Thesis
Reference
On the transmission of prices and of market access shocks
MC LAREN, Alain
Abstract
Two of the main shocks that affect international trade are price and tariff changes. The present thesis covers these two topics. The first chapter shows that the transmission of international agricultural prices to local producers is asymmetric, with downward price movements being transmitted more strongly than upward price movements. This seems to be due to the market power that intermediaries have on these markets. The second chapter looks at the way Peruvian firms are affected by market access conditions and it shows that it isn't only the direct tariff that firms face that plays a role but also the tariff he faces relative to the tariff that competitors from other countries face. The last chapter shows that the effect of these relative tariffs depends on the distance to the destination country, with the effect being stronger when exporting to closer countries.
MC LAREN, Alain. On the transmission of prices and of market access shocks. Thèse de doctorat : Univ. Genève, 2014, no. SES 863
URN : urn:nbn:ch:unige-402089
DOI : 10.13097/archive-ouverte/unige:40208
Available at:
http://archive-ouverte.unige.ch/unige:40208
Disclaimer: layout of this document may differ from the published version.
1 / 1
market access shocks.
Th`ese pr´esent´ee `a la Facult´e des
Sciences ´Economiques et Sociales de l’Universit´e de Gen`eve par
Alain MCLAREN
pour l’obtention du grade de
Docteur `es sciences ´economiques et sociales mention ´economie politique
Membres du jury de th`ese:
Prof. Marcelo OLARREAGA, Directeur de th` ese, Universit´ e de Gen` eve
Prof. Fr´ ed´ eric ROBERT-NICOUD, Pr´ esident du Jury, Universit´ e de Gen` eve
Prof. C´ eline CARRERE, Universit´ e de Gen` eve
Prof. Bernard HOEKMAN, European University Institute de Florence
Th`ese no 863 Gen`eve, le 4 septembre 2014
La Facult´e des sciences ´economiques et sociales, sur pr´eavis du jury, a autoris´e l’impression de la pr´esente th`ese, sans entendre, par l`a, n’´emettre aucune opinion sur les propositions qui s’y trouvent ´enonc´ees et qui n’engagent que la responsabilit´e de leur auteur.
Gen`eve, le 4 septembre 2014
Le doyen Bernard MORARD
Impression d’apr`es le manuscrit de l’auteur
R´esum´e ix
Summary xi
Acknowledgements xiii
General Introduction 1
1 Motivation . . . 1
2 Overview . . . 2
1 Asymmetry in price transmission in agricultural markets. 3 1 Introduction . . . 3
2 Theoretical background . . . 5
3 Data . . . 10
4 Empirical strategy . . . 11
4.1 Relationship between the international price and the pro- ducer’s price . . . 11
4.2 Asymmetry of price transmission . . . 12
4.3 Dealing with endogeneity . . . 13
5 Results . . . 15
5.1 Robustness checks . . . 15
5.2 Oligopsony and asymmetry . . . 16
6 Conclusion . . . 17
2 Market Access, Export performance and Survival: Evidence from Peruvian Firms. 31 1 Introduction . . . 31
2 Related literature . . . 33
2.1 Survival . . . 34
2.2 Export Performance . . . 35
3 Data . . . 36
3.1 Stylised Facts for Peruvian Exporting Firms . . . 36
3.2 Measuring market access . . . 38
4 Empirical strategy and results . . . 41
4.1 Trade Relationships Survival . . . 41
4.2 Export Value . . . 46
4.3 Sensitivity Analysis . . . 48 iii
5 Concluding remarks . . . 50
3 The discriminatory effect of a non-discriminatory relative prefer- ence margin. 61 1 Introduction . . . 61
2 Relative tariffs and overall costs . . . 63
3 Data . . . 64
4 Empirical specification . . . 65
5 Results . . . 66
6 Adding more destination specific variables . . . 68
7 Robustness checks . . . 68
8 Conclusion . . . 70
General Conclusion 85 1 Main findings . . . 85
2 Policy issues . . . 85
3 Future research . . . 86
Bibliography 87
1 Proportion of population in agriculture and wealth. . . 27 2 Price transmission for various international prices. . . 28 3 Relationship between international price and producer price. . . 28 4 Rainfall IV. In parenthesis the export share and the annual standard
deviation of rainfall in millimeters. . . 29 1 Market access conditions in MERCOSUR countries. . . 57 2 Market access conditions in non MERCOSUR countries. . . 57 3 Positive versus Negative Preference Margins (# of trade relationships) 58 4 Trade Relationship Status (frequency) . . . 58 5 Entry versus Continuing Trade Relationships: share in current year
exports . . . 59 1 Marginal effect of the RPM on the exports of firms, depending on
distance. . . 80 2 Top 20 destinations in terms of export spells. . . 81 3 Top 20 destinations in terms of export value. . . 82 4 The relationship between the RPM and distance for agricultural ver-
sus non-agricultural products. . . 83 5 Marginal effect of the RPM for non-agricultural products. . . 83 6 Marginal effect of the RPM for agricultural products. . . 84
v
1 Asymmetry of agricultural price transmission. . . 24
2 First stage regressions of the 2SLS. . . 25
3 GMM regressions. . . 26
4 Long and short run transmission. . . 26
5 Asymmetry explained by the export importance. . . 26
6 Cargill presence. . . 27
1 Survival of exports. . . 52
2 Survival of exports (weighted and unweighted TTRI and RPM). . . 53
3 Regression on the value of exports. . . 54
4 Survival of exports with marginal effects. . . 55
5 Survival of exports with marginal effects. . . 56
1 What explains firms’ exports. . . 71
2 Including sector-destination dummies. . . 72
3 Adding destination-year specific variables. . . 73
4 Excluding the elasticity of import demand and trade in the RPM. . 74
5 Excluding partners that have an RTA with Peru and closest partners. 75 6 Including outliers (with and without the elasicity of import demand). 76 7 Excluding the top export destinations in terms of number of export spells. . . 77
8 Excluding the top export destinations in terms of export values. . 78
9 Agricultural and non-agricultural products. . . 79
vii
Des chocs sur le commerce international sont fr´equents et peuvent fortement in- fluencer les partenaires commerciaux. Par exemple, des changements dans les prix internationaux vont non seulement affecter l’exportateur, mais potentiellement le producteur aussi. Vu la haute volatilit´e des prix il est important de comprendre les cons´equences de ces changements. Un autre exemple est celui des changements dans les conditions d’acc`es aux march´es, tels les tarifs, qui ont eu lieu ces derni`eres ann´ees avec la mondialisation. Des tarifs changeants influencent les partenaires commerciaux, et les m´ecanismes en jeux sont cruciaux pour comprendre comment ils sont affect´es par ces chocs. L’utilisation de plus en plus r´epandue de donn´ees au niveau de la firme sont ´egalement la source d’une meilleure compr´ehension de la fa¸con par laquelle le commerce est affect´e par des chocs au travers des firmes.
