Tuesday, October 30, 2012

Hurricane Sandy - 5th, 6th Most Severe?


How will Hurricane Sandy rank in severity, for example, compared with Hurricane Irene? A number of estimates are already appearing and estimates of insured losses have been growing beyond the first estimate of $10 billion, which would imply a total cost of approximately $25 billion using a rule of thumb described below.

In 2011, Hurricane Irene was described as being the fifth-costliest hurricane in U.S. history. I didn't believe it, and I checked it out. This number should not be used. Hurricane Irene does not even rank among the ten most costly. A dollar ranking that does not adjust for inflation is just not useful.

The suggestion has already been made that Hurricane Sandy will be the costliest East Coast hurricane ever. Some superlatives have been used:
-          The MTA says Sandy is its worst disaster ever because of the corrosive nature of the salt water from floods on its tracks.
-          Con Ed says this its worst disaster ever.
-          On Long Island, 90 percent of LIPA customers are without power.
-          The Wall Street Journal posted satellite photos showing Sandy is larger than Irene.
-           
The WSJ concludes that Sandy will cost more than its estimate of a $15 billion cost for Irene. NY Governor Cuomo has observed that the cost of Irene was mostly upstate, whereas the cost of Sandy is downstate.

Prospective Measures of Severity

The significance of an oncoming storm is now estimated by meteorologists based on wind-speed categories and barometric pressure. The two measures are interrelated and point to likely wind speeds. The public needs also an indicator of the likely economic impact of flooding.

1. Five Wind-Speed Categories. A Category 1 hurricane means wind speeds of 74-95 mph on the Saffir/Simpson Hurricane Scale. The categories go up to 5 for wind speeds above 155 mph. Hurricane Irene petered out on its way north. The warm air carried by Hurricane Sandy on its way north met another storm with cold air from the northwest.

2. Millibars - Barometric Pressure. The Christian Science Monitor has posted a lucid summary of the importance of this measure of hurricane severity. (It also repeats the error cited above about the cost of Hurricane Irene - I will return to this.) Ordinarily, the barometric pressure is related to wind speed. The normal sea-level barometric pressure is 1013.5. During a hurricane the eye of the storm shows the lowest barometric pressure. The lower the pressure, the higher the winds. During the afternoon before Sandy hit landfall, the barometric pressure at its eye fell from 943 to 940, which is a level associated with Category 3 or Category 4 winds. The lowest barometric pressure that has been measured in a U.S. hurricane is 882 for Hurricane Wilma. Hurricane Carla was the tenth-lowest, 931. The National Hurricane Center list of the most intense Atlantic hurricanes does not follow the Millibars ranking exactly, since Katrina and Wilma are not in the order one would expect.

      Ten Windiest (“Most Intense”) Hurricanes

Name (after 1953) or Location
Year
Category
Millibars
1.Florida Keys
1935
5
892
2.Camille
1969
5
909
3.Katrina
2005
3
920
4.Andrew
1992
5
922
5.Indianola, TX
1886
4
925
6.Keys, FL
1919
4
927
7.Lake Okeechobee, FL
1928
4
929
8.Donna
1960
4
930
8.Miami, FL
1926
4
931
10.Carla
1961
4
931
Notes: Wind Category is at Landfall. Category 5 on the Saffir/Simpson scale means 155 mph winds for at least one minute. Category 4 means 131-154 mph for at least one minute. Category 3 means 111-130 mph for at least one minute.  Millibars are mercury readings for barometric pressure. Source: Based on NOAA, National Weather Service, National Hurricane CenterBlake and Gibney, 2011.

3. Flood Surge Impact. However, most of the damage is caused by the delayed impact of the flooding surge (the hurricane equivalent of a post-earthquake tsunami). We need a new indicator of likely flood damage, which would have to take into account the economic value of property in the track of the hurricane, the sea level of the land, and the size of the expected surge.  The Flood Surge Impact index could take into account the timing of the tides – Hurricane Sandy hit landfall near high tide and the full moon added to the height of the tide and therefore to the surge. The geography of the surge was important in the case of New York City because the surge came from two directions – down the Connecticut coastline through the Long Island Sound and northward through the funnel of New York Harbor.

Retrospective Measures of Cost

There are at least six basic ways to measure or adjust the cost of a hurricane. They overlap:

1. Loss of Life, or Injury.  While every life is precious, on the simple measure of number of lives lost to a hurricane, Hurricane Irene's 24 lives lost did not even rank among the 100 most costly hurricanes. Preparedness is much better than it used to be, and evacuation is widely recommended. Mayor Bloomberg has made clear that the overriding priority of the City of NY is to avoid loss of life among residents and emergency workers. The mayor has taken a scientific approach to evacuation based on flood probability maps that use feet above sea level a proximity to water to create three evacuation zones. These maps proved highly predictive. Also contributing to reduced fatalities is the steady improvement in (a) U.S. Government warning systems via NOAA and its National Weather Service and National Hurricane Center, and (b) the FEMA network of state notification and assistance.  Loss of life can be converted to a dollar figure via life insurance losses or a value that economists impute to a person's remaining working life. Injuries also represent a cost either to the individual or to health insurance plans (private or governmental), and injuries that result in a disability have a working-life cost that can be attached.

