Students / Subjects



Forgot password?



A B C D E F G H I J K L M N O P Q R S T U V W X Y Z (Show all)

H index

Stands for Herfindahl-Hirschman index, which is a way of measuring the concentration of market share held by particular suppliers in a market. It is the sum of squares of the percentages of the market shares held by the firms in a market. If there is a monopoly -- one firm with all sales, the H index is 10000. If there is perfect competition, with an infinite number of firms with near-zero market share each, the H index is approximately zero. Other industry structures will have H indices between zero and 10000.
Tirole's version is bounded between zero and one because each of the market shares is between zero and one.

Source: econterms

Habakkuk thesis

That high wages and labor scarcity stimulated technological progress in the U.S. in the 1800s, and in particular brought about the American system of manufacturing based on interchangeable parts. (This description from Mokyr, 1990; idea from Habakkuk, 1962).

Source: econterms


Generally, habits are conceptualized as the learning of sequences of acts that have become automatic responses to specific situations, which may be functional in order to achieve a given result, or to obtain certain goals or end states (e.g. James, 1890, Triandis, 1977, Watson, 1914). Habits thus comprise a goal directed type of automaticity; they are instigated by a specific goal directed state of mind on the presence of triggering stimulus cues, for instance taking the car to travel to the supermarket. Once evoked, the behavior will run to completion without the need for attentional control of the process. Habit strenght is proposed to increase as a result of repetitions of positive reinforcements.

Source: SFB 504

Hahn problem

Hahn (1965) question: when does there exist an equilibrium in a model in which money has positive value?

Source: econterms

Hansen's J test

See J statistic

Source: econterms


A synonym for labor-augmenting, in practice.

Source: econterms

Hausman test

Given a model and data in which fixed effects estimation would be appropriate, a Hausman test tests whether random effects estimation would be almost as good. In a fixed-effects kind of case, the Hausman test is a test of H0: that random effects would be consistent and efficient, versus H1: that random effects would be inconsistent. (Note that fixed effects would certainly be consistent.) The result of the test is a vector of dimension k (dim(b)) which will be distributed chi-square(k). So if the Hausman test statistic is large, one must use FE. If the statistic is small, one may get away with RE.

Source: econterms

hazard rate

escape rate; rate of transition out of current state

Source: econterms

Heaviside function

Is a mapping from the real line to {0, 1}, denoted (at least sometimes) hv(x). hv(x) is zero for x<0, and is one for x>=0.

Source: econterms


An occasional name for generalized Tobit. This approach allows a different set of explanatory variables to predict the binary choice from those which predict the continuous choice. (The data environment is one in which the continuous choice is measured only when the binary choice is nonzero -- e.g., if we have data on people, whether they bought a car, and how expensive it was, we can estimate a statistical model of how expensive a car other people would buy, but only on the basis of the ones who did buy a car in the data sample.) A regular, non-generalized Tobit constrains the two sets of variables to be the same, and the signs of their effects to be the same in the two estimated equations. 'Heck' is for James Heckman.

-- Christopher Baum, Boston College economics department, 20 May 2000, in a broadcast to the statalist, the email list of people interested in the software Stata.

Source: econterms

Heckman two-step estimation

A way of estimating treatment effects when the treated sample is self-selected and so the effects of the treatment are confounded with the population that chose it because they expected it would help -- the classic example is that college educations may be selected by those most likely to benefit.

Taking that example, we wish to advance past the following regression:
wi = a + bXi + dCi + ei
where i indexes people, wi is the wage (or other outcome variable) for agent i, Xi are variables predicting i's wage, and Ci is 1 if i went to college and 0 if not. ei is the remaining error after least squares estimation of a, b, and d.

Source: econterms

Heckscher-Ohlin model

A model of the effects of international trade. "The Heckscher-Ohlin framework typically is presented as a two-country, two-good, two-factor model. The two countries are assumed to share identical, homothetic tastes for the two substitutable goods and identical, constant-returns-to-scale technologies with some factor substitutability. Perfect competition prevails in each market with zero transport costs and no artificial barriers to international trade in goods, although factors are internationally immobile. In this framework, each country will (incompletely) specialize in production and export the good using intensively in production the factor that the country has in relative abundance." That effect is called factor-price equalization across countries, and is used sometimes to explain how rising international trade would lead to greater income inequality in the most developed countries. (from Bergstrand, Cosimano, Houck, and Sheehan, 1994, p 3)
The reference in the name is to "Scandinavian economists Eli Heckscher and Bertil Ohlin early in [the twentieth century]" in work that is rarely cited directly. (from Bluestone, 1994, p 336).

