Employment and Workers Compensation Claims in the Aftermath of Hurricanes
InsightsEconomic & Financial
By NCCI Insights April 09, 2018

KEY TAKEAWAYS

  • Over the past 15 years, we observed fewer workers compensation claims in the year following major hurricanes for affected states
  • After Hurricane Katrina, workers compensation claims dropped 10% in Q3 2005 and returned to normal levels after about one year
    • Claims dropped on impact in every industry group except manufacturing
  • For the most severe non-Katrina hurricanes, claims fell only about 3% in the quarter of impact and remained slightly below expectation for the following year
  • Besides slight growth for contracting during rebuilding efforts, hurricanes do not appear to have a lasting impact on industry mix of employment or workers compensation claims in affected states
 
 

Employment and Workers Compensation Claims in the Aftermath of Hurricanes

Last year, the United States suffered a record amount of damages from natural disasters, led by Hurricanes Harvey, Irma, and Maria. The National Centers for Environmental Information (NCEI) estimates that since 2001, the United States has incurred over $1 trillion in damages from natural disasters. This estimate is for direct effects, primarily damage to buildings, vehicles, infrastructure, and agricultural assets.

Property damage is the most obvious and straightforward measure of a disaster’s impact. But how the local economy recovers is also important, particularly to workers compensation. In this edition of the QEB, we examine local economic outcomes in the aftermath of hurricanes, and focus on these questions especially:

  • How do hurricanes disrupt the economies of affected states, and for how long?
  • In the aftermath of a hurricane, what is the effect on employment and work injuries?

What Do We Know Already?

Previous research evidence shows significant lingering effects to regional economies after major hurricanes, and that the net loss to a region goes well beyond initial damages. Some of that economic activity is simply displaced to other counties or states, while some is a net loss to the overall economy.

One study on all hurricanes to strike Florida from Andrew in 1992 to Wilma in 2005 showed that hurricanes affect both the place directly hit and the surrounding areas. The authors found faster earnings growth but slower employment growth in counties hit by hurricanes. They also found that in the short term, earnings decreased in counties not struck themselves, but which neighbored counties that were hit. While property damage mostly affected relatively small areas, lingering economic effects were more widespread across nearby counties.

The economic effects of a natural disaster can be large. A study by the Texas Engineering Extension Service estimated the decreased economic activity attributable to Hurricane Ike over the 12 months after the 2008 storm to be $142 billion—four times larger than NCEI’s estimated $35 billion in direct damages.

The most striking recent US example of a disaster’s long-run impact is Hurricane Katrina, which struck the Gulf Coast on August 29, 2005. The Bureau of Labor Statistics (BLS) reported in 2006 that there were year-over-year employment losses of over 9% for all of Louisiana, concentrated in the affected areas. More recent estimates suggest that employment dropped by over 40% in Orleans Parish. Total statewide employment did not reach pre-storm levels until December 2007, just before the onset of the Great Recession.

There are many studies measuring different aspects of Katrina’s long-term effects on New Orleans’ economy, from education to construction to inequality, up to the present date. New Orleans today is different than it was in 2005 before Katrina.

However, another study found that victims of Hurricane Katrina made full income recoveries within a few years. This conclusion seems to contradict studies showing persistent long-term effects—how can there be a full recovery within a few years, but also lasting effects over a decade later? In fact, these studies illustrate that the perceived magnitude of a natural disaster’s impact depends on how the question of impact is framed. In this instance, the key disparity is migration. The study on income recovery followed pre-Katrina New Orleans residents over time, whether they returned to New Orleans or not. Many did, but many others relocated permanently. What happens to local residents in an area affected by a natural disaster is not the same as what happens to the area itself. Both perspectives are relevant.

Research on the effectiveness of federal aid on displaced residents focuses on individual outcomes, and could conclude that long-term effects of hurricanes are small. However, when focusing on state and regional economies, as is most relevant to workers compensation, the body of research to date suggests that hurricanes can have significant, regionally widespread, and lasting effects.

What Can We Add?

In keeping with our focus on outcomes relevant to workers compensation, we pose the following questions:

  • How large are the employment impacts after a natural disaster? How long does it take for an affected state’s economy to get back to normal?
  • In the months following the hurricane, is there a significant impact on the industry mix of a state’s workforce? What about composition by class code? How long do these changes persist?
  • Are changes in employment and worker mix, if any, also reflected in changes to workers compensation claims and incurred losses?

Our analysis is restricted to the effects of hurricanes, because these are the disasters that do the largest amount of localized damage in a short time frame. Although only 22 of the 142 events tracked since 2001 (15%) were hurricanes,1 they were responsible for 65% of the total damages. They are also best suited to a study focused on timing and local impacts.

