Specification Tests in Econometrics
Citations
28,298 citations
Cites background or methods from "Specification Tests in Econometrics..."
...Hausman (1978) suggested comparing the OLS and 2SLS estimators of b1 1 ðd 0 1; a1Þ 0 as a formal test of endogeneity: if y2 is uncorrelated with u1, the OLS and 2SLS estimators should di¤er only by sampling error....
[...]
...See Lahiri and Schmidt (1978) and Hausman, Newey, and Taylor (1987). Chapter 9 228...
[...]
...A classic example of selection based on yi is Hausman and Wise’s (1977) study of the determinants of earnings....
[...]
...Breusch, Mizon, and Schmidt (1989) studied the Hausman-Taylor model under the additional assumption that Eðx 0 it2ciÞ is constant across t. This adds more orthogonality conditions that can be exploited in estimation. See IASW (1999) for a recent analysis....
[...]
...In the Hausman and Wise (1977) example, a2ðxÞ was a function of family size because the poverty income level depends on family size....
[...]
26,580 citations
25,689 citations
8,131 citations
Cites background or methods from "Specification Tests in Econometrics..."
...Second, to assess the magnitude of any simultaneity between High Performance Work Practices and firm profits, I used Hausman's (1978) test to evaluate the ordinary-least-squares (OLS) regression assumption that the High Performance Work Practices scales are exogenous in the profitability models. In analyses whose results are not shown, I generated a predicted value for the employee skills and organizational structures and employee motivation scales using a reduced-form model. Then I included each scale and its predicted value in an OLS model for Tobin's q and GRATE. A significant coefficient on the predicted value for each scale would indicate it is endogenous in the model being estimated (Hausman, 1978). Although these results showed that the High Performance Work Practices scales were not in fact endogenous in the profitability models, the general controversy surrounding the use of this test (Addison & Portugal, 1989) led me to estimate two-stage least-squares models for each dependent variable as a formal correction for simultaneity. Not only were these results consistent with the OLS results, but they were in each case somewhat larger than the OLS results presented here. Survey response bias was also considered directly. The presence of response bias implies that unobserved determinants of the decision to respond to this study’s survey are related to both firm performance and High Performance Work Practices. Given the extensive control variables included in my models, such bias is unlikely. However, to formally test this possibility, I used Heckman's (1979) procedure, which generates an inverse Mills' ratio that I then included in the OLS and two-stage least-squares regression models for each dependent variable to control for selectivity bias....
[...]
...Second, to assess the magnitude of any simultaneity between High Performance Work Practices and firm profits, I used Hausman's (1978) test to evaluate the ordinary-least-squares (OLS) regression assumption that the High Performance Work Practices scales are exogenous in the profitability models....
[...]
...A significant coefficient on the predicted value for each scale would indicate it is endogenous in the model being estimated (Hausman, 1978)....
[...]
7,768 citations
Cites background or methods from "Specification Tests in Econometrics..."
...(1981), Chapman and Staelin (1982), and Hausman and Ruud (1987). A mixed logit model can be estimated on ranked data with the same explosion....
[...]
...Following the practice suggested by Hausman (1997), they used the prices in other cities by the same company as instuments for each city....
[...]
...Berndt, Hall, Hall, and Hausman (1974), hereafter referred to as BHHH (and commonly pronounced B-triple H), proposed using this relationship in the numerical search for the maximum of the log-likelihood function....
[...]
...Hausman and McFadden (1984) provide an appropriate statistic for this type of test....
[...]
References
4,733 citations
"Specification Tests in Econometrics..." refers background in this paper
... Однако возникает важная проблема спецификации, что было отмечено в (Maddala, 1971, p. 357) и затем подчеркивалось в работе (Mundlak, 1976)....
[...]
...Mundlak Y. (1976)....
[...]
...357] and has been further emphasized by Mundlak [14]....
[...]
...27 Работа была позже опубликована как: Mundlak Y. (1978)....
[...]
[...]
3,429 citations
Additional excerpts
... В работах (Scheffé, 1959) и (Searle, 1971) можно найти превосходные обсуждения этой темы в рамках дисперси- 16 Такой метод оценивания модели со случайными эффектами остался незамеченным в литературе....
[...]
2,976 citations
"Specification Tests in Econometrics..." refers methods in this paper
... Проверка того, что 0a= в уравнении (2.20) при нулевой гипотезе, основана на статистике 2 2 ˆ ˆˆ ˆ( )xx Q x s = a a. Однако 10 Тест инструментальных переменных может быть также рассмотрен как формализация и улучшение предложений Sargan (1958), который рекомендовал проверять, лежат ли оценки метода наименьших квадратов вне доверительных интервалов для IV оценок....
[...]
...9The instrumental variable test can also be considered a formalization and an improvement of a suggestion by Sargan [20] who recommended checking whether the least squares estimates lie outside the confidence regions of the IV estimates....
[...]
...Sargan J. D. (1958)....
[...]
... Проверка того, что 0a= в уравнении (2.20) при нулевой гипотезе, основана на статистике 2 2 ˆ ˆˆ ˆ( )xx Q x s = a a. Однако 10…...
[...]
... Таким образом, предложенная процедура дает более короткие доверительные интервалы, чем процедура Sargan....
[...]
2,464 citations
Additional excerpts
... 5 Утверждение этой леммы для случая конечной выборки и одного параметра содержится в статье (Fisher, 1925), ссылка на которую получена от W. Taylor....
[...]
...Fisher [8], a reference supplied by W....
[...]
[...]
1,472 citations