SOL-306-13-000034_-_RADP_NW_-_2FBO_-_Attachment_J.17_-_Wheat_Markets_and_Wheat_Availabiltiy.pdf

Wh eat M

Food S

Market

Security Re ts and Afgh esponse An d Whe hanist nalysis Sup

(RASTA)

Fin eat Av tan pport Team al Draft:

vailab m Afghanist

Septemb bility tan ber 17 20 in

012.

Table of Contents

EXECUTIVE SUMMARY

1. INTRODUCTION

2. OTHER WORK DONE IN THIS AREA

3. THE CONTRIBUTION OF THE WHEAT SECTOR TO FOOD AVAILABILITY AT THE NATIONAL LEVEL IN

AFGHANISTAN

3.1 WHY LOOK AT WHEAT?

3.2 TRENDS IN PRODUCTION OF WHEAT

Incentives for wheat production have increased over the recent past

3.3 REGIONAL TRADE, FORMAL AND INFORMAL IMPORTS OF WHEAT AND WHEAT FLOUR INTO AFGHANISTAN

The magnitude of informal and formal sector imports Sources of informal imports Regional Wheat trade flows and domestic wheat availability

4. DOMESTIC MARKETS FOR WHEAT AND WHEAT FLOUR IN AFGHANISTAN

4.1 MARKET INTEGRATION

How well integrated are markets in Afghanistan compared to neighboring countries?

4.2 PRICES AND SEASONALITY

Wheat Surplus and Deficit Areas and Major Commodity Flows The Structure of Wheat Flour Imports

MARKET STRUCTURE AND THE FEASIBILITY OF CASH BASED TRANSFER MECHANISMS TO ADDRESS FOOD

INSECURITY – GENERAL INDICATIONS

5.1 INTRODUCTION

5.2 MARKETS IN THE NORTH AND NORTHEAST

5.3 MARKETS IN THE CENTRAL EAST AND SOUTHEAST

5.4 MARKETS IN THE CENTRAL HIGHLANDS

5.5 MARKETS IN THE SOUTH

5.6 MARKETS IN THE WEST

5.7 SUMMARY OF PARTIAL MIFIRA ANALYSIS

6. CONCLUSIONS

REFERENCES

APPENDIX I

Executive Summary

This study reviews the structure and dynamics of wheat markets in Afghanistan. The study compiles the findings of two surveys of urban and rural markets in Afghanistan undertaken in March and April 2012 in collaboration with UN and NGO partners, and also reviews available secondary data on wheat and wheat flour production and trade in the country. The objective of the work was to establish a contemporary understanding of the functioning of these markets with a view to better understanding the role that they play in addressing food security concerns in the country. The major findings of the study have been as follows:

From available data, the level of market integration between urban markets is relatively good. It is at least comparable to the level of integration for wheat markets in neighboring Pakistan. Despite large shifts in domestic production, private markets appear to be responsive in supplying adequate supplies of grain.

There are strong sub‐national dimensions to structure of grain trade in Afghanistan. Markets in the northern areas of the country have a very different trade orientation than do markets in other parts of the country. In northern markets, trade with the Central Asian republics, and particularly with cereals imported from Kazakhstan, are much more predominant. Markets in the East and South have a greater orientation to supplies from Pakistan.

• Given the large degree of variability in existing data sets, it is hard to draw any significant conclusions about the magnitude of informal wheat imports, except to say that informal imports undoubtedly exist and do vary according to domestic production levels, regional and domestic pricing and trade policy decisions. This highlights the importance of improvements in data collection systems and also further applied research and monitoring work in this area.

The study finds that wheat prices are influenced by domestic production, and relative prices in major import supplying countries (i.e. there is little or no evidence that food aid programmes have created significant disincentives to domestic production). The study also finds that incentives for wheat production have increased over the recent past, and finds that under most assumptions, wheat production for Afghanistan indeed holds a comparative advantage vis a vis imported wheat.

An analysis of price data does not show strong seasonal trends throughout the year. However results of a field survey undertaken as part of this study indicate that the issue of seasonality is a significant issue affecting availability of cereals in some rural markets. Specific markets are reported to be generally cut off from access to markets during the winter months when excessive snowfall can impede the normal flow of agricultural goods.

From the perspective of market functioning, the study reaches some very general indications about the appropriateness of cash / vouchers and food aid as transfer modalities to address food insecurity in different parts of the country. These indications would need to be verified and be made more specific by further empirical work. Notwishtanding this, the study finds some evidence that a cash transfer or voucher program would be an appropriate option to address problems of food insecurity at all times of the year in most Eastern South Eastern and Southern parts of the country . Cash /vouchers would also be approporate for most of the year (i.e outside of winter months) in the North and North East as well as West and the Central Highliands. In these parts of the country, food aid would be approapriate during the winter months as the weather and topography constrains the ability of markets to supply sufficient quantities of food to the population.

1. Introduction

This paper evaluates the effectiveness of agricultural markets for wheat flour and wheat grain in addressing food security in Afghanistan. The emphasis on wheat is due to the importance of this cereal in the Afghan diet1.

The paper has three main objectives.

1. To review the role of the wheat sector and its role in contributing to food security at the national level. Within the definition of ”wheat sector” is included all imports of wheat and wheat flour into Afghanistan (Afghanistan does not export wheat or wheat flour);

2. To examine the effectiveness and efficiency of wheat and wheat flour markets in

Afghanistan as distribution mechanisms;

And;

3. To give a general indication of the appropriateness of different types of resource transfer (food aid and/or cash/vouchers) as ways of addressing food insecurity at the sub-national level given existing market dynamics for wheat.