Cette th`ese regarde comment des chocs au niveau des prix internationaux en agriculture sont transmises aux producteurs locaux ainsi que comment des change- ments dans les tarifs affectent des firmes sur le march´e des exportations. Le premier chapitre de la th`ese montre que la transmission des prix internationaux vers les pro- ducteurs locaux est asym´etrique, `a savoir que les prix sont transmis plus fortement
`
a la baisse. Il met aussi en avant une explication pour ceci, `a savoir `a cause du pouvoir de march´e des interm´ediaires. Le deuxi`eme chapitre regarde comment les firmes p´eruviennes sont affect´ees par des changements dans les conditions d’acc`es au march´e quand elles exportent, et que ce n’est pas seulement l’effet direct des tarifs qui a son importance, mais aussi le tarif relatif auquel fait face un exporta- teur par rapport `a des comp´etiteurs d’autres pays. Le troisi`eme chapitre montre que ces tarifs relatifs auxquels font face les firmes exportatrices p´eruviennes ont un effet qui diff`ere selon la distance jusqu’`a leur destination, `a cause de diff´erences de coˆuts li´ees `a la destination en question dont notamment les coˆuts de transport.
Ces r´esultats sont int´eressants du point de vue de la politique ´economique. Le fait que l’analyse ainsi que les r´esultats montrent que les producteurs locaux peu- vent souffrir plus de baisses des prix du fait d’un pouvoir monopsonistique des interm´ediaires peut contribuer `a aider les gouvernements `a comprendre comment leurs producteurs souffrent de variations internationales de prix et comment mieux les soutenir. La compr´ehension des effets de l’acc`es au march´e sur les firmes au-del`a de l’habituel effet direct des tarifs mettent ´egalement de la lumi`ere sur ce qui influ- ence les firmes sur les march´es ´etrangers. Le fait que la distance ait un impact sur la magnitude de l’effet de ces tarifs relatifs tarif montre qu’une politique ´economique impos´ee d’une mˆeme fa¸con sur plusieurs pays peut avoir un effet discriminatoire envers certains pays exportateurs en particulier.
ix
Shocks to international trade are frequent and can strongly affect trading partners.
For instance, changes in prices will typically affect an exporter and may even affect local producers. Since price volatility is high it is of strong interest to know what the consequences of these changes are. Another example is changes in market access conditions, such as tariffs, particularly with the widespread globalisation that has taken place in recent years. Evolving tariffs affect trading partners, and what mechanisms are truly at play is crucial to understanding how they will be affected.
New insights are also coming to light with the use of firm level data for empirical analysis. This relatively new tendency, mainly due to new data availability, gives some new means to understanding how trade is affected by shocks, by understanding how these shocks are affecting firms.
This thesis looks at how price shocks in international agricultural prices trans- mit to local producers as well as how tariff changes affect the presence of firms on the export market. The first chapter of the thesis shows that the transmission of in- ternational agricultural prices to local producers is asymmetric, such that prices are transmitted more strongly for downward movements than for upward movements.
It also puts forward an explication for this, namely the market power of interme- diaries. The second chapter looks at how Peruvian firms are affected by changes in market access conditions when exporting, showing that it isn’t only the direct effect of tariffs that will affect firms, but the relative tariff they face compared to competitors from other countries. The third chapter shows that the effect of these relative tariffs on Peruvian exporting firms is different depending on how far they are exporting, due to destination specific costs.
These results are interesting from a policy perspective. The fact that the analy- sis and results point to the idea that local producers may suffer more from declines in prices due to monopsonistic power of intermediaries may help governments under- stand how their producers may be suffering from variations in international prices and how to support them better. The understanding of how firms are affected by relative market access conditions and not just by the usual direct market access measures also gives new insights on what affects firms on a foreign market. The fact that distance also has an impact on how strongly a firm is affected by relative tariffs sheds light on the fact that a same policy imposed on all countries could discriminate some exporting countries in particular.
xi
I would like to thank everybody who has contributed to this work by helping me in many different aspects of my work. First of all, I am grateful to Professor Marcelo Olarreaga, my thesis supervisor, for helping me along the way. Having done my Master’s dissertation with him in 2009, it was with delight that I learned that he had confidence in my work and the will to be my thesis supervisor should I want to pursue with a PhD. He knew how to find the right balance between letting me work on my own and setting regular meetings, either the two of us or with other PhD students, in order to guide me in my research. He was always very available despite having a huge workload when becoming the Director of the Economics Department.
I would also like to thank Doctor Marco Fugazza of UNCTAD, for enabling me to write my second paper during my consultancy period in UNCTAD during the summer of 2011 with him as co-author. I first met him as my teacher in the Bachelor course entitled ”Economie du D´eveloppement” and afterwards in the Master course called ”Economic analysis in international organizations”. When writing the paper we worked very well together and I learned a lot from him. He was always open to help me, even taking time to give me advice in my third paper.
I would like to thank Professor Fr´ed´eric Robert-Nicoud and Professor C´eline Carr`ere for actively participating in our weekly PhD student informal meetings and and for having many very excellent ideas and propositions concerning my research.
I thank them for being members of the jury for this thesis and in particular to Fr´ed´eric for taking up the role of the time consuming Director of the jury. I am also very grateful to Professor Bernard Hoekman for accepting to be my external examiner.
I would like to thank all participants of the annual ProDoc Trade PhD Work- shop, in particular Professor Jean-Louis Arcand for reading a draft of my first paper and coming up with many useful comments on how to improve it. Professor Richard Baldwin must also be thanked for some useful insights on my first paper while pre- senting at an FNRS meeting. I would also like to thank the participants of the Geneva Young Researchers Seminar for their comments and discussions, in partic- ular Professor Stephan Sperlich for some technical details concerning my second paper.
I also owe great thanks to the participants of the ”Bari Third International Workshop on Economics of Global Interactions: New Perspectives on Trade, Factor Mobility and Development” for their useful comments and suggestions.
xiii
1 Motivation
With the strong globalisation that has taken place over the last decades, the impor- tance of trade for countries’ economies is becoming particularly large. According to data from theWorld Development Indicators(2010), the average country’s share of exported value of goods and services with respect to Gross Domestic Product (GDP) has gone from 33% in 1985, to 37% in 1995, to 46% in 2005. This shows why the understanding of trade and more particularly what affects trade is important for our economies, therefore our welfare. In some cases, such as agriculture, poorer countries are potentially at a high risk when facing international negative trade shocks, especially for the farmers living close to the poverty line. This thesis will concentrate on two types of shocks, one on prices and the other on tariffs. In the first case the idea will be to look at how a price change is transmitted to producers, in the other on how tariff changes affect exporting firms.
