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 PaperThe Dynamics of House Price Capitalization and Locational Sorting: Evidence from Air Quality Changes
September 2012
Working Paper Number:
CES-12-22
Despite extensive use of housing data to reveal valuation of non-market goods, the process of house price capitalization remains vague. Using the restricted access American Housing Survey, a high-frequency panel of prices, turnover, and occupant characteristics, this paper examines the time path of capitalization and preference-based sorting in response to air quality changes caused by differential regulatory pressure from the 1990 Clean Air Act Amendments. The results demonstrate that owner-occupied units capitalize changes immediately, whereas rent capitalization lags. The delayed but sharp rent capitalization temporally coincides with evidence of sorting, suggesting a strong link between location choices and price dynamics.View Full Paper PDF
-
Working PaperHow is Value Created in Spin-Offs? A Look Inside the Black Box
July 2005
Working Paper Number:
CES-05-09
Using a unique sample of plant level data from the Longitudinal Research Database (LRD), we identify (for the first time in the literature), how (the precise channel and mechanism), where (parent or subsidiary), and when (the dynamic pattern) performance improvements arise following corporate spinoffs. We identify the source of value improvements in spin-offs by comparing the magnitude of post-spinoff changes in the wages, employment, materials costs, rental and administrative expenses, sales, and capital expenditures in the plants belonging to firms undergoing spin-offs relative to the magnitude of such changes in a control group of plants belonging to firms not undergoing spin-offs. We show that the total factor productivity (TFP) of plants belonging to spin-off firms (parent or spun-off subsidiary) increase, on average, following the spin-off. This increase in overall productivity begins immediately, starting with the first year following the spin-off, and continuing in the years thereafter. This performance improvement can be attributed to a decrease in workers' wages, employment at the plant, decrease in the cost of materials purchased, as well as a decrease in rental and office expenditures, but not from improved product market performance by these plants. Further, such productivity improvements arise primarily in plants that remain with the parent; plants belonging to the spun-off subsidiary do not experience such productivity increases. However, contrary to speculation in the previous literature, plants that are spun-off do not underperform parent plants prior to the spin-off. Finally, in our split-sample study of plants that were acquired subsequent to the spin-off and those that were not, we find that productivity increases for both groups of plants: while such productivity increases start immediately after the spin-off for the nonacquired plants, for the acquired plants they occur only after being taken over by a better management team.View Full Paper PDF
-
Working PaperTHE RELATIONSHIPS AMONG ACQUIRING AND ACQUIRED FIRMS' PRODUCT LINES
September 1990
Working Paper Number:
CES-90-12
This study develops detailed information on the relationships among the activities of acquiring and acquired firms at and near the time of merger for a sample of 94 takeovers undertaken between 1977-1982. We focus on takeovers for two reasons. First, takeovers are an important and controversial phenomenon. Second, takeovers allow us to look at marginal changes, admittedly large ones, in the firm's boundaries. Thus, they provide a useful way of examining relationships among activities of the firm without having to go into great detail regarding the historical decisions that generated the firm's current structure. While the individual establishment is our basic data unit, in this study we aggregate the activities of the firm to the line of business (LOB) level. Each LOB of an acquired firm is classified as to its relationship horizontal, vertical (upstream or downstream), and conglomerate to the LOBs of the acquiring firm. Using these categorizations we aggregate the LOB-level information to the firm level to investigate the degree to which our sample of mergers is specialized to particular types of relationships. While we find a significant group of unspecialized takeovers, most appear to fit a specific category. We also look at the pattern of closed operations immediately following the takeover. Closings are generally concentrated in operations involving horizontal relationships. Finally, we consider the pattern of relationships between hostile and friendly takeovers and whether takeover premiums vary by type of merger. Merger premiums are not related to the type of relationship between the acquiring and acquired firm, but they are tied to whether the takeover is friendly or hostile.View Full Paper PDF
-
Working PaperNeighborhood Income and Material Hardship in the United States
January 2022
Working Paper Number:
CES-22-01
U.S. households face a number of economic challenges that affect their well-being. In this analysis we focus on the extent to which neighborhood economic conditions contribute to hardship. Specifically, using data from the 2008 and 2014 Survey of Income and Program Participation panel surveys and logistic regression, we analyze the extent to which neighborhoods income levels affect the likelihood of experiencing seven types of hardships, including trouble paying bills, medical need, food insecurity, housing hardship, ownership of basic consumer durables, neighborhood problems, and fear of crime. We find strong bivariate relationships between neighborhood income and all hardships, but for most hardships these are explained by other household characteristics, such as household income and education. However, neighborhood income retains a strong association with two hardships in particular even when controlling for a variety of other household characteristics: neighborhood conditions (such as the presence of trash and litter) and fear of crime. Our study highlights the importance of examining multiple measures when assessing well-being, and our findings are consistent with the notion that collective socialization and community-level structural features affect the likelihood that households experience deleterious neighborhood conditions and a fear of crime.View Full Paper PDF
-
Working PaperLong Term Effects of Military Service on the Distribution of Earnings
August 2009
Working Paper Number:
CES-09-17
