Dynamic Corporate

Transcript

1993 General Insurance Convention Dynamic Corporate Management MaintainingSolvency Over Time Report ofThe Dynamic Solvency Working Party tothe General Insurance StudyGroup Membership: RichardBulmer Francis Chacko James Dean Andrew Macnair RaviManjrekar Nicholas Michaelides, chair Peter Rains August 1993 1 5 5 1 5 6 CONTENTS 1 to Dynamic Solvency Introduction 2 Why Solvency Matters 3 What isDynamic Solvency Testing? 4 Where isCorporate Planning Today? 5 Integration ofSolvency and Planning 6 Modelling the Company in a Market Context 7 Select Bibliography 157 Ifwe consider the personality profile of a typical insurance company to be task manager, (s)he islikely oriented, witha short-term perspective to be pragmatic. and a tendency Over-anxious tofind solutions, (s)he will make decisions before understanding the problem or the implications of what is decided. Itis thenlittle wonder that theindustry has the problems that it does.An idealistic, visionary leadership which sees through the ups and downs ofthecycle to the real needs of the customer might generate some immunity from theworseexcesses ofprice competition. I fi ti st r u e that“people will alwaysneed insurance”, we really have no excuse for losing so much money. The primary purpose of this paperistolengthen the horizon of thinking aboutthebusiness so as to encourage realistic decision making that will create sustainable advantage without first crippling the balance sheet. The goldenrule isthat there areno golden rules so do not expect to find to answerthequestions all theanswers buttry we pose. R e a do n..... 158 Dynamic Corporate Management - Maintaining Solvency Over Time 1 Introduction to Dynamic Solvency The workingparty came to the subject in the belief that this was to be virgin territory and thata simple, introductory paper would break new ground. It came as quitea shock to discover justhow insular t h i s perspective was. Withinthebibliography are some of the papers we have considered but each of theseisthe tipof the extant literature. Each carries its own list of papers and the Brender et al paper is an extremely valuable bibliography and commentaryinitself and highly recommended reading. Itquickly became evident that we could all too easily replicate what has gone on elsewhere. However,if this paperachieves nothing else, it will at least cause us toconsider ouractivities on a wider stage and point towards previous achievements. Thisgroupis an amalgam of two;it incorporates those who expressed an interest in the costof maintaining year-end solvency. While this isof interest, it is perhapssubsidiary. The cost of meeting the statutory hurdles by (ab-) using certain accounting mechanisms, financial instruments and so forth may represent the tactics of a company which has not,in previous years, fully incorporated the conceptof dynamic solvency. To be solvent isto be able to pay alldebts. In the shortterm,any company witha positive cashflowshould therefore be solvent. A strong flow ofcashinto thecompany today may, however, simply be indicative of underpriced business which will generate a large outflow at some future date.Thus there isa strong responsibility placed on the management to ensure that atall timesthecompany currently has the resources to meet all the commitments it has incurred. In most markets, 1 5 9 managementsare not trusted to do this without somebody looking over their shoulder. Beingfinancial institutions of theform that incomeisknown-ish but costs are not,it is imperative for many interested parties that insurance companies can demonstrate publicly that theyhave theresources to meet their commitments inreasonably foreseeable circumstances. In the UK at present, most of the focushas been on the situation the company finds itself inatsome arbitrary date.Ingeneral, thequestion asked is whether thecompany can meet all its liabilities from its existing assets. In the UK and therest of theEC, thetest is whethernetassets exceed a minimum required margin. To calculate this margin, both premium written and claims incurred are taken as measures of exposure. Fixed percentages areapplied tothese figures and thelarger of the two outcomes is taken to be theminimum margin. The effect ofthis method is generally to require thesame margin forall companies ofa given size irrespective oftheir mix of business. Long-tailed lines will, over time, build up proportionately larger technical reserves thanshort-tailed ones.As reserves aresubject t o fluctuation, a given percentage change in those of a company with a longtailed bias would have a bigger impact on the margin than in, say, a personal lines office. This isbut one limitation of the static, year-end approach. Dynamic solvency takes thehorizon out to a future date when we can try to assess the company’s strength giventhe execution of its business plan in the meantime. In dynamic solvency testing, we try to understand the circumstances which would placethe company under thegreatest s t r e s s givenitscurrent strategies. It can then decide the extent to which strategies should be amended to balancestress potential with profit potential. Thisbrings solvency assessment from a concern with keeping supervisors happytouseasa management tool. We need, though, to look at both perspectives. 160 1.1 Supervisor’s Role United Kingdom The DTI assesses thereturns submitted by a company within six months of itsfinancial year-endto determineitsability to continue writing soundly. The EC requirement fora solvency margin of at least 16% of netpremiums was (perhaps optimistically) intended to protect an insurer against underwriting losses sustained betweenits financial year-end and thepoint atwhich thecompany’s returns havebeen prepared, submitted and analysed. The quasibreak-up cost and valuation basis required by UK law does not takeexplicit account of thecompany’s intentions which iswhy the DTI has started a series of informal visits to companies to understand their future plans. Canada in Canada,plansare afootto implementdynamicsolvency throughthe establishment of appointed actuaries. Here the focuswill be on the business plan and the ability of the company’sresources to sustain planned activities over at least the next twelve months. The stated objective ofCanadianDynamic Solvency Testing or DST is to “enable theactuary to provide advice abouttrends insurplus and threats to thecompany’s solvency, and to identify courses of action which may mitigate the threats”. This isamplified in one of thefootnotes which says, “It isimportant to realize thatthe primarypurposeof this exercise isnot to findout ifa company will be solvent or notata future date, buttheextent that certain factors or elements can adversely affect thesolvency of a company”. This pro-active stance should helpcompanies avoid strategic dead ends as long as management is prepared to amend itsstrategy and introduce appropriate contingency plans. It is probably at this point that the supervisor and management may be expected toconverge, if not clash. 161 The United States ofAmerica The USA isapproaching solvency from the so-called risk-based c a p i t a l approachwhere capital requirements are based on the scale of the balance sheet risk, both of assets and liabilities. Another working party hasthebrief tocover this development and we do not attempt to emulate them here. 1.2 Manager’s Role Most managers are interested in keeping their jobs and getting promotion. Determining the impact tomorrow of what they decide today has not alwaysbeen one of their strong points yet isessential if their primaryinterests are to be fulfilled. Business planning is still in its formative years inthegeneral insurance business and has mostly been the preserve of theaccountant. Hence, much of the work has been on point estimating and budgetary control. The horizon of thewisemanagerwill be much longer than that of either theDTI or theCanadianauthorities. Also, it is imperative that even the sacred cows aresacrificed intheinterest ofan objective assessment of the future. For example, history teaches us that nobody ever learns from history. So,who istosaythat thenext downturnoftheinsurance cycle will not hit us or be even more severe thanlast time?What corporate plan everincluded sucha scenario? They all seem to show growth and chirpy loss ratios. We hope that we may helptoopen a chink oflight inthis dark corner. 1.3 Our Objectives Within this paperwe shall endeavour tolink thesupervisor’s interest with thatof the management. It is clearly vital that the company remains solvent in the eyesof the supervisor. Also,in orderthatdirectors can continue to allow their company to trade, they must at all times reasonably believe that it iscurrently solvent. However, that may not be sufficient forlongterm viability. Everyfailure was once solvent! HOW do we ensure that ouremployer or client goesfrom strength to strength? 162 In other words, how do We manage our company capital strength in the light of therelative position of our competitors and the absolute rest of solvency adoptedby theparticular regulatory framework we are working within? Given that capital is relatively scarce, how do we best allocate our resources in order to maximisethevaluecreated by their use? To what extent should risk be managed by theinsurer rather than the shareholder? As buying reinsurance reduces thelong-term profitability of the business, itwould not appearto be in the interests of the shareholder so what factors should motivate its purchase ? If reinsurance is bought for capacity purposes, why notgetthe capital from the current investors and reward them rather thantheshareholders of the reinsurer? We consider briefly the perspectives of thevarious stakeholders to show boththecommon interest insolvency aswellassome ofthedifferences in perception as to what makes a company solvent. We thenexaminethe roleof corporate planning in helping the management to ensurethe prosperity of the company. Next is a section on the ways in which assessment ofsolvency can be incorporated into theplanning process. It is this section which addresses the key issues raised in the previous paragraph. The features of a market-oriented model are then briefly discussed. Finally therefollows a select bibliography which includes comments on thepapers listed. We suggest that we raise more questions thananswers but perhaps there isa lesson there. Let us not presume to know but be ready to find out. 1 6 3 2 Why Solvency Matters While we can define solvency in this context as “at all times currently having the resources to meet all thecommitmentsalready incurred”, the relevance and even understanding of this willvary according to the perspective of the stakeholder. 