Contracts and information in the firm

Our fictional story of person A’s building project was designed to highlight some of the difficulties everyone encounters  at some time or other in the process of contracting. It is now necessary to look at the issues involved from a more analytical viewpoint. Perhaps the most important point about the hapless A is that  his problems  all derive from various  forms of infor- mation deficiency. Were information costlessly available  and  all transac- tions costlessly enforceable  most of his worries would be over. Consider now the various points at which A confronts  the problem of his own igno- rance.

1.    Adverse Selection or ‘Hidden Information’

1.1    Examples of adverse selection

The first problem  was that  A did not know the location,  skills or reliabil- ity of the tradespeople he required.  Finding  out this type of information requires time spent in search. More consideration will be given to the ques- tion of search in Chapter 3, but it should  be evident from the discussion in Chapter 1 that  to search exhaustively  – that  is to search until informa- tion is complete – would be to do nothing  else. At some stage the costs of further  search  in terms  of  the perceived opportunities forgone  will out- weigh the benefits in terms of the expected new opportunities potentially discoverable.  Further, certain types of ignorance are in their nature  costly to dispel through search.  Ignorance of  the price of  a very well-defined product can be mitigated  by asking for quotes from an increasing number of sellers, but overcoming  ignorance  of product quality  is more difficult. It may be worth paying more for the services of a more skilled person, but how is person A to tell a good and reliable craftsperson from a poor  one ex ante? Clearly, this problem ultimately derives from the difficulty of pre- cisely specifying what  services are  required in a contract. If  a contract between person A and a craftsperson were clearly and unambiguously specified the ‘reliability’ or ‘skill’ of  that  person  would  not  be an issue. Either the provisions  of the contract are fulfilled and the craftsperson demonstrates sufficient skill or they are not fulfilled in which case the absence of sufficient skill results in a specified penalty.  A skilled and reli- able craftsperson is valuable to person A because he may not have the tech- nical  knowledge  required  to  specify precisely what  is to  be done.  The absence of this knowledge, however, will make it difficult for A to discover the credentials  of his potential workforce.  All craftspeople, skilled or oth- erwise, will have an incentive to overstate  their expertise in the process of negotiation, and  the  truly  skilled will have  difficulty  in communicating their status to the doubting person A.

The problem which we have uncovered here is in fact of very widespread interest in economics. Essentially it relates to any transaction in which one of the parties is better informed than the other. Such transactions are said to be characterised by a structure  of information which is asymmetric. Akerlof  (1970) gives as an example of a market with asymmetric informa- tion; transactions in second-hand cars. It will often  be very costly for a buyer of a second-hand car to determine accurately its true quality. He or she may or may not buy a ‘lemon’ (an American expression for a bad car) but ex ante there is not much that can be done to avoid it. Ignorance  on the part  of buyers will imply therefore  that  both  good and bad second-hand cars sell for the same price. Sellers of these cars, however, will have much better knowledge of their history and characteristics and therefore corre- spondingly  better  judgement  about  the probability of obtaining good or bad service in the future. The upshot will be that owners of good cars will tend to feel that the second-hand market  value seriously understates their (better informed) valuation. The owners of ‘lemons’, on the other hand, are more likely to sell. As the average quality of second-hand cars offered for sale falls, the price that they fetch falls with it, and this accentuates  the ten- dency for only the worst to be offered. It is at least possible to develop a model in which this adverse selection problem is so serious that no transac- tions will take place at all, even though better-informed buyers would stand to gain.2  Note the implied assumption, however, that contracts  cannot  be drawn up in such a way that failure of a car to meet the standards claimed for it would elicit penalties from the seller.

Perhaps the classic instance of adverse selection brought about by asym- metric information is in the realm of insurance.  A provider  of insurance against  some  undesired  contingency  (say  ill  health)  may  be  less  well informed  about  the probability of  the occurrence  of  this event than  the person seeking insurance. The terms quoted by an insurance company will be based upon certain basic pieces of information such as the age and medical history of the person buying the insurance; information which may be obtained  at relatively low cost. This information may not  be detailed enough, however, to distinguish  with sufficient subtlety between relatively good  and  bad  risks. Once more,  people  with good  health  prospects  will regard the insurance terms offered as rather unfavourable while people with bad prospects  will find the terms attractive. The people who are therefore most inclined to take out health insurance are those most likely to require health care. There is an ‘adverse selection’ problem based upon asymmetric information. Ex ante it may be very costly for an insurance company to dis- tinguish between good and bad risks, just as, by assumption, buyers of cars in the last paragraph could not distinguish good and bad cars, and person A in our story found it difficult to distinguish between good and bad crafts- people.

