Policy implications of the 1999 earthquakes.
In the light of the above, the aim of this chapter is to draw out the implications of the 1999 earthquakes and of Turkey's susceptibility to high rates of material damage, especially to residential infrastructure, for the budgetary and structural policy reform issues discussed in the previous chapters. The chapter has the following structure. The first section presents the general factual background, including geological characteristics and impact effects of the earthquake. The second section investigates the economic and budgetary costs of the earthquakes and their implications for economic growth and the budget. Understanding such costs is necessary for near-term budget planning and the longer-term re-allocation of national resources. The third section describes the regulatory and governance factors behind the country's vulnerability to heavy earthquake damage, including factors related to building codes and zoning requirements. Coordinating central and local government roles here is obviously essential. Bu t achieving the ultimate goal of damage minimisation also implies changing private-sector behaviour and culture through a combination of education, regulatory reform and private incentives. In this context the fourth section looks at the policy issues surrounding better private sector risk management, through the development of the insurance and mortgage markets. The final section contains an assessment of the catastrophe, and the main lessons it holds for enhanced disaster preparedness.
Geological background to the earthquakes
The earthquakes struck western Turkey on one of the world's longest and best-recorded strike- slip (horizontal motion) faults: the east-west trending North Anatolian fault, This fault is very similar to the San Andreas fault (Map 1), which has led to active scientific collaborations between scientists in Turkey and the US aimed at understanding the hazards both countries face.
Turkey has a long history of large earthquakes, which have occurred in progressive adjacent historical phases. The August 1999 event was the eleventh, with a magnitude greater than or equal to 6.7, of a sequence of major earthquakes which started in 1939. By 1944 the earthquake locations had moved westward, rupturing 600 km of contiguous fault. An additional adjacent 100 km of fault then ruptured in the events of 1957 and 1967. The August and November 1999 events filled in a 100 to 150 km long gap between the 1967 event and two smaller disturbances which took place further west during 1963-64 (Map 2). The severity of the first, magnitude 7.4, earthquake is underlined by the fact that there was as much as five metres of horizontal fault slip and two metres of vertical slip. Geologists now expect that Southern Istanbul will, with very high probability, experience an event with significantly larger intensity of ground shaking than in the Marmara, within at most 30 years and probably within the next decade.
Despite the fact that the epicentre of successive earthquakes seems to be moving westward along the North Anatolian fault with a high degree of regularity, a comprehensive national study of disaster risk is currently unavailable. The scientific study of the geological fault lines in Turkey has only been undertaken in recent decades and geological surveys conducted by professionals who follow established international practice apparently exist for certain localities only. Geological mapping has suffered from a lack of coherent direction, unreliable compilation, inadequate financing, and a shortage of trained personnel. A national assessment of areas of risk has been undertaken using proxy data, including the recorded levels of earthquake frequency and intensity in recent decades. This generated a 1996 map by the Ministry of Public Works, which divides the country into five zones. This division is not detailed enough to distribute insurable risks (except over entire regions) or determine land use plans, althoug h more than half the population, numerous dams and three-quarters of the nation's industry are found in the two most hazardous zones. At the local level, risk assessment is compromised by a gap in coverage at a scale large enough (minimum 1: 1 000) to take property lines and the outline of buildings into account.
A high mortality and injury rate
The two earthquakes caused considerable damage to housing, public facilities and infrastructures, but the numbers of dead and injured dominate the tragedy (Table 33 and Figure 19). Over 18 thousand people are estimated to have died, and around 50 thousand were injured, of which perhaps two-fifths will be left permanently disabled. Large portions of the area were devastated, with around 113 thousand housing units and business premises completely destroyed, and another 264 thousand damaged to varying degrees. Numerous schools, health facilities, roads, bridges, water pipes, power lines, phone lines, and gas pipelines were severely damaged. Up to 600 thousand people were forced to leave their homes, of which perhaps half became homeless and had to stay in tents, and many of the survivors, especially children, were left deeply traumatised. The characteristics of this population and of the survivors are not known: how many parents lost children, how many children are orphaned, or lost one parent or a sibling, etc. Information of this kind is needed to understand better the relation between the constructions and vulnerability (e.g., what types of household tend to live on what types of structure), as well as long-term health and social welfare costs.
As in any natural disaster, the level of the earthquake's destructiveness reflected not only the intensity of the shock, but also the vulnerability of structures subjected to this shock. The maximum intensities of ground shaking in Turkey were less than what was recorded in the earthquakes in Northridge, California in 1994, and Kobe, Japan in 1995, yet the loss of human life was at least an order of magnitude higher.  Since population density was similar, this reveals the far higher vulnerability of structures in Turkey. Surface fault opening, ground shaking, and soil-liquefaction caused structural damage that was dramatically exacerbated by poor construction quality. Expert evaluations of the post-earthquake devastation confirm that much of it could have been avoided with proper siting and construction practices (Box 9).
In most other OECD economies, good construction techniques, enforcement, land use planning and emergency planning usually mean that injuries and deaths are low, while property losses can be very significant. Property losses have been rising throughout the 1990s, at least in part because natural disasters in highly urbanised regions have been more frequent and more intense. In both developed and developing countries, some social groups, and especially the poor, are most at risk, their exposure often being a function of the location and quality of housing. However, most of the people affected in the Turkish earthquakes were not the poor but the middle class. The earthquakes occurred in a fast-growing region with high per capita incomes, and where population growth had been managed through the construction of multi-storey housing which had replaced traditional buildings. The ratio of numbers of deaths to numbers of houses destroyed in August (over 17 000 deaths and over 60 000 houses destroyed) is exceptional wh en compared to other earthquakes in OECD Member countries (Table 33).
... and an initially difficult relief effort
A contributory factor in the heavy loss of human life may have been the difficulties in mounting a timely and effective official emergency response to the first earthquake. To a large extent, this was due to the fact that it struck at 3 a.m. in a heavily populated and large area. Many people were asleep and trapped inside collapsed buildings, including the officials responsible for implementing emergency response efforts.  Emergency response resources were largely destroyed or damaged. In the critical first hours, rescue efforts were provided by on-site survivors mobilising themselves in an ad hoc way. Telecommunications, transport, and electricity infrastructures were severely damaged by the earthquake and initially overwhelmed by outsiders trying to phone or drive to the region.  This in turn prevented the arrival of civil defence rescue units and medical teams (dispatched by Ankara) until early evening. Even after arrival, the lack of non-telephone communications (radios) hampered the effectiveness of the rescuers. Full telecommunications and electric power was restored only two to three days after the earthquake. 
Economic impact of the disaster
The region affected by the earthquake is both geographically extensive and economically dynamic. It forms the industrial heartland of Turkey. The major industries are autos, petrochemicals, manufacturing and repair of motor (and railway) vehicles, basic metals, production and weaving of synthetic fibres and yarns, paint and lacquer production, and tourism.  The four districts most severely affected (Kocaeli, Sakarya, Bolu and Yalova) contribute over 7 per cent of the country's GDP and 14 per cent of industrial value added (Table 34). Per capita income is almost double the national average. Though containing only 4 per cent of the nation's population, the region contributes over 16 per cent of budget revenues. The immediately surrounding districts (of Bursa, Eskisehir, and Istanbul) have been mainly affected indirectly by their close economic linkages with the former area, e.g., industries and small businesses supplying services or material inputs to each other's production processes. They also are subject to a shared seismic risk and so face magnified uncertainty for the future as a fall-out of the recent events. Taking all seven cities together, the wider earthquake region accounts for 35 per cent of national GDP and almost half of the nation's industrial output.
Damage to economic infrastructure
Heavy damage was sustained in the energy, transport, and communications sectors. In electricity, an estimated 3 400 distribution towers and 490 km of overhead lines were damaged or destroyed, and there was extensive damage to underground cable lines. Oil and gas production facilities suffered extensive damage, though the fiscal cost has been held down by insurance coverage of fire damage to the Tupras oil refinery. Modest oil and gas pipeline damage was sustained to municipal distribution systems, and there were clean-up costs due to oil and chemicals discharged into the Sea of Marmara. Telecommunications damage included ruptured transmission lines, station damages, buildings and network facilities. Office buildings, water pipes and supplies, wastewater treatment, sewerage systems and other structures accounted for additional damage to municipal infrastructure.
Damage to the transport infrastructure included 60 km of the Ankara-Istanbul highway, the Gebze-Izmit-Arifiye railroad, the railcar factory in Adapazari and rolling stock, Derince Harbour, local streets and provincial highways. Traffic on the railway and motorway connections between Istanbul and Ankara was restored quite quickly. Damage to the industrial facilities and port or jetty structures located along the northern shores of the Gulf of lzmit was concentrated and varied from small displacements to settlement and total collapse. With immediate intervention, a large part of these infrastructure facilities were soon serviceable again. However, SPO estimates that at least $500 million will be needed both to repair damage and to meet the infrastructure requirements of new construction, without counting the cost of relocating cities.
Effects on the enterprise sector
Private and public sector estimates of the damage to the enterprise sector as a whole range from $1.1 to $4.5 billion (Table 35 below). The agricultural sector suffered little damage (SPO estimates $25 million in financing needs in the sector). In manufacturing, damage to large enterprises was lighter than in smaller enterprises, though by no means trivial (SPO estimates an $880 million total loss just for the 19 affected state-owned enterprises in the region). Human capital losses sustained by industry have been more serious, but harder to estimate. Besides immediate disruptions to labour supply due to deaths, injuries, and trauma, SMEs and large enterprises in the region are concerned about possible out-migration of qualified employees. Consequently, many of the larger enterprises are participating in the provision of shelter, care, and housing for their employees, which appears to have persuaded many to stay in the area. The tourism industry (based in Yalova) has been virtually destroyed and tourists may n ot return for many years, so that a fundamental restructuring will be needed.
Microenterprises (retail shops, artisan workshops, and services employing up to ten people) suffered the most, losing most of their working capital and premises (often situated on the ground floors of collapsed buildings, see Box 9), and key family workers. They accounted for the major part of the more than 15 000 destroyed and nearly 31 000 damaged business premises (Figure 19). Smaller firms were also hurt by close economic linkages with larger firms.
The government has attempted to help businesses in two ways. First, it has announced the deferral of all tax payments for individuals and businesses living in the earthquake area. A substantial part of these deferred payments may never be recovered, due to the large-scale loss of tax records. Second, a debt rescheduling and new subsidised credit programme via the state banks has been introduced to support particularly hard-hit businesses and individuals (see below). Nevertheless, the World Bank estimates that 50 per cent of self- and SME employment in the region will be lost permanently.
Burden on the financial sector
Insurance coverage among small enterprises and in the housing sector is very limited. Payments of claims are estimated to have amounted to $750 million. The bulk (95 per cent) of these losses have been covered by international reinsurers, entailing an upward adjustment in reinsurance premia though probably not affecting the availability of coverage (in part reflecting over-capacity in the international reinsurance industry). Given the dependence on foreign reinsurers, claims processing has been subject to some months of delay, resulting in liquidity constraints for insured businesses facing major repair and replacement costs. It does not appear that banks were willing to provide bridge financing.
In the banking sector, the emergence of bad loans due to uninsured earthquake losses constitutes a concern. The direct exposure (cash loans outstanding) of banks in the region has been estimated to be about $733 million, of which $119 million is held by public banks and the remainder by private banks.  The World Bank estimates that roughly one-third of such outstanding loans could be directly affected by the earthquake, which could in turn lead to defaults and affect capital adequacy of the system -- itself coming under closer scrutiny in the context of banking sector reform.
