Functional Form Misspecification Research Articles

Digital age imperatives and firm internal control quality: evidence from CEOs personal trait and type of state-owned enterprises

PurposeIn the context of an evolving digital-based global economy, this study aims to investige the impact of digital transformation (DT) on a firm’s internal control (IC) quality. It also explores how the personal traits of (CEOs) – such as age, gender and educational background – intersect with DT to shape the IC quality in various types of state-owned enterprises (SOEs).Design/methodology/approachThe study uses the data from China A-shares non-financial enterprises, listed on Shanghai and Shenzhen stock exchanges between 2007 and 2020. Using the fixed effect regression method alongside various statistical techniques, such as propensity score matching, alternative analysis and instrumental variables analysis, yields robust findings. These methods effectively address issues related to functional form misspecification and potential biases from omitted explanatory variables.FindingsThe findings reveal a positive impact of DT on firm IC quality, and this impact is more pronounced in firms when the CEO is female, young and possesses a higher level of education. Notably, the study also distinguishes between central and local state-owned enterprises (SOEs), highlighting that DT has a greater influence on IC quality in central SOEs, where CEOs often have higher political ranks and closer to government monitoring. Overall, the findings are robust and consist to alternative variable and other statistical methods.Research limitations/implicationsFollowing are the significant implications for both academia and business. First, firms that effectively adopt DT to enhance IC not only gain a strategic advantage over competitors but also establish efficient risk management practices and a robust IC system. Second, better IC resulting from DT can enhance investor and stakeholder confidence. This is particularly important for publicly traded companies, where investors and analysts closely scrutinize the robustness of IC systems. Third, DT could result in cost savings over time, as automation and streamlined processes may reduce the need for manual efforts and resource-intensive tasks associated with IC.Originality/valueThe findings are contributed to the literature in multiple ways. It enhances our comprehension of the intricate DT-IC quality relationship, and provides valuable insights into the transformative impact of DT on organizational operations and risk management. It also introduces a novel perspective by investigating how CEOs personal traits intersect with DT to shape IC quality, contributing to upper echelons theory. Furthermore, it expands the discussions on firm ownership by considering the types of SOEs (central vs. local), in the DT-IC quality context.

Inclusive bank based financial development in countries with special needs: A semiparametric analysis

What is the role of bank based financial development in inclusive sustainable development? Our paper tries to answer this question with a novel methodology that provides an estimate of the inequality-finance relationship for every country-year observation, allowing us to examine the distribution of the relationship of interest for various country groups. Unlike other papers in the literature, our estimates are free of both functional form mis-specification bias as well as bias from endogenous regressors. Our results indicate that financial development lowers income inequality in low income countries and countries that invest in primary education. This inequality-finance relationship is also sensitive to a country’s size of financial sector and quality of institutions. Our findings should be of particular interest to scholars and domestic policy makers of Countries with Special Needs, who face unique challenges meeting the goals set by United Nation’s 2030 Agenda for Sustainable Development.

Comparing the Incomparable? Issues of Lacking Common Support, Functional-Form Misspecification, and Insufficient Sample Size in Decompositions

Decompositions make it possible to investigate whether gaps between groups in certain outcomes would remain if groups had comparable characteristics. In practice, however, such a counterfactual comparability is difficult to establish in the presence of lacking common support, functional-form misspecification, and insufficient sample size. In this article, the authors show how decompositions can be undermined by these three interrelated issues by comparing the results of a regression-based Kitagawa-Blinder-Oaxaca decomposition and matching decompositions applied to simulated and real-world data. The results show that matching decompositions are robust to issues of common support and functional-form misspecification but demand a large number of observations. Kitagawa-Blinder-Oaxaca decompositions provide consistent estimates also for smaller samples but require assumptions for model specification and, when common support is lacking, for model-based extrapolation. The authors recommend that any decomposition benefits from using a matching approach first to assess potential problems of common support and misspecification.

