2008 6 SA 264 (Ck)

2010 4 SA 122 (SCA)


Ostensible authority, reliance and the implied authority of an attorney to compromise a suit

  CJ Pretorius
  University of South Africa


  44 Volume 1 2011 pp 189-203
  Download Case Note in PDF
  Read the judgment of the court on SAFLII


1 Introduction

Estoppel by representation is a flexible doctrine capable of application in a variety of circumstances, including the question of contractual liability. The courts have recognised that this doctrine may be employed to hold a party bound to the impression created of assenting to a contract, despite an absence of true agreement between the parties (Van Ryn Wine & Spirit Co v Chandos Bar 1928 TPD 417 422-424; Peri-Urban Areas Health Board v Breet 1958 3 SA 783 (T) 790; cf Benjamin v Gurewitz 1973 1 SA 418 (A) 425; Saambou-Nasionale Bouvereniging v Friedman 1979 3 SA 978 (A) 1002ff). Although over the years some writers have strongly advocated the use of estoppel to maintain the fiction of a contract in such instances (see De Wet and Van Wyk Die Suid-Afrikaanse Kontraktereg en Handelsreg 1 (1992) 22-23; cf De Vos “Mistake in Contract” 1976 Acta Juridica 177 180-81), the fact is that the need for estoppel has been greatly diminished by the rise of the doctrine of quasi-mutual assent, a version of reliance-based liability giving rise to an actual contract rather than the mere fiction of one (Rabie and Sonnekus The Law of Estoppel in South Africa (2000) 196; Van der Merwe et al Contract: General Principles (2007) 36-37 41).

Nonetheless, while estoppel in all probability will not directly be used to uphold a fictitious contract where consensus is in question, it has been instrumental in determining contractual liability in another sense, and that is where a principal has been held bound to a contract on the basis of ostensible authority. In the latter situation a principal incurs contractual obligations, despite a lack of actual authority on the part of his representative, on the basis that the principal created the impression that the representative actually had the requisite authority to conclude the juristic act in question on behalf of the principal (Van der Merwe et al 256-57; Rabie and Sonnekus 156). In such instances estoppel, in the guise of ostensible authority, is pivotal in affirming contractual liability. The recent matter of MEC for Economic Affairs, Environment and Tourism, Eastern Cape v Kruizenga 2008 6 SA 264 (Ck); MEC for Economic Affairs, Environment and Tourism, Eastern Cape v Kruizenga 2010 4 SA 122 (SCA) provides an apt illustration of the working of estoppel in this manner and also deals with the pressing question whether an attorney has the implied authority to compromise a suit, a matter which it seems has yet to be settled in our law. This note examines both aspects, as well as whether the reliance theory could be applied in circumstances traditionally suited to invocation of estoppel by agency.


2 Facts

In an action for damages in delict, the applicant's (MEC for Economic Affairs, Environment and Tourism, Eastern Cape) legal representatives, at the pre-trial conference (in terms of rule 37 of the Uniform Rules of Court) and later at the trial, conceded liability on the merits of the case and undertook to pay amounts claimed under certain heads of damage. An order of court that the applicant was to pay the admitted damages was made by consent and the hearing postponed. The applicant (appellant on appeal), with a view to reopening his case on the merits, launched the present application for rescission of the aforementioned judgment and an order withdrawing the admissions that his legal representatives had made as recorded in the pre-trial minutes.

The application was brought on the ground that, consistent with departmental practice, the state attorney, who had represented the applicant, had lacked specific authority to concede the merits of the action or to agree to payment of the amounts claimed. The attorney concerned testified that he had been instructed, in general, to conduct the litigation, but conceded that he had not received any express instruction to settle the merits and had simply made a mistake in doing so. The respondents resisted the application firstly on the grounds that the concessions made by the applicant's legal representatives, and the agreement that the matter should proceed only on the remaining issues, constituted a compromise of issues that were no longer in dispute; and secondly that the applicant was estopped from denying his legal representatives’ actual authority.


3 Decisions

In the court of first instance, Van Zyl J, in a comprehensive judgment concluded that by virtue of his appointment to conduct the applicant's defence, a representation had been made to the outside world that the state attorney (and counsel) had the usual authority that applies to that office (including it seems the authority to compromise or settle a case). In the absence of informing the respondents that a limitation had been placed on the authority of his legal representatives, the applicant must reasonably have expected that persons who dealt with them would believe that they had the authority to compromise the claims. Accordingly, the applicant was estopped from denying the authority of his legal representatives to enter into the settlement agreement and the application was dismissed with costs (par 69). In this regard Van Zyl J duly took stock of Glofinco v Absa Bank Ltd (t/a United Bank) 2002 6 SA 470 (SCA) 480D-E wherein Nienaber JA observed that the appointment of a person to a certain position of authority is a factor that is not to be underestimated.