According to a government decree of 28 August 1999, the outstanding debts to Ziraat, Halk and Emlak banks owed by individuals or firms (in the seven cities of the broader region) who have sustained serious damage from the earthquake can be deferred for three years, with grace period of one year for both principal and interest, the latter being set generally at half the current interest rate. New subsidised loans would also be made available to the same applicants, including working capital loans for up to one year, and investment loans up to five years with a grace period of one year for both principal and interest.  However, applications for both reschedulings and new loans under the scheme amounted to merely $56 million and $42 million, respectively, i.e. well below the amount of known damage. Along with the decline of market interest rates to below the subsidised lending rates in early 2000, this means that the implied addition to state banks' duty losses will have been minor, but also that the financi al help to stricken enterprises will be limited. This could reflect some problems with the design of the subsidised credit scheme. Only applicants who filed for the scheme within three months after the earthquake were considered. Also, it is not clear that the system of independent verification of damage by provincial commissions was rigorous and transparent enough. Scarce public resources might not have been directed to those most in need, particularly small and micro businesses who have fewer personal connections and access to public banks. Indeed, small businesses in the region complain that they are not able to obtain credit to proceed with restructuring. For these firms, access to credit is at least as big a problem as its cost (Chapter III).
Aggregate economic costs of the disaster
Several assessments of the earthquake's aggregate macroeconomic impact have been made, notably by the World Bank, the State Planning Organisation, and the Turkish Businessmen's Association (TUSIAD), shown in Table 35. They are based on different methodologies, which makes comparisons difficult but nonetheless allows for a range of plausible quantification. Also, for the most part the estimates were made in the month following the earthquake, when information was still sparse while the effects of the devastating second shock to the Bolu region on 12 November were not factored in. Figure 19 suggests that additional costs of perhaps 50 per cent might be implied by the Bolu shock. Subject to these caveats, the results of these studies suggest the following possible magnitudes of macroeconomic impact:
- Wealth and income losses range from $5 to $14 billion. Destruction of physical capital accounts for the greater part, $3 to over $10 billion, of which housing and enterprise sectors each account for roughly 40 to 50 per cent, and infrastructure the remainder. The greatest uncertainty in these estimates (i.e., the widest ranges) appears to lie in the extent of damage to the enterprise sector (particularly to small and micro businesses). The associated income losses range from 1/2 to 3 per cent of GDP, affecting mainly 1999 and including not only the loss of output due to supply and demand disruptions, but also the cost of emergency relief -- a "dead-weight cost" which does not replace damaged structures but diverts resources from other uses all the same.
- Job losses could range from 20 to as much as 50 per cent of the pre-earthquake labour force in the affected region, due to both damage to business premises (demand side) and loss of life and health and out-migration (supply side). Most of these job losses are concentrated in self-employed and small business jobs, of which in turn a large proportion is expected to be semi-permanent.
- Growth in 1999 may have been up to 1 percentage point lower than the SPO baseline (Figure 20), reflecting disruptions to both supply (loss of physical capital and labour force) and demand (loss of inventories, temporarily depressed consumption and investment activity, interrupted input-output linkages across firms) -- albeit offset to some extent by the mobilisation of spare capacity in the rest of the country. Initially, these impacts were expected to result in a 2 to 2 1/2 per cent point drop in final GDP. However, the official estimates for 1999 indicate a drop of 5 per cent, suggesting that either the estimates of the earthquake impacts were too optimistic, or that the recession deepened by more than had been initially anticipated (i.e., worsening the baseline). In 2000, by contrast, growth may be up to 1 1/2 points above baseline due to the demand effects of the reconstruction effort which gets underway in that year.
- Inflationary effects will be small or negligible, as substantial excess capacity in the wider economy, together with imports, should diffuse any excess demand pressure arising from the reconstruction. Thus, the government does not consider that the reconstruction poses any risks to the disinflation programme. Even so, price pressure in housing and pockets of the construction materials industry cannot be excluded, as existing spare capacity may not well match the areas of excess demand.
Social costs of the disaster: shelter, social infrastructure and rebuilding needs
Mitigating the displacement of the population by temporary shelter
With around 330 000 housing units destroyed or damaged by the earthquake, about 600 000 people were forced to find emergency shelter. They had three options: i) shelter with friends or relatives, ii) move to an undamaged second home or rent accommodation, and iii) tent shelter in organised camps (more likely for people whose homes were demolished), or in the neighbourhood of the damaged but usually still-standing house.  The latter option could also involve the use of vacant public buildings. Only people in category iii) were considered to be "homeless": of these, by 1 December 1999, about 200 000 people were registered in 121 tent communities, while another 80 000 to perhaps 200 000 were estimated (by various sources) to be living in individual tents or in public buildings. By implication, at least 200 000 people opted for shelter with friends or alternative second-home or rented accommodation; many of these moved out of the region. Those who chose not to live in a tent town received a "rent allowance" o f about $175 per month, on condition that their house was mid- to heavily damaged. However, the cash value of assistance received in camps was far higher -- including free health care, food, cooking, clothing, counselling, skills training, schooling, and pocket money -- which provided a strong incentive to go there rather than accept the rent allowance, especially for those from the lower economic strata. It could be argued that the large gap between the value of the benefit-in-kind option and the alternative cash benefit, in essence, restricted peoples' choices, including the decision to out-migrate, while making them dependent on the state. 
With the cold and damp winter weather approaching, the government decided early on that tents had to be substituted by sturdier temporary accommodations. The search for a solution generated much internal debate. In the end, it was decided to build 30 772 prefabricated homes by public means, while another 10 696 were to be funded and built by the private sector. With an average household size of four, these prefabs were able to house around 165 000 people, so that the units had to be rationed. The cost of a 2-unit prefab (each 30 [m.sup.2] unit fully equipped) was $3 300,  rising to around $5 000 inclusive of infrastructure costs. Construction of the units was completed by end-November 1999, and by end-year the shift of people to the prefabs was complete. Tent dwellers were given the choice of moving into such a house or accepting the rent allowance in order to find their own arrangements, either in the same or in another city. As of April 2000, around 65 000 people were still living in 54 tent communities , while a total of 300 000 to 400 000 people (100 000 families) were receiving the rent allowance.
Besides assistance associated with shelter, government sought to help people in other ways, notably by extending special social insurance benefits for earthquake victims, insured and uninsured alike. An "earthquake amendment" to the new pension law has reduced the period of mandatory contributions in order for insured members and their dependants to be eligible for death and disability benefit, to just one year (with Treasury covering the shortfall in minimum contributions from the normal four years). This affects mainly monthly payments to dependants of deceased breadwinners as well as disability payments to insured members.  The law also provided for the following lump sum payments to be paid regardless of length of service or wage earned: $1 500 for each lost direct relative, and $1 000 for each second degree relative. Financial help was further given in the form of a deferral of all taxes until 2000. Finally, the government has provided extra community services, such as orphanages, child-care faciliti es, elderly accommodations, and training centres.
Longer-term housing replacement obligations
The latest estimates of the Ministry of Public Works (MPW) indicate that nearly 97 thousand houses were completely destroyed and over 231 thousand damaged, of which 107 thousand moderately and 124 thousand lightly (Table 36). According to the provisions of the (former) Disaster Law, the government is responsible for replacement of destroyed stock and rehabilitation of lightly to moderately damaged stock, though only owner-occupied primary residences are eligible for this guarantee. The World Bank estimates that 55 to 75 per cent of the affected units might satisfy the government criteria.  The public cost of replacing a destroyed house is approximately $20 000 (not including land acquisition costs), covering a modest (80 [m.sup.2]) apartment construction of standard design and location. The average cost of repairing a moderately damaged house is estimated to be $8 000, and a lightly damaged one, $3 000. The MPW is planning to build some 40 thousand new units, with construction due to have started in July 2000 and to be completed by end-year.  The preparatory work has involved the rapid mobilisation of engineers and other experts from all around the country to complete the work of damage assessment, determination of holders of property rights, geological and geotechnical surveys to ensure the proper siting of permanent housing settlements, project preparation, and infrastructure construction and repair. The process of tendering the construction contracts has been broadened to allow for bids by foreign construction firms, as well as speeded up, compared with past practices. But despite the scale and pace of reconstruction, there will remain a further 10 to 30 thousand units potentially eligible for full restitution and many more still in need of repair. Ultimately, the government may face $1.7 to $2.4 billion in housing reconstruction costs (Table 36). 
In lieu of direct provision, the government is also offering a system of cash benefits towards building or repairing one's own home, which could help to hold down fiscal costs and further speed up the work of reconstruction and rehabilitation. Those who lost their house and are willing to leave the region or to build their own home in the region are given a 20-year no-interest loan of about $10 500. Around 18 thousand people are estimated to have taken this set of options. Owners of damaged homes receive a loan of around $1 000 toward repair of a lightly damaged house and $3 500 for a mid-damaged house, with eligibility for the latter subject to certification by an engineering company. About 88 thousand households have received the former, and 50 thousand the latter benefit. While allowing for more freedom of choice and flexibility -- hence more rational redevelopment patterns -- the cash benefits are (once again) only at best around half as generous as the corresponding direct government provision. The absol ute gap is high in the case of a new home, where the take-up rate may remain low.
Repairing damage to social infrastructure
As regards social infrastructure, damage to schools was extensive: 43 school buildings were demolished and 377 endured severe damage. Until these are rehabilitated, around 25 000 schoolchildren will need to be transported to different school facilities, which requires extra payments for transport, uniforms, books, schoolteachers, and food. The quality of education is also likely to suffer, as classroom size in the schools receiving the overflow will double during the interim period, aggravating capacity bottlenecks already present in this sector (Chapter III). On the other hand, quick mobilisation of extra resources meant that little school time was lost after the earthquake. In the health area, 28 health centres and 10 hospitals were severely damaged, depriving the area of health infrastructure just when it was needed the most. This required deploying temporary prefab health care units and replacing damaged medical equipment, in addition to rehabilitating damaged capacity.
Impact on tile Budget
The above earthquake costs will be borne to a large extent by the government budget, namely:
- extra consumption/transfer spending for the relief effort and extra social security spending due to extraordinary death and disability benefits, as well as credit subsidies and tax deferrals/losses for affected businesses and individuals;
- new investment spending for the construction of interim prefabricated homes and the progressive reconstruction and repair of permanent housing and associated infrastructure;
- repairs to damaged transport and communications infrastructure, schools and hospitals.
Table 37 provides the government estimates of these various direct fiscal costs, which are seen to total about I per cent of GNP in 1999 and 2 per cent in 2000, $5.9 billion in all. It is seen that expenditures for replacement housing (inclusive of prefabs), at $1.4 billion, accounts for less than one quarter of the total, although as noted, there is an upside risk to the final cost of the government housing guarantee.
Official foreign funding will meet much of the budgetary costs. Total commitments to date stand at $3.8 billion (Table 38): $2.6 billion is in the form of project finance, to be disbursed mainly in 2000 and beyond as the reconstruction proceeds, while $1.1 billion is in the form of budget support, $107 million being already disbursed in 1999.A large part of the project money ($1.0 billion) is being administered through the World Bank-financed Project Implementation Unit (PIU).  After official foreign funding, there remains a financing gap of $2.1 billion. To bridge this gap, on 26 November 1999 the government announced an "earthquake package" of one-off tax measures (see Box 3, Chapter II). The total expected revenues from this package are $189 million in 1999 and $1.5 billion in 2000, sufficient to offset a large part of the gap.
