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 PaperScience, R&D, And Invention Potential Recharge: U.S. Evidence
January 1993
Working Paper Number:
CES-93-02
The influence of academic science on industrial R&D seems to have increased in recent years compared with the pre-World War II period. This paper outlines an approach to tracing this influence using a panel of 14 R&D performing industries from 1961-1986. The results indicate an elasticity between real R&D and indicators of stocks of academic science of about 0.6. This elasticity is significant controlling for industry effects. However, the elasticity declines from its level during the 1961-1973 subperiod, when it was 2.2, to 0.5 during the 1974-1986 subperiod. Reasons for the decline include exogenous and endogenous exhaustion of invention potential, and declining incentives to do R&D stemming from a weakening of intellectual property rights. The growth of R&D since the mid-1980s suggests a restoration of R&D incentives in still more recent times.View Full Paper PDF
-
Working PaperComparing Earnings Outcome Differences Between All Graduates and Title IV Graduates
August 2021
Working Paper Number:
CES-21-19
Recently, two public data products have been released that publish earnings outcomes for college graduates by program of study and institution: Post-Secondary Employment Outcomes and College Scorecard, from the Census Bureau and U.S. Department of Education, respectively. While the earnings data underlying the data products is similar, persons eligible for the frames of the two products is different, with College Scorecard restricted to only students that receive Title IV aid. This paper documents how these differences in the population studied affect the published earnings outcomes. I show that at an institution, of the institutions in my sample, an average of sixty percent of baccalaureate graduates receive Title IV aid, and that the lower the coverage, the large the difference in earnings measurement. Additionally, I show that short-run earnings outcomes are very similar for these two samples, while longer-run outcomes (10 years after graduation) are significantly lower for the Title IV population. I also show that program ranking can change significantly when considering the Title IV population rather than the entire graduate population.View Full Paper PDF
-
Working PaperA Guide To R&D Data At The Center For Economic Studies U.S. Bureau Of THe Census
August 1994
Working Paper Number:
CES-94-09
The National Science Foundation R&D Survey is an annual survey of firms' research and development expenditures. The survey covers 3000 firms reporting positive R&D. This paper provides a description of the R&D data available at the Center for Economic Studies (CES). The most basic data series available contains the original survey R&D data. It covers the years 1972-92. The remaining two series, although derived from the original files, specialize in particular items. The Mandatory Series contains required survey items for the years 1973-88. Items reported at firms' discretion are in the Voluntary Series, which covers the years 1974-89. Both of the derived series incorporate flags that track quality of the data. Both also include corrections to the data based on original hard copy survey evidence stored at CES. In addition to describing each dataset, we offer suggestions to researchers wishing to use the R&D data in exploring various economic issues. We report selected response rates, discuss the survey design, and provide hints on how to use the data.View Full Paper PDF
-
Working PaperGender Segregation Small Firms
October 1992
Working Paper Number:
CES-92-13
This paper studies interfirm gender segregation in a unique sample of small employers. We focus on small firms because previous research on interfirm segregation has studied only large firms and because it is easier to link the demographic characteristics of employers and employees in small firms. This latter feature permits an assessment of the role of employer discrimination in creating gender segregation. Our first finding is that interfirm segregation is prevalent among small employers. Indeed men and women rarely work in fully integrated firms. Our second finding is that the education and gender of the business owner strongly influence the gender composition of a firm's workforce. This suggests that employer discrimination may be an important cause of workplace gender segregation. Finally, we estimate that interfirm segregation can account for up to 50% of the gender gap in annual earnings.View Full Paper PDF
-
Working PaperChanges in Workplace Segregation in the United States Between 1990 and 2000: Evidence from Matched Employer-Employee Data
June 2007
Working Paper Number:
CES-07-15
