Explore the vast range of titles available.

To identify the estimand in missing data problems and observational studies, it is common to base the statistical estimation on the ‘missingness at random’ and ‘no unmeasured confounder’ assumptions. However, these assumptions are unverifiable by using empirical data and pose serious threats to the validity of the qualitative conclusions of statistical inference. A sensitivity analysis asks how the conclusions may change if the unverifiable assumptions are violated to a certain degree. We consider a marginal sensitivity model which is a natural extension of Rosenbaum’s sensitivity model that is widely used for matched observational studies. We aim to construct confidence intervals based on inverse probability weighting estimators, such that asymptotically the intervals have at least nominal coverage of the estimand whenever the data-generating distribution is in the collection of marginal sensitivity models. We use a percentile bootstrap and a generalized minimax–maximin inequality to transform this intractable problem into a linear fractional programming problem, which can be solved very efficiently. We illustrate our method by using a real data set to estimate the causal effect of fish consumption on blood mercury level.

Policymakers are increasingly using norm-based messages to influence individual decision making. We partner with a metropolitan water utility to implement a natural field experiment to examine the effect of such messages on residential water demand. The data, drawn from more than 100,000 households, indicate that social comparison messages had a greater influence on behavior than simple prosocial messages or technical information alone. Moreover, our data suggest that social comparison messages are most effective among households identified as the least price sensitive: high users. Yet the effectiveness of such messages wanes over time. Our results thus highlight important complementarities between pecuniary and nonpecuniary strategies.

In the context of a dynamic, stochastic, general equilibrium model, we perform classical maximum likelihood and Bayesian estimations of the contribution of anticipated shocks to business cycles in the postwar United States. Our identification approach relies on the fact that forward-looking agents react to anticipated changes in exogenous fundamentals before such changes materialize. It further allows us to distinguish changes in fundamentals by their anticipation horizon. We find that anticipated shocks account for about half of predicted aggregate fluctuations in output, consumption, investment, and employment.

PurposeThe purpose of this paper is to investigate what factors can affect people’s continuous watching and consumption intentions in live streaming.Design/methodology/approachThis research conducted a mixed-methods study. The semi-structured interview was deployed to develop a research model and a live streaming typology. A survey was then used for quantitative assessment of the research model. Survey data were analyzed using partial least squares-structural equation modeling.FindingsThe results suggest that sex and humor appeals, social status display and interactivity play considerable roles in the viewer’s behavioral intentions in live streaming and their effects vary across different live streaming types.Research limitations/implicationsThis research is conducted in the Chinese context. Future research can test the research model in other cultural contexts. This study can also be extended by incorporating the roles of viewer gender and price sensitivity in the future.Practical implicationsThis study provides managerial insights into how live streaming platforms and streamers can improve their popularity and profitability.Originality/valueThe paper introduces a novel form of social media and a new business model. It illustrates what will affect people’s behavioral intentions in such a new context.

