Reassessing Marshall’s Producers’ Surplus: A Case For Protectionism

Download full paper

This paper is closed for comments.


The rationale for liberal economic policies refers inter alia to the so-called producer and consumer surpluses, namely welfare concepts which were proposed by Alfred Marshall in his seminal work Principles of Economics, first published in 1890. In the case of trade policy, relying on surpluses and referring to the ‘small country case’, it is recommended to remove tariffs barriers imposed on the imports of commodities because it should increase welfare and, in theory at least, the losers of such a trade policy orientation can be compensated with the use of adequate transfers from winners.

Despite extensive use, the concepts of surpluses still raise key-questions that may alter the case for free trade. Thus, from a purely semantic perspective, the concept of producer, as presented in Marshall’s work, seems to be broader than what is proposed in the dominant economic discourse; in other words, workers should (also) be seen as producers.

Assuming the workers are considered as producers, at least their wage rents should be taken into account when discussing the impacts of trade liberalization, and added to profit losses – as a result, the case for free trade weakens considerably, it could even vanish.

Posted for comments on 12 May 2017, 11:00 am.

Comments (4)

  • Eithne Murphy says:

    I believe that this paper makes a simple central important point; that is, that the impact of trade liberalisation on workers as producers is insufficiently recognised, both in theoretical work and applied scholarship. Moreover, he makes the argument that in evaluating the impact of trade liberalisation on workers’ welfare, we need to broaden beyond an aggregate income approach. Alternatives put forward are a focus on the impact of such policies on society’s least well off (Rawlsian perspective), or on workers’ capabilities; which depend, inter-alia, on their income, but also on their employment status. Even Nozick’s theory of justice allows for an emphasis on those who may be deemed victims of past government injustices.

    This is the key message of the paper, which is that such welfare effects have been rendered invisible by the tools that economists deploy. Saying that I found the paper to be too long and rambling and believe that the key central message would be clearer if the paper was shorter and more focused.

    My suggestions are the following:

    1. Keep the section on trade liberalisation using demand and supply curves for individual markets. However, explain better the presumed gains from trade (areas 2 and 4 in Figure 1), not just confining your analysis to the difference between the gain in consumer surplus minus the loss of tariff revenue and producer surplus. Explain that area 2 is categorised as a production gain from the presumed better deployment of resources elsewhere in the economy, while area 4 is the consumer gain caused by the expansion of the market. In this regard, you could also show, that the removal of export subsidies can be analysed deploying the same set of tools and concepts and yielding similar welfare outcomes (production and consumption triangles), though in this instance it would be a case of fiscal and consumer gains outweighing producer loss.

    2. The section on WITS (World Integrated Trade Solution) is really interesting as it is an applied trade tool, which addresses the impact of different trade policy scenarios and, moreover, it is widely used. As the authors of the User Manual acknowledge, one of its main advantages is that it is easily deployed, as it simply requires knowledge of market demand and supply elasticities for individual products, the demand substitution elasticity for similar products from different countries and the proposed trade policy measure. In other words, its data requirements are not too onerous and the data are widely available.
    What stood out for me about WITS was what Linotte rightly refers to; that is to say, the lack of focus on domestic supply conditions. This I discovered is because, in most instances, the assumption is made that the elasticity of supply is infinite. Therefore the welfare gains from trade are not even areas 2 and 4 in figure 1, but simply area 4; that is to say, the consumption effect only. So an application of partial equilibrium analysis to evaluate trade impact goes further than its textbook equivalent by completely ignoring the supply side in its welfare calculations. (Producer surplus does of course still feature in a distributive sense but not in an aggregate sense, as an increase (decrease) will be completely offset by changes in consumer surplus).

    I would exclude the comments on import elasticities used by WITS, not because the issue of more accurate demand elasticities do not matter, but because the focus of the paper is on the neglect of the production side of the economy.