The large percentage of poor countries’ population working in agriculture and the high price volatility of such goods are two very important reasons in seeking to understand the outcome of such a situation. Especially important is the fact that farmers tend be relatively small and use the services of intermediaries to sell their products abroad. It is important for policymakers to understand what this market structure implies in terms of exposure to conditions on international markets. This new knowledge may enable them to help the farmers in difficult times, especially those that will suffer the most.
The other shock often encountered and widely studied in international trade is the one created by tariff changes. Tariffs imposed on importers make the access to markets more difficult, and a decrease in tariffs tends to facilitate the situation of exporting firms. The availability of firm level data is becoming more widespread, which is helping researchers rethink as well as discover new aspects of the func- tioning of trade. In particular, the study of the so called intensive and extensive margins are at the heart of a new trend of theoretical and empirical analysis. The second and third chapters of this thesis use such a dataset of Peruvian firms to better understand how tariff changes affect trade at the firm level. On one hand by assessing the importance of not just tariffs but also of tariffs faced by firms relative to those faced by firms from other countries. On the other, by shedding light on the influence of distance on the impact of these relative tariffs, due to the existence of destination specific costs. The share related to tariffs in overall costs will be smaller in more distant countries, such that variations in relative tariffs will not induce the
1
same variation of overall costs depending on the destination country. This arises interesting policy questions, knowing that there are relative aspects to take into account.
2 Overview
The first chapter, entitled ”Asymmetry in price transmission in agricultural mar- kets”, explores the asymmetries in price transmission from international to local markets. In theory, the presence of large intermediaries in agricultural markets could lead to a stronger price transmission when international prices decline than when they rise. The empirical evidence confirms the presence of asymmetric price transmission consistent with the presence of large intermediaries with monopsony power.
The second chapter, ”Market Access, Export performance and Survival: Evi- dence from Peruvian Firms”, looks at the effect of market access on firms’ export performance and their survival on foreign markets. The data used covers all Peru- vian exporting firms between 2002 and 2008, a period during which Peru was active in joining the global economy. This is done using two indices, one that summarizes the tariffs faced by exports, the other that measures the preferential margin at the bilateral level. Results show that more than market access conditions per se, it is market access conditions relative to those faced by competitors that significantly influences export performance and survival. About a fifth of the increase of exports directed to Mercosur countries is due to improvement in preference margins.
As shown in the second chapter, tariffs will tend to have a negative impact on trade. The third chapter, ”The discriminatory effect of a non-discriminatory relative preference margin”, shows that that since the share related to tariffs in overall costs will be lower in more distant countries, due to higher destination specific export costs, the effect of a change in the relative tariff will be lower when exporting to further away countries. Empirical results using firm-level data from Peru confirm that firms exporting further away are less affected by changes in relative tariffs.
Asymmetry in price transmission in agricultural markets. ∗
”While grocery store shelves appear to provide abundant choices, most of these products are marketed by a small and decreasing num- ber of firms. Gigantic multinational corporations are consolidating their control over our food system [...]. The trend raises concerns about how this power is exercised, as most of these corporations are accountable to their shareholders, not to communities in which they operate.” Phil Howard (Howard (2006))
1 Introduction
Many poor countries have a large proportion of their active population working in agriculture. In many cases, the amount of revenue they receive will be crucial for their survival. Using data from theWorld Development Indicators (2010) one sees that in the year 2000, the world’s 25 percent poorest countries had at least a quarter of their population working in agriculture, and half of these 25% poorest had over 70% of their population active in agriculture. An illustration of this is given in Figure 1. Thus poorer countries will be more exposed to large falls in agricultural prices. Understanding the determinants of changes in agricultural prices is therefore crucial for the poorest countries.
In this paper the focus is on the extent of asymmetry in the price transmission from international to local markets. Whether a farmer is selling locally or exporting, the price he will receive for his production will be directly or indirectly affected by prices determined in world markets. Indeed, Mundlak & Larson (1992) show that variations in local agricultural prices are mainly explained by variations in world prices. But the transmission from international to local markets may not necessarily be symmetric. Depending on market conditions falls in international prices may be
∗A similar version of this paper has been accepted for publication in the Review of Develop- ment Economics. Earlier versions were presented at the ”Bari Third International Workshop on Economics of Global Interactions: New Perspectives on Trade, Factor Mobility and Development”
in September 2012 as well as at the ”ProDoc Trade PhD Workshop” in September 2010 and the
”University of Geneva Young Reseachers Seminar” in April 2010.
3
better transmitted to local markets than increases in international prices. The consequences of this asymmetric price transmission could be particularly harmful in poor countries where farmers often live close to the poverty line. In fact, Mosley
& Suleiman (2007) suggest that a portion of the small farmers are below the poverty line in some of the poorer countries. They also put forward that the one sector that has had a strong ability to stimulate pro-poor growth processes, especially in East and South Asia, is smallholder agriculture.
Why would one expect a better price transmission when agricultural prices fall?
Agricultural markets are characterized by the presence of large international inter- mediaries, with strong monopsony power over often small and numerous producers.
Murphy (2006) shows that in the United States two companies (Cargill and Archer Daniels Midland) export 40 percent of all U.S. grains. Rogers & Sexton (1994) show that in the United States more than 60 percent of all food and tobacco markets can be considered as noncompetitive when measured by their top four-firm concentra- tion ratio (with a threshold at 50 percent). Figures for other countries are similar.
For example in Vink & Kirsten (2002), concentration ratios for the four largest firms for South Africa are also large: 47 percent in slaughtering, dressing and packaging livestock, 65 percent for vegetables and animal oils and fats, 43 percent for flour, 37 percent for animal feeds, 99 percent for sugar, golden syrup and castor sugar, 80 percent for coffee, coffee substitutes and tea.
In this paper it is shown that in the presence of strong monopsony power of agri- cultural intermediaries with sufficiently convex marginal cost functions one should expect an price transmission that is stronger following a price decline than for a price rise. This is consistent with the use of this monopsony power by intermedi- aries. Indeed as international price falls, local prices will fall proportionally more than when international prices increase. This prediction is confirmed when con- fronted to a sample of 161 agricultural products produced in 117 countries over a period of 35 years. Moreover, the asymmetry seems to be driven by the results for markets where large international intermediaries are present or when exports represent a large share of total production which increases the monopsony power of international intermediaries.