Taxes on sugar-sweetened beverages (SSBs) meant to improve health and raise revenue are being adopted, yet evaluation is scarce. This study examines the association of the first penny per ounce SSB excise tax in the United States, in Berkeley, California, with beverage prices, sales, store revenue/consumer spending, and usual beverage intake. Methods included comparison of pre-taxation (before 1 January 2015) and first-year post-taxation (1 March 2015-29 February 2016) measures of (1) beverage prices at 26 Berkeley stores; (2) point-of-sale scanner data on 15.5 million checkouts for beverage prices, sales, and store revenue for two supermarket chains covering three Berkeley and six control non-Berkeley large supermarkets in adjacent cities; and (3) a representative telephone survey (17.4% cooperation rate) of 957 adult Berkeley residents. Key hypotheses were that (1) the tax would be passed through to the prices of taxed beverages among the chain stores in which Berkeley implemented the tax in 2015; (2) sales of taxed beverages would decline, and sales of untaxed beverages would rise, in Berkeley stores more than in comparison non-Berkeley stores; (3) consumer spending per transaction (checkout episode) would not increase in Berkeley stores; and (4) self-reported consumption of taxed beverages would decline. Main outcomes and measures included changes in inflation-adjusted prices (cents/ounce), beverage sales (ounces), consumers' spending measured as store revenue (inflation-adjusted dollars per transaction) in two large chains, and usual beverage intake (grams/day and kilocalories/day). Tax pass-through (changes in the price after imposition of the tax) for SSBs varied in degree and timing by store type and beverage type. Pass-through was complete in large chain supermarkets (+1.07¢/oz, p = 0.001) and small chain supermarkets and chain gas stations (1.31¢/oz, p = 0.004), partial in pharmacies (+0.45¢/oz, p = 0.03), and negative in independent corner stores and independent gas stations (-0.64¢/oz, p = 0.004). Sales-unweighted mean price change from scanner data was +0.67¢/oz (p = 0.00) (sales-weighted, +0.65¢/oz, p = 0.003), with +1.09¢/oz (p < 0.001) for sodas and energy drinks, but a lower change in other categories. Post-tax year 1 scanner data SSB sales (ounces/transaction) in Berkeley stores declined 9.6% (p < 0.001) compared to estimates if the tax were not in place, but rose 6.9% (p < 0.001) for non-Berkeley stores. Sales of untaxed beverages in Berkeley stores rose by 3.5% versus 0.5% (both p < 0.001) for non-Berkeley stores. Overall beverage sales also rose across stores. In Berkeley, sales of water rose by 15.6% (p < 0.001) (exceeding the decline in SSB sales in ounces); untaxed fruit, vegetable, and tea drinks, by 4.37% (p < 0.001); and plain milk, by 0.63% (p = 0.01). Scanner data mean store revenue/consumer spending (dollars per transaction) fell 18¢ less in Berkeley (-$0.36, p < 0.001) than in comparison stores (-$0.54, p < 0.001). Baseline and post-tax Berkeley SSB sales and usual dietary intake were markedly low compared to national levels (at baseline, National Health and Nutrition Examination Survey SSB intake nationally was 131 kcal/d and in Berkeley was 45 kcal/d). Reductions in self-reported mean daily SSB intake in grams (-19.8%, p = 0.49) and in mean per capita SSB caloric intake (-13.3%, p = 0.56) from baseline to post-tax were not statistically significant. Limitations of the study include inability to establish causal links due to observational design, and the absence of health outcomes. Analysis of consumption was limited by the small effect size in relation to high standard error and Berkeley's low baseline consumption. One year following implementation of the nation's first large SSB tax, prices of SSBs increased in many, but not all, settings, SSB sales declined, and sales of untaxed beverages (especially water) and overall study beverages rose in Berkeley; overall consumer spending per transaction in the stores studied did not rise. Price increases for SSBs in two distinct data sources, their timing, and the patterns of change in taxed and untaxed beverage sales suggest that the observed changes may be attributable to the tax. Post-tax self-reported SSB intake did not change significantly compared to baseline. Significant declines in SSB sales, even in this relatively affluent community, accompanied by revenue used for prevention suggest promise for this policy. Evaluation of taxation in jurisdictions with more typical SSB consumption, with controls, is needed to assess broader dietary and potential health impacts.

This paper investigates the relation between growth forecast errors and planned fiscal consolidation during the crisis. We find that, in advanced economies, stronger planned fiscal consolidation has been associated with lower growth than expected. The relation is particularly strong, both statistically and economically, early in the crisis. A natural interpretation is that fiscal multipliers were substantially higher than implicitly assumed by forecasters. The weaker relation in more recent years may in part reflect learning by forecasters and in part smaller multipliers than in the early years of the crisis.

We explore empirically models of aggregate fluctuations in which consumers form anticipations about the future based on noisy sources of information and these anticipations affect output in the short run. Our objective is to separate fluctuations due to changes in fundamentals (news) from those due to temporary errors in agents' estimates (noise). We show that structural VARs cannot be used to identify news and noise shocks, but identification is possible via a method of moments or maximum likelihood. Next, we estimate our model on US data. Our results suggest that noise shocks explain a sizable fraction of short-run consumption fluctuations. (JEL D84, E13, E21, E32)

This article provides a simple shrinkage representation that describes the operational characteristics of various forecasting methods designed for a large number of orthogonal predictors (such as principal components). These methods include pretest methods, Bayesian model averaging, empirical Bayes, and bagging. We compare empirically forecasts from these methods with dynamic factor model (DFM) forecasts using a U.S. macroeconomic dataset with 143 quarterly variables spanning 1960-2008. For most series, including measures of real economic activity, the shrinkage forecasts are inferior to the DFM forecasts. This article has online supplementary material.

In this paper we modify the expectation maximization algorithm in order to estimate the parameters of the dynamic factor model on a dataset with an arbitrary pattern of missing data. We also extend the model to the case with a serially correlated idiosyncratic component. The framework allows us to handle efficiently and in an automatic manner sets of indicators characterized by different publication delays, frequencies and sample lengths. This can be relevant, for example, for young economies for which many indicators have been compiled only recently. We evaluate the methodology in a Monte Carlo experiment and we apply it to nowcasting of the euro area gross domestic product.