The final number for lives lost from Hurricane Irene appears to be 24. Just one of the deaths was in New York City. The deadliest Atlantic hurricane since 1900 was in Galveston, in 1900, with a range of between 8,000 and 12,000 deaths.

    Five Deadliest Atlantic Hurricanes Since 1900
Name or Location, Category
Year
Deaths
1. Galveston, TX, 1900, Category 4
1900
8,000-12,000
2. Lake Okeechobee, FL, 1928, Cat. 4
1928
2,500-3,000
3. Katrina, Category 3
2005
1,200
4. Florida Keys, Category 4
1919
600 (287 land)
5. Long Island Express (Great NE), Cat. 3
1938
600 (256 land)
Source: NOAA, National Weather Service, National Hurricane CenterBlake and Gibney, 2011. The National Weather Service started giving names to hurricanes in 1953. 

The fifth-deadliest was the 1938 Long Island Express. On the criterion of deadliness, Hurricane Irene did not qualify as one of the five most severe.

2. Loss of Physical Property.  Property can be destroyed by wind or flooding or a combination. This means a loss of wealth of the property owner. If the loss is charged against revenue, it means a loss of revenue. (A building may be a depreciated asset; loss of inventory is likely to be expensed.) The first impact may be flying debris, the lifting off of roofs, the flattening of flimsily constructed buildings. The delayed effects include (a) loss of electricity from downed power lines, which means that many perishables have to be thrown out, and (b) flooding, which destroys or rends temporarily useless all kinds of property such as books and electronics, especially if the flooding is from salt water, which MTA Chairman Joe Lhota and Con Ed Chairman Kevin Burke say is especially damaging to power connections, create huge problems for electricity supply and transportation infrastructure.

3. Business Interruption. The delayed effects of a hurricane also include business interruption. Increasingly, businesses insure not only against loss of property but the loss in profits that comes from an interrupted business. When a restaurant or a theater remains closed because of floods that prevent people from showing up, it is hard to make up the loss because the business space has a limited capacity. That is something that is not fully taken into account by those who look for a large rebound after a disaster, as might be true of a retail store that offers a post-hurricane sale. Some kinds of losses are much harder to make up Figures on the cost of hurricanes increasingly include business-interruption losses, which bias upward the later numbers – another reason it is so important to adjust for inflation as discussed below.

4. Insured vs. Uninsured Private Losses. Insurance companies are most interested in the total of insured losses. But from an economic perspective, losses to individuals (e.g., workers paid by the hour) are real. The money that would have been spent in the community by the individuals is missing. The National Hurricane Center uses a simple formula to estimate uninsured losses - it doubles the number for insured losses.

5. Government Losses. At the national level, flood insurance is provided by the National Flood Insurance Program. Individuals pay a premium for this insurance, which would otherwise not be available. After a hurricane, there will be payouts and a loss that may exceed cumulative premiums. The National Hurricane Center in its estimate of damage adds in the number for flood damage provided by the National Flood Insurance Program. FEMA programs provide relief to local governments and individuals. Other Federal bodies (the dewatering unit of the Army Corps of Engineers, for example), states (emergency response teams) and localities (police, fire, sanitation, ambulance) must also be factored in as costs of a disaster.

6. Adjustments for Inflation or Growth in Business Activity. Two kinds of adjustments are typically made to comparisons among hurricanes. One is to adjust cost figures for inflation. The Christian Science Monitor story cited above incorrectly describes Hurricane Irene as the fifth mostly costly hurricane in U.S. history. As I explained last year, that label only works if we are under the delusion that a dollar 100 years ago should be valued the same as a dollar today. (Apart from the fact that business-interruption costs are increasingly included in hurricane losses, adding to the size of the numbers.)  There are widely available cost-of-living indicators to refer to, such as this one from the BLS. Business activity measures are used to relate hurricane damage to the value of the real estate through which the hurricane travels. This is a good predictor of cost and is also a factor to consider in comparing the impact of a hurricane traveling the same path in different years.

Total Economic Damage

The deadliest hurricanes are not always the costliest in terms of property loss or business interruption.