By nature the above costs are temporary, so that they do not really impinge on questions of fiscal sustainability; but they come at a critical time for the public accounts. They have to be assimilated against the background of a severe underlying deterioration in 1999, and the subsequent adoption of a very tight stabilisation programme. The overall fiscal effort, needed to offset both baseline budget deterioration and earthquake costs while achieving programme targets, amounts to some 7V2 percentage points of GDP in 2000 alone -- a speed of adjustment almost unheard of in the OECD context. The stabilisation strategy is buttressed by an impressive number of structural reforms, and the disaster has made these even more difficult yet also more urgent, in part because of the need for re-evaluating the priorities in the use of national resources.
Overall, the economic and financial repercussions of the disaster have been managed in such a way as to preserve and even enhance confidence in the programme. It was initially feared that indirect budgetary impacts might arise as a result of a higher risk premium on government debt due to the worsened budgetary position and hence a higher debt service burden.  Around $1 billion of private capital flowed out of the country in the week after the earthquake. However, capital flows quickly stabilised with the passage of a social security reform bill on 25 August 1999, accompanied by the government's announcement that it would pursue other structural reforms without delay and that it would not finance earthquake costs by new domestic borrowing. This was followed in short order by a series of important legislative initiatives that served as prior actions for agreement with the IMF. The government's refusal to allow the tragedy to delay reforms or to soften budget targets impressed financial markets and contribu ted strongly to the turnaround of sentiment which culminated in the sharp fall in interest rates at the turn of the year.
Regulatory issues: towards more effective governance
Uncontrolled urban development
The problems of poor construction and siting described at the beginning of the chapter need to be understood in the context of the extremely rapid process of urbanisation which Turkey has been undergoing, in which the population has been moving en masse from the countryside to the cities (Figure 21). The process of urbanisation itself has been part of a momentum to industrialisation which has relied extensively on the informal economy for its underlying dynamic and in doing so may have encouraged a trade-off between economic expansion and orderly development. In practice, despite a "legally governed" real-estate core catering to the needs of the formal business sector and the traditional urban population, the bulk of land development in Turkey seems to have occurred through two informal mechanisms:
- occupation and expropriation of government land by new city-dwellers; and
- de facto development of agricultural lands (outside official settlement plans) around cities.
Regular amnesties created moral hazard and thus expanding illegal settlement of the urban peripheries. The proportion of Turkey's urban population living in such settlements is currently close to 40 per cent and since the late 1970s its growth has been explosive (Figure 22).
In effect, Turkey has coped with urban growth through migration by tolerating the illegal construction of housing, often on publicly-owned land,  thereby encouraging the construction sector to supply housing at lower cost by eliminating the need to purchase and improve land. Almost by definition, such "gecekondu" (overnight houses) settlements have been built without respect for planning rules and in contravention of the building codes. They were not exclusively for the poor, as is often the case in some developing countries; even middle class people have had to compromise on the quality of construction, infrastructure and public services. High inflation also played a role, since it meant that real estate was the only viable long term savings vehicle. This fuelled excess demand for housing and urban land speculation which made the better urban sites too expensive for low and moderate income groups to purchase. The regularisation of illegal settlements was initiated by their residents, who used the process to secure investment in infrastructure and public services as municipalities incorporated their district, and who acquired certification of conformity for their buildings through amnesty even if these still did not meet the standards of the codes. All involved, local and national officials, builders, and property owners, found it convenient to ignore legal measures for town planning and construction.  The benefits of this process have been the gradual extension of municipal jurisdiction over illegal settlements, and the avoidance of major social tensions during a sustained period of rapid urban growth through migration, while helping to avoid totally informal, illegal, unequipped and unserviced settlements around Turkish cities that may be seen in favela-type communities in some developing countries. It has been a major factor minimising housing and living costs in the country, and consequently fostering industry's competitiveness. In the process, there has been implicit agreement among all concerned tha t economic development was the highest priority. Licences have been granted for plots that carry a high risk, with tacit collusion between government, builders and homeowners to bypass the (weak) zoning! construction ordinances. 
The earthquakes have made clear the unsustainability of this framework. In view of Turkey's inherent seismicity and the high costs of the earthquakes, a reckoning of longer run costs and benefits would undoubtedly argue in favour of a model which accords significantly greater weight to public safety and rule of law in economic development. The remainder of this section looks at the regulatory framework behind Turkey's earthquake vulnerability, in the context of existing legislation and its enforcement, which involves accountability and performance at all levels of government. The succeeding section then presents a discussion of the policies and institutions needed to govern the development of a property market in which individuals gradually assume responsibility for finance and insurance.
The need for regulatory reform
Urban planning and land use management, construction quality, and disaster management are all subject to legislation and regulation. Four major laws govern the underlying structure of public responsibilities, legal liability and professional standards: the "Disaster Law" in respect of disaster response; the "Planning Law" in the area of construction and land use control; the 'Tender Law" in public construction project contracting, and the "Municipality Law" in urban planning implementation (see Annex IV). The current legal framework has not succeeded in preventing unsafe development, however. The relevant laws and regulations have suffered from both inadequate implementation and noncompliance due to the perverse public and private incentives that they often imply. The lack of clarity concerning responsibilities and principles of co-ordination has also hampered effective co -operation within and across levels of government, as well as between civil society and government. Also, as evidenced above, there is a p revailing culture of an official tolerance of illegal or informal activities in the construction and land use areas. Achieving better safety standards and ensuring that the principles of orderly urban development are more closely followed will entail both an overhaul of legislation, to ensure modern best practice, and better law enforcement. The need for reform is now generally recognised, and a review of regulatory and supervisory practices is under way.
Regulatory reform in the disaster planning and response area
The Disaster Law establishes arrangements for post-disaster management which are highly centralised but poorly co-ordinated both within and across levels of government. In the event of a major natural disaster, fifteen central government ministries - each having its own disaster management unit and plan authorised by the Ministry of Public Works (MPW) - establish a crisis centre in the Prime Minister's office, which mobilises resources to the affected area.  The MPW is in charge of overall emergency planning, assistance to victims, temporary shelter and reconstruction (Annex IV). Since disasters, by definition, are of such a magnitude as to overwhelm local and even regional resources, there is a strong argument in favour of such national responsibility. However, due to the concentration of all disaster planning competence and authority within the MPW, most cities in high-risk earthquake zones of Turkey do not have an earthquake response master plan, and there have been no awareness campaigns to inform cit izens of risks and appropriate actions to take in an emergency. International best practice indicates the importance of municipalities and provinces being able to develop their own capabilities in disaster management, the lessons of recent global disasters highlighting the contribution of civil society in an emergency, the value of local knowledge, and the benefits of international co-operation.
The Marmara earthquake made clear the inadequacies in the procedures for disaster planning and response, and reforms to address these followed quickly. On 22 November 1999, the government issued a decree creating a new agency, the Emergency Management Agency of Turkey (EMAT), to be housed in the Prime Ministry and charged with comprehensive emergency management. This involves, first, co-ordinating and rationalising hitherto dispersed emergency response functions and responsibilities. To this end, it will be equipped with modern communications, database, and earthquake monitoring facilities, benefit from a comprehensive training and exercise programme (search and rescue activities, fire-fighting, etc.), and engage the public via information and awareness campaigns. Secondly, it will be in charge of developing a national earthquake mitigation plan in co-operation and co-ordination with five pilot municipalities, the results of which will later be applied to other cities in high seismic risk areas.
A related defect of the Disaster Law concerns its rather sparse reference to pre-disaster mitigation, with a resulting lack of coherence between development and emergency planning. This problem is partially addressed with the creation of EMAT. Most seriously, however, the law has dulled risk awareness in the population, while creating a huge contingent fiscal liability, through its broad guarantee of replacement housing. In the immediate aftermath of the Marmara earthquake, decrees were issued to roll back the guarantee and so limit the government's financial exposure. On 1 September 1999, two groups of building owners were excluded from the guarantee: those whose buildings were constructed on others' land without a construction permit, and those who (or their spouses) owned another undamaged building. In December 1999, a decree was issued to remove the guarantee altogether for urban areas in the event of future disasters, being replaced by a new compulsory national earthquake insurance scheme (discussed furt her below). The decrees on the introduction of compulsory national insurance and of EMAT together form the core of a future revised Disaster Law.
Regulatory reform in the construction and land use area
The Planning Law has suffered from a number of defects which play a key role in undermining the principles of risk mitigation and development control - even though these principles figure among the law's main objectives (Annex IV). As in the case of the Disaster Law, it gives the various central ministries authority for planning and land development, but without a clear delineation of responsibilities and virtually no sanctions for failure to carry these out. In conjunction with a weak institutional backdrop for effective local authority implementation of planning law (discussed below), there is a severe failure to co-ordinate and almost no incentive to implement the law. The attendant regulations and practices have, as a consequence, failed to impose technical and professional standards in the building and planning sector, to keep building codes up to date and prevent collusion between builders and inspectors, or to govern appropriate land use. Most notably:
- There has been little or no criminal and tort liability, and no supervisory agency responsible for professional standards and liability for structural engineers, architects, contractors, building inspectors and city planners. Professional training for those in the construction sector, building contractors, sub-contractors, foremen and apprentices, has been left unsupervised.
- Building codes have only been periodically updated, but at a pace that has not kept up with international earthquake technology. The majority of the structures subjected to the earthquake were built under the old codes (established in 1975). But even for those built since 1996 (when the codes were revised), the main problem has remained the implementation, monitoring, and control of these codes at the local level. Furthermore, whereas legal standards have been issued for seismic design of principal structural systems, there is no comprehensive, unified building code which would serve as a clear reference for both building engineers and regulatory authorities.
- Perverse incentives have arisen insofar as inspecting engineers are traditionally chosen and paid by building contractors or owners. This practice stems from insufficient technical manpower at the municipal level. Low pay and lack of performance management, in addition to the lack of legal sanctions, have further increased the susceptibility of municipal inspectors to illegal payments. The process of tendering government contracts has been so loose as to allow similar abuses in the domain of public construction projects.
- Zoning ordinances are unclear and usually flaunted. Urban, regional, and national plans often intersect due to the absence of an established hierarchy among plans, a clear delineation of authority, or a clear mechanism for resolving conflicts. For example, an area designated for parks or greenery in the plans of the local authority can be forced to host the building of a commercial structure due to the intervention of the MPW (or other ministries, such as Tourism or Trade). Added to this is the problem of the informal gecekondu settlements on the outskirts of cities, discussed above.
The Planning Law is now under re-evaluation and some of the related regulations have already been changed. An important innovation has been the setting up (by 10 July 2000) of "construction controlling firms" in cities and towns. These firms, composed of groups of engineers having majority ownership, are to take over design-review and construction-site responsibilities from the municipal building departments. A process for the certification of such supervisory firms is being established. The introduction of professional standards and liability for these firms is expected to enhance the rigour of plan-review and inspection procedures significantly. The construction standards to be applied will parallel those of the EU, and those of the Turkish Standards Institute on the use of materials (for example, ready cement must be used in lieu of "own-mixed" cement, elastic instead of rigid steel must be used in frame-work, and steel must be used in the base of industrial buildings, schools and hospitals). The MPW estim ates that application of the standards will add 4 to 8 per cent to the cost of new buildings. There will also for the first time be specific municipal and professional responsibilities and penalties for failure to carry out those responsibilities.