This study reviews secondary data related to agricultural production and trade together with primary data from a recent survey of urban and rural market traders. The study benefited significantly from the participation and collaboration of a large number of partners who assisted in study design and execution of the two surveys. Specifically these were UNWFP, CRS, ActionAid, ACTED, FGA, FEWS NET, Oxfam, Afghan Aid, and the Afghan Ministry of Agriculture Irrigation and Livestock. Without their participation, the current study would not have been possible.

2. Other Work Done in this Area

Over the past decade, there has been a considerable amount of research and analytical work undertaken to look at different agricultural markets in Afghanistan and the impact they have on farm household welfare, rural livelihoods, and food security in both urban and rural contexts. Some of the more notable pieces are briefly reviewed in this section.

A rather extensive review of multiple aspects of agricultural production in Afghanistan was undertaken by FAO in 2003 with the study ‘Agriculture and Food Production in Afghanistan:

The Results of the 2002-2003 Winter Farm Survey’ by Maletta and Favre. This work involved a quite extensive data collection exercise which evaluated many different facets of agricultural production, input use, land ownership structures and other variables. This study did not focus on agricultural markets directly. There were a number of major findings from this study, one of which was the relatively low level of marketable surplus of wheat among farm households surveyed, estimated at 11 percent of total supply, and the finding that 57 percent of all farm households were net consumers of wheat.

1 Wheat accounts for approximately 60% of kilocalories consumed per person per day in Afghanistan (source:

NRVA 2003).

In a similar area, Favre 2003 also evaluated the behavior of wheat markets in Afghanistan and their relationship to international markets over the period 1996-2003. One of the findings of this study was wheat and wheat flour markets in Afghanistan track movements in international markets surprisingly well, considering all that was occurring in Afghanistan over the period of the 1990’s. The current study will also address these issues in an updated context.

A Crop and Food Supply Assessment mission was conducted by FAO and WFP in 2004. This CFSAM found a sharp decline in domestic wheat production and estimated 35 percent of the population would face increased food insecurity in response to reduced domestic production.

In addition, data collected by the CFSAM and other FAO monitoring efforts have been used as background material and data in the current study.

The World Bank conducted a relatively extensive review of food security and food aid options for Afghanistan that culminated in the South Asia discussion paper ‘Enhancing Food Security in Afghanistan: Private Markets and Public Policy Options. One of the main findings of this paper was that private sector international trade had in fact helped to stabilize supplies of wheat and wheat prices and as such, policies and institutions that further support the functioning of these international markets would likely lead to improved food security outcomes for the country.

In a follow on study Dorosh and Chabot 2007 evaluated wheat prices and market flows in Afghanistan based on results of surveys of wheat traders and millers, and econometric analysis of price movements in major markets. The paper found that in spite of food aid imports, domestic prices were not lowered below import parity levels in most major Afghan markets suggesting that in most instances food aid programs did not have major price disincentive effects on domestic production, at least until 2003. However, following the 2003 bumper harvest, the analysis suggests that continued food aid inflows may have depressed producer prices by as much as about 15%.

The Famine Early Warning Systems Network (FEWS NET) has also conducted a considerable amount of analytical work in Afghanistan to better understand the relationship of market dynamics and food security variables in Afghanistan. In 2007, FEWS NET published ‘Northern Wheat Trader Survey and Afghan Food Security’. This study looked at wheat markets and trade dynamics in the western and northern regions of the country, which was an element of national grain markets which had not been looked at very closely in previous studies of this type. One of the findings of this study was the problem of identifying accurate and timely agricultural production and price data, comparable between countries, for establishing a complete picture of food security conditions at the regional level. Since this period (2007) this situation has been improved upon somewhat for purposes of early warning systems for food security.

Similarly, in 2007 FEWS NET also conducted the study ‘The Pakistan Wheat Subsector and Afghan Food Security’. This study evaluated the structure of Pakistan wheat production, trade and trade policy structures as Pakistan wheat trade to Afghanistan is such an important element of food security for the country.

More recently, in 2011 FEWS NET conducted the study ‘A Review of Wheat Markets and Food Security in Central Asia: With a Special Focus on Afghanistan and Tajikistan’ by Chabot and Tondel. This study was somewhat similar to the two FEWS studies conducted in 2007 on the Pakistani wheat sub-sector. However the 2011 work expanded on that done in 2007 to also include wheat market dynamics Central Asia and incorporated how the wheat trade in Central Asia impacted upon food security variables in Afghanistan and Tajikistan. This study found that, with a few exceptions the trade of wheat functioned relatively effectively.

It also noted that, in instances where policy interventions disrupted this trade, as was the case in 2008 when both Pakistan and Kazakhstan placed trade restrictions of various types on wheat, this had a very significant impact on markets in Afghanistan and on the effective functioning of grain markets at the regional level.

One of the more lengthy activities to evaluate rural livelihoods and rural household welfare variables is the Afghanistan National Risk and Vulnerability Assessment (NRVA). This activity is not focused on agriculture and food security matters directly. It is a multi-faceted survey of between 12,000 and 30,000 households (depending on the year undertaken). In the area of nutrition and food security the NRVA does compile and evaluate household food consumption to evaluate both poverty and potential food deficits in households. The survey has been implemented in 2003, 2005, 2007/8, and most recently in 2011/12. Results of the initial 2003 NRVA found that roughly 20 percent of the rural population experienced chronic food insecurity, the second poorest 20 percent of rural population was prone to periodic food insecurity. The follow on NRVA survey of 2005 found that 30 percent of the population of Afghanistan was unable to meet is daily calorie requirement of 2,100 calories on a regular basis, while 44% of households perceived themselves to be food insecure at different times in that they sometimes, often, or mostly have problems meeting their food needs.