1.2    Adverse selection and ‘reputation’

In this context it is useful to note the importance of ‘reputation’ or ‘good- will’ in markets with asymmetric information. As was seen in the exchange game of Chapter  1, if knowledge  of cheating  becomes widely known  ex post, traders  will have an incentive to remain in good standing.  A seller of second-hand cars who has established a reputation for providing good cars will be able to charge higher prices or establish markets where none existed before. Similarly, person  A found  it expedient  to use craftspeople with a good local reputation. The existence of ‘goodwill’ economises on transac- tions costs by reducing search, thus enabling trade to take place in higher- quality products  and services than might otherwise be possible.

1.3    Adverse selection and ‘signalling’

Another response to the adverse selection problem is to think of ‘signalling’ or ‘screening’ mechanisms.  A signal is an activity which convinces buyers of the quality of a seller’s wares. It is convincing because it is so structured that a seller of poor-quality products  would be irrational to undertake the activity. As Hirshleifer and Riley (1979, p. 1406) express it: ‘Signalling takes place when sellers of truly higher-quality products  engage in some activity that   would  not  be  rational   for  those  selling  lower quality  products.’ Examples of possible signalling devices include the following:

• Advertising    Advertising  may  be  more  rational   for  producers   of high-quality   goods  than  for  providers  of  low-quality  goods.  By engaging in an expensive advertising campaign, producers are expressing confidence in the ability of their product  to attract and keep customers.  A shoddy  product  that  may fool customers  on  a single occasion  but  for no longer is unlikely to be worth  extensive advertising.
• Education qualifications    Educational credentials can act as a signal if the marginal cost of education  is lower for higher-quality workers. Offering higher wages for people with certain  qualifications will be most attractive  for those who find these qualifications the least costly to achieve. A more detailed exposition of this mechanism will be pre- sented in Chapter  6.

• Insurance deductibles    In insurance markets, higher quality risks can signal their status to insurance companies by a willingness to accept a big deductible (that is, less than full coverage).

Signalling may develop in the presence of information asymmetry and adverse  selection.  It  should  not  be assumed  that  the results  are  always socially advantageous. As will be seen later, signalling can lead to a waste of resources. If the signal is costly to transmit  and the private gains to sig- nalling are all derived from someone else’s private losses, the end result will be  a  simple  redistribution  of  income  with  resources  dissipated  in  the process. Where, therefore,  the information conveyed by a signal permits a more effective allocation of resources, signalling may be both privately and socially beneficial. Some signals, though,  may have the opposite  effect. In Chapter 9, for example, we will discuss how ‘signalling’ to the stock market by managers of joint-stock companies might result in a form of ‘short termism’.

1.4    Ignorance of contractual results

Where information remains hidden ex post and the buyer can never really tell the quality of the services he or she has received, the adverse selection problem  is yet more difficult to overcome.  Suppose,  for example, that  A attempts  to draw up a set of state-contingent contracts  with his workpeo- ple. He agrees, for example, that if certain geological conditions  are found to prevail, the foundations will need to be strengthened or extra drainage installed. He agrees that if weather conditions  are unfavourable, construc- tion may be delayed and specified extra expenses incurred and so forth. This type of contract  clearly requires that  A and his workpeople  can agree on what ‘state of the world’ actually  pertains.  If, for example, it proved  very costly for A to verify the correct position, the workers would have an incen- tive to ‘observe’ any state of the world which they felt was most favourable to  their  own  interests.  Where  it is expedient  to  discover  problems,  the worker  will duly discover them,  and where it is inexpedient  they will be ignored. When workers inform person A that they have ‘hit problems’, they are saying ‘I have observed a state of the world which permits me to take the following actions under the terms of our contract  and which commitsyou to extra expenditure.’ In many cases, the problem  will be sufficiently obvious  to  both  parties,  but  in others  person A may  have  to  trust  the worker. Asymmetric information therefore turns out to be at the root of A’s difficulties once more.

In the circumstances of person A’s building project it is perhaps unlikely that this extreme form of information asymmetry  will pose intractable difficulties.  Nevertheless,  there  are  situations  in which  the  quality  of  a service is difficult to assess by the purchaser even after it has been delivered. As will be seen in Chapter 11, this possibility lies behind some recent think- ing concerning the rationale  of the non-profit enterprise.

2.   Moral Hazard or ‘Hidden Action’

A second important problem facing A was that even after striking a bargain he did not know whether the other parties were fulfilling their obligations. The electrician  seemed to have wired his house, but was it safe? Was the architect  actually exerting him or herself on A’s behalf ? Was the plumber using unnecessarily  expensive equipment?  Whereas  under  section 2.1 the problem was that ex ante a buyer might be ill-informed about the qualities of a potential  purchase  or the difficulties encountered by people working on their behalf, the problem  being considered now concerns the difficulty of observing or deducing the actions of the supplier.