Source: econterms

Hedge strategies

A hedge strategy is the intentional reduction of the loss risk (downside risk) of an underlying asset to the debit of the gain chance. Usually hedge strategies are done with derivative securities, e.g. options.

Source: SFB 504


of or relating to utility. (Literally, pleasure-related.) A hedonic econometric model is one where the independent variables are related to quality; e.g. the quality of a product that one might buy or the quality of a job one might take.

A hedonic model of wages might correspond to the idea that there are compensating differentials -- that workers would get higher wages for jobs that were more unpleasant.

"A product that meets several needs, or has a variety of features ... generates a number of hedonic services. Each one of these services can be thought of as generating its own demand, along with a resulting hedonic price. Although each separate component is not observable, the aggregation of all the components results in the observed product demand and equilibrium price.... [Q]uality improvements will appear to an observer as an outward shift of the product demand curve, as consumers are willing to purchase more at the prevailing price." -- William J. White, "A Hedonic Index of Farm Tractor Prices: 1910-1955", Ohio State University working paper, October 1998, pp. 3-4.

Source: econterms


A list of fields contained here is below. There is some other advice at this help page:

Most terms are in one of these categories. You can click on one to see a list of terms relevant to it.

Source: econterms

Herfindahl-Hirschman index

See 'H index'.

Source: econterms

Hermite polynomials

The Hermite polynomials are a series of polynomials defined for each natural number r, used for statistical approximations I believe. Click here for the equation and graphs of the first several.

Source: econterms


The matrix of second derivatives of a multivariate function. That is, the gradient of the gradient of a function. Properties of the Hessian matrix at an optimum of differentiable function are relevant in many places in economics: 1) In maximum likelihood estimation, the information matrix is (-1) times the Hessian.

Source: econterms

heterogeneous process

A stochastic process is heterogeneous if it is not identically distributed every period.

Source: econterms


An alternate spelling of heteroskedastic. McCulloch (1985) argues that the spelling with the k is preferred, on the basis of the pronunciation and etymology (Greek not French derivation) of the term.

Source: econterms


An adjective describing a data sample or data-generating process in which the errors are drawn from different distributions for different values of the independent variables.
Most commonly this takes the form of changes in variance with the magnitude of X. That is, in
y = Xb + e
that the e's vary in magnitude with the X's. (An example is that variance of income across individuals is systematically higher for higher income individuals.)
If the errors are drawn from different distributions, or if higher moments of the error distributions vary systematically, these are also forms of heteroskedasticity.

Source: econterms


A heuristic is a strategy that can be applied to a variety of problems and that usually ? but not always ? yields a correct solution. People often use heuristics (or shortcuts) that reduce complex problem solving to more simple judgmental operations. Three of the most popular heuristics are discussed by Tversky and Kahnemann (1974):

Source: SFB 504

Hicks-Kaldor criterion

For whether a cost-benefit analysis supports a public project. The criterion is that the gainers from the project could in principle compensate the losers. That is, that total gains from the project exceed the losses. The criterion does not go so far as the Pareto criterion, according to which the gainers would in fact have to compensate the losers.

Source: econterms


An attribute of an effectiveness variable in a production function. The attribute is that it does not affect labor differently from the way it affects capital.

The canonical example is the Solow model production function Y=AF(K,L). There Y is output, L labor, K capital, F a production variable, and A represents some kinds of effectiveness variable. In Y=F(AK,L) the effectiveness variable affects capital but not labor. In Y=F(K,AL) it affects labor but not capital. These two cases can be described as Hicks-biased. In Y=AF(K,L) it is Hicks-neutral.

Source: econterms

Hicks-neutral technical change

Given a production function AF(K,L) changes in A are Hicks-neutral, meaning that they do not affect the optimal choice of K or L. The subject comes up in practice only for aggregate production functions.

Uzawa, H. 'Neutral Inventions and the Stability of Growth Equilibrium,' The Review of Economic Studies 28:2 (Feb., 1961), 117-124 contains the first known published use of the adjective 'Harrod neutral' According to it, the criterion of Harrod-neutrality comes from

Harrod, Roy F., 'Review of Joan Robinson's Essays in the Theory of Employment,' Economic Journal, vol. 47 (1937), 326-330.

Uzawa also proves that AF(K,L) and F(K,AL) are the right functional forms to meet Hicks and Harrod-neutrality, and that only the Cobb-Douglas form accomplishes both.