Factors influencing export performance of Ginger (Zinbiger officinale) in Nigeria

This study evaluated factors influencing export performance of ginger (Zinbiger officinale) in Nigeria. Specifically, the research study was designed to achieve the following objectives: examine the short run equilibrium relationships between test variables; evaluate the long run equilibrium relationships between test variables; and determine the linear – interdependence between factors influencing export of ginger. The test variables under considerations were export quantity of ginger (Zinbiger officinale), exchange rate, interest rate, ratio of producer price to domestic price, and ratio of producer price to export price. Data were of secondary sources covering 1995 -2020.The statistical and econometrics tools used to achieve the stated objectives were: Augmented Dickey Fuller (ADF), Phillips-Perron (PP), Johansen Co-Integrating Test and Vector Error Correction Model (VECM). The results of Augmented Dickey-Fuller (ADF) unit root test shows that all test variables were non-stationary at level. All test variables were statistically significant at first difference at 5% level of probability. Phillip-Perron (PP) unit root test shows that all test variables were stationary at first difference and were statistically significant. Johansens co-integration test revealed the presence of co-integration among variables and long-term relationships among variables. Vector Error Correction Model (VECM) shows that exchange rate (EXR), interest rate (INR), had negative coefficients and was statistically significant in influencing export performance of ginger (Zinbiger officinale) in Nigeria in the short run. Ratio of producer price to domestic price (PP), producer price to export price (PX), quantity of ginger exported (Y) had positive coefficients and were statistically significant in influencing export performance of ginger (Zinbiger officinale) in Nigeria in the short run. In the long run, exchange rate (EXR), interest rate (INR), ratio of producer price to domestic price (PP), ratio of producer price to export price (PX) were statistically significant in influencing export performance of ginger (Zinbiger officinale) in Nigeria. The coefficients of error correction model (ECM) was negative, this measures the speed of adjustment towards the long run equilibrium. The coefficients of multiple determinations (R^2) was 0.6710. This confirmed presence of goodness of fit. The F-statistics of 43.06 was significant at 1% level of significance. This confirmed the explanatory power of the entire model. The RESET value of VECM was 1.510 (P < 0.01) shows no evidence of misspecification of functional form. The research study recommends policy formulations and implementations that will stabilize exchange rate. Interest rate should be single digit. Export market should be developed along the ginger (Zinbiger officinale) value chains and other crops to expand export market base and earn foreign exchange

The impact of research and development (R&D) on economic growth: new evidence from kernel-based regularized least squares

PurposeResearch and development (R&D) is increasingly considered to be a key driver of economic growth. The relationship between these variables is commonly examined using linear models and thus relies only on single-point estimates. Against this background, this paper provides new evidence on the impact of R&D on economic growth using a machine learning approach that makes it possible to go beyond single-point estimation.Design/methodology/approachThe authors use the kernel regularized least squares (KRLS) approach, a machine learning method designed for tackling econometric models without imposing arbitrary functional forms on the relationship between the outcome variable and the covariates. The KRLS approach learns the functional form from the data and thus yields consistent estimates that are robust to functional form misspecification. It also provides pointwise marginal effects and captures non-linear relationships. The empirical analyses are conducted using a sample of 101 countries over the period 2000–2020.FindingsThe estimates indicate that R&D expenditure and high-tech exports positively and significantly influence economic growth in a non-linear manner. The authors also find a positive and statistically significant relationship between economic growth and greenhouse gas emissions. In both cases, the effects are higher for upper-middle-income and high-income countries. These results suggest that a substantial effort is needed to green economic growth. Internet access is found to be an important factor in supporting economic growth, especially in high-income and middle-income countries.Practical implicationsThis paper contributes to underlining the importance of investing in R&D to support growth and shows that the disparity between countries is driven by the determinants of economic growth (human capital in R&D, high-tech exports, Internet access, economic freedom, unemployment rate and greenhouse gas emissions). Moreover, since the authors find that R&D expenditure and greenhouse gas emissions are positively associated with economic growth, technological progress with green characteristics may be an important pathway for green economic growth.Originality/valueThis paper uses an innovative machine learning method to provide new evidence that innovation supports economic growth.

Convertible Debt Usage and the Pricing of Audit Services

SYNOPSIS This study examines the association between convertible debt usage and the pricing of audit services. We test the (nondirectional) hypothesis that convertible debt usage is associated with audit effort and therefore fees, due to its association with client business risk and its dilutive effect on earnings per share. We find a positive association between audit fees and convertible debt, suggesting that auditors view convertible debt as a source of risk. We also find that audit fees related to convertible debt are sensitive to CEO bonus incentives and to market valuation incentives. Our results suggest that following the Public Company Accounting Oversight Board (PCAOB) regulation, auditors exert greater effort on convertible debt, but no additional effort on straight debt. Our inferences are robust to using a change in audit fees specification, controlling for litigation risk, and controlling for functional form misspecification. Data Availability: Data are available from the public sources cited in the text. JEL Classifications: M42; G32.

