Winners and losers of the global food crisis : the case for Sri Lanka

Thesis BSc. Economics and Business

Faculty of Economics and Business

Winners and Losers of the Global Food Crisis

The C

Supervisor: Menno Pradhan

Amsterdam, 29

Thesis BSc. Economics and Business

Economics

Faculty of Economics and Business

Winners and Losers of the Global Food Crisis

The Case for Sri Lanka

By

Leah Situnayake

10652949

Supervisor: Menno Pradhan

Amsterdam, 29

June 2016

i

Statement of Originality

This document is written by Student Leah Situnayake who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

ii

Abstract

The spike in global food prices from 2006 to 2009 has raised concerns about the implications on the poverty and welfare of developing countries. Many were pushed into a situation of food insecurity resulting from the break out of the global food crisis. In this paper, the impacts of a rice price increase on Sri Lanka are investigated. This provides novel results as it adds to the understanding of whether Sri Lanka is sensitive to such local rice price

changes, which has not been applied to this country in the past. By analyzing household data, the impacts on welfare and poverty for rural, urban, estate, and different income deciles are measured. Unlike previous research, results indicate that rural and estate sectors suffer from rising rice prices in terms of welfare and poverty. These results are determined by the consumption patterns of Kekulu rice. As a majority of the households are net rice consumers they are negatively affected by a rise in their consumption bundles.

iii

Table of contents

Statement of Originality ………..i

Abstract ………ii

1. Introduction ……….1

2. Literature Review ………...4

3. Research Methodology and Data ………...10

3.1 Methodology ………...10

3.2 Household Data ………..11

4. Results and Discussion ……….14

4.1 Household Expenditure and Welfare……….14

4.2 Household Poverty Impacts………18

4.3 General Discussion………..20

5. Conclusion ………22

Bibliography ………..23

1

1. Introduction

In 2008, rising world food prices triggered the break out of a severe global food crisis (Cudjoe, Breisinger, & Diao, 2010). The Food and Agriculture Organization (FAO) indicated that by April 2009, cereal prices were at a record high in 10-30 percent of the countries studied, confirming that this crisis was of a widespread nature (Swan, Hadley, & Cichon, 2010). A sharp increase of 24 percent in the FAO food price index and a 43 percent increase in the cereal price index were witnessed (Vu & Glewwe, 2011). In addition, many key staples such as wheat, soya beans, rice, and corn experienced a price doubling (Clapp & Helleiner, 2012). This sudden spike in international food prices has largely been termed “the global food crisis”, nomenclature that instilled panic which resulted in the outbreak of numerous food riots (Headey, 2014).

The rising food prices have had several detrimental effects. As food and fuel prices increased, millions have been pushed into a situation of food insecurity; estimates of the number of undernourished individuals from the FAO have increased from 850 million to 963 million from 2005 to 2008 (Swan, Hadley, & Cichon, 2010). For the first time in mid 2009, this estimate topped 1 billion resulting in a striking delay of the UN goal of 2015 of keeping this number under 420 million (FAO, 2009 (as cited in Clapp & Helleiner, 2012)). This outstanding deviation from the UN goal was a great cause for concern. In addition, the World Bank estimated that 105 million people have sunk into poverty since the crisis hit (Estrades & Terra, 2012). Such price volatility has been particularly disruptive to those involved in food production and to consumers in developing countries (Clapp & Helleiner, 2012). Concerns that the developing world could fall deeper into poverty were raised following the food crisis (Vu & Glewwe, 2011).

This global food crisis can be attributed to a combination of various factors. According to Estrades and Terra (2012), these factors include the increase in demand for some crops, falling supply, and severe weather conditions. Across several studies, the major contributing factor is the rise in oil prices, which in turn increased transport and production costs, and pushed up food prices (Estrades & Terra, 2012). Multiple explanations for rising demand have been discussed. Energy price rise, globalization, urbanization, income and population growth, and subsidized bio fuel production are examples of some (Robles & Keefe, 2011). On the other hand, the reduction in supply of crops may be due to underinvestment in rural

2

infrastructure, weather conditions, land and water constraints, and agricultural innovation which have impaired the necessary production response to meet rising demand (von Braun & Torero 2009; see also von Braun et al. 2008; UNCTAD 2008).

In this paper, the effects of the global food crisis on Sri Lanka, a developing country in South Asia, are discussed. The beginning of the global food crisis from 2006 to 2007 will be analyzed in the present research. In order to measure such effects, the implications on poverty and welfare are analyzed. Welfare is a measure of utility that is calculated using the equation introduced by Vu and Glewwe (2011), which is presented in the methodology (section 3.1). This measure may not be representative across different countries. However, it is appropriate when looking into Sri Lanka (for reasons which will be clarified in the

subsequent sections). On the other hand, poverty changes are analyzed based on the widely used standard poverty measure, which will be explained in the following sections.

As Sri Lanka has been recovering from a 30 year war that ended in May 2009, sharp increases in food prices will affect the economic recovery of the country and could result in a poverty rise in the country’s war torn regions (Naranpanawa & Bandara, 2012). As a result of the civil war experienced in Sri Lanka, poverty has been of great concern, mainly concentrated in rural areas and other under-developed regions (Naranpanawa & Bandara, 2012). The implications of the global food crisis on Sri Lanka have hardly received much attention. It is therefore important to investigate the effects of rising food prices on welfare and poverty in order to provide assistance in government policy analysis in Sri Lanka. Further, this study aims to assess the applicability of the equations presented by Vu and Glewwe (2011) to the case of Sri Lanka, and to identify what data processing needs to be carried out in order to make this approach appropriate for Sri Lanka.