On appeal, Cachalia JA, in delivering the unanimous decision of the Supreme Court of Appeal, accepted that by agreeing to the settlement, the State attorney not only exceeded his actual authority, but did so against the express instructions of his principal. However, it mattered not whether an attorney acting for a principal exceeded his actual authority, or acted against the principal’s express instructions. The consequence for the other party, who is unaware of any limitation of authority and has no reasonable basis to question the attorney’s authority, is the same. That party is entitled to assume, as did the respondents, that an attorney attending a pre-trial conference clothed with an “aura of authority” has the necessary authority to agree on compromises and settlements on behalf of his principal. In the respondents’ eyes the state attorney quite clearly had apparent authority to agree to a settlement (par 20). Consequently, concluded Cachalia JA, the applicant was estopped from denying the authority of the State attorney and the appeal was dismissed.

That these decisions appeal to one’s sense of justice and are entirely appropriate is quite apparent, but several aspects invite closer scrutiny, more specifically whether an attorney has the implied authority to compromise a suit and whether there is any room for applying the reliance theory directly in the circumstances.


4 Commentary

4 1 The Implied Authority of Legal Representatives to Compromise a Suit

It appears as if the question whether an attorney of record (or counsel) has the implied authority to compromise a case has not authoritatively been settled. In the present matter Cachalia JA expressed himself as follows in this regard:

It is settled law that a client's instruction to an attorney to sue or to defend a claim does not generally include the authority to settle or compromise a claim or defence without the client's approval.

That seems straightforward enough, but compare on the other hand the following dictum of Plewman JA in Hlobo v Multilateral Motor Vehicle Accidents Fund 2001 2 SA 59 (SCA) 65C-E:

What is more, in this country (as in England) the conduct of a party’s case at the trial of an action is in the entire control of the party's counsel. Counsel has authority to compromise the action or any matter in it unless he has received instructions to the contrary. In England his apparent authority to compromise cannot be limited by instructions unknown to the other party. .... At the stages prior to the assumption of control by counsel the attorney of record stands in the same position.

The contradiction between these two dicta is fairly obvious and both emanate from the level of the Supreme Court of Appeal, suggesting a measure of ambiguity at the very least (see further Midgley “The nature and extent of a lawyer’s authority” 1994 SALJ 415 419ff). Cachalia JA was prepared to admit that there was some uncertainty in the way the principle he enunciated had been applied by the courts (127B-C). However, our law has been heavily influenced by English law on this point (Kruizenga court a quo 296B; Klopper v Van Rensburg 1920 EDL 239 242; Hlobo v Multilateral Motor Vehicle Accidents Fund supra 65C; Midgley 1994 SALJ 420ff) and there it has been plainly said that the general authority to conduct a case provides the attorney with authority to compromise (Chown v Parrot (1863) 14 CB (NS) 74 83). Similar powers extend to counsel in English law (eg Swinfen v Lord Chelmsford (1860) 5 H & N 890 921).

Midgley (1994 SALJ 420) concludes that our courts, under the influence of English law, have distinguished between settlements on the one hand made prior to trial proceedings and those made during the course of litigation on the other, “and have accepted that the power to settle a claim is one of the usual powers afforded a legal representative when acting as an advocate in an action” (referred to by Cachalia JA par 8). Moreover, he observes (422), attorneys who conduct trials should have similar powers. Midgley cautions, however, that in the English matter of Matthews v Munster (1888) 20 QBD 141 (CA) 143-144 145 the court held that the authority to compromise does have certain limitations (420-423): first, the compromise must be incidental to conducting the case; second, counsel must act reasonably within the limits of his authority; third, the compromise must not be unjust; fourth, the legal representative must not have acted under a mistake of fact; and fifth, the client must not have been in court. Above all it seems that counsel has the power to do what he considers to be in the best interest of his client.

Although there is authority to the effect that it is within the general power of a legal representative to compromise his client’s case (eg Mfaswe v Miller (1901) 18 SC 172; Alexander v Klitzke 1918 EDL 87; Klopper v Van Rensburg supra), in other instances the courts have taken a far more restrictive view of the matter (eg Goosen v Van Zyl 1980 1 SA 706 (O); Bikitsha v Eastern Cape Development Board 1988 3 SA 522 (E); Hawkes v Hawkes 2007 2 SA 100 (SE) (which indicates that the case law is fairly ambiguous); see further Midgley 1994 SALJ 419ff). In contrast to the position in English law, Roman-Dutch law was seemingly adverse to permitting agents, including attorneys, to compromise suits without a special mandate (eg Voet 3 3 18; Grotius 3 4 3; and see further Kruizenga court a quo par 58; Midgley 1994 SALJ 415). Midgley (1994 SALJ 428), however, suggests that cases where legal representatives have settled without notifying the client are not necessarily in conflict with Roman-Dutch law, but are instances where these agents have what Voet 2 15 3 refers to as “free administration” (translation by Gane The Selective Voet (1957)). He also refers to Van der Keessel (Praelectiones ad Gr 3 4 3) who notes that a procurator is within his power to compromise a case where it is in the interests of the principal (423) (see further Silke The Law of Agency in South Africa (1981) 158).

