Property rights and reputation in the dairy agro-industrial system

2.1 Governance structure

For TCE and MCE, the alignment generates savings in transaction costs, with different emphases. While TCE is concerned with quasi-rent protection in an ex-post perspective (Williamson, 1985), for MCE, the focus is property right protection with an ex-ante perspective (Barzel, 2001).

In TCE’s view, transaction costs economizing prediction considers transaction attributes and governance structures alignment and its effectiveness in limiting potential opportunistic behaviors among agents (Williamson, 1985). According to the author, the governance structures are market (discontinuous and impersonal); hybrid (formal or informal contracts); vertical integration (internal assets production).

Transaction attributes involve asset specificity, uncertainty and frequency. Asset specificity is the key variable that indicates the appropriate governance structure and involves value loss when the transaction does not materialize or in a contractual breach case (Williamson, 1985). Thus, asset specificity generates quasi-rent (the difference between the value resulting from a specific activity and its best alternate use), which requires an adequate governance structure to coordinate transaction, as agents can dispute for quasi rent appropriation (Williamson, 1985). In this guideline, the adequate governance structure serves as a protection mechanism against these opportunistic behaviors that can occur when a bilateral dependence is created, promoting adaptation of autonomous and cooperative kinds (Williamson, 1985).

In the MCE perspective, an alignment hypothesis can also be inducted. In this case, the principal condition in which transactions differ is related to measurement. Measurement affects organization form or protection mechanisms (Barzel, 2001). As pointed by Barzel (1987, p. 105):

[…] among factors contributing to the value of common effort, the greater the difficulty in measuring one factor’s contribution vis-à-vis that of others, the more likely is the owner to that factor to assume the position of the residual claimant.

Thus, as the attributes measurement becomes complex to measure and their costs rise, the use of reputation (long-term relationships) and, at the limit, vertical integration is justified. Contracting is justified when attributes measurement and verification can be established objectively in the contractual clauses (Barzel, 2005). Though, transaction costs arise when the absence of legal apparatus leads to economic rights disputes, in which parties seek to negotiate losses of rights.

2.2 Measurement and property rights protection

By protecting property rights in the transaction, besides reducing opportunistic behavior (Williamson, 1985), efficiency is perceived by the more productive choices made by agents, generating a positive impact on economic growth and development (Auerbach & Azariadis, 2015). Thus, measurement is presented as a monitoring mechanism to generate information about the dimensions of assets transacted, contributing to the protection of property rights (Barzel, 2005). Therefore, information is created by decomposing transactions into measurable dimensions that influence the agreements and enable evaluation and control (Barzel, 2001).

3In this context, measurement is established as an ex-post instrument to enforce agreements between agents. So, property rights indications are defined ex-ante in contractual relationships (which generates transaction costs) to make enforcement mechanisms action feasible. Thus, property rights study is associated with information costs, with a focus on protecting rights related to transaction dimensions (Zylbersztajn, 2018).

Following the old property rights tradition, Barzel (1994) considers that an individual’s property right over resources consists of the right or power to consume, obtain income and dispose of these resources through an exchange. The author distinguishes legal rights and economic rights, considering legal rights as “[…] individuals’ rights that the state helps enforce” (Barzel, 2001, p. 4). Economic right, on the other hand, is characterized as the ability to directly consume the asset’s services or indirectly through exchange.

In this sense, according to Barzel (1994), the key question about property rights is not on assets possessed by the actors, but what can they do with the attributes of the assets. For Foss and Foss (2001), this occurs, as most assets have several attributes, which cannot be specified and the notion of ownership of assets is vague. As the transaction is expensive and obtaining complete information on the dimensions involved is difficult, property rights are never fully delineated, making space for income capture.

2.3 Reputation

Reputation is informal or indirect enforcement, without a third party to arbitrate where punishment can be inflicted instantly. Is related to a person’s propensity to keep his promises and is associated with the continuity of transactions (Macleod, 2007). As stated by Williamson (1993), reputation must be acquired in the relationship, enable reducing risks and generate efficient contractual relationships by reducing enforcement costs. Consequently, reputation is sustained by agents when honoring agreements and not engaging in the capture of value (Barzel, 2005).

