Applicability of Corporate Risk Management Techniques to Catastrophe Risk Management at the Country Level Rodney Lester Lead Specialist Contractual Savings and Insurance Business group FSEVP Financing the Risks of natural Disasters: A New Perspective on Country Risk Management World Bank Washington DC June 2-3, 2003 I sometimes think that there are as many meanings for the words ‘risk management’ as there are pairs of ears to hear them. Every person who is conscious of their surroundings and who can think ahead deals with risk. If there were any pie eyed optimists left in the world by the late 1990s the recent combination of collapsing markets and man made and natural disasters will have thinned their ranks considerably. Given that the somewhat perverse genius of mankind has rapidly increased the potential for catastrophic losses in the last century, and the last few decades in particular, there is little doubt that the world as a whole, including governments, is more conscious of risk. We also know that people are not consistent in their approach to risk. People, including politicians, generally have trouble dealing with small probabilities, the area of risk where most people in this room find their primary interest. On the other hand people tend to overestimate risk where a dread factor exits, such as nuclear accident. Often people also assume that they will be bailed out in the event they are struck by a major natural disaster, a sometimes rational judgment given the typical knee jerk reactions of politicians following natural calamities, and earthquake in particular. In the World Bank we have been grappling with the question of natural hazard sourced risk for upwards of two decades. In a paper written some years ago we pointed out that by some measures the Bank is one of the World’s leading catastrophe reinsurers. This arises out of a specific section of our operating rules which covers the provision of loan funds post disaster. This may sound strange but if you accept the premise that cat. excess of loss reinsurance is a form of inter temporal funding then the comparison becomes quite valid. If I wished to be provocative I could argue that the Bank may in fact be a preferred source of funding in this regard as its pricing tends to be more consistent over time! Having said this I should hasten to add that, contrary to the presumptions of some, we do believe in markets – whole heartedly in fact: if the markets are doing their job there should be no need for Bank involvement. If we are present in a market in a way which is obvious to the insurance sector it is because risk transfer markets are yet to develop, or have become distorted in some manner. In fact the model for country level management of risk which we have developed, and which will be disseminated in the coming months, owes much to the private sector corporate risk management model which has been developed in the last few decades. I first became aware of this model in the hiatus between leaving the AMP in Australia and joining the World Bank, when I had some free time and decided to bring myself up to date on the latest thinking on risk, as well as losing some weight. In particular I discovered the ARM certificate and the associated text books, which lay out a clear and logical approach to risk management for large and complex enterprises. Subsequently we have seen the publication of Neil Doherty’s book, Integrated Risk Management, which providing an intellectual foundation for the common sense approach already in place, at least for the more enlightened organizations. I should perhaps add as an aside here that the reinsurance pricing cycle plays havoc with the introduction of good risk management practices and Neil’s book in particular has been most timely. The final pieces of the puzzle came together with Paul Freeman’s work in Latin America on post disaster fiscal gaps, our own disaster insurance work in Turkey, India and the Caribbean, and Bob Litan’s efforts at Brookings to develop an economic taxonomy for the impact of natural disasters at the country or state level. These have all convinced us that the concepts developed for large corporations can be adapted to country level risk management. The corporate risk management model, particularly as elaborated upon by Neil Doherty is a complex animal indeed. It is effectively a new element of the prevailing corporate finance model within shareholder capitalism, which feeds into such matters as optimal leverage, the cost of capital, investment choice and valuation. For our purposes however we are concerned about the process which leads to the optimal risk management approach. This can be summarized by the following figure: • • Risk Management IDENTIFY RISKS - ASSESSMENT DEVELOP STRATEGY – RETENTION – AVOIDANCE – CONTROL – TRANSFER MITIGATION Experience indicates that corporate management tends to be more risk averse than finance theory would say is ideal – I should perhaps emphasize here that we are talking about shareholder rather than managerial capitalism, its unhealthy offshoot. In any case the key stage is the first one, identifying and measuring the risks. This is a subtle and at times frustrating exercise, particularly where there are many interlinking processes, and the ARM text books lay out a range of techniques to identify what can go wrong. Deciding how much risk to retain is also not always immediately obvious as the costs associated with mitigating and arranging ex ante funding will determine options. Those of you have had to decide retentions when establishing annual reinsurance programs will know exactly what I mean. In the past we have presented this model many times to senior government officials and received little response. In retrospect this was entirely predictable – if senior management are a hard sell, senior bureaucrats who have to survive under politicians with short term and often economically destructive agendas are going to be even worse. In addition there are many countries which have broad tax bases, established insurance sectors and strong borrowing capacities for which the do nothing strategy may in fact make sense. These countries tend not to be World Bank clients by definition. Thus we have often found ourselves in a similar situation to the poor corporate risk manager trying to get the message up the line to the decision makers. However things are changing. In particular the frequency and severity of major calamities is increasing. This chart for India makes the point eloquently. Reported Catastrophe Losses in India, 1965-2001 Nominal US $million at then applying exchange rates 5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 01 20 98 19 95 19 92 19 89 19 86 19 83 19 80 19 77 19 74 19 71 19 68 19 19 65 0 Source: CRED, International Disaster Database, Université Catolique de Louvain, Belgium. The reasons for this increase in loss potential can be debated over the next two