2.1 Managersand Employees Good management will consider the aims and needs of policyholders, shareholders, supervisors, etc., and so will have an eye to all the considerations mentionedin sections 2.2-2.5 below. It will also have regard totheposition of thecompany workforce and of thedirectors and managers themselves. Generally, managersand workforce will be interested in the continued solvency of the company,so thatsalaries continue to be paid and the pension fund topped up. Some bonus schemes can encourage shorter term thinking althoughcontinuedemployment should be a sound incentive to manage a financial institution on a long term basis. Also senior managersin a company which has failed financially may find it difficult to obtaina similar post elsewhere. The degree to which reputations suffer will depend on thereasons forfailure but it israrely a wise move to have such an association, either from the market’s viewpoint, or that ofthesupervisor where, as intheUK, thelatter has the power ofveto oversenior appointments. Many of theconsiderations will suggest conflicting courses ofaction, and management willsometimes need a pragmatic solution in order to balance theneedsofthevarious stakeholders. Thisis most clearly seenin the distinction between a company’sabsolute position and its position relative toits main competitors. When considering, forexample, the level of exposure to equities in the company’s investment portfolio, management may decide that the dangers 164 from overor underexposure are smaller than the risks from a level of exposure whichis outof line with that of the company’s peers. If a stock marketcrash hits all major insurers equally, the supervisor cannot (the argumentruns)closethem alldown; the suppliers of capital will not desert thesector entirely; themore talented managerscannot all flee; and so on. There isa subjective elementto this which can be seen in the Canadian approach of dynamicsolvency testing usinga rangeof assumptions. In some scenarios all insurers will be insolvent, while inothers practically a l l will be solvent. The importance of assessing the relative position i s reflected in thegrowthof agencies whichevaluate the financial strength of insurance companies and allocate a rating. Most companies will have passed thestatutory tests but some will have cleared the hurdles witha greater marginthanothers. The assessment by these agencies is taken into account by those who buy insurance and it islikely that better quality business will go to thestronger companies, thusaccelerating the decline of thosewhich are already weak. Management of allaspects of risk commensurate withcapital resources istherefore vital to avoid becoming frail relative tothemarketand starting a steadily falling s p i r a l . 2.2 Shareholders Ithas been arguedthat in a perfect market companies shouldnot, from theperspective of their shareholders, takeactions toprotect solvency on thegroundsthat itisforshareholders to diversify their own investment portfolios if they wishto reducerisk. The argument runs that any actions taken toprotect solvency imply extra costs, suchasthereduction in return from investments which do not maximiseyield or the profit margin element ofa reinsurance premium. In the realworld, dealing costs prevent shareholders from diversifying their portfolios completely, so theylookforcertain companies toexercise a degreeof prudence.Ifa shareholder invests in companieswithrisktaking profiles, thenhe/sheshouldnot be too discontent ifthere isthe 165 occasional failure. However, insurance companiesnecessarily present quite the opposite profile, and so servetheir shareholders by following low-risk strategies. Ifinsurance futures everaregenerally available, t h i s will become even more true, as investors will then have a choice between investing directly in insurance through futures or investing through the risk-filtering medium ofan insurance company. It would notbe impossible to imagine an insurance marketconsisting of majorand minorinsurers, all following low-risk strategies, plus a rangeof specialist insurers following high-risk strategies such as not seeking to reduce underwriting risk(by reinsurance) any further than occurred naturally through the pooling of risks. While this would require special treatment by policyholders (who would themselves have to diversify by spreading their cover)and regulators, such a market couldservethe interests of shareholders better by making available a spread of risk l e v e l s . However, it shouldbe a feature of such a market that most of the insolvencies would occur among the specialist companies and be attributable to identified high-risk causes.In reality, the majority of failures seems to result from poor management practice of one sort or another, rather than from following a highrisk/reward strategy. This is reflected in theJune 1991 report of AM insolvencies from 1969to 1990: Deficient loss reserves/inadequate pricing Rapid growth Alleged fraud Overstated assets Significant change in business Reinsurance failure Catastrophe l o s s e s Miscellaneous Beston primarycauses of US 28% 21% 10% 10% 9% 7% 6% 9% 1 6 6 Whilethere will be elements ofhigh risk and poor management in at least thesecondand sixth of these, we could, somewhat arbitrarily, classify t h e first fouraspoor management and thelatter fouras highrisk/reward. ( l t istruethat to invest in a company with poor management is a high risk, butit isalso likely to leadtolow rewards.) Thus we coulddivide f a i l u r e s into 69% poor management and 31% highrisk strategy, and say that the insurance worldofthelast twenty years does not appear to have rewarded an investor seeking to invest efficiently in high-risk/reward insurancebased ventures.In view of this, shareholders of insurance companies should be presumed to be expecting the management to seek a low-risk s t r a t e g y . Finally, shareholders may need a steady streamof dividends in order to budget accurately or tomeet their own capital commitments. This should be less ofa problemfora shareholder witha diversified portfolio, b u tw i l l be particularly important for one which is a holding company for the insurer and which has no other assets. If it is financed by debt or cumulative preference shares, it will be dependent on a reliable incoming dividend streamto pay its own interest or preference dividends. Clearly, theability of an insurance company to maintain its dividend i sr e l a t e dt o its financial strength and toits ability togenerate both cash and profit. 2.3 Policyholders Itnegates the whole point of insurance ifan insurance company cannot pay claims fully and promptly. Even where policyholders are protected in theevent ofinsolvency, such as in the UK through the PPB, there is likely to be an increased delay, the loss of any claimsadviceservice, and perhaps a deduction from the amount refunded (10% in UK for noncompulsoryinsurance).However, such protection does reduce the importance of solvency to the policyholder, which perhapsexplains t h e predominanteffect of pricein the current UK private motor market. Whether suchemphasis on cost rather thaninsurer quality is appropriate isquestionable, since itseems likely that a strong company will assess claims more fairly than a weak one. 1 6 7 From a statistical viewpoint, what an insurance company is selling isnot the repaying of the costof insured incidents, but the removalof the variation inthis cost. In effect, the part of the premium which represents theaverage cost of incidents whichtheinsured mightexpect tosuffer over thepolicy period can be separated from therest, and indeed is sometimes reduced or evenremoved completely by deductibles, retentions and other forms of self-insurance. From the insurer’s viewpointthe remainingpart of the premium represents brokerage, expenses, profit loading, reinsurer’s loading, and othercosts, but forthe insured itismuch more straightforward. Itis simply that amount whichhe/she isprepared to pay totransform the cost of the insured incidents from an unknown, highly variable figure to a known quantity. It follows that an insured will onlybe prepared to pay sucha sum ifthe uncertainty isgenuinely removed; ifthere isa significant remaining risk thatthe insurer willfail to pay claimsthroughinsolvency, then the company isfailing the policyholder to the extent thatthe latter might decide that there is no value inpaying theextra cost. A large commercial policyholder will be abletospreadinsurance among a number of companies, making the solvency of individual companies a less critical factor thanfora small commercial firmwitha single insurer. Equally, thesecurity of individual reinsurers becomes less important if a broker cangetseveral signatures t ot h es l i p . In the case of a mutual,where the members are in effect both policyholders and shareholders, their interests fall somewhere between those setout in sections 2.2and 2.3. However, theyare likely to lean towards the latter since their prime concernwill be protection against insured risks and the remarksaboutdiversification of shareholdings do not come into play. 168 2.4 Regulatory Authorities A regulator’s primary aim islikely to be the protection of policyholders against therisk that insurance companies will notmeet their liabilities, or will notfulfil reasonable expectations, thoughinsome countries there is a secondary, and sometimesconflicting, aim of controlling pricing. The regulatory organisation may alsohave a role insponsoring the insurance industry of its country and hence in helping it to appear sound to prospective policyholders. There are parallels between these aims (protection and fair pricing) and the twin aims of the policyholder (security and value), thoughtheattitudes towards the use of the country’s protection fundwill diverge. There should, however, be a strong distinction between the aims of a regulator with regard to a company and the aims of the company’s managers.As discussed above, the latter will regard thepreservation of the company as their prime concern, and so will sometimesbe more conscious of their company’srelative, than of its absolute, position. In contrast, the regulatory authority should look mainly at the absolute position and shouldbe prepared to encourage or require companiesin difficulties tocease towrite business if such is the only means of avoiding an otherwise significant risk of future insolvency, even ifseveral other companies areinsimilar positions. Such culling isactually partof a healthy insurance market since, by reducing capacity and hence permitting the survivors to raise premium rates, it canbring abouttheend ofa soft market. Alsoit creates room for providers of fresh capital to move in to the market. There is a Darwinian theme to this, with the survivors being the companies with the fittest capital, marketing, underwriting, claims handling, investment management and ultimately management, The approaches used by supervisors to try to detect insolvency risks vary considerably. It is clear that supervision must combinean assessment ofa company’s reportedposition with a view as to its possible future 1 6 9 development. In Canada this will be done by requiring actuarial projections on various b a s e s . In UK the DTI is currently updating itsmethods of analysis of the Financial Year end position and isusing company visits to try to assess future plans, with the introduction fortwo typesof company of more frequent short visits to enhance this aim. The Institute of Actuaries i s promoting certification of reserves by actuaries or other suitably qualified persons, and should perhaps be advocating thefull Canadianapproach. 