Poisson Quasi-maximum Likelihood Research Articles

Epidemic change-point detection in general integer-valued time series

ABSTRACT In this paper, we consider the structural change in a class of discrete valued time series, where the true conditional distribution of the observations is assumed to be unknown. The conditional mean of the process depends on a parameter θ ∗ which may change over time. We provide sufficient conditions for the consistency and the asymptotic normality of the Poisson quasi-maximum likelihood estimator (QMLE) of the model. We consider an epidemic change-point detection and propose a test statistic based on the QMLE of the parameter. Under the null hypothesis of a constant parameter (no change), the test statistic converges to a distribution obtained from increments of a Browninan bridge. The test statistic diverges to infinity under the epidemic alternative, which establishes that the proposed procedure is consistent in power. The effectiveness of the proposed procedure is illustrated by simulated and real data examples.

A NEGATIVE BINOMIAL AUTOREGRESSION WITH A LINEAR CONDITIONAL VARIANCE-TO-MEAN FUNCTION

A general integer-valued time-series model with a conditional variance proportional to the conditional mean is proposed. Specifically, the conditional distribution is a Poisson mixture with a dependent mixing sequence, which results in a negative binomial distribution with a linear conditional variance-to-mean relationship. In addition, the conditional mean is specified as a general parametric function of past observations. We first propose stationarity, ergodicity, and finite moment conditions for the model. Furthermore, the parameters are estimated using the Poisson quasi-maximum likelihood estimate, whose asymptotic properties are studied under weak conditions. Illustrations of the proposed methodology on simulated and actual time series of counts are given.

Poisson QMLE for change-point detection in general integer-valued time series models

We consider together the retrospective and the sequential change-point detection in a general class of integer-valued time series. The conditional mean of the process depends on a parameter \(\theta ^*\) which may change over time. We propose procedures which are based on the Poisson quasi-maximum likelihood estimator of the parameter, and where the updated estimator is computed without the historical observations in the sequential framework. For both the retrospective and the sequential detection, the test statistics converge to some distributions obtained from the standard Brownian motion under the null hypothesis of no change and diverge to infinity under the alternative; that is, these procedures are consistent. Some results of simulations as well as real data application are provided.

Statistical Myths about Log-Transformed Dependent Variables: When, Why, and How

Strategy and management researchers often employ log-transformed dependent variables (LTDVs). Such a transformation means that researchers estimate an exponential model, which implies estimating a non-linear relationship. We review 10 years of research published in SMJ and find the wide use of LTDVs to be based on statistical myths with possible detrimental effects for the validity of research findings. We identify three myths of major concern that are widely shared: 1) LTDVs should be used to make distributions more normal, 2) linear hypotheses can be tested with LTDVs, and 3) LTDVs are the best way to estimate an exponential model. We find that many researchers use LTDVs for the wrong reasons, and very often in a way that is misaligned with the hypothesis they intend to examine. Researchers also appear unaware of the severe shortcomings of LTDVs. We call researchers to exhibit caution when planning to use LTDVs. They are not relevant when a linear hypothesis is tested, and when the interest is in modeling exponential relationships, the Poisson Quasi Maximum Likelihood Estimator is a better alternative.

Impacts of regional integration and maritime transport on trade: with special reference to RCEP

