1. IntroductIon¹

The main theoretical argument against tariff and non-tariff trade barriers is that they cause social welfare loss; hence, the rhetoric for most trade agreements focuses on their elimination so that welfare gains will improve and enhance opportunities for all people.²

On the other hand, recent developments in world trade barriers, mainly driven by the United States of America (USA) -for instance Trump’s steel and aluminum tariffs on Canada, Europe and Mexico-,³ indicate that an era of world trade wars and protectionism could be knocking on the door. While the European Union (EU) can be considered a peer to the USA in a given degree, and retaliation measures might be handled on an eye-for-an-eye basis, the USA-Mexico relationship has clearly being an asymmetrical one and, as a Less Developed Country (LDC) what Mexico needs more than trade wars and/ or retaliations, is a committed government to develop the Mexican productive apparatus using, among other things, barriers to protect the construction and development of strategic and/or key sectors.

Another frequent argument in favor of “free trade” is that it would protect consumers from domestic market power, but if a domestic specific industry is too underdeveloped or does not exist, then simply there is no whom to protect the consumers from.

It remains, for a broader research agenda, to estimate something that mainstream free traders seem not to take into account too often: if a quota is imposed to protect and develop a given (sub)sector, to what extent gains from new employment, from integration and development of that (sub)sector (thus from synergies), increases in multiplier effects, scale economies, spillover effects, and others, would countervail welfare losses caused by that quota?

The main purpose of this paper is to carry out an estimation exercise to asses a single non-tariff barrier -an import quota on intermediate goods for a single (sub)sector-, quantify its welfare effects, and to see whether it would be advisable, on the grounds of an equivalent variation measure, to impose that import quota in the case of the Mexican economy. To that end, we design and implement a full-fledged, but parsimonious, Applied General Equilibrium Model for Mexico (Agem-Mx12q) using a 2012 Social Accounting Matrix (SAM-Mx12).⁴

This paper advances the literature, we believe, by clearly showing that a quota on imported inputs of a given key/strategic (sub)sector does not necessarily generate a significant welfare loss and would stimulate its development and consolidation.

The paper is organized as follows. In section 2, the literature on import quotas and welfare is reviewed. In section 3, we describe the data, the model and discuss macro-closures. In section 4, simulations are explained and implemeted, and results are analyzed. Section 5 concludes with some final comments.

2. BrIef lIterature revIew on Import quotas

The Great (World) Depression (1929), among other changes, marked a new era in world trade practices and, as early as 1933, a paper by Dietrich analyses the system of import quotas in France, noting that “Though import quotas are a hindrance to the free flow of international trade, there is the question whether they are not an instrument of economic control which may be of value in economic planning” (^{Dietrich, 1933, p. 661}). Dietrich’s main conclusion is that “There can be no question but that France instituted her quota system as a first-aid measure to safeguard the domestic market for French producers, agricultural and industrial, and to reinforce her protective tariff system, crippled by consolidated tariffs in commercial treaties. As such they can be considered a qualified success […] [it] would seem to indicate that France is beginning to look upon the quota system as a permanent instrument of her trade policy”. Dietrich finishes stating: “If, in truth, we have reached ‘the end of laissez faire’, the corollary of which has always been free trade, it would seem as if import quotas merit consideration, not as a weapon but as an instrument of economic control” (^{1933, p. 674}).

Import quotas in the USA were analyzed in a ^{1937 paper by Whittlesey}. Since then, research on import quotas has covered a wide spectrum, reaching sometimes well-stablished conclusions, but not always. In what follows we mention some of the issues just to illustrate that wide spectrum, showing the importance of import quotas, and how they might have diverse effects.

With respect to effects on quality, it is well-known theoretically and empirically that import quotas lead to higher quality products being exported (^{Feenstra, 1988}; ^{Harrigan and Barrows, 2009}; ^{Khandelwal et al., 2013}; ^{Upward and Wang, 2016}).