More recently, a joint government of Afghanistan-World Bank publication, ‘Poverty Status in Afghanistan’, released in 2012 used the results of the 2007-2008 NRVA to evaluate food security variables in different regions of the country. Similar to the 2005 NRVA, the study found that 29 percent of the population on average was not able to meet its daily calorie requirement of 2,100 calories. Food insecurity was more prevalent in rural areas where calorie deficiency affected 30 percent of the population as compared to 24 percent in urban areas. The study also found that food security outcomes are closely associated with terrain characteristics. Populations living in areas classified as mountainous areas had a calorie deficiency rate of 33 percent as compared to 24 percent for those living in areas classified as ‘plains’, or non-mountainous. Terrain is an issue associated with market access and so has bearing on the current study.

Key issues identified

Three main themes are evident from the literature review: First, wheat markets in Afghanistan appear to function relatively well as distribution mechanisms; Second, large proportions of the population are net food buyers thus benefit from lower food wheat prices and are disadvantaged by higher prices; Third, policy decisions in neighboring countries regarding trade flows can have a major positive or negative effect on wheat availability and market functioning.

3. The contribution of the wheat sector to food availability at the national level in Afghanistan

3.1 Why look at wheat?

This study will focus on wheat due to its importance in the diet in Afghanistan. Figure 1 contains data from the National Risk and Vulnerability Assessment that estimated the relative contribution of different foods to the average diet in Afghanistan. According to this estimate, wheat contributes just slightly less than 60% of total calories to the diet on average, while rice and maize collectively contribute another 10 percent.

Figure 1. Food Consumption by Product Category (% cal. per day)

Source: 2003 National Risk & Vulnerability Assessment (NRVA)

Table 1. Per Capita Wheat Consumption selected countries and regions Per Capita Wheat

Consumption kg/yr

Afghanistan 170 Pakistan 106 Iran 160 Uzbekistan 164 Asia Region 66

Source: FAOSTAT; for Afghanistan NRVA 2005

Per capita wheat consumption in Afghanistan is estimated at 170kg/yr. This rate of annual consumption is roughly comparable to neighboring Iran and Uzbekistan but rather higher than the consumption rate in Pakistan which is only 106kg/yr. It is considerably higher than the Asia region average of 66kg/yr. It is due to this very high level of consumption and high prevalence in the diet that the current study is focused on wheat.

Wheat 59%

Rice 7%

Maize 3%

Sugar 4%

Pulses 2%

Vegetabl e Oils 10%

Vegetables 1%

Fruits 2%

Meat 1%

Milk 6%

Other 5%

3.2 Trends in Production of Wheat

Table 2 shows the level of production of wheat rice maize and barley over the period 2003- 2011.

Table 2. Cereal Production in Afghanistan 2002-2010 (1,000 MT) 2003 2004 2005 2006 2007 2008 2009 2010 2011 Avg.

Irrigated Wheat 3,017 1,867 2,728 2,902 2,878 2,406 3,433 3,082 3,067 2,820 Rainfed Wheat 1,345 426 1,537 809 1,606 217 1,682 1,450 321 1,044 All Wheat 4,362 2,293 4,265 3,711 4,484 2,623 5,115 4,532 3,388 3,864 Milled Rice 291 310 325 361 425 410 432 481 450 387 Maize 310 234 315 359 360 280 300 285 301 305 Barley 410 220 337 364 370 333 486 428 305 361 Total Cereals 5,373 3,057 5,242 4,495 5,639 3,646 6,333 5,726 4,444 4,917 Source: Afghanistan Ministry of Agriculture Irrigation and Livestock

Figure 2. Irrigated and Rainfed Wheat Production in Afghanistan 2002-2010

Source: Afghanistan Ministry of Agriculture, Irrigation, and Livestock

The table shows how wheat is by considerably margin, the dominant crop grown relative to rice, maize and barley. It is grown in every province of the country, and is a major livelihood base for many rural households. There has been considerable variability in wheat production in Afghanistan over the years. From Figure 2 above, one of the major drivers of this variability is the very large shifts in the level of production of rainfed wheat. By contrast, irrigated wheat production has been considerably more stable.

Incentives for wheat production have increased over the recent past

Incentives for wheat production in Afghanistan have increased over the past 5-7 years. A detailed discussion of this issue is contained in Appendix I. Table 3 presents the summary results of an evaluation that looks at costs and returns for wheat production which were done for irrigated wheat in Kunduz province in northeast Afghanistan. The initial estimate was done in 2005 to evaluate the profitability of wheat production during that period. This

1,000

2,000

3,000

4,000

5,000

6,000

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

W he at P ro du ct io n 1, M T

Irrigated Wheat Rainfed Wheat estimate was updated during the current study to evaluate how costs and returns may have changed over the period.

Table 3. Estimated Costs and Returns for Wheat Production in

Northeast Afghanistan 2005 and 2012 2005 2012 Afs/Jerib US$/Ha Afs/Jerib US$/Ha Estimated Total Revenues 4,900 490 10,800 1,080 Estimated Input Costs 3,635 364 7,000 700 Estimated Net Income 1,265 127 3,800 380

Estimated Unit Cost 7.3Afs/Kg 146US$/Ton 13.75Afs/kg 275US$/Ton

Source: Author’s estimates calculated from secondary data and discussions with key informants

What the evaluations show is that incentives for wheat production have increased considerably since 2005 from a net return of 127US$/Ha in 2005 to US$380US$/Ha in 2012.