    3. Marshall’s view of surpluses and their use in trade analysis.

    The standard textbook use of producer and consumer surpluses to evaluate the effects of trade, derive from Marshall’s partial equilibrium analysis. While it was Marshall’s successors ‘Henry Cunyghame, who is credited with applying Marshallian curves to analyse the impact of trade on a single market in two countries, and Enrico Barone, who drew the now familiar welfare conclusions, Marshall did apply his analysis to trade using consumer and producer surplus to evaluate the effects. This was however not published at the time, for reasons that are not clear. Possibly Marshall was not satisfied with this application? [Whitaker, J. ed. (1975). The Early Writings of Alfred Marshall, 1867-1890, Volume 2, London and Basingstoke: Macmillan Press Ltd.]

    However, it is the case that in Appendix K, Marshall talks at some length about surpluses, where he identifies two types of producers; workers as producers and owners of equipment as producers. In the context of this paper, what is of note is that he questioned if producer and consumer surpluses could be evaluated independently ‘when we have reckoned the producer’s surplus at the value of the general purchasing power which he derives from his labour or savings, we have reckoned implicitly his consumer surplus too … (Marshall, 1920, p. 683).

    He also implies that much of the remuneration of workers and capitalists may take the form of a surplus. In the case of the former, he identifies the following reasons: (i) compensation is at the marginal rate and; (ii) work may give the worker pleasure (p. 683). Rather intriguingly (and relevant given Linotte’s advocacy of a broader approach to welfare), he later says that a worker’s work or ‘man’s activities are ends in themselves as well as means of production’ (p. 684). Similarly, much of capitalist earnings may be in the form of a surplus since compensation is at the marginal rate and many savers would actually pay to keep their savings safe, as opposed to requiring positive remuneration. However, ultimately he sees producer surplus, as a remuneration that is not necessary to induce supply.

    As Neil Hart indicated in his earlier comment, in Marshall, one can find support for different perspectives.
    So it is worth stressing that one interpretation of Marshall is that he had a broader conception of producer rents/surplus than subsequent textbook and applied model applications, and that he asserted that producer surplus subsumed consumer surplus so that measuring both could amount to double counting. So WITS modelling, with its exclusive focus on consumer surplus as welfare measure is a far cry from Marshall’s more nuanced discussion of the issue.

    Back to the leitmotif of Linotte’s paper, if much of the earnings of workers (and capitalists) are in the form of rents, then the impact of trade is much greater than implicitly assumed when measuring areas above the market supply curve. Import competition may cause a sector to shrink, and even if workers find alternative work, it may be less well remunerated. So, displaced workers will suffer loss of earnings (or wage rent), that is, if they actually find alternative employment. If they do not find alternative employment, their earnings loss may be all of their previous wages (less any social assistance), while the welfare consequences could be much more devastating than simple income loss. Conversely, given the focus on the importance of production, an expansion of a sector as a result of trade liberalisation, may confer gains beyond traditional producer surpluses, if we accept that worker and capitalist earnings contain a large proportion of surplus.

    I would recommend that Linotte eliminates section 4 and merges 5 and 6 with an extended discussion on how earnings may contain large elements of rents, which can be lost (or enhanced) with trade, and that contemporary trade analysis glosses over these important supply side realities, with its excessive focus on consumption and consumption gains.

    4. I would also suggest the elimination of sections 7, 8 and 9, to be replaced by a single section that distinguishes between the impact of trade on earnings (or rents) as opposed to its effect on employment status. It is too diffuse as it is currently written. In the revised new section, I would recommend ignoring discussion of effective rate of protection (ERP), it is not necessary and distracts from the main thesis. His point (as I see it) is that workers have, by and large, not done well from globalisation, as evidenced by poor wage growth in developed countries over the last 40 years, and by high levels of unemployment. Loss of wages is a serious producer loss, while unemployment can have social consequences for wellbeing that are not visible from simple income statistics.

    I think the penultimate section is very thought provoking and challenging; moving the discussion of production status and welfare beyond the facile concept of rents. It would link back to Marshall too, where he didn’t deny the complexity of identifying the part of workers’ earnings that were rent, and where he alluded to the importance of work, beyond being a means to an end. He explicitly said ‘The problem is a much philosophical as economic’ (p. 684). Deploying philosophical concepts, whether it be Sen’s capabilities approach, or the importance of ‘doing’ in for Nozick’ would give more insightful understanding of the effect of policies, such as trade policy on citizens. I would also suggest that this be the concluding section.