With respect to effects at the firm level, ^{Upward and Wang (2016, p. 71)} assert: “We study how Chinese textile and clothing firms adjusted the product structure of their exports to the US, as triggered by the termination of Multifiber Arrangement (MFA) quotas. We find that the removal of MFA quotas induced firms to expand their product scope while reducing the concentration on their core product”.

^{Sievers and Schaffer (2016, p. 497)} analyze “the impacts of the German biofuel quota on sectoral domestic production and imports of the German economy. […] The results indicate a clear increase of domestic production and a decline of net imports in the first case. In the second case gains in domestic production are smaller and net imports are, in contrast to the first case, increasing”.

Equivalence among several taxes, volume quotas and share quotas, has also been studied at large (^{Okumura, 2015}; ^{Francois and Woerz, 2009}). Effects on exports of the incumbent country (^{Ülengin et al., 2015}). Effects on domestic firms’ market power (^{Blonigen et al., 2013}). Best import quota policy (^{Teimoury et al., 2013}), etc.

Studying the economy wide impacts of South African quotas on the imports of clothing and textile products from China, ^{Naudé and Rossouw (2008, p. 737)} find that “contrary to the motivations apparently underlying the quota implementation, the macro-economic, sectoral and household effects are negative and result in greater inequality between poorer and richer households. […] The policy implications are the imposition of these quotas could come to be seen as a policy mistake, and that South Africa may benefit more from considering a free trade agreement with China”.

Another alternative to import quotas are “voluntary restrictions”. According to ^{Johnston and Parajuli (2017, p. 135)}, while studying the share of Canadian lumber in the U.S. lumber market, in order to protect its domestic industry: “Alternatively, Canada could voluntary limit its share of the U.S. market through voluntary exports restrictions, and capture rents that more than offset any losses to its domestic producers”.

The fact is then that import quotas continue to be of interest, and they are in place for several reasons in several countries. For Mexico, the “Trump Era” has several implications related to the NAFTA and other issues, such as Foreign Direct Investment (FDI).

This paper is mainly concerned with the effects of quotas on welfare following the line of works such as ^{Naudé and Rossouw (2008)}, and with the quantitative effects of an import quota following the line of ^{Krugman (1995, p. 31)}:

Economists have […] long preached the virtues of free markets […] free trade in particular. The logic that says that tariffs and import quotas almost always reduce real income is deep and has survived a century and a half of often vitriolic criticism nearly intact. […] Yet there is a dirty little secret in international trade analysis. The measurable costs of protectionist policies -the reduction in real income that can be attributed to tariffs and import quotas- are not all that large. […] These costs are very real, but when you try to add them up, they are usually smaller than the rhetoric of free trade would suggest. For example, most estimates of the cost of protection in the United States put it well under one percent of GDP. […] That is, the standard estimates suggest that a highly protectionist developing country, by moving to completely free trade, would get a one-time economic boost equal to the growth China achieves every five or six months.

^{Chao et al. (1990, p. 160)} stated that “Despite the fact that there has been a rich literature on the issue of the equivalence between tariffs and quotas, very little has been written on the welfare effect of quotas”, and in their paper they show that “[…] a tightening of import quotas generally reduces welfare. Quotas can improve welfare of a small economy under a stringent set of conditions”.

^{Schweinberger (2003, p. 696)} studied import quotas on intermediate goods in relation to special economic zones, and under a set of assumptions,⁵ he concludes: “It is proven that there generally exist allocations of binding quotas to the domestic and special economic zones which improve welfare.

Then it is shown that the abolition of the quotas on imports into the special economic zone may reduce welfare”.

On the other hand, under international oligopoly, ^{Furusawa et al. (2003, p. 31)}, analyze welfare-enhancing policies, concluding: “We show that even if policy makers have no idea about costs and demand, they can raise welfare by introducing a small production subsidy. If the government knows that demand is not very convex, a small tariff can be used to enhance welfare. With strategic complements, a small import reduction by an import quota deteriorates welfare”.