Revenues have increased due to the fact that increases in wheat prices (from 8 Afs/kg in 2005 to 18Afs/kg in 2012 have outweighed increases in input costs (labour and fertilizer).

Whilst unit costs have increased (from US$146 / ton to US$275 / ton), costs of production for Afghan wheat was found to be quite cost competitive with wheat from imported sources, suggesting that Afghanistan indeed would hold a comparative advantage in wheat production vis a vis wheat imported from abroad. This issue is also addressed in greater detail in Appendix I.

3.3 Regional Trade, Formal and Informal Imports of Wheat and Wheat Flour into Afghanistan

The magnitude of informal and formal sector imports

As a perennial food deficit country with large swings in domestic production, imports play a critical role in ensuring adequate supplies of cereals for domestic consumption. In most years, formal imports as a portion of total supply range between 15 and 25 percent depending on domestic production conditions. In years of particularly poor production, this level of import dependency can be higher still.

It is widely acknowledged that estimating imports of wheat into Afghanistan is made problematic by the absence of agreed official figures. To show the variability in the import data more explicitly, Figure 3 charts import data from three different sources, the Ministry of Agriculture Irrigation and Livestock (MAIL), the Central Statistics Office (CSO), and the Afghanistan Customs Office.

Figure 3. Annual formal sector wheat Imports into Afghanistan 2006 – 2011: Figures from Ministry of Agriculture, Central Statistics Office (CSO) and the Afghanistan Customs House

The large degree of variation in official estimates of formal imports has a knock-on effect in terms of estimates of informal imports.

Tables 4, 5 and 6 provide estimates of the level of informal imports using import data from these three data sources (MAIL, CSO, and the Afghanistan Customs Office). Production and estimated national consumption requirement data in all tables are from the Ministry of Agriculture, Irrigation, and Livestock. Informal import estimates are derived as the residual between the estimated national consumption requirement, and production and formal imports.

Table 4. Production, Imports and Estimates of Informal Wheat Imports into Afghanistan Import Data from the Ministry of Agriculture Irrigation and Livestock (MAIL)

2006 2007 2008 2009 2010 2011 Mean

Production 3,711 4,384 2,623 5,115 4,532 3,388 3,871

Imports 980 325 1,402 143 525 1,430 801

Food Aid (Cereals) 84 174 212 218 82 n.d. 154

Total Supply 4,775 4,983 4,237 5,476 5,139 4,818 4,903

MAIL Est. Cons. Requirement 4,820 4,917 4,722 5,260 5,231 5,163 5,019

Estimated Deficit (Informal Import Estimate)

45 -66 485 -216 92 345 116

Est. Deficit as % of Total Supply 1% -1% 11% -4% 2% 7% 3% Est. Deficit as % of Official Imports

5% -20% 35% 0 18% 24% 12%

2006 2007 2008 2009 2010 2011 Average

1, M

MAIL CSO Customs Office

Table 5. Production, Imports and Estimates of Informal Wheat Imports into Afghanistan Import Data from the Central Statistics Office (CSO)

2006 2007 2008 2009 2010 2011 Mean

Production 3,711 4,384 2,623 5,115 4,532 3,388 3,871

Imports 555 316 324 583 1135 803 619

Food Aid (Cereals) 84 174 212 218 82 ND 154

Total Supply 4,350 4,874 3,159 5,916 5,749 4,191 4,644

MAIL Est. Cons. Requirement 4,820 4,917 4,722 5,260 5,231 5,163 5,019

Estimated Deficit (Informal Import Estimate)

470 43 1,563 -656 -518 972 312

Est. Deficit as % of Total Supply 11% 1% 49% -11% -9% 23% 7% Est. Deficit as % of Official

85% 14% 482% 0 0 121% 50%

Table 6. Production, Imports and Estimates of Informal Wheat Imports into Afghanistan Import Data from the Afghanistan Customs Office

2009 2010 2011 Mean

Production 5,115 4,532 3,388 4,345

Imports 1338 1032 1158 1,176

Food Aid (Cereals) 218 82 ND 150

Total Supply 6,671 5,646 4,546 5,621

MAIL Est. Cons. Requirement 5,260 5,231 5,163 5,218

Estimated Deficit (Informal Import Estimate)

-1,411 -415 617 -403

Est. Deficit as % of Total Supply -21% -7% 14% -5% Est. Deficit as % of Official

65%

There is considerable variability in the data, both between years, and between data sets.

Meanwhile, all data sets indicate that there is a net surplus of wheat in the country in certain years- 2009 for MAIL, 2009 and 2010 for CSO and the Customs Office. There could be a number of reasons for this. The first is that the estimated national consumption requirement of 4.9-5.2 MMT is too low. In support of this, there is indeed some uncertainty regarding population data in Afghanistan. The 2011 population estimate from the Central Statistics Office is 24,987,700 and this is what has been used to calculate the consumption requirement. However, other sources suggest that the population could be much closer to 30 million2. Were the CSO population estimate to be increased, this would be reason for raising the consumption estimate. Similarly, there is a fair degree of uncertainty concerning the issue of post-harvest losses. In the event that these are higher than existing estimates, this would also reduce the level of these estimated surpluses.