The common law’s reluctance to permit attorneys to compromise suits is of course not without merit. The settling of a case is hardly a trivial matter and in the normal course of events one would expect any reasonable attorney to confer with his client prior to reaching a settlement (Midgley 1994 SALJ 428). But in the court a quo Van Zyl J observed that our law has been largely influenced by English law in this regard and that the requirement of a special mandate to compromise a suit is to be found in the historical development of the office of attorney in the Roman and Roman-Dutch law rather than as a general principle (par 58). Consequently, he concluded, the rationale for a special mandate is no longer relevant and that presently an attorney and counsel are substantially in the same position as their counterparts in English law (par 59). These observations are not to be taken lightly, coming from whence they did, but regrettably in the Supreme Court of Appeal Cachalia JA did not deal with this aspect of the judgment by Van Zyl J.

Whatever the history, it nonetheless appears as if the position at present is anything but clear, a situation that is graphically demonstrated by different versions of the law stated in the present matter. In the court a quo Van Zyl J followed dictum by Plewman JA in Hlobo v Multilateral Motor Vehicle Accidents Fund (supra 65C-E) as being a correct reflection of the legal position (par 57), while after reviewing the authorities Cachalia JA concluded as follows (par 11):

To summarise, it would appear that our courts have dealt with questions relating to the actual authority of an attorney to transact on a client's behalf in the following manner: attorneys generally do not have implied authority to settle or compromise a claim without the consent of the client. However, the instruction to an attorney to sue or defend a claim may include the implied authority to do so, provided the attorney acts in good faith. And the courts have said that they will set aside a settlement or compromise that does not have the client's authority where, objectively viewed, it appears that the agreement is unjust and not in the client's best interests. The office of the State attorney, by virtue of its statutory authority as a representative of the government, has a broader discretion to bind the government to an agreement than that ordinarily possessed by private practitioners, though it is not clear just how broad the ambit of this authority is.

Unfortunately, this statement does not take the matter much further. It may, for instance, be rather difficult to determine whether the instruction to an attorney to conduct a suit may include the implied authority to settle and, furthermore, if attorneys generally do not have such implied authority then why would the issue of setting aside an unsanctioned settlement which is unjust and not in the client’s best interests arise at all? This dictum could perhaps even be construed as meaning that for all practical purposes attorneys do have the implied authority to compromise a claim. All of which again suggests that the situation at present is far from satisfactory.

There are probably several good reasons why attorneys should have implied authority to compromise a claim: for one thing, they are no ordinary agents, but trained professionals and officers of the courts. One should be able to rely on the fact that an attorney (or counsel) who offers or accepts a compromise of a legal suit has the authority to do so. In other words, the very status of an advocate or attorney justifies an inference of such authority (see Midgley 1994 SALJ 421-22; cf R v Matonsi 1958 2 SA 450 (A) 456A-H; Benjamin v Gurewitz 1973 1 SA 418 (A) 428E-F). For another, the implied authority to settle a matter is necessary to facilitate the proper administration of justice (Midgley 1994 SALJ 421 428; cf Kruizenga court a quo 302D-E). In this regard it is opportune to note that it is the attorney of record who is the mouthpiece for a litigant. If exchanges between litigants usually flow through their legal representatives, it makes little sense to saddle a party with the duty of verifying a concession or compromise made by an opposing legal representative during the course of legal proceedings. That would render a process which is already somewhat cumbersome even more difficult to manage. And, as noted by Van Zyl J in the court a quo (296D), an agent’s implied authority and his ostensible authority normally coincide. So, in the absence of a communication or circumstances indicating the contrary, it appears reasonable to take an attorney who conducts himself as having the authority to compromise a suit at his word. What is further apparent is that the state Attorney has a broader discretion to agree to a settlement of a suit than private practitioners, although as indicated by Cachalia JA the exact ambit of this authority is not clear (see further Dlamini v Minister of Law and Order 1986 4 SA 342 (D); Moult v Minister of Agriculture and Forestry, Transkei 1992 1 SA 688 (Tk) 692; SCA par 10; Kruizenga court a quo par 62; Midgley 1994 SALJ 424-426).

Interestingly, despite Cachalia JA’s statement of principle regarding an attorney’s authority, or lack thereof, to compromise a suit, he had little trouble in finding that in the circumstances the relevant attorney was clothed with an “aura of authority” to do what attorneys usually do at rule 37 conferences, including agreeing to compromises (132D). That seems very similar to accepting that attorneys normally have the implied authority to settle matters at pre-trail conferences, a fortiori because the very purpose of such proceedings is to limit or settle points of contention between litigating parties. Nonetheless, it is suggested that when again confronted with the question of the implied authority of an attorney to compromise a case, the Supreme Court of Appeal should authoritatively deal with the matter and dispel the apparent ambiguity present in case law.

