Wage insurance that compensates those directly negatively affected by trade can cost-effectively help workers who most need assistance and allow them to move with less pain to new jobs. This column examines the wage insurance components of US Trade Adjustment Assistance, which offers wage subsidies to workers who find reemployment at a lower wage. The programme boosted workers’ employment probabilities and their earnings, and workers returned to work more quickly with no declines in job quality. Wage insurance appears to pay for itself, with tax receipts on increased earnings and reduced unemployment insurance payments fully offsetting the programme’s costs.
The extraordinary rise in US and Chinese import tariffs that began in 2018 reflects one of the most significant shifts in global trade policy in recent memory. While US consumers have benefited from freer trade with China through lower goods prices (Sager and Jaravel 2019), bipartisan concerns around declining US competitiveness and a view of unfair trade practices have been invoked to justify protectionist policies (Lamy 2023).
An often-overlooked aspect of the debate over trade is the extent to which social insurance programmes can cushion those harmed by trade and allow them to move with less pain to jobs in growing sectors of the economy (including those sectors that benefit from trade). Such programmes hold the promise of delivering widespread gains to the far larger group of consumers who value access to cheaper imported goods while protecting the (typically smaller) group whose livelihoods are more directly threatened by them. The potential for such insurance is sizeable, given that economists have long associated large economic costs with harsh tariffs (Cockrell 2018).
In our recent paper (Hyman et al. 2024), we analyse the economic effects of an innovative policy for compensating those directly negatively affected by trade – wage insurance. We show that such a programme circumvents many of the concerns related to prior safety net policies and offers a cost-effective approach to trade adjustment, one that focuses on workers while at the same time allowing consumers to avail themselves of the benefits of trade.
Compensating those hurt by trade: What went wrong?
Trade Adjustment Assistance (TAA) has been the US’s flagship programme for trade adjustment, which offers workers negatively affected by import competition and offshoring with retraining subsidies and extended unemployment insurance. But in a prominent analysis of the effects of Chinese import competition, Autor et al. (2013) found that exposed workers received 30 times as much money in payments from Social Security Disability Insurance than payments from TAA. This finding shows that TAA may not have provided sufficient coverage for trade-affected workers.
One reason coverage may have been incomplete was that the return to TAA job training programmes was thought to be meagre at best: comparing roughly 7,500 Trade Adjustment Assistance trainees to a matched sample of 12,500 unemployment insurance claimants, Schochet et al. (2012) found limited effects on employment and earnings outcomes. However, their sample period at the onset of the Great Recession had an unanticipated limitation:
It is possible that trainees could not take full advantage of their new skills, because many re-entered the labour market during the peak of a recession, whereas their matched comparisons (many fewer of whom undertook training) mostly re-entered the labour market earlier (Scochet et al. 2012: xvii).
In a more recent analysis, Hyman (2018) studied 300,000 workers applying for TAA across a wider set of years (1990–2011) and found more favourable effects of retraining, using an examiner design comparing workers who were quasi-randomly certified for TAA versus those who were denied. Though estimated earnings returns were large and positive, the programme remained costly on an individual basis, with average annual expenditures on training and extended unemployment insurance exceeding $20,000 per person.
The general takeaway from much of this literature had thus become that the US social safety net for trade-impacted workers compensated, at best, only some workers, and at worst generated moral hazard – providing workers perverse incentives to stay out of the labour market while receiving generous unemployment insurance benefits or Social Security Disability Insurance payments.
Wage insurance: Generous displacement insurance but with a work requirement
One innovative approach to trade adjustment is wage insurance (Kletzer and Litan 2001), which provides additional income to displaced workers who find re-employment at a lower wage. In our paper, we studied the wage insurance provisions of the TAA programme. Displaced workers in the traditional TAA programme participate in job training and receive extended unemployment insurance payments. Workers aged 50 or older are additionally eligible for a wage insurance programme, Reemployment TAA, which does not require job training and instead pays a wage subsidy of up to half the difference between workers’ pre- and post-separation wages for up to two years. Because the amount of the subsidy is proportional to the earnings decline, the policy makes re-employment more attractive, particularly in lower-wage jobs, and directs larger benefit payments to workers who lose the most following displacement.
Estimating the causal effect of any voluntary social insurance programme is challenging. Those receiving benefits may be self-selected on various characteristics that relate to future outcomes. For example, workers receiving wage insurance payments must experience wage declines to be eligible. Therefore, comparisons between those who receive wage insurance payments and those who don’t could reflect underlying differences in the two groups rather than the causal effect of the policy on either group of workers.
To circumvent these challenges, we leverage the requirement that workers must be age 50 or older when re-employed to be eligible for wage insurance. After a worker’s TAA petition for a given layoff event is certified by the US Department of Labor, the associated workers qualify for the baseline benefits of training and extended unemployment insurance payments described above. Those aged 50 or older are eligible for both standard benefits and wage insurance, while younger workers qualify only for standard TAA.