Notwithstanding this, the maximum estimate of informal imports derived from MAIL data would be 485,000 MT in 2008 equivalent to about 10% of national consumption

2 Worldpopulationreview.com estimates a 2011 population figure of 29,835,392 and the CIA’s World Factbook estimates a population of 30,419,928 at July 2012 (see https://www.cia.gov/library/publications/the‐world‐ factbook/geos/af.html).

requirements, slightly more than the next highest year, 2011 (477,000 MT). From CSO data, the maximum estimates are much higher: 1,563,000 MT in 2008 (32% of national consumption requirements) and 1,104,000 MT in 2011 (23%). One other source of data comes from a year long survey conducted by WFP in 2008 – 2009. This estimated that informal wheat and wheat flour imports were around 1,500,000 MT. In other words, the WFP survey was in line with the CSO data for that year (2008)3.

Given the large degree of variability in existing data sets, it is hard to draw any meaningful conclusions about the magnitude of informal wheat imports, except to say that informal imports undoubtedly exist and do vary according to domestic production levels, regional and domestic pricing and trade policy decisions. This highlights the importance of improvements in data collection systems and also further applied research and monitoring work in this area.

Sources of informal imports

Unlike estimates of magnitude, there is a relatively higher degree of certainty about the sources of informally imported wheat and wheat flour. Pakistan is the major source of informal wheat imports into Afghanistan. Informal routes through northern channels via Uzbekistan, Tajikistan are much less feasible as these borders are more tightly controlled.

Indeed Uzbekistan, which would have by far the largest share of imports via northern sources (as this would be wheat and wheat flour transshipped largely from Kazakhstan) places very strict controls on transshipped goods, rendering non-formal imports of a modest valued commodity like wheat or wheat flour generally unattractive.

There are two ways in which wheat or wheat flour is imported informally into Afghanistan.

The first is through underreporting of imports at the customs agency at the border. For example, when the truck arrives at the border, 8 tons of wheat are declared to customs and are subject to import duties when there are in fact 10 tons on the truck. This way, 2 tons of wheat, or 20 percent of the shipment, will enter the country unreported. The second way is by bringing it through routes not covered particularly well by the government of Afghanistan.

There are reportedly as many as 15 different entry points for wheat and wheat flour (among other things) to be brought into Afghanistan through informal channels from Pakistan.

Discussions with the trading community and other key informants indicate that the majority of this trade is centered in the Central Eastern provinces of the country, predominantly Paktia, Khost, Paktika, Nanghar, and Kunar. Figure 4 contains a map of estimates of the most significant routes.

3 It should be noted that July 2008‐June 2009 was precisely during the period of when Pakistan had a ban placed on wheat flour exports, and low domestic production in Afghanistan precipitated a very high import requirement.

As such, the figure of 1.5 MMT would not appear to be representative of the level informal imports prevalent on average over the longer period.

Figure 4. Informal Entry Points of Wheat and Wheat Flour

Table 7. Main Informal Border Entry Points-

Entry Point Province Nawa Pass Kunar Gandao Nangahar Khardani Paktia Ghulam Khan Khost Shkin Paktika

Regional Wheat trade flows and domestic wheat availability

This section compares price trends in Afghan wheat markets to an import parity price of wheat imported from Pakistan and Kazakhstan. Import parity prices for both Pakistan and Kazakh wheat prices were derived from discussions within the trader community during the execution of the market survey fieldwork. Figure 7 maps the Kabul wheat price against an estimated import parity price for imported wheat from Lahore Pakistan (green price series), and from Akmola, Northern Kazakhstan (blue price series)

Figure 5. Kabul Wheat Prices and Kazakh and Pakistani Import Parity Wheat Prices 2007-2012

Over the period 2007-2008 the Kabul wheat price tracks both import parity wheat prices relatively well. Prices diverge significantly in April-May 2008, which was driven in large part by the ban Pakistan placed a ban on wheat and wheat flour exports from the country at the time and the ban that Kazakhstan placed on grain exports the same year. Prices remained very high in Kabul until at least early 2009.

In the period starting in September 2009 until May 2010 Kabul prices were below the import parity price for wheat imported from Pakistan. One might presume that this may have been due to an influx of food aid. However, it would appear that low prices in Kazakhstan very likely could have been exerting a significant downward pull on prices over this period. In addition, a healthy harvest domestically would have had wheat prices in Afghanistan trending downward during this period (February 2009- May 2010).

From the figure above, it would appear that price changes were the result of relative prices and availability of wheat from different importing markets i.e. either Pakistan or Kazakhstan. The figure does not show significant evidence that prices would have dropped due to a large influx of food aid. During the period of 2010, when Kabul wheat prices dropped below the Pakistan import parity price, food aid cereal shipments totaled only 82,559 MT which was to be distributed throughout the entire country. This represents less than 2 percent of total available supply for Afghanistan. At that level of food aid, it seems doubtful that food aid would have been the reason for the drop in prices. The low prices of wheat available from Kazakhstan appear to be a more likely explanation for this occurrence.

U S$ /T on

Pak IPP Kabul Kaz IPP

A. Kabul wheat prices dip below Pakistani import parity price‐ September 2009‐ May 2010

B. Kazakh wheat prices putting downward pressure on Afghan prices

Box 1.1 The Winter of 2011/12: The Importance of Regional Trade

The key importance of regional trade to domestic food availability was highlighted very clearly in the winter of 2011 / 2012. In the spring of 2011, wheat production projections for Afghanistan were not favorable. Production was very low in most of the rainfed areas in the north of the country. The estimated average rainfed yield was only 0.2MT/Ha, a figure very near what would be considered crop failure. Prices in April increased sharply at a time when the expected trend should be down in anticipation of the coming harvest. There was concern in some areas that the situation could approach circumstances seen in the spring of 2008.