2.5 Other stakeholders 2.5.1 Third Parties A third party’s interest lies onlyin thesecurity of theinsurer protecting theparty against which the claim is made, since the cost of the cover is paid by the latter. Itwill normally have had no say in the choiceof insurer, so its interests are represented by the laws which make certain types of insurance compulsory, by any protection funds which will allow the claim to be paid shouldthe insurer fail, and by the supervisor. Politically itmight be difficult to pass laws enforcing certain typesof insurance if thecountry’s insurance industry is not seen to be sound. 2.5.2 Reinsurers Reinsurers shouldavoidcompanies which theyconsider suffer a serious risk of becoming insolvent. It is a poor use of management time to cultivate relationships with customers which have no long-term future, and itcan harm a reinsurer’s reputation to be associated with a direct insurer which becomes insolvent. A further feature isthat, in the eventof administration or liquidation, there will be a desire to come to termsas quickly as possible. While reinsurers sometimesbenefit initially from delaying payment of debts to 1 7 0 customers indifficulties, there may eventually be a significant speeding-up ofcashoutflow at the same time as one source of revenue has dried up. 2.5.3 Coinsurers/Market Ifan insurer fails, therepercussions can affect thewhole market.In the UK, this could be through the levies made by the Policyholders’ Protection Board whichguarantees 90% of all personal insurance claims and 100% of compulsory insurance claims. The nasty shock that even the Weavers business couldbe deemed to fall within PPB realms reinforces the value even to the stronger companiesof participating in a well regulated market. Where the claimsfall outside PPB scope,therecan be moral and commercial pressure on co-insurers to pickup the unprotected share. There is a suggestion, for example, that ILU members should meet claims made against Andrew Weir to maintain the reputation of that market. Lloyd’s, despite claims to the contrary, operates on a “mutual” basis in extremis through thecentral guarantee fundforthat veryreason. Trade Associations generally mightbe expected to want their good name to be linked withcompanies which remain solvent. 2.5.4 Intermediaries Security isbeingtakenverymuch more seriously now than even in the recent pastand it will tarnish a client relationship if expensive insurance policies turnoutto have been worthless. Intermediaries may need to be able to demonstrate that they have taken suitable precautions in recommending a particular insurance company should it fail in due course, iftheyareto avoidbeinginvited to make good the non-existent cover. 2.5.5 Rating Agencies These organisations endeavour to add value tothepublished accounts by analysis and direct contact with both the market and individual companies. By getting beneath thefigures and applying certain tests, they 171 attribute to insurance companiesa rangeof security ratings from those which are thought to represent very good security to those which are considered to be barely solvent. Through this feedback to insurers of the agencies’ reports, companiesmightbe expected to giveconsideration to balance sheetmanagement. Thiswill be both on a continuous basis as well as to address suchissues as “how much capital doesmy parent need toinject togo from BBB toA?” History showsthat these ratings have a veryshort shelf life - a significant proportion of the 1992 and 1993 UK failures and withdrawals had “A” ratings as recently as March 1992.(Eg English and American, Municipal Mutual and General, NW Re, Orion Insurance, Prudential (broking o n l y ) ) . Even where the ratings may be successful in warning of insecurity, t h i s may have somethingto do with self-fulfilling prophecies. English and American, forinstance, publicly claimedthat their withdrawal from the London Market in 1993 was directly attributable to their downgrading from “A” to “B”. They felt that they would only be shown second rate business which would be unprofitable and hinder rather than assist a recovery oftheir previously good security r a t i n g . 1 7 2 3 What isDynamic Solvency Testing? In the bibliography, we outline the current state of the Canadian approach[seesection 7.4].The intention here would appear mainly supervisory and to provide early warning that a company is heading for problems, ifnot insolvency, unless itchangestack.The natureisof a prospective audit overthe nest twelve months. While theintention of the process is to aid management and the tone of the guidance notes encourages the actuary to be part of the management team, his whistleblowing responsibility totheBoard mightputa strain on that relationship, What might be of greater value to the industry and aid better management of capital would be to incorporate dynamicsolvency testing intothe strategic planning process. We outline below the current perceived state of corporate planning and thensuggest how this could be enhanced by what is effectively scenario t e s t i n g . Efforts have been made to allocate capital to lines of business reflecting their particular qualities. Scenario testing, making allowance for these features, shouldhighlight any strain on the overall business and enable particularly capital intensive ventures or strategies to be identified. We could contrast thedirection proposed insection 5 to the development in reserving foroutstanding claims. We started there by producing point estimates that were often more reliable thaneither theclaims department or accountants couldmanage, particularly when inflation was rampant. This has developedinto statistical models of the claims settlement function whichenable us tounderstand thenature of anyvolatility within the process. Until now,planning hasfocused on single figures and it is often unclear a s to whether these are best estimates, optimistic targets or simply arithmetic extrapolations. The principles of dynamic solvency testing, 1 7 3 principally throughvarying the base planning assumptions about the future, provide thestarting point for a range. The Daykin and Hey work [seethe Bender et al paper referred to in section 7.1for summaries oftheir studies] initially assessed the variability of outcomes given break-up from a year end position. By extending trading for one year and repeating the exercise, theystarted down the pathof modelling various futures. Thisenables us to tryto answerthe question astowhether that extra yearhas made thecompany more orless risky.It might not,though, tell us so much about what could have occurred during theyearitself, Resources havenotthusfarenabled us to reassess this work to consider whetherit provides a stepping stone inthe direction we are heading for. One of the drawbacks, though, of any “black-box” technique isthat very few peopleinthe company will understand what goeson inside thebox. The transparency of deterministic scenario testing may prove of greater valueto management simply through its verycomprehensibility. Ifthey seeand accept theoutcomesofa given setof circumstances, theymay be more ready toact. So,whilewe couldconcentrate purely on thecontext of the supervisory perspective, we consider that a more fruitful initial paper might look at thefundamental management problem: “how do we manage the business and its resources -people and capital - tostay inbusiness forever?“. Any scenario testing needsto be firmly planted in the real world. We keep coming back to “management” - hence the title of this paper,Two management questions whichtheCanadians include are: The nature of management information available to detect changes in experience and 1 7 4 Management’s ability and willingness to implement changes inkey management policies. Put another way, does management know what is going on and would it do something about bad news if it noticed? 