Analytic Derivatives for Estimation of Linear Dynamic Models
November 1988
Working Paper Number:
CES-88-05
Abstract
This paper develops two algorithms. Algorithm I computes the exact, Gaussian, log-likelihood function, its exact, gradient vector, and an asymptotic approximation of its Hessian matrix, for discrete-time, linear, dynamic models in state-space form. Algorithm 2, derived from algorithm I, computes the exact, sample, information matrix of this likelihood function. The computed quantities are analytic (not numerical approximations) and should, therefore, be useful for reliably, quickly, and accurately: (i) checking local identifiability of parameters by checking the rank of the information matrix; (ii) using the gradient vector and Hessian matrix to compute maximum likelihood estimates of parameters with Newton methods; and, (iii) computing asymptotic covariances (Cramer-Rao bounds) of the parameter estimates with the Hessian or the information matrix. The principal contribution of the paper is algorithm 2, which extends to multivariate models the univariate results of Porat and Friedlander (1986). By relying on the Kalman filter instead of the Levinson-Durbin filter used by Porat and Friedlander, algorithms 1 and 2 can automatically handle any pattern of missing or linearly aggregated data. Although algorithm 1 is well known, it is treated in detail in order to make the paper self contained.
Document Tags and Keywords
Keywords
Keywords are automatically generated using KeyBERT, a powerful and innovative
keyword extraction tool that utilizes BERT embeddings to ensure high-quality and contextually relevant
keywords.
By analyzing the content of working papers, KeyBERT identifies terms and phrases that capture the essence of the
text, highlighting the most significant topics and trends. This approach not only enhances searchability but
provides connections that go beyond potentially domain-specific author-defined keywords.
:
Tags
Tags are automatically generated using a pretrained language model from spaCy, which excels at
several tasks, including entity tagging.
The model is able to label words and phrases by part-of-speech,
including "organizations." By filtering for frequent words and phrases labeled as "organizations", papers are
identified to contain references to specific institutions, datasets, and other organizations.
:
Similar Working Papers
Similarity between working papers are determined by an unsupervised neural
network model
know as Doc2Vec.
Doc2Vec is a model that represents entire documents as fixed-length vectors, allowing for the
capture of semantic meaning in a way that relates to the context of words within the document. The model learns to
associate a unique vector with each document while simultaneously learning word vectors, enabling tasks such as
document classification, clustering, and similarity detection by preserving the order and structure of words. The
document vectors are compared using cosine similarity/distance to determine the most similar working papers.
Papers identified with 🔥 are in the top 20% of similarity.
The 10 most similar working papers to the working paper 'Analytic Derivatives for Estimation of Linear Dynamic Models' are listed below in order of similarity.
-
Working PaperForeign-Born and Native-Born Migration in the U.S.: Evidence from IRS Administrative and Census Survey Records
July 2018
Working Paper Number:
carra-2018-07
This paper details efforts to link administrative records from the Internal Revenue Service (IRS) to American Community Survey (ACS) and 2010 Census microdata for the study of migration among foreign-born and native-born populations in the United States. Specifically, we (1) document our linkage strategy and methodology for inferring migration in IRS records; (2) model selection into and survival across IRS records to determine suitability for research applications; and (3) gauge the efficacy of the IRS records by demonstrating how they can be used to validate and potentially improve migration responses for native-born and foreign-born respondents in ACS microdata. Our results show little evidence of selection or survival bias in the IRS records, suggesting broad generalizability to the nation as a whole. Moreover, we find that the combined IRS 1040, 1099, and W2 records may provide important information on populations, such as the foreign-born, that may be difficult to reach with traditional Census Bureau surveys. Finally, while preliminary, the results of our comparison of IRS and ACS migration responses shows that IRS records may be useful in improving ACS migration measurement for respondents whose migration response is proxy, allocated, or imputed. Taking these results together, we discuss the potential application of our longitudinal IRS dataset to innovations in migration research on both the native-born and foreign-born populations of the United States.View Full Paper PDF
-
Working PaperCross Sectional Variation In Toxic Waste Releases From The U.S. Chemical Industry
August 1994
Working Paper Number:
CES-94-08
