Indicative CBA Model for Earthquake prone building review Summary of methodology and results

Transcription

1 Indicative CBA Mode for Earthquake prone buiding review Summary of methodoogy and resuts Fina report - September 2012

2 CBA Mode PURPOSE AND OUTPUTS 2

3 CBA Mode: Purpose and Outputs 3 The purpose of the Mode is to cacuate the impact (costs and benefits) of possibe changes to the earthquake prone buiding poicy reated to the target %NBS to be appied and the timing of compiance. The mode measures the incrementa difference between a status quo position (assuming existing timeframes for compiance and the current 33% NBS target) and three Poicy Options: A 33% NBS target but a shortening of the timeframe to compy A 67% NBS target with or without a reset to the timeframe to compy, and A 100% NBS target with or without a reset to the timeframe to compy. For the status quo and each of the Poicy Options, the mode cacuates costs and benefits for each Territoria Authority (TA) and sums the resuts to provide a tota Cost/Benefit resut for New Zeaand. In addition, the difference between the current poicy and timing and a no strengthening option is aso measured. Costs that are modeed Buiding strengthening costs An option to incude or excude additiona costs incuding Fire Safety and Disabed Access improvements Benefits from strengthening that are modeed (i.e. reduced costs in the event of an earthquake) Reduced buiding damage Reduced oss of ife Reduced injury Outputs produced A present vaue net cost/benefit for each poicy option based on a 75 year modeing period and a range of earthquake sizes with specific probabiities for each earthquake in each TA. A cost/benefit resut, as above, but measured for a singe event annua earthquake and a singe average probabiity factor. A per event cost/benefit for each poicy option assuming an earthquake actuay occurs at set dates (eg at years 10, 15, 20, 30).

4 CBA Mode: Purpose and Outputs As the CBA Mode measures the NPV of costs and benefits, tota benefits decine each year after the end of strengthening due to annua discounting of the impacts. For exampe, the benefits accruing from an earthquake that occurs in year 25 have a greater present vaue than an earthquake of simiar size that occurs in year 50. Costs and Benefits are incrementa. The CBA Mode measures the annua differences between the current strengthening and timetabe and the poicy strengthening and timetabes. Why a 75 year mode? 4 75 Years is seen as a conservative estimate of the remaining usefu ife of the current earthquake prone buiding stock. It is possibe that buiding ives wi be shorter, but using 75 years makes it more ikey that a potentia benefits are captured. The CBA (cost-benefit anaysis) provides just one piece of the information to be considered in deveoping the overa EPB poicy. The resuts of the CBA shoud be used aongside a other poicy factors and considerations.

5 CBA Mode: Economic & socia costs Economic and socia costs The CBA Mode has not incuded an estimate of benefits that might accrue, after an earthquake event, from ower overa economic or socia costs as a resut of having strengthened earthquake prone buidings (EPBs). The basis for excuding such costs is as foows: If EPBs perform to design then in a arge earthquake they shoud ensure that the occupants wi survive. However, as evidenced in Christchurch, the buidings may sti need to be demoished after the earthquake. It is this subsequent demoition that coud cause an economic oss - and strengthening for the purpose of preserving ife may do itte to prevent this. There is ikey to be a very narrow enveope of earthquake strengths where a strengthened EPB woud remain standing and be useabe after the event, but where it woud not have survived if it hadn t been strengthened. This owers the probabiity of there being a materia reduction in impact on economic and socia outcomes from strengthened EPBs compared to other impacts (such as infrastructure damage). A materia impact (on a New Zeaand scae) on economic and socia outcomes after an earthquake is ony ikey to occur in cities/towns where there is a reativey high probabiity of a arge earthquake and where there is a arge commercia/industria area. This is ony the case in Weington, Hutt City, Christchurch and Pamerston North. 5 Our assumption is that ony very arge earthquakes wi cause economic or socia costs (e.g. Christchurch was a 1 in 2,500 year earthquake). The probabiity of very arge earthquakes is so ow that any probabiity based benefits from strengthening buidings are sma. Aso, with very arge earthquakes, the positive impact of a few strengthened EPBs is ikey to be dwarfed by more significant impacts on infrastructure and residentia dweings. We woud not expect to attribute much (if any) economic or socia impacts to the demise or otherwise of EPBs. Simiary, in a arge earthquake the damage to non- EPBs is ikey to be arge, and the damage to strengthened EPBs wi aso sti be a major factor. These impacts may aso dwarf whatever improvements may have been made by strengthening a sma number of oder buidings.

6 CBA Mode NEW ZEALAND S BUILDING STOCK 6

7 Tota buiding stock Tota New Zeaand buiding stock An estimate of the tota buiding stock by ocation and age for commercia, industria and other nonresidentia buidings, and high-rise residentia buidings, was obtained from Quotabe Vaue (QV). 7 The data received from QV for numbers of buidings was higher than expected. It appeared that the data might incude, in some cases, the numbers of units for muti-unit properties. QV were unabe to easiy separate out the actua buiding numbers from their database. As a resut, buiding numbers have been reported as buidings/units in a mode outputs and reports to take account of this possibe issue with the QV data. Importanty, however, the CBA mode bases the key cacuations on the buiding area (rather than number) so the CBA Mode is not affected by the recording of the number of buidings/units Note: 41% of buidings had age unknown because it was either not recorded or because the buiding was made up of severa parts with different ages. These were aocated pro-rata in the mode. The foowing charts show the make-up of the base data for a New Zeaand buidings:

8 Tota buiding stock The buiding stock is reativey concentrated, with 50% of the tota buiding area for New Zeaand in the top 7 TAs and 80% in the top 25. (There are 74 TAs in the QV database as it incudes seven separate Auckand areas and the Chatham Isands). Tota area for a buidings in New Zeaand is 171,000,000m2. 8

9 Tota buiding stock According to the QV database, the tota number of buidings/units in New Zeaand is 194, % were buit pre 1935 and 31% were buit between 1935 and 1976: The average area per buiding/unit is 884m2, but this differs depending on the age of the buiding. Newer buidings are on average arger than oder buidings, with post 1976 buidings being on average amost twice as arge as pre 1935 buidings. 9

10 Earthquake prone buiding stock (EPBs) Existing data on the numbers of EPBs in New Zeaand is scarce. Severa rough estimates have been made (e.g. Jason Ingham, David Hopkins) but a have recognised the serious ack of data and have caveated their resuts accordingy. The estimates for the CBA Mode are no different. The work has considered and interpreted the avaiabe data, incuding the previous studies, but at best the numbers are ony indicative. To take account of this uncertainty around the makeup of the buiding stock the CBA Mode appies a robust sensitivity anaysis across a the resuts. The estimates for EPBs were deveoped as foows: In March 2012, DBH carried out a survey of a TAs across New Zeaand asking for detais of EPBs (among other things). Unfortunatey, this survey provided very itte data on EPBs with most councis either just starting to gather data on their EPBs or not yet having started. The exception was Weington City Counci which had made significant progress in identifying the city s at-risk buidings. Survey resuts are shown on the foowing page: 10

11 11 Earthquake prone buiding stock (EPBs) Summary of initia TA survey response. Incompete data from a TAs with some usefu data from Weington and Whanganui. (The green boxes show the QV voumes. The bar charts for each TA show data provided in the survey. For data for a TA to be usefu, the red, orange and green bars shoud be cose to the green box) 6,000 5,000 Auckand :Tota QV Buidings/Units 19,050 Tota TA data 17,096 1,400 QV and TA-provided Information on Buiding Stock (pre-1976) 4,000 1,200 1,000 3, , , Tota Buidings / Units Auckand Weington City Christchurch City Dunedin City Lower Hutt City Hamiton City Whangarei District Invercargi City New Pymouth District (Larger Scae) 0 Rotorua District Far North District Southand District Pamerston North City Timaru District Hastings District Napier City South Taranaki District Wanganui District Tauranga City Gisborne District Marborough District Ashburton District Matamata-Piako District Waikato District Waitaki District Tasman District South Waikato District Neson City Horowhenua District Waipa District Thames-Coromande District Cutha District Whakatane District Tararua District Rangitikei District Porirua City Masterton District Taupo District Manawatu District Centra Otago District Grey District Ruapehu District Gore District Kaipara District Sewyn District Upper Hutt City Waimakariri District Kapiti Coast District Hauraki District Wairoa District Buer District Stratford District Centra Hawke's Bay District South Wairarapa District Waitomo District Hurunui District Western Bay of Penty District Westand District Queenstown-Lakes District Opotiki District Carterton District Otorohanga District Waimate District Mackenzie District Kaikoura District Kawerau District Chatham Isands Territory <33% NBS 33-67% NBS >67% NBS Not Assessed Tota QV Pre 1976 Buidings/Units

12 Earthquake prone buiding stock (EPBs) Foowing the poor response from the survey, in Juy 2012, councis of ten of the major cities and towns in New Zeaand were contacted directy by phone and the atest EPB data was requested. Data was obtained between 9 Juy and 19 Juy Dunedin and Rotorua had no current data. Christchurch was not approached and Hamiton had a imited amount of data. As a resut, this approach provided useabe data and estimates from seven councis - on buidings <33%NBS as we as on buidings between 33% and 67%NBS. This group of councis were: Auckand, Weington, Hastings, Neson, Whangarei, Whanganui, and Whakatane. The data coected reated ony to pre-1976 buiding stock. This is because it is most ikey that buidings <33% NBS wi have been buit before a buiding code change in that year and buidings beow 33% have been the focus of the councis to date. 12

13 Earthquake prone buiding stock (EPBs) The foowing tabe shows the detaied information received from the seven councis isted above. The extrapoated data appies the resuts from the assessments to date to the whoe popuation, or in other cases shown beow, estimates for the pre-1976 popuation were provided by the councis More representative data was received from a second source for Weington and this was used in the cacuations. The Councis in the tabe represent 36% of the tota number of pre-1976 buidings and overa the Councis have competed, on average, ony 22% of their assessment programmes. The imited competion of assessments and the risks of extrapoating the data were discussed with each of the Councis. As noted above, even though the resuts are based on the atest and most compete information that is avaiabe at best they can ony be considered as indicative. Detaied information from major centres, and extrapoation to estimate buidings <33%NBS and between 33% and 67% NBS Per Counci Data/Estimate Numbers assessed or on register (pre 1976) Extrapoated to a pre-1976 buiding stock < 33% NBS 33% to 67% > 67% NBS Tota Assessed Pre 1976 QV % Assessed < 33% NBS 33% to 67% > 67% NBS Tota Estimate Auckand ,465 19,050 8% 4,941 5,104 9,005 19,050 Weington 239 1,607 1,865 3,711 5,139 72% 750 3,160 1,229 5,139 << Uses aternative Counci data Hastings ,330 15% ,330 Neson << Counci estimate (not assessment) Whangarei ,616 9% ,616 Whanganui ,237 69% ,237 << Counci estimate (not assessment) Whakatane % Sub-tota 1,317 2,436 2,851 6,603 29,800 22% 8,202 10,214 11,384 29,800 20% 37% 43% 100% 28% 34% 38% 100% A pre 1976 Buidings 81,838 Percent of pre-1976 in tabe 36% 13