The ten costliest Atlantic hurricanes are listed below with their estimated damage. Total estimated damage includes insured and uninsured losses. A rule of thumb is that uninsured losses equal insured losses. So if an insurance association or forecaster estimates that insured losses are $10 billion, then total private losses (insured plus uninsured) are commonly estimated at $20 billion. In addition, losses are borne by Federal, state and local governments – the cost of special national insurance programs like flood insurance, or the cost of FEMA assistance, or the cost to states and localities of the overtime of emergency assistance personnel or the damage to public infrastructure.

Note that earlier estimates are generally based on physical damage only, whereas later economic impact numbers, after WWII, include impacts such as business-interruption costs because these became widely insured events.

In addition, dollar-value rankings must be adjusted for inflation. There is no sense in unadjusted dollar numbers that go back to 1900. The most costly U.S. hurricane ever was the 1926 Miami Hurricane, which cost $165 billion in 2010 dollars according to the National Hurricane Center.

       Ten Costliest Hurricanes (pre-Sandy)
Rank
Hurricane
Year
$bil. (2010 $)
1
Great Miami
1926
164.8
2
Katrina
2005
113.4
3
Galveston
1900
104.3
4
Galveston
1915
71.4
5
Andrew
1992
60.5
6
L.I. Express
1938
41.1
7
SW Florida
1944
40.6
8
Lake Okeechobee
1928
35.3
9
Ike
2008
29.5
10
Donna
1960
28.2
Source: NOAA, National Weather Service, National Hurricane Center, Pielke et al. (R. A. Pielke, Jr., J. Gratz, C.W. Landsea, D. Collins, M. Saunders, and R. Musulin, 2008: "Normalized Hurricane Damages in the U.S.: 1900-2005." Natural Hazards Review, 9, 29-42, cited in Blake and Gibney, 2011). Pielke et al. adjust historical data for inflation to 2010, wealth per capita and population. The adjustment for inflation is essential. 

Hurricane Irene's estimated cost was $15 billion. Clearly, it does not rank among the ten most costly hurricanes. However, Hurricane Sandy will surely do so. The key is insured losses and government losses. If Hurricane Sandy ends up with insured losses above $15 billion, the economic impact could end up at about $38 billion, and it will rank #8, between the SW Florida Hurricane of 1944 and the Lake Okeechobee Hurricane of 1928. But if insured losses approach $20 billion, the cost of Hurricane Sandy will end up above the 1938 Long Island Express, and would rank #6. If insured losses exceed $25 billion, it could rank ahead of Hurricane Andrew, at #5.

Monday, October 29, 2012

How to Measure Hurricane Sandy's Damage

Indian Wells Beach, East Hampton, NY, 10-29-12.
Photo by JT Marlin.
What is the cost of Hurricane Sandy compared with Hurricane Irene? A number of estimates are already appearing and the dollar amounts are already very large. My concern is that erroneous statements were made about Hurricane Irene that should not be repeated, for example that it was the fifth-costliest hurricane on record. A ranking based on dollars that does not adjust for inflation is just not useful.

The suggestion has already been made that this will be the costliest East Coast hurricane ever. The Wall Street Journal has compared satellite photos of Sandy and Irene and shows that Sandy is much larger.  The WSJ concludes that Sandy will cost more than its estimate of a $15 billion cost for Irene. On the other hand, television interviews with officials in Westchester indicate that Hurricane Irene was more severe.

Advance Measures of Danger

At the moment two main measures of the significance of a storm are provided by meteorologists. They are interrelated and both point to likely wind speeds. The public needs at least one more  measure that shows the likely economic impact of flooding, and I suggest under the third heading below what the measure might look like.

1. Wind-Speed Category. A Category 1 hurricane means wind speeds of 74-95 mph on the Saffir/Simpson Hurricane Scale. Hurricane Irene petered out on its way north. Hurricane Sandy on its way north has been affected by two other independently originating storms, as well as by tidal factors.

2. Barometric Pressure. The Christian Science Monitor has posted a lucid summary of the importance of this measure of hurricane severity. (It also repeats the error cited above about the cost of Hurricane Irene - I will return to this.) Ordinarily, the barometric pressure is related to wind speed. The normal sea-level barometric pressure is 1013.5. During a hurricane the eye of the storm shows the lowest barometric pressure. The lower the pressure, the higher the winds. During the afternoon today, the barometric pressure at the eye of Hurricane Sandy fell from 943 to 940, which is a level associated with Category 3 or Category 4 winds. The lowest barometric pressure that has been measured in a U.S.hurricane is 882 for Hurricane Wilma. Hurricane Carla was the tenth-lowest, 931. The National Hurricane Center list of the most intense hurricanes does not follow the Millibars ranking exactly, since Katrina and Wilma are not in the order one would expect.