This paper measures and examines the 1987 cross sectional variation in toxic releases from the U.S. chemical industry. The analysis is based on a unique plant level data set of over 2,100 plants, combining EPA toxic release data with Census Bureau data on economic activity. The main results are that intra-industry variation in toxic releases are as great as, or greater, than inter-industry variation, and that plant, firm, and regulatory characteristics are important factors in explaining observed variation in toxic releases. Even after controlling for primary product and plant characteristics, there are some firms that generate significantly lower toxic waste due to managerial ability and/or technology differences.View Full Paper PDF
-
Working PaperPrimary Versus Secondary Production Techniques in U.S. Manufacturing
October 1994
Working Paper Number:
CES-94-12
In this paper we discuss and analyze a classical economic puzzle: whether differences in factor intensities reflect patterns of specialization or the co-existence of alternative techniques to produce output. We use observations on a large cross-section of U.S. manufacturing plants from the Census of Manufactures, including those that make goods primary to other industries, to study differences in production techniques. We find that in most cases material requirements do not depend on whether goods are made as primary products or as secondary products, which suggests that differences in factor intensities usually reflect patterns of specialization. A few cases where secondary production techniques do differ notably are discussed in more detail. However, overall the regression results support the neoclassical assumption that a single, best-practice technique is chosen for making each product.View Full Paper PDF
-
Working PaperAge, Sex, and Racial/Ethnic Disparities and Temporal-Spatial Variation in Excess All-Cause Mortality During the COVID-19 Pandemic: Evidence from Linked Administrative and Census Bureau Data
May 2022
Working Paper Number:
CES-22-18
Research on the impact of the COVID-19 pandemic in the United States has highlighted substantial racial/ethnic disparities in excess mortality, but reports often differ in the details with respect to the size of these disparities. We suggest that these inconsistencies stem from differences in the temporal scope and measurement of race/ethnicity in existing data. We address these issues using death records for 2010 through 2021 from the Social Security Administration, covering the universe of individuals ever issued a Social Security Number, linked to race/ethnicity responses from the decennial census and American Community Survey. We use these data to (1) estimate excess all-cause mortality at the national level and for age-, sex-, and race/ethnicity-specific subgroups, (2) examine racial/ethnic variation in excess mortality over the course of the pandemic, and (3) explore whether and how racial/ethnic mortality disparities vary across states.View Full Paper PDF
-
Working PaperAllocation of Company Research and Development Expenditures to Industries Using a Tobit Model
November 2015
Working Paper Number:
CES-15-42
This paper uses Census microdata and a regression-based approach to assign multi-division firms' pre-2008 Research and Development (R&D) expenditures to more than one industry. Since multi-division firms conduct R&D in more than one industry, assigning R&D to corresponding industries provides a more accurate representation of where R&D actually takes place and provides a consistent time-series with the National Science Foundation R&D by line of business information. Firm R&D is allocated to industries on the basis of observed industry payroll, as befits the historic importance of payroll in Census assignments of firms to industry. The results demonstrate that the method of assigning R&D to industries on the basis of payroll works well in earlier years, but becomes less effective over time as firms outsource their manufacturing function.View Full Paper PDF
-
Working PaperMatching State Business Registration Records to Census Business Data
January 2020
Working Paper Number:
CES-20-03
We describe our methodology and results from matching state Business Registration Records (BRR) to Census business data. We use data from Massachusetts and California to develop methods and preliminary results that could be used to guide matching data for additional states. We obtain matches to Census business records for 45% of the Massachusetts BRR records and 40% of the California BRR records. We find higher match rates for incorporated businesses and businesses with higher startup-quality scores as assigned in Guzman and Stern (2018). Clerical reviews show that using relatively strict matching on address is important for match accuracy, while results are less sensitive to name matching strictness. Among matched BRR records, the modal timing of the first match to the BR is in the year in which the BRR record was filed. We use two sets of software to identify matches: SAS DQ Match and a machine-learning algorithm described in Cuffe and Goldschlag (2018). We find preliminary evidence that while the ML-based method yields more match results, SAS DQ tends to result in higher accuracy rates. To conclude, we provide suggestions on how to proceed with matching other states' data in light of our findings using these two states.View Full Paper PDF