14 14 Earthquake prone buiding stock (EPBs)

15 Earthquake prone buiding stock (EPBs) From this data, the ratio of buidings that are 33-67% NBS compared to the buidings that are <33% NBS is 1.2. (This was based on an average of the TA ratios, excuding Weington. Weington was excuded because it appeared to be an outier compared to the other TA resuts). This ratio is appied to a other TAs to estimate the number of buidings between 33% and 67% NBS, based on the recorded or cacuated data for buidings beow 33%. Additiona data sources In addition to the direct contact made with major councis, an internet search was carried out for any further information about EPBs across New Zeaand. This produced additiona information on buidings beow 33% NBS from Counci ists of EPBs, as we as from press reports. (Searches were made by DBH, the Ministry of Cuture and Heritage and MartinJenkins). Additiona data was sourced from 16 TAs: Tauranga, Lower Hutt, Manawatu (Fieding), Pamerston North, Centra Otago, Opotiki, Sewyn, Hamiton, SouthTaranaki, Tararua, Masterton, Marborough, Hurunui, Waimakariri, Cutha, and Westand. This data was of ow quaity, comprising estimates that incuded either most of the pre-1976 buiding stock, or were from partiay competed assessment programmes. The combined data from a the 23 councis with usabe information provided the overa indicative estimate of the EPB stock. This data suggested that, across New Zeaand, 24% of a pre-1976 buidings coud be beow 33% NBS. In comparison, the average percentage for the seven councis with higher quaity data was 28%. An indicative estimate of the EPB status of the New Zeaand buiding stock is shown in the chart on the foowing page Note: buidings buit after 1976 have not been assessed, but it woud be fair to assume that a significant number of such buidings wi be between 33% and 67% NBS. 15

16 Earthquake prone buiding stock (EPBs) Across New Zeaand there are estimated to be about 82,000 pre-1976 buidings/units out of the tota buiding stock of 194,000. This is 42% of the tota. Buidings beow 33% are estimated to be 10% of the tota, being 19,000 buidings/units. 16

17 Earthquake prone buiding stock (EPBs) Summary of point estimates and indicative range of buiding stock numbers (before appying attrition rates for buidings that woud not be strengthened regardess of the poicy settings): EPBs - Buiding Stock Assumptions Number of Buidings / Units - Point Estimates Number of Buidings / Units - Range of Estimates EPB Status Age EPB Status Number Percent of Tota Pre 1976 Post 1976 Tota < 33% NBS 15,000 to 25,000 8% to 13% < 33% NBS 19,000 19, % NBS 15,000 to 30,000** 8% to 16% 34-67% NBS 23,000 0* 23,000 > 68% NBS 135,000 to 160,000 71% to 84% > 68% NBS 40, , ,000 Approx Tota 190,000 Tota 82, , ,000 ** Excuding post 1976 buidings * Not assessed Numbers for 34-67% NBS (post 1976 buidings) are not incuded in the resuts. No actua data or Counci estimates are avaiabe for this group of buidings, but anecdota evidence (from MBIE-BH) suggests that up to an additiona 50,000 post buidings coud be between 33% and 67% NBS. 17

18 Earthquake prone buiding stock in CBA mode CBA Mode - use of buiding stock The buiding stock used in the CBA mode is reduced by an attrition rate of 10% for <33% NBS buidings and 5% for 33%-67% NBS buidings to take account of buidings not expected to be strengthened regardess of the future poicy. These rates are judgement-based, and the ower rate for 33%-67% NBS refects that these buidings are more ikey to be newer (and therefore more ikey to be strengthened) than the <33% NBS buidings. As noted above, the arge uncertainty around the actua make-up of the buiding stock is managed in the CBA Mode by appying a robust sensitivity anaysis across a the resuts. 18

19 Strengthening costs for individua buidings Costs of strengthening buidings A range of strengthening costs per m2 of foor area for 33%, 67% and 100% strengthening options was provided by engineers (Adam Thornton and Win Cark; and separatey by John Hare). The ranges were necessariy arge because of the wide nature of the buidings, as we as to account for some regiona variation in buiding costs. The tabe beow shows the costs used in the base case of the CBA Mode. The tabe shows the costs of strengthening pre 1935 buidings, and buidings buit between 1935 and 1976, to 33%, 67% and 100% NBS. The charts on the foowing pages show the avaiabe data-sets and the amounts used for the CBA Mode base case (as aso shown in the preceding tabe). The data seected for the mode was based on the mid-points of the Adam Thornton and Win Cark estimates. (The estimate used the weighted average resuts of the and data). Cost uncertainty was managed in the CBA mode by appying a robust sensitivity anaysis to the resuts 34%NBS 67%NBS 100%NBS Pre 1935 buidings $300 $510 $ buidings $416 $640 $807 19