Questions of asymmetric price transmission have been widely studied for oil markets, known as the literature on ”Rockets and Feathers”, where prices rise like rockets but fall like feathers. Many researchers have been analyzing the evolution of oil output prices. Tappata (2009) analyzed the theoretical aspect of potential asymmetric responses of oil retail prices. Many empirical studies come to the con- clusion that asymmetry does exist. This is the case of studies such as Galeotti et al.
(2003) or the main study on the topic done by Borenstein et al. (1997). The latter has several explanations for asymmetry. The one than seems to be closest to what is found for agriculture is one of costly search of consumers. The idea is that the consumer will believe that a change in price at the retail station during a period of volatile crude oil prices will really be due to a change in this price, whereas in less volatile periods he will believe that the station’s margin is changing. He will put more effort into searching for a lower price elsewhere if he believes that it is that specific station increasing its price. This is because the consumer’s expected gain of search is higher in this case. This, in turn, leads to a higher market power
of retailers who can dampen the rate of passthrough of upstream price decreases.
This paper shows similar effects for agriculture, however affecting the producer side rather than the consumer side.
The results are interesting from a policy perspective and give room for ac- tive competition policies when facing large international intermediaries. Levinsohn (1994) suggests that many countries are more lax in their competition policy when dealing with export markets, perhaps because anticompetitive practices in an ex- port market will not be harmful for domestic consumers. However, the asymmetric price transmission identified in this paper would lead to increased losses for often poor farmers. More generally, as argued by Murphy (2006), market power in inter- national markets is not factored into the models and assumptions that inform the trade and agriculture debate, which can mislead policy makers in terms of the dis- tribution of the gains from trade liberalization in these markets. This result can also offer some interesting perspectives in terms of aid-effectiveness. Some studies such as Mosley & Suleiman (2007) have made clear some important aspects of aid, such as the importance of food crops due to their high poverty leverage. The result of this paper concerning intermediaries may complement such studies, proposing some strategies that may be adopted within the agricultural sector. An asymmetric price transmission could also alter the usual way of approaching the analysis of welfare linked to price changes. De Hoyos & Medvedev (2011) study the changes in wel- fare created by the price transmission from international prices to domestic prices.
They put forward the importance of the degree of price transmission in explaining short to medium-term poverty effects. The presence of asymmetric transmission can give new insights to such theories and a new way of thinking about the effects of international price changes for welfare.
The remainder of the paper is organized as follows: section 2 presents a simple model to illustrate how a producer’s price decline may be larger than the potential rise that follows an international price change. Section 3 presents the data used to empirically explore this question and section 4 presents the empirical strategy.
Results are presented in section 5 and section 6 concludes.
2 Theoretical background
This section presents a basic model in which a one time departure from equilibrium will lead to a stronger producer price fall following an international price decrease than the producer price rise that would follow an international price rise. This effect will be due to a concave relationship between both prices.
Agricultural markets are characterized by a large dispersion of farmers, as noted by Sexton (1990), which are numerous and therefore act as price takers. He also emphasizes the bulkiness and/or perishability of raw products that will have an influence on market structure. This could lead to what he calls spatial oligopsony power of processors or wholesalers. As quoted in OECD (2008) it is the difficulty for sellers to find other buyers which determines the extent of a buyer’s monopsony power. Murphy (2006) says that most farmers lack the storage and capital needed to get their goods to distant markets, so they are left selling locally, to middle-men
who now have more suppliers to choose from.
Market power is modeled based on Sexton (1990). Wholesalers are considered homogeneous and act as price takers in their selling markets, notably due to the size of international markets. The fixed cost associated with exporting is too high for the individual farmer to face, such that he has to pass by a wholesaler. This is consistent with the literature, notably with Gopinath et al. (2007) who say that agriculture is unique as farmers often do not export directly since it is marketing firms that make the export decision.
The international price of a good is denote by P∗ and is considered as being exogenous. The wholesalers therefore face an infinitely elastic demand and it is assumed that they benefit from economies of scale, such that there will be much fewer of them than farmers. Defining the price received by the farmer aspf i, such that one can write the supply of one farmer as in equation (1).
qis = qi(pf i) with qi(pf i)0 >0 and qi(pf i)00 ≤0 (1) By summing over all farmers one gets the aggregate supply of farmers to a processor, presented in equation (2).
Qs = Qs(pf) with Qs(pf)0 >0 and Qs(pf)00≥0 (2) For later use one can define the inverse supply function as in equation (3).
(Qs)−1(Q) = pf (3)
This inverse supply will be called w(Q) which leads to the expression given by equation (4).
w(Q) = pf with w(Q)0 >0 and w(Q)00 ≥0 (4) The wholesaler must either package the products or incur some extra costs to export the product. These costs will be increasing in the quantity of raw productQ and will be given by the function m(Q). For higher quantities, the wholesaler will have to pay even higher costs to get products from farmers that are further away.
Besides these costs the wholesaler must incur the cost of buying the product, w(Q)Q as well as a fixed cost to export. Total cost is given by equation (5).
c(Q) = w(Q)Q+m(Q) +f (5) The fixed cost to export mentioned above is denoted by f and is added to the model such that there are economies of scale. This will spread out the different wholesalers geographically, instead of having one at each farm site.
The next step is to depart from the analysis of Sexton (1990) to see what will happen if there is an exogenous shock to P∗, as illustrated in Figure 2.
Due to the fact that an increased quantity must at least in part be supplied by farmers that are further away, which comes on top of the usual increasing supply in the presence of monopsony power, the marginal cost will be increasing and convex inQ and is given by equation (6).
mc(Q) = ∂c(Q)
∂Q = dw(Q)
dQ Q+w(Q) + ∂m(Q)
∂Q (6)
The supply curve of the wholesaler on the international market is his marginal cost. On Figure 2 one can see that a change in international price leads to a lower increase in quantity than the decrease in quantity associated with a decrease in price of the same magnitude. This larger decrease in quantity will induce a larger change inpf for a decrease in international price if the farmers supply has a constant slope or if it doesn’t, as long as marginal cost is sufficiently convex, as will be presented below. This intuitive mechanism will be formally developed below, with a model linking international price to producer price.
One can then formulate the wholesaler’s profit function as given by equation (7).1
max π = P∗Q−w(Q)Q−m(Q)−f (7) In this modelw(Q) = pf is the inverse supply function for the farm-based input andm(Q) is the cost function associated with preparing the farm product for home consumption or export, andf is the fixed cost mentioned above.
By maximizing profit with respect to Q one gets the First Order Condition (FOC) as presented in equation (8).