For Williamson (1993), those who violate contracts and behave opportunistically lose their reputation. When performance is substandard “[…] the seller is punished via the effect of the breach on his future payoff,” ending the transaction (Macleod, 2007, p. 602). So, reputation directly influences the frequency of the transaction. According to North (1991), reputation is related to path dependence. The stability of institutions reinforces reputation, reduces information costs and generates potential gains. In that case “[…] adaptative expectations arising from the prevalence of contracting based on the existing institutions” (North, 1991, p. 109). The prevalence of contracting acts as a safeguard in transactions given experiences sharing, reduction of opportunistic behavior and, consequently, in transaction costs (Williamson, 1985).

The measurement and information generated and shared, provide transaction-specific protection in the absence of reputation. Thus, when considering Williamson’s (1985) proposition, transaction frequency can generate reputation among agents, in an alignment condition (attributes and governance structure); in the same way for Barzel (2005) reputation is associated with the use of short-term contract (for measurement) and long-term relationships that reduce moral hazard and the need for measurement. Correspondingly, considering that coordination also reduces conflicts in relationships (Farina, 1999), it is possible to present the following propositions:

4.1 Profile of respondents

The profile of interviewed producers in France and Brazil is presented in Table 1. Table 2 shows the characteristics of the processors interviewed in both countries.

4.2 Governance structure alignment in Paraná

In Paraná (Central-Eastern and Western regions), specific assets transacted (site-specificity, temporal specificity and physical assets) are aligned with a formal contract, as a governance structure. This alignment is based on the assumptions of TCE (Williamson, 1985) and MCE (Barzel, 2001). Measurable specific dimensions (volume, bacteria, somatic cells, temperature, fat, protein, cryoscopy, capacity or flexibility of storage in the property, accreditation of good practices on the farm, access to the property) and their impact on price, whether in bonuses or discounts, are defined in a booklet which confirms the enforcement by contract and low opportunism risk.

Identified verbal agreements (Western and Northern regions) indicate that there is no alignment on the TCE and MCE assumptions. This increases the possibility of opportunistic behavior, generating transaction costs in case of the retest in milk evaluation by the agents, ratifying Williamson (1985). Also, although measurement information is generated, there are difficulties in obtaining protection from the state from the producer side, given that only economic rights are present and they depend on negotiation, which is not always positive. The governance structure used and their analysis is shown in Table 3.

4.3 Property rights and reputation in Paraná

In the Central-Eastern region, the proposition that property rights protection occurs from measurement and information sharing (Barzel, 2001) is convergent to empirical data. The alignment between governance structure (formal contract) with the assumptions of TCE (Williamson, 1985) and MCE (Barzel, 2001, 2005) promotes reputation among agents (Macleod, 2007) and product value is distributed. It was empirically identified that measurement is performed by a third party continuously and information is shared between producers and processors, which makes it possible to protect property rights. Therefore, the contract guarantees the additional price for product dimensions (legal right), encouraging the production of quality milk in the region. The quality is considered a legal right, in terms of Barzel’s proposition. That’s because the measurable dimensions related to quality and how it will be encompassed by the price established in the contract (formal or informal). However, as the price is given by the buyer, based on a base price, its enforcement has to be negotiated (economic right).

Thus, the relationship between measurement (Barzel, 2001) and reputation (Macleod, 2007) was observed by all producers interviewed in the region, when considering that measurement generates trust in the relation. Furthermore, the confidence in the relationship is recognized in the evaluation that producers made of the processor. On a scale from zero to ten, seven as a minimum score was pointed out by interviewees. Path dependence also was highlighted, as evidenced by North (1991) and was empirically identified in the relationship, which contributes to system improvement. It is observed that measurement and information sharing reduce opportunistic attitudes related to possibilities of income appropriation, regarding values of milk dimensions (legal right). It is also noted that recurring frequency is established by contractual conditions but also enhances reputation.