Purpose The proposed Regional Comprehensive Economic Partnership (RCEP) will become a large trade agreement in Asia, which has brought together the ten members of Association of Southeast Asian Nations (ASEAN) and five of the neighbors’ countries. Under the trend of globalization, the progress of the transportation industry and regional integration will increase the volume of trade, therefore maritime performance is intrinsically linked to trade. In fact, few studies have examined regional integration in the context of seaborne. This paper aims to use the cluster analysis and Poisson quasi-maximum likelihood (PQML) gravity model to investigate the trading bloc phenomenon and relation between trade and marine transportation. Design/methodology/approach In this paper, hierarchical clustering analysis and tree diagrams are used to identify functional areas characterized by bilateral trade intensity and bilateral liner shipping connectivity indices. Regional reorganizations that have occurred within Asian countries were studied. This study illustrates that these trading blocs have a positive impact on trade when maritime transport, production and trading networks have developed between regions. A gravity model was constructed using worldwide trade data for 2007, 2010 and 2015. The study considered free trade agreement (FTA)/common market (CM) of EU, RCEP and North American Free Trade Agreement (NAFTA) as regional dummies and designed a real trade bloc induction variable. In addition, the study did not use the commonly adopted ordinary least squares (OLS) estimation but used the PQML method to estimate the gravity equation to overcome the problem of a large number of zero trade observations. Preliminary results show that regional integration cannot guarantee the establishment of intraregional trade but depends on the stage of economic development and regional industrial characteristics. Findings The major findings are summarized as follows. Both liner shipping connectivity and logistics performance have significant advantages with positive coefficients in each regression results. The creation of intraregional trade is not guaranteed, depending on the characteristics of the trade and the stage of economic development of the region. For RCEP, the effect created by intra-regional trade is better than the EU. Instead, the “nominal” intra-RCEP trade was significantly below the “real” trading blocs. For RCEP, the effect created by intra-regional trade is better than that of the EU. Instead, “nominal” intra-RCEP trade is much lower than “real” trading blocs. The real trading bloc between East Asia and Taiwan clearly exists, and the bloc phenomenon is becoming more and more significant. This result shows that Taiwan’s trade flow with East Asia is higher than the normal level relationship implied by its corresponding economic and geographical conditions. Originality/value This paper focuses on new empirical work done for this study is on the potential impact on trade. Earlier studies that have discussed and/or provided estimates of the benefits to the RCEP plan from improved transport and supply chain connectivity are cited. Marine transportation performance inherently links to economies of commerce. Few studies have examined regional integration in the context of maritime transportation, which reflects the lack of a mix of trade economists and maritime logistics research in the existing literature. This paper attempts to investigate the trading bloc phenomenon formed by regional integration (such as RCEP) and the relation between trade and marine transportation. With the official entry into force of the RCEP in 2020, it will promote increased trade and demand for logistics and maritime transport services in East Asia.

Poisson Quasi-Maximum Likelihood Estimator-based CUSUM Test for Integer-Valued Time Series

This study considers the parameter change test for integer-valued time series models based on the Poisson quasi-maximum likelihood estimates. As a change point test, we consider the score vector-based CUSUM test and show that its limiting null distribution takes the form of a function of Brownian bridges. Moreover, the residual-based CUSUM tests are considered as alternatives. For evaluation, we conduct a Monte Carlo simulation study with Poisson, zero-inflated Poisson, negative binomial and Conway-Maxwell integer-valued generalized autoregressive conditional heteroscedastic models andPoisson integer-valued autoregressive models, and compare the performance of the proposed CUSUM tests. Our findings confirm that the proposed test is a functional tool for detecting a change point when the underlying distribution is unspecified.

Aggression and sleep: a daylight saving time natural experiment on the effect of mild sleep loss and gain on assaults

The purpose of this study was to test the effect of a mild, short-term sleep loss/gain on assault rates. Using National Incidence Based Reporting System data and city-reported data from Chicago, New York, Philadelphia, and Los Angeles, we calculated the difference in assault rates on the Monday immediately following daylight saving time (DST) as compared to the Monday a week later using a Poisson quasi-maximum likelihood estimator model. The same analyses were performed to examine effects of the return to standard time in the fall. We employed several falsification checks. There were 2.9% fewer (95% CI: –4.2%, −1.6%, p < 0.0001) assaults immediately following DST, when we lose an hour, as compared to a week later. In contrast, there was a 2.8% rise in assaults immediately following the return to standard time, when an hour is gained, as compared to a week later (95% CI: 1.5%, 4.2%, p < 0.0001). Multiple falsification analyses suggest the spring findings to be robust, while the evidence to support the fall findings is weaker. This study suggests that mild and short-term changes in sleep do significantly affect rates of assault. Specifically, there is support for the theory that mild sleepiness possibly associated with an hour loss of sleep results in reduced assaults. This contradicts the simple inverse relationship currently suggested by most of the correlational literature. This study and the mixed findings presented by experimental studies indicate that measurement variability of both sleep and aggression may result in conflicting findings.