Droughts and wildfires are classified by season,2 spreading their damage over a wide geographic and temporal range. Similarly, severe storms and winter storms’ damage tends to be spread over many states as the storm front moves through, while freezes tend to primarily cause agricultural losses. Among different kinds of natural disasters, hurricanes have the greatest impact on local economies.

Economic and Workers Compensation Impacts

Table 1

Our analysis relies on a combination of government data on employment by state and NCCI’s Unit Statistical data on workers compensation claims and incurred losses. Because of the nature of our data, and the nature of workers compensation, we analyze outcomes at the state level for areas that are affected by hurricanes. The state-level approach is motivated by research showing that while property damage is localized, economic impacts are felt more widely.

We examine a set of hurricanes that struck the United States between 2001–2015, shown in Table 1. From NCEI, we know the date, primarily affected states, and severity of each hurricane. We merge this with our BLS and NCCI data sets, thus obtaining a panel data set of economic and workers compensation variables aligned with the timing of hurricanes and severe tropical storms. Note that “primarily affected states” are restricted to NCCI states. States for which we do not have Unit Statistical data, such as New York and New Jersey, are excluded from the analysis, as are hurricanes that had their most severe impacts in those states, such as Hurricane Sandy.3

The research design follows a simple intuition. After adjusting employment and workers compensation claims for state and industry group-specific seasonality and national trends, we compare changes in economic and workers compensation outcomes in states that experienced a natural disaster to the experience of other unaffected states at the same time, and to the state’s own experience in periods without a natural disaster. In this way, we seek to separate the effects of the disaster from nondisaster- related differences between states and over time.

In the following sections, we analyze the impact of Hurricane Katrina and other hurricanes on state employment and claims. To facilitate comparisons, we normalize adjusted employment and claims to 100% for the calendar quarter just before a hurricane. We interpret deviations from 100% as resulting from the hurricane, net of our adjustments for seasonality and national growth trends.

Hurricane Katrina: A Case Study

Hurricane Katrina was the most disruptive hurricane between 2001 and 2015. If any lasting effects of storms are observable in economic data, they should be apparent for Katrina. We compare pre- and post-Katrina employment levels and workers compensation claims and losses. If regions affected by Katrina quickly got back to business as usual, then we do not to expect to see major changes either in employment or in workers compensation claims. But, if regional economic effects were lasting, then we would expect to see significant changes in both. How big were these impacts, in fact? And are they of similar magnitude for workers compensation claims as for employment?

How Much Did Employment Fall?

Hurricane Katrina’s employment shock to New Orleans and other Gulf Coast counties was large enough to significantly affect the entire state of Louisiana. In Figure 1, we show Louisiana employment levels, adjusted for seasonality and national employment growth. Hurricane Katrina made landfall on the Gulf Coast on August 29, so we would expect to see immediate impacts in the third quarter of 2005 and its aftermath beginning in the fourth quarter.4

In fact, state employment dropped in September and for the rest of 2005. Louisiana employment declined about 8% in the fourth quarter of 2005—about 120,000 workers—which is consistent with the raw 9% year-over-year drop in employment reported by the BLS. Employment slowly returned to normal throughout 2006 and 2007, but remained 3% below pre-hurricane levels two years later.

What Happened to Workers Compensation Claims?

We also see significant effects on workers compensation claims. However, claims do not exactly follow employment trends. Instead, we see a sharper and quicker dip of about 10% in the quarter of the hurricane. Claims remained down by 8% in the fourth quarter of 2005. By the third quarter of 2006, a year after the hurricane, claims had returned to near pre-hurricane levels.

Figure 1

Which Industry Groups Were Affected Most?

Figure 2 breaks down Louisiana’s change in employment by industry group. Every industry group’s adjusted employment declined in the two quarters including and immediately following Katrina, but their magnitudes differed. In the fourth quarter of 2005, goods and services saw the largest drop, declining 12%, while the other four industry groups fell 3–5%.

Figure 2

Goods and services was also the only industry group not to approach full recovery by two years out, remaining about 7% below pre-hurricane levels. Contracting employment increased throughout 2006, due largely to rebuilding efforts, and remained high after that. The other industry groups recovered more slowly, but started drifting back toward pre-hurricane employment levels by the end of 2006.

How are these employment changes reflected in workers compensation outcomes? Figure 3 shows employment by industry group along with total workers compensation claims. As before, changes from pre-hurricane levels are adjusted by industry group for seasonality and for national trends.

