Testing for Cointegrating Relationships with Near-Integrated Data

Testing theories about political change requires analysts tomake assumptions about the memory of their time series. Appliedanalyses are often based on inferences that time series areintegrated and cointegrated. Typically analyses rest on Dickey–Fullerpretests for unit roots and a test for cointegration based onthe Engle–Granger two-step method. We argue that thisapproach is not a good one and use Monte Carlo analysis to showthat these tests can lead analysts to conclude falsely thatthe data are cointegrated (or nearly cointegrated) when thedata are near-integrated and not cointegrating. Further, analystsare likely to conclude falsely that the relationship is notcointegrated when it is. We show how inferences are highly sensitiveto sample size and the signal-to-noise ratio in the data. Wesuggest three things. First, analysts should use the singleequation error correction test for cointegrating relationships;second, caution is in order in all cases where near-integrationis a reasonable alternative to unit roots; and third, analystsshould drop the language of cointegration in many cases andadopt single-equation error correction models when the theoryof error correction is relevant.     

Experimental Intelligence

This article explores the use and application of experimental methodologies in intelligence analysis. Previous experimental work has much to tell us about how people process information in making judgments. Calibration, and how appropriate levels of confidence might be improved with systematic feedback, is addressed. Training around uncertainty biases is covered. Extant experimental work suggests mechanisms and tests which might be used to better screen potential analysts for personality characteristics that might make them more suited for some tasks over others. The paradigm of experimental manipulation serves as a useful template for exploring alternative conceptualizations of uncertain environments.     

Taking Time Seriously

Dramatic world change has stimulated interest in research questions about the dynamics of politics. We have seen increases in the number of time series data sets and the length of typical time series. But three shortcomings are prevalent in published time series analysis. First, analysts often estimate models without testing restrictions implied by their specification. Second, researchers link the theoretical concept of equilibrium with cointegration and error correction models. Third, analysts often do a poor job of interpreting results. The consequences include weak connections between theory and tests, biased estimates, and incorrect inferences. We outline techniques for estimating linear dynamic regressions with stationary data and weakly exogenous regressors. We recommend analysts (1) start with general dynamic models and test restrictions before adopting a particular specification and (2) use the wide array of information available from dynamic specifications. We illustrate this strategy with data on Congressional approval and tax rates across OECD countries.     

Have Your Cake and Eat It Too? Cointegration and Dynamic Inference from Autoregressive Distributed Lag Models

Although recent articles have stressed the importance of testing for unit roots and cointegration in time‐series analysis, practitioners have been left without a straightforward procedure to implement this advice. I propose using the autoregressive distributed lag model and bounds cointegration test as an approach to dealing with some of the most commonly encountered issues in time‐series analysis. Through Monte Carlo experiments, I show that this procedure performs better than existing cointegration tests under a variety of situations. I illustrate how to implement this strategy with two step‐by‐step replication examples. To further aid users, I have designed software programs in order to test and dynamically model the results from this approach.     

Narrative analysis for the policy analyst: A case study of the 1980–1982 medfly controversy in California

Narrative policy analysis provides a way of analyzing those highly uncertain and complex policy issues whose truth-value cannot be ascertained and about which the only thing practicing policy analysts know are the stories policymakers use in articulating these issues. One extremely well-documented controversy, the 1980-82 California Medfly Crisis, illustrates how a comparison and analysis of the structure of the dominant stories in that controversy reduces some of its underlying uncertainty. By focusing on the differential risk perceptions reflected in the stories, narrative policy analysis identifies an important area in which conventional policy analysis could have contributed to the controversy's resolution.     