It is essential that these steps to improve safeguards and incentives in the building sector are fully enforced, i.e. that the practices of the past do not simply reassert themselves within the new framework. In this sense, it will be necessary to fully enforce the new sanctions on the inspecting professions; have full judicial back up of legal liability; put more resources into the training of construction and design engineers; move forward with government procurement and municipal reforms (below), and monitor results. However, beyond this, more work is needed in order to establish a really coherent legal framework for satisfactory land and urban development. Four factors seem essential:
- More rigorous standards must be established.  But also, it will be essential that plan standards for closed and open areas, social and recreational facilities, etc. be improved. These will need to take into account planning goals which promote certain environmental, economic, and social outcomes. Direct citizen involvement in development plans, as exists in most other OECD countries, is an important precondition, and will depend to a large extent on progress in municipal reforms (below).
- The problem of lack of co-ordination and of clear delineation of responsibilities of the bodies responsible for different aspects of planning need to be resolved by a careful redrafting of the Planning Law. In parallel, a hierarchy of plans needs to be established nation-wide, including feedback mechanisms to handle spill-over effects.
- Better information about land ownership and a clearer distribution mechanism for land are needed for proper zoning and more orderly land and housing markets. Information about land quality is also needed for better risk assessment in the public and private sectors. Therefore, in addition to a planned World Bank project to improve the land registration and cadastre system in the Marmara region, a national geological survey along with wider public access to detailed geophysical information is advisable. The latter need could be met by establishing a nation-wide computerised Geographical Information System (GIS).
- The problem of the gecekondu needs to be resolved. This will certainly require eschewing the repeated use of amnesties. As the problem is strategically related to the economic development process, however, solutions will further need to be part of a global approach to bringing the informal economy within the ambit of state regulations. But even if the inflow into such settlements can be arrested, dealing with the large inherited stock of vulnerable structures is likely to require budgetary resources and effective management.
Reform of government procurement rules
The Tender Law has not succeeded in preventing abuses in public procurement. The large amount of money awarded for public construction projects - with little competition or control - has greatly exacerbated unethical practices in the construction sector, not to mention being a considerable waste of taxpayers' money. A major loophole in the law was that state economic enterprises were exempt, being subject to their own (unclear) tender procedures. Another problem was the inefficient structure of the bidding process. Once a tender was announced, only ten days were allowed to enter bids; bidding was based on unrealistic price estimates of the tendering authority; feasibility studies were not required; multiple "commissions" obfuscated procedures; "deferred" tenders (build as money becomes available) were used extensively (Annex IV). These methods favoured insiders, and led to the awarding of contracts where no budgetary appropriations or realistic cost estimates existed. This led in turn to massive misuse of res ources, cost overruns, inefficient projects running for years, and indeed unfinished projects. These problems were compounded by the failure to assign responsibility for construction control. The final result was that public buildings such as schools and hospitals had many of the building defects of private houses, and suffered from major devastation in the earthquake.
Since the earthquake, the government has completed a draft new Tender Law, in accordance with EU and WTO norms. The main objectives of the proposed changes are i) to implement tenders in a more transparent and competitive way by covering the whole of the public sector, including state enterprises, under a single law; ii) to increase the speed and efficiency of project implementation by more realistic tender procedures and pricing mechanisms including: a single "tender commission" and streamlined bureaucratic procedures, a requirement for minimum percentage budget appropriation to open a tender, a 52-day period to enter bids after tender is open, the requirement for feasibility studies, competitive bidding based on the delivery cost of the project, and the abolition of "deferred" tenders; and iii) to allocate responsibility for control over the construction process. These reforms would appear to address all of the above serious defects in the old law. Moreover, public construction projects will be subject to t he new surveillance mechanism that has been put into place for construction projects in general.
Local authority reform
The local level has, in principle, substantial autonomy over land-use decisions, but within an overall framework of rules and procedures set at higher levels. However, local government capabilities have proved insufficient to respond flexibly to local needs. It may be argued that one of the governance problems affecting disaster outcomes is that functional responsibilities have been devolved to the local level without extending the necessary decision-making powers or accountability. Despite the trend toward more decentralisation embodied in the Municipality Law and its various adaptations and updates, the central government retains considerable powers of control (Annex IV). Another problem is a confusion in the lines of authority due to inconsistencies across the different laws. For example, under the Municipality Law, municipalities are responsible for implementing development control and are under the administrative responsibility of the Ministry of Interior; however the latter does not have the appropriate capabilities to monitor functions related to land use or building safety, which by the provisions of the Planning Law reside with the MPW. Local capabilities also suffer from a lack of internal structural reforms: patronage counts in hiring, while low wages and insufficient incentives for performance reduce efficiency and productivity.
It follows that any empowerment of the municipalities through greater devolution of decision-making would have to be accompanied by greater accountability and citizen involvement to provide the necessary checks and balances. Indeed, when responsibility for development planning was devolved to the municipalities after 1984, it was done without the preconditions laid down in most OECD countries, namely, direct general-public control of development plans and the establishment of mechanisms to provide members of the public with the relevant technical information in the course of exercising this control. In light of these problems, the OECD has previously identified among the principal needs of local government in Turkey as follows: i) a more decentralised and less interventionist policy in the planning area with a more responsive attitude to divergent local circumstances; ii) greater efficiency, via the development of human and financial resource management aimed at transparency, accountability and control; iii) co-operation and joint ventures to deal with the large-scale needs of major cities; and iv) the participation of local citizens in government and the development of an urban consciousness. 
The government's 5-year Development Plan -- the main instrument for co-ordinating government policies including the local level -- echoes similar concerns as it emphasises the need to address the absence of decision-making powers in the decentralisation process, the insufficient resources allocated to local administrations, and the lack of consultative mechanisms at the local level (citizen-participation). The process of integration with Europe has given further momentum to the need to adopt new ideas concerning local administrations, including the idea of citizens' rights to participate in the self-government process as a democratic right.  With this understanding and parallel to the objectives of the Development Plan, a local authority reform law was drawn up by the government and submitted to Parliament in June 2000. The latter, however, has not yet acted upon it and it would appear crucial that such legislation be passed. Indeed, in its absence, there is a risk that reforms to the Planning Law could f ail to be fully effective.
Strategic issues in urban planning and reconstruction
The current reconstruction will have a weighty influence on the future direction for urban development in the affected region, since many of its cities and towns are being virtually built anew. However, this process will profit very little from the above reforms, which will take some time to fully bear fruit, so that important strategic issues may need be addressed within the constraint of the present institutional framework. With respect to the current reconstruction of areas suffering large-scale destruction, including Izmit, Adapazari and Duzce, the responsibility of the government to replace housing gives the central government the initiative, but the decision-making about what is built, where and when would need to involve local levels for a strong component of informed consumer choice to be built into the reconstruction programme. Engaging local government and civil society in land use planning and facilitating measures to reduce risk and improve preparedness at the local level is, however, a complex ch allenge.
Optimising the reconstruction process
In the aftermath of the 1999 earthquakes, the immediate challenges have been to ensure coherent development/reconstruction of the affected areas and towns, some of which have had their economic viability put in question, while trying to ensure that the defects which magnified the losses in the recent earthquakes are not so catastrophic when the next earthquake hits. Developing a coherent strategy in this respect requires the reconciliation of two potentially conflicting considerations: the desire to move people into new, permanent houses as quickly as possible, thereby reducing the emotional as well as financial cost of interim solutions, versus the opportunity provided by the reconstruction process to introduce improvements in infrastructure, civic facilities, etc., while implementing new measures of code enforcement, insurance, etc. for rebuilt property and districts. Some cities will need to develop a whole new economic base (e.g. Yalova, which lost its entire tourist industry) and there needs to be a stra tegy to cope with this shift.
The emphasis so far has been on moving people into new, permanent housing and reducing government support for temporary housing as quickly as possible. Some observers have criticised expenditure on temporary housing as wasteful and costly. Indeed, there are reasonable concerns about fostering a dependency mentality, and about people trying to claim benefits on dubious grounds. But in the absence of better maps of the affected sites, of intensive consultations with local populations, and of an assessment of economic and environmental deficiencies in the cities that have been destroyed, the rush to move people into permanent housing may be restricting the options available. In particular, the opportunity to include broad measures of public participation into the strategic planning may be lost. The objective of reconstruction on sustainable lines would be compromised if decisions are made prematurely, on the basis of inadequate information and consultation. For example, because the government's plan to replace h ouses is limited to owner-occupied housing, landlords will get only one building rebuilt. In the immediate future, therefore, there will be a shortage of rental housing, and the rebuilding strategy does not appear to indicate where rental housing might be built. A shortage of rental housing might compromise mobility at a time when population movement must be expected (with consequences for employers who need to attract workers with certain skills); a failure to plan for rental housing may lead over time to the de facto separation of owner-occupied housing from rental properties with a resulting "ghetto-isation" of the latter and a lack of choice for the population.
More generally, the current approach to reconstruction is highly centralised and interventionist. This may be justified by the urgency of the situation. But consideration should be given to the introduction of market forces with the redevelopment programme. This could be based on tradeable vouchers for land auction and partnerships with the private sector for commercial development. It is important for the new situation to be used to progress towards one where central and local interests, state and private, are balanced. The experience of the city of Izmir is instructive in this context (Box 10). Interacting with municipal planners and scientific experts from other countries facing earthquake risks, it has developed a plan of action which brings together different organisations in society and integrates the critical elements required for sale urban development: tools for risk assessment, the principles of selective "retrofitting" (i.e. improving to be more earthquake resistant) of existing urban structures, a nd emergency planning and training procedures. The important lesson here is that the city was able, with international support, to obtain a high degree of co-ordination with other levels of government and civil society, despite the highly centralised legal structure of disaster management in Turkey.
Preparing for the next earthquake
Given the earthquake movement along the North-Anatolian fault, Istanbul is perhaps the city-region now most at risk. As noted above, seismologists expect a severe earthquake in the southern Istanbul region, bordering the Sea of Marmara at some time in the next quarter-century, with the strong possibility of such an occurrence within a decade. In Istanbul, there are known areas of high risk (where liquefaction of soil would occur). A question arises as to whether the people living in these areas should be resettled, or their housing retrofitted. Low-cost improvements may yield significant benefits in terms of resistance to destruction, and could be phased in while preserving existing districts. In the meantime, emergency planning needs to take account of the possibility that some parts of the city will suffer far greater destruction than others. Finally, because the city continues to grow (at a rate estimated at 300 000 newcomers per year), and in the absence of a coherent, enforceable strategy to plan for thi s growth, illegal settlements and sub-standard construction will continue to be the norm, thus increasing the magnitude of the problem facing those with the responsibility of emergency preparedness.
The elements required to meet any future earthquake emergency are numerous. However, good communications will be a priority. The experience of Kobe should encourage the creation of more areas of open space within the city, where people can assemble, as well as redundant infrastructure, networks and pathways for public services (i.e., excess capacity and multiple pathways), together with massive campaigns of public information and training. Since centralised public services can be easily crippled in time of a disaster, priority will need to be given to retrofitting schools and hospitals, the build-quality of which is generally poor. These could be used as focal points during emergencies.