P∗−w0(Q)Q−w(Q)−m0(Q) = 0 (8) where the quantity that optimizes profit is denoted Q. What we are interested
1I owe special thanks to Henry Kinnucan for proposing a simpler and shorter route in finding the relationship between farmer price and international price. The development below, from equations (7) to (15), closely follows the development that he proposed.
in here is how optimal quantity will vary for a given change of P∗. Since one can’t directly isolate Q in equation (8), an implicit function is used. Denoting F(Q) as being the left hand side of the FOC leads to the identity presented in equation (9).
F(Q, P) = P∗−w0(Q)Q−w(Q)−m0(Q) ≡ 0 (9) The implicit function is the one given in equation (10), which is the wholesaler’s supply function.
Q = Q(P∗) (10)
The use of the Implicit Function Theorem is now useful to differentiate optimal quantity with respect to price, as in equation (11).
dQ
dP∗ = −
dF dP∗
dF dQ
=− 1
−w00(Q)Q−w0(Q)−w0(Q)−m00(Q) (11)
= 1
w00(Q)Q+ 2w0(Q) +m00(Q) > 0
The relationship between farm and wholesale prices that was given in equation (4) can now be written, at the optimum, as in equation (12).
pf = w(Q) =w[Q(P∗)] (12) Using the chain rule, the change of farmer price for a change in the international price is given in equation (13).
dpf
dP∗ = dw(Q(P∗)) dQ
dQ
dP∗ = w0(Q)
w00(Q)Q+ 2w0(Q) +m00(Q) > 0 (13) Two important remarks can be said concerning the relationship between inter- national price and farmer price. The first is that the relationship is increasing, and the second is that 0 < dPdpf∗ < 1. 2 It basically means that a one dollar increase in
2I thank Henry Kinnucan for pointing out that in much of the literature, the term ”imperfect price transmission” is used in situations such as this, which is not entirely correct. For a detailed explanation of this, see Kinnucan & Zhang (2013) in which it is shown that a farm-retail price transmission of 1 isn’t a prerequisite for competitive market clearing and that one must distinguish
wholesale price causes farm price to rise by less than a dollar. If the farm supply and m(Q), the cost of packaging as well as other extra costs to export, are linear functions, then dPdpf∗ = 12, and the one dollar increase in international price mentioned above would lead to a 50 cents increase in farm price.
Any asymmetry in the price transmission will take place if the price transmission relation is non-linear in wholesale price. This can be evaluated by taking the second derivative with respect to P∗ of equation (12). This is given by equation (14).
d2pf
(dP∗)2 = w00 dQ
dP∗ 2
+w0 d2Q
(dP∗)2 (14)
Asymmetric price transmission requires non-linearity of either the farm supply or the wholesale supply functions. If both functions are linear, then (dPd2p∗f)2 = 0, and price transmission is symmetric.
Whether the sign of the right hand side of equation (14) is negative or positive is indeterminate. Although w0(Q) and w00(Q) are positive as defined higher up, and the squared term
dQ dP∗
2
> 0 is also positive, the last term dPd2Q∗2 is negative (assuming convexity of the marginal cost curve, if it were linear then this term would be zero).
The sign of equation (14) can be made determinate by placing restrictions on the shapes of the farm and wholesale supply functions. This becomes clearer when rewriting it withw00 on the left hand side, as in equation (15).
w00 < −w0(dPd2Q∗)2
dQ dP∗
2 =
w0(dQ)d2P∗2 dQ
dP∗
2 (15) One such restriction is that the inverse wholesale supply function must be suf- ficiently convex compared to the inverse farm supply function. If this is the case, then there will be a negative sign in equation (15).
For the wholesalers inverse supply function to be convex, that is d2P∗
(dQ)2 >0, its ordinary supply function must be concave, meaning that (dPd2Q∗)2 >0. The latter is a necessary condition for equation (14) to have a negative sign, and thus produce the concave function for the price transmission relation shown in Figure 3.
To summarize, the relationship between P∗ and w is increasing and concave as long as the marginal cost curve is sufficiently convex with respect to the inverse supply curve of farmers. If this is the case, the relationship between the two will be such that transmission will be larger if there is a one time downward international price movement from the equilibrium than if there is an upward movement.
between the elasticity of price transmission and the slope of price transmission.
3 Data
In order to estimate the impact of international agricultural price variations on producer prices, yearly data on export and producer prices from the Food and Agriculture Organization of the United Nations (FAO) is used. The producer price is the price received by farmers as collected at the point of sale for primary crops, live animals and livestock primary products. FAO’s export data is produced according to the International Merchandise Trade Statistics Methodology and mainly comes from national authorities and other international organizations. The export values are reported as Free-on-Board (FOB) such that insurance and transport costs are not included. It is an unbalanced panel of 161 items, 117 countries and 35 years, ranging from 1966 to 2000. The list of items and countries are presented in the Appendix.
Higher frequency data are usually used in the Feathers and Rockets literature as well as some studies on agricultural commodities’ price transmission. However, the latter are usually case studies, covering a few items or a specific country. This paper is aimed at a more general approach. Tomek & Myers (1993) say that farmers often make annual decisions. Crops are an example, where the farmer decides at the beginning of the period the area where he is going to plant, then can’t change it. Other similar decisions are taken by the farmers which can’t then be changed, whether international prices change or not.
The data used for the meteorological IV’s comes from Mitchell et al. (2003). For the natural catastrophies what is used isThe OFDA/CRED International Disaster Database (n.d.). The neighbouring countries’ data was taken from the World CIA Factbook (2011).
As the interest here is on price transmission, exporters who are considered large are excluded from the sample. A benchmark of 1% of total exports in each item- country pair is set as a small exporter. This is to stay in line with the idea of a
”small country” being a price taker. This is also a first step towards dealing with endogeneity.
Data is also cleaned in a way such that extreme values are excluded. Even though intuition leads to expect a price of exports above the domestic price, due to processing and handling costs, we however allow an export price as low as 80% of the producer price in order to include different possible scenarios, such as exceptional price variations. A ratio of export price over domestic price that is above 20 is also excluded, as producer prices reported as being less than 5% of the export price suggests that what is captured is probably that the product may be slightly different from what is sold locally.3 Beyond product quality differences and potential mistakes in the data, the exclusion of export prices that are considerably lower than producer prices can also be justified by the fact that it may reflect some ”dumping”
3There is a relatively large amount of observations for which the ratio is below 0.8. These cases don’t seem to be coming from any countries or items in particular, but the frequency does drop somewhat after 1990. This seems to be coming from the fact that FAO producer prices come from two databases, one being considered as historical price data and entitled ”Price archive”. The latter is probably of lower quality than the more recent data. Regressions without any cleaning were run but lowR2 and Hansen test values confirm that some cleaning is necessary.
mechanisms.4
To study the link between oligopsony and asymmetry, data on one of the main companies that trades agricultural products worldwide is included. This data was collected by looking at Cargill’s worldwide website. The company reports each country in which it is present and have either a page or a specific website for each.