Modeling Equilibrium Relationships: Error Correction Models with Strongly Autoregressive Data

Political scientists often argue that political processes movetogether in the long run. Examples include partisanship andgovernment approval, conflict and cooperation among countries,public policy sentiment and policy activity, economic evaluationsand economic conditions, and taxing and spending. Error correctionmodels and cointegrating relationships are often used to characterizethese equilibrium relationships and to test hypotheses aboutpolitical change. Typically the techniques used to estimateequilibrium relationships are based on the statistical assumptionthat the processes have permanent memory, implying that politicalexperiences cumulate. Yet many analysts have argued that thisis not a reasonable theoretical or statistical assumption formost political time series. In this paper I examine the consequencesof assuming permanent memory when data have long but not permanentmemory. I focus on two commonly used estimators: the Engle-Grangertwo-step estimator and generalized error correction. In my analysisI consider the important role of simultaneity and discuss implicationsfor the conclusions political scientists have drawn about thenature, even the existence, of equilibrium relationships betweenpolitical processes. I find that even small violations of thepermanent memory assumption can present substantial problemsfor inference on long-run relationships in situations that arelikely to be common in applied work in all fields and suggestways that analysts should proceed.     

An Evaluation of the Recent Move to Centralize the Finance of Public Schools in Michigan

This article examines Michigan's 1994–95 shift from its system of K–12 public education being funded largely by local property taxes, to it being funded predominately by statewide revenue sources. We briefly describe events that led up to the adoption of Michigan's school finance reform and then go into some detail about the tax and revenue-sharing specifics. Since the structure of Michigan's school finance reform is complex, somewhat unique, and likely to be considered as a model for other states to duplicate, this examination will assist policy analysts and policymakers throughout the United States. Other states considering similar reforms will benefit from the simulations we provide with regard to the likely long-run results of Michigan's school finance reform.     

Structure and Process: Examining the Interaction between Bureaucratic Organization and Analytical Requirements

Attempts by politicians to control bureaucratic decisions include both structural (how is the agency making the decision organized?) and procedural (what rules must they follow when making the decision?). But how do these two modes of influence interact? This article examines the interaction between bureaucratic structure and one procedural control, the requirement that agencies conduct an analysis of their decisions prior to their issuance. I look at this interaction in the context of two types of analysis, cost‐benefit analysis and environmental impact assessment. I interview 16 individuals in each field and draw from their experiences of conducting and reviewing more than a thousand analyses. The conduct of analysis is affected by where analysts are placed in agencies. In particular, independence of analysts has a trade‐off. The more independent analysts are, the more likely they can challenge preferred decisions in their agency. But independent analysts are brought into decisions later and their independence may limit their long‐term impacts on agency culture. Despite this trade‐off, analysts expressed a clear preference for independence. The interaction between different controls of bureaucratic behavior is a potentially fruitful line for further research.     

Do Participants and Observers Assess Intentions Differently During Bargaining and Conflict?

Political actors in conflict settings are often uncertain about their counterparts' intentions. This article explores the psychology of how intentions are assessed using a novel experimental design that randomly assigns subjects to one of three roles—"proposer,""recipient," or "observer." Recipients and observers are given identical noisy information about proposers' actions, and make postplay assessments of proposers' intentions that are rewarded based on accuracy. A first experiment explores a context of ambiguity, while a second experiment explores a context of uncertainty . The results suggest that actors' perceptions can sometimes be directly affected by the set of strategic alternatives they possess. When signals about proposer behavior appear "negative," recipients' assessments of proposers' intentions are more negative than observers' assessments if recipients have the ability to respond to the proposer's action—but not if recipients lack this ability. The ability to respond to proposer behavior appears to cause recipients to make more negative inferences about the proposer than circumstances warrant. Interestingly, recipients' and observers' assessments are indistinguishable when signals about proposer behavior instead appear "positive."     