The objective of this research is to analyze the effects of the rising rice prices from the beginning of the global food crisis, 2006 to 2007 on the welfare and poverty of households in Sri Lanka. Prior to the global food crisis, Sri Lanka experienced higher non-food prices compared to food prices, which changed after mid 2006 when the spike in prices resulted in food price inflation exceeding non-food price inflation (The World Bank, 2010). Domestic rice price spikes were also experienced. As a large number of households in Sri Lanka are net rice buyers a decrease in welfare is expected (The World Bank, 2010).This may be caused by an increase in the cost of the (regular) consumption bundle resulting from an

3

increase in food prices, particularly in the price of rice, a major food staple which accounts for a large proportion of the consumption bundle (The World Bank, 2010). The food crisis may particularly threaten the welfare of poorer household which generally spend a larger portion of their income on food (The World Bank, 2010). In Sri Lanka, rice is the major staple and the most important crop occupying 34 percent of the total cultivated land area (Department of Agriculture, 2006). As stated by the Department of Agriculture (2006), Sri Lanka produces around 95 percent of the domestic requirement and spends a large

proportion of income on rice, indicating its importance. Data retrieved from the World Bank represented that most of the population live in rural areas and are employed in the agricultural sector, suggesting the relevance of measuring the impacts of commodity prices on Sri Lanka. In this study, the type of rice analyzed is red Kekulu rice (red rice). The Household Income and Expenditure Survey (2006/07) indicated that this type of rice is the most consumed among households in comparison to other types of rice which is why this has been chosen.

As rice is a principal staple food item, the focus of this research will be on rice prices in order to investigate the implications of the global food crisis. Almost half of the average calorie intake of the Sri Lankan population consists of rice, signifying its importance in Sri Lanka (The World Bank, 2010). As Sri Lanka has not been looked into in the past and is highly dependent on rice, this has motivated the research question of the paper: To what extent do the rising kekulu rice prices in the wake of the global food crisis impact welfare and poverty in Sri Lanka?

In order to tackle this research question this study will be structured as follows. Firstly, an overview of the literature will be presented, where various existing models which assess poverty and welfare are compared, and the model chosen for this study will be justified. Here, the hypotheses of this research are also presented. Secondly, the methodology and data used is discussed, exploring the equations from Vu and Glewwe (2011) as well as the steps taken to make it applicable for the data available for Sri Lanka. Thirdly, the results are presented, where the findings are offered to the aforementioned research question. Finally, the results are analyzed and discussed, and the data is concluded, with suggestions for further research.

4

2. Literature Review

Sri Lanka has been subject to many natural disasters. In addition to the Tsunami experienced in 2004, floods seem to be more of a common occurrence in comparison to other natural disasters ("Sri Lanka Disaster Knowledge Network", n.d.). As previously touched upon, weather conditions may be a contributing factor of the increase in food prices. Such disasters may cause a spike in food prices, demonstrating the importance of measuring their impact on welfare and poverty.

In the paper by Vu and Glewwe (2011) they stressed upon the fact that the impact of rising food prices depends on the household characteristics, on whether the household is a food consumer or producer. The food crisis may also have effects on welfare, particularly on poor households just above the poverty line (De Hoyos & Medvedev, 2009). This will be clarified in the subsequent sections. As stated previously, the focus of this paper will be on rice, Sri Lanka’s major staple and the structure is as follows. First, the welfare outcomes found in various studies are discussed. Second, poverty effects are analyzed and finally the models used will be reviewed and compared in order to enrich the background for the method used in this study.

2.1 Welfare Effects

Much has been studied on the effects of the global food crisis on welfare in various countries. Estrades and Terra (2012) conducted a study in the case of Uruguay, a small developing agriculture-oriented economy. Uruguay is an atypical case as it is a net food exporting country and a crude oil importer, with its poverty mainly concentrated in urban areas. Welfare of households’ is dependent on consumption possibilities that fluctuate with income and prices of commodities. The study showed that the wealthier the household, the larger was the increase in welfare, as wealthier households spend a lower proportion of their income on food (Estrades & Terra 2012).

Wodon and Zaman (2010) conducted a study on the effect of the global food crisis on net food importing countries. The authors looked at Sub-Saharan Africa during the year 2008 and expected that most of the countries would be negatively affected as they are predominantly net cereal importers. Unlike net producers of food, net consumers tend to be

5

negatively affected from commodity price rises as their consumption bundle becomes more costly (Wodon & Zaman, 2010). Wodon and Zaman (2010) suggested that in most cases the urban poor are net consumers and thus would be hurt by an increase in commodity prices. Indeed, their results show that although many in rural areas are net consumers of food, urban households were hit harder as urban areas are predominantly net consumers of food. Finally, consistent with global estimates, the authors concluded that those who were already below the poverty line experienced a loss of welfare, indicating that it is the present poor who suffer most.

Dimova and Gbakou (2013) analyzed the impact of the rise in prices of a particular food item on welfare of net food buyers and sellers in Côte d’Ivoire. Similar to Sub-Saharan Africa, Côte d’Ivoire represents another case for a net food importing country, enjoying a comparative advantage in cash crops (Dimova & Gbakou, 2013). Using household data from Côte d’Ivoire from the peak of the crisis, Dimova and Gbakou (2013) found that the urban sector loses out and those next to the poorest in rural areas benefit, which is in line with the results from Wodon and Zaman (2010). By determining whether an individual is a net food consumer or producer, a trend can be observed. Dimova and Gbakou’s (2013) findings indicate that those able to generate income from cash crops enjoy an improvement of welfare in rural areas, but this also helps smooth consumption patterns of the urban households.

Furthermore, research carried out by Vu and Glewwe (2011), whose methodology has been used as the basis of the present research, provides valuable insights into this topic. In this study the authors focused on Vietnam, being a poor developing country where the average Vietnamese household spends half its income on food. This indicated that food price changes will have adverse effects on welfare (Vu & Glewwe, 2011). Similar to the previous findings, Vu and Glewwe (2011) mainly focused on net food purchasers or net food

sellers. They found that welfare is more likely to reduce when a household is a net food

purchaser than a net food seller. The effects of both rice and food prices were analyzed as rice is both produced and consumed by most rural households (Vu & Glewwe, 2011). Vu and Glewwe (2011) find that, with the use of domestic prices and data from the 2006 Vietnam Household Living Standards Survey (VHLSS), a rise in food prices increased the average household welfare. However, as a result of higher food prices, urban households’

6

welfare almost always reduces as they are net purchasers of food (Vu & Glewwe, 2011). Therefore, it is clear that the findings regarding the welfare implications seem to correspond in many studies.