The information given includes the year from when it has been active in that country and the products that it covered in that country. Matches were done so that the product names were comparable to the FAO database. The list of items covered by Cargill are presented in the Appendix.
4 Empirical strategy
4.1 Relationship between the international price and the producer’s price
As was shown in section 2, one expects a concave relationship between international prices and producer prices due to the presence of intermediaries with market power.
Therefore, the first empirical step is to test this relationship. However, one must bear in mind that this is a necessary condition to be consistent with the theory, but not a sufficient one. 5 There will be concavity between the two prices if the equality of equation (16) holds for π <1 and $ > 0. In the empirical section international prices are proxied by export prices.
Pi,c,tprod = $(Pi,c,texp)π (16)
A logarithmic transformation of this expression leads to the expression given in equation (16).
ln(Pi,c,tprod) = ln($)
| {z }
≡σ+
+πln(Pi,c,texp) (17)
⇔ln(Pi,c,tprod) = σ+πln(Pi,c,texp) +
To this end, the empirical specification used is the one presented in equation (18), where σ ≡ α+λic+λct +λit. These three last parameters are respectively item-country, country-year and item-year fixed effects.
4According to the World Trade Organisation, if a company exports a product at a price lower than the price it normally charges on its own home market, it is said to be ”dumping” the product.
We can therefore believe that if companies are engaged in dumping, the causality and mechanism of price transmission isn’t as clear.
5I thank Fr´ed´eric Robert-Nicoud for pointing this important fact out, as elements such as decreasing returns on the good that is in the producion function of the intermediary in perfect competition would also lead to such a result.
ln(Pi,c,tprod) = α+πln(Pi,c,texp) +λic+λct+λit+i,c,t (18) Items are noted by i, with i = 1, ..., N, countries by c, with c = 1, ..., C and years byt, with t= 1, ...T.
To get rid of the item-country, country-year and item-year fixed effects the transformation given by equation (19) is used.
ln(Pgi,c,tprod) = ln(Pi,c,tprod)−ln(Pi,c,·prod)−ln(Pi,·,tprod)−ln(P·,c,tprod) (19) + ln(Pi,·,·prod) + ln(P·,i,·prod) + ln(P·,·,tprod)−ln(P·,·,·prod)
with
ln(Pi,c,·prod)≡ 1 T
X
t
Pcit
ln(Pi,·,·prod)≡ 1 CT
X
c
X
t
Pcit
ln(P·,·,·prod)≡ 1 CN T
X
c
X
i
X
t
Pcit
The other means being defined in the same way.
This transformation is applied to all variables individually, which will eliminate the fixed effects and the constant such that what is left is presented in equation (20).
ln(Pgi,c,tprod) = πln(Pgi,c,texp) +gi,c,t (20) For the relationship to be concave, and therefore consistent with the theory presented in section 2, we must have that π < 1. Results for the Ordinary Least Squares (OLS) estimation are given in column 1 of table 1. Standard errors are clustered by item-country in this regression as well as for the Two-Stage Least Squares (2SLS) regression presented below.
4.2 Asymmetry of price transmission
The next step is to depart from the theory and to look at a dynamic specification, in which a dummy taking a value of 1 if the export price for a certain good in a
specific country increased from the previous year and 0 if it decreased is added to the regression. This will enable us to distinguish increasing prices from decreasing prices.
The specification used is given by equation (21).
ln(Pi,c,tprod) = α+γ(price upi,c,t) +βln(Pi,c,texp) (21) +δ(price upi,c,t) ln(Pi,c,texp) +λic+λct+λit+i,c,t
The same transformation as above is applied to all variables such that what is left is presented in equation (22).
ln(Pgi,c,tprod) = γ(price up)g t+βln(Pgi,c,texp) (22)
+δ(price up)gtln(Pi,c,texp) +gi,c,t
The β coefficient is the elasticity of price transmission. The δ coefficient will indicate two things. The first is that if the coefficient is different from zero, the producer price percentage change will be different on the upward and the downward side, if we are considering a given percentage change of the international price. The second is that in a dynamic setting, if once again the coefficient is different from zero, this will lead to hysteresis.6 The following example illustrates this by assuming that δ < 0. If the producer price is at an initial level P0prod, after an international price increase of ∆Pexp, the producer price will now be at the new level P1prod. However, if the international price then decreases of the same amount ∆Pexp, the producer price will then drop to a level P2prod, which is lower than the initial level P0prod.
Results for the Ordinary Least Squares (OLS) estimation are given in column 3 of table 1. Standard errors are once again clustered by item-country in this regression.
4.3 Dealing with endogeneity
As mentioned above, the export price may be endogenous. This implies that the price up dummy and the interaction may also be endogenous. For this reason the method of estimation that will be used is that of 2SLS. For this one must find a variable that will affect export price without however being directly correlated to producer price. This can be done by using climatological phenomena in other coun- tries exporting the same good. Three meteorological Instrumental Variables (IV’s)
6I thank Fr´ed´eric Robert-Nicoud for pointing out this hysteresis.
will be used, which includes the amount of rainfall, temperature and cloud cover.
A variable of catastrophic events is also used as an instrument. To avoid any cor- relation between the weather conditions in the country considered and the weather in other countries, all neighbours are excluded when creating the instrument.
The three meteorological IV’s will be created in the same manner. As it is hard to know what the effect of a certain change of each variable on prices will be and since the study contains many different types of goods, it will be considered that being far away from the mean is bad for the farmer. This is close to Shaw (1964) where it is said that a reasonable way of seeing the relationship between meteorological factors and yield is by representing it as a bell-shaped curve. In our case the effect of the three IV’s will be seen in this way such that what will be used will be the standard deviation of the monthly temperature, monthly rainfall and cloud cover each year.
What is considered is the effect of these variables on all other countries that are exporting the good, multiplied by their share of exports with respect to the rest of the world in that good, and excluding neighboring countries. This exclusion is to avoid any direct correlation between the neighbouring countries’ weather conditions and the producer price considered. This will give us exogenous shocks to all other countries’ prices, therefore influencing international price but not the own country’s price directly.
An illustration of the construction of these variables is given in Figure 4 for bananas in Zimbabwe in the year 2000. This illustrates that all neighbours and non exporters of the good aren’t included in the instrument. All others are counted, the weight in the instrument depending on their export share. This is explicited in equation (23).
rain IVb = X
c
rainf allict∗export shareict, c6= n and c 6= b (23)
with: b: own country c: country n: neighbours
Finally, a count variable of the number of climatological (including extreme temperature, drought, wildfire) or meteorological (storm) events in other countries is added as another instrument. It is constructed in the same way as before but this time the number of disasters is used directly rather than the standard deviation.