It appears that the courts are prepared to interfere with compromise agreements concluded without a client’s consent in certain circumstances (see Kruizenga SCA 127A-B; Midgley 1994 SALJ 420ff; cf Matthews v Munster supra). However, the real question is whether the grounds for striking down a compromise that has not specifically been authorised will be limited to the existing grounds for setting aside a contract (or declaring it void), or extend further. So, for instance, it may be perfectly feasible to declare void a compromise on the basis of (justifiable) mistake (compare eg De Vos v Calitz and De Villiers 1916 TPD 465; Chevallier v Protea Versekeringsmaatskappy Bpk 1972 2 PH A47 (T); Bam v Rafedam Boerdery BK 2004 1 SA 484 (O); and see further Midgely 1994 SALJ 420-423; Pretorius “Reasonable reliance and the duty to enquire” 2005 THRHR 122 124ff), but will the courts generally be prepared to extend this grace to the vaguer ground of agreements that are “unjust and not in the client's best interests” emphasised by Cachalia JA (129D-E)? After all the Supreme Court of Appeal has frequently declared that the courts will not strike down contracts merely because they are unfair (eg Sasfin (Pty) Ltd v Beukes 1989 1 SA 1 (A) 9B-C; Brisley v Drotsky 2002 4 SA 1 (SCA) par 24; Afrox Healthcare Bpk v Strydom 2002 6 SA 21 (SCA) par 32), and one might wonder why compromise agreements should be treated any differently. So guidelines regarding any extraordinary circumstances in which an unsanctioned settlement agreement may be set aside are also required.

4 2 Limitation of an Agent’s Authority

It is of course open to a principal to expressly curtail the implied authority of his agent but, as Van Zyl J duly noted in the court a quo (par 60), if a litigant limits the implied authority of his attorney to compromise a case (if it is accepted that generally such an authority exists and even more so in the case of the state attorney), unless the limitation of authority is communicated to the opposing litigant or legal representative, or is implicit from the principal’s conduct or the surrounding circumstances, he may be estopped from relying on the lack of authority (as to the nature of the relationship between attorney and client see Midgley Lawyers’ Professional Liability (1992) 5-10). A litigant (or any principal for that matter) cannot by way of private instructions to his legal representative (or any representative for that matter) curtail the latter's authority as far as third parties are concerned (see Hlobo v Multilateral Motor Vehicle Accidents Fund 2001 2 SA 59 (SCA) 65D; Glofinco v Absa Bank Ltd (t/a United Bank) 2002 6 SA 470 (SCA) 482B; Sonnekus “Akojee v Sibanyoni and another” 1977 TSAR 176 179; De Wet and Van Wyk 114).

De Wet and Du Plessis LAWSA (ed Joubert) 1 (1993) par 119 seem to maintain that where the tacit authority of a representative has been limited by private instructions which have not been communicated to the other party, the principal’s liability is based on actual tacit authorisation and not estoppel as such. The distinction between tacit authorisation and ostensible authority is rather fine, but more importantly it is suggested that where authority, whether express or implied, is curtailed by private instructions there is in fact no authority to bind the principal to the prohibited juristic consequence, and so it seems artificial to speak of “actual tacit authorisation.” The principal is in fact held bound on the basis that it reasonably appears to third parties as if the representative has the requisite authority, which conceptually amounts to no less than estoppel or the maintaining of an impression of authority attributable to the principal (see generally Du Bois (gen ed) Wille’s Principles of South African Law (2007) 990-991; Visser and Potgieter Estoppel: Cases and Materials (1994) 290ff; Van der Merwe et al 256-257; Kahn, Lewis and Visser Contract and Mercantile Law: A Source Book 1 (1988) 855).

Nevertheless, it is trite that ostensible authority (estoppel by agency) requires a representation by the principal in words or conduct (and not merely the assurance of the agent) that the agent has the required authority to bind the principal to the juristic act in question (see eg NBS Bank Ltd v Cape Produce Co (Pty) Ltd 2002 1 SA 396 (SCA) 412C-E; Glofinco v Absa Bank Ltd (t/a United Bank) supra par 13; South African Broadcasting Corporation v Coop 2006 2 SA 217 (SCA) pars 63-64; and the authorities cited by Du Bois et al 991 fn62; Rabie and Sonnekus 156ff). But it seems that a principal can represent by conduct merely by appointing a person to a position that usually clothes the incumbent with certain powers (Kerr The Law of Agency (2006) 96). And where, as in the present matter, an attorney is appointed to defend a claim, those powers would include attending pre-trial conferences where parties habitually make concessions and often agree to settlements (Kruizenga SCA pars 16-20). In other words, if an attorney attending a pre-trial conference conducts himself as if having the authority to settle the case, his counterpart would in most instances be reasonable in relying on such conduct, and the former’s principal would be bound by a compromise, irrespective of any express limitation on the attorney’s authority to conclude a settlement (Kruizenga SCA par 20). All of which tends to reinforce the notion that an attorney of record generally should have the implied authority to compromise a case. At the very least the distinction between actual implied authority and ostensible authority virtually evaporates in such instances and one simply cannot fault a party who relies on a settlement agreed to by an attorney at a pre-trial conference (Kruizenga SCA par 19). It should be remembered that although the principal is bound by the compromise agreement, he has recourse against the attorney who has breached his mandate (see Midgley Professional Liability 75-76; Midgley 1994 SALJ 428).