2.2 Poverty Effects

Similar to welfare, the effects of the global food crisis on poverty was also extensively studied. The micro-simulation results of Estrades & Terras’ (2012) study on Uruguay, a net exporter of food, indicate that the rise in agricultural commodity prices led to extreme poverty because consumption goods became more expensive. Therefore, individuals who are already poor suffering from lower/no income were hit the hardest. However, overall poverty declined due to an increase in wages (especially for unskilled labor) for those benefitting from higher farm incomes (Estrades & Terra, 2012).

Other cases, such as Warr and Yusuf’s (2009) study on Indonesia, a net importer of food, explained that the poverty increase during the global food crisis was because the owners of agricultural land and capital benefited at the expense of the poor, who suffered from the price increases of staple food items. The benefiting parties from this price rise did not outweigh those who lost out, which caused the worsening of the poverty rate in

Indonesia (Warr & Yusuf, 2009). Exploring the magnitude of poverty benefits and losses is therefore important when considering the overall impact of the food crisis. Consistent results were found by Cudjoe, Breisinger, and Diao (2010) for Ghana and by Rocchi,

Romano, and Hamza (2013) for Syria. Finally, all these studies suggested that an increase in poverty was experienced by urban households that are generally net food consumers.

A study carried out by Lu and Yu (2011) on China, concluded that the impact on poverty varies by crop production, but in general poor households are more likely to lose from higher output prices due to lower productivity and market participation. Therefore, poverty effects may change according the crop produced. In addition, longitudinal survey data on Bangladesh showed that the high food prices during 2007 and 2008 resulted in a higher cost of living, which was counteracted by an increase in farm income (Balagtas, Bhandari, Cabrera, Mohanty, & Hossain, 2014). Overall, a 25 percent decline in the poverty incidence is experienced as the majority benefit from higher farm income. These results were found by estimating panel models of household income on a subsample of 964

7

households, with the use of four survey waves where the poverty effects during each survey wave were measured (Balagtas, Bhandari, Cabrera, Mohanty, & Hossain, 2014).

A study conducted by Valero-Gil and Valero (2008) states that when the price hike was more severe, poverty increases were larger. Such results were obtained using a partial equilibrium analysis where first order impacts were accounted for, for the years 2006 to 2007 and 2007 to 2008. In general, many studies have indicated that the poor allocate a larger proportion of their income to the consumption of food products, which in some situations may hit the poor the hardest, affecting welfare and poverty (Valero-Gil & Valero, 2008).

In contrast, the study conducted by Vu and Glewwe (2011) on Vietnam showed that with a moderate rise in food prices, a reduction in rural poverty and an increase in urban poverty were experienced, lowering the national poverty rate. However, when food prices experience a sharp increase, urban poverty rose dramatically causing the national poverty rate to increase. Overall, the authors found that middle income households in rural areas benefit the most while poorest households in urban and rural areas do not benefit from an increase in rice prices. Poorest households generally don’t benefit as they do not profit from higher farm incomes and suffer from an increase in their expenses (Vu & Glewwe, 2011). Thus, it is important to weigh the effects on urban and rural areas in order to determine the overall effect.

Such findings have been consistent across the literature. De Hoyos and Medvedev (2009) collated the results from various case studies and examined the effects of higher food prices from a more global, wider context. Their results concluded that with higher food prices, poverty increased in developing countries (Sub-Saharan Africa and South Asia), while it reduced significantly in Latin America, where farm incomes were higher. Overall, however, those living in extreme poverty rose by 32 million people globally. Therefore, it is clear that when considering a particular case, the influence of those working in the agricultural sector must not be overlooked when studying the influence of food prices on welfare and poverty effects. Taking a global stance on this crisis, such findings indicate that poverty is likely to increase for those that do not benefit from higher farm income.

8

2.3 Methodological Variety

Multiple approaches have been used in the past to provide some insight on the global food crisis in order to predict poverty and welfare responses to food price fluctuations. One widely used model across the literature is the Computable General Equilibrium (CGE) model. This is a single country static model assuming perfect competition in goods and factor markets (Estrades & Terra, 2012). The advantage of this model is that controlled experiments focusing on household incomes, poverty, welfare, and economic shocks can be conducted (Warr & Yusuf, 2009). In the study carried out by Estrades and Terra (2012), the authors analyzed the effects of international price changes, aiming to design policies to alleviate such effects with the CGE model with micro-simulations. This model adopts assumptions on the nature of price changes and examines the channels through which the prices affect welfare (Estrades & Terra, 2012). However, the CGE model used in these papers only provides a static view of the economy. This is not reflective of reality as

consumption and production patterns are constantly changing which can differ between and within countries and cultures (Karwala, 2005). Thus, the use of other methods may be useful in assessing these effects with respect to different angles.

Many other studies employing diverse models have been conducted with regard to the implications of the global food crisis. The methodology used in the study conducted by Vu and Glewwe (2011) is the compensating variation. This is utilized in order to measure welfare effects, where the change in utility is considered in order to obtain the results. It was the approach from their study that was applied in the current research due to its unique perspective that had not been presented previously. Despite the vast amount of research that has been carried out regarding the food crisis, the present study will provide insight on Sri Lanka, a country that has not yet been considered in this context. By applying a novel model to such an unexplored case, this study aims to explore the applicability of the model by Vu and Glewwe (2011), whilst simultaneously providing fresh insight on the responses of Sri Lanka’s economy to food price fluctuations.