We therefore use a regression-discontinuity design that compares employment and earnings outcomes for workers who are slightly older than age 50 when displaced to those for workers who are slightly younger. For example, Figure 1 shows the relationship between the proportion of displaced workers employed two years after displacement and the workers’ age at separation. The visible jump at age 50 reflects the positive effect of wage insurance on workers’ employment probability.
Figure 1 Employment probability jumps for those eligible for wage insurance
Notes: The chart shows a regression discontinuity plot for the quarterly employment probability eight quarters following displacement. Hollow dots represent an omitted range in which recipients are only partially eligible for wage insurance (see Hyman et al. 2024 for details). RD denotes regression discontinuity estimate of an 8.8-percentage-point higher proportion of workers employed, with standard error from the regression in parentheses.
Sources: US Department of Labor; US Census Bureau; authors’ calculations.
To empirically estimate the effect of wage insurance eligibility using this approach, we merge administrative data on TAA petitions with the US Census Bureau’s Longitudinal Employer-Household Dynamics dataset from 2007 to 2014, which enables us to track worker earnings and employment outcomes for several years before and after job loss. Our sample covers 76,500 workers separating from approximately 1,000 TAA-petitioning firms.
Compared to the average displaced worker in the US, workers in our sample are older, have longer tenure at the displacing firm, and have lower educational attainment – all characteristics that make the consequences of job loss particularly severe. Indeed, the share of lost earnings ‘replaced’ after job loss (left-hand panel of Figure 2) and the share of workers employed (right-hand panel) remain stunted for several years beyond initial layoffs in our sample.
Figure 2 Workers eligible for TAA experience large and persistent earnings and employment losses
Notes: The left-hand panel shows quarterly earnings over mean earnings from 5 to 8 periods prior to separation, in which workers are required to be employed to be in the sample. The right-hand panel shows the proportion of displaced workers (those employed in quarter 0 and non-employed in quarter 1) who are employed in each quarter relative to separation.
Sources: US Department of Labor; US Census Bureau; authors’ calculations.
The impact of wage insurance on US workers
We find that wage-insurance eligibility boosted workers’ employment probabilities by 8 to 17 percentage points in the two years after displacement before fading to zero after four years (left-hand panel of Figure 3). Programme eligibility also persistently increased earnings (even omitting the value of the subsidies). On average, wage insurance eligibility increased workers’ earnings by 10% of their pre-displacement earnings, amounting to an increase of more than $18,000, or 26%, over the four years following a layoff (Figure 3, right-hand panel).
Figure 3 Effects on employment and cumulative earnings over time
Note: Each dot represents a separate regression coefficient in each quarter relative to separation using our main estimation strategy, with 95% confidence intervals denoted in the shaded areas.
Sources: US Department of Labor; US Census Bureau; authors’ calculations.
Overall, workers eligible for wage insurance tend to return to work more quickly following job loss. Shortening unemployment spells explains most of the positive effects of wage insurance eligibility on earnings, consistent with prior evidence that extended periods of unemployment harm worker outcomes.
Our study finds minimal effects on other employment outcomes, including industry-switching rates and workers’ number of unique employers, geographic mobility, job quality (measured by firm age, firm size, and earnings growth rates), or the length of employment at the first job after displacement. The absence of declines in earnings or job quality allays concerns that wage insurance might have led to worse job matches and persistently lower wages.
Wage insurance appears to be a very cost-effective policy in our setting. In fact, we estimate that it pays for itself; the tax receipts on increased earnings and reduced unemployment insurance payments fully offset the costs of the programme. This means that the programme led to net government savings while benefiting eligible workers. By contrast, most other insurance and training programmes that target displaced workers are less cost-effective (Hendren and Sprung-Keyser 2020). Moreover, our results suggest that wage insurance is likely to reduce the take-up of Social Security Disability Insurance, which would make the government savings even larger than we estimate.
A cost-effective trade adjustment policy with broad legislative potential
Wage insurance provides a cost-effective alternative to trade disruption that has the potential for broad legislative support: it addresses the central problem of worker protection without generating disincentives to search for a job. This design improves upon current programmes, while still allowing for the benefits brought through access to imports that can save US consumers money.
Despite these compelling findings, important questions about wage insurance remain. For example, it would be valuable to understand how wage insurance affects other important outcomes like mortality, which has been shown to increase after job loss. Also, as with the Earned Income Tax Credit (Rothstein 2010), a larger-scale wage insurance programme might lead employers to lower wage offers with the knowledge that some of their applicants will receive subsidies (Davidson and Woodbury 1993). This response would blunt the favourable effects for workers that we find in a smaller-scale programme. Nonetheless, the existing programme was cost-effective, and our results suggest that wage insurance could be an effective policy to support economically vulnerable workers impacted by trade shocks.
Source : VOXeu