Besides, producers’ reputation is also highlighted, ratifying Williamson (1993) that reputation has to be acquired. The interview convergence revealed that reputation is generated by the transparency that exists in the relationship and information sharing that allows the reduction of possible opportunistic behaviors. Besides, it can be asserted that the continuity of transactions, related to frequency, is also a reflection of the reputation created.

Empirically, it was not verified that reputation decreases all measurement costs. This can be justified by milk characteristics, which have variability in evaluation results. Besides, the high possibility of sanitary problems in its dimensions demands strict control. When comparing to other regions of Paraná, however, there were no reports of retests execution, which reduces transaction costs by double measurement, mainly by producers. Although measurement, in all milk collections, is necessary to protect agents’ property rights. In short, it is observed that reputation reduces transaction costs, as opportunistic attitudes are limited and double measurement (Barzel, 2005) is not mentioned.

In this sense, it was induced that reputation, associated with path dependence, recurring frequency and governance structure alignment with attributes and dimensions (TCE and MCE), leads to system improvement. It can be identified by the national representativeness that the region’s milk production has in terms of productivity and product quality (SEAB/DERAL, 2019), which demonstrates empirically what Auerbach and Azariadis (2015) claim about the system development. The statement of reinvestments in the activity by interviewed milk producers ratifies this notion.

In the West and North regions of Paraná (verbal agreement cases), the protection of property rights is limited. Measurement is made by a third party, continuously and results are shared, but respondents do not believe in these results from the processor’s lab. As an example of distrust, Producer 18 claimed to have done a retest and the results of the same sample were different from the results of the buyer’s lab. Thus, reputation was not identified as a consequence of measurement. However, in the West region, producers claim to trust the processor because it is cooperative and payments are always made.

It was observed in the interviews that measurement did not influence reputation, as reports of unreliable results put the producer on alert. As reported, reputation comes from path dependence (North, 1991). In this case, path dependence and reputational capital reinforce the stability of the relationship. Although there is no doubt about the payment, the governance structure is not sufficient to limit opportunistic attitudes (unpaid dimensions) and the measurement results are not reliable, say interviewed producers. As consequence, transaction costs are higher (retests cases have been identified) due to negotiation possibilities and the greater potential for opportunistic attitudes (unpaid dimensions), which is consistent with TCE and MCE.

It can be inferred that the fact of a formal contract does not exist puts a large part of the transaction under economic rights protection, just as Barzel (2005) predict, which reduces the potential for generating reputation. Also, the history of measurement errors limits this possibility, although confidence payment is effective. The uncertainties associated with milk variability, inherent to the activity, in turn, allows the persistence of this form of governance structure (informal contract).

According to the interviewed producers, measurement helps in negotiations with the processors. Despite this, the reputation was not identified in the North region. Thus, it is clear that reputation is not identified due to the lack of property rights protection, (legal and economic rights) to producers. It is observed that the lack of guarantee of rights also negatively impacts the income distribution in the chain, ratifying Barzel (2005) and Williamson (1985). According to producers, with the milk income, it is not possible to reinvest in milk production.

In this sense, it is noted that by not observing the reduction of opportunistic attitudes and property rights protection, the reputation is not verified, which reinforces the main statement proposed. Despite the sharing of information, this is not reliable by producers. Accordingly, property rights are not protected, as pointed out by Barzel (2005). As consequence, quality milk production is not encouraged. The absence of alignment between the transacted attributes and governance structure and the existence of economic rights not well defined, make governance structure inefficient. Transaction costs are thus generated, due to the need for double measurement and the need for negotiation (ex-post costs).

It is observed that the small number of processors reduces possibilities for producers to transact with reliable buyers, who consider quality in the payment of the product. No reductions in transaction costs were identified, as the governance structure does not limit the possibilities for opportunistic attitudes and there are negotiation costs, beyond the additional measurement costs for carrying out retests, just as Williamson (1985) and Barzel (2005) predicted. Table 4 presents the highlights from the interviews and their analysis.