Implementing Performance-Based Program Budgeting: A System-Dynamics Perspective

The article proposes a model for evaluating budget reforms that combines insights from budgeting, policy implementation, and system-dynamics literatures. System-dynamics modeling combines both quantitative and qualitative research techniques to provide a new framework for applied research; its use is illustrated using performance budgeting as an example. Applied to the implementation of Florida's performance-based program budget, the model identifies actions in the short run that will increase the reform's likelihood of success: providing clear communications; facilitative budget and accounting routines; reliable performance information. The model also identifies critical legislative behaviors that influence executive implementation: how the legislature in the long-run uses performance information to inform resource allocation and how it applies incentives or sanctions to programs that achieve or fail to achieve their performance standards. The legislature has the opportunity to use program reviews prepared by legislative staff to invigorate the executive branch's resolve to continue implementing the reform.     

Beware of econometricians bearing estimates: Policy analysis in a “unit root” world

Current statistical approaches to modeling many economic relationships are grounded in traditional ideas of deterministic trends. Some of the failures of these approaches are due to inappropriate models using time series with “unit roots.” After a shock, unit root processes do not revert to some time trend, but rather can drift up or down without bounds. A “random walk” is a well-known example of a unit root process. The purpose of this paper is to explain the importance of unit root processes to policy analysts who make or rely upon econometric models using time series dat. In particular, the presence of unit root processes in GNP, energy and electricity consumption exports, imports, and other variables suggests that modifications to the way economic relationships are estimated may be necessary. Once these modifications are made, many important parameters turn out to be much different, with substantive implications for both forecasting and policy.     

NEW ISSUES IN THE LOGIC OF POLICY ANALYSIS

Case studies of social interventions need not exchange relevance for rigor, provided analysts keep three methodological issues in mind. First, certain critical tests can and should be performed to help establish causal relations even when empirical data are limited. Second, "causell is an explanatory category representing a moral judgment. Third, case studies should provide some systematic attempt to evaluate the program under scrutiny. The article entitled "Social-Problem Solving in a Revolutionary Setting: Nicaragua's Pesticide Policy Reforms" by Douglas L. Murray (Policy Studies Review, November 1984) is used to illustrate these issues.     

Interpretation in the Analysis of Policy

Interpretation is a critical element in both the analysis and the accomplishment of governing, and taking an interpretive approach to policy should not be seen as a methodological option for consenting adults, but as an integral element of political analysis. Understanding governing (whether by practitioners or analysts) raises questions about the shared meaning which makes action valid, and makes a reality of ‘the government’. This calls for attention to the ways in which this is done, and how this interpretive construction has been recognised in political science (and other social sciences) in Australia and elsewhere, and for a consideration of what this means for political scientists’ agenda of inquiry.     

Chinese Geopolitics: Space Program Cooperation among China,Brazil, and Russia

With the Obama Administration’s foreign policy pivot to Asia, the Chinese government is concerned about the possibility of new challenges in their foreign policy sphere of influence. Since the 1990s, many political and military analysts have explored the eventual rivalry between these two states. However, events in the Middle East preoccupy U.S. attention, leaving China moderately free in Southeast Asia, though it is uncertain for how long this will persist and if the United States will reallocate its attention to contain China. In this case, the Chinese government will need to develop countermeasures to sustain its geopolitical interests. There are two focuses in this analysis: (1) the Chinese geopolitical naval situation, exploring the problems in the geosphere surrounding China and the eventual competition with other regional powers and the United States in East Asia; and (2) the potential for space program cooperation with Russia and Brazil, which, in turn, dramatically changes the balance of power in outer space through positive aspects of Russian technical capabilities and strategic Brazilian geographic (Alcântara launch site).     

Favorite INTs: how they develop,why they matter

Collection of important information is a critical part of the intelligence business. Less recognized and studied is the differential use of types of intelligence information based on personal and organizational preferences for, and biases against, specific intelligence collection disciplines, or ‘INTs’. This article presents a framework for assessing the implications of ‘favorite INTs’ for policy-making, policy implementation, and intelligence analysis. The record shows that favorite INTs negatively influence analysts and the use of intelligence by senior political leaders and military commanders. Practitioners can improve intelligence support and scholars can better understand how intelligence influences decision-making by appreciating how and why favorite INTs develop and influence decision-makers and analysts.     

