Simple Statistical Models for Discrete Panel Data Developed and Applied to Test the Hypothesis of True State Dependence against the Hypothesis of Spurious State Dependence

Persistent lack of health insurance might arise from two distinct sources. First, individuals who lack insurance might possess certain traits, also exhibiting persistence, that associate with lower probabilities of having insurance. Second, persistence might arise if lack of insurance in the present period directly increases the probability of lacking insurance in future periods. This second channel, called "true state dependence", implies a causal mechanism from past uninsurance. State dependence can also arise if insurance status "feeds back" to future factors that affect insurance status in subsequent periods. This paper extends standard dynamic models to incorporate feedback from uninsurance to future health and employment. Results indicate that, after incorporating feedback effects, lacking insurance six months ago increases the probability of current uninsurance by 57-59 percentage points. Similarly, lacking insurance two years ago increases the probability of current uninsurance by 18-22 percentage points. The large estimates of state dependence suggest that future uninsurance can be effectively reduced by enacting blanket policies that increase insurance rates now. On the other hand, estimates also reveal that unmeasured heterogeneity contributes more to persistence than observed heterogeneity, especially in the long run. This finding draws into question the effectiveness of implementing new, or expanding upon existing, targeted reforms.

Download Full-text

Time and State Dependence in an Ss Decision Experiment

American Economic Journal Macroeconomics ◽

10.1257/mac.20130267 ◽

2016 ◽

Vol 8 (1) ◽

pp. 285-310 ◽

Cited By ~ 4

Author(s):

Jacopo Magnani ◽

Aspen Gorry ◽

Ryan Oprea

Keyword(s):

Time Dependence ◽

Laboratory Experiment ◽

State Dependence ◽

Optimal Level ◽

Time Dependent ◽

Fixed Cost ◽

Optimal State ◽

State Variable ◽

State Dependent ◽

True State

Flow earnings in a laboratory experiment decline the further a Brownian state variable, z, evolves from its optimal level, z *. Optimal state dependent models predict subjects will pay a fixed cost to return z to z * only when z strays outside a critical inaction region around the optimum. On average, subjects adjust at states remarkably close to optimal threshold levels but, as in the field, do not establish true “state dependent” inaction regions, suggesting significant “time dependent” components in adjustment rules. Structural estimates of subjective observation cost qualitatively account for variation in time dependence observed across treatments. (JEL C91, D21, D80)

Download Full-text

True versus spurious state dependence in firm performance

Empirical Economics ◽

10.1007/s00181-007-0149-y ◽

2007 ◽

Vol 35 (2) ◽

pp. 207-228 ◽

Cited By ~ 11

Author(s):

Ulrich Kaiser ◽

Hans Christian Kongsted

Keyword(s):

Firm Performance ◽

State Dependence ◽

Spurious State

Download Full-text

The Future Strikes Back. Using Future Treatments to Detect and Reduce Hidden Bias

10.31235/osf.io/yhtpu ◽

2018 ◽

Author(s):

Felix Elwert ◽

Fabian T. Pfeffer

Keyword(s):

State Dependence ◽

Treatment Variable ◽

New Approach ◽

Parental Income ◽

Unmeasured Confounder ◽

Omitted Variable Bias ◽

True State ◽

Data Generating Process ◽

Future Treatments ◽

Variable Bias

Conventional advice discourages controlling for post-outcome variables in regression analysis. Here, we show that controlling for commonly available post-outcome (i.e. future) values of the treatment variable can help detect, reduce, and even remove omitted variable bias (unobserved confounding). The premise is that the same unobserved confounders that affect treatment also affect future values of the treatment. Future treatments thus proxy for the unmeasured confounder, and researchers can exploit these proxy measures productively. We establish several new results: Regarding a commonly assumed data-generating process involving future treatments, we (1) introduce a simple new approach to reduce bias and show that it strictly reduces bias; (2) elaborate on existing approaches and show that they can increase bias; (3) assess the relative merits of approaches; (4) analyze true state dependence and selection as key challenges; and (5) demonstrate that future treatments can test for hidden bias, even when they fail to reduce bias. We illustrate these results empirically with an analysis of the effect of parental income on children’s educational attainment.

Download Full-text

A Note on “Sequential Neighborhood Effects” by Hicks et al. (2018)

Demography ◽

10.1215/00703370-9000711 ◽

2021 ◽

Author(s):

Mark S. Handcock ◽

Andrew L. Hicks ◽

Narayan Sastry ◽

Anne R. Pebley

Keyword(s):

Neighborhood Effects ◽

Causal Model ◽

State Dependence ◽

Causal Modeling ◽

Math Skills ◽

New Approach ◽

Adjustment Procedure ◽

True State ◽

Reading And Math ◽

Over Time

Abstract We revisit a novel causal model published in Demography by Hicks et al. (2018), designed to assess whether exposure to neighborhood disadvantage over time affects children's reading and math skills. Here, we provide corrected and new results. Reconsideration of the model in the original article raised concerns about bias due to exposure-induced confounding (i.e., past exposures directly affecting future exposures) and true state dependence (i.e., past exposures affecting confounders of future exposures). Through simulation, we show that our originally proposed propensity function approach displays modest bias due to exposure-induced confounding but no bias from true state dependence. We suggest a correction based on residualized values and show that this new approach corrects for the observed bias. We contrast this revised method with other causal modeling approaches using simulation. Finally, we reproduce the substantive models from Hicks et al. (2018) using the new residuals-based adjustment procedure. With the correction, our findings are essentially identical to those reported originally. We end with some conclusions regarding approaches to causal modeling.

Download Full-text