In the Central-Eastern region, theoretical propositions are empirically evidenced, as formal contracts protect legal rights over milk properties and measurement generates information that favors the protection of property rights. Besides, reputation is created by the trust generated in the frequency of the relationship and some positive effects are identified such as reduced coordination costs, milk production with superior quality, continuity in the relationship and reinvestments in production. In the West region, reputation is created by path dependence and the relation with cooperatives (history of payment reinforces stability in the relationship) and in the North region, reputation is not created. In these two last regions, respondents stated that they were unable to invest in production. This analysis is summarized in Table 5.

4.4 Governance structure alignment in the old Midi-Pyrénées

In France, Williamson’s alignment prediction was observed. Supported by the institutional environment, a medium level of specific assets (site, temporal and human specificity) are transacted under formal contract governance. This alignment, in turn, is supported and justified under the assumptions of TCE and MCE (Williamson, 1985; Barzel, 2001, 2005) propositions. The dimensions are defined in the contract (fat, protein, bacteria, cells, germs, antibiotics and butter content) and measurement cost is low, divided between producers and processors interviewed.

In complement, primary and secondary data indicate that this governance structure is mandatory, established by the institutional environment to regulate transactions (Trouvé et al., 2014). According to the authors, a formal contract became mandatory in 2015, with a five years term. Thus, the contract sets the price determination mode considering the quality of milk, the volume to be delivered, the collection modalities and the conditions for reviewing and terminating the contract (Trouvé et al., 2014).

Producers 1 and 3, beyond selling milk to processors, also process the milk for sale directly to consumers, adding value to the product. Both producers differentiate their production, with the label bleu blanc coeur and yogurt production (producer 1) and with bio milk and cheese production (producer 3). In these cases, the use of vertical integration, predict by TCE is ratified, considering the increase in asset specificities, according to Williamson (1985). For MCE, the governance structure is not aligned given that measurement costs are not high and the dimensions are specified between agents.

Although the use of contracts is mandatory in France, it is observed that producers 3 and 4 have a tacit agreement with buyers. Producer 3, for direct selling to consumers, delivers to the processor only what is leftover from his production, which is interesting to the buyer due to product differentiation. Producer 4, on the advice of his lawyer, did not sign the contract because the base price of milk is not previously agreed, as he says:

We haven’t signed a contract because the price is not established. It has the volume, the quality, but not the price. We try hard to make milk of the established quality, but we don’t know the price that will be paid?

Thus, according to the producer, the processor is obliged to collect the milk, even without a contract, as the relationship is old.

Producers 3 and 4 said that they intend to terminate the transaction with buyers, process all the milk and sell directly to the final consumers. The intention is to seek the protection of their property rights and get better financial returns on their products. Failure to protect the base price of milk (economic right) generates mistrust in the relationship and the possibility of opportunistic behavior discourages transaction continuity, which confirms the assumptions of Barzel (2005). The governance structures used and their analysis is shown in Table 6.

4.5 Property rights and reputation in the old Midi-Pyrénées

In the old Midi-Pyrénées, property rights are protected by the institutional environment (mandatory contract) which guarantees the additional price for the product quality (legal right), which encourages producers to produce milk with superior quality. The economic rights are observed in the milk base price definition that is made by the processor and is not guaranteed by the contract. Thus, the use of contracts and the lack of economic rights protection do not generate reputation. The lack of property rights protection can favor the producer’s rent appropriation (Williamson, 1985) and according to Barzel (2005), the contract reduces the need for reputational capital.

Also, in the milk transaction case, measurement does not generate reputation, as the high possibility of problems per batch does not allow reducing the frequency of measurement, which is required in all batches/collections (by the processor), as empirically identified in the interviews. Thus, this characteristic does not allow the reduction of measurement costs due to the need to control the dimensions. Table 7 presents the highlights and observations identified in the interviews.