^{Kreickemeier (2005, p. 194)} analyses the welfare effects of tariffs and quotas in the presence of involuntary unemployment, concluding that “[…] employment effects constitute only one part of the welfare effects triggered by a change in the trade policy. It is clear from the analysis that reforms that yield negative employment effects can nevertheless increase overall welfare”.

Also, ^{Fan and Fan (2005, p. 1)}, found that “[…] there are substantial differences in the welfare effects of an equivalent tariff and quota designed to achieve identical import restrictions. When there is an increase in demand, the deadweight loss from an import quota will be greater than that of a tariff”.

In a more recent paper, ^{Winchester (2009, p. 820)} stated that “A growing literature focusing on welfare implications following the removal of [nontariff barriers] FDI has emerged. We contribute to this research by estimating ad valorem tariff equivalents (TES) of FDI and simulating reductions in tariffs and FDI in a computable general equilibrium (CGE) model”. ^{Winchester (2009, p. 820)} found that “Our simulations reveal that bilateral negotiations between New Zealand and four possible FTA partners (China, Japan, Korea, and the ASEAN block) will increase New Zealand’s welfare by 1.5% when tariffs are eliminated and 16.3% when tariffs and FDI are abolished”. Therefore, the study concludes “[…] we find trade economists do not have a dirty little secret”.

Even assuming Winchester’s calculations are correct,⁶ that conclusive remark does not seem very fair, since ^{Krugman (1995)} was talking only about quotas, while Winchester considered tariffs and NTB’s together.⁷

Then, ^{Chen et al. (2011, p. 549)} found that “[…] quotas can result in a welfare advantage over tariffs for an importing country despite that its government does not capture any quota rents. Specifically, the conditions under which an equivalent quota yields higher expected welfare than a tariff are shown to depend on a set of economic variables”.

3. Data, model, and macro-closures

Based on Input-Output Table (IOT) published by ^{INEGI (2018a)} for 2012 and additional data from national accounts, we build a SAM in order to design and calibrate an AGEM.⁸

The Social Accounting Matrix of the Mexican Economy for 2012 (SAM-Mx12), was built following the North American Industry Classification System (NAICS), this SAM has two versions, the first one with 19 productive Activities and the second one with 47 Activities.

This SAM-Mx12 provides a transparent dataset built using official data from the Mexican institute of statistics, and a clear, parsimonious methodology. It can be extended to include more detailed data on energy, environment or any other data disaggregation of interest. It can also be modified, corrected, and/or extended by other researchers to perform diverse experiments trough structural analysis, applied general equilibrium, or other methodologies.

The SAM-Mx12 provides information on the structure of the Mexican economy as of 2012 at the 19 and 47 productive Activities level of disaggregation. It comprises the whole set of elements constituting the yearly circular flow of the economy, specifying levels of income and expenses for every agent: Households, Government, Firms, and the Rest of the World.

The SAM-Mx12 was entirely elaborated using official data from the INEGI. To begin with, we took the IOT published by ^{INEGI (2018a)} and rearrange its data into a social accounting matrix format (see Table 1).

To get rid of the statistical discrepancies we divide them, proportionally, among the elements of final demand. Then we add accounts for Income tax and Private consumption and use data from the Mexican System of National Accounts (^{INEGI, 2014} y ^2018b), to fill in the gaps and balance the matrix presented in Appendix 1.

According to the SAM, the AGEM is specified as follows: we design a neoclassical model (^{Barro and Sala-i-Martin, 2004}), where the economy has a fixed endowment of non-differentiated capital and labor. Under full employment, total demands of factors always equal total endowments. This also implies full-factor mobility among activities and, therefore, a long-term horizon since, eventually, capital can be transformed and allocated to other sectors. The same applies to labor.

Aggregated households are the first account of our SAM, obtaining income from capital, wages, government transfers, and transfers (remittances) from the Rest of the World (RoW). On the other hand, households pay taxes, save, and buy goods for present consumption, which comprise one aggregated domestic good and one imported good.