1 7 s 4 4.1 Where isCorporate Planing Today? Corporate Planning in General Itisgenerally regardedthatthe level and sophistication of corporate planning in General insurance companies fall behind that of other industries. This issummed up in the management consultant’s lament: “To run an insurance company you need to know a lot about insurance and that does not leave room for anything else.” The special complications of the insurance industry have given rise to additional difficulties to management theorists in applying conventional strategic a n a l y s i s . The theoretical development oftheapplication ofcorporate planning and corporate strategy developedas a separate topic in the 1960’s. This occurred withthepublication ofa series ofbooksmainly emanating from authors associated withtheHarvardBusiness school. Itwas backedup in the 1970’s by a substantial amount of empirical work predominantly relating to American companies. One such programme- the PIMS programme- comprises detailed information from over twenty-six hundredbusiness units and allowed researchers to analyse theimpactof market conditions on profitability, By theearly 1980’s thevast majority of thelargest US corporations had separate units responsible for strategic planning. Analysis of themarketsinwhicha company operates had always been a part of the strategic management process. This aspect of strategic management,however,gaineda major impetusfollowing the work of Michael Porter published in 1980.He proposed methodsofanalysing the industry inwhich a company operates and thecompany’s position within it. In particular competitive pressures were split into five sources (i) current competitors (ii) potential new entrants tothemarket(iii) product substitutes (iv) suppliers to the company including the supply of labour and (v)buyers.This method of analysis has formed a major part of subsequent strategic management approaches. Although there is little 176 agreement among thespecialists inthis area, on thedefinition of strategy for example,several toolshave been developed. These include the strategy model,theproduct-market/portfolio model and the risk/return model. Other topics in strategic management have included the appropriate organisational structure for different types of company and the development of thestrategy function itself. McKinsey hassuggested four stages ofstrategic management development:(i) company budgeting ( i i ) company forecasting (iii) analysis of the external market (iv) strategic management- including a well defined strategic framework and widespread strategic thinking c a p a b i l i t y , 4.2 Corporate Planning in General Insurance. A variety of planning techniques has been used and various company approaches have been documentedby theGeneva Association. Most of thedocumentedapproaches havebeen unsophisticated. There havebeen some exceptions, perhapsunsurprisingly mostlyin the United States where many insurance companieshave been using conventional strategic management methods. Survey datashowsthat intheearly 1980’s veryfew US companies were takingaccount of external influences on the company. Also,a survey in the UK in 1987 suggested only a few UK insurance companies usedconventional strategic management techniques. Turningto modelling, most surveys show a majority of companiesmake some use of modelling techniques. However,surveys bothin theUK and the US show a low use of modelling of overall insurance company a c t i v i t i e s . A previous GeneralInsurance StudyGroup working party (Akhurst et al 1988) looked at corporate planning using conventional strategic management techniques. However,this doesnotaddress the difficulty of integrating solvency testing models into conventional strategic management theory. 177 4.3 Diffrences between Insurance and other Industries One reason fortheslowdevelopment of corporate planning in insurance is themajordifferences betweeninsurance and other industries. As such, much of management theory has to be modified before being applied. Also, thecomplexity oftheinsurance business means that it is notan easy areaformanagement theorists to encompass. As stated earlier, unlike most otherindustrieswhere costs can be easily predicted butincome is more difficult toforecastininsurance notonlycan income be difficult to predict, but costs are also uncertain. To be consistent withthe approach in otherindustries income couldbe defined to include solely premium income. Then the prediction of costs includes not only the unknown claims experience, but alsoa reduction in costs reflecting the return on the invested premiums and any associated solvencymargin-which commonly includes equities. Despitethe additional uncertainty involved, the insurance business excluding the solvency margin can be analysed with the same methods usedfor other industries, The solvency margin, however, adds a particular difficulty. Once equityinvestment is considered, the application of corporate management theory tothis part ofthebusiness is more difficult, especially Porter’s theories on competitive positioning. Nevertheless, the importance of theperformance of the invested assets in determining the company’soverall returnmakes a suitable treatment necessary. The 1980’s provided a salutary lesson inthis regard. Itispossible to have a more efficient insurance business excluding a s s e t management, but still lower overall returns. Also, those firms that producehighasset returns may be encouraged to lowertheir premiums making other firms unprofitable. One approach is to consider the shareholders’ invested assets asan investment trust and adopt approaches used by the investment management industry withparticular regardto competitor positioning. The effective result would be that overall management of costswhich in the insurance case would include the reduction incosts from thereturn on invested assetswould pay regard to competitive pressures. Strategic cost analysis would thenbe setin terms 1 7 8 of theinvestment policies of competitors as wellas thecosts associated withpureinsurance s t r a t e g i e s . Objectives for Corporate Plans. 4.4 Setting Whilemanagement will need tobe aware of theneedsofvarious parties, its main aim will be satisfying theshareholders of a company. To this end itmay be useful when considering management objectives to consider directly how shareholders are likely to gauge company performance retrospectively. Insurance analysts commonly use dividend growthand share price performance compared withcompetitors as the main criteria for assessing historic performance. Usually risk adjustment isnotmade to the returns. This approach implies that not only is performance compared with competitors the main influence on the survival of a company but also one of the main criteria by which shareholders judge company performance. This has some common ground with policyholders’ reasonable expectations in Life assurance. If thisis accepted then thesetting of,forexample, return on capital in absolute termshaslimited use as a guidetosatisfying shareholders. One reason is that viewson acceptable returns are likely to changebecause of market circumstances. Some authors havesuggested that thecoreobjective ofa company should be to increase its appraisal value and that executive remuneration should be setinthose terms.One consequence of this isthat theadoption of a suitable risk discount ratebecomes crucial injudging performance both prospectively but alsoin the periodbetween valuations. However the benefits of, for example, lowerrisk may be lost on investors ina company. There are usually few objective criteria for establishing risk discount rates. Therefore, itmay be difficult in practice to refute claims that the incremental returns were the result of superior management insight rather thanincreased risk. Furthermore shareholders’ risk will depend on their own objectives, which may have little connection with the risk assessment usedintheusual appraisal value approach. The objectives for individual classes of business which are setneed to be consistent with 179 overall objectives. Here again, ifa competitor positioning approach is takenthen companieswill attemptto adopt new strategies that differ from their competitors to the extent thatthey are confident thatthe returns are above or below the average. Nevertheless, incremental activity will need to be considered by deciding the effect on theoutlook for thecompany asa whole. 4.5 RiskAssessment Many approaches have been suggested forrisk assessment some ofwhich havebeen referred to elsewhere in thepaper. An approach adopted by US regulators intheir attempts to setappropriate premium rate levels is toconsider theexposure tothenon-diversifiable risk ofa line of business. The argumentisthat shareholders inan insurance company should only be rewardedfor risk thatthey cannotdiversify away by havingother investments in their portfolios. One example of an insurance line with significant non-diversifiable risk is Mortgage IndemnityGuarantees (MIG). The exposure oftheperformance ofthis line togeneral economic conditions may be difficult todiversify away to theextent that most other enterprises are alsoexposedto general economic conditions to some extent. The estimation process forthenon-diversifiable risk of individual lines is extremely imprecise.When considering individual insurance companies the non-diversifiable risk (orBeta)isequally hard to assess. Thismakes risk assessment of different strategies basedon their forecast Betas problematical. Anotherapproach istoconsider theshareholders’ funds as an investment trustthat the policyholders have a right to in certain adverse circumstances. IFthis concept is adopted then it is in the shareholders’ interests to minimize their investment in the company. Thiswill reduce the extent of the policyholders’ right on their funds. In practice the financial strength of a company affects its ability todo business. Thismay be reflected in the profit margin itcan include in itspremiums,or alternatively the riskiness of itsassetmix. Thereforereductions in solvency do notnecessarily leadtohigher returns on capital. 180 There are otherrisk assessment processes that look from a regulatory standpoint. These include thecurrent solvency regulations, therisk based capital approach, and various stochastic models of dynamic solvency. While these are very useful in considering policyholder risk, some adaption is required if full allowance is to be made for competitor positioning asdescribed by Porter. 4.6 Consistent ApproachtoAssets and Liabilities One particular difficulty in corporate planning for General insurance companiesis in producingconsistent projections for the assets and liabilities. Factors suchas economic growthand inflation have an impact on boththeassets and liabilities. One specific exampleistheeffect of the UK recession on UK property. This had an impact on both asset portfolios and throughmortgageindemnity contracts the liabilities as well. Such factors are, however, likely to have an impactacross the full rangeof asset and liability classes. Although some work has been done on allowing forthe effect of inflation across the assets and liabilities in some stochastic models - theoretical work in this area is limited. Therefore itisusually necessary toadopta scenario approach - possibly based on specific periodsof history - when attempting to produce consistent scenarios ofthis type. 1 8 1 5 Integration ofSolvency and Planning A phrasewhich provides a link betweensolvency and planning is “the practical management of capital requirements in a dynamic framework“. Thissection isintended to be practical to the extent that it will include very little theory. It should provide some further answers to the question, “why plan?