Overall, all the models discussed present similar results. The general consensus appears to be that poverty worsens in urban areas as they experience an increase in their

consumption expenditure. However, those working in the agricultural sector benefit from higher farm incomes, simultaneously reducing poverty in this sector. In addition, extreme

9

poverty seems to increase as those who are already poor are hit harder with increased food prices and fall deeper into poverty. Furthermore, the overall impact on poverty depends on the economic situation of the country - if the country is a net food exporter or importer and the individual poverty impact on the urban and rural areas (dependent on being a net food consumer and producer). For welfare, the impact depends on the consumption possibilities of the household. For a small developing country with a majority of net food buyers, a

reduction in welfare can be expected.

In Sri Lanka, a decline in welfare is expected as most households are net rice consumers (The World Bank, 2010). In addition, the protection of the domestic rice market has

reduced the motive to export rice and therefore the country does not export rice which also may hurt the welfare of households as they do not benefit from exporting rice (The World Bank, 2010). Finally, as many of the households consume rise it is predicted that poverty in rural and estate sectors (where rise consumption is the highest) increase as such households suffer from a rise in the cost of their consumption bundle. Whereas, poverty may decline in urban areas as net consumption of rice is lower. This leads to the hypotheses tested in this study. On the backbone of the present literature, the following hypotheses for the case of Sri Lanka are formulated as follows:

Hypothesis 1: As a result of the increase in rice prices, a rise in rural and estate poverty and a decrease in urban poverty are expected resulting in an increase in overall poverty, as majority of households are net rice consumers.

Hypothesis 2: Due to rice price increases, welfare is expected to fall in rural and estate areas, where majority of the population are net food buyers and rise in urban areas as net consumption of rice is not as high.

10

3. Research Methodology and Data

3.1 Methodology

This study uses the model in the research carried out by Vu and Glewwe (2011) and applies it for the case of Sri Lanka. Vu and Glewwe (2011) used the compensating variation in order to measure the impact of price changes on household welfare. The compensating variation represents the amount of money necessary to maintain the previous level of utility, prior to the rise in food prices. Thus, the implications that the food crisis has on welfare in Sri Lanka will be measured using the compensating variation introduced by Vu and Glewwe (2011). This equation measures the impact of food price changes on welfare, by determining the monetary value households would require in maintaining their previous standard of living (Deaton, 1989). As explained by Boianovsky (2014), welfare should be based on cardinal utility. Welfare must be directly observable, which may be in situations where utility can be expressed as consumption. Thus the change in welfare is measured using the equation below. Short-run effects are not measured in this study, as consumption patterns may change due to substitution of goods. Therefore, the equation below measures the immediate impact.

(1)

=

−

Δ ln Bi measures the compensation required to maintain the previous level of utility, which is the money necessary to achieve the previous standard of living (Deaton, 1989). This equation measures the impact on welfare by analyzing the change in price for one single commodity, which is kekulu rice in this case. As mentioned by Vu (2008), the variable wi is a measure for the budget share of good i. This is the consumption expenditure of good i, as a fraction of total consumption expenditure. The variable (ppi*yi)/X represents the production of good i as a fraction of total consumption expenditure (X). The other variables ppi and pci measures producer and consumer prices respectively and yi represents the

production of good i.

The intuition of the first equation is as follows: as explained by Deaton (1989), wi Δln(pci) represents the producers profit or the consumer cost resulting from a change in consumer prices. This is the change in consumption expenditures resulting from a rise in consumer prices. The second half of the equation, ((ppi*yi)/X)*Δln(ppi) is the producers cost

11

(value of production) resulting from a change in producer prices (Deaton, 1989) – a measure of the change in the production cost of the good. Thus, it is clear that the difference between the gain and cost of production resulting from changing prices and production patterns will give us the change in income necessary to maintain the previous level of utility (Deaton, 1989).

Finally, in order to measure the poverty impacts of the global food crisis the indices introduced by Foster, Greer, and Thorbecke (1984) are used. Mathematically, this is calculated as:

(2) ( )

= ∑

1 −

Where α ≥ 1 for ki < z

The parameter ki measures the per capita expenditure or the calorie intake of household i, z measures the poverty line, and N represents the total number of persons (Vu, 2008). The index P(0) is the head-count poverty index, which is the percentage of people that are poor and P(1) measures the poverty gap, representing the average gap between the income of the poor and the poverty line, where the non-poor and poor are assigned a gap of zero (Vu, 2008). The third index P(2), measures the severity of poverty which is the squared poverty gap (Vu, 2008). The poverty measure presented is appropriate as the changes in the poverty rate provides an indicator of society's success in alleviating suffering among those with relatively low incomes (Ellwood & Summers, 1985), which makes it suitable for the case of post-war Sri Lanka.

3.2 Household Data

The data used was gathered from the Department of Census and Statistics (DCS) – Sri Lanka. The household income and expenditure survey (HIES) was conducted from July 2006 to June 2007 and carried out in all districts in Sri Lanka, excluding the Trincomalee district and the Northern Province. From a total of 21,790 housing units, 26.6% were from the urban sector, 63.9% from the rural sector and the remaining 9.5% were from the estate sector. The effects of welfare and poverty will be investigated for the whole of Sri Lanka and for each of the three sectors individually. In addition, the welfare effects on the different income deciles are measured. The various income deciles have been distributed according to

12

the HIES mean income threshold for the purpose of this study. According to the DCS, urban sectors are defined as areas governed by the municipal council and estate sectors are defined as plantation areas with over 20 acres of extent. Rural areas are characterized as residential which do not fall under the urban and estate criterion.

However, as almost all the districts in Sri Lanka (including those with land over 20 acres and low population densities) are governed by the municipal council, this does not seem to be a fair representation of the three sectors in Sri Lanka. Despite personal contact with the DCS, this allocation was not clarified and the criteria for how districts were divided

amongst the sectors were not made available. Because of this, this study reclassifies districts between the sectors based on their population densities. Densities are an appropriate

measure for this form of classification and have also been used in great metropolitan contexts (Clark, 1951). Accordingly, districts that are among the highest density will fall under urban areas (around 500 to 3000 persons per square km), estate sectors will be the plantation districts and the rest will fall under rural areas. Similar to the definition of the HIES, plantation districts are defined as areas with over 20 acres of land and with 10 or more resident laborers. These divisions will provide a more accurate representation resulting in better results compared to the divisions stated by the DCS. As paddy production has not been provided in the HIES, paddy production retrieved from the DCS have been allocated to rural, urban, and estate sectors using this method.