days, however an increase in exposures and vulnerabilities must almost certainly be an important part of the picture. The potential for loss has now reached a stage where many countries cannot adopt the null position and effectively ignore loss potential from natural disasters. The technical economic argument for countries to take a null position was set out by Arrow and Lind when they argued that the natural economic portfolio of a country ensures that governments only need to set aside the expected cost of natural disasters each year in the budget process and that variability will not cause too much trouble. This in fact is what many countries have done, either implicitly as in most jurisdictions, or explicitly as in Mexico, India and some European countries. The unpleasant emerging fact is that even the United States is not large enough for the variability of current potential natural disaster losses to be lost in the economic or fiscal noise. There are economists who even argue that natural disasters can have a positive economic impact, presumably through replacement of outdated physical capital, but the funds need to be there to do so. Another factor we have discovered in ex post situations is that politicians cannot resist handing out money, even if the economic returns are meager. The classic case is middle class housing, which should ideally be dealt with through private insurance markets. Typically around 50% of World Bank ex post disaster lending goes on housing, when much of it would have been better spent on building essential infrastructure and improving the lot of the poor. Finally we find that the tax ratio, the ratio of tax collections to GDP, in many client countries is pathetic – typically around 17% or less. Thus even if the GDP impact is not huge the impact on government finances of a suddenly crystallized obligation can be significant. In India we estimate the burning cost of direct natural disaster losses is running at about 12% of federal government revenues, with high variability. Thus we have a basic product and an emerging need – where to from here? At this point I should make a confession – when we designed the Turkish earthquake pool, known as TCIP, with the Turkish Treasury we were all under enormous time pressure and only had sketchy historical loss data available. In fact we modeled the earthquake risk based on some back of the envelope assumptions which now make me blush – although it is amazing how close the back of the envelope can get to what pops out of the sophisticated black boxes now available from the modeling companies. In addition we did not have a country risk management model available at the time – essentially we intuited that it was wrong that the Government of Turkey should be using scarce government funds to rebuild houses for those who could afford to buy insurance, and that we needed to build a mechanism to get disaster insurance to the people effectively , at the lowest possible cost and with the widest possible distribution. Given the circumstances it turned out pretty well. However there have been some lessons learned, the first of which is to never take an opportunity away from a politician to look like a good guy. And post disaster situations are the ultimate in this regard. So there will need to be something there to deal with this reality if TCIP is to remain a success. In addition we underestimated the need for people to have fairly constant reinforcement if they are going to buy an intangible like insurance, and it is not to be seen as a tax. My colleague Eugene Gurenko will elaborate on this shortly. Our thinking has moved on considerably since the Turkish TCIP was designed and we are about to publish a study on India which includes a fully fledged risk country risk management model. This is similar to the earlier corporate model with some additional feedback loops (see Figure below). In particular the country risk management model conceptually recognizes that there will be both short and longer term fiscal and economic effects which are dependent on the risk management approach adopted. For example if it is decided to fund reconstruction through debt rather than taxes there will be timing effects and there are also likely to be re-distributional effects between sections of society. Given the manifest problems in measuring broad economic effects our current model focuses largely on fiscal issues, which are much more tractable. We are hopeful that a full economic risk management model will emerge at some stage but the challenges are considerable, and we have found that civil servants react well to a fiscal approach in any respect. Country/ State Risk Management Model STATE OF ECONOMY – Stocks (Capital), Flows (GDP) IDENTIFY LOSS POTENTIAL DEVELOP STRATEGY RETENTION AVOIDANCE CONTROL FUNDING Mitigation Direct Losses and ST Revenue and Cost Effects Redistributional and LT Debt Effects The key tools are similar to those that would be used by a corporate risk manager, although perhaps a little more formalized. These include the development of loss exceedance curves for the major hazards and centers of exposure. These curves provide potential loss levels by cumulative probability level (usually called return periods in the insurance world) and these in turn can be fed into relatively simple table. Then, based on the fiscal stance of the government concerned decisions can be taken about required levels of expenditure on mitigation and the extent to which ex ante funding mechanisms, including insurance, should be adopted. We have begun to prepare some policy notes on the trade offs between the various ex ante funding tools and can perhaps discuss these in the next few days. This table includes a forced ranking of post disaster priorities and you will note that housing for the non poor comes a distant last. Whether this can be enforced in practice is debatable, but a formalized approach may at least reduce the prospects of politicians creating moral hazard amongst those sections of the citizenry who should be able to look after themselves. Funding Gap Reconstruction and Rehabilitation 50 Year Event 100 Year Event 150 Year Event Government’s post disaster commitment 1.Lifeline infrastructure replacement 2.Provision of living needs of the poor 3. High return and sustainable mitigation investment 3.Other infrastructure 4. Other housing Ex post sources of funds Aid Central funds transfer Budget reallocation Tax surcharge Domestic credit Development Banks Other external Credit Ex post funding gap Thus in summary we now have a fairly well developed country risk management model in place and are looking forward to working with our colleagues in the government, reinsurance and investment banking sectors as well as with colleagues within the World Bank, on making the approach operational. We look forward to your contributions to further developing this approach in the coming sessions.