“. A barrier to planning ininsurance has been that we are here to protect against theunknown - that beingthecase, we cannot, by definition, plan! Such conservatism seems to missa fundamental point. If an insurance company cannot planits future witha reasonable degree ofcertainty then it has possibly become a concentrator of risk, whether through accepting excessive volumesof non-diversifiable risk or inappropriate reinsurance structures. Ifthe insurance company isproperly structured then it has transferred in the right mix of “unknowns” and converted them into an aggregation which is “knowable” within reasonable bounds. The integration of solvency testing withplanning will measure the extent to whichthecurrent strategy is vulnerable and hasconcentrated r i s k . Although thedynamicsolvency testing referred to insection 7.4 is largely supervisory, thegeneral approach does at least provide a framework for integration. The starting pointisthe base scenario which is generally taken to be the corporate plan. From there, a variety of alternative futures can be considered and their impacton financial viability can be examined.The extent to which the company is dependent on a specific future ornarrowrange offutures can then be determined. Thisprocess can be relatively simple withone variable being considered at a time or complications incorporated. At one level, thesewould be where there could be interactions betweenvariables. A classic exampleis where an insurer is providing earthquake coveron properties inwhichit also holds significant investments. 1 8 2 Also of particular interest to this working party is where these relationships are direct as with mortgageindemnity insurance and the property portfolio. The same harsh economic environmentof high interest rates and growing unemploymentwhichreduced property values on theasset sideof the balance sheet, alsotriggered MIG coverage on 100% mortgages issued in theheady property boom of the 1980’s. This near-disaster illustrates the danger of failing to address scenarios imaginatively and without deep understanding of the verynature of the business and its d r i v e r s . The nextlevel iswhere thescenarios are embedded in the market place and thecompany considers the actions of its competitors and potential new entrants. It also needstothink through theimplications ofa series of smallchanges which might eventually result in structural change. As a parallel, thetechnology of thefaxmachinehad existed fordecades before itsuddenly become genuinely popular.What were the features which brought it overthe threshold from the arcane worldof newspapers into everyday usage developments which appear to be ? Are theresimilar ticking over but which could soon erupt into daily use and drastically change our way of doing business? 5.1 Integrity oftheBase Plan In our experience, thestandard of corporate planning varies significantly throughout the industry. This is not just a comment on the final documentation buton theprocess itself and how closely related t or e a l i t y it is or is expected to be. The plan may be the mechanism by which strategic decisions aretakenor reflected; it couldbe a forecast produced by a staff department insplendid isolation; it might be for the directors aloneor widely circulated. The outputcouldbe thousands of detailed numbersatone extreme or a few visionary words at the other. Anotherimportant question istheaccuracy of theplanitself as a model ofcompany behaviour. Thisis nota reflection of theuncertainties within thedataitself, so much as how well constructed theplanning model is. In 183 other words, if all theassumptions actually heldtrue, would thebalance sheet and profit and loss account actually be thesame asthat appearing in theplan? Ifitisto have value, theplan needsto reflect what the business really wants to achieve and considers to be possible. The planning process shouldtherefore involve thosewho will make ithappen so theyknow their partin itand believe thatitisachievable. Itneedsto be as up to dateas possible. The figures should extendto balance sheet itemsand any other aspects which are considered significant from a solvency perspective. In other words, the components of the plan and the process usedto produceit, need to relate to those aspects of thebusiness which havean effect on solvency. For example, there seems tobe good evidence that exceptional growth is quite highly correlated withsubsequent financial weakness[section 2.2] Within theplanning process itself, any aggressive growthplans shouldbe carefully checkedthrough to examinewhethertheyshare thefeatures of observed failures. Does management understand the business? Isthis a new area? What controls areinplace toensure that poorquality business will be keptatbay?Thus,some atleast of thelessons of thepast can be incorporated within the planning process itself. Once the plan is complete, a Canadian-style DST can be carried out. We have thought longand hard aboutthe nature of the testing - “what ifs?” vs “Monte Carlo”. None of us isnaturally drawn to deterministic approaches because of their arbitrariness and lack ofvigour but theyare probably theonly practical way toapproach a variety ofscenarios. Thisis becauseof the current state in the development of stochastic models theyare notconsidered accurate enough fortheuserto be surethat the outcomeisnotdue to thecrudeness of themodel rather than a potential event. There issome progress beingmade inbuilding stochastic models of some elements but we are a longway from filling thegaps and then integrating the components. 1 8 4 One of thedrawbacks of deterministic models is the tendency to confine theviewtoa relatively narrowrange aboutthebest estimate. One way to introduce some .stochastic element would be to express some of thetests in termsof the underlying variability, thestandard deviation of the item in question. Itistheunexpected that causes strain buta 20% rangeforthe valueof cashwould have a verydifferent level of validity than20% for equities and thestandard deviation basis recognises t h i s . 5.2 How WillManagement RidetheDownturn? The literature, experience and common sense suggest that thekey tolong term solvency in the majority of companiesisthe quality of the current management.Even those which have inherited insoluble problemsfrom the pasthave some responsibility to recognise thatsoonerrather than later.Testing the plan in a variety of outcomes willidentify those circumstances which placethe company under the greatest stress. The baseplanmay or may notrecognise theinsurance cycle but most certainly thescenario testing should do so.Therefore, although even an unrealistic planmightbe accepted asthebasescenario, thenature of its detachment from reality should become apparent through thedynamictesting process itself. The Canadianapproach, by considering specific scenarios rather thansimply claims “plus a bit”, doeshelptoachieve t h i s . Thisprocess should helpmanagement todevelop its strategies rather than simply tryto trip them up. For example, although the capital structure may appearperfectly adequate in the basescenario, what happensto it after theimpact of thenextprice war? Would management do better to maintain marketshare or keep its prices high? How would these answers varyif we lookatindividual territories and/orclasses? Do certain areas emerge as more capital hungrythanothers, perhapsindicative of nondiversifiable or under-diversified risk? Does management have some convincing answers tothese questions? Can theyadapt tothedynamics of thebusiness? Can it build defensive alliances withthedistribution chain and end customers to increase entry costs forinvaders and protect the existing business without having toresort toprice cuts? 185 5.3 Capital Implications ofPlan Much thinking on solvency has been focused on the relative position. There issome validity in this. For example, the whole UK insurance industry cannot, politically, all be forced to stoptrading. The same must be true ofthebanks; intheevent ofa crisis, blind eyes are turned or rules re-written, theGovernment has to do something. In an increasingly global market, this focus on the local market rather thantheworldwide marketbegins tolookmyopic.While theUK market frittered away its capital in the yearsfollowing 1988,each company consoled itself withthethought that its competitors were doingmuch the same. While some of the weaker insurers may have been eradicated, the remaining market has been left open to the predators - new capital unencumberedby thescars of therecent past. Thisis particularly true of reinsurance where millions of dollars have poured intoBermuda and some new or revitalised subsidiaries of foreign companies havearrived in London. So eachcompany needstoconsider absolute strength and, when checking on its relative strength, ensure that thereal peer groupis being evaluated. Italso means that thecompany needstobe considering its strategies and potential strategies and thepossible impact on its own capital. 5.4 Has Management theAbility toSeeit Through? Financial theory has shown that there is no premium for accepting diversifiable risk, hence the costof reducing variability of results from such business impingeson profitability. If the shareholder has a diversified portfolio, he should prefer therisk to be retained rather than seeprofit diluted through reinsurance. However, there are other interests, notleast those of thepolicyholder, tobe considered. As we have already 1 8 6 noticed, thereisalsoself-interest. Provided thatthis isenlightened, it may he healthy. A competent managershould havesome notion ofbeing abletooffer the marketsomethingunique, something thatwill be valuedby at least a segment of that market. He can only continue to offer that service/product if hisemployer remains inbusiness, hencean incentive to protect thesolvency of thecompany. Given a true and therefore valued competitive advantage, profitability should be assured. It is profitability that is theprerequisite ofsustained solvency. It is far tooeasytoforget that insurance is a business and,tothrive, needs tohavesomething attractive to offer its customers inreturn for the price they are prepared to pay. As a generality, middling companies trying to sell at middling prices have returns on capital so low that they cannot sustain themselves.It is likely that this is also true of insurance companies so,has the company a clear competitive strategy? I si tw e l l understood and consistently implemented?Isit trying to mix strategies, tobe all things toall people? Thereis widespread agreementthat thequality and style of management matters in considering the riskiness and future solvency of an insurance company. Determining whetherthe management isappropriate is also commonly accepted to be notoriously difficult. How can we judge them before we find out thehardway? At a seminar organised by the Society of Fellowsof the CharteredInsurance Institute in May 1993 on the subject of insurance company solvency, a number of suggested questions emerged. [Journal of the Society of Fellows, Vol 8 part1 July1993] A selection f o l l o w s : - Ismanagement too“racy”, tooconcerned withits own comforts and public image? 