By contrast, for that year Pakistan had a very strong wheat crop, one of the highest on record. When the wheat crop in Kazakhstan came in in the fall, it turned out to indeed be its highest on record at 22MMT, far higher than the decade average of 13MMT and over double the previous year’s production level of 10MMT that had been strongly affected by drought. The result of this strong increase in production was that prices in Kazakhstan declined precipitously from over $300/MT in the spring of 2011 to $150-160/MT in September. The feeling was that despite the feeble harvest in Afghanistan in 2011, regional grain availability was quite good. Pakistan, with a healthy harvest, would not have significant incentives to place trade restrictions on wheat exports. Meanwhile large exportable surpluses from Kazakhstan, coupled with the very low prices prevalent there, should make substantial quantities of grain available for export to Afghanistan and Central Asia, and should put a significant downward pull on prices.

And, this is indeed what transpired. While the winter months of November-March are typically the lean period in Afghanistan where there is upward pressure on prices, despite the poor harvest domestically, prices over the period remained quite flat. This experience highlights the importance of regional trade in influencing staple food price levels in Afghanistan and indeed for much of the Central/South Asia region. It reinforces the importance of open trade policies for providing adequate levels of affordable wheat.

Wheat Prices North Kazakhstan (Akmola), Mazar Sharif, and Faizabad 2009-2012

U S$ /T on

N. Kazakhstan Mazar Faizabad

B. Prices up in response to poor production in rainfed areas

D. Prices flat through the fall and winter

C. Record harvest in Kazakhstan leads to large price fall

A. Poor production year in Kz puts upward pressure on prices

4. Domestic Markets for Wheat and Wheat Flour in Afghanistan

4.1 Market Integration

How Well Integrated are Agricultural Markets in Afghanistan?

For markets to serve their purpose they must function effectively. If an agricultural production shortfall in one region leads to a price increase there but does not lead to a price increase in a neighboring market, it suggests there is no effective market linkage between these two points. If that is the case, then market mechanisms alone would not be enough to rely upon for addressing food deficit problems in that region. One way of evaluating how effectively they are operating is to evaluate market efficiency, or how well markets are integrated. A poor degree of integration implies a lack of efficiency, and a situation where market mechanisms could not be relied on to address problems of food deficits. To evaluate this issue, the following section evaluates the level of market efficiency and market integration for wheat markets in Afghanistan.

Figure 6 shows trends in wheat prices over the period January 2004 to February 2012 for the markets of Kabul, Mazar Sharif, Herat, Kandahar, Faizabad, and Nili. Data is from the Afghanistan WFP VAM unit.

Figure 6. Wheat Prices Afghanistan, Selected Markets 2004-2012

Source: WFP

Four observations can be made from the figure:

1. In general across the 7 year period, prices tend to move together.

2. This includes during the price spike of 2008.

3. The main outlier is the rural market, Nili. This is still relatively well integrated with the other markets but not as well integrated as the urban markets. The reason is presumably transport and terrain issues. A key question is how indicative is Nili of other rural markets? If it is reasonably indicative (and this would need to be

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

A fs /k g

Kabul Mazar Faizabad Herat Kandahar Nili established) then overall one could say that urban and rural markets in Afghanistan are generally well integrated.

4. After 2008, all markets appear to be slightly less well integrated than before

To evaluate the question of market efficiency and market integration further, correlation coefficients were derived4 to evaluate the level of co-integration between these markets.

The results of this exercise are contained in Table 8.

Table 8. Correlation Coefficients Major Wheat Markets Afghanistan

Kabul Mazar Faizabad Herat Kandahar Jalalabad Nili

Kabul 1

Mazar 0.89 1

Faizabad 0.91 0.95 1

Herat 0.93 0.95 0.92 1

Kandahar 0.93 0.85 0.89 0.91 1

Jalalabad 0.97 0.90 0.92 0.95 0.95 1

Nili 0.80 0.67 0.70 0.83 0.89 0.84 1

Results of that exercise suggest that markets in Afghanistan are quite well integrated. Most correlation coefficients show a value of .8 or greater suggesting that the level of market integration between markets is relatively good. The exception to this is Nili, a market in the Central Highlands of Afghanistan. However, even for Nili, the markets where one would believe there would be a strong relationship, specifically Kabul and Kandahar as the major supplying markets, these relationships are relatively good. Clearly Nili is confronted with some challenges in terms of how well it is integrated at the sub-national level, with other markets in the country, but for its main supplying markets, market linkages appear to be relatively strong.

How well integrated are markets in Afghanistan compared to neighboring countries?

To evaluate the degree of market integration in Afghanistan, correlation coefficients for major markets in Afghanistan were compared to those derived for major wheat markets in neighboring Pakistan. The results of this exercise are contained in Table 9 below.

Table 9. Correlation Coefficients for Major Pakistani Wheat Markets

Lahore Karachi Multan Peshawar Quetta

Lahore 1

Karachi 0.93 1

Multan 0.99 0.93 1

Peshawar 0.87 0.97 0.88 1

Quetta 0.91 0.93 0.92 0.91 1

4 A correlation coefficient between two variables is defined as the covariance of the two variables divided by the product of their standard deviations. It is a value between 0 and 1 that provides an indication of the relationship between two variables. A value of ‘1’ implies complete correlation. A value of ‘0’ implies no correlation between the two variables.