Moral  hazard  exists when the probability of a given ‘state of the world’ occurring  is influenced  by one of the parties  to a contract but when the behaviour of  this  contractor cannot  be  observed.  Insurance contracts again  supply  the  classic case. Suppose  that  person  A,  in  return  for  a specified payment now (an insurance premium), promises to pay to person B (who we will assume to be female) another specified sum in the event of person  B being  robbed.  Our  discussion  so far  has  been limited  to  A’s problem of deciding whether B is telling the truth  when she claims to have been robbed or whether in reality the only person being ‘robbed’ is person A. Even when the prevailing ‘state of the world’ is easily verified, however, there remains  the possibility  that  the outcome  was materially  influenced by the activities of B. In other  words, the probability of being robbed  is not entirely independent of B’s behaviour. She is obviously more likely to be robbed if she spends long periods of time away from her house and habitually leaves the door open than  if she installs a system of locks and alarms and never leaves her property unattended. This suggests the possi- bility that  the insurance  contract could specify conditions which commit person  B to take  certain  precautions. Once more,  the problem  of asym- metric  information is confronted, however.  If the contract states  that  B must  always  lock her door  when leaving her house,  how  is A to  know whether this provision  was or was not complied with when B was robbed? Further, although it is easy enough  to think  of  a few basic precautions against robbery, it would be extremely costly to investigate the detailed cir- cumstances of person B in order to establish the actions required of her in every particular. Only  B can  have  the  kind  of  knowledge  concerning specific circumstances necessary to determine  all the options available  to discourage  thieves. Once the insurance  contract is agreed, however, B will clearly have a much smaller incentive to engage in thief-discouraging activ- ities than before.

This general problem of verifying ex post whether a person’s actions have been compatible  with the provisions of a contract  is called the problem of moral  hazard.  In the context  of  insurance  markets,  Arrow  (1962) sum- marises the issue as follows: ‘The general principle is the difficulty of dis- tinguishing  between a state of nature  and a decision by the insured. As a result, any insurance policy and in general any device for shifting risks can have  the  effect of  dulling  incentives’ (p. 145). As we have already  seen, however, these problems  are not confined to insurance  markets.  Any con- tracts drawn up in conditions  of asymmetric information may give rise to moral hazard.  As Demsetz (1969) puts it: ‘Moral hazard  is a relevant cost of producing insurance; it is not different from the cost that arises from the tendency  of  men  to  shirk  when  their  employer  is not  watching them’ (p. 167).

3. Bounded Rationality

The third broad  class of transactional problem  facing person A which we may identify was the simple magnitude  of the potential task of coordinat- ing the activities of his workpeople.  This problem  would exist even were information symmetrical; that is, available equally to A and the people he employs;  and  it is therefore  logically distinct  from  the  issues discussed earlier in this section. Not only were the provisions  of each person’s con- tract interdependent but they would vary with all sorts of possible contin- gencies which might arise as work proceeded. The capacity of person A to imagine all possible future contingencies and then process the information required to allow for these different contingencies in the contracts  of each person  he hires is obviously  limited.  Person  A faces, in other  words,  a problem which is now usually referred to as ‘bounded rationality’.

The idea of ‘bounded  rationality’  is especially associated with the work of H.A. Simon (1957, 1969, 1979) and O.E. Williamson (1975, 1985). Both writers  use  the  example  of  the  game  of  chess  to  illustrate  the  issues involved. Given the rules which govern the movement  of the pieces on a chessboard, we might in principle consider constructing a list of all possi- ble games. We might  start  by recording  all possible opening  moves and then, for each one, record all possible legal responses, and so on, until we have built  up an entire  ‘decision tree’. The problem,  of  course,  and  the factor which prevents chess becoming a totally trivial pastime,  is that  the decision tree would be of such size and complexity that it beggars the imag- ination.  Even  the  best chess players  must  make  their  decisions  in  the absence of  a complete  list of  future  contingencies  which might  possibly flow from them, and resort must be made to a limited set of considerations which experience has suggested are important.

If rationality is conceived as selecting the best possible course of action for achieving a specified objective, chess moves evidently do not qualify. Yet most  people  would  baulk  at  describing  chess  decisions  as  irrational. Indeed,  chess is widely regarded  as the board  game requiring  powers of reason in the highest degree. Chess problems  are susceptible to the appli- cation of reason, but the complexity of the game is such that decisions are effectively taken under conditions  of uncertainty. This is why Williamson (1975, p. 23) argues that ‘the distinction  between deterministic  complexity and  uncertainty is inessential . . . As long as either uncertainty or com- plexity  is present  in  requisite  degree,  the  bounded  rationality problem arises.’

