Quantifying<i>R</i><sup>2</sup>bias in the presence of measurement error

Majeske, Karl D.; Lynch-Caris, Terri; Brelin-Fornari, Janet

doi:10.1080/02664760902814542

Cited by 7 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, R2 is biased downward because of the noise from mismeasurement. A simple derivation (see Majeske, Lynch-Caris, and Brelin-Fornari 2010) shows that in expectation,…”

Section: E Comparisons When the Benchmark Policy Is Imperfectmentioning

confidence: 99%

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

Jacobsen

Knittel

Sallee

et al. 2020

Journal of Political Economy

View full text Add to dashboard Cite

Corrective taxes can solve many market failures, but actual policies frequently deviate from the theoretical ideal because of administrative or political constraints. We present a method to quantify the efficiency costs of constraints on externality-correcting policies or, more generally, the costs of imperfect pricing, using simple regression statistics. Under certain conditions, the R 2 and the sum of squared residuals from a regression of true externalities on policy variables measure relative welfare gains from policies. We illustrate via four empirical applications: random mismeasurement of externalities, imperfect electricity pricing, heterogeneity in the longevity of energy-consuming durable goods, and imperfect spatial policy differentiation.

show abstract

“…However, R2 is biased downward because of the noise from mismeasurement. A simple derivation (see Majeske, Lynch-Caris, and Brelin-Fornari 2010) shows that in expectation,…”

Section: E Comparisons When the Benchmark Policy Is Imperfectmentioning

confidence: 99%

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

Jacobsen

Knittel

Sallee

et al. 2020

Journal of Political Economy

View full text Add to dashboard Cite

show abstract

“…For example, if property rights protection in the host-country dictates a firm's decision to invest there in the first place, it should also influence its choice to reinvest or repatriate foreign-earned income. And while typical treatments of the linear model tend to conclude that measurement error in the dependent variable is rather unproblematic (Greene 2008, 326), this conclusion does not hold where, as is the case here, the error is correlated with regressors (Wooldridge 2010, 76-82), or when the quantity of interest is the coefficient of determination (Majeske, Lynch-Caris, and Brelin-Fornari 2010). Therefore, it seems plausible that the weak results found above are due to measurement error.…”

Section: What Could Explain the Weak Explanatory Power Of The Countrymentioning

confidence: 65%

The Political Determinants of Foreign Direct Investment: A Firm-Level Analysis

Arel‐Bundock

2016

International Interactions

View full text Add to dashboard Cite

Many large-N cross-national studies claim to show that political institutions and phenomena determine where foreign direct investment (FDI) flows. In this paper, I argue that these studies tend to overemphasize statistical significance, and often neglect to assess the explanatory or predictive power of their theories. To illustrate the problem, I estimate variations of a statistical model published in an influential article on Political Risk, Institutions, and FDI. I find that none of the political variables that the authors consider account for much of the variation in aggregate FDI inflows. To ensure that this underwhelming result is not driven by misspecification or measurement error, I leverage a large firm-level dataset on the investment location decisions of thousands of multinational firms. Using non-parametric machine-learning techniques and out-of-sample tests, I show that gravity variables can help us develop very accurate expectations about firm behavior, but that none of the 31 "political determinants" of FDI that I consider can do much to improve our priors. These findings have important implications, because they suggest that governments retain some room to move in the face of economic globalization.

show abstract

“…They also inflate the standard errors of the regression coefficients, resulting in an understatement of the significance of these coefficients (Wooldridge, , pp. 71–72; Majeske et al ., ). To see this, let us consider the following regression model where y is the variable that we would like to explain:

y = β_{0} + β_{1} x_{1} + β_{2} x_{2} + \dots + e

…”

Section: Hypothesis Developmentmentioning

confidence: 97%

“…Following Majeske et al . (), we subtract Equation ) from ), and we obtain the bias in ρ 2 due to measurement error in the dependent variable, denoted by D .

D = ρ_{y}^{2} - ρ_{y^{*}}^{2}

…”

Section: Hypothesis Developmentmentioning

confidence: 99%

See 1 more Smart Citation

Return windows and the value relevance of earnings

Tan

Wong

2018

Accounting & Finance

View full text Add to dashboard Cite

Research on the value relevance of annual earnings commonly accumulate stock returns over a 12-month period starting from the fourth month of the fiscal year, resulting in a mismatch between the return window and the earnings period (i.e. the fiscal year). By comparing this return window with alternative windows, we show that the mismatch produces a downward bias in the estimated R 2 from the regression of stock returns on earnings, especially for firms that announce earnings early. Our results also show that both profits and losses are more value relevant when announced earlier, supporting regulatory calls for timely disclosure.

show abstract

QuantifyingR²bias in the presence of measurement error

Cited by 7 publications

References 25 publications

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

The Political Determinants of Foreign Direct Investment: A Firm-Level Analysis

Return windows and the value relevance of earnings

Contact Info

Product

Resources

About

QuantifyingR2bias in the presence of measurement error

Cited by 7 publications

References 25 publications

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

The Use of Regression Statistics to Analyze Imperfect Pricing Policies

The Political Determinants of Foreign Direct Investment: A Firm-Level Analysis

Return windows and the value relevance of earnings

Contact Info

Product

Resources

About

QuantifyingR²bias in the presence of measurement error