The data provided in this survey has been averaged out for the time period from July 2006 to June 2007 which will be used for the analysis of this research. This time frame is significant, as a rise in prices is witnessed from mid 2006. Consumer and producer prices are obtained from the DCS. According to the DCS, the all-island producer prices are provided for the year 2006 and 2007 for each agricultural product. The consumer prices, measured as consumer price index (CPI), are available on a monthly basis but have not been

distinguished per food item. In order to provide more accurate price data monthly producer prices have been gathered from the Hector Kobbekaduwa Agrarian Research and Training Institute (HKARTI) which will be used for the computations of this study. Similar to the study conducted by Vu and Glewwe (2011), consumer and producer prices are assumed to increase at the same rate due to the lack of data available. This assumption is suitable because a price increases in the wholesale market is instantly transmitted to the retail level,

13

as Sri Lankan markets are integrated (Korale Gedaraa, Ratnasiria, & Bandara, 2016). Data regarding monthly paddy production has been obtained from the paddy statistics of the DCS. The averaged monthly paddy production per household (yi) was computed in order to measure the welfare implications for 2006 to 2007.

In the poverty analysis section, data regarding calorie intake/per capita expenditure and the poverty line were obtained from the DCS. The official poverty line (OPL) for 2006/07 was Rs. 2233 real total expenditure per person per month and the nutritional anchor for the OPL is 2030 kilo calories. The OPL represents the minimum level of income necessary to afford basic necessities (Department of Census and Statistics, 2006/07). The DCS (2006/07) chooses the absolute poverty line as a threshold, an approach used in many developing countries. This way, changes in poverty over time can be checked with reference to the fixed poverty line. This ensures that the same standard of living is represented across time, whilst accounting for variations in the cost of living and changes in inflation as measured by the Colombo consumer price index (CCPI). Many individuals may move above the poverty line due to inflation, therefore it is crucial to account for inflation, as has been accounted for in the OPL. The change in the poverty gap is calculated by computing the change in per capita consumption expenditure using the change in rice prices from July 2006 and June 2007. Due to data limitations only the change in rice prices are taken into account assuming that the same quantity of rice is consumed. However, this may not necessarily be the case as some households may substitute Kekulu rice for a cheaper alternative – as already mentioned, substitutions of goods have not been accounted for as immediate effects are taken into account. In the preceding section the implications of the crisis in numerical terms is presented. This measures the degree of the effect rice price changes have on welfare and poverty. Thus, with the use of the equations discussed above, the results will be computed.

14

4. Results and Discussion

4.1 Household Expenditure and Welfare

In the following section the impact of rice price increases on welfare for each sector is discussed. Welfare implications on the separate income deciles have also been computed. Focusing on the time period from July 2006 to June 2007 the results have been compiled.

The food items presented in table 1 resembles the major staples consumed and the price change experienced for each food commodity. There is an increase in prices for each food commodity from 2006 to 2007 for all island producer prices (Table 1). The largest increase in producer prices is in the commodity Kekulu/red rice. Kekulu rice prices experienced a rise of 52.8 percent. This is a considerable increase, foreboding detrimental effects on welfare and poverty. According to the HIES 2006/07, the rural and estate sectors’ budget share of

kekulu rice is around 2.7 and 3.2 percent of total expenditure, compared with a much lower

budget share of only 1.2 percent in urban areas (See appendix C). This is indicative of the higher expenditure on kekulu consumption in rural and estate areas.

Data compiled in the 2006/07 HIES regarding the percentage of average monthly household income from agricultural activities, is presented in table 2. Such results present the importance of agricultural income for each sector. As these show, agricultural income is lowest in the urban sector and highest in the estate sector, demonstrating that unlike the urban sector, majority of the income earned in the estate sector is from agricultural activities (Table 2). However, this does not simply mean that welfare is likely to improve in estate sectors. The welfare effects are dependent on the commodity studied.

Furthermore, the kekulu rice producer price increase from Rs. 11.97/kg in July 2006 to Rs. 18.29/kg in June 2007 has resulted in a 52.8 percent increase in kekulu rice expenditure as well. In this study immediate impacts are considered while substitution effects resulting from price changes are not. Therefore, the change in consumption expenditure is

representative of the change in kekulu rice prices. As rice expenditure has experienced a sharp increase from July 2006 to June 2007, discussing the impacts this price increase may have is important.

15

Table 1

All Island Producers' Prices for Selected Agricultural Products: 2006‐2007

Item Unit 2006 2007 % Change 2006- 2007

Red Rice Kg. 11.97 18.29 52.80 Cereals Kg. 46.95 55.93 19.13 Vegetables Kg. 25.9 28.31 9.31 Fish Kg. 176.12 241.72 37.25 Dried Fish Kg. 247.91 315.04 27.08 Coconuts 100 1291.3 1577.99 22.20 Fruits Each 13.42 17.16 27.87 Condiments Kg. 191.03 251.49 31.65 Milk 1L 23.94 30.11 25.77

Source: Authors’ calculations with data from HIES 2006/07

Table 2

Percentage of Average Monthly Household Income by Sector: 2006/07

Sector Total income (%) Agricultural activities (%) Non agricultural activities (%) Wages/ salaries (%) Other cash income (%) Income by chance/adhoc gains (%) Non monetary income (%) Monetary income (%) Sri Lanka 100 12.1 17.1 35.8 11.9 9.1 14 86 Urban 100 3.3 20.3 35.9 14.4 10.4 15.7 84.3 Rural 100 13.5 16.9 35.6 11.3 8.9 13.7 86.3 Estate 100 34.1 4.2 39.9 8.5 4.8 8.4 91.6 Source: HIES 2006/07

16

Using equation 1 presented in the methodology section, the welfare effects of rice price increases have been computed and compiled in table 3. As mentioned previously, due to data limitations the focus will be on producer prices. The immediate welfare change

measured in table 3 suggests that the average household welfare decreased by 1.78 percent. This is a moderate decline in welfare. Focusing on separate sectors, a rise of 0.42 percent in urban welfare is seen whereas, welfare decreases by 5.79 percent and 1.13 percent in rural and estate sectors respectively. Although estate sectors earn the majority of their income from agricultural activities (see table 2), they also experience a reduction in welfare as they are net consumers of rice. As a result, the rise in rice prices will increase their expenses, which in turn reduces welfare. The same is seen for rural areas as they are also net consumers of rice.