4.  Asset Specificity and ‘Hold-up’

There is a final contractual problem  which plays an important role in the transactions costs theory  of economic organisation. Where situations  are complex and contracts  are not absolutely ‘cast iron’ and perfectly enforce- able, there will be room for adjustments to contractual terms over time in response  to  changes  in  the  bargaining  power  of  the  contractors.  This problem is particularly serious when assets become ‘transaction specific’.

Consider once more the relationship between person A and his architect. It is likely that, over time, the architect has accumulated all sorts of special information about  the circumstances  and preferences of A, the nature  of his property, and the engineering and other difficulties associated with con- struction  work in the area. Were A to change his architect, the new person would not have access to the same stock of information and might, there- fore, be much less effective whilst undergoing  the initial process of ‘learn- ing’ about the background. The existing architect has an advantage over an outsider  in serving person A – a so called ‘first-mover advantage’.  Notice that, by assumption, person A has financed the accumulation of this useful information by paying the architect a fee at least as great as could have been achieved elsewhere during the initial period of their association, but the information is in the form of ‘human capital’, and the human form is that of the architect,  not person A. The architect,  realising that  he or she is a much  more  productive   resource  than   any  outsider,   may  therefore   be tempted to raise his or her fee. Person A will not like this, but may be pre- pared to pay some increase because the alternative of employing a new architect carries an even higher cost.

The architect  is more productive  than  any outsider,  and  this return  in excess of what would be achievable using alternative  human  resources is a form of ‘rent’.3  There is a surplus associated  with the use of the existing architect, a surplus accruing to a type of specific capital (specialised knowl- edge), and there is a danger that much effort can be taken up in a fight over who should receive it. Although person A has financed it, the architect may attempt to take some of it by implicitly threatening termination. If person A succumbs to this threat,  he is the victim of ‘hold-up’.

It is not necessarily always the buyer in a transaction who is vulnerable to ‘hold-up’, however. Suppose, for example, that A has such eccentric ideas that the plumber has to invest in material and equipment which he or she is most unlikely ever to find a use for again and which has a very low value on the second-hand market. Once the plumber has invested in these assets, s/he in turn will be vulnerable to contract  renegotiation. They are transaction- specific physical assets, and any return above their value in alternative  uses is again a ‘rent’. Person A may try to push down the agreed price of the work thus appropriating some of the rent which accrues to these assets. The plumber  will feel aggrieved but may be in a poor  bargaining  position  for resisting some adjustment to contract  terms. Any return  on the assets in excess of their value elsewhere is, in principle, better than terminating the contract.

In  practice,  of  course,  these  problems   are  unlikely  to  be  of  great significance in a local community where reputational effects are significant. Further, a simple solution to the specific physical assets problem is for the plumber  to  insist that  person A should  finance  them  in the  first  place. Person  A will buy the specialised equipment  and the plumber  will use it when undertaking the work. The plumber cannot ‘hold up’ person A, since a threat to raise the agreed fee will be met with the use of another  plumber. Person A cannot ‘hold up’ the plumber because a threat to lower the fee will result in the plumber working elsewhere. In other words, there is no longer a ‘rent’ element associated with the plumber’s services. The rent accrues to person A’s services as a provider of capital to himself, and it does not matter whether the return is seen as accruing to the provider or the user. That par- ticular transaction has been ‘internalised’.

Although there  are circumstances,  therefore,  in which asset specificity poses no great problem,  vulnerability  to ‘hold-up’ in general plays a large part in the theory of economic organisation. The existence of transaction- specific assets (either human  or physical) will play a significant role in our discussion of the firm’s relationship with its employees (Chapter  6), its sup- pliers (Chapter  7) and its financiers (Chapter  9). It is also associated with the question of the viability of cooperative and other forms of enterprise (Chapter  10). As will be seen, the modern approach to the firm emphasises that  its capabilities  are highly specific and related to the information and experience accumulated in its membership.  This implies that  a significant part of any firm’s return is rent on its specific human assets. A high degree of trust is required if these rents are not to be dissipated in distributional squabbles.  Casson  (1991) argues  that  it is a crucial role of  the business leader to engineer this trusting  environment. We will concentrate more on the  power  of  repeat  dealing  and  reputation. There  can  be  no  doubt, however, that business structure both reflects the degree of trust that exists between transactors and influences its future development.

In the following section, consideration is given to a number of different institutional responses to the problems  we have been considering.  Search costs,  asymmetric  information leading  to  adverse  selection  and  moral hazard,  bounded  rationality, and vulnerability  to  ‘hold-up’  are  factors which  inevitably  influence  the  ways in which  people  contract  with  one another. If gains from trade are potentially  available, it might be expected that institutions will be developed to facilitate their realisation, and this will involve mitigating the effects of some of the forces which stand in the way.

Source: Ricketts Martin (2002), The Economics of Business Enterprise: An Introduction to Economic Organisation and the Theory of the Firm, Edward Elgar Pub; 3rd edition.