4 3 Ostensible Authority and Reliance

When an agreement is struck by an agent outside his authority the principal usually raises the representative’s lack of authority as a defence to being bound to the agreement. Although fairly common it appears that the nature of this defence is not entirely clear. The accepted grounds for holding a party contractually liable in South African law provide an apt point of departure, because notionally a principal can only incur contractual responsibility if one of these grounds is present. The bases for contractual liability have been dealt with elsewhere and do not bear repeating in any detail (see Pretorius “The basis of contractual liability in South African law” 2004 THRHR 179, 383, 549; Hutchison and Pretorius (eds) The Law of Contract in South Africa (2009) 13-20 91-107), but may be summarised as comprising the will theory qualified by the doctrine of estoppel (eg Van Ryn Wine & Spirit Co v Chandos Bar 1928 TPD 417 422-424; Peri-Urban Areas Health Board v Breet 1958 3 SA 783 (T) 790) or its relative the reliance theory (eg Hartley v Pyramid Freight (Pty) Ltd t/a Sun Couriers 2007 2 SA 599 (SCA) par 6; Pillay v Shaik 2009 4 SA 74 (SCA) par 55), and a modified form of declaration theory as corrected by the iustus error approach (eg George v Fairmead (Pty) Ltd 1958 2 SA 465 (A) 470-473; National and Overseas Distributors Corporation (Pty) Ltd v Potato Board 1958 2 SA 473 (A) 479).

Now since actual agreement is usually regarded as the primary basis for contractual liability (see eg Saambou-Nasionale Bouvereniging v Friedman 1979 3 SA 978 (A) 993E-F; Mondorp Eiendomsagentskap (Edms) Bpk v Kemp en De Beer 1979 (4) SA 74 (A) 78G-H; De Wet and Van Wyk 9; Van der Merwe et al 38-39; Hutchison and Pretorius 20), it tends to follow that the defence of lack of authority of an agent amounts to no more than an assertion by the principal that he has simply not assented to the contract in question and therefore cannot be liable. That seems simple enough, if not self-evident, because clearly the principal does not wish to be contractually bound on the basis that the agreement brokered by his agent does not accord with his intention. In other words there is dissensus between the principal and third party, and possible liability falls to be decided on objective grounds, including the doctrine of estoppel (compare eg Van Ryn Wine & Spirit Co v Chandos Bar supra 422ff; cf Benjamin v Gurewitz 1973 1 SA 418 (A) 425). In terms of the law of agency, however, a principal’s liability is determined with reference to consensus reached between the agent and third party since it is the agent who acts and not the principal (see Joubert Die Suid-Afrikaanse Verteenwoordigingsreg (1979) 44; De Wet and Du Plessis LAWSA par 109). In the court a quo Van Zyl J explained as follows (292 note 105):

As it is the representatives and not the principal who concludes the juristic act, the question whether the parties were in agreement has to be determined with reference to the intention of the representative. The principal’s intention is not relevant to the question whether consensus necessary to produce a contract existed or did not exist. Knowledge required [acquired] by the representative is imputed to the principal and a contract concluded by the representative on terms not intended by the principal does not affect the issue of consensus.

One may nevertheless question whether the intention of the principal is necessarily irrelevant to the determination of consensus, because surely the juristic act in question should generally accord with the mutual intention of the contractual parties (cf Van Ryn Wine & Spirit Co v Chandos Bar supra 422; Goldberg v Carstens 1997 2 SA 854 (C) 859). After all the contractual obligations arise between the principal and the third party (albeit through the endeavours of the agent), something of which all the parties should be aware (cf Du Bois et al 989; Van der Merwe et al 254 260-261). If, for instance, the third party knew that the agent exceeded his authority in binding the principal one could hardly say that consensus existed (cf De Wet and Du Plessis LAWSA par 119). A further indication that the intention of the principal is relevant is the fact that he can exert his will on an unauthorised juristic act by affirming its validity by way of ratification. Ratification is a unilateral juristic act amounting to a declaration of intention to be bound by an otherwise unauthorised (and hence non-binding) legal act (see generally Kerr Agency 80-94; De Wet and Du Plessis LAWSA 123-29; Van der Merwe et al 257; De Wet and Van Wyk 114-15). The relationship between principal, agent and third party seems to entail more than mere linear acts between principal and agent on the one side and agent and third party on the other (cf De Wet and Du Plessis LAWSA par 119). Ultimately, the principal is bound because there is a basis for liability as between the agent and third party, and because the principal actually authorised or ostensibly authorised the conclusion of the juristic act. But the question is whether this two-pronged approach can be merged into one in typical instances of estoppel by agency.