                                        Ten Most Intense Hurricanes
Name (After 1953) or Location
Year
Category
Millibars
1.Florida Keys
1935
5
892
2.Camille
1969
5
909
3.Katrina
2005
3
920
4.Andrew
1992
5
922
5.Indianola, TX
1886
4
925
6.Keys, FL
1919
4
927
7.Lake Okeechobee, FL
1928
4
929
8.Donna
1960
4
930
8.Miami, FL
1926
4
931
10.Carla
1961
4
931
Notes: Wind Category is at Landfall. Category 5 on the Saffir/Simpson scale means >155 mph winds for at least one minute. Category 4 means 131-154 mph for at least one minute. Category 3 means 111-130 mph for at least one minute.  Millibars are mercury readings of barometric pressure.
Source: Based on NOAA, National Weather Service, National Hurricane Center, Blake and Gibney, 2011.

3. Flood Surge Impact. However, most of the damage is caused by the delayed impact of the flooding surge (the hurricane equivalent of a post-earthquake tsunami). We need a new indicator of likely flood damage, which would have to take into account the economic value of property in the track of the hurricane, the sea level of the land, and the size of the expected surge.  The Flood Surge Impact index could take into account the tides (the full moon is adding to the surge) and the geography of the surge.

Retrospective Measures of Cost

There are seven basic ways to measure and adjust the cost of a hurricane. They overlap:

1. Loss of Life, or Injury.  On the simple measure of number of lives lost to a hurricane, Hurricane Irene did not even rank among the 100 most costly hurricanes. Preparedness is much better than it used to be and saving lives is a priority. Mayor Bloomberg has made clear that the overriding priority of the City of NY is to avoid loss of life. Hurricane Irene took one life and as far as Mayor Bloomberg is concerned that was one too many. The mayor has generated a scientific approach to evacuation based on flood probability maps. Also contributing to reduced fatalities is the steady improvement in U.S. Government warning systems via the National Weather Service and the National Hurricane Center,  and the FEMA network of state notification and assistance.  Loss of life can be converted to a dollar figure via life insurance losses or a value that economists impute to a person's remaining working life. Injuries also represent a cost either to the individual or to health insurance plans (private or governmental), and injuries that result in a disability have a working-life cost that can be attached.

2. Loss of Physical Property.  Property can be destroyed by wind or flooding or a combination. This means a loss of wealth of the property owner. If the loss is charged against revenue, it means a loss of revenue. (A building may be a depreciated asset; loss of inventory is likely to be expensed.) The first impact may be flying debris, the lifting off of roofs, the flattening of flimsily constructed buildings. The delayed effects include (a) loss of electricity from downed power lines, which means that many perishables have to be thrown out, and (b) flooding, which destroys or rends temporarily useless all kinds of property such as books and electronics, especially if the flooding is from salt water, which MTA Chairman Joe Lhota says is especially damaging to rail signal and power connections. It also includes business interruption, which is the next category.

3. Business Interruption. Increasingly, businesses insure not only against loss of property but the loss in profits that comes from an interrupted business. When a restaurant or a theater remains closed because of floods that prevent people from showing up, it is hard to make up the loss because the business space has a limited capacity. Figures on the cost of hurricanes increasingly include business-interruption losses, which bias upward the later numbers.

4. Insured vs. Uninsured Losses. Insurance companies are most interested in the total of insured losses. But from an economic perspective, losses to individuals (e.g., workers paid by the hour) are real. The money that would have been spent in the community by the individuals is missing. The National Hurricane Center uses a simple formula to estimate uninsured losses - it doubles the number for insured losses.

5. Private vs. Public Insurance. Flood insurance is provided by the National Flood Insirance Program as a last resort because private insurers have been reluctant to provide it. Individuals pay a premium for this insurance. After a hurricane, there will be payouts. FEMA programs also provide relief to individuals. The National Hurricane Center in its estimate of damage adds in a number for flood damage provided by the National Flood Insurance Program.

6. Federal, State and Local Assistance. Emergency assistance by FEMA, other Federal bodies (the Army Corps of Engineers, for example), states (emergency response teams) and localities (police, fire, sanitation, ambulance) must be factored in as costs.

7. Adjustments for Inflation, Business Activity. Two kinds of adjustments are typically made to comparisons among hurricanes. One is to adjust cost figures for inflation. The Christian Science Monitor story cited above incorrectly describes Hurricane Irene as the fifth mostly costly hurricane in U.S. history. As I explained last year, that label only works if we are under the delusion that a dollar 100 years ago should be valued the same as a dollar today. (Apart from the fact that business-interruption costs are increasingly included in hurricane losses, adding to the size of the numbers.)  There are widely available cost-of-living indicators to refer to, such as this one from the BLS. The business activity measure relates the hurricane to the value of the real estate through which the hurricane travels. This is a good predictor of cost and is also a factor to consider in comparing the impact of a hurricane traveling the same path in different years.

I have written on this topic last year on Huffington Post and my CityEconomist blogsite.