    Marshall, A. (1920), Principles of Economics, MacMillan Press, 8th edition.
    Whitaker, J. ed. (1975). The Early Writings of Alfred Marshall, 1867-1890, Volume 2, London and Basingstoke: Macmillan Press Ltd.

  • Daniel Linotte says:

    Eithne Murphy’s well-documented comments are most interesting and stimulating. The paper can certainly be shortened and focused differently.

    Murphy provides additional information about Alfred Marshall’s thoughts; as she mentions, he was almost ready to publish his views about the welfare dimensions of international trade.

    She also underlines the influence of Enrico Barone, who proposed a welfare analysis of international trade, relying on Marshall. However, Barone excluded the workers from his paradigm; they were left out, made ‘invisible’. Such a view could have a ‘psychological/behavioral’ explanation: suppliers and buyers can be seen as active agents who decide about the level of production and consumption – workers would rather be passive agents, they ‘just’ produce what is requested and have no say at all regarding the level of production(1).

    With a new reading of Marshall, the workers are made explicitly and fully visible; they come into existence – a status that corresponds to … reality.

    As Murphy confirms, their welfare condition is also related to their employment status, which implies the need to move beyond income rents (when assessing the impacts of trade liberalization) by adding ethical perspectives – which can be found in the contributions of Rawls, Sen, Nozick, and others.

    (1) It is perhaps worth mentioning that, according to Barone (in the “The Ministry of Production in the Collectivist State”, published in 1908), socialism can be conciliated with economic efficiency by mimicking the functioning of the market (as explained in textbooks, which does not necessarily correspond to realities). The so-called theory of decentralized planning is largely based on Barone’s contribution – see G. M. Heal, The Theory of Economic Planning, North Holland Publishing Co, 1973. Thus, paradoxically, one of the intellectual fathers of what could be called ‘mathematical socialism’ ignored the workers when writing about international trade and welfare.

  • Daniel Linotte says:

    Following Eithne Murphy’s comments, I would propose the following changes:
    1. Provide more explanations about the gains of trade liberalization, referring the areas 2 and 4 in Figure 1.
    2. The section on WITS (World Integrated Trade Solution) can be shortened by erasing the paragraph on the linkages between estimated elasticities and countries GDP.
    3. The section on effective protection can be deleted.
    4. The sections on wages dynamics and the costs of unemployment can be merged and shortened.
    5. As suggested by Murphy, the paper may end with a short “discussion of production status and welfare beyond the facile concept of rents.”
    I think these changes may help better focus the paper and, as a result, strengthen its arguments.

  • Constantinos Repapis says:

    Linotte’s paper is an interesting take on Marshall’s producer and consumer surplus approach in relation to trade liberalisation. It discusses how partial equilibrium arguments are employed by international organisations like the WTO and operationalised in WITS- which is a trade liberalization simulation tool proposed on the website of the World Bank.

    Eithne Murphy has done an extensive review of the paper above, and I only intend to focus on a few minor revisions. As Linotte himself acknowledges, the paper is long, and some cuts will be needed. Therefore section 7 can be erased and 8 and 9 merged.

    Furthermore, the discussion on the two figures (figure 1 and 2) needs to be more detailed. For example in section 4 Linotte mentions that “by introducing S2, the costs of trade liberalisation for the producer is larger than the traditional one, as reported in figure 1.” Then an explanation on why they are larger should be discussed.

    The only addition to the above that needs to be considered is what Murphy notes in her last paragraph, and also is something that I also thought the paper needs more clarity on, which is the way Marshall uses these tools and how this has changed in the subsequent literature. I do not think this needs to be discussed at length, or, indeed we need to go into a history of the changes of the term, but, a paragraph or two on Marshall’s use and his methodology of the context of these tools would add clarity to the paper and some depth in the argument. It also will signal the difference between Marshal’s analysis and that employed by international organisations- which do not pay respect to context of application of partial equilibrium analysis in the same way.

    Other minor notes:
    Graph 2’s interpretation is abit unclear. It shows the relation to import demand rather than size of GDP. In any case I agree that this paragraph can be erased without loss.
    Point III in final remarks. More detail is needed on the distinction between the two curves.