Beyond regulatory capture: Coproducing expertise for critical infrastructure protection

Complex industries such as petroleum production, civil aviation, and nuclear power produce “public risks” that are widely distributed and temporally remote, and thus tend to be ignored by the risk producers. Regulation is perhaps the most common policy tool for governing such risks, but requires expert knowledge that often resides solely within the industries. Hence, many scholars and policymakers raise concerns about “regulatory capture,” wherein regulation serves private interests rather than the public good. This paper argues that regulatory capture framing has tended to limit understanding of expertise and its role in governing public risks. Most studies of regulatory capture treat expertise as a source of knowledge and skills that are created exogenously to political processes, and which can therefore be politically neutral. By contrast, we draw on work in science and technology studies that highlight the value‐laden and relational nature of knowledge and expertise, showing how its formation is endogenous to political processes. Thus, we argue for both broadening analyses of regulatory capture to consider the historically contingent and uncertain process of creating expert knowledge, and going beyond the capture framing by considering the challenge of negotiating different epistemologies and ways of life. We illustrate this analytic strategy by examining the history of and current debate about critical infrastructure protection standards to protect the United States electric power grid from cyberattack. We conclude by considering the broader implications of these findings for governing public risks.     

Cloth without a weaver: power,emergence and institutions across global value chains

Abstract

In studies of the fragmentation and internationalization of production, most value chain approaches consider the inter-firm balance of power as the critical dynamic in development. With the firm as the primary unit of analysis, research long held out two promises: first, bridging the ‘micro-macro gap’ in development theory, meaning making valid inferences from micro-level actors (firms) to macro-sociological outcomes; and second, reconciling its firm-level organizational approach with institutionalism. This paper argues, first, that the literature is artificially constrained in bridging the micro-macro gap due to its delimited conceptualization of ‘power’, based on the ‘agentic-strategic’ behaviour of firms. It argues for broadening the notion of power to bridge the levels of analysis, based on the concept of ‘emergence’. Second, while institutional critics are correct in criticizing value chain scholarship for its neglect, this paper finds that the effects of institutions are not as consistent or determinative as suggested, and hence it seeks to expand the scope for incorporating institutionalism. These points are illustrated through an intra-industry comparative study of three textile agro-industries in China.     

Fusing algorithms and analysts: open-source intelligence in the age of ‘Big Data’

In the age of ‘Big Data’, the potential value of open-source information for intelligence-related purposes is widely recognised. Of late, progress in this space has increasingly become associated with software that can expand our ability to gather, filter, interrelate and manipulate data through automated processes. The trend towards automation is both innovative and necessary. However, techno-centric efforts to replace human analysts with finely crafted algorithms across the board, from collection to synthesis and analysis of information, risk limiting the potential of OSINT rather than increasing its scope and impact. Effective OSINT systems must be carefully designed to facilitate complementarity, exploit the strengths, and mitigate the weaknesses of both human analysts and software solutions, obtaining the best contribution from both. Drawing on insights from the field of cognitive engineering, this article considers at a conceptual level how this might be achieved.     

Time Series Intervals and Statistical Inference: The Effects of Temporal Aggregation on Event Data Analysis

While many areas of research in political science draw inferencesfrom temporally aggregated data, rarely have researchers exploredhow temporal aggregation biases parameter estimates. With somenotable exceptions (Freeman 1989, Political Analysis 1:61–98;Alt et al. 2001, Political Analysis 9:21–44; Thomas 2002,"Event Data Analysis and Threats from Temporal Aggregation")political science studies largely ignore how temporal aggregationaffects our inferences. This article expands upon others' workon this issue by assessing the effect of temporal aggregationdecisions on vector autoregressive (VAR) parameter estimates,significance levels, Granger causality tests, and impulse responsefunctions. While the study is relevant to all fields in politicalscience, the results directly apply to event data studies ofconflict and cooperation. The findings imply that politicalscientists should be wary of the impact that temporal aggregationhas on statistical inference.