20 Strengthening costs for individua buidings Summary of engineers cost estimates and the CBA Mode seection: 20

21 Costs of strengthening buidings - and timing to compete Additiona costs Costs of fire and disabed access upgrades are abe to be incuded in the CBA mode, but were set at zero for the base case. Phasing of strengthening timetabe The CBA Mode incudes a phasing of the strengthening timetabe that can be set to various distribution shapes incuding fat, norma and a ate skew. The base case uses a ate skew, with greater strengthening occurring towards the end of the egisated timeframe, as foows: Strengthening Cost Methodoogy Cost is a function of tota <33% NBS and 33%-67% NBS buidings per TA, ess attrition (those not strengthened), times cost/m2 to strengthen (pus additiona costs if seected), aocated across the poicy timeframe based on a seected distribution type (skewed, norma, inear)

22 CBA Mode TOTAL COSTS OF STRENGTHENING 22

23 Summary of CBA Mode Strengthening Costs Three base case options were modeed: (1)33% strengthening with shortening of the timeframe from the current average period of 28 years to a maximum timeframe of 15 years (2)67% strengthening with a timeframe of 15 years (3)100% strengthening with a timeframe of 15 years As noted earier, the status quo assumes that the current standard of 33% is impemented according to existing timeframes. This is an average of 28 years. The three options above are compared to the status quo (current 33% case) to provide an estimate of incrementa costs. The foowing page summarises the base case modeed resuts: 23

24 Summary of CBA Mode Strengthening Costs (3 strengthening options: (1) 33% with reduced timing; (2) 67%; and (3) 100% Buiding Strengthening Costs Attrition Rate (being percent of stock not strengthened for any reason (incuding aowing for some aready strengthened, and aowing for tota repacement where this is <33% NBS - Attrition Rate 10% economic in a scenarios). 33% -67% NBS - Attrition Rate 5% Buiding Stock for Strengthening (tota stock ess Attrition) Number Area Average m2 per buiding <33% NBS 17,424 11,994, % -67% NBS 22,012 15,617, Tota 39,435 27,611, Base Strengthening Costs To 33% NBS To 67% NBS Ratio To 100% NBS Ratio Incude Additiona Fire/Access Costs? <33% NBS $300/m2 $510/m $615/m No 33% -67% NBS $640/m2 $807/m Tota Strengthening Costs for 3 options Tota Rea $m NPV $m Strengthening Timing Assumptions (used for NPV) Maximum Time to Strengthen <33% to 33% - Current Timing 28 Years $3,598m $958m Status Quo (weighted average of TAs) 28 Years <33% to 33% - Poicy Timing 15 Years $3,598m $1,717m Maximum Years to Improve buidings <33% NBS 15 Years Incrementa Cost vs Current 33% Case > $0m $760m Maximum Years to Improve buidings 33% -67% NBS 15 Years <33% to 67% $6,117m $2,919m 33% -67% to 67% $10,000m $4,772m Tota Cost to Strengthen to 67% $16,117m $7,692m Incrementa Cost vs Current 33% Case > $12,519m $6,734m 24 <33% to 100% $7,376m $3,520m 33-67% to 100% $12,599m $6,012m Tota Cost to Strengthen to 100% $19,975m $9,533m Incrementa Cost vs Current 33% Case > $16,377m $8,575m

25 CBA Mode BENEFITS 25

26 Benefit Inputs and Cacuations Benefit Inputs and Cacuations Information on the benefits arising from earthquakes is imited and difficut to estimate. GNS has provided a the base information for the CBA Mode and have run a number of specific modes to provide the required inputs. The CBA Mode takes the GNS resuts and extrapoates and modifies these for a the TAs using, where reevant, popuation, buiding areas, geography, seismicity or earthquake probabiities. The three modes were expanded by GNS to six modes by tweaking the inputs to produce ow seismicity equivaents. The modes are: Weington HIGH Popuation: 179,000 Weington LOW Hastings HIGH Popuation: 62,000 Hastings LOW Greytown HIGH Popuation: 2,000 Greytown LOW 26 Three representative high seismicity cities/towns (Weington, Hastings, Greytown) were modeed for earthquakes by GNS, producing estimates of expected workday deaths and injuries, non-workday deaths and injuries and damage costs. The modes aso estimated impacts on the pubic outside of buidings, on the city/town streets. These centres were seected by GNS due to the data and modes that were avaiabe at the time. They were aso seected as they gave a fair representation of three difference sizes of New Zeaand centres: arge city, medium centre, sma town. For exampe, the LOW seismicity Weington mode is designed to represent the impacts in a city such as Auckand. (Wgtn Z-factor 0.40; Auckand Z-factor 0.13). The six GNS mode resuts were scaed and appied to other cities/towns across NZ based on reative popuations for major towns and cities in each TA as we as seismicity. An aternative aocation using reative buiding areas in each TA was aso cacuated but this did not produce a materia difference so the popuation aocation driver was retained.

27 27 GNS Modes Exampe Output (Workday deaths MM8 MM11) This chart of workday deaths demonstrates the expected ow impact of MM8 and MM9 earthquakes and the tota scae of a arge earthquake. (Note that the three LOW seismicity scenarios have weaker buidings and therefore suffer greater harm than HIGH seismicity scenarios, given the same size of event. E.g. Wgtn Low > Wgtn High).

28 28 GNS Modes Exampe Output Repairs ($m) MM8 MM11 This chart for Repairs shows a simiar effect as the chart showing deaths, except there is some damage experienced in the smaer MM8 and MM9 earthquakes.