We present evidence on changes in workplace segregation by education, race, ethnicity, and sex, from 1990 to 2000. The evidence indicates that racial and ethnic segregation at the workplace level remained quite pervasive in 2000. At the same time, there was fairly substantial segregation by skill, as measured by education. Putting together the 1990 and 2000 data, we find no evidence of declines in workplace segregation by race and ethnicity; indeed, black-white segregation increased. Over this decade, segregation by education also increased. In contrast, workplace segregation by sex fell over the decade, and would have fallen by more had the services industry - a heavily female industry in which sex segregation is relatively high - not experienced rapid employment growth.View Full Paper PDF
-
Working PaperPirate's Treasure
January 2017
Working Paper Number:
CES-17-51
Do countries that improve their protection of intellectual property rights gain access to new product varieties from technologically advanced countries? We build the first comprehensive matched firm level data set on exports and patents using confidential microdata from the US Census to address this question. Across several different estimation approaches we find evidence that these protections affect where US firms export.View Full Paper PDF
-
Working PaperReleasing Earnings Distributions using Differential Privacy: Disclosure Avoidance System For Post Secondary Employment Outcomes (PSEO)
April 2019
Working Paper Number:
CES-19-13
The U.S. Census Bureau recently released data on earnings percentiles of graduates from post secondary institutions. This paper describes and evaluates the disclosure avoidance system developed for these statistics. We propose a differentially private algorithm for releasing these data based on standard differentially private building blocks, by constructing a histogram of earnings and the application of the Laplace mechanism to recover a differentially-private CDF of earnings. We demonstrate that our algorithm can release earnings distributions with low error, and our algorithm out-performs prior work based on the concept of smooth sensitivity from Nissim, Raskhodnikova and Smith (2007).View Full Paper PDF
-
Working PaperCapturing More Than Poverty: School Free and Reduced-Price Lunch Data and Household Income
December 2017
Working Paper Number:
carra-2017-09
Educational researchers often use National School Lunch Program (NSLP) data as a proxy for student poverty. Under NSLP policy, students whose household income is less than 130 percent of the poverty line qualify for free lunch and students whose household income is between 130 percent and 185 percent of the poverty line qualify for reduced-price lunch. Linking school administrative records for all 8th graders in a California public school district to household-level IRS income tax data, we examine how well NSLP data capture student disadvantage. We find both that there is substantial disadvantage in household income not captured by NSLP category data, and that NSLP categories capture disadvantage on test scores above and beyond household income.View Full Paper PDF
-
Working PaperBusiness Dynamics Statistics of High Tech Industries
January 2016
Working Paper Number:
CES-16-55
Modern market economies are characterized by the reallocation of resources from less productive, less valuable activities to more productive, more valuable ones. Businesses in the High Technology sector play a particularly important role in this reallocation by introducing new products and services that impact the entire economy. Tracking the performance of this sector is therefore of primary importance, especially in light of recent evidence that suggests a slowdown in business dynamism in High Tech industries. The Census Bureau produces the Business Dynamics Statistics (BDS), a suite of data products that track job creation, job destruction, startups, and exits by firm and establishment characteristics including sector, firm age, and firm size. In this paper we describe the methodologies used to produce a new extension to the BDS focused on businesses in High Technology industries.View Full Paper PDF
-
Working PaperSpecialization, Firms, and Markets: The Division of Labor Within and Between Law Firms
May 2003
Working Paper Number:
CES-03-13
What is the role of firms and markets in mediating the division of labor? This paper uses confidential microdata from the Census of Services to examine law firms' boundaries. We find that firms' field scope narrows as market size increases and individuals specialize, indicating that firms' boundaries reflect organizational trade-offs. Moreover, we find that whether the division of labor is mediated by firms differs systematically according to whether lawyers in a particular field are mainly involved in structuring transactions or in dispute resolution. Our evidence is consistent with hypotheses in which firms' boundaries reflect variation in the value of knowledge-sharing or in the costs of monitoring, but not in risk-sharing. Our findings show how the incentive trade-offs associated with exploiting increasing returns from specialization help lead the structure of the industry to be fragmented, but highly-skewed.View Full Paper PDF