-
Working PaperMortality in a Multi-State Cohort of Former State Prisoners, 2010-2015
February 2022
Working Paper Number:
CES-22-06
Previous studies report that individuals who have been imprisoned have higher mortality rates than their demographic counterparts in the general population, particularly non-Hispanic white former prisoners. Most of these studies have been based on a single state's prison system, and the extent to which their findings can be generalized has not been established. In this study we explore the role that race/Hispanic origin, other demographic characteristics, and custodial/ criminal history factors have on post-release mortality, including on the timing of deaths. We also assess whether conditional release to community supervision or reimprisonment may explain the higher post-release mortality found among non-Hispanic whites. In the second part of the analysis, we estimate standardized mortality ratios (SMRs) by sex, age group, and race/Hispanic origin using as reference the U.S. general population. The data come from state prison releases from the Bureau of Justice Statistics' (BJS) National Corrections Reporting Program (NCRP). The NCRP records were linked to the Census Numident to identify deaths occurring within five years from prison release. We also linked NCRP records to previous decennial censuses and survey responses to obtain self-reported race and Hispanic origin if available. We found that non-Hispanic white former prisoners were more likely to die within five years after prison release and more likely to die in the initial weeks after release compared to racial minorities and Hispanics. Reimprisonment, age at release, and a history of multiple prison terms had a similar influence on the odds of dying across all race/Hispanic origin groups. Other factors, such as the type of release and the duration of the last term in prison, were associated with higher risks of mortality for some groups but not for others.View Full Paper PDF
-
Working Paper2010 American Community Survey Match Study
July 2014
Working Paper Number:
carra-2014-03
Using administrative records data from federal government agencies and commercial sources, the 2010 ACS Match Study measures administrative records coverage of 2010 ACS addresses, persons, and persons at addresses at different levels of geography as well as by demographic characteristics and response mode. The 2010 ACS Match Study represents a continuation of the research undertaken in the 2010 Census Match Study, the first national-level evaluation of administrative records data coverage. Preliminary results indicate that administrative records provide substantial coverage for addresses and persons in the 2010 ACS (92.7 and 92.1 percent respectively), and less extensive though substantial coverage, for person-address pairs (74.3 percent). In addition, some variation in address, person and/or person-address coverage is found across demographic and response mode groups. This research informs future uses of administrative records in survey and decennial census operations to address the increasing costs of data collection and declining response rates.View Full Paper PDF
-
Working PaperEvaluation And Use Of The Pollution Abatement Costs And Expenditures Survey Micro Data
January 1996
Working Paper Number:
CES-96-01
The Pollution Abatement Costs and Expenditures Survey (PACE) is an annual survey of manufacturing establishment=s operating costs and capital investment expenditures for pollution abatement purposes. This paper provides a description and evaluation of the PACE micro data available at the Center for Economic Studies (CES). The paper provides an overview of the survey, how the sample is drawn, how the survey questionnaire has changed over time, an assessment of the data quality, and suggestions for the use of the data, as well as its limitations. Also included are suggestions for modifying the survey design and data processing procedures. The PACE data series, linked to the economic data in CES= Longitudinal Research Database (LRD), covers the years 1979-1993, excluding 1983 and 1987.View Full Paper PDF
-
Working PaperForeign vs. U.S. Graduate Degrees: The Impact on Earnings Assimilation and Return Migration for the Foreign Born
June 2019
Working Paper Number:
CES-19-17
Using a novel panel data set of recent immigrants to the U.S., we identify return migration rates and earnings trajectories of two immigrant groups: those with foreign graduate degrees and those with a U.S. graduate degree. We focus on immigrants (of both genders) to the U.S. who arrive in the same entry cohort and from the same country of birth over the period 2005-2015. In Census-IRS administrative data, we find that downward earnings trajectories are predictive of return migration for immigrants with degrees acquired abroad. Meanwhile, immigrants with U.S.-acquired graduate degrees experience mainly upward earnings mobility.View Full Paper PDF