According to the World Bank (2010), consumption of Kekulu rice is highest among rural and estate households but lowest in urban households. This explains the small rise in

welfare experienced in the urban sector – as rice consumption (budget share spent on rice is much lower) is not as high in comparison to rural and estate sectors. Production data indicated that the urban sector produces a small amount of paddy, which translates into a moderate rise in welfare (0.42 percent). This result may be explained by the low

consumption of kekulu rice and the production of paddy, which increases welfare due to a small rise in farm income. A reduction in welfare that is witnessed among poorer

households is due to the proportion of income spent on food – again, poorer households tend to be hit harder (Estrades & Terra, 2012). An overall welfare decrease by 1.78 percent suggests that majority of the households are net consumers of Kekulu rice and generally suffer from a higher cost of their consumption bundle. According to the DCS consumption data, as Kekulu rice is mainly consumed by the poorer population, welfare may reduce among poorer households. This indicates the importance of considering which commodity has been analyzed, as the welfare effects may change according to the commodity studied. These effects are subject to change when looking into different types of rice as well.

Rice price increases tend to improve the welfare of middle income earners unlike the higher and lower income earners (decile 4). The welfare loss of 0.5 and 0.29 percentages in decile 1 and decile 2, suggests that lower income households experience a reduction in welfare as a larger proportion of income is spent on food. This may be due to the higher

17

consumption costs experienced in these households compared to household income. The welfare decline experienced in the income deciles is quite small. Considering this was the beginning of the global food crisis, the rise in rice prices was moderate which explains the small change in welfare. Upper middle income earners (decile 3) seem to benefit the most, as they are more likely to be spending a lower proportion of their income on Kekulu rice. Therefore, a 0.21 percent increase in welfare is seen in decile 3. This indicates that upper middle income earners are less likely to suffer from an increase in the cost of their

consumption bundle, as welfare increases. However, the highest income earners (decile 4) seem to experience a reduction in welfare of 0.39 percent. This is an unanticipated result as they earn a higher income and it is expected that the rise in rice prices may not affect these income earners as harshly. This unusual result may be explained by the production

computations and price assumptions made in this research.

Table 3

Household Welfare Changes Due to Increases in Rice Prices (%): 2006/07

Sector, Decile Group Change in Welfare 2006/07 (%)

Sri Lanka -1.78 Sector Urban 0.42 Rural -5.79 Estate -1.13 Income Deciles Decile 1 -0.50 Decile 2 -0.29 Decile 3 0.21 Decile 4 -0.39

18

4.2 Household Poverty Impacts

To better understand the poverty situation during the food crisis, using the poverty gap index, the change in poverty from 2006/2007 is measured (Table 4). The poverty gap index (P(1)) is a measure of the average gap between the income of the poor and the poverty line. This is a good measure of the depth of poverty (Department of Census and Statistics, 2006/07). The change in poverty is measured by analyzing at the impacts on per capita consumption resulting from a rise in rice prices. Initial poverty is computed using the average per capita consumption or daily caloric intake of the poor, during the time period 2006/07. The change in poverty is then computed by calculating the difference in per capita consumption resulting from the change in rice prices.

These findings indicate that the poverty in rural and estate areas have been subject to an increase. As these sectors are relatively poor, they suffer from an increase in the cost of their consumption bundle, resulting in a rise in the poverty gap index. However, when focusing on the urban sector, a decline in poverty is witnessed. The urban sector benefited from a 2.02 percentage point drop in poverty while the rural and estate sectors experience a 0.22 and 15.47 percentage point increase in poverty respectively. This is explained by the rice consumption patterns of these households.

As stated by the World Bank (2010), the majority of households in Sri Lanka, including a large number of rural households are net rice buyers. The type of rice focused in this study is Kekulu rice, and as stated by the Department of Census and Statistics (2006/07),

households in rural sectors indicate the highest consumption of Kekulu rice, whereas urban households consume the lowest quantity of Kekulu rice. This may explain the trend seen in poverty. Consistent with these findings, a reduction in urban poverty is experienced, as urban areas consume the least amount of Kekulu rice. Urban sectors are not hurt by the increase in the rice prices as their consumption bundle may not be subject to a severe rise in cost. Unlike urban areas, rural and estate areas consume a higher portion of rice and may fall into poverty or deeper into poverty as they suffer from an increase in their daily

consumption costs. An unexpected result here is a very low percentage change in the overall poverty in Sri Lanka. This could be due to the suboptimal classification of households into sectors by the DCS, overestimating households in urban areas. Data from the HIES were used in this poverty analysis which explain this, as many districts are considered urban. Due

19

to this, the decrease of poverty in urban areas is overstated relative to the increase in rural and estate sectors, which may finally result in an extremely small/insignificant rise in poverty. Therefore, the degree of the poverty effect may not be entirely accurate as the data is not very precise in terms of per capita expenditure and the divisions of the sectors. These limitations will be discussed further in the following section. As a result, when analyzing the welfare and poverty impacts, the findings may be more precise for each sector as overall results may be subject to measurement errors.