Determinants of pharmaceutical innovation: the role of technological opportunities revisited

Recent empirical contributions emphasize the importance of (potential) market size for the development of new pharmaceuticals. At the same time many scholars point out the importance of of scientific advances for the industry’s R&D activities. Against this background I analyze the relationship between (potential) market size, technological opportunities, and the number of new pharmaceuticals in the United States. Technological opportunities are operationalized as growth rates of the relevant knowledge stock as proposed by Andersen (1999, 1998). I analyze a unique d ataset by using an “entry stock” Poisson quasi-maximum likelihood estimator. The results reveal a rather robust and significantly positive response of the number of new pharmaceuticals, i.e., new molecular entities or new drug approvals, to market size and technological opportunities.

Score-driven dynamic patent count panel data models

In this paper, we propose the use of Dynamic Conditional Score (DCS) count panel data models. We compare the statistical performance of the static model with different dynamic models: finite distributed lag, exponential feedback and different DCS models. For DCS, we consider random walk or quasi-autoregressive dynamics. We use panel data for a large cross section of United States firms for period 1979–2000, and the Poisson quasi-maximum likelihood estimator with fixed effects. The empirical results suggest that DCS has the best statistical performance.

Regional Trade Agreements and Bilateral Trade: An Eclectic View of the Hidden Trade Costs

Trade costs associated with institutional failures and political risks, continue to serve as a big barrier to regional and international trade. Despite this strong impediment, limited work has been done to investigate the effect of weak institutions and type of political regime on bilateral trade. In an attempt to bolster this literature, this paper investigates the effects of these hidden trade costs on bilateral trade. The paper employs an augmented gravity equation derived from a monopolistic competition model of product differentiation to examine this effect for a sample of 46 countries in three distinct regional trade blocs. Using the Poisson Quasi Maximum Likelihood, the Heckman two-stage, Tobit and Negative Binomial estimation techniques, this paper concluded that stronger institutions of trading partners leads to an increase in trade. I argue that stronger institutions of both the exporting and importing countries in the form of good regulatory and legal framework reduce the uncertainty in contract enforcement, transactions costs and ultimately increase trade. Differences in political regimes of trading partners also have a strong effect on trade. Results also indicate that regional integration boosts bilateral trade, but there is evidence of trade diversion in WAMZ and MRU trading blocs.The proposition that geographically proximate partners traded more is supported according to the results, but there is little or no empirical support for the hypothesis that differences in factor endowments predict patterns of trade in the sample of countries. The sensitivity and robustness tests results indicate that the parameter estimates are robust to different estimation techniques and specification of the gravity model.

Extent and drivers of intra-regional trade in food products in ECOWAS

Agriculture is the mainstay of the economies in most of ECOWAS countries. Ironically, however, persistent food shortage arising from constraints on agriculture is exposing many of them to threats of extreme hunger and starvation. Trade in food products has been identified as a very critical factor in safeguarding food security for the world's poor. This study therefore examined the extent and drivers of intra-regional trade in food products in ECOWAS and the two monetary zones (WAEMU and WAMZ) in the sub-region. Poisson Quasi Maximum Likelihood of Santos-Silva and Tenreyro was used to estimate the gravity model. While exporter income discourages intra-regional trade in food products, importer income promotes it. Landed area irrigated and the land available for cultivation promotes intra-regional trade in food products in ECOWAS. It is only in WAMZ that colonial ties and border enhance intra-regional trade in food products.

The Comparison of the Computer Numbers in the Houses in Turkey with the Pooled Count Data Models&lt;sup&gt;1&lt;/sup&gt;

The effect of informatics technologies upon our lives gradually increases in parallel with the significant inclusion of computer and internet in every area of life. As technological developments have a positive effect upon the informatics sector, the use of computers has become widespread in houses. The objective of this study is to explain the factors affecting the number of computers owned by the household with the help of Counting Data Models. In this context, the most convenient method was tried to be determined through comparing the Standard Poisson, Poisson Quasi Maximum Likelihood and Negative Binom regression models. Being prepared by the Turkish Statistical Institute (TUIK) between 2002-2010, the data of the ‘Household Budget Survey’ were used in the study.