29 Benefit Inputs and Cacuations (cont) 29 Aocation of GNS modes to TAs The six modes were assigned to TAs as foows: Weington uses Weington High, Hastings uses Hastings High, Greytown uses Greytown High Z-factors for these centres are 0.40, 0.39 and 0.42 respectivey If a Z-factor for a TA is ess than 0.20 it uses a LOW mode, otherwise it uses a HIGH mode If tota popuation of a TA is greater than 75,000 it uses a Weington mode, if average town popuation in a TA is beow 5,000 it uses a Greytown mode, otherwise it uses a Hastings mode The Z-factor and popuation steps were varied as part of the sensitivity anaysis Each of the 6 modes provided by GNS incuded resuts (number of deaths/injuries in a workday/nonworkday and costs of damage) under earthquake strengths from MM8 to MM11*. This band was seected because: MM12 earthquakes (the maximum) have a huge impact but are not predicted to occur in New Zeaand s towns or cities Earthquakes ower than MM8 have a higher probabiity of occurrence but are not expected to cause materia damage for the purposes of the CBA anaysis * The MM scae is the Modified Mercai Index. The size of an earthquake is often described using magnitude, which is the amount of energy reeased during an earthquake. However, not a of the energy reeased in an earthquake wi necessariy be fet at the surface, depending on the earthquake's depth. In New Zeaand, where earthquakes occur from near the surface right down to a depth of over 600 km, the Modified Mercai intensity scae is a better indicator of an earthquake's effects on peope and their environment The MM scae is described in summary form on the foowing page:

30 Background information: Modified Mercai Intensity Scae 30 Modified Mercai Intensity scae: MM8 to MM11 have been modeed by GNS for the CBA MM 1: Imperceptibe Barey sensed ony by a very few peope. MM 2: Scarcey fet Fet ony by a few peope at rest in houses or on upper foors. MM 3: Weak Fet indoors as a ight vibration. Hanging objects may swing sighty. MM 4: Light Generay noticed indoors, but not outside, as a moderate vibration or jot. Light seepers may be awakened. Was may creak, and gassware, crockery, doors or windows ratte. MM 5: Moderate Generay fet outside and by amost everyone indoors. Most seepers are awakened and a few peope aarmed. Sma objects are shifted or overturned, and pictures knock against the wa. Some gassware and crockery may break, and oosey secured doors may swing open and shut. MM 6: Strong Fet by a. Peope and animas are aarmed, and many run outside. Waking steadiy is difficut. Furniture and appiances may move on smooth surfaces, and objects fa from was and sheves. Gassware and crockery break. Sight non-structura damage to buidings may occur. MM 7: Damaging Genera aarm. Peope experience difficuty standing. Furniture and appiances are shifted. Substantia damage to fragie or unsecured objects. A few weak buidings are damaged. MM 8: Heaviy damaging Aarm may approach panic. A few buidings are damaged and some weak buidings are destroyed. MM 9: Destructive Some buidings are damaged and many weak buidings are destroyed. MM 10: Very destructive Many buidings are damaged and most weak buidings are destroyed. MM 11: Devastating Most buidings are damaged and many buidings are destroyed. MM 12: Competey devastating A buidings are damaged and most buidings are destroyed.

31 1,200 1,100 1, A TAs - Workday deaths MM10 (MM11 has higher fataities than MM10, MM9 has fewer and MM8 has ni) This chart shows exampe resuts of appying the 6 modes to the TAs for an MM10 event. Significant impacts are ony expected in arge centres with high popuations. Aso note: non-workday impacts (not shown beow) are approximatey 1/10 th of workday impacts as fewer peope are in the EPBs. Expected Workday Deaths MM10 (before event probabiity appied) Strengthening 0% Strengthening 33% Strengthening 67% Strengthening 100% Far North District Whangarei District Kaipara District Rodney District North Shore City Waitakere City Auckand City Manukau City Papakura District Frankin District Thames-Coromande District Hauraki District Waikato District Matamata-Piako District Hamiton City Waipa District Otorohanga District South Waikato District Waitomo District Taupo District Western Bay of Penty District Tauranga City Rotorua District Whakatane District Kawerau District Opotiki District Gisborne District Wairoa District Hastings District Napier City Centra Hawke's Bay District New Pymouth District Stratford District South Taranaki District Ruapehu District Wanganui District Rangitikei District Manawatu District Pamerston North City Tararua District Horowhenua District Kapiti Coast District Porirua City Upper Hutt City Lower Hutt City Weington City Masterton District Carterton District South Wairarapa District Tasman District Neson City Marborough District Kaikoura District Buer District Grey District Westand District Hurunui District Waimakariri District Christchurch City Sewyn District Ashburton District Timaru District Mackenzie District Waimate District Chatham Isands Territory Waitaki District Centra Otago District Queenstown-Lakes District Dunedin City Cutha District Southand District Gore District Invercargi City 31