Results would suggest the level of co-integration of markets in Afghanistan is at least roughly comparable to those in Pakistan. This is a relatively favorable development. Both countries show correlation coefficients of broadly between 0.85 and above for their major markets. The exception to this would be Nili, where the level of integration with other markets is somewhat lower. This issue of how well rural markets are integrated into national wheat markets will be evaluated in further detail in the next section which presents the results of survey work conducted in a number of major urban and rural wheat and wheat flour markets around Afghanistan.

4.2 Prices and Seasonality

There are very strong seasonal factors affecting food production and consumption in Afghanistan. The major wheat harvest in May-June puts most of the food on the market.

This is followed by a second crop of rice, maize and other minor crops, all of which are harvested in the late summer/early fall in August-September. After this there is the preparation for winter, followed by winter itself. During this time there are little if any additional foodstuffs coming onto the market from domestic production. Stocks need to be sufficient to last through these lean winter months. As a result of these seasonal trends, one might expect there to be seasonal behavior in prices in agricultural markets in Afghanistan to reflect these variations in supply and food availability. To evaluate this question, Figure 7 takes the same wheat price data set and matches up average wheat prices for each month over the calendar year to evaluate whether there are strong seasonal trends in prices.

Figure 7. Average Monthly Wheat Prices 2004-2007

For the markets of Kabul, Mazar, Herat, and Kandahar, average monthly prices through the year are quite flat suggesting that there is not a significant seasonal trend in prices in these markets. Given these strong seasonal cycles in production and consumption, one would expect that prices would reflect these changes. However, for the data available, it appears that they do not. One possible explanation for this is that management of stocks by both

0.00

5.00

10.00

15.00

20.00

25.00

30.00

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Kabul Mazar Herat Kandahar Nili the trader community and private households is quite effective in Afghanistan. Another possible explanation might be that regional trade has a strong role in influencing prices seasonally.

By contrast the market of Nili in the very rural province of Daikundi does in fact have a seasonal element to its price trend over the year; it is work noting however, that these price differences were not found to be statistically significant. Nevertheless there would appear to be somewhat of a seasonal price trend in this market with prices falling in the summer after the late spring harvest, and prices rising during the winter months. There were no data available for other markets in rural areas with conditions similar to those of Nili in order that this issue could be evaluated further.

Wheat Surplus and Deficit Areas and Major Commodity Flows

Figure 8 contains a map that shows both the wheat surplus and deficit regions of Afghanistan, and the major flows of wheat and wheat flour in the country. Major surplus areas are defined as provinces having on average a net surplus of 40,000 MT of wheat on average over the period 2007-2011 using wheat production data from the Afghan Ministry of Agriculture Irrigation and Livestock and population data from the Afghan Central Statistics Office (CSO). Minor Surplus areas are defined as those provinces having a surplus of between 20,000-40,000 MT on average over this same 2007-2011 period. Minor deficit areas are defined as areas having between 20,000-40,000 MT deficits on average over this period, and major deficit areas are defined as having a net deficit of more than 40,000 MT on average over this same period.

Figure 8. Wheat Surplus and Deficit Areas and Major Commodity Flows

Wheat surpluses are concentrated in the northern provinces of the country. Mazar Sharif in northern province of Balkh is a major market for both domestically produced wheat and imported wheat and wheat flour, the very large majority of which is coming from Kazakhstan via rail to the Termez-Hairaton border. There may be additional points of entry points for wheat coming in from Tajikistan via Kunduz, and to Faryab province from Turkmenistan at Andohy. However, these entry points are considerably smaller as compared to the Termez-Hairaton crossing.

From the markets of Mazar Sharif and Kunduz in the surplus northern areas, wheat is flowing east and south to the markets in Central Afghanistan, the major market for which is Kabul. Wheat is also flowing east from Kunduz into Badakshan. Wheat is also flowing west from Mazar Sharif to the market of Maimana in Faryab province.

The East and Southeast of the country are predominantly wheat deficit regions. In addition to the wheat and wheat flour coming from points north, large quantities of wheat flour are also coming into this region as imports from Pakistan. The overwhelming form of imports from this source is wheat flour rather than wheat grain. From the data available, Kabul city and province has an estimated wheat deficit of 628,000 MT annually. This deficit represents approximately 12 percent of total wheat consumption in the country.

From Kabul, imported Pakistani wheat also flows south to the Central East provinces, the major market for which is Ghazni. Imported wheat is also flowing northwest to the northern Central Highlands, the major market for which is Bamyan. In summer months when there is no snow, this wheat flour may also be transported further south to Daikundi province.

In the south of the country, Pakistani wheat is also imported via Quetta to Kandahar.

Importantly, Quetta and Baluchistan province in Pakistan are also wheat deficit areas. As such, Pakistani wheat flour imported to Southern Afghanistan is also very likely originating in the grain surplus areas of the Punjab and Sindh. Imported Pakistani wheat flour is consumed in Kandahar, it is also moved on to markets northeast into Ghazni and surrounding provinces. It also moves northward to Uruzgan and Daikundi provinces in the southern Central Highlands.

Additionally, there are quantities of Pakistani wheat flour moving west all the way to Herat province, with quantities also moving all the way up to Faryab province and the Maimana market. Herat is also the major supplying market for Chagcharan and other markets in Ghor province. The market linkage between Ghor province and the eastern Central Highland provinces of Bamian and Daikundi is quite weak due to challenging topography and poor infrastructure. These eastern Central Highland provinces instead are sourced with wheat from markets east from Kabul and Ghazni and from the south from Kandahar.