When looking into the implications of the global food crisis on poverty, the focal commodity has great influence: the impacts on poverty vary with crop type (Lu and Yu, 2011). It has been made clear in this section that households may be net consumers of certain commodities and net producers of others, therefore changes in poverty are sensitive to which commodity has been considered. Although many households do produce rice and may enjoy an increase in farm incomes, they are also hit by an increase in the cost of their consumption bundle, as rice is the primary staple for a large majority of the population. Thus, this may outweigh the rise in farm income and result in negative implications on poverty when focusing on the commodity kekulu rice. However, it is important to remember that this trend may not be consistent across all food items.

Table 4

Change in the Poverty Gap Index P(1): 2006/07

Sector P(1) Before price change P(1) After price change P(1) Change

Sri Lanka 3.66% 3.66% 0.01

Sector

Urban 2.75% 0.73% -2.02

Rural 3.85% 4.06% 0.22

Estate 3.57% 19.04% 15.47

20

4.3 General Discussion

As presented above, the majority of households experience a reduction in welfare as they are net consumers of Kekulu rice. In addition, an increase in poverty on the households that spend most of their income on kekulu rice is seen, as they experience a rise in their expenses. Considering the consumption patterns of rice, such findings are consistent with the majority of past research. However, a few important points of discussion must be made with regard to the steps taken in this study. Due to the data restrictions experienced during the data process, there are some limitations that need to be addressed.

First, it has been assumed in this study that consumer and producer prices increase at the same rate. As mentioned previously, although this assumption is stronger in a country similar to Sri Lanka with a more integrated market, this may not be entirely true. There may be small differences among consumer and producer prices of Kekulu rice (resulting from travel costs etc) that have not been accounted for in this study. As a result, this assumption may be harmful, distorting the welfare effects and resulting in inaccurate results. Next, when allocating paddy production of the certain districts into urban, rural, and estate sectors a different classification, unlike the classification mentioned in the HIES were used. Because the classifications of the different sectors were misleading and unclear in the household survey presented by the DCS, this procedure was carried out. Therefore, the data used from the survey may not be representative of the three sectors as the classification used in the survey was based upon municipalities, a very vague, ambiguous representation. As most districts are governed by a municipal council, this classification for urban areas is not accurate. Therefore, it is important to account for possible misrepresentations when conducting further research as this may change the results.

It is clear that the underlying assumptions regarding prices and the divisions of the various sectors have been made. Being aware of such assumption is helpful when

conducting further research in this field. If more detailed data is compiled, the validity of the results can be improved upon. It is also important to note that the welfare model presented in the methodology of this paper may not be applicable to all countries. It is crucial to look into the cost of living of the country studied as welfare measures may change across

countries. If this measure were to be applied to a different country it may not be an accurate measure due to expenditure costs. For example, in some countries the state provides

21

different services that are not paid for (such as food stamps), which will not be taken into account by the welfare equation. Also, culturally, food plays different roles in different societies and higher expenditure on food may be a result of cultural differences. Some cultures may spend heavily on luxury items, which don’t necessarily make welfare worse off, although this may be suggested by the equation. Thus, the welfare effects resulting from food price changes may be over or under estimated. This shows the importance of carefully evaluating the suitability of applying the model to another country.

22

5. Conclusion

The present research demonstrates that there are many implications of the rise in Kekulu rice prices on Sri Lanka in the wake of the global food crisis. While analyzing the beginning of the crisis from July 2006 to June 2007 it can be seen that rising rice prices have had detrimental effects on welfare and poverty in Sri Lanka. A 52.8 percent increase in Kekulu rice prices during this time period has resulted in an increase in the consumption

expenditure of households across urban, rural, and estate sectors. The rise in rice prices has contributed to the reduction in welfare seen in rural and estate sectors. This reduction in welfare is explained by the consumption patterns of these households as they are

predominantly net consumers of rice. As kekulu prices increase, their consumption bundle becomes more expensive increasing their consumption expenditure, which in turn has resulted in a reduction in welfare. However, the opposite is seen in urban households as a rise in welfare is experienced. This is because they consume lower quantities of Kekulu rice and engage in a small amount of paddy production, which explains the different result. Therefore, urban household may not be negatively affected by the rise in rice prices

compared to rural and estate sectors. Lower middle income earners tend to be hit harder as a larger proportion of their income is spent on rice, negatively affecting their welfare, while middle income earners do not experience this fall in welfare as they tend to earn more and spend a lower share of income on food. However, the highest income decile experienced a reduction in welfare which was an unexpected result. Such a result may be due to the allocation of production into the separate sectors.

With the rise in rice prices, negative poverty implications were witnessed. An increase in rural and estate poverty resulting from the kekulu rice price increase may be explained by the increase in the cost of the consumption bundles as those of a lower income are hit the

hardest. However, as urban areas consume the lowest quantity of Kekulu rice, a reduction in poverty is seen. Such effects change with the commodity considered in the research.

Therefore, it is crucial to identify the consumption and production patterns first, as this varies across different commodities and different districts. Overall, to answer the research question, I conclude that welfare and poverty seem to be negatively affected from the rise in

kekulu rice prices from July 2006 to June 2007 as many households are hit by an increase in

23

Bibliography

Balagtas, J. V., Bhandari, H., Cabrera, E. R., Mohanty, S., & Hossain, M. (2014). Did the commodity price spike increase rural poverty? Evidence from a long‐run panel in Bangladesh. Agricultural Economics, 45(3), 303-312.

Boianovsky, M. (2014). Robertson and the Cambridge approach to utility and welfare.

Cambridge Journal of Economics, 38(4), 961-985.

Braun, J. V., Ahmed, A., Asenso-Okyere, K., ShengGen, F., Gulati, A., Hoddinott, J., ... & Rheenen, T. V. (2008). High food prices: The what, who, and how of proposed policy actions. IFPRI-Policy Brief, (May).

Clapp, J., & Helleiner, E. (2012). Troubled futures? The global food crisis and the politics of agricultural derivatives regulation. Review of International Political Economy, 19(2), 181-207.

Clark, C. (1951). Urban population densities. Journal of the Royal Statistical Society. Series A

(General), 114(4), 490-496.