1 8 7 Ifthecompany seems to be theonlyone with the right answers, how come they are not being copied? What is their underwriting philosophy? Is it well known? Isits internal structure efficient and appropriate? These and many others need to be incorporated in any assessment of whether the company will respond to emergent problems in such a way as to see them through. 5.5 External Factors analysis to try In traditional planning, companies will carry out a SWOT to identify thosefactors which couldadversely affect its fortunes. For example,it may consider changesin regulations - eg EC Directives concerning discounting and thefreedomtodo business - and in the wider environment:technological, political, social, financial and economic aspects. These areasare obviouscandidates forscenario testing. What would happen ifone or more of the identified threats materialised? The company may already havecontingency plans inplace butthis process w i l l help to identify priority areaswhere the survival of the company is potentially at risk unlessstrategy is changed to reflect the new circumstances. Withinthe scenario testing, any contingency plans could be tested for validity, thus aiding their refinement. 5.6 Planning Outputs If the planning process iseffectively extended to incorporate scenario testing, itwould be helpful if the content of the planning documents reflected thenature of thetesting. Those ratios which are considered to be reflective of financial strength shouldcertainly be generated. In addition to those considered important internally, window dressing i sa l s o 188 a feature and outputcouldinclude pro-formas of the various security agencies’ ratio tests. Itisalso essential that trends ata lowerlevel areidentified - targets f o r people attheoperating level. What rate increases, how much cash to be collected, thevalue atwhichclaims can be settled, whether income should be allocated towards equities, cash or fixed interest etc? Qualitative targets areveryhardtodevelop butshould be worked on. Certain types of business are likely to holdgreater appealthanothers; new business targets can reflect this. Settlement targets may indicate productivity but, when combined with targets for minimising the number of claims reopenedand values for settlements, canalso indicate effectiveness. 5.7 Competitive Pressures/Pulls Everycompany is pronetocompetitive pressures. Each one only prospers if itoffers valueformoney ata price that exceeds its costs. Generally, a company can adoptone of two strategies: it can try to avoid competition by targeting a specialist area thatnobody elseis interested in and developing sufficient expertise that others are daunted by the “learning curve”. Consider, for example, Westminster Motor Insurance Association Ltd. In the 5 years to 1991, its underwriting profit averaged 27% of net premium and was insurplus forevery one ofthose years. By contrast, the company market’s underwriting loss formotorinsurance averaged 13% of netpremium and nevershowed a surplus! However,Westminster writes blackcab business and has virtually cornered this market, The niche strategy isgenerally of appealto thesmaller companies but there is no reason why larger insurers should notoperate ina whole series of niches. That is Zurich’s stated global plan. The secondstrategy is to stand out from the crowd, whether by offering something special whichnobodyelse can match or by beingcheaper fora given product/service level. Driving down costs and thusprices tends to be of limited longterm success becauseitisso easyto copy. Making customers overly price conscious may not be in the supplier’s interests 189 longer term - ifthey onlyeversee value in termsof cheapness, how can thesupplier everhe rewarded foradding value? Providing a unique- or atleast different - and people driven service is harder tocopy.People are nota formula: togenerate and sustain theright culture is such a huge task that few try and fewer still seeit through. The planning process should reflect the validity and viability of the strategy beingadoptedand the likely responses from the competition. What mechanisms and desires areinplace tostrive constantly to develop theadvantage so that thecompany remains atleast one step ahead of its rivals inits chosen arena? A further key issue istheextent to which management will take account of thescenarios inits strategy selection. The whole point of testing the Planis torefine it. The true test ofanything, though, is toputit into effect and this isthe test of time. The control loop of monitoring actual outcomes against thoseexpected should result in further refinement not justof forecasts but,more importantly, of activities to optimise the company’s position. Therefore, it isvital that both theplanning process and thescenario testing areintegrated into thedecision making process if it is tomake a difference to performance. We can bringthis closer to home: the essential feature of insurance business is that the product is sold before knowing its cost. The plan, whichmay include obtaining certain business volume and/or profitability targets will have takeninto account the pastclaims experience and the market environment. On subsequently writing business i t : 5.7.1 may notbe able togetbusiness atthedesired rate 5.7.2 may getbusiness butsubsequently find its rates havebeen toolow 5.7.3 may obtain thebusiness it expected, on profitable terms. 190 For all cases it is vital that theCompany estimate its true cost as soon as possible. A monitoring system could incorporate a dynamicmodel of the office withits business targets and expected claims, investments, expenses etcand withthe assumptions beingregularly up-dated according to the emerging experience. The difficult cases are5.7.1 and 5.7.2 and highlight themain problemwith theindustry: that itisplaguedby “Bad Competitors” la aPorter. Many companiesinsist on doingbusiness at uneconomicrates, either through deliberate policy to“shake outthemarket” or lack ofunderstanding ofthe business. The logical coursefor an office in this situation is not to competeinunprofitable lines, thussharing thelosses and prolonging the situation. If it truly believes, and this is thecrux ofthecase, that the rates are too low,then itmust withdraw.Ifthe “good” offices followed t h i s course then losses sustained would be accentuated and drive the others out sooner. Itisnotusually practical wholly to withdraw, hoping to re-enter the class infuture years. Itis also difficult tobe sure oftheloss makingposition as there will be uncertainty on claims and investment returns. However,the company needs to have a clearstrategy; can it discern and enter profitable niches inthemarketand doesit havetheresources to enable it do this, financial and technical. If it remains in the loss making market, how long will it continue? What sort of losses does it think its competitors cansustain? What needs to happen before it stops? 5.8 Integration into a Market Model One of the achievements of MichaelPorter was to providea way of looking at the market which aidedcompanies to developtheir strategy. This model isgenerally applicable and can be used forinsurance - we would be unwiseto consider thatour business isso different thatwe cannot apply generalities t oi t . 1 9 1 However,thereare differences and this isrecognised in the concessions made inits accounting treatment and need forsupervision throughout the world. Through a better understanding of how individual insurance companies work and what marks outthesuccessful from theloss making, thesurvivors from thewithdrawals, we may be ableto suggest the most important factors whichshould be examined intheplanning process. 192 6 6.1 Modelling theCompany in a Market Context General Comments Models arenotoriously tricky to build, so why bother?The objective we proposeissufficiently worthwhile as to make at least some preliminary work justifiable: to aid the practical management of the capital requirements in a dynamic framework. Specifically, to be able to help answerthequestion, if themanagement doesX, will thecompany still be here and in what shape? A secondary objective is to explain how insurance companieswork - the business beyond the investment t r u s t . The target audience in this casewould be investment analysts who often regard insurance companies as responsible for managing the investment of the freereserves with a volatile gearing element - the insurance business itself whichgenerates huge losses and,occasionally, some profit. We would hope that we could show that there is some value in the core business itself but the industry has not been a good advocate of its own cause lately. Hence the primary objective! First ofall, though, we need todefine themarket.Inthefamiliar worldof private motor insurance, do we mean the sum total of all private motor policies sold, all private motorists, all policies sold through a particular distribution channel, or a more specific and narrow grouping? In different circumstances, we probably mean any one of these things. We have,though,to start somewhere and a practical constraint is the availability ofdata.So we probably mean thefirst and would turnto the DTI returns and Lloyd’s data. We can collate the figures from individual participants and sum them to obtainmarket figures. Although it is tempting to ignore the smaller companies, privately available datashowssignificantly stronger growth for mostclasses among those companies outside the top ten over the past five years or so. These more aggressive companies are often those which are 1 9 3 considered to affect profit prospects for the market as a whole - a thesis whichwe wouldhope totest through themodel building process. Givenan aggregation, we can thenanalyse key datato consider whether there are correlations between individual company and market performance and whether this is on a sustained basis. Do companies “pop up”, make a nuisance of themselves and disappear again? What are the features of those who join and survive in finding ? ifwe have any success some parameters to build a model,we couldattempt to project into the future. There are at least threeways to do this. The first is demand based-just how many private cars will there be? What will be the impact oftheincreased taxcharges on company cars?... The second is to look at theaggregate marketas it currently exists. What can its current capital and profitability sustain?... The third would be toconsider each company on its own, basedon track records and any clues as to future intentions that can be gleaned. The extentto which the threeapproaches convergeor diverge may indicate thenature ofthedynamics of themarketand whether prices will tendtorise or fall and when this might be expected. One way to test whetherthis exercise isof any value would be to tryto predict the previous five years basedon what we knew five or six years ago. Given thatthe business environment appears to be moving at an ever increasing pace, if we could add little or nothing tothelast five years, thelikelihood ofdoingso forthefuture must be low. This would, though, give us some indication as towhether insurance cycles are predictable. ( I t isprobably reasonable to assume that inflated asset values generated by bullish stockmarketswill tend to exacerbate competitive pressures as companies attempt to maintain their return on capital through growing marketshare. As they have allsharedin the same good fortune, t h i s common strategy can have only one outcome!). 