Where an agent exceeds his authority in binding his principal to an agreement there certainly appears to be a measure of dissensus: on the one hand the third party surely labours under some form of misapprehension as to the intention of the principal to enter into the legal relations in question, while on the other hand the principal lacks the intention to enter into the contract (cf Christie The Law of Contract in South Africa (2006) 29-32; Kerr The Principles of the Law of Contract (2002) 41-45; Van der Merwe et al 23-24; Hutchison and Pretorius 85). Consequently, one may very well query where an agent exceeds his authority, and hence misrepresents (by act or omission) his authority to bind his principal, whether true consensus between the representative and the third party exists. Rather it seems as if an element of consensus is absent; namely whether the principal has the requisite intention to enter into the specific legal relations with the third party. It is important to note that although one is dealing with consensus between the agent and third party, their mutual intention relates to the incurring of obligations as between the principal and third party (cf Du Bois et al 989; Van der Merwe et al 254; De Wet and Du Plessis LAWSA par 101) and the common intention must at least by implication incorporate the principal’s intention (to incur contractual obligations). It is this latter intention that is misrepresented to the third party. Whichever way one looks at the situation there seems to be dissensus occasioned by the conduct of the representative. Although of course the agent himself may incur personal liability for unauthorised acts (see De Wet and Du Plessis LAWSA par 138ff), the question for present purposes is under what circumstances the principal should be held accountable.

It is clear that a principal must bear the brunt of the misrepresentation of his agent who causes an operative mistake on the part of a third party; this results in the undoing of the agreement on the basis of a iustus error (eg Allen v Sixteen Stirling Investments (Pty) Ltd 1974 4 SA 164 (D); Spindrifter (Pty) Ltd v Lester Donovan (Pty) Ltd 1986 1 SA 303 (A); Maresky v Morkel 1994 1 SA 249 (C); Goldberg v Carstens supra; cf Kok v Osborne 1993 4 SA 788 (SE)). In such circumstances the misrepresentation of the agent activates successful invocation of reliance in its indirect or negative form; that is that no contract exists because of material and reasonable error on the part of the third party. Conversely, one could argue that where an agent misrepresents (in as many words or by other conduct) to a third party that he has the necessary authority to conclude an agreement of specific content on behalf of his principal, whereas in fact he has exceeded his actual authority, it could trigger an application of reliance in its direct or positive form; that is that despite dissensus a contract exists based on the reasonable reliance in consensus of the third party (cf Van Ryn Wine & Spirit Co v Chandos Bar supra). However, a prerequisite for rendering the principal accountable in either instance is that the agent must at least have acted within the general scope of his authority, although he has exceeded the private instructions of the principal unbeknown to the third party (cf Du Bois et al 998; Van der Merwe et al 259; Van der Walt and Midgley Principles of Delict (2005) 38; Neethling, Potgieter and Visser Law of Delict (2010) 371). Generally the courts have been loath to come to the aid of a principal who clothes his agent with apparent authority and then falls back on private instructions in an attempt to escape the consequences of the agent’s conduct (compare eg United Cape Fisheries (Pty) Ltd v Silverman 1951 2 SA 612 (T) 616C; Akojee v Sibanyoni 1976 3 SA 440 (W) 442H).

Notionally the issues of an intention to incur legal relations on the part of the principal and the authority to enter into such relations by the agent on behalf of the principal tend to overlap. If the representative’s misrepresentation is attributed to the principal then the latter may be held liable on the basis of an objective theory of contract, such as the reliance theory, and provided further that the principal was complicit in some or other manner in creating the impression that the agent had the authority to bind him to the particular transaction. Indeed the law reports contain clear cases of the analogous situation where an employee has exceeded his authority in representing that his employer has assented to an agreement, and the employer has incurred contractual liability despite the absence of an actual intention to enter into the particular transaction (eg National and Overseas Distributors Corporation (Pty) Ltd v Potato Board supra; Hlobo v Multilateral Motor Vehicle Accidents Fund 2001 2 SA 59 (SCA); cf Cecil Nurse (Pty) Ltd v Nkola 2008 2 SA 441 (SCA)). In such instances the courts have stressed that as long as the employee was the appropriate person within the organisation to usually deal with the matter at hand, the lack of actual authority for the transaction could not be used to escape liability (eg National and Overseas Distributors Corporation (Pty) Ltd v Potato Board supra 480C-D; Hlobo v Multilateral Motor Vehicle Accidents Fund supra 66I-J). Now it may be said that in these cases it is in fact the employer acting through its functionary, while in the case of an independent agent it is the agent acting albeit on the behalf of the principal, but the principle nevertheless seems to be very similar.