Figure 3

Overall, Louisiana experienced a sharp drop in claims during the quarter of Katrina’s impact, but claims returned to pre-hurricane values during 2006. By industry group, we see generally similar patterns, with a few exceptions:

  • Goods and services matches the overall shape of the recovery most closely, although drops in both claims and employment for goods and services were more severe than for other industry groups.
  • Claim decreases in contracting, office and clerical, and miscellaneous roughly tracked adjusted employment in 2006. The miscellaneous industry group did not experience a disproportionately large drop in claims in the latter half of 2005, but the other two industry groups did.
  • Manufacturing claims rose beginning in the quarter of hurricane impact,5 even as employment fell.

Summary

Hurricane Katrina significantly affected Louisiana’s economy. Employment and workers compensation claims both dropped significantly during the quarter of impact. Workers compensation claims returned to pre-hurricane levels over the next year, but employment did not entirely recover. We do not show the series for incurred losses from workers compensation, but patterns for losses generally follow the series for claims. Compared to the quarter before the hurricane, normalized losses fell about 12% in Q3 2005, similar to the 10% drop in claims shown in Figure 1.

Results From Other Major Hurricanes

The last section showed important effects on employment and workers compensation in Louisiana after Hurricane Katrina. But Katrina was a singular event in many ways. Did other hurricanes have similar impacts on employment and WC claims patterns?

The solid lines in Figure 4 show the paths of employment and workers compensation claims averaged over all of the hurricanes in Table 1. As with Hurricane Katrina, both data series are normalized to 100% as of the quarter before the hurricane. The dotted lines show the paths averaged over only the most severe hurricanes, which we define as those that the NCEI estimated caused at least $10 billion in damages.

Figure 4

For the four quarters following a hurricane’s landfall, the average effect is greater on workers compensation claims than on employment, although all effects are very small. Across all hurricanes, claims in the average state6 fell about 1% in the quarter of impact and returned to normal immediately after. For the most severe hurricanes, claims fell by more than 2% and stayed below pre-hurricane levels for the next four quarters.

Table 2

For employment, the effects are virtually zero. Katrina displaced many more people than typical hurricanes do. This corresponds to others’ previous findings that showed employment changes of only 2–5% on impact in affected counties—effects that can be diffused to tiny blips across entire states.

These negative effects on claims are fairly consistent across the most severe hurricanes. In Table 2, we show the claim impact in the quarter of the storm for each of the major hurricanes and tropical storms shown in Figure 4. Katrina had two of the three largest effects of hurricane-state pairs, with the third being Hurricane Irene in Vermont, a small state that experienced massive flooding in the wake of the storm. Claims fell between 1.6–3% in Florida and Texas in the quarters when one or more large hurricanes made landfall. Tropical Storm Allison and Hurricane Ivan were two storms with high overall direct damages for which we did not observe a drop in workers compensation claims.

Other than a slight uptick in contracting employment and claims in quarters following the hurricane—presumably capturing rebuilding efforts—we did not observe consistent patterns in claims or employment by industry group. The increased manufacturing claims in Louisiana post-Katrina were not seen after the other disasters. Our evidence suggests that most hurricanes affect local economies across the board.

1 We use the term "hurricane" colloquially to mean all tropical cyclones that appear in our data set. In the time period we studied, NCEI classified two tropical storms, Allison and Lili, as billion-dollar disasters. We used these in our analysis exactly the same way as the other tropical cyclones, which were all classified as hurricanes.

2 For example, Western Wildfires, June–Nov 2012, or Western/Plains Drought/Heatwave, Mar–Nov 2013.

3 Note also that our NCCI Unit Statistical data is necessarily restricted to employers that use private workers compensation insurance. Self-insured employers will not appear in our claims data but will be represented in the BLS employment data. If self-insured businesses experience dramatically different claims impacts of hurricanes than privately insured businesses in the same state, or if rates of self-insurance change significantly after a hurricane, that could be a partial explanation for differences we see between employment and claims effects. However, data from the National Academy of Social Insurance on self-insured rates by state and year suggests that no such structural changes are key drivers of our findings.

4 Hurricane Rita made landfall in September 2005, so the effects discussed in this section include effects from Rita as well.

5 Unadjusted manufacturing claims fell slightly in Louisiana in Q4 2005, but during that quarter they also fell in other NCCI states by a similar amount.

6 One potential concern is that this sample includes the large states of Florida and Texas, which would see lower effects as a percentage of the state economy than smaller states such as Louisiana or Mississippi, given a storm of similar magnitude. In an alternate specification, we estimated total changes in claims and employment and found patterns that were similar to those reported here.

​This article is provided solely as a reference tool to be used for informational purposes only. The information in this article shall not be construed or interpreted as providing legal or any other advice. Use of this article for any purpose other than as set forth herein is strictly prohibited.


TOP