I estimate the long term effect of military service on quantiles of earnings and education using the Vietnam draft lottery eligibility status as an instrument. I compare the local quantile treatment effect estimator studied by Abadie, Angrist, and Imbens (2002) to the instrumental variables quantile regression technique developed by Chernozhukov and Hansen (2008). Ordinary quantile regression shows a large negative association between service in Vietnam and earnings of white men, with the effect increasing in magnitude for the upper quantiles. Quantile treatment effects estimates show the opposite pattern, although much smaller in magnitude, with a small negative effect at the lower end of the distribution, and a small positive effect at the upper end. This suggests the ordinary quantile result is due to heterogeneous selection effects. The two methods of quantile treatment effects estimation give similar results.View Full Paper PDF
-
Working PaperIntroduction of Head Start and Maternal Labor Supply: Evidence from a Regression Discontinuity Design
January 2016
Working Paper Number:
CES-16-35
I use the non-public decennial censuses in 1970 to investigate the effect of the Head Start program on maternal labor supply and schooling in its early years. I exploit a discontinuity in county-level Head Start funding beginning in the late 1960s to explore differences in countylevel maternal employment and maternal schooling. The results provide suggestive evidence that the more availability of Head Start led to an increase the nursery school enrollment of children and a decrease in maternal labor supply. In addition, the ITT estimates imply a relatively large, negative effect of enrollment on maternal labor supply. However, the estimates are somewhat sensitive to addition of covariates and the standard errors are also large to draw firm inferences.View Full Paper PDF
-
Working PaperLearning by Doing and Plant Characteristics
August 1996
Working Paper Number:
CES-96-05
Learning by doing, especially spillover learning, has received much attention lately in models of industry evolution and economic growth. The predictions of these models depend on the distribution of learning abilities and knowledge flows across firms and countries. However, the empirical literature provides little guidance on these issues. In this paper, I use plant level data on a sample of entrants in SIC 38, Instruments, to examine the characteristics associated with both proprietary and spillover learning by doing. The plant level data permit tests for the relative importance of within and between firm spillovers. I include both formal knowledge, obtained through R&D expenditures, and informal knowledge, obtained through learning by doing, in a production function framework. I allow the speed of learning to vary across plants according to characteristics such as R&D intensity, wages, and the skill mix. The results suggest that (a) Ainformal@ knowledge, accumulated through production experience at the plant, is a much more important source of productivity growth for these plants than is Aformal@ knowledge gained via research and development expenditures, (b) interfirm spillovers are stronger than intrafirm spillovers, (c) the slope of the own learning curve is positively related to worker quality, (d) the slope of the spillover learning curve is positively related to the skill mix at plants, (e) neither own nor spillover learning curve slopes are related to R&D intensities. These results imply that learning by doing may be, to some extent, an endogenous phenomenon at these plants. Thus, models of industry evolution that incorporate learning by doing may need to be revised. The results are also broadly consistent with the recent growth models.View Full Paper PDF
-
Working PaperPRODUCTIVITY AND ACQUISITIONS IN U.S. COAL MINING
December 1999
Working Paper Number:
CES-99-17
This paper extends the literature on the productivity incentives for mergers and acquisitions. We develop a stochastic matching model that describes the conditions under which a coal mine will change owners. This model suggests two empirically testable hypotheses: i. acquired mines will exhibit low productivity prior to being acquired relative to non-acquired mines and ii. extant acquired mines will show post-acquisition productivity improvements over their pre-acquisition productivity levels. Using a unique micro data set on the universe of U.S. coal mines observed from 1978 to 1996, it is estimated that acquired coal mines are significantly less productive than non-acquired mines prior to having been acquired. Additionally, there is observable and significant evidence of post-acquisition productivity improvements. Finally, it is found that having been acquired positively and significantly influences the likelihood that a coal mine fails.View Full Paper PDF
-
Working PaperReady-to-Mix: Horizontal Mergers, Prices, and Productivity
January 2017
Working Paper Number:
CES-17-38
I estimate the price and productivity effects of horizontal mergers in the ready-mix concrete industry using plant and firm-level data from the US Census Bureau. Horizontal mergers involving plants in close proximity are associated with price increases and decreases in output, but also raise productivity at acquired plants. While there is a significant negative relationship between productivity and prices, the rate at which productivity reduces price is modest and the effects of increased market power are not offset. I then present several additional new results of policy interest. For example, mergers are only observed leading to price increases after the relaxation of antitrust standards in the mid-1980s; price increases following mergers are persistent but tend to become smaller over time; and, there is evidence That firms target plants charging below average prices for acquisition. Finally, I use a simple multinomial logit demand model to assess the effects of merger activity on total welfare. At acquired plants, the consumer and producer surplus effects approximately cancel out, but effects at acquiring plants and non-merging plants, where prices also rise, cause a substantial decrease in consumer surplus.View Full Paper PDF
-
Working PaperMaterials Prices and Productivity
June 2012
Working Paper Number:
CES-12-11
There is substantial within-industry variation, even within industries that use and produce homogeneous inputs and outputs, in the prices that plants pay for their material inputs. I explore, using plant-level data from the U.S. Census Bureau, the consequences and sources of this variation in materials prices. For a sample of industries with relatively homogeneous products, the standard deviation of plant-level productivities would be 7% lower if all plants faced the same materials prices. Moreover, plant-level materials prices are both persistent across time and predictive of exit. The contribution of net entry to aggregate productivity growth is smaller for productivity measures that strip out di'erences in materials prices. After documenting these patterns, I discuss three potential sources of materials price variation: geography, di'erences in suppliers. marginal costs, and suppliers. price discriminatory behavior. Together, these variables account for 13% of the dispersion of materials prices. Finally, I demonstrate that plants.marginal costs are correlated with the marginal costs of their intermediate input suppliers.View Full Paper PDF