32 Benefit Inputs and Cacuations (cont) Earthquake probabiity modeing Once a mode was assigned to each TA, and scaed up or down according to reative popuations, an earthquake probabiity specific to the TA was appied to each of the MM8 to MM11 resuts. The data for the earthquake probabiities for MM8 to MM11 was provided by GNS. This was for the 30 major towns/cities of New Zeaand (and updated earier work provided for a report by David Hopkins in 2002). Data for each of the 30 towns/cities was matched to the reevant TA. For TAs not covered in the ist, probabiities were appied to each TA based on the nearest isted data - taking into account both geography & seismicity. Where any judgement was required between two options the higher probabiity factor was appied. This is a conservative approach for the CBA Mode, refecting the preiminary resuts that showed very ow benefits compared to costs. This conservative approach was adopted throughout the anaysis in order to provide greater comfort over the cacuations - and greater confidence around the reationship between estimated costs and estimated benefits. For each TA, the CBA mode sums a the impacts of the major earthquakes (MM8-MM11) taking into account their respective probabiities. This produces average annua workday and non-workday deaths/injuries and average damage per TA for a major earthquakes. The probabiity based resuts of the earthquake modeing are combined into a tota probabiity factor for each TA, together with outputs for weighted average workday and non-workday deaths/injuries and average costs of damage. These are used as inputs into the main part of the CBA Mode. 32

33 Probabiity percentages for 30 main towns and cities across NZ In summary, the GNS percentages show very few areas are expected to have earthquakes of strength MM10 or MM11. Whie there is a much higher probabiity of an MM8 earthquake across the country, an event of that size is not expected to cause any deaths and ony a sma amount of damage. (Probabiities are shown as additive for presentation purposes). 30 Main Cities and Towns Additive Annua Earthquake probabiities for MM8 to MM11 (data abe for MM11) 1.600% 1.400% 0.005% 1.200% 0.011% 0.003% 0.012% 0.003% 0.002% 1.000% 0.002% 0.800% 0.003% 0.001% 0.001% MM11 MM % MM9 MM % 0.200% 33

34 Expected deaths after appying probabiity annua average by TA (orange = off the scae) Once probabiities are appied to the tota number of deaths that coud be caused by an earthquake event, the expected deaths are very sma in a TAs except Pamerston Nth, Lower Hutt, Weington and Christchurch. The Auckand cities a have very ow probabiity of earthquakes, giving ow expected deaths in spite of high popuations. Weighted Average Expected Deaths p.a. (MM10 Quake) Zero % Strengthening Far North District Whangarei District Kaipara District Rodney District North Shore City Waitakere City Auckand City Manukau City Papakura District Frankin District Thames-Coromande District Hauraki District Waikato District Matamata-Piako District Hamiton City Waipa District Otorohanga District South Waikato District Waitomo District Taupo District Western Bay of Penty District Tauranga City Rotorua District Whakatane District Kawerau District Opotiki District Gisborne District Wairoa District Hastings District Napier City Centra Hawke's Bay District New Pymouth District Stratford District South Taranaki District Ruapehu District Wanganui District Rangitikei District Manawatu District Pamerston North City Tararua District Horowhenua District Kapiti Coast District Porirua City Upper Hutt City Lower Hutt City Weington City Masterton District Carterton District South Wairarapa District Tasman District Neson City Marborough District Kaikoura District Buer District Grey District Westand District Hurunui District Waimakariri District Christchurch City Sewyn District Ashburton District Timaru District Mackenzie District Waimate District Chatham Isands Territory Waitaki District Centra Otago District Queenstown-Lakes District Dunedin City Cutha District Southand District Gore District Invercargi City 34

35 Benefits CBA mode cacuations 35 Benefit Cacuations in the CBA Mode The data provided by the GNS modes and probabiity cacuations (as described in the previous section) resuts in annuaised estimates of deaths, injuries and damage in the event that an earthquake occurs in each of the TAs across New Zeaand. These resuts are appied in the CBA mode as foows: In the first year of the CBA mode, amost no strengthening has been competed. Therefore, the impact if there was an earthquake woud be equa to the impact measured in the GNS work for a no strengthening scenario. Once strengthening is competed (at the end of the maximum aowed time under the prescribed poicy) the annua benefits wi be equa to those measured in the reevant poicy option provided by the GNS modes. Over the interim period, as strengthening graduay moves from zero to fuy competed, the benefits increase from zero to the fu amount of the benefit - in proportion to the cumuative amount of strengthening competed at the end of each year. 1,200 1, (200) (400) Once strengthening is competed the benefits remain fat, in rea terms, for the rest of the mode. As the CBA mode measures the Net Present Vaue of costs and benefits over the term of the mode (75 years) the benefits are discounted to current day using the rea discount rate. The impact of benefits in outer years is therefore ess than the impact of the same benefits were they to occur in nearer years. The highest benefits therefore occur in the year that strengthening is competed and they then decine over time unti the concusion of the mode. The foowing schematic of a singe TA shows the peaks of poicy options for 33%, 67% and 100% NBS at 15 Years timing. Current timing of 30 years is aso shown in green (this chart is shown in more detai ater in this report). Auckand City: Incrementa Net Cost / Benefit of a MM10 Earthquake occuring during a workday in each year (NPV$m) (Peaks in Year that strengthening is competed) Current Timing 33% vs No Strengthening New Timing 33%NBS vs No Strengthening New Timing 67%NBS vs No Strengthening New Timing 100%NBS vs No Strengthening