Most importantly, while options for the future can still be explored in the reconstruction process, there is an urgent need for guidance about metropolitan regional development and appropriate policy instruments. A metropolitan regional review, perhaps in the context of a forum bringing in expert knowledge from OECD Member countries could complement efforts in Turkey to analyse regional conditions. Such a forum could provide an opportunity for the lessons of success and failure in recovery from disaster to be fashioned into best-practice procedures, and could help shape the agenda for the newly-formed national emergency management agency as well as the elaboration of city earthquake master plans.
Enlarging the scope for market-driven economic development
Involving the private sector: the mortgage market
The administrative deficiencies noted above have operated against the background of a perverse set of incentives for the private sector, which have acted to make both the population and the business sector short-sighted with respect to risks. A government guarantee to replace housing after an earthquake has been behind the lack of pressure for better mapping and risk analysis, and an underdeveloped insurance market. A system of private incentives to risk minimisation has thus failed to develop. At the same time, chronic inflation and associated high real interest rates have prevented the development of a mortgage market. Standard urban planning, micro-zoning according to geophysical information, building code development and enforcement will increase the cost of housing, creating a demand for both a mortgage and insurance market. Indeed, a longer-term consequence of the earthquakes will need to be the development of the two, so far minimally-developed markets for earthquake-resilient and infrastructure-equipp ed suburban habitable land and that for pricing the natural risk vulnerabilities of individual assets.
The creation of housing-credit and insurance markets should be assisted by a low inflationary environment, which would allow for a longer-term financial planning horizon and help Turkey to attract long-term foreign capital. Moreover, the need to create the conditions for more sustainable urban development comes at a time when the opportunity to make profits from buying government paper are diminishing, promoting a need for alternative domestic savings instruments. However, for mortgage and insurance markets to emerge, legal/institutional impediments to their development will have to be reviewed, giving an impetus to financial market reform. A new action plan, including the securitising of mortgage credits, would need to be developed in order to help form a sound system. Important reforms to the financial system are already taking place that should facilitate financial market diversification (Chapter III).
Reform of the banking sector as it adapts to a low-inflation regime could unleash the development of a mortgage market, which may be instrumental in providing financing for new housing construction. Such a market is already developing, albeit only in the upper segment of the property market. High inflation has thus far prevented the creation of TL-denominated instruments beyond one year, effectively ensuring that the housing industry operates on a pay-as-you-build principle. However, there are some US dollar-denominated transactions in the middle-upper market, the main market vehicle for property investment being Real Estate Investment Companies. These are regulated by the Capital Markets Board since 1995 and are tax-favoured (to allow them to compete with the informal economy which dominates the building industry). These companies seem set to expand, not least because earthquake risks have directed market interest quite clearly towards property compliant with building and zoning codes, which such companies c an guarantee. Indeed, the role played by mortgage companies in insisting on building standards and insurance will be a growing stimulus to more orderly development of the housing market as far as new building is concerned.
Managing risks: developing the private sector insurance market
While a natural corollary of a growing mortgage market would be the development of a private insurance market, this would be slow to impact on the current housing stock in Turkey, which is currently severely underinsured. Property owners are usually reluctant to incur the up-front costs of risk mitigation measures because they either misperceive risks, are myopic and/or face severe budget constraints. And judged by its past record and present financial capacity (capital adequacy), the technical abilities of the local insurance industry are limited for handling the earthquake risks. Moreover, when it comes to insuring catastrophic risks, market failures are well-known even in highly developed insurance markets, so that the government has a role to play, as indeed is the case in other OECD economies where the concurrent provision of public and private earthquake insurance is considered a necessity. 
Forming the core of a future revised Disaster Law (see above), a government-sponsored "Turkish Catastrophic Insurance Pool (TCIP)" is now being put in place, which will be subject to a coverage limit of $25 000 per house, and will permit additional private insurance coverage (Box 11). The new scheme is to be operational by 28 September 2000.  It will require the services of insurance companies as information and collection agents, and could help stimulate development of the sector in the longer run. However, for this to be really effective, the challenge is to reconcile low-cost insurance with the need to generate economic signals which allow individuals to respond to risks. Public provision alters incentives and creates a moral hazard. To enhance the role of insurance in encouraging property owners to take steps to reduce losses from natural hazards such as earthquakes, private incentives are needed that internalise risks. This may be the only real solution to enforcement failures noted above, since ince ntives to favour development over safety will change only slowly and uncoordinated regulatory responsibilities between local and national levels are not yet rationalised. Uncontrolled building on dangerous land may well continue, since it will take great political courage to call out the bulldozers against squatters who have erected buildings. The effectiveness of the proposed new supervisory companies, with professional standards and liability, is uncertain. The greatest boost to home safety in Turkey might thus well lie in private risk-based enforcement to standards in land use and construction. For this to be the case, private insurance should ideally assign risk-related premia to the risk exposures of different types and locations of buildings. The lack of premium differentiation is a drawback here, which should be corrected as cartographic and insurance sector deficiencies are remedied.
An active insurance industry would be instrumental in ensuring the enforcement of building codes, relieving the burden on government in this respect. Indeed, it is because reinsurers do not utilise information on risk to price their products so as to create an incentive to invest in cost-effective mitigation measures that the responsibility of the public sector for enforcing building codes is so high. There is also limited interest by engineers and builders in designing safer structures if it means incurring costs that will hurt them competitively. Interviews with structural engineers concerned with the performance of earthquake-resistant structures indicate that they have no incentive to build structures that exceed existing codes because they have to justify these expenses to their clients and would lose out to other engineers who did not include these features in the design. This would change with the development of insurance and mortgage markets.
Summing Up: implications for future disaster readiness
The human and material toll of the two earthquakes have been severe. International assistance, however, has reduced the financial burden to manageable proportions, and the economy is rebuilding and recovering. The disaster has not diverted policy attention from the need for sound macroeconomic policies and the structural reform process has been strengthened by it. On the other hand, many displaced persons remain in temporary shelters, many of the governance factors responsible for the more than 18 thousand deaths and 50 thousand injuries remain in place, and there are important questions with respect to Turkey's preparedness in the event of a probable future major earthquake. The 1999 earthquakes have revealed a relative slowness of response and a human vulnerability to shocks greatly in excess of that observed in other OECD member countries subject to earthquakes.
The World Bank has identified three critical elements that are needed for a comprehensive disaster management strategy: risk identification, risk-reduction, and risk-transfer.  This chapter has shown arrangements in Turkey to be deficient on all three levels.
- The process of risk identification has been inadequate. The delineation of hazard zones indicating both soil quality and exposure to risk (earthquake zone maps) is fairly rudimentary and geological surveys are out of date and incomplete. Awareness campaigns to inform citizens of risks and appropriate actions to take in an emergency have been absent.
- Risk reduction and mitigation have been impeded by a systemic failure to enforce building codes and implement appropriate land use and planning policies, even in relation to known risks. The laws governing development contain loopholes, while local level implementation and monitoring has been inadequate. Planning/construction standards do not correspond to potential hazard exposure, land use and building certification have been compromised, with a lack of a rigorous system of training, licensing and liability in the engineering, building and planning professions.
- Many of the deficiencies stem from the fundamental fact that personal incentive structures have been perverse. The obligation of the government under disaster law to rebuild damaged residences, and the lack of a mortgage market, operate as disincentives to the development of a market for risk pricing and transfer. Individual insurance is virtually absent, and even small businesses are under-insured. The failure to properly price and internalise risk has thus amplified the systematic failure to mitigate risks through appropriate behaviour.
Measures are now being put in place to correct many of the above deficiencies (Table 39). With respect to the planning legislation, these include the out-sourcing of design-review and construction-site responsibilities from the municipal building departments, and the introduction of a process for the certification to ensure construction standards parallel to those of the EU. There will be specific municipal and professional responsibilities and attached penalties. There are proposals for state tender reform and for local municipality reform to strengthen governance in the construction and development area. The severe disincentives to proper risk management that are embodied in the disaster legislation are likewise being addressed. The government has introduced a new mandatory national insurance scheme, while abolishing its former guarantees in the housing area. National risk will be transferred to world-wide risk-sharing pools. However, for the moment, the differentiation of risks will remain rather rudimenta ry and individual incentives to take precautions against earthquake will be inadequate. In this sense, a mortgage market, which would serve to encourage a property insurance market, would be needed. It has been prevented from developing inter alia by endemic high inflation, so that the transition to a low inflation regime, now set in train, should be beneficial for individual earthquake awareness. However, it could only impact on the current housing stock with a very long lag.
Whatever measures are taken to ensure better governance and to promote private insurance provision, disaster preparedness, especially in the threatened region of Istanbul, needs to rely on improving the earthquake resistance of the existing housing stock ("retrofitting"). Retrofitting will be enormously expensive, and could only be achieved very gradually, but it would seem inevitable that the government will have to undertake more spending beyond the reconstruction phase in order to avoid yet greater budgetary and human costs in the future. This might involve tax breaks to complement the new national insurance scheme, but the full cost for rehabilitation of low-income housing may have to be borne by the state, especially those in illegal settlements that are not covered by insurance. And in the event of a major earthquake before the new insurance fund has matured, the state will also have to bear a large portion of the insured claims. At the same time, schools, hospitals, and other public buildings in the zo nes at risk urgently need to be reinforced and earthquake-proofed, so as to become useable as shelters for the most vulnerable members of society. The demands on the budget will thus remain heavy - coming on top of those identified to meet the needs of human capital development, requiring redoubled structural reforms to eliminate wasteful spending, prioritising of expenditures and continuing efforts to strengthen tax administration.
The general policy implications of the recent earthquakes are that their human and material effects are a function of the governance and incentive structures that are in operation, and policy responses to future threats need to take this into account. In this sense, the structural reform responses to Turkey's earthquake vulnerability are part of the same process of modernisation and enhanced transparency which is required for economic modernisation. The progress already underway within the government in enhancing transparency and accountability in the public sector and creating a rules-based market-oriented framework for private-sector initiative should make for an environment which allows a more orderly, earthquake-resistant pattern of urban development than has been apparent in the past two decades.
(67.) See Wiss, Janney, Elstner Associates, Inc. (1999).
(68.) A Parliamentary Earthquake Investigation Commission report commented that the 110 people of the Civil Defence Institution had "virtually become lost amongst the 13 600 damaged buildings" after the Marmara quake and that insufficiency of personnel had resulted in a low success rate for rescuing persons from collapsed buildings.
(69.) There were critical failures in the national infrastructure: i) the main fibre optic cable governing telephone connections into the earthquake region was cut where it crossed the fault line just east of Ismit; ii) two main substations on the electric power grid were damaged, causing a widespread power blackout across Turkey; and iii) an overpass on the motorway between Ismit and Ankara collapsed. Compounding these failures was widespread cellular phone use by outsiders trying to get information from the region, which caused the telephone system to break down completely, and by people driving to the region, which clogged the roads. See World Bank (1999b).
(70.) The response to the second (November) earthquake was much better co-ordinated, as reflected in part in the far lower ratio of deaths to houses destroyed (.03 versus .26 in August). This was because of lessons learned from the mistakes made in first earthquake, but also because the geographical and seismic magnitude of the shock was not so large as to overwhelm the rescue capacities of the region and nation.
(71.) Many foreign companies have affiliates nearby in the region, including Goodyear, Pirelli, Honda, Hyundai, Toyota, Renault, FIAT, Ford, Bridgestone, Mannesmann, Lafarge and Bayer.