The courts have filled the interstitial space between an agent exceeding his authority and possible dissensus with the construct of ostensible authority, which boils down to a specific application of estoppel by representation (Pretorius “The basis of tacit contracts” 2010 Obiter 518 532-533; Van der Merwe et al 256-257; De Wet and Du Plessis LAWSA 129). In such instances for liability to lie there must be a basis for contractual liability (at the very least by implication) and the requirements for ostensible authority must have been met (compare eg Big Dutchman (South Africa) (Pty) Ltd v Barclays National Bank Ltd 1979 3 SA 267 (W) 281E, 282E-F; NBS Bank Ltd v Cape Produce Co (Pty) Ltd 2002 1 SA 396 (SCA) 408E, 412C-E; South African Eagle Insurance Co Ltd v NBS Bank Ltd 2002 1 SA 560 (SCA) 573E-H; cf Hlobo v Multilateral Motor Vehicle Accidents Fund supra 65G-I, 66D-I). On the other hand if the principal is directly accountable for the agent’s misrepresentation, the process may be short-circuited by merely applying the reliance theory against the principal. And in both scenarios the issues central to establishing liability tend to converge, the underlying theme being the protection of reasonable reliance on the part of the third party (cf Hutchison and Pretorius 92-93; Pretorius 2004 THRHR 190-91; Van der Merwe et al 41). Furthermore, in either instance the law would require some form of misrepresentation on the part of the principal as to the agent’s authority to hold him accountable for the acts of the agent. Generally, such a misrepresentation (by conduct) would constitute appointing a representative to a position with certain inherent or implied powers (see Glofinco v Absa Bank Ltd (t/a United Bank) 2002 6 SA 470 (SCA) par 1; Kerr Agency 96; cf Southern Life Association Ltd v Beyleveld 1989 1 SA 496 (A) 503B-D) or instructing an agent belonging to a particular class which has customary powers (Inter-Continental Finance and Leasing Corporation (Pty) Ltd v Stands 56 and 57 Industria Ltd 1979 3 SA 740 (W) 748D), and then limiting those powers by way of private instruction or later denying them. Notionally then liability stems from two misrepresentations: a misrepresentation on the part of the principal in clothing his representative with apparent authority to conclude the juristic act in question (contrary to the agent’s actual authority) and a misrepresentation by the agent that he has authority to bind the principal to the juristic act, whereas in fact he has exceeded his actual authority. Cumulatively these misrepresentations engender reasonable reliance in consensus on the part of the third party, justifying the ascription of contractual liability to the principal (contrast also Kerr Agency 94ff).

In the present matter Cachalia JA framed the test for holding the principal bound by way of estoppel as follows (132B-C):

The proper approach is to consider whether the conduct of the party who is trying to resile from the agreement has led the other party to reasonably believe that he was binding himself.

Now this approach is on all fours with the authoritative statement of the reliance theory in Sonap Petroleum (SA) (Pty) Ltd v Pappadogianis 1992 3 SA 234 (A) 239I-J (see also eg Constantia Insurance Co Ltd v Compusource (Pty) Ltd 2005 4 SA 345 (SCA) par 17; Cecil Nurse (Pty) Ltd v Nkola supra par 15), and the practical result in the circumstances was no different than applying the reliance theory directly to the facts (cf Kerr Agency 94-95). In sum: where an agent exceeds his actual authority in concluding a contract on behalf of his principal, depending on the circumstances, one could either determine whether a basis for the contract is present and then look to estoppel to counter a plea by the principal that the agent lacked the actual authority to bind him, or accept that there is dissensus between the agent and third party and apply the reliance theory directly. Circumstances permitting, the end result should for all practical purposes be the same. Hardly surprising then that whenever lack of authority is raised by a party as a defence to being contractually bound elements of estoppel (ostensible authority) and/or the reliance theory (indirectly or directly) tend to feature and are sometimes interwoven (compare eg Van Ryn Wine & Spirit Co v Chandos Bar supra 422-23; National and Overseas Distributors Corporation (Pty) Ltd v Potato Board supra 478ff; Hlobo v Multilateral Motor Vehicle Accidents Fund supra 64ff; cf Kerr Agency 94ff).

In general terms it seems that a principal will be liable for the misrepresentation of his agent where the latter has the (usually implied) authority to make representations in relation to the transaction concerned (Van der Merwe et al 259; Neethling et al 371-72; cf Van der Walt and Midgley 38). The basis for such liability is not clear (see Wicke “Vicarious liability for agents and the distinction between employees, agents and independent contractors” 1998 THRHR 609 612ff; Van der Merwe et al 260 note 103), but the possibility of risk creation seems to be gaining favour (see Neethling, Potgieter and Visser 371). Likewise the most plausible explanation for attributing an agent’s misrepresentation (as to his authority) to his principal within the context of the reliance theory is probably risk. The risk theory has been suggested as a possible corrective to the will theory in certain instances of dissensus (see eg Meijers “De grondslag der aansprakelijkheid bij contractueele verplichtingen” Verzamelde Privaatrechtelijke Opstellen 3 (1955) 81 84-85; Hofmann “The basis of the effect of mistake on contractual obligations” 1935 SALJ 432 436), but this principle is poorly developed and simply cannot stand independently as contractual basis (see Stewart v Zagreb Properties (Pvt) Ltd 1971 2 SA 346 (RA) 350-51; Lubbe and Murray Farlam and Hathaway Contract: Cases, Materials and Commentary (1988) 163 note 1; Pretorius 2004 THRHR 389 note 44). However, risk can potentially function as a possible determinant of liability within the context of an objective theory of contract, such as for instance the reliance theory (Pretorius “The basis of contractual liability (2): Theories of contract (will and declaration)” 2005 THRHR 441 453; Floyd and Pretorius “A reconciliation of the different approaches to contractual liability in the absence of consensus” 1992 THRHR 668 672-73).