The government collects revenues from income tax, from products and production taxes, and from capital imports. Then the government makes transfers to households (mainly poverty alleviation programs), saves, and consumes domestic goods and an aggregated imported good.

Savings come from households, activities, the government, and the RoW. Total savings are invested in domestically produced capital goods and one aggregated capital good from the RoW.

Production of total supply (gross product) is carried out as follows: in the first place, value added (VA) is produced combining capital and labor (composite factor), then VA is combined with domestic inputs to obtain the domestic product, and finally, foreign inputs are added to get the total gross product.

The RoW obtains income from imports: inputs, capital goods, final consumption of households and government, and capital payments. Then spends on transfers (remittances), documented labor, exports, and savings.

Following ^{Lofgren et al. (2002, p. 16)}, we specify a modified Johansen closure to avoid “[…] the misleading welfare effects that appear when foreign savings and real investment change in simulations with a single-period model”.

Therefore, we fix real investment so that households’ marginal propensity to save will adjust to meet any changes in prices. As for the RoW, we fix the exchange rate (er) and foreign savings, and, also, transfers to households and payments to labor, assuming that non-documented and documented labor will remain constant, leaving exports as the adjustment variable. Government expenditures are fixed in real terms, including savings, in such a way that when a quota is imposed, the rent generated will accrue to government revenues so that a surplus will be generated. In turn, this surplus will be used to reduce taxes on production and products.

Following ^{Hosoe et al. (2010)}, we model quotas through a complementarity condition in order to obtain quota rents, which in turn will accrue to government revenue and used to implement a revenue neutral decrease in production taxes (Appendix 2, equations 34-36).

The usual caveats apply along with the results from ^{Fugazza and Maur (2008, p. 475)}: “We find that serious estimation and modeling efforts remain to be undertaken in order to make CGE modelling a useful policy tool to analyze FDI. Casual policy inferences from loose specifications may, indeed, lead to analytical mistakes”.

Appendix 2 contains the mathematical formulation.

4. SimulatIon and results

Assume that the government decides to protect and foster the development of a given productive sector. The main simulation we implement is the imposition of a quota on intermediate consumption imported by the Manufacturing industry, while at the same time, the rents generated by this quota are used to reduce taxes paid by the Manufacturing industry itself. To this end, we use the sam-Mx12 with 19 productive sectors (Activities), where Activity 5 (Manufacturing industries) amounts to 34.3% of total gross product, and its imports to 76.6% of total input imports. Since our focus is on welfare, after solving for the new equilibrium, we compute Hick’s Equivalent Variation (HEV). See Table 2.⁹

Indeed, according to theory, households’ welfare diminishes, but if we observe the present consumption of domestic goods, it barely decreases by 0.43% (not even half a percentage point). Present consumption of imported goods shows an increase, although very small.

Let’s assume that stronger protectionism is enforced and the quota goes to 50%, instead (see Table 3). As expected, subsidies to the Manufacturing industry increase up to 9.7% since quota rents will increase when lowering the quota. On the other hand, as one would expect, households’ welfare decreases, since a greater distortion is introduced into the economic system, but even in this case, it only amounts to about 2.6% in terms of domestic present consumption. But on the other hand, consumption of imported good increases by 0.5% and, most importantly, Household’s savings increase by 3.8%.

However, quotas are generally imposed on much smaller subsectors. Assume then, that the government desires to protect and develop the Chemicals industry, which was chosen mainly because it is a key subsector¹⁰ in most major economies of the world, which is to say that, the Chemical industry is a strategic upstream (downstream) sector that supplies (demands) a wide range of intermediate inputs to manufacturing activities, amplifying its potential spillover effects throughout the economy, and therefore it is a good candidate to protect through a quota. Besides, international experience identifies the chemical sector as a core component of industrial policy in both developed and emerging economies, due to its role in innovation and value-chain integration. Although the Automotive industry or other subsectors could also be considered, especially those with strong import reliance, such extensions are left for future research.