As the price up dummy may also be endogenous, it is instrumented by a dummy taking the value 1 if the IV variable in question went up and taking the value 0 otherwise. This dummy is also interacted with the IV variable itself in order to instrument the interaction of price up with the export price.
5 Results
Results of the regression that tests the concavity of the relationship between inter- national prices and producer prices are presented in columns 1 and 2 of table 1. The coefficients on the export price are below 1 in both cases, and significantly different from 1, which is consistent with the necessary condition for concavity. The p-value of the test of underidentification as well as the one for overidentifying restrictions suggest that the instruments are valid.7 The results of the first stage regressions are given in column 1 of table 2.
Results of the regression using the OLS and IV’s to test for asymmetry are given in columns 3 and 4 of table 1 respectively. The results of the first stage regression are given in columns 2 to 4 of table 2. The interaction term clearly points to some asymmetry and subsequent hysteresis. The sign of the asymmetry term is the same as in the OLS regression given in column 3 and the p-value of the test of underidentification as well as the one for overidentifying restrictions suggest that the instruments are valid. Empirical results suggest that when FOB export prices rise by 1% farm prices rise by 0.60%,8 and when FOB export prices fall by 1% farm prices fall by 0.98%. In other words, there is near perfect transmission of declines in wholesale price, but imperfect transmission of rises in wholesale price.
5.1 Robustness checks
The results above show that there is an asymmetry in the long run price transmis- sion of international prices to local prices. However, it is important to check the immediate response as well. The inclusion of the lagged value of producer price as an explanatory variable will enable one to do the aforementioned interpretation.
The specification used is given by equation (24). λi andλc are respectively item and country fixed effects.
ln(Pi,c,tprod) = α+ρln(Pi,c,t−1prod ) +γ (price up)t+βln(Pi,c,texp) (24) +δ (price up)tln(Pi,c,texp) +λi+λc+i,c,t
First-differencing will eliminate the fixed effects as well as the constant. How- ever, one then has ln(Pi,c,t−1prod ) on the left hand side which will be correlated with the error term i,c,t−1. To deal with this problem the estimation method used is that of Arellano & Bond (1991) and Arellano and Bover / Blundell and Bond (Arellano
& Bover (1995); Blundell & Bond (1998)) who developed a Generalized Method- of-Moments (GMM) estimator that instruments the differenced variables that are not strictly exogenous with all their available lags in levels or in first differences.
It is implemented using the approach of Roodman (2006) based on the Arellano and Bond (Arellano & Bond (1991)) and Arellano and Bover / Blundell and Bond
7Using a benckmark for the p-value of 0.05.
8From table 1 we can work out the upward transmission as being 97.83%−37.47% = 60.36%.
(Arellano & Bover (1995); Blundell & Bond (1998)) dynamic panel estimators. As mentioned above, the explanatory variables used here cannot be considered strictly exogenous and the change in the export price cannot be considered uncorrelated with item and country unobservable fixed effects, therefore the original equation in levels is not added to the system. The export price, the price change as well as the interaction term are considered as being predetermined such that the first lag that is used as an instrument for these variables int is t−2. The first lag for producer prices that is used int−1 (the lag of the explained variable) ist−3. When running the regression on the whole data the N dimension is composed of countries and items and is therefore much larger than T. Even though this is good in the case of these types of estimators requiring small T and largeN, only half of the periods in the sample are used as instruments to limit the number of instruments used. Time dummies are also included in the regression.
As the sufficiency of the variables’ past values being used as instruments may be questionable we also use a specification with the above mentioned economic instruments, used in the 2SLS specification. The results are shown in table 3.
These results support the ones found in the OLS and 2SLS specifications, that is a negative and significant asymmetry term. Due to the lag of producer price on the right hand side of equation (24) the interpretation of the coefficients is slightly different. The short run price transmission is given byβ in equation (24) whereas the long run transmission, comparable to the β in equation (22), is computed as follows: 1−ρβ . The long run asymmetry is obtained by computingρ−β−δ1−ρ. This yields the results presented in table 4. Standard errors are computed using a calculation technique based on the ”delta method”.9
One sees that the initial transmission, although relatively small, is asymmetric.
This then leads to a long run transmission that is, albeit slightly smaller than what is found with the 2SLS approach, not different in terms of statistical significance.
5.2 Oligopsony and asymmetry
The next step is to see whether measures of market power are linked to asymmetry.
For this two variables are used. The first is the importance of exports in the mar- ket. By looking at the quantity exported with respect to the quantity produced, we will be able to see whether asymmetry is influenced by a larger share of ex- ports.10 One might expect the fact of exporting more of local production to lead to more price transmission. However, this would not explain more asymmetry. What is of interest here is whether the mechanism specific to exporting plays a role in explaining asymmetry. A larger local presence of intermediaries will be a prereq- uisite for exporting a larger share of local production. A difference in asymmetry for high versus low shares of exports will point to the impact of intermediaries in the asymmetric pattern of price transmission. We will therefore use two groups of countries, those exporting a large share of their production and those exporting a small share. Results are given in table 5. One sees that in markets where exports
9This is implemented in the statistical package STATA using the nlcom command.
10A benchmark value of 30 percent is used to distinguish a large share from a small share.
consist of a large share with respect to local production there is a significant amount of asymmetry whereas in the other markets there is no significant asymmetry.
This is complemented by another measure of market power. This measure is a regression run specifically on items where the presence of one of the main wholesalers is present, namely Cargill. As mentioned in section 3, this data has been collected from Cargill’s country websites. This is an item specific information that will enable us to separate the regressions into two groups, one where Cargill is present and the other where it isn’t. A larger asymmetry for the group where Cargill is present will point to an influence of the wholesaler on asymmetry. The results of these regressions are given in table 6. We see that market power plays a role in explaining asymmetry. The presence of Cargill significantly increases the asymmetry of price transmission whereas when it is absent from a market, the asymmetry term isn’t significantly different from zero.
The presence of one of the main intermediaries as well as the share of local production exported abroad are both explaining asymmetry, supporting the theory presented above.
6 Conclusion
A model of price transmission from international agricultural prices to producer prices is presented in order to understand the mechanisms behind price transmission.
Due to the geographical dispersion of farmers and economies of scale in wholesaling, agricultural markets will we characterized by market power on the demand side. The model predicts that this power of intermediaries buying the products from farmers leads to an asymmetric price transmission when intermediaries have sufficiently convex marginal cost curves. The asymmetry is such that there is more price transmission when prices fall.