Cudjoe, G., Breisinger, C., & Diao, X. (2010). Local impacts of a global crisis: Food price transmission, consumer welfare and poverty in Ghana. Food Policy, 35(4), 294-302. Deaton, A. (1989). Rice prices and income distribution in Thailand: A non-parametric

analysis. The Economic Journal, 99(395), 1-37.

De Hoyos, R. E., & Medvedev, D. (2009). Poverty effects of higher food prices: A global perspective. World Bank Policy Research Working Paper Series, Vol.

Department of Agriculture. (2006). RICE. Retrieved from

http://www.doa.gov.lk/index.php/en/crop-recommendations/808

Department of Census and Statistics. (2006/07). Household Income and Expenditure Survey

2006/07. Retrieved from

http://www.statistics.gov.lk/HIES/HIES2006_07Website/Publications/HIES200 07Final%20ReportWeb%20.pdf

Dimova, R., & Gbakou, M. (2013). The global food crisis: disaster, opportunity or non-event? Household level evidence from Côte d’Ivoire. World Development, 46, 185-196. Ellwood, D. T., & Summers, L. H. (1985). Poverty in America: Is welfare the answer or the

24

Estrades, C., & Terra, M. I. (2012). Commodity prices, trade, and poverty in Uruguay. Food

Policy, 37(1), 58-66.

Foster, J., Greer, J., & Thorbecke, E. (1984). A class of decomposable poverty measures.

Econometrica: Journal of the Econometric Society, 761-766.

Headey, D. (2014). Food prices and poverty reduction in the long run.

Karwala, S. (2005). Changing Consumption Patterns.[WWW document]. URL

http://mentoring. com. pl/Changing_consumption_patterns. pdf (accessed 10 June 2010).

Korale Gedara, P. M., Ratnasiri, S., & Bandara, J. S. (2015). Does asymmetry in price transmission exist in the rice market in Sri Lanka?. Applied Economics, 1-15.

Lu, K., & Yu, B. (2011). The impact of high food prices on poverty in China. Development

in Practice, 21(4-5), 679-690.

Naranpanawa, A., & Bandara, J. S. (2012). Poverty and growth impacts of high oil prices: Evidence from Sri Lanka. Energy policy, 45, 102-111.

Robles, M., & Keefe, M. (2011). The effects of changing food prices on welfare and poverty in Guatemala. Development in Practice, 21(4-5), 578-589.

Rocchi, B., Romano, D., & Hamza, R. (2013). Agriculture reform and food crisis in Syria: Impacts on poverty and inequality. Food Policy, 43, 190-203.

Sri Lanka Disaster Knowledge Network. (n.d.). Retrieved from http://www.saarc sadkn.org/countries/srilanka/hazard_profile.aspx

Swan, S. H., Hadley, S., & Cichon, B. (2010). Crisis behind closed doors: Global food crisis and local hunger. Journal of Agrarian Change, 10(1), 107-118.

The World Bank. (2010). Food Price Increases in South Asia National Responses and Regional

Dimensions. Retrieved from

http://siteresources.worldbank.org/SOUTHASIAEXT/Resources/223546 1269620455636/6907265

1287693474030/South_Asia_Regional_Food_Prices_Final.pdf

United Nations Conference on Trade and Development (UNCTAD) (2008) ‘Tackling the Global Food Crisis’, Policy Brief 2, available at

www.unctad.org/en/docs/presspb20081_en.pdf.

Valero‐Gil, J. N., & Valero, M. (2008). The effects of rising food prices on poverty in Mexico. Agricultural Economics, 39(s1), 485-496.

25

Von Braun, J., & Torero, M. (2009). Exploring the price spike. Choices, 24(1), 16-21. Vu, L., & Glewwe, P. (2011). Impacts of rising food prices on poverty and welfare in

Vietnam. Journal of Agricultural and Resource Economics, 14-27.

Vu, L. H. (2008). Essays on the Economics of Food Consumption and Production in Vietnam. Unpub. Ph. D. diss., Dept. of Appl. Econ., University of Minnesota.

Warr, P., & Yusuf, A. A. (2009). International food prices and poverty in Indonesia. Australian national university (ANU). College of Asia and the Pacific. Arndt-Corden division of economics.

Wodon, Q., & Zaman, H. (2010). Higher food prices in Sub-Saharan Africa: Poverty impact and policy responses. The World Bank Research Observer, 25(1), 157-176.

26

Appendix

Household Calorie Consumption and Consumption Expenditure

Sector Daily Kcal for poor New Household Consumption

Expenditure Sri Lanka 1696 1695.129661 Sector Urban 1316 1990.597759 Rural 1686 1655.318853 Estate 1984 904.4505495

Source: Authors’ calculations with data from HIES 2006/07

Appendix B

Division of Districts into Urban, Rural, and Estate sectors, Underlying the Production Analysis

Urban Rural Estate

Colombo Nuwara Eliya Kurunegala

Gampaha Moneragala Badulla

Galle Anuradhapura Matale

Polonnaruwa Kandy Kalutara Matara Hambantota Batticaloa Ampara Puttlam Ratnapura Kegalle Housing Units 6660 11250 3880

27

Appendix C

Compensating Variation Inputs

Sector, Decile Group Total Expenditure (Rs) Consumption Expenditure on Rice (Rs) Household Production (Kg) Budget Share (PP*Y)/X Sri Lanka 70170 1425 288.436915 0.0203078 0.06219254 Sector Urban 35274 420 4.73451944 0.0119067 0.00203076 Rural 21440 570 231.3386 0.0265858 0.16325340 Estate 13456 435 52.3637959 0.0323275 0.05887813 Quintile Quintile 1 51239 577 78.143565 0.011260 0.02307445 Quintile 2 198168 569 127.355438 0.002871 0.00972350 Quintile 3 91355 597 9.13612588 0.006534 0.00151310 Quintile 4 58213 581 73.8017857 0.009980 0.01918164