Source: econterms

Hicksian demand function

h(p,u) -- the amount of a good that demanded by a consumer given that it costs p per unit and that the consumer will have utility u from all goods. h(p,u) is the cost-minimizing amount.

Source: econterms

High School and Beyond

A panel data set on U.S. high school students.

Source: econterms

high-powered money

reserves plus currency

Source: econterms

Hilbert space

A complete normed metric space with an inner product. So the Hilbert spaces are also Banach spaces. L2 is an example of a Hilbert space. Any Rn with n finite is another.

Source: econterms

Hindsight bias

It is a common observation that events in the past appear simple, comprehensible, and predictable in comparison to events in the future. Everyone has had the experience of believing that they knew all along the outcome of a football game, a political election or a business investment. The hindsight bias is the tendency for people with outcome knowledge to believe falsely that they would have predicted the reported outcome of an event. After learning of the occurrence of an event, people tend to exaggerate the extent to which they had foreseen the likelihood of its occurrence.

Source: SFB 504

Hindsight bias biased reconstruction

Contemporary models suggest that the hindsight bias appears due to reconstructive processes while people try to generate their original estimates.
One is loosely based on the response bias hypothesis, originally developed in eyewitness testimony research: People are assumed either to have remembered or to have forgotten their original judgements. Those that do remember their original estimates are likely to reproduce them. Those who have forgotten them are forced to guess and, in the presence of outcome information, are likely to utilize this information as an anchor assuming that their estimates must have been somewhere in the proximity of the true outcome. But since people are generally optimistic about their capacities, they will locate their presumed prior estimates closer to the real outcome than it had actually been, resulting in the hindsight bias.
Other authors concentrate their assumptions on three different stages in the reconstructive process: Selective retrieval, prejudiced interpretation and weighing of different cues.
Finally, the hindsight distortion may be triggered by the heuristic of anchoring and adjustment.

Source: SFB 504

Hindsight bias implications for further research

Hindsight Bias is a strong phenomenon which has been observed in many circumstances, and appears to influence every-day decision-making. Researchers are trying to develop hindsight bias models to provide a better theoretical explanation. Another aspect is practical relevance, whereby real-life cases need to be examined for evidence on distortion influence.

Source: SFB 504

Hindsight bias memory impairment

Fischhoff´s original explanation of the hindsight bias, Immediate Assimilation Hypothesis, states that memory for original predictions is altered by subsequent outcome (Fischhoff, 1975). When learning about the actual or alleged outcome, the person re-interprets the original evidence in the light of the outcome. They are therefore inadvertently modifying what had been previously stored in memory. Subsequent outcome knowledge is integrated immediately into the existing knowledge structure. This results in a permanent modification of the person´s prior representation of the event. Other variations of the memory impairment hypothesis suggest that the origins of hindsight biases lay in the retrieval stage. The Selective Retrieval Hypothesis maintains that known outcome serves as a retrieval cue for relevant case material. Once an outcome has been learned, information congruent with this outcome will become highly accessible. Incongruent information cannot be retrieved with the same ease. The authors of the Dual Memory Traces Model (Hell, Gigerenzer, Gauggel, Mall & Müller, 1988) suggested an extension of Fischhoff´s model. They assume two separate memory traces for own judgements and subsequent outcome information. The strength of hindsight biases is determined by the relative strength of the memory traces.

Source: SFB 504

Hindsight bias motivational explanations

Motivational explanations suggest that judgement and decision processes are not only affected by rational cognitions. They are also influenced by actual needs and motives, including need for control, need for cognition, self-relevance and, most importantly self-presentation concerns. In the latter approach, people are motivated to make others believe that their predictions were close to the actual outcome, in an attempt to maintain a high level of public self-esteem. Contrary to the memory impairment hypothesis, this explanation interprets hindsight distortions as adjustments during the response generation state. Several authors showed, that empirical evidence for motivational underpinnings of the knew-it-all-along effect is rather weak.

Source: SFB 504

Hindsight bias practical relevance

Every-day life provides numerous examples on the practical relevance of hindsight bias. This was shown in a medical context. A GP´s second opinion does not differ completely from another GP´s opinion, if he/she is aware of the first opinion. This seems to be of serious consequence if one considers that a second opinion is only required when serious illnesses have been diagnosed.

In a legal context, hindsight bias was found to occur when a jury makes a final decision in court. In the course of a trial, the judge is empowered to order the jury to ignore certain testimonials, by disallowing them. It has been proven impossible to ignore such information.