194 Any model building shouldbe as simpleas possible - forexample, we would start by keeping asset and liability risks separate. For individual company modelling, we would need to incorporate the UK solvency hurdle so,if theyfailed, theywould withdraw. The comparison between the market as a whole and the sum of individual companies might be indicative of thepressure fornew capital or the need to shed excessive c a p i t a l . Given thata model can be developed, we could extend its value by varying an individual company’s strategy. We havemade a start by compiling some dataand looking at outstanding claims provisions unearned premium reserve incurred loss ratios paidloss r a t i o s acquisition cost ratios market share. ) ) as% netwritten premiums ) ) ) absolute/relative tomarket However,we wouldappreciate broader input and thoughts before seeking toconsider thenature ofthemodel. Isit worthwhile? Are we on the right direction? Isthemarkettoocomplicated and unknowable to be modelled successfully? 1 9 5 6.2 Comments on Stochastic Modelling We have already briefly discussed the nature of appropriate modelling: deterministic compared with stochastic. Both approaches have their champions aswell astheir virtues. The key issues are that models need to be sufficiently accurate to present valid outcomes and that management responds positively to thosemodels. Model building should focus on causative elementslikeinflation, accident levels, number of vehicles, growthin theeconomy,theperformance of assets, weatherpatterns and so on. Inthis way,managementwill come tounderstand something of the dependenceof the company’sperformance on theseexternal events, except to the extent that they add value to the company by taking appropriate counter measures. The events will occur come what may: in a givenscenario, some companieswill fail whileothersthrive and the difference betweenthetwo must be largely down tothestrategies carried out. Stochastic modelling is but one tool to aid management to develop those strategies. Below we indicate an alternative approach. However, any outcomeshould be tested against two benchmarks. Are both the scenario and the outcome intuitively valid? Some scenarios could be mutually exclusive and thusbe indicative of a weakness in the model. Equally, a poor outcome might ignore some of the control mechanisms in the company. Iffailure of the company results from failure to react to the indicators, that would layemphasis on a rapid and right response to that particular adverse trend and hencetheimportance of those controls. We see this development as a longprocess but one which will shed light on the company as a whole in a similar way to the revelations which stochastic claims analysis hasbrought. These methods have not replaced judgement or deterministic modelling but more probably improvedthem both. We anticipate a similar trendin the modelling of companiesas whole entities. 1 9 6 6.3 Scenario Building “You cansee alot by observing” - Yogi Berra. We havealready intimated that features which can have the most serious and widespread adverse effects are those which represent “discontinuities”. Sometimes, they can be arbitrary and appear as if from nowhere. In othercases, the result of a series of visible incremental changes can be toalter a marketindramatic and unexpected ways. What links these two outcomesinthemajority of cases is that the outcome was largely pre-determined.The eventswhich eventually led to the discontinuity were already inplace, if only we knew it! Far toooften our 20-20 vision is most apparent when we view our market withhindsight. The causes of the MIG losses are now well understood butwere notintimeformostunderwriters totaketimely and appropriate action. Had they seen what was there to be seen, they would have found huge MIG losses were a foregone conclusion because of these events (ie 100% mortgages, theLawson boom, notice of withdrawal of 2x mortgage relief inAugust 1988, control of general inflation unless ...) they changed their underwriting stance. Understanding the key determinants at a giventime foreach class and thenaggregating them may reveal to companies how to keep in balance becausemany of thosedeterminants will cross classes. For example, interest rates, trends in the law, the price of property and the EDI (Electronic Data Interchange) installed base could all impact a number of classes. Inother words, modelling should be basedon inputs rather than outputs, the causes of premium growth rather than premium forecasts from underwriting departments. There isoften a lagbetweenchanges intheoriginal eventand its impact on insurance so timing thediscontinuity isnevergoingto be easy, Or important: what does matter isto identify thecausative events and their 1 9 7 likely consequences. When considering strategies, alternative viewson timing can be considered and balanced against a “donothing/head inthe sand” scenario. There will almostalwaysbe a costto changeand,by considering thevarious scenarios, thecompany can compare this to the possible costs of missing an opportunity or, say, meeting an avalanche of c l a i m s . The point of modelling isnot to be ableto predict that on the23 June 1998company X will be insolvent. Analysis oftheexternal factors should reveal the natureof the forces which are takingthe industry in a particular direction. Modelling should helpus tounderstand the business better and so be able to manage the company so it produces a growing and robust stream of profits. 1 9 8 7 7.1 Select Bibliography A Synopsis and Analysis ofResearch on Surplus Requirements f o r Property and Casualty Insurance Companies A Brender, R Brown, H Panjer Institute of Insurance and PensionResearch, Waterloo Ontario, 1992. Thispaperrepresents an excellent starting point for anyone interested in the subject of solvency. A “Readers’ Digest” of books and papers, it guides the reader throughthe material. The papersarepresented with brief summariesboth in alphabetical orderby (first) authorand also by subject. The eleven subject headings areclassical risk theory, projection simulation models, financial economics, loss reserving, statistical methods, regulation. financial reporting and surplus management,life insurance, investment models, rate making, and “general”. From thepaper, relying on thesummaries alone, we refer to three papers whichsuggest key factors at least correlated to if notdeterminants of the financial health of insurance companies:SALZMANN, RE “RLS Yardsticks to Identify Financial Weakness”. Proceedings oftheCasualty Actuarial Society 68 (1951) 172-194 Salzmann’s Yardsticks:• • l • • l reserve l e v e l surplus l e v e l quality of assets operating results excessive growth reinsurance protection Model for • liquidity TRIESCHMANN, JS and PINCHES, GE ” A Multivariate Insurance 40 (1973) 327-338 Trieschmann and Pinches’ Ratios:- Predicting Financially Distressed PL Insurers”. Journalof Risk and 1 9 9 • agents balance : total assets • stocks cost :stocks market value • bonds cost : bonds market value loss adjustment underwriting & : net premium • • combined ratio premiumswritten direct :s u r p l u s • TAYLOR, GC and BUCHANAN, RA “The Management of Solvency”. International Conference on Insurance Solvency I ,1 9 8 6 . Taylor and Buchanan’s Variables:• • • • • ratio of risk premium tovalue of liabilities (exposure to insolvency due tofuture claims fluctuation and asset fluctuation) estimated v a l u eo fl i a b i l i t i e s expected future rate ofincrease and variability ofunit asset values number ofclaims (proxy for size of portfolio) contribution torisk ofeachline of business Theseelements wouldappeartobe worthconsidering in any investigation into thefinancial health ofan insurance company and should play a key part intheday today management if that health is to be sustained. 7.2 Strategic Financial Management in a General Insurance Company JWE Mariathasan, PF Rains ThirdAFIR Collquium 1993 The material in this paper has already been summarised in section 4. Ratherlike the Cummins paper- see section 7.6- this acts as a beacon towards where tomorrow’s advances shouldoccurand encourages us in that direction. The authors start by quoting Nietzsche, “To make plans and project designs brings withit many good sensations; and whoeverhad thestrength to be nothing but a forger of plans his whole life longwould be a veryhappy man: but he would occasionally have to takea rest from this activity by carrying outa plan- and thencomes thevexation and the sobering u p ” . 