The Hybrid Open Access Citation Advantage: How Many More Cites is a $3,000 Fee Buying You?

We study the hybrid open access (HOA) citation effect. Under HOA pilot agreements, HOA is assigned for all articles of eligible authors. We use unique data on 208 (1,121) HOA (closed access) economics articles. We control for the quality of journals, articles and institutions and citations to RePec pre-prints. Performing Poisson quasi-maximum likelihood regressions, HOA turns out to be a significant predictor of citations with marginal effects ranging between 22% and 26%. However, once we additionally control for institution quality and citations to RePec pre-prints, the marginal HOA citation advantage turns out to be insignificant and drops to 0.4%.

The Poisson quasi-maximum likelihood estimator: a solution to the ‘adding up’ problem in gravity models

This article shows that the Poisson Quasi-Maximum Likelihood (QML) estimator applied to the gravity model produces estimates in which, summing across all partners, actual and estimated total trade flows are identical. Other methods such as OLS do not have this desirable property. Indeed, Poisson is the only QML estimator that preserves total trade flows. This result is an additional reason for preferring Poisson as a workhorse gravity model estimator.

RESEARCH UNIVERSITIES AND REGIONAL HIGH‐TECH FIRM START‐UP AND EXIT

If localized knowledge spillovers are present in the university setting, higher rates of both start‐up and/or survival would be observed in areas that are geographically proximate to the university. Using a detailed industry data set for Texas for 1999:3–2005:2, we analyze start‐up and exit rates for high‐tech firms. Based on a Poisson quasi‐maximum likelihood estimation, we find evidence that the level of R&amp;D and proximity of a research institution positively affects the likelihood of technology start‐ups. However, using both the Cox proportional hazards model and Kaplan–Meier approach, our results suggest that geographic proximity to knowledge centers does not reduce hazard rates. (JEL R12, R53, O18)

Research Universities and Regional High-Tech Firm Start-up and Exit

If localized knowledge spillovers are present in the university setting, higher rates of both start-up and/or survival would be observed in areas that are geographically proximate to the university. Using a detailed industry dataset for Texas for 1999:3-2005:2, we analyze start-up and exit rates for high-tech firms. Based on a Poisson Quasi-Maximum Likelihood estimation, we find evidence that the level of R&D and proximity of a research institution positively affects the likelihood of technology start-ups. However, using both the Cox Proportional Hazards Model and Kaplan-Meier approach, our results suggest that geographic proximity to knowledge centers does not reduce hazard rates.

Estimating gravity equations: to log or not to log?

In this study we compare the traditional OLS approach applied to the log-linear form of the gravity model with the Poisson Quasi Maximum Likelihood (PQML) estimation procedure applied to the non-linear multiplicative specification of the gravity model. We use the trade flows for all products, for all manufacturing products as well as for manufacturing products broken down by three-digit ISIC Rev.2 categories. We base our conclusions on the gravity model of Bergstrand (Rev Econ Stat 71(1):143--153, 1989) for disaggregate trade flows that allows us to investigate differences in factor-proportions and home-market effects at the industry level. In addition, we compare the effects of other explanatory variables such as exporter and importer total income, distance, preferential trade agreements, common border, historical ties, and common language on the volume of trade. Our main conclusion is that both estimation results as well as results of the regression mis-specification tests provide supporting evidence for the PQML estimation approach over the OLS estimation method.

Using the gravity equation to differentiate among alternative theories of trade: another look

This study compares two alternative approaches–the traditional ordinary least square (OLS) and the Poisson Quasi Maximum Likelihood (PQML) procedures–to estimation of the home market effects for the heterogeneous, the reference priced and the homogeneous goods in the gravity model derived in Feenstra et al. (2001). Despite our findings that the PQML estimation results oftentimes (significantly) differ from those obtained by the OLS procedure, both estimation procedures largely lead to the qualitatively similar conclusions on the existence of the home market effect for the different types of goods with several exceptions.