36 Summary of CBA Mode Benefits Data The foowing page shows exampe data from the CBA mode for deaths and buiding coapses for 4 cities (Weington, Christchurch, Auckand and Hastings). This data iustrates the impact of the probabiity assumptions on the estimated resuts. For exampe, athough a arge earthquake woud have a very arge impact in Auckand, the probabiity of this occurring is considered by GNS to be ni (or too ow to measure for the purposes of modeing the benefits). The base case probabiity resuts produced expected annua deaths for New Zeaand in a no strengthening case of approximatey 1 per annum. This appeared ow in comparison to historica deaths over the ast 169 years which were an average of 3 per annum. Since 1929 the average has been 6 per annum (incuding Canterbury) and an approximate popuation adjusted estimate of the annua figure is 8.5 per annum. A stretch case sensitivity has been run which repaces the GNS probabiities for each TA with the highest evidenced probabiity factor for each earthquake size. On this basis, the expected annua deaths are approximatey 8 per annum However, whie this case significanty increases the impact of earthquakes and the eve of benefits, the benefits sti remain consideraby ess than the costs, particuary for the 67% and 100% strengthening options. As with other assumptions, uncertainty in benefit cacuations was managed by appying a robust sensitivity anaysis to the resuts 36

37 Benefits modeing - Exampe resuts Estimated Deaths - daytime work-day Earthquake Estimated Buiding Coapses Return No Strengthen 33% NBS 67% NBS 100% NBS Period (Years) No Strengthen 33% NBS 67% NBS 100% NBS Weington MM MM MM , MM11 1,521 1, , Expected Annua Impact Christchurch MM MM , MM , MM11 3,071 2,372 1, Expected Annua Impact Auckand City MM , MM , MM10 1, MM11 4,467 3,204 2,063 1, Expected Annua Impact Hastings MM MM , MM , MM Expected Annua Impact Tota New Zeaand (p.a.) Sensitivity anaysis resuts - maximum probabiity (return periods for MM8 to MM11 for a cities of 99, 350, 1499 and 8475) Tota New Zeaand (p.a.)

38 Benefits modeing resuts assuming 15 year strengthening timetabe Tota benefits for New Zeaand and contribution from each earthquake band for MM8 to MM11 33% New Timing vs Current Poicy 67% New Timing vs Current Poicy 100% New Timing vs Current Poicy Benefits from MM8 earthquakes $3m $13m $25m Benefits from MM9 earthquakes $2m $9m $18m Benefits from MM10 earthquakes $1m $6m $10m Benefits from MM11 earthquakes $0m $2m $3m Tota Benefits a earthquakes $6m $30m $56m Much greater benefits are reaised from ower intensity earthquakes because, even though they cause ess damage and deaths, they have a much higher probabiity of occurring. $70m $60m $50m $40m $30m Benefits by Earthquake size MM8 to MM11 $30m $56m Benefits from MM11 earthquakes Benefits from MM10 earthquakes Benefits from MM9 earthquakes $20m $10m $6m Benefits from MM8 earthquakes 38 $0m 33% New Timing vs Current Poicy 67% New Timing vs Current Poicy 100% New Timing vs Current Poicy

39 CBA Mode COST / BENEFIT RESULTS 39

40 Cost / Benefit Cacuations in the CBA Mode Cacuation of net cost/benefit resuts in the CBA Mode The fina step in the CBA mode cacuations is to combine the annua costs and benefits and to cacuate the Net Present Vaue of the expected resuts for the 75 years of the mode. The foowing tabe sets out the tota costs and benefits for the base case, assuming a 15 year maximum timeframe for strengthening. It aso shows net costs and benefits assuming 10 year and 5 year maximums. 40

41 CBA Resuts high costs, ow benefits (Indicative resuts based on the assumptions described in this report) Indicative Cost / Benefit under 3 new timing options (15, 10 and 5 years) COST BENEFIT NET NET Increase COST BENEFIT - Mid point estimate based on extrapoated Counci data NPV $m NPV $m NPV $m $m Cumuative on Prior Cumuative Cumuative Cost/benefit to achieve 33% NBS at current timeframes (average 28 years) $958m $25m ($932m) ($932m) $958m $25m Additiona cost/benefit to achieve 33% NBS in 15 years $760m $11m ($748m) ($1,680m) 80% $1,717m $37m Pus additiona cost/benefit to achieve 67% NBS in 15 years $5,974m $52m ($5,922m) ($7,603m) 352% <See beow> $7,692m $89m Pus additiona cost/benefit to achieve 100% NBS in 15 years $1,841m $56m ($1,785m) ($9,388m) 23% $9,533m $144m NPV $m NPV $m NPV $m $m Cumuative Cumuative Cumuative Cost/benefit to achieve 33% NBS at current timeframes (average 28 years) $958m $25m ($932m) ($932m) $958m $25m Additiona cost/benefit to achieve 33% NBS in 10 years $1,237m $22m ($1,215m) ($2,147m) 130% $2,194m $47m Pus additiona cost/benefit to achieve 67% NBS in 10 years $7,635m $67m ($7,568m) ($9,716m) 352% $9,829m $114m Pus additiona cost/benefit to achieve 100% NBS in 10 years $2,353m $71m ($2,281m) ($11,997m) 23% $12,182m $185m NPV $m NPV $m NPV $m $m Cumuative Cumuative Cumuative Cost/benefit to achieve 33% NBS at current timeframes (average 28 years) $958m $25m ($932m) ($932m) $958m $25m Additiona cost/benefit to achieve 33% NBS in 5 years $1,840m $35m ($1,805m) ($2,738m) 194% $2,798m $60m Pus additiona cost/benefit to achieve 67% NBS in 5 years $9,735m $85m ($9,649m) ($12,387m) 352% $12,533m $145m Pus additiona cost/benefit to achieve 100% NBS in 5 years $3,000m $91m ($2,908m) ($15,296m) 23% $15,532m $237m Expanation of Cumuative Cost Increases (15 year case) After Attrition After Attrition Buiding NPV Costs Numbers Ratios Area Avg Size m2 Nomina Cost Base costs to strengthen to 33% over average 28 years $958m 17,424 11,994, $3,598m Additiona timing impact of strengthening more quicky $760m 17,424 11,994,163 $0m Tota Cost of strengthening to 33% over 15 years a $1,717m 17,424 11,994,163 $3,598m Additiona costs to strengthen <33% buidings to 67% $1,202m 17,424 11,994,163 $2,519m Tota cost to strengthen <33% buidings to 67% b $2,919m 17,424 b / a 1.70 Note 1 11,994,163 $6,117m Additiona costs to strengthen 33% -67% buidings to 67% c $4,772m 22,012 c / b 1.63 Note 2 15,617, $10,000m Tota Cost to strengthen both <33% and 33%-67% to 67% $7,692m 39,435 $16,117m Note1: This ratio refects the reative costs of strengthening: Note 2: This ratio refects the greater number of buidings and the higher cost per m2 to repair 33% -67% compared to <33%. To 33% $300/m2 Ratio 33% -67% numbers to <33% numbers 1.30 (Being 15,617,133 / 11,994,163) To 67% $510/m2 Ratio 33% -67% Cost to <33% Cost 1.26 (Being $640/m2 / $510/m2) Ratio: 1.70 Tota Ratio