200 7.3 A New Look at Evaluating the Financial Condition of Properly and Casualty Insurance and Reinsurance Companies TM Redman, CE Scudellari CAS 1992 Thispaper iswritten from an American perspective and so some of its conclusions may not bewidely applicable. The essential conclusions are thathistoric information relating to causes of insolvency - such as that included in section 2.2- can be incorporated intosolvency evaluation reviews of insurance companies. Thiswould seem toprovide a solid base for the considerations which should apply in any Dynamic Solvency Testing. Havinganalysed thefailures from 1969to 1990, they make some interesting observations including Insolvency by policyholder’s surplus:Surplus < $5m > $5m < $50m > $50m Proportion of Insolvencies 63% 34% 3% But, thegreatest frequency ofinsolvencies came among the middle group. Stockcompaniesmade up 75% of all insolvencies although representing (Ie Mutuals were more secure than stock only 50% of allinsurers. companies). Roughly 50% ofinsolvencies occurred in companies 15 years old or less. 81% of insolvencies involved premium growth of more than 25% or a decline ofmore than5% The lessons thewriters draw from this 22 yearhistory include:a) insurance isa commodity- withprice so crucial, companies need the strength tosurvive thestress ofbothhard and soft markets. Where we are 201 in cycle may have an impact on the level of resources expecting. we should be b) company management may be critical but its quality is hard to evaluate. c) rapid growth means additional in new and unfamiliar areas. exposure which may be underpriced or d) company characteristics including ownership should be considered. e) weaknesses in past rating agency ratings can be a guide particularly such areas as overstated assets and unprotected catastrophe exposure. f) the future will not follow the past precisely - new issues will emerge Based on their analysis, the authors suggest certain industry norms:in i) loss reserves to policyholders’ surplus property 50% - 150% casualty 200% - 300% ii) reserve development to policyholders’ surplus < 25% iii) net leverage property 250% - 400% casualty 500% - 580% [(net premiums + net liabilities)/policyholders’ surplus] iv) gross leverage [the sum of net leverage and ceded reinsurance leverage] property 300% - 500% casualty 500% - 700% 202 Havingcarried outthese tests, thepaperexpects thecompany to focus on those areas presenting thegreatest exposure. In other words, to actas a catalyst and starting point for management action. 7.3 Standardof Practice on Dynamic Solvency Testing for Property and Casualty Insurance Companies Discussion Draft Canadian Institute ofActuaries, May 1993 The current draft was circulated to members of theCanadianInstitute of Actuaries. Itattempts todefine good actuarial practice inDST butdoes not actually require thatitshould be carried out,nor does itguideon when thatmightbe a good idea.Simply, ifaskedto carry out DST, the actuary would be expected to follow these guidelines. Theirtimetable is looking for finalisation by the end of 1993 so we shouldexpectan exposure draft tobe available aroundconference time. At present, itisnot entirely clear whetherDST isa requirement of the CanadianSupervisor or a professional issue.Rather than tryhere to second-guess theCanadians, it is suggested that we keep a watching brief on their approach and judgeits efficacy over time. Clarity and consistency betweenthesupervisor, professionals and trade bodies would appearto be a requisite of a smooth introduction of any changeso significant. (It isinteresting toobserve, though, that Canada is alsotrying to define“provisions for adverse deviations” in respect of claims development, reinsurance recoverable and interest rates. Thisisa subject in itself but itappears to be their intention that thesePFAD’s wouldbe included inthebasescenario for DST -seesection 7 . 5 . ) The general approach isto start with a base scenario - which would usually be thebusiness plan- and thentest theresilience ofthecompany in a variety of unpleasant circumstances. The draft guidance notesare practical and realistic eg “Since the base scenario may be seen by the boardas a most likely scenario, discrepancies withtheplanmay put the actuaryat odds with management, thus jeopardizing the effective performance oftheactuary’s role” and “The report totheboardshould be 2 0 3 an interpretive report, not a report brimming with statistics for each scenario that was tested.” The current practice notes include a basic list of thescenarios to be tested in relation to their impacton policy liabilities and other balance sheet items. It is not intended to be exhaustive and omits changes in the external environment:I Reinsurance Programme The occurrence of multiple catastrophic losses in a givenyear including the cost of any reinstatement. The occurrence of one catastrophic loss equal to the probable maximum catastrophic loss given theexposure of the company. A significant increase inthefrequency ofindividual large l o s s e s . The default r i s k . II Loss Ratio Claimfrequency and severity - what if worse than planned? RateAdequacy-what if planned rate increases not implemented? III Investment Significant changeininterest r a t e s Significant adverse changes ininvestment values Liquidity ofassets tomeet cashflow requirements of other adverse scenarios IV Expense Level and Volume Business levels below expectations Rapidgrowth inexcess of the base scenario 204 V Unpaid Claim Liabilities Impact ofadverse deviation from expected values 7.5 Provision for Adverse Deviations Exposure Draft CanadianInstitute ofActuaries, May 1993 Thispaperrepresents an interesting development. Many actuaries have talked aboutthesubject of PFAD’s but the Canadians have dared to set out not onlythe features which mightrequire such a provision but also their scale. The paperis succinct so theinterested reader is directed t oi t hutwe can give a flavour. Itis split into 6 sections: background, overview oftheapproach, definitions, and thenthevariables: claims development, reinsurance recovery and interest rate. For each of thevariables, t h e r ei s a list of the considerations, a description of “high margin” and “low margin”situations and the suggested upper and lower levels of these margins. For example,in considering company claimshandlingpractice, t h e features of a low marginsituation are listed as:“stable claims handling environment, few significant changes in claims staff and handling procedures, no majorsystems changes, and casereserves established i na consistent and responsive manner.” In selecting a margin for the variability in claimsdevelopment, the guidance includes the following: “Ifthe claimsare discounted, an appropriate margin needs to be considered forpayoutpattern. Ifclaims are not discounted, the margin should be reduced appropriately. “The member should be guided by thefollowing range: Low MarginFactor High MarginFactor 0% 15% 2 0 5 “When two or more of thesignificant considerations exist, the member should useatleast theaverage of the high and low situations.” Chris Daykin- theGovernmentActuary - haswritten a brief commentary on the exposure draft as follows: ” A study by the Insurance Bureau of Canada has suggested thatdiscounted reserves plus the PFAD could come out on average3% below undiscounted reserves. However, this clearly depends on thetype ofbusiness and theextent of the PFAD. “ThePFAD should be a fair risk charge for uncertainty. However, unlike thecorresponding discussions intheUS, it seemstobe acknowledged that uncertainty itself demands a PFAD and that there is no need to classify risk assystematic or diversifiable. “TheCanadianInstitute of Actuaries’ viewisthat thediscounted reserve, together withthe PFAD, couldsoon be expected regularly to exceed the undiscounted reserve, having regard tofalling rates o fi n f l a t i o n ” . These two Canadianpapershave a common purposein improving the solvency ofgeneral insurance companies. DST looks a little way into the future while PFAD recognises that some numbersare more reliable than others. One interesting common feature isthe need to understand the dynamics of the business both instatistical terms as well as the “softer” elements. DST considers the risk involved in pursuing the current strategy while PFAD reflects theimpact ofpast business decisions. I ti st o be expected that there will, overtime, be a blurring of these distinctions. By reacting to the outcome of DST, management should find itself tending towards the“lowmarginsituations” of PFAD as itrecognises t h i s cost element inits business decisions. A further general question about the two papers is the wisdom of producing guidance notes that havetheappearance of manuals. Is there a danger that practitioners will come toregard this guidance as the ceiling 2 0 6 rather than the floor of their explorations? As well as the theory, a worthwhile discussion could be on themeans ofimplementing i t . 7.6 Statistical and Financial Models of Insurance Pricing and the Insurance Firm J D Cummins The Journal ofRiskand Insurance, 1991 Thisis a fascinating paperwhichismore of a signpost thanjourney’s end. Cummins endeavours to reconcile theactuarial and financial models of insurance. There issome veryhelpful background material on both the statistical and financial models; particularly for those of us with only limited understanding of financial theory, this paper is accessible. Financial theory gives useful insights that may be missing from purely statistical models. For example, using theCapital AssetPricing Model,theequilibrium rate ofreturn on any asset i s : where ri= rf= theexpected return o on f asset i therisk-free rate interest thesystematic risk coefficient or betaofasset i = Cov(ri,rm)/Var( rm) rm = theexpected return on themarketportfolio portfolio ßi = The CAPM model implies thatinvestors will be rewarded for bearing systematic or betarisk but not fortaking unsystematic risk because this risk can be diversified by properly structuring theportfolio. But this is onlythestarting point! From hereCummins shows theinsurance CAPM and points to some of the impractical assumptions which limit the viability of some of the conclusions thatare oftendrawn. One hope isthatintegration might improve the quality of the financial modelling which tends to make oversimplistic assumptions as withtheclaim distribution functions. The 207 current reliance on log-normal distributions may understate the riskiness ofinsurance ascurrently presented inthefinancial l i t e r a t u r e . Within CAPM work, there is no recognition of the lag between a particular opportunity arising and themarketadjusting to the equilibrium. An examplehere isthe 1993 hiatus in the reinsurance market. At the moment there is most certainly - atleast in the eyes of the “Storm Rating inthe90’s” working party - a premium forcatastrophe even though it is a diversifiable risk. As new capital flows into the market, we might expect the predictions of financial modelling to come good. In other words, CAPM may tell us abouttheequilibrium position but not be too helpful about what happens on theswitchback that will take us there. 2 0 8