Nonparametric Specification Testing of Conditional Asset Pricing Models

This article presents an adaptive omnibus specification test of asset pricing models where the stochastic discount factor is conditionally affine in the pricing factors. These models provide constraints that conditional moments of returns and pricing factors must satisfy, but most of them do not provide information on the functional form of those moments. Our test is robust to functional form misspecification, and also detects any relationship between pricing errors and conditioning variables. We give special emphasis to the test implementation and calibration, and extensive simulation studies prove the functioning in practice. Our empirical applications show a conditional counterpart of a well-known problem of unconditional models. The lack of rejection of consumption based conditional models seems to be due to a poor conditional correlation between consumption and stock returns.

The Environmental Kuznets Curve: A Semiparametric Approach with Cross-Sectional Dependence

This paper proposes a new approach to examine the relationship between CO2 emissions and economic developing. In particular, we propose to test the Environmental Kuznets Curve (EKC) hypothesis for a panel of 24 OECD countries and 32 non-OECD countries by developing a more flexible estimation technique which enables to account for functional form misspecification, cross-sectional dependence, and heterogeneous relationships among variables, simultaneously. We propose a new nonparametric estimator that extends the well-known Common Correlated Effect (CCE) approach from a fully parametric framework to a semiparametric panel data model. Our results corroborates that the nature and validity of the income–pollution relationship based on the EKC hypothesis depends on the model assumptions about the functional form specification. For all the countries analyzed, the proposed semiparametric estimator leads to non-monotonically increasing or decreasing relationships for CO2 emissions, depending on the level of economic development of the country.

Does the Feldstein–Horioka relationship vary with economic policy uncertainty?

ABSTRACT This research addresses the omitted-uncertainty bias and functional form misspecification problem when estimating the saving-investment relationship. We apply a semiparametric varying-coefficient cointegration approach to test for the presence of an ‘uncertainty-varying’ cointegrating relationship between savings and investment. Based on quarterly time series data for the U.S. from 1947Q1 to 2014Q4, we find that the saving-retention coefficient is a U-shaped function of economic policy uncertainty. This indicates that domestic investment is more constraint by domestic savings at very high and very low levels of uncertainty. Given the continued rise in economic policy uncertainty, our estimate predicts that domestic investment will become more constrained by domestic savings. Further sub-period analysis shows that the null hypothesis of the presence of an uncertainty-varying cointegrating relationship between savings and investment cannot be rejected in the recent sub-period of 1999Q1-2014Q4, suggesting that the U.S current account is sustainable.

Effect of corporate tax avoidance activities on firm bankruptcy risk

Corporate tax avoidance has been shown to raise the cost of bank debt and lower credit and bond ratings. However, it is unclear whether tax avoidance actually increases a firm’s bankruptcy risk or whether it is just viewed negatively by banks and rating agencies. We find that firms engaging in tax avoidance and firms that are thinly capitalized face higher bankruptcy risk. To account for endogeneity and functional form misspecification, we verify our results using instrumental variable and propensity score matching methods. Our findings are consistent with the view that tax avoidance is a risk-enhancing activity.

Modelling Healthcare Costs: A Semiparametric Extension of Generalised Linear Models

The empirical and methodological efforts in using the generalised linear model to model healthcare costs have been mostly concentrated on selecting the correct link and variance functions. Another type of misspecification - misspecification of functional form of the key covariates - has been largely neglected. In many cases, continuous variables enter the model in linear form. This means that the relationship between the covariates and the response variable is entirely determined by the link function chosen which can lead to biased results when the true relationship is more complicated. To address this problem, we propose a hybrid model incorporating the extended estimating equations (EEE) model and partially linear additive functions. More specifically, we partition the index function in the EEE model into a number of additive components including a linear combination of some covariates and unknown functions of the remaining covariates which are believed to enter the index non-linearly. The estimator for the new model is developed within the EEE framework and based on the method of sieves. Essentially, the unknown functions are approximated using basis functions which enter the model just like the other predictors. This minimises the need for programming as the estimation itself can be completed using existing EEE software programs. The new model and its estimation procedure are illustrated through an empirical example focused on how children’s Body Mass Index (BMI) z-score measured at 4-5 years old relates to their accumulated healthcare costs over a 5-year period. Results suggest our new model can reveal complex relationships between covariates and the response variable.