The results are shown using a Two Stage Least Squares estimator to control for endogeneity problems. This approach has the advantage, as noted by Acharya et al.
(2011), of avoiding some of the disadvantages of many studies in the recent litera- ture that focus on the time-series properties of the data such that it is sometimes unclear whether rejection of symmetry isn’t simply due to specification error. The instruments used are variations in rainfall, temperature, climate disasters and cloud cover in other geographic regions. The exclusion of neighbouring countries within the region ensures that these instruments aren’t correlated to the local weather conditions, thus local prices. The results are clear in pointing out the presence of asymmetry, with the transmission being stronger for price decreases. Robustness checks using two different specifications of a Generalized Method of Moments esti- mator confirm the presence of an asymmetric price transmission. These results also imply hysteresis, where an international price increase followed by a decrease of the same magnitude will not bring the producer price back to its initial equilibrium.
The link between market power and asymmetry is then tested. A variable indicating the presence of intermediaries on specific markets is used to test this link.
More specifically it is the presence of Cargill, one of the largest intermediaries in commodity markets, that is used as a variable and the results show that asymmetry
is stronger when Cargill is present. This result is supported by another regression where the larger the share of local production exported abroad, the higher the degree of asymmetric price transmission.
This points towards some policy issues, notably the fact that governments should be aware of the effect of market power of intermediaries and the role they play in influencing the price received by farmers, and therefore the gains from trade liberal- ization in agricultural markets. Abuse of monopsony power by large intermediaries in agricultural markets can be particularly harmful in poor countries where farmers often live close to the poverty line.
Future research could be done in the collection of market power data at the item and country level on a yearly basis. Other private companies than the one looked at here could also be integrated into such studies. Measures of search costs should also be considered.
Appendix
Appendix 1: Sample of items.
Item Cargill Item Cargill
1 Almonds, with shell No 27 Cherries No
2 Anise, badian, fennel, corian. No 28 Chestnuts Yes
3 Apples Yes 29 Chick peas No
4 Apricots No 30 Chicken meat Yes
5 Arecanuts No 31 Chicory roots No
6 Artichokes No 32 Chillies and peppers, dry No
7 Asparagus No 33 Chillies and peppers, green No
8 Avocados No 34 Citrus fruit, nes No
9 Bananas No 35 Cloves No
10 Barley Yes 36 Cocoa beans Yes
11 Beans, dry Yes 37 Coconuts No
12 Beans, green Yes 38 Coffee, green Yes
13 Beeswax No 39 Cotton lint Yes
14 Berries Nes No 40 Cottonseed Yes
15 Blueberries No 41 Cow milk, whole, fresh No
16 Broad beans, horse beans, dry No 42 Cow peas, dry No
17 Buckwheat No 43 Cranberries No
18 Cabbages and other brassicas No 44 Cucumbers and gherkins No
19 Canary seed No 45 Currants No
20 Carobs No 46 Dates No
21 Carrots and turnips No 47 Duck meat Yes
22 Cashew nuts, with shell No 48 Eggplants (aubergines) No
23 Castor oil seed No 49 Fibre Crops Nes No
24 Cattle meat Yes 50 Figs No
25 Cauliflowers and broccoli No 51 Flax fibre and tow No
26 Cereals, nes No 52 Fruit Fresh Nes No
Appendix 1: Sample of items (continued).
Item Cargill Item Cargill
53 Fruit, tropical fresh nes No 90 Natural rubber No
54 Game meat No 91 Nutmeg, mace and cardamoms No
55 Garlic No 92 Nuts, nes Yes
56 Ginger No 93 Oats Yes
57 Goat meat Yes 94 Oilseeds, Nes Yes
58 Goose and guinea fowl meat Yes 95 Okra No
59 Gooseberries No 96 Olives No
60 Grapefruit (inc. pomelos) Yes 97 Onions (inc. shallots), green No
61 Grapes No 98 Onions, dry No
62 Groundnuts, with shell Yes 99 Oranges No
63 Hazelnuts, with shell No 100 Other Bastfibres No
64 Hemp Tow Waste No 101 Other bird eggs,in shell No
65 Hempseed No 102 Other melons (inc.cantaloupes) No
66 Hen eggs, in shell Yes 103 Palm kernels Yes
67 Hops Yes 104 Palm oil Yes
68 Horse meat Yes 105 Papayas No
69 Jute No 106 Peaches and nectarines No
70 Karite Nuts (Sheanuts) No 107 Pears No
71 Kiwi fruit No 108 Peas, dry Yes
72 Leeks, other alliaceous veg No 109 Peas, green Yes
73 Leguminous vegetables, nes No 110 Pepper (Piper spp.) No
74 Lemons and limes No 111 Persimmons No
75 Lentils No 112 Pig meat Yes
76 Lettuce and chicory Yes 113 Pigeon peas No
77 Linseed No 114 Pineapples No
78 Lupins No 115 Pistachios No
79 Maize Yes 116 Plantains No
80 Maize, green Yes 117 Plums and sloes No
81 Mangoes, mangosteens, guavas No 118 Poppy seed No
82 Manila Fibre (Abaca) No 119 Potatoes No
83 Mat´e No 120 Pulses, nes No
84 Meat nes Yes 121 Pumpkins, squash and gourds No
85 Millet Yes 122 Pyrethrum,Dried No
86 Mixed grain No 123 Quinces No
87 Mushrooms and truffles No 124 Rabbit meat Yes
88 Mustard seed No 125 Ramie No
89 Natural honey No 126 Rapeseed Yes
Appendix 1: Sample of items (continued).
Item Cargill
127 Raspberries No
128 Rice, paddy Yes
129 Roots and Tubers, nes No
130 Rye Yes
131 Safflower seed No
132 Sesame seed No
133 Sheep meat Yes
134 Silk-worm cocoons, reelable No
135 Sisal No
136 Sorghum Yes
137 Sour cherries No
138 Soybeans Yes
139 Spices, nes Yes
140 Spinach No
141 Stone fruit, nes No
142 Strawberries No
143 Sugar beet Yes
144 Sugar cane Yes
145 Sunflower seed Yes
146 Sweet potatoes No
147 Tangerines, mandarins, clem. No
148 Taro (cocoyam) No
149 Tea Yes
150 Tobacco, unmanufactured No
151 Tomatoes No
152 Triticale No
153 Turkey meat Yes
154 Vanilla No
155 Vegetables fresh nes No
156 Walnuts, with shell No
157 Watermelons No
158 Wheat Yes
159 Wool, greasy No
160 Yams No
161 Yautia (cocoyam) No
