Recalls and unemployment insurance taxes

Abstract

The US unemployment insurance (UI) system draws its funds from a payroll tax on employers. The tax rate varies directly with an employer's layoff history. There exists extensive evidence on the effect of this so-called experience rated tax on layoff decisions. However, since firms are typically liable for each dollar of regular UI benefits paid to laid off former employees, experience rating may also affect recall behaviour. This note therefore measures the effect of the UI financing system on the duration of unemployment. This article finds some evidence that higher layoff tax costs shorten the duration of recall unemployment.

Acknowledgements

The author is grateful to John Ham, Frederick Tannery, and Randall Filer for helpful comments and to John Engberg for generously providing the raw data.

Notes

 The TAA programme was designed to compensate workers harmed by market fluctuations resulting from a rise in imports. The programme was amended several times and remains active. The data comes from California, Indiana, Massachusetts, New York, Ohio, Pennsylvania and Virginia.

 Their entitlement was extended by up to 52 additional weeks of UI coverage. Also, their dollar benefits were calculated as 70% of their previous wages as opposed to the typical 50%.

 I assume that workers are able to correctly calculate their UI compensation even in future UI claims. For further details see Jurajda (2002).

 With the exception of Pennsylvania and Virginia, all states covered by the data use the reserve ratio system. Pennsylvania has a hybrid system with the total tax rate being a sum of tax rates determined by both schedules. The combination of the benefit ratio and reserve ratio systems makes it difficult to compute the measure of experience rating for some values of taxes. For these tax values, it is still possible to derive tight bounds of the extent of experience rating. Virginia had a benefit ratio system in the sample period.

 A typical MTC computation (e.g. Topel, 1983; Card and Levine, 1994) uses insured unemployment rates to compute a ‘steady state’ tax rate which would equate UI tax payments with UI benefits for a given industry.

 Such a comparison is interesting as most previous studies analysing the effect of UI taxes on layoffs relied on variation across industries.

 Individual i subscript is not used in any of the formulae in order to streamline notation.

 Apparel and footwear, other nondurable manufacturing, automobile, steel, other durable manufacturing, finance, retail trade, and construction.

 The first two steps, i.e. 1–13 and 14–26, also control for collection of regular UI benefits.

 The estimates of the baseline hazard coefficients are all significant at the 1% level in both hazards and are available from the author upon request. Both hazards exhibit negative duration dependence.

 In particular, the data allow for within-industry comparison of workers with different level of experience rating, which is particularly important in absence of firm-specific tax rates when the imputed extent of experience rating is industry-specific.