(72.) In addition, there are some large-scale enterprises which are established in Istanbul and use loans from Istanbul branches to make investments both in and out of the earthquake region. Including such indirect exposures, the World Bank initially estimated total banking exposure at about double the given estimate.
(73.) The new loan programme differs from regular subsidised lending operations (Ziraat to the agricultural sector, Halk to small businessmen and artisans, and Emlak for housing) mainly in terms of ease and breadth of access (e.g., large firms as well as SMEs could apply for Halk Bank credits).
(74.) In many cases the houses were habitable, but the residents were afraid to enter them for fear they would collapse. This reflected the very low level of credibility among the population of official damage assessments (see "Temporary Shelter Report" in World Bank, 1999c).
(75.) Surveys indicate that the tent camp population was relatively poor and poorly educated. Nearly half had owned their houses and 61 per cent had lived in an apartment building; 38 per cent were under treatment for illness, and 64 per cent had experienced psychosocial problems since the earthquake, rising to 79 per cent for children. They tended to develop focused strategies for assuring that they were eligible for as many benefit options as possible, but felt great uncertainty. They exhibited classical symptoms of dependency and victimisation. A significant number were unwilling to move to prefabs because they did not want to lose the fringe benefits associated with camp living. By contrast, those living next to friends and relatives were more likely to have financial resources and emotional support, and more likely to have a plan that the cash from the rental subsidy would support (ibid.).
(76.) This was below market value of $4 200, as the Ministry of Public Works (the sole purchaser) was in a position to set prices and standards for these units.
(77.) Reflecting the pattern for Turkey as a whole, about three-quarters of the affected population is either a member of a social security fund or the dependent of one.
(78.) According to the MPW, 70 per cent of the population having lost their shelters are homeowners. Of these, around 75 per cent will have legal beneficiary. This would tend to support the lower band estimate of the World Bank.
(79.) Original plans were for construction to start in March. However, the MPW later imposed a moratorium on new construction until July in order to allow more time for new inspections and control systems to be put into place (see below). Also, another 6 500 units are being built directly by the World Bank, but approval of the sites (by the MPW) has been subject to some delay. The World Bank is also repairing some 55 000 units.
(80.) These estimates do not include the cost of relocating entire cities, which could substantially raise the upper bound. For example, the town of Adapazari (capital of Sakarya province) has been planning to relocate all housing 20 km outside the city, due to poor soil quality there (which liquifies easily in earthquakes).
(81.) The PIU was set up in the Prime Minister's office in the context of an earlier loan project and has now been strengthened to co-ordinate all paperwork, tenders, and disbursements of aid in a transparent way. Appointments of the staff of the PIU must be approved by the Bank, and Bank procurement rules must be followed. Official disbursements via the PIU are moreover largely conditional upon Bank-specified policy reforms in the disaster preparedness area. These include the setting up of: i) a national co-ordinating disaster management agency while enhancing the capabilities of the local authorities in coping with a future disaster, and ii) a national mandatory earthquake insurance plan; with iii) appropriate modification of the Disaster Law, Urban Planning Law, and State Tender Law (see below).
(82.) See World Bank (1999b).
(83.) Public land ownership is quite high in Turkey (55 per cent of the territory). This reflects the legacy of Ottoman times when the government owned essentially all the land, together with the subsequent failure to distribute it in a methodical way, but rather by default through ex post regularisation of squatting.
(84.) Often, mayoral candidates promised such amnesties in return for the sizeable number of votes the residents of these squatter communities could deliver.
(85.) Anecdotal evidence suggests that in some cases building quality and legal observance have deteriorated in the formal sector as well (e.g., builders would add a number of stories to high rise buildings beyond those stipulated in the construction permits, and then secured ex post certification from inspectors). This could reflect both competitive pressure from the informal sector and the general attitude of regulatory laxity with which it is associated.
(86.) For smaller-size crises with more localised effects, the Ministry of Public Works is put in charge of crisis management, with local authorities administering its directives and "Provincial Committees" (purely administrative organs of the central government at the regional level) mobilising resources to the region.
(87.) The World Bank will be helping Turkey to set up a process for the development, review, and promulgation of a comprehensive unified building code, while clarifying the legal status of the codes, reference standards and public commentary.
(88.) See OECD (1996).
(89.) See State Planning Organisation (2000).
(90.) The New Zealand and California Earthquake Authorities, for example, have found ways to price residential earthquake insurance products, transferring large portions of the risk via reinsurance.
(91.) However, it did not become fully operational until the end of November 2000.
(92.) See World Bank (1999a).
Table 33. Major disasters in OECD countries with Implications for regional development Date Place Agent 1906 San Francisco Earthquake 1908 Messina, Italy Earthquake 1923 Kanto Plain Earthquake Japan (Tokyo) 1939 Erzincan, Turkey Earthquake 19-20 September 1985 Mexico City, Earthquake Mexico 10-17 September 1988 Gulf of Mexico Hurricane Gilbert 15 September 1989 Texas Hurricane Hugo 7 October 1989 California Earthquake (Loma Prieta) 28 December 1989 Newcastle NSW, Earthquake Australia 1990 Sicily Earthquake 25 January-26 February 1990 NW Europe Storms, Gales, Daria, Vivian 13 March 1992 Erzincan, Turkey Earthquake 23-26 Aug. 1992 Florida Hurricane Andrew July-August 1993 Midwest, US: Flooding Mississippi Valley 17 January 1994 Northridge, CA Earthquake 17 January 1995 Kobe, Japan Earthquake February 1995 NW Europe Flooding 7-10 October 1997 Mexico Hurricane Pauline 17 August 1999 Izmit, Turkey Earthquake 26 December 1999 France Wind storm Date Deaths Damage cost 1906 2 000 $6 billion in 1987 $500 million in 1906 1908 150 000 1923 140 000 + 1939 32 000 19-20 September 1985 7000 + $4 billion 10-17 September 1988 In US, $10 billion In Mexico, $880 million 15 September 1989 $5 billion 7 October 1989 61 $7 billion 28 December 1989 12 Australian $1 billion 1990 20 + $10 to 15 billion 25 January-26 February 1990 120 $10.5 billion 13 March 1992 540 $1.5 billion 23-26 Aug. 1992 34 $16 to 30 billion July-August 1993 50 $12 billion 17 January 1994 56 $30 billion 17 January 1995 6 000 + $100 billion (+) February 1995 40 + $3 billion 7-10 October 1997 230 + $100 million insured value 17 August 1999 18 000 + $5 to 14 billion 26 December 1999 80-90 $8 to 10 billion Date Additional comments 1906 1908 1923 1939 230 000 homeless 19-20 September 1985 40 000 Injured 30 000+ homeless. General hospital collapsed, burying 600 staff, patients, widespread ad-hoc, collective solidarity efforts. Reconstruction fostered outer-edge development at expense of city centre 10-17 September 1988 400 000 homeless in Mexico 15 September 1989 7 October 1989 Collapse of Cypress Freeway in Oakland 28 December 1989 Reconstruction to take 5 to 10 years; 10 000 houses damaged in city of 300 000 people 1990 25 January-26 February 1990 13 March 1992 180 000 homeless, 3 850 injured 23-26 Aug. 1992 Reconstruction changed socio-economic structure of many south Florida communities; greater segregation and polarisation by income, race, ethnicity July-August 1993 100 000 evacuated; 150 000 homeless; 40 000 business or homes damaged or lost; changes afterward in settlement pattern 17 January 1994 25 000 homeless; 8500 iniured thousands without water or power, Major disruption to highway system (detours cost $1 million/day in congestion, accidents) 17 January 1995 Massivedevastation to port city, trade disrupted. 50 000 injured, 300 000 homeless. City rebuilt with new infrastructure for economic activities, social integration, disaster preparedness February 1995 250 000 evacuated. Flooding raised questions about land-use changes and planning practices, which increased vulnerability and intensity. Worst storms in the Netherlands since 1953, after which the Dutch initiated a maior civil engineering project of protection. In 1953, 2 000 drowned, 300 000 displaced people 7-10 October 1997 50 000 Homeless. Worst affected were Acapulco, Guerrero, Oaxaca 17 August 1999 50 000 injured; 600 000 homeless. Death and damage include impact of second shock on 12 November 1999 in Bolu province 26 December 1999 Major damage to electricity distribution network, railroads and forestry industry Notes: All dollars ($) are US$ unless otherwise stated. Major emerging risks: * Urbanisation in coastal zones. * Increasing gap between value of losses and insurance cover. * Lack of capacity to plan for reconstruction within a strategic vision for regional development. Source: OECD. Table 34 Selected indicators for the earthquake region  Population Share Share Per capita in GDP in industrial income value added Thousands Percent US $ Kocaeli 1 177 4.8 11.3 7 845 Sakarya 732 1.1 1.1 2 734 Yalova 164 0.4 0.7 4 966 Bolu 553 0.9 0.7 3 104 Bursa 1 959 3.5 5.0 3 434 Eskisehir 661 1.2 1.1 3 335 Istanbul 9 199 22.8 26.8 4 728 Kocaeli + Sakarya + Yalova + Bolu 2 626 7.2 13.8 5 243 Total of 7 Cities 14 444 34.7 46.7 4 581 Turkey 62 866 100.0 100.0 3 031 Share Share Share in budget in bank in banking tax revenues deposits credits Per cent Kocaeli 15.8 1.4 0.9 Sakarya 0.4 0.5 0.2 Yalova 0.1 0.2 0.1 Bolu 0.3 0.3 0.2 Bursa 3.0 2.4 3.2 Eskisehir 0.8 0.7 0.7 Istanbul 37.5 44.1 41.0 Kocaeli + Sakarya + Yalova + Bolu 16.6 2.4 1.4 Total of 7 Cities 58.0 49.6 46.3 Turkey 100.0 100.0 100.0 (1.)1997 or 1998. Source: Turkish authorities. Table 35 Macroeconomic costs of the earthquake US$ billion TUSIAD  SPO  World Bank  Direct costs 10 6.6 to 10.6 3.1 to 6.5 Housing 4 3.5 to 5 1.1 to 3 Enterprises 4.5 2.5 to 4.5 1.1 to 2.6 Infrastructure 15 0.5 to 1 0.9 Indirect costs 2.8 2 to 2.5 1.8 to 2.6 Value-added loss 2 2 to 2.5 1.2 to 2 Emergency relief expenditures 0.8 .. 0.6 Total damage costs (rounded) 13 9 to 13 5 to 9 Secondary effects Current account losses 2 .. 3 Fiscal costs 2 5.9 3.6 to 4.6 Job losses (per cent of labour force in the region) .. .. 