot distinguish demand from supply. Tintner specifies separate linear relationships by including income in demand but not supply, and a production-cost variable in supply but not demand; both equations include time trends. These exclusions depend on substantive assumptions that are only approximately justified. Treating agriculture separately also risks overlooking interactions with the wider economy.

Using twenty-four annual observations from 1920–1943, Tintner estimates measurement-error variances through the Variate Difference Method and fits equations by weighted regression. The implemented procedure allows errors in variables while neglecting errors in equations, leaving the adequacy of the model dependent on the importance of omitted influences.

The results are uneven. Demand coefficients are significant at the five-percent level, but supply coefficients are not. Tintner attributes the supply failure to inadequate specification, particularly omitted weather conditions and the possible need for lagged prices. Estimated demand elasticities are −0.123 for price and 0.307 for income. Tests reject both unit and zero price elasticity and indicate that income responsiveness exceeds price responsiveness in magnitude.

These estimates suggest different consequences for price supports and income-expanding policies, but their interpretation requires unchanged conditions and other things remaining equal. Tax and subsidy illustrations are particularly tentative because they also rely on the unsuccessful supply estimates. The article thus links theoretical specification, identification, estimation, and uncertainty to conditional policy inference. Its failed supply equation is central to this argument: econometrics exposes where models need revision rather than simply furnishing numerical authority for policy.