Hindsight bias also plays a role in an economic context. An economic expert may, for example, analyze certain share activity, as if he/she had always known what would happen. This results in them forming a higher opinion of their own judgement. This in turn can have a long-term, restrictive effect on their individual learning capability as well as future decision-making. In addition to this intra-personal effect, it is possible for inter-personal effects to arise. Example: A supervisor may no longer be able to make an undistorted judgement on his employees´ decision-making if he/she got information about some results of their performance. This is a special problem in the case of poor outcome because it could happen, despite them having acted correctly on the information they were given at the time.

Source: SFB 504

Hindsight bias response bias hyothesis

This model was originally conceived during eyewitness testimony research. This was to account for the fact that witnesses receiving misleading information about a previously observed event, show a poorer memory for it. Similar to the hindsight bias, the misleading information effect was originally attributed to memory impairment.

Subsequent interpretations offered the following explanation: Rather than altering existing memory traces, the new information may be used as a reference point by those who cannot remember the original information and therefore need to guess. Misleading information does not alter the original representation, but simply serves as an anchor to perceivers who are unable to retrieve it. Parallels between the hindsight bias and the misleading information paradigms are obvious. Both lines of research query whether information stored in the memory might be less accessible after being confronted with inconsistent new information. The experimental designs of both research traditions show strong similarities. Important differences: In the misleading information paradigm, the original information is presented by the experimenter. In hindsight bias studies however, the original estimates are generated by the subjects. Misleading information is presented unobtrusively without the subjects being aware of its misleading nature. Outcome information given in hindsight studies is explicitly labeled as the correct information.

Source: SFB 504

Hindsight bias theoretical and empirical work

More recent studies have come up with models that enable precise forecasting of the strength and direction of the hindsight distortion, using a quantitative basis. Researchers are also attempting to determine how far hindsight decisions affect a person´s trust in their own judgements. They are also trying to determine how far such decisions affect a person´s impression of their competency in future decision-making on the same subject. Another interesting point to be examined is the relationship between hindsight bias and the attribution of responsibility.

Source: SFB 504

Hindsight bias theoretical explanations

Although there is a wealth of literature on hindsight distortions, the underlying mechanisms are not yet fully understood. Attempts of explanation were given in three major theoretical areas: Models of memory impairment state that outcome knowledge affects memory for previous judgements by either altering or erasing existing memory traces, or by rendering them less accessible. Other attempts to explain the hindsight bias are based on assumptions that these distortions are driven by motivations. An alternative explanation is that people use distorting heuristics while reconstructing their original judgements.

Source: SFB 504


The subject of economic history is anything in history that is subject to economic explanations. Application of formal theory or statistical analysis of data may be relevant, although it is possible to make a contribution without either, e.g. with a case study or a contextual reinterpretation. Historians tend to be focused on what happened, how, and why, not on the question of whether a model fits the evidence.

Source: econterms


Statistical software for Hierarchical Linear Modeling, from Scientific Software International.

Source: econterms

Hodrick-Prescott filter

Algorithm for choosing smoothed values for a time series. The H-P filter chooses smooth values {st} for the series {xt} of T elements (t=1 to T) that solve the following minimization problem: min { {(xt-st)2 ... etc. } Parameter l>0 is the penalty on variation, where variation is measured by the average squared second difference. A larger value of l makes the resulting {st} series smoother; less high-frequency noise. The commonly applied value of l is 1600. For the study of business cycles one uses not the smoothed series, but the jagged series of residuals from it. See Cooley, 1995, p 27-29. That H-P filtered data shows less fluctuation than first-differenced data, since the H-P filter pays less attention to high frequency movements. H-P filtered data also shows more serial correlation than first-differenced data. For l=1600: "if the series were stationary, then [this choice] would eliminate fluctuations at frequencies lower than about thirty-two quarters, or eight years."

Source: econterms

hold-up problem

One of a certain class of contracting problems.

Imagine a situation where there is profit to be made if agents A and B work together, so they consider an agreement to do so after A buys the necessary equipment. The hold-up problem (in this context) is A might not be willing to take that agreement, even though the outcome would be Pareto efficient, because after A has made that investment, B would have the power might decide to demand a larger share of the profits than before, since A is now deeply invested in the project but B is not, so B has some bargaining power that wasn't there before the investment. B could demand all of the profits, in fact, since A's alternative is to lose the investment entirely.

Other hold-up problems are analogous to this one.

Source: econterms

Holder continuous

An attribute of a function g:Rd->R. g can be said to be Holder continuous if there exist constants C and 0<=E<=1 such that for all u and v in Rd:
|g(u)-g(v)| <= C||u-v||E

So if g is Holder continuous for C=1 then it is Lipschitz continuous? And if g is Holder continuous then it is continuous.