A Simple App to Teach Regression

This paper introduces a simple free web app that can teach regression to anyone who can point and click. Originally designed to teach Justice Department attorneys enough about regression so that they could cross examine rival experts, the app ``inverts'' the usual pedagogy: instead of showing users how to run regressions on data, it asks them to click on a graph to ``create'' data to achieve a given outcome, like a statistically significant line. Successful completion of each task is rewarded with immediate feedback that reveals the principle behind the exercise. This paper describes short, intuitive exercises to teach: (i) hypothesis testing, statistical significance and confidence intervals, (ii) the difference between correlation and causality, and (iii) how to diagnose functional form mis-specification. These exercises can be completed in just a few minutes.

Factors that Determine Aflatoxin Infestation in Ghana: Application of Cragg's Double Hurdle Model

Abstract In our application of Cragg's (double hurdle) model, we explicitly address functional form misspecification by first using nonparametric regression to identify the functional relationship between aflatoxin infestation and a vector of agronomic and socioeconomic variables. Based on the established functional forms, we then apply Cragg's model to estimate the determinants of farm-level aflatoxin infestation. As expected, we find that aflatoxin infestation in northern Ghana is determined by pre-production and post-production factors. However, we estimate insignificant effects for farmer characteristics. An assessment of the importance of functional form shows that an inappropriate function can lead to biased estimates of the Cragg model.

'Choosing Factors' by Fama and French (2018): A Comment

This paper features a statistical analysis of the monthly three factor Fama/French return series. We apply rolling OLS regressions to explore the relationship between the 3 factors, using monthly data from July 1926 to June 2018, that are available on Ken French's website. The results suggest there are significant and time varying relationships between the factors. We then switch to a sub-sample from July 1990 to July 2018, also taken from Ken French's website. The three series and their inter-relationships are analysed using two stage least squares and the Hausman test to check for issues related to endogeneity, the Sargan over-identification test and the Cragg-Donald weak instrument test. The relaationship between factors is also examined using OLS incorporating Ramsey's RESET tests of functional form misspecification, plus Naradaya-Watson kernel regression techniques. The empirical results suggest that the factors, when combined in OLS regression analysis, as suggested by Fama and French (2018), are likely to suffer from endogeneity. OLS regression analysis and the the application of Ramsey's RESET tests suggest a non-linear relationship exists between the three series, in which squared and cubed terms are significant. This non-linearity is also confirmed by the Kernel regression analysis. We use two instruments to estimate the market betas, and then use the factor estimates in a second set of panel data tests using a small sample of monthly returns for US firms that are drawn from the on-line data source “tingo”. These issues are analysed using methods suggested by Petersen (2009) to permit clustering in the panels by date and firm. The empirical results suggest that using an instrument to capture endogeneity reduces the standard error of market beta in subsequent cross-sectional tests, but that clustering effects, as suggested by Petersen (2009), will also impact on the estimated standard errors. The empirical results suggest that using these factors in linear regression analysis, such as suggested by Fama and French (2018), as a method of screening factor relevance, is problematic in that the estimated standard errors are highly sensitive to the correct model specification.

The Effect of Corruption on Energy Security and the Analysis of Outlying Countries in East Asia

With annual data for 2010-13 from fifty-five countries, we explore the possibility that differences in corruption are causal to differentials in energy security. For the purpose of policy formation diagnostic double-checkings are conducted to evaluate all the models in the sense that the more tests that are carried out, the less the chance of accepting a poor model. Among those performed are: the RESET tests for functional form misspecification and the tests for heteroskedasticity. The fixed effect (FE) estimates suggest that corruption has a detrimental effect on energy security. It is also evident from the Spearman rank correlation coefficient that corruption and energy security are negatively dependent on one another, implying that higher corruption correlates with lower energy security. The FE estimates suggest that corruption has a detrimental effect on energy security. Political stability energy price, energy use, energy imports, energy production, total carbon dioxide emissions, and military expenditure variables are controlled for. Each of the control variables presents a statistically significant coefficient with the predicted sign. Results from Granger causality test suggest that corruption is Granger-causal to energy security risk in both all countries as well as East Asian countries (i.e., China, Japan, and Korea). More specifically, the FE estimates also suggest that the harmful effects of corruption on energy security is greater in the three core East Asian countries than in other countries. This suggests that the three core East Asian countries can be regarded as outliers. The ceteris paribus mean elasticity of energy security risk with respect to corruption in East Asia, indicating that a 10 percent increase in corruption enhances energy security risk by 2.46 percent in East Asia. This implies that energy security risk is less sensitive to a change in corruption. Given the importance of energy security for sustainable development in East Asia, therefore, a key element of ensuring energy security is that for more transparent physical and financial energy markets the three East Asian countries enhance energy cooperation, to make energy market data more accurate and available, and to take steps to support the development of cleaner and more efficient energy technologies to enhance the efficiency of markets and shift towards a more sustainable energy future. For example, compliance with the United Nations Convention Against Corruption may be facilitated by reforms aimed at reducing corruption.