In other words the representative’s misrepresentation as to his authority binds the principal simply because the latter should bear the potential risk occasioned by appointing the agent to a position which usually carries certain powers or appointing an agent of a class which customarily has certain powers, even if those powers have been curtailed by private instruction. It seems that the courts have been prepared to embrace the notion of risk in agency situations (see Neethling et al 371). For instance, in Randbank Bpk v Santam Versekeringsmaatskappy Bpk 1965 4 SA 363 (A) 372 Steyn CJ noted:

Dit is redelik dat die prinsipaal wat sy verteenwoordiger kies en hom voorhou as ’n betroubare person, en nie die ander party wat geen seggenskap by die keuse het nie, die risiko van sy moontlike oneerlike voorstellings of verswygings sal dra.

Ultimately, the principal may be liable, despite a lack of intention to engage in the specific contractual relations in question, because the third party was led to reasonably believe that a consensual juristic act had been validly concluded. Although the specific circumstances of each case will no doubt play an important role in determining whether the principal and agent have been instrumental in inducing reasonable reliance on the part of the third party, within the context of legal practice, the compromising of a suit at a pre-trial conference by an attorney of record will be hard to refute by the principal irrespective of the reliance-based mechanism invoked by the third party (cf Kruizenga SCA par 19).


5 Conclusion

Whether or not an attorney (or counsel) has the implied authority to compromise a suit on behalf of his client still seems to be an unresolved issue in our law, and there are seemingly contradictory statements of principle in this regard at the elevated level of the Supreme Court of Appeal. In the present matter the court a quo was quite prepared to acknowledge such authority, but the Supreme Court of Appeal came to the opposite conclusion. Roman-Dutch law generally was adverse to the notion of implied authority on the part of an attorney to settle a matter without the client’s consent, but English law developed along different lines and broadly supports this proposition. Although there probably are good reasons for holding either way, it is suggested that the evidence that our law has borrowed rather heavily in this regard from English law, and generally is in accordance with the position in that legal system, is convincing. Consequently, the approach of the court a quo is preferable to that adopted in the Supreme Court of Appeal. What is still required though is an authoritative statement of principle by the Supreme Court of Appeal to dispel the present uncertainty. Above all it appears as if the general implied authority to compromise a suit is vital for the effective administration of justice: it simply is too much to saddle a litigant with the burden of having to determine whether each concession or settlement proposed by the opposing attorney bears the express approval of the latter’s principal.

Although estoppel these days is rarely seen as a mechanism for holding a party bound to an apparent agreement, it continues to be instrumental in the apportionment of contractual liability by way of ostensible authority where seemingly an agreement has been concluded by an agent who has exceeded his authority. In such circumstances the principal invariably relies on the representative’s lack of actual authority to escape liability under the agreement, to which the other contractual party usually raises a plea of estoppel by agency. The question is whether the principal represented to the other contractant that the agent had the required authority to bind the principal to the juristic act in question. Although estoppel has done sterling service in this regard and its application here attests to its continued vitality within the context of contract formation, it is suggested that the process may be short-circuited simply by applying the reliance theory in appropriate circumstances. Such an application is essentially founded on a dual premise: firstly, the issue of lack of authority in actuality amounts to a contention that the principal lacked the required intention to enter into the juristic act in question, which boils down to an absence of actual agreement as contractual basis as between the agent and third party and potentially activates an objective theory of contract, such as reliance. Secondly, a misrepresentation on the part of the principal in appointing his representative to a position which usually carries certain powers or appointing an agent of a class which customarily has certain powers, and then privately limiting the implied authority of his representative, and a further misrepresentation on the part of the agent in exceeding his actual authority (attributed to the principal on the basis of risk creation) cumulatively induce reasonable reliance in consensus on the part of the third party.

Although the estoppel by agency and direct reliance approaches eventually achieve the same result, the latter may be preferable in that it combines the elements of contractual liability and ostensible authority in one approach, whereas estoppel by agency requires a separte basis for liability on the one hand and the requirements for ostensible authority to be met on the other. Both approaches, however, are largely contingent upon elements typical of reliance-based liability in the form of inducement and reasonable reliance. In the present matter the Supreme Court of Appeal framed the test for ostensible authority in a manner suggestive of direct reliance, perhaps portending that the reliance theory may yet readily be applied in circumstances traditionally suited to estoppel by agency.