42 CBA Resuts high costs, ow benefits Moving to the 67% option is the argest step with the highest net cost. This is mainy due to the greater voume of buidings that are incuded in a move to 67% compared to 33%. A move to 67% woud be expected to require strengthening of a arge number of buidings buit between 1936 and 1976, whereas a 33% poicy mosty incudes buidings buit before If a new poicy is based on shorter strengthening timeframes than those currenty in pace, this is ikey to resut in the additiona costs being greater than the additiona benefits. This is unfavourabe from CBA perspective, but may be desired for other reasons. Benefit/cost ratios of the options are as foows. A are unfavourabe and greater strengthening and shorter timeframes have ower BCRs: Benefit / cost ratios BCR to achieve 33% NBS at current timeframes (average 28 years) BCR to achieve 33% NBS in 15 years BCR to achieve 67% NBS in 15 years BCR to achieve 100% NBS in 15 years BCR to achieve 33% NBS in 10 years BCR to achieve 67% NBS in 10 years BCR to achieve 100% NBS in 10 years BCR to achieve 33% NBS in 5 years BCR to achieve 67% NBS in 5 years BCR to achieve 100% NBS in 5 years

43 43 As noted earier, GNS base probabiity data ony generates average deaths p.a. of 1.0. (This forms part of a the benefits bocks in the chart except the fina bock). A modified sensitivity produces 8.5 deaths p.a. and tota benefits on this basis are shown in the ast bock on the chart. This sensitivity (for benefits) is more in ine with actua historica deaths, adjusted for popuation growth. In a cases, even with extreme sensitivity scenarios, costs substantiay exceed benefits. This is mainy because arge earthquakes that cause significant damage are very rare, and smaer more common earthquakes don t cause very much damage.

44 Aternative actua event resuts Not based on probabiities This chart shows the combined tota impact of separate actua events occurring in each year for a TAs in New Zeaand. The resuts are for an indicative scenario - an MM10 earthquake on a Workday. Tota New Zeaand: Incrementa Net Cost / Benefit of a MM10 Earthquake occuring during a workday in each year (NPV$m) (Peaks in Year that strengthening is competed) 7,000 6,000 5,445 5,000 4,000 3,000 2,000 1,000 0 (1,000) (2,000) Current Timing 33% vs No Strengthening New Timing 33%NBS vs No Strengthening New Timing 67%NBS vs No Strengthening New Timing 100%NBS vs No Strengthening 44 Net benefits peak in the year strengthening is competed (poicy set at 15 yrs above, existing average is 28 years) 67% and 100% NBS options are positive ony between years 5 and 47 The 33% new timing case is more favourabe than the 67% or 100% cases after year 32 The shape of these curves differs across TAs. However, the underying principe is that that where net benefits are positive, there is a reativey short window of time where this is the case. The effect of discounting decreases the vaue of future benefits in comparison to the costs (which are incurred up front).

45 Concusions of CBA modeing There are significant data imitations... The TAs currenty have very poor records of both their tota buiding stock and in particuar the number of earthquake prone or earthquake risk buidings. Based on survey returns, Weington City is the ony TA where reasonabe data on %NBS of buidings has been coected to date. However, even in Weington the process is ony partiay competed. The costs of strengthening buidings is difficut to estimate on a genera a of New Zeaand basis. This is because of the wide range of buiding types that need to be strengthened and the imited amount of data that is pubicay avaiabe on costs of competed work. There is currenty no New Zeaand-wide mode that measures the impact of earthquakes, athough GNS is in the process of expanding RiskScape. The CBA mode has had to be constructed using avaiabe data from the imited modeing work aready undertaken by GNS.... But cacuations are possibe - and sensitivity anaysis provides comfort around concusions On a probabiity basis, costs are we in excess of benefits Even under extreme sensitivities, this reationship does not change On an actua event basis, there is ony a sma time window where higher strengthening options show net benefits. This window wi shrink and may disappear if higher assumptions were used for buiding stock numbers. The CBA aone does not support higher eves of strengthening - or shorter timeframes. The CBA mode base case has been deveoped on a conservative basis to provide increased robustness over the mode outputs and greater comfort over the concusions above. 45