Pass‐Through and the Prediction of Merger Price Effects

We use Monte Carlo experiments to study how pass‐through can improve merger price predictions, focusing on the first order approximation (FOA) proposed in Jaffe and Weyl []. FOA addresses the functional form misspecification that can exist in standard merger simulations. We find that the predictions of FOA are tightly distributed around the true price effects if pass‐through is precise, but that measurement error in pass‐through diminishes accuracy. As a comparison to FOA, we also study a methodology that uses pass‐through to select among functional forms for use in simulation. This alternative also increases accuracy relative to standard merger simulation and proves more robust to measurement error.

Investment Crowding-Out: Firm-Level Evidence from Germany

The main objective of the present paper is to estimate the extent to which firm investment is substituted (crowded-out) by investment support policies granted under the EU Rural Development Programme (RDP). In the empirical analyses we employ the difference-in-difference propensity score matching approach, which allows us to address several important sources of bias, such as selection bias, the simultaneity bias, and functional form misspecification, from which many previous studies suffer. Using panel data of 1,333 firms from the Schleswig-Holstein region in Germany, we find that the crowding-out effect of the RDP is close to 100%, implying that firms use public support to substitute for private investments. Furthermore, no evidence was found that, due to RDP programme support, firms would have brought forward their investments planned originally in a later period, rejecting the f inter-temporal substitution of investments.

Investment Crowding Out: Firm-Level Evidence from Northern Germany

Michalek J., Ciaian P. and Kancs d’A. Investment crowding out: firm-level evidence from Northern Germany, Regional Studies. The main objective of this paper is to estimate the extent to which firm investment is substituted (crowded out) by investment support policies granted under the European Union Rural Development Programme (RDP). The empirical analyses employed the difference-in-difference propensity score-matching approach, which allows several important sources of bias to be addressed, such as selection bias, simultaneity bias and functional form misspecification, from which many previous studies suffer. Using panel data of 1333 firms from the Schleswig-Holstein region in Germany, it is found that the crowding-out effect of the RDP is close to 100%, implying that firms use public support to substitute for private investments. Furthermore, no evidence was found that, due to the RDP programme support, firms would have brought forward their investments planned originally for a later period, rejecting the inter-temporal substitution of investments.

Education, Intelligence, and Well-Being: Evidence from a Semiparametric Latent Variable Transformation Model for Multiple Outcomes of Mixed Types

This paper uses a semiparametric latent variable transformation model for multiple outcomes to examine the effect of education and maternal education on female multidimensional well-being and proposes a procedure to build a well-being index that is less susceptible to functional form misspecification. We model multidimensional well-being as an unobserved common factor underlying the observed well-being outcomes. The semiparametric methodology allows us to alleviate misspecification bias by combining multiple indicators into a latent construct in an unspecified, data-driven way. Using data from female participants of the 1974–2010 waves of the US General Social Survey, we find that education, intelligence, and maternal education contribute positively to multidimensional well-being. However, the effects of education and maternal education on female multidimensional well-being declined steadily between the mid-1970s and the 1990s, and have not rebounded since.

PRICING KERNEL ESTIMATION: A LOCAL ESTIMATING EQUATION APPROACH

This paper investigates a general semiparametric stochastic discount factor formulation that avoids functional form misspecification. A new semiparametric estimation procedure is proposed which combines orthogonality conditions and local linear fitting to give a semiparametric generalized estimating equation approach. Asymptotic properties of the estimators are established and we explore the empirical usefulness of the proposed approach to value-weighted stock returns.