The Structure of Wheat Flour Imports

Field survey work afforded the possibility of estimating the structure of imports of wheat flour into Afghanistan. These results are tabulated in Table 10.

Table 10. Where are you usually buying your wheat flour?

Local Kazakhstan Pakistan

Kabul 0 % 52 % 48 % Mazar 35 % 53 % 12 % Jalalabad 0 % 18 % 82 % Herat 57 % 29 % 14 % Kunduz 21 % 79 % 0 % Kandahar 13 % 0 % 87 %

For urban markets there is considerable variability between locations in terms of how wheat flour is sourced. In Mazar Sharif, Kunduz and Kabul, over half the wheat flour was sourced from Kazakhstan. In Kandahar and Jalalabad, clearly Pakistani wheat flour is the predominant product traded. For Herat the results are more mixed. The high prevalence of locally procured wheat flour for the city is perplexing. However, for imported wheat flour in Herat, there was roughly twice as much Kazakh wheat flour in the market as Pakistani flour suggesting that northern sources are the dominant supply of flour in this market. The results of the survey would suggest that markets in different parts of the country have different trade orientations with neighboring countries.

To provide an alternative picture of the structure of these different markets, Figure 9 provides a schematic of market share of different types of wheat flour- Local, Kazakh, or Pakistani- by major markets around the country using the same data as in Table 9.

Figure 9. Market Share of Local, Kazakh, or Pakistani Wheat Flour by Market

Market structure and the feasibility of Cash Based transfer mechanisms to address food insecurity – General Indications.

5.1 Introduction

The preceding sections of this paper have reviewed trends in domestic production, informal imports, market integration, commodity flows of wheat within the country and other variables. In light of this preceding analysis, this section of the paper uses a partial Market Information and Food Insecurity Response Analysis (MIFIRA) framework to evaluate the feasibility of food vs non-food aid programs to address problems of food insecurity in Afghanistan. Much of the data used to do the MIFIRA analysis was collected during two field surveys in March – May 2012. A survey of urban markets was conducted in the markets of Kabul, Mazar Sharif, Herat, Kunduz, Kandahar, and Jalalabad in March 2012. A rural market survey was conducted in the markets of Saripul, Samangan, Ghor, Daikundi, Faizabad, and Khost in April 2012. All surveys were conducted in markets in the provincial capital and did not include survey interviews at the district level. For the urban market survey, between 25 and 30 traders were surveyed in each market. For the rural market survey, this number was reduced to 20 for the reason that for some markets, it may be challenging to find up to 30 respondents due to size. In sum, the exercise represents a survey of 290 traders in 12 urban and rural markets around the country.

The MIFIRA framework is an analytical tool developed as a collaborative effort between Cornell University, CARE International, and Tufts University that is designed to evaluate how well markets are functioning with a view to informing the feasibility of response options regarding cash transfer, voucher programs, or food aid. Effectively, if markets are found to be functioning under the framework, it suggests that a cash transfer, voucher, or other market based mechanism is likely the preferred response option. If markets are found to be not functioning either due to lack of market access, civil conflict, or other variable, then food aid or other in-kind assistance programs are a resulting preferred response option. To date, the MIFIRA framework has been applied to food security response analysis circumstances in Kenya and Somalia. At the national or sub-national level, the MIFIRA framework is a set of five questions-

1. Are food insecure households well connected to local markets?

2. How much additional food can traders supply at or near current costs?

3. How will local demand respond to transfers?

4. Do local food traders behave competitively?

5. Do food insecure households have a preference over the form of aid they receive?

These questions can be broken out into a decision tree that assists in determining the preference of the form of aid received to address food insecurity.

Some of question In relatio any distr 3, some elasticity should in preferen ‘Do food

In the lig need to

In order sheds ar survey a in follow Afghanis

5 Data avai for wheat / spent on fo

Figure the questio s in the field on to questi ricts were cu national lev y of demand ncrease in re ces with reg insecure ho ght of this, t be complete to structure re distinct ge allowed the i ing the resp stan. Table 1

Table

Ka Ma Ja He Ku ilable from the / wheat flour s ood and within e 10. MIFIRA ons are supp d survey ins on 1, the su ut-off from vel secondar d in relation esponse to gard to tran ouseholds h the following ed with furth e the MIFIRA eographic re identificatio ponse to the 10 tabulates

11. Where abul azar lalabad eart unduz e NRVA (2003?

specifically of 0 n this Af 0.4 wo

A Decision Tr ply side que strument sp urvey gener markets at ry data was to food. Th cash based nsfer types a ave a prefe g is best vie her primary

A analysis a egions in wh n of five dis e question: W s the results are you sellin

Kabul, Par Mazar/Bal Nangaha

K

???) indicates

0.4. This means

ould be spent o ree estions, and pecifically de rated indirec certain time s used to ga e data indic transfers5.

and was the rence over ewed as a p y and second a “market-sh hich traders stinct marke Where are y s:

ng your whea rwan, Logar,G kh, Saripul, J ar, Laghman.

Herat, Gho

Kunduz, Bada a household in s that for every on wheat / wh some are d esigned to a ct answers b es of the ye in a very ge cates that ho The survey erefore unab the form of artial MIFIR dary data.

hed ” appro s operate. S et sheds in t you selling y at/flour in Afg

Ghazni, Bam Jawzjan, Fary Kabul, Kuna or, Badghis akshan, Balkh ncome elasticit y increase in in eat flour.

demand side answer ques by asking tr

ar. In relat eneral indica ousehold de y did not ad…