Source: econterms


An alternate spelling of homoskedastic. McCulloch (1985) argues that the spelling with the k is preferred, on the basis of the pronunciation and etymology (Greek not French derivation) of the term.

Source: econterms


An adjective describing a statistical model in which the errors are drawn from the same distribution for all values of the independent variables. Contrast heteroskedastic.
This is a strong assumption, and includes in particular the assumption in a linear regression, for example,
y = Xb + e
that the variance of the e's is the same for all X's.

(The observed variance will differ in almost any sample. But if one believes that the data-generating process does not in principle have greater variances for different values of the independent variable, one would describe the sample as homoskedastic anyway.)

Source: econterms


Let u(x) be a function homogeneous of degree one in x. Let g(y) be a function of one argument that is monotonically increasing in y. Then u(g()) is a homothetic function of y.

So a function is homothetic in y if it can be decomposed into an inner function that is monotonically increasing in y and an outer function that is homogeneous of degree one in its argument.

In consumer theory there are some useful analytic results that can come from studing homothetic utility functions of consumption.

Source: econterms

Household behavior

In traditional microeconomics, household behavior is understood narrowly as the theory of consumer demand for commodities, i.e., household consumption. There are, however, other aspects of household behavior that have also been investigated in the microeconomics literature, such as the household´s supply of labor, the production of commodities (mainly, services) within the household (household production), saving decisions, retirement decisions, and many more.

Source: SFB 504


Health and Retirement Study, a longitudinal panel of older Americans studied by the Survey Research Center at the University of Michigan. Their Web site is at

Source: econterms


High School and Beyond, a panel data set on U.S. high school students.

Source: econterms

Huber standard errors

Same as Huber-White standard errors.

Source: econterms

Huber-White standard errors

Standard errors which have been adjusted for specified assumed-and-estimated correlations of error terms across observations.

The implicit citations are to Huber, 1967, White, 1980, and White, 1982.

Source: econterms

human capital

The attributes of a person that are productive in some economic context. Often refers to formal educational attainment, with the implication that education is investment whose returns are in the form of wage, salary, or other compensation. These are normally measured and conceived of as private returns to the individual but can also be social returns.

''Human capital' was invented by the economist Theodore Schultz in 1960 to refer to all those human capacities, developed by education, that can be used productively -- the capacity to deal in abstractions, to recognize and adhere to rules, to use language at a high level. Human capital, like other forms of capital, accumulates over generations; it is a thing that parents 'give' to their children through their upbringing, and that children then successfully deploy in school, allowing them to bequeath more human capital to their own children.' -- Traub (2000)

Source: econterms

hyperbolic discounting

A way of accounting in a model for the difference in the preferences an agent has over consumption now versus consumption in the future.

For a and g scalar real parameters greater than zero, under hyperbolic discounting events t periods in the future are discounted by the factor (1+at)(-g/a).

That expression describes the "class of generalized hyperbolas". This formulation comes from a 1999 working paper of C. Harris and D. Laibson, which cites Ainslie (1992) and Loewenstein and Prelec (1992).

In dynamic models it is common to use the more convenient assumption that agents have a common discount rate applying for any t-period forecast, starting now or starting in the future. Hyperbolic discounting is less convenient but fits the psychological evidence better, and when contrasted to the constant-discount-rate assumption can get models to fit the noticeable fall in consumption that U.S. workers are observed to experience when they retire. In a constant-discount-rate model the worker would usually have forecast the fall in income and their consumption expenses would be smooth.

One reason hyperbolic preferences are less convenient in a model is not only that there are more parameters but that the agent's decisions are not time-consistent as they are with a constant discount rate. That is, when planning for time two (two periods ahead) the agent might prepare for what looks like the optimal consumption path as seen from time zero; but at time two his preferences would be different.

Contrast quasi-hyperbolic discounting.

Source: econterms


a hypothesized property of unemployment rates -- that there is a ratcheting effect, so a short-term rise in unemployment rates tends to persist.
Theories that would lead to hysteresis:
-- an insider/outsider model of decisionmaking about employment; insiders such as the unionized workers ratchet up wage rates beyond where it is profitable to hire the unemployed; outsiders who are unemployed don't get to be part of the negotiation process.
-- behavioral and human capital changes among the unemployed, such as forgetting the details of work or work behavior, or losing interest or skill in getting new jobs, could lead to declining chances of becoming employed.

Source: econterms

Copyright © 2006 Experimental Economics Center. All rights reserved. Send us feedback