20 to 50% (1.)TUSIAD first estimated the value of the loss of national wealth by surveys of its members and in co-operation with SPO. It then estimated the associated loss of national income by assuming that economic activity in the region came to a halt for two to three months (with about $50 million lost each day), due not only to loss of physical capacity, but also employee absenteeism, lack of water and energy, supply shortages and transportation difficulties, which depressed overall output regionally as well as nationally. (2.)SPO estimated wealth losses on the basis of information given to the government from various sources (including a physical count of destroyed properties) and preliminary estimations based on certain assumptions. (3.)The World Bank used an enumerative technique to estimate physical damages (on-site inspections by Bank staff). The GNP impacts are estimated by: a) assuming that the percentage of value added lost due to disruption to industry and services in the four most severely affected regions is 50, 30, 15, and 8 per cent 1999 Q3 to 2000 Q2, respectively, b) further assuming that one-third of the disruptions in the first two quarters are offset by increased economic activity in other areas; c) multiplying the net disruption by the weight of the region (7.2 per cent) in national value added. Source: TUSIAD (Turkish Industrialisation and Businessmen's Association), "Economic Impact of the Turkish Earthquake", 1 September 1999; SPO (State Planning Organisation), "The Impact of the Turkish Earthquake on the Turkish Economy (A Brief Assessment)", 23 September 1999; World Bank (1999b); and OECD staff estimates. Table 36. Housing reconstruction cost  As of 17 August 2000 Fiscal Cost Units eligible Number Cost per unit Total damage Lower of units (US$) (US$ million) bound Collapsed 96 808 20 000 1 936 53 244 Medium 107 331 8 000 859 59 032 Light 124 033 3 000 372 68 218 Total 328 172 9 650 3 167 180 495 Fiscal Cost Units Total cost($ million) eligible Upper Lower Upper bound bound bound Collapsed 72 606 1 065 1 452 Medium 80 498 472 644 Light 93 025 204 279 Total 241 629 1 741 2 375 (1.)This table replicates the method utilised by the World Bank (applied to the most recent damage assessment reports), which assumes that 55 percent of the damaged housing is eligible for restitution under government criteria in the lower bound estimate, while 75 per cent is eligible in the upper boun bound estimate. The latter represents the actual proportion of primary homeowners and assumes that none of them takes the cash benefit in lieu of direct housing benefits. Source: World Bank (1999b); Turkish authorities; OECD. Table 37. Fiscal impact of the earthquake 1999 US$ million Per cent of GNP I. Consolidated budget 1 402 0.8 Revenue loss 739 0.4 Expenditure 663 0.4 Current 133 0.1 Transfers 306 0.2 Investment 225 0.1 II. Housing 122 0.1 Prefabricated houses 122 0.1 Permanent houses Payment in cash Aid for medium-damaged houses III. Duty losses of public banks  Ziraat Halk Emlak IV. Local governments  81 0.0 V. Funds 169 0.1 Social aid and solidarity fund  169 0.1 Mass housing fund VI. SOEs  Total public sector 1 774 1.0 2000 US$ million Per cent of GNP I. Consolidated budget 1 571 0.7 Revenue loss 314 0.1 Expenditure 1 257 0.6 Current 445 0.2 Transfers 463 0.2 Investment 349 0.2 II. Housing 1 242 0.6 Prefabricated houses Permanent houses 509 0.2 Payment in cash 283 0.1 Aid for medium-damaged houses 450 0.2 III. Duty losses of public banks  531 0.2 Ziraat 328 0.2 Halk 94 0.0 Emlak 109 0.1 IV. Local governments  65 0.0 V. Funds 371 0.2 Social aid and solidarity fund  261 0.1 Mass housing fund 110 0.0 VI. SOEs  452 0.2 Total public sector 4 248 1.9 (1.)The most recent estimates of fiscal costs of the earthquake credit subsidy scheme are considerably lower than the initia initial estimates shown in the present table. (2.)Includes expenditure for sewerage, water, mapwork and development plan, water and sewerage for temporary settlements, and equipment (latter three categories in 1999 only). (3.)Includes outlays for death aid, aid for disabled people, emergency aid, shelter aid, restore and shelter aid, and business aid. (4.)Refers to the damage recovery costs and includes all SOEs in the region. Moreover, production and sales losses at SOEs are estimated at $632 million. Source: Turkish authorities. Table 38. External financing for Marmara earthquake US$ million Type of assistance Project loans Donor Programme Managed Managed loans by PU by the (World Bank) implementing agency IMF 500 World Bank Import and budget finance 252.5 MEER project finance 505 Reallocation 62.5 173.8 International Finance Corporation 50 European Investment Bank  455 152 Council of Europe Development Bank  303 Italy  18 Spain  60 Belgium  4 Islamic Development Bank Import trade finance 150 Project finance 150 Gulf co-operation Council 400 Black Sea Trade and Development 10 Bank Japan Commodity loan 200 Project finance (SMEs) 250 South Korea  30 EU  Germany Total 1 102.5 1 022.5 1 600.8 Type of assistance Donor Grants Total IMF 500 World Bank 993.8 Import and budget finance 252.5 MEER project finance 505 Reallocation 236.3 International Finance Corporation 50 European Investment Bank  607 Council of Europe Development Bank  303 Italy  18 Spain  60 Belgium  4 Islamic Development Bank 300 Import trade finance 150 Project finance 150 Gulf co-operation Council 400 Black Sea Trade and Development 10 Bank Japan 450 Commodity loan 200 Project finance (SMEs) 250 South Korea  30 EU  35 35 Germany 12 12 Total 47 3 772.8 (1.)Assumes $1 = 1.01 euro (as of 24 February, 2000). (2.)Tied. Source: Turkish authorities. Table 39. The post-earthquake agenda Phase/activity Turkish action plans Emergency relief Phase/activity Turkish action plans Locate people to shelters, provide By end-1999 121 tent towns social services and food until converted into 41 500 prefabricated permanent houses are built houses for the winter. Emergency cash to those who find own housing. Preconditions for reconstruction Complete land surveys of geological Completed spring 2000. Needed for conditions Complete new land land valution, housing planning based on geological construction, and resumption of surveys Establish ownership rights commercial activities. Ankara for housing guarantee largely in charge, but local mayors want to plan for local needs. Transition during reconstruction Housing construction and Housing construction highly installation of basic interventionist. State is building infrastructure in new areas 40 000 homes, with supporting infrastructure, entirely designed and allocated from Ankara. Target date for completion end-2000. Compensation for lost or damaged Mix of interventionist (housing houses swaps) and bottom-up approaches (cash payments for housing repair or relocation). About 125 000 households affected by the latter. Post-reconstruction and normalisation Move to market-driven economic Regulatory reforms to banking development sector. Mandatory insurance scheme for buildings. Removal of housing replacement guarantee. More efficient government land Outsourcing of municipal planning and better building inspections function to private standards engineering companies with professional standards and legal liability; upgrading of building codes to EU standards; specification of municipal and professional responsibilities and penalties; reform of public contract tendering rules. Better disaster response management National emergency management agency set up, consolidating hitherto dispersed emergency planning and response functions. Phase/activity Recommendations for further action Emergency relief Phase/activity Recommendations for further action Locate people to shelters, provide Better temporary accommodation social services and food until (e.g. vacant building) for people permanent houses are built still living in 54 tent towns. Preconditions for reconstruction Complete land surveys of geological Encourage national/local conditions Complete new land co-operation in urban planning. planning based on geological Assemble international team of surveys Establish ownership rights urban planning advisors. Local for housing guarantee mayors have ideas, but no expertise in urban planning. Need also to plan for adequate rental housing and needs of business. Transition during reconstruction Housing construction and Introduction of market forces in installation of basic redevelopment, for example, giving infrastructure in new areas tradeable vouchers for land auctions, and partnering with private sector for commercial redevelopment. Use the reconstruction to improve cities and their economic base. Compensation for lost or damaged Compensation should permit as much houses choice as possible. Cash is better than houses. Tradeable vouchers are a possibility. Incentives better than standardised approaches. Post-reconstruction and normalisation Move to market-driven economic Risk-based insurance for new development buildings. Government should provide risk information on land use to sellers/buyers/financiers. More efficient government land Responsibilities in planning law planning and better building must be clarified with standards corresponding sanctions. Urban planning standards should be more rigorous with informed citizen input and more methodical distribution of land. Local authority reform to improve implementation of planning. More training of engineers. Better disaster response management Follow up on plans to educate the public and to engage the local authorities. Need to retrofit schools, hospitals, and possibly houses, in the short to medium run. Source: OECD.
Box 9. Engineering, construction and siting errors in the earthquake zone
In the past, builders in Turkey used traditional, time-proven construction methods which minimised destruction in earthquakes. However, the earthquakes revealed several defects in urban planning, building design and construction which may have amplified the material and human toll of the disaster. The resulting loss of life is especially tragic because the majority of the affected buildings were less than twenty years old, by which time significant advances in international earthquake engineering knowledge would have been available. For the 30 years previous to August 1999, Turkey had suffered far fewer earthquakes than would be expected from its seismological track record. This lull reduced the awareness of seismological risks lust at the time when huge numbers of buildings were being erected. The following types of error were common:
Siting on liquefiable soils: large-scale urbanisation has been permitted on liquefiable soil. In the absence of adequate foundation systems, this was a primary cause of the destruction in cities like Adapazari.
Siting on the fault line: many other structures were also destroyed by virtue of their having been built directly over, or immediately adjacent to, the ruptured segment of the North Anatolian Fault. Entire villages and developments (including the Golcuk naval base) straddling the fault were destroyed.
Construction engineering problems were another major contributing factor to the disaster. These included weak ground level stories of multi-storey buildings, related to their usage for commercial purposes, with fewer retaining walls and higher ceilings, causing ground floor collapses. Widespread use of unreinforced hollow clay tiles to construct interior and exterior walls had deleterious effects on the performance of the structural frame during the earthquake, with an ensuing failure to provide adequate shear resistance.
Poor construction materials and construction errors: visual inspections of collapsed buildings suggest that the concrete in the typical building was often of a very poor quality and unacceptably weak. The presence of seashells in the concrete suggests that contaminated beach sand was used in the mix. Also, smooth (undeformed) steel reinforcing was used; causing degradation of the bond between the concrete and the steel reinforcing, which is not permitted in much of the world especially in construction expected to withstand the shock of an earthquake. Poor or incomplete use of reinforcing details could also be observed.
Source: Wiss, Janney, Elstner Associates, Inc. (1999); BBC Science News (15 November 1999).
Box 10. The Izmir plan of action
The city of Izmir in western Turkey has developed an action plan for managing seismic risk. The main obstacle to co-ordination of such a plan was the highly centralised legal structure of disaster management in Turkey. In light of this constraint, the international "RADIUS" project (Risk Assessment Tools for Diagnosis of Urban Areas against Seismic Disaster) acted as a catalyst in bringing together governmental and municipal organs, the professional chambers, NGOs, etc. [*] Four working groups were formed in order to deal with the following subjects:
The assessment of earthquake risks group (representatives from the chambers of geologists, geophysicists, civil engineers and related departments of the universities):
- the assessment of geological and tectonic structure of Izmir Metropolitan Area, including the study of active faults, landslides, and liquefaction;
- the assessment of past earthquakes in the vicinity of Izmir, the determination of their destructiveness, and the filing of seismic instrumentation records;
- in view of past earthquake experience, the determination of possible geological hazards and the conditions of buildings situated in these areas; evaluation of seismic factors and other disastrous events triggered by earthquakes such as fires, destruction of dams, etc.
The assessment of buildings and infrastructure group (representatives of the chambers of civil engineers, architects, related departments in universities, governmental agencies such as Ministry of public Works, Waterworks, Railways, Turk Telekom, Electricity, municipality subway and construction departments):
- the evaluation of the building stock within city boundaries, classifying them according to their age, construction material, height, etc.;
- the classification of buildings with regard to architectural configuration, foundation type, structural features and existing soil conditions;
- the evaluation of buildings destructed in past earthquakes;
- the evaluation of the infrastructure, including data on location, condition, construction date and materials of structures such as bridges, motorways, subway, airports, seaports, railways, dams, clean water and sewage systems, electrical energy transmission lines and transformers, PTT centres, fire stations, industrial complexes.