This time we look directly at a quota of 50% on intermediate inputs imported by this subsector, using the same SAM but now disaggregated to 47 subsectors, where the Chemical industry (Activity 19 in the 47 subsectors SAM) accounts for only 3% of gross product (see Table 4).

Not surprisingly, welfare loss is dramatically smaller, and present consumption of domestic good decreases in only 0.24% (not even one fourth of a percentage point), while at the same time consumption of imported good increases by 0.16% and, more importantly, Households savings increase by 0.23%.

Although welfare losses remain negative across all simulations, which is theoretically correct, we consider their magnitude is modest in the sense stated by ^{Krugman (1995, p, 31)}:

These costs are very real, but when you try to add them up, they are usually smaller than the rhetoric of free trade would suggest. For example, most estimates of the cost of protection in the United States put it well under one percent of GDP. […] That is, the standard estimates suggest that a highly protectionist developing country, by moving to completely free trade, would get a one-time economic boost equal to the growth China achieves every five or six months.¹¹

The finding that welfare losses are modest can be explained by the model’s transmission mechanisms. In the AGEM-Mx12q, quota rents are fully transferred to the government and recycled as production tax reductions, partially offsetting the initial distortion generated by the import restriction. Moreover, the assumptions of full employment and perfect factor mobility allow capital and labor to be efficiently reallocated toward sectors with higher domestic demand, which mitigates welfare losses at the aggregate level. The overall negative impact stems mainly from a mild increase in domestic goods’ prices; however, this is counterbalanced by a slight expansion in imported consumption and a fiscal stimulus to the protected sector. As a result, the economy experiences only a limited decline in real welfare. The simulations thus highlight that the welfare impact of import quotas depends crucially on their fiscal recycling mechanism and the structural features of the economy. When quota rents are reinvested to reduce production taxes and factor markets adjust flexibly, welfare costs become minimal.

5. Final comments

In this paper we build a SAM and design an AGEM for Mexico to analyze the welfare effects from quotas on imports of intermediate goods.

Contrary to some authors (v.g.^{Winchester, 2009}), we find that import quotas do not generate significant welfare losses, hence, the “dirty little secret” stated by ^{Krugman (1995)} might be in place.

Policy implications in terms of international trade and the development of Mexico’s productive apparatus are of the greatest importance: protection and development of specific industries through import quotas on intermediate goods, should not be discarded by default on the grounds of “The logic that says that tariffs and import quotas almost always reduce real income” (^{Krugman, 1995, p. 31}). Rather, a sound evaluation of the case at hand should be carried out.

It remains the fact, of course, that other negative effects might unfold (retaliation among them), but also, other positive effects should be considered (employment, wages, development and integration of the productive apparatus), and then again, “[…] serious estimation and modeling efforts remain to be undertaken in order to make cge modelling a useful policy tool to analyze FDI” (^{Fugazza and Maur, 2008, p. 475}).

At any rate, development requires a decided impulse from committed governments to protect and foster specific key (sub)sectors if an efficient and sufficient national productive apparatus is to be developed.

The point is not that small gains are not worth obtaining, but that an import quota could be a helpful tool in fostering and developing specific productive activities. At the end of the day, the question of whether restrictions would be beneficial in practice, even though they can be in theory, is an empirical question.

Summarizing, the essential point is that an import quota, per se, does not necessarily imply a significant loss of welfare (if any at all), and it can be a useful policy tool to foster economic development, particularly in LDCS where some key productive activities are far from a sensible state of development.

Consequently, the central finding of this analysis is not that import quotas are devoid of costs, but rather that under a framework where the resulting rents are strategically recycled-as simulated through a revenue-neutral reduction in production taxes, which could be effectively implemented through import permit auctions-these static welfare costs could be so modest that they could effectively be offset by the prospective dynamic benefits of industrial development. While this model does not capture long-term gains such as employment generation, knowledge spillovers, enhanced sectoral linkages, and innovation, it is precisely these dynamic advantages that constitute the core rationale for such industrial policy.