For the value distribution among agents in the region, it was observed that the measurement of milk attributes is done by a third party, continuously (three times a month for additional price calculation) and the information is shared between agents. Thus, it was noted that measurement of each product dimension and information sharing act just as Barzel (2005) predicted, limiting opportunistic behavior, reducing transaction costs and ensuring payment for quality milk. Therefore, the appropriating income possibility generated by milk attributes is restricted. Besides, in the region, transaction frequency is high due to contractual clauses and the difficulty that producers have in exchanging buyers. Despite this, the relation between measurement (Barzel, 2001) and reputation (Macleod, 2007), has not been identified, as this has not been confirmed by producers. This context is presented in Table 8.

4.6 Alignment, property right and reputation – analyze in the investigated regions

When observing the information from the Central-Eastern region of Paraná, it is noted that path dependence and appropriated governance structure (formal contract) generate reputation and increase transaction frequency, promoting coordination costs (transaction) reduction (according to P1). Measurement (by a third party, continued and with information distribution) responds to reputation in the relationship, which implies a reduction in coordination costs and positively influences the continuity of the relationship (according to P2). Increased returns are perceived by agents (individual performance) through superior quality milk production and reduced transaction costs, as a result of dimensions measured (guarantee of property rights) and reputation created (according to P3). In the old Midi-Pyrénées, the reputation has not been identified, the transaction continuity and its frequency result from contractual reasons (mandatory) and informal restrictions (agreement between processors). This makes it impossible for producers to replace buyers, indicating limits on the ability of the structure alone to get coordination improvements. Figure 2 summarizes the results in these two regions.

In the Western region of Paraná, reputation was identified due to path dependence, as discussed by North (1991) and it leads to an ongoing relationship between producers and cooperatives (according to P1). Measurement is performed on a recurring basis by a third party, but the information is partially distributed, which possibility processors value appropriation, limiting property rights protection (according to P2). As low financial returns are made, as a result of minimal quality milk transacted, modest investments in production by the producer are identified (validating P3).

In the Northern region of Paraná, reputation has not been found because there is a scenario of lack of confidence in measurement results, which is used to generate discounts in cases of non-compliance. The recurrence in transactions is not based on efficiency but is due to the low number of available and reliable processors for purchasing raw milk, which does not negate P1. There is an absence of property rights protection (legal and economic) by processors, generating the possibility of the capture of individuals’ rights (P2). The lack of rights protection and reputation reflects the inefficiency in the system, increase in transaction and measurement costs and less investment in production (P3). Figure 3 presents the results when property rights are partially protected, specific to Paraná (Western and Northern regions).

By relating empirical information with the theoretical assumptions presented, the propositions proved to be valid.

Regarding reputation created, it is noted that the reduction of opportunistic behavior and information sharing is directly linked to it. The frequency was also identified as a factor that influences reputation, but it can be associated with the governance structure established (a mandatory contract or lack of option). In cases where reputation has not been identified as previously existing, there are opportunistic attitudes in relationships and partial sharing of information, generated by measurement. The investment decision identified in conditions of structural alignment and reputation created shows that the performance is present when lack of rights protection and reputation is not present.

Efficiency (reduction in transaction and coordination costs) and increases returns (superior quality milk) as a guideline was identified in cases where: property rights protection was effective; frequency was high (mandatory or not); reputation was present (Central-Eastern region of Paraná) or reputation costs were replaced by mandatory contract (old Midi-Pyrénées). It is important to note that in studied dairy AGS, measurement costs are present, due to milk specific characteristics and its variability that requires measurement in all collections, however, an increase in that costs could be identified (duplication of measurement). Where reputation was absent, higher measurement costs were anticipated (coordination, negotiation and retest).

Thus, the main proposition presented theoretically was ratified. Empirical research indicated that, in cases where efficiency is present (reduction in coordination and transaction costs), there is reputation and property rights protection through measurement and information sharing, as predicted by Barzel (2005), as well as governance structure alignment, as hypothesized by Williamson (1985).