The assessment of social and economic situation group (representatives of chambers of architects, commerce, industrialists, medical doctors, NGOs, related departments of universities, and the media):
- the evaluation of city community with regard to cultural background, education, age, economic level, living conditions, immigration, etc.;
- determination of the locations of the trade centres, hypermarkets, transportation systems, recreational areas and open spaces, hospitals and clinics, schools, universities, police and army stations, which people frequent.
The relief activities and rehabilitation of the buildings group (representatives of the chambers of civil engineers, architects, city planners, commerce in particular builders and developers, medical doctors and city's health board, civil defence board, NGOs such as mountain rescue teams, amateur wireless network organisation, Agenda 21, fire brigade, army, police force, related university departments):
- the determination of guidelines for urban planning and new development areas which shall be least affected by natural disasters;
- the determination of efficient quality control systems in the construction sector and proposals for new building regulations, etc.;
- a guide to show communities how to behave during and after and earthquake;
- preparation of a guide and documentary films concerning the renovation or retrofitting of hazardous buildings after a major earthquake.
(*.) See www.geohaz.org/project/radius/radiusintro.htm.
Box 11. Compulsory earthquake insurance
The prospects of the local insurance industry expanding its coverage, if earthquake insurance were made compulsory, are limited, due to its low capital base (partly reflecting high inflation-protected returns from low risk activities) and the reluctance of leading international reinsurers to provide more capacity to the industry in its present state. In addition, a pure private sector approach is unlikely to succeed, as some insurers will attempt to underwrite only those risks that are overpriced, leading to adverse selection and instability for those who get the wrong part of the portfolio. Well-founded concerns over Turkish building standards are an aggravating factor.
Under these circumstances, it was perceived that a joint public/private sector solution was necessary to reduce the risk borne by the government and the property owners. On 27 December 1999, the Turkish Government approved a Decree-Law on the introduction of compulsory earthquake insurance. This measure represents an important and clear break with the past, when under the Disaster Law the government was liable for rebuilding housing destroyed by natural disasters almost free of charge. When the Decree goes into effect on 28 September 2000, those who fail to get insurance will no longer be able to benefit from government aid in the event of a natural disaster.
The Decree paves the way for the creation of the Turkish Catastrophic Insurance Pool (TCIP) that will be the stand-alone provider of earthquake property insurance in the country for up to $25 000.[*] Coverage in excess of that amount will be provided by private insurers. While initially, in the first several years of its operations, the TCIP will be supported by a contingent line of credit from the World Bank, its major sources of capital support will come from the collected premiums (a flat annual fee of roughly $50 per household), reinsurance and excess of loss treaties and, possibly, issuance of catastrophe bonds. It is envisaged that in ten years, the TCIP will have enough capital to protect Turkish homeowners against catastrophic events larger than that occurred on 17 August 1999, which would effectively relieve the Turkish Government from its large contingent liability due to frequent natural disasters. The risk that the government will be unable to resist political pressure to turn TCIP into an indiscr iminate liquidity conduit to victims of a major future earthquake, regardless of their terms of coverage, is limited by the financial design of the programme.
In addition to its core earthquake insurance functions, it is expected that the TCIP would also greatly contribute to the better enforcement of building codes through a commercial arrangement with independent engineering firms that would be retained to certify the construction quality of new residential dwellings to be insured under the plan. However, since there will be a single undifferentiated premium, this will need to come via pressure from the reinsurers rather than a built-in incentive structure. Indeed, care will need to be exercised to ensure that the insurance premium is not treated as a tax and ultimately assimilated into the central government budget. To this end, it is envisaged that after two years' operation, TCIP will introduce either a reduced premium or greater coverage if the insured adopts recommended mitigation practices.
(*.) To reduce moral hazard and encourage top-up cover from the private sector, claims payments will be subject to "average", a method by which loss claims are scaled down in inverse proportion to the value of the property. Thus, only houses worth $25 000 would be able to get the full maximum coverage.
Source: World Bank (1999c).
Ahn, S. and P. Hemmings (2000),
"Policy Influences on Economic Growth in OECD Countries: an Evaluation of the Evidence", OECD, Economics Department Working Papers, No. 246.
Atiyas, I. and S. Sayin (1997),
Political Accountability, Managerial Accountability and Budget System: Towards a Proposal for Restructuring, in Turkish, TESEV Publications No. 4.
Atiyas, I. and S. Sayin (1998),
"Budgetary Institutions in Turkey", presented at Mediterranean Development Forum, Morocco, 3-6 September.
Atiyas, I., Y. Gunduz, F. Emil, C. Erdem and B. Ozgun (1999),
"Fiscal Adjustment in Turkey: the Role of Quasi-Fiscal Activities and Institutional Reform", T.C. Hazine Mustesarligi, Arastirma Inceleme Dizisi 22.
Blondal, S. and M. Pearson (1994),
"Unemployment and Other Non-employment Benefits", Oxford Review of Economic Policy, Vol. 11, No. 1.
Cilli, H. and C. Kaplan (1998),
"Analysing the Impact of Disinflation on the Banking System: a Conceptual Note", in Macroeconomic Analysis of Turkey: Essays on Current Issues (ed. M. Ucer), Central Bank Research Department Publications, pp. 220-241, December.
Demirguc-Kunt, A. and H. Huizinga (1999),
"Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence", The World Bank Economic Review, Vol. 13, No. 2 (May), pp. 379-408.
Fischer, S. (1982),
"Seignorage and the Case for a National Money", Journal of Political Economy, Vol. 90, No. 2.
Fischer, S. (1993),
"The Role of Macroeconomic Factors in Growth", Journal of Monetary Economics, Vol. 32, No. 3.
Folkerts-Landau, D., G.J. Schinasi, M. Cassard, V.K. Ng, C.M. Reinhart and M.G. Spencer (1995),
"Effects of Capital Flows on the Domestic Financial Sectors in APEC Developing Countries", in Capital Flows in the APEC Region (eds. MS. Khan and C.M. Reinhart), International Monetary Fund, 3 1-57. Washington.
Forster, M.F. (assisted by M. Pellissari) (2000),
"Trends and Driving Factors in Income Distribution and Poverty in the OECD Area", OECD, Labour Market and Social Policy Occasional Papers, No. 42, August.
Ghosh, A. (1999),
"Turkey's External Current Account", in IMF, Turkey: Selected Issues and Statistical Appendix, December.
Gonenc, R., M. Maher and G. Nicoletti (2000),
"The Implementation and the Effects of Regulatory Reform: Past Experience and Current Issues", OECD, Economics Department Working Papers, No. 251.
Government of Turkey (1999),
"IMF Stand-By Agreement: Letter of Intent dated 9 December 1999", www.tcmb.gov.fr or www.treasury.gov.tr.
International Monetary Fund (IMF) (2000),
Report on the Observance of Standards and Code: Fiscal Transparency, Washington.
Jarvis, C, and A. Selassie (1999),
"The 1999 Social Security Reform", in IMF, Turkey: Selected Issues and Statistical Appendix, December.
Kiguel, M.A. and N. Liviatan (1992),
"The Business Cycle Associated with Exchange Rate-Based Stabilizations", The World Bank Economic Review, Vol. 6, No. 2.
Kogar, C. (1998),
"Operational Budget Balance: Methodological Issues and Application to Turkey", in Macroeconomic Analysis of Turkey: Essays on Current Issues, (ed. M. Ucer), Central Bank Research Department Publications, pp. 171-211, December.
Mishkin, F.S. (1999),
"International Experiences with Different Monetary Policy Regimes", NBER Working Paper, No. 7044, March.
Moalla-Fetini, R. (1999),
"Inflation as a Fiscal Problem", in IMF, Turkey: Selected Issues and Statistical Appendix, December.
Mussa, M., P. Masson, A. Swoboda, E. Jadresic, P. Mauro and A. Berg (2000),
"Exchange Rate Regimes in an Increasingly Integrated World Economy", IMF Occasional Paper, No. 193.
National Science and Technology Policy Review: Turkey, Paris.
Managing Across Levels of Government, Paris.
Energy Policy Review: Turkey, Paris.
Fostering Entrepreneurship, Paris.
OECD Economic Surveys: Turkey, Paris.
Agricultural Policies in OECD Countries: Monitoring and Evaluation 2000, Pads.
Ozmen E. and C. Kogar, (1998),
"Sustainability of Budget Deficits in Turkey with a Structural Shift", METU Studies in Development, 25(1), pp. 107-128.
Oxley, H., T.T. Dang and P. Antolin (2000),
"Poverty Dynamics in Six OECD Countries", OECD Economic Studies 30, 2000/1.
Ozatay, F. (1997),
"Sustainability of Fiscal Deficits, Monetary Policy and Inflation Stabilisation: the Case of Turkey", Journal of Policy Modelling, 19(6), pp. 661-681.
Ozgun, B. (2000),
"An Empirical Approach to Fiscal Deficits and Inflation: Evidence From Turkey", T.C. Hazine Mustesarligi, Arastirma Inceleme Dizisi 25.
Parker, R., A. Kramer and M. Munsinghe (eds. 1995),
Informal Settlements, Environmental Degradation, and Disaster Vulnerability: The Turkey Case Study, The International Decade for Natural Disaster Reduction (IDNDR) and the World Bank, Washington.
Rajan, R. and L. Zingales (2000),
"The Great Reversals: The Politics of Financial Development in the 20th Century", OECD, Economics Department Working Papers, No. 265.
Saygili K. (2000),
"Funds System in Turkey" mimeo, SPO.
Selassie, A. (1999),
"Implications of Disinflation for Banks' Profitability in Turkey", in IMF, Turkey: Selected Issues and Statistical Appendix, December.
Selcuk, F. and E. Yeldan (1999),
"On the Macroeconomic Impact of the August 1999 Earthquake in Turkey: A First Assessment", Department of Economics, Bilkent University, Bilkent-Ankara, Turkey, September.
State Planning Organisation (SPO) (2000a),
"Restructuring of Public Finance Management and Fiscal Transparency", Report, Special ad hoc Committee, SPO, March (in Turkish).
Global Forum on Local Governance and Social Services for All: Turkish Case for Local Administrations, Stockholm, May.
Van Rijckeghem, C. (1977),
"The Political Economy of Inflation: Are Turkish Banks Potential Losers from Stabilization?", Istanbul Stock Exchange Review, Vol. 3, pp. 1-16.
Wiss, Janney, Elstner Associates, Inc. (1999),
WJE Earthquake Engineering Update, September.
World Bank (1999a),
Managing Disaster Risk in Mexico: Market Incentives for Mitigation Investment, Washington, June.
World Bank (1999b),
Turkey: Marmara Earthquake Assessment, Turkey Country Office, September.
World Bank (1999c),
Project Appraisal Document of a Proposed Loan in the Amount of US$505 million to the Republic of Turkey for a Marmara Earthquake Emergency Reconstruction Project, Washington, November.
World Bank (2000a),
Turkey: Economic Reforms, Living Standards and Social Welfare Study, Washington, January,
World Bank (2000b),
Turkey: Country Economic Memorandum, Structural Reforms for Sustainable Growth, Washington, September.
|Printer friendly Cite/link Email Feedback|
|Publication:||OECD Economic Surveys - Turkey|
|Article Type:||Statistical Data Included|
|Date:||Feb 1, 2001|
|Previous Article:||The structural reform programme.|
|Next Article:||Annex I: Links between growth and inflation: evidence for Turkey.|