1. Introduction
As pointed out by [
1], smallholder value chains are exposed to a number of shocks. On the supply side, production may be affected by climatic events that leaves smallholders with little to sell or having to choose between selling their produce or using it for their own consumption. On the demand side, an increase in the price faced by final consumers reduces the demand for the upstream product. Additional shocks may come from sudden changes in government policy and may affect both supply and demand. Ref. [
2] tells the story of sorghum beer in Kenya, which illustrates the effect that government policy may have on the development of smallholder value chains. In 2013, the Government of Kenya increased the tax revenue by imposing an excise duty of 50 percent on sorghum beer, which resulted in economic losses for smallholders, the brewery industry, and for the government itself, as it effectively obliterated the value chain. In 2015, the government reversed the policy decision and reduced the excise duty to 10 percent.
A similar story can be told for the case of Malawi, although regarding the value added tax, which is a consumption tax placed on a product whenever value is added at each stage of the supply chain, from production to the point of sale. The amount of VAT that the user pays is on the cost of the product, less any of the costs of materials used in the product that have already been taxed [
3]. VAT introduction was short lived. Dairy (farming and processing) is a key investment sector for the Government of Malawi [
4], and donors have also been committed to the development of the supply chain [
5,
6,
7]. Several value chain studies have discussed strategies to improve the dairy sector performance and raise its contribution to poverty alleviation and food security [
5,
6,
7,
8,
9].
On 1 October 2016, the Malawi Revenue Authority announced that domestic pasteurised milk, which had been exempted from VAT previously, was going to be taxed at the standard rate of 16.5 percent (the measure excluded infant milk). This was the third time that the Government of Malawi imposed VAT to milk, after two earlier similarly fleeting attempts.
The tax was announced in July 2016 and published in October 2016 but was not implemented until November of the same year due to complaints against it. Note that the introduction of the VAT was not only on milk but on almost all products previously not taxed, and thus the probability that the Government of Malawi was going to make an exception for milk was low. The VAT added 90 Kwachas to the price per litre of milk sold at the retail level. The introduction of VAT on domestically produced pasteurised milk increased the retail price and further decreased the demand during a period of low seasonal demand. In addition, processors cut the farmers’ price by 9 percent to reduce the increase in the price of the product at the retail level. The measure was met with strong opposition and thus, was short lived i.e., eliminated in May 2017.
The purpose of this paper is to analyse the incidence of such a tax along the domestic dairy supply chain. This provides unique information about the functioning of the Malawian dairy supply chain, in particular its sensitivity to government policy and any related changes in the demand. The purpose of this paper is to use the brief VAT episode to extract potential lessons for the domestic dairy industry.
This paper responds to the [
10] question on whether taxes levied on goods are completely shifted to consumer prices, or if the incidence also falls on firms. The question is important for both policy and scientific research, as consumption taxes make up a large part of the fiscal revenue in most developed and developing countries and may impact fairness and consumers’ welfare (e.g., [
11]). The typical assumption is that consumer prices fully reflect taxes; however, according to [
12], the evidence supporting this is limited.
In our view, the topic is important because it shows how taxation in developing countries can affect an inclusive supply chain such as the dairy one, the welfare of poor farmers, and government’s plans to foster growth in the agri-food sector. In this sense, by introducing the VAT, the government led to a contraction in the demand for domestic milk, which had implications both upstream the supply chain and for the consumers. An inclusive business model is a commercially viable model that benefits low-income communities by including them in a company’s value chain on the demand side as clients and consumers, and/or on the supply side as producers, entrepreneurs, or employees in a sustainable way [
13,
14]. This highlights the vulnerability of supply chains in developing countries to economic policy and the extent to which this may affect their economic and social sustainability.
Following a brief literature review focused mostly on the implementation of VAT in developing countries and particularly in Africa, the paper describes the situation of the Malawian dairy sector pre- and post-introduction of the tax and presents the multimarket model used to simulate the impact of the rise in the VAT, data used, and model calibration details. The implications of the findings are discussed in the context of the ongoing debate around taxation policies for agri-food value chains in developing countries.
2. Literature Review
This section provides a brief overview of the taxation literature related to the introduction and implementation of VAT in developing countries, Africa in particular, and those issues that pertain to this paper, i.e., effects of taxation of the final product of an agri-food supply chain such as the tax incidence.
According to [
15], since its introduction in France in the 1950s, VAT has been adopted in more than 160 countries, including a majority of low- and middle-income countries (LMICs). Its purpose is to tax final consumption but leave intermediate transactions between VAT-registered firms untaxed in order to avoid distorting production and supply-chain decisions.
VAT is an important source of revenue for countries. According to the Organisation for Economic Co-operation and Development Centre for Tax Policy and Administration and the African Union Commission [
16], tax revenues increased by 1.5 percent of the gross domestic product (GDP) on average between 2008 and 2017 as a result of increases in revenues from VAT (0.7 percentage points) and from taxes on income and profits (0.3 percentage points) driven entirely by a 0.7 percentage points increase in revenues from personal income taxes. The increase in the African (based on 26 countries) average tax-to-GDP ratio was higher than the increase for Latin America and the OECD averages (both 1.3 percentage points). The African average has plateaued at 17.2 percent since 2015, with increases in tax-to-GDP ratios in some countries being offset by decreases in others. The greatest source of tax revenues among the countries featured in this publication were taxes on goods and services, which accounted for 53.7 percent of the total tax revenues on average in 2017, with VAT alone contributing 29.4 percent.
Ref. [
17] provides a recent overview of the VAT situation in Africa (see also [
18,
19]). According to him, two kinds of VAT were introduced on the African continent. The Southern African Customs Union (SACU) countries (i.e., Botswana, Eswatini, Lesotho, Namibia, and South Africa) modelled their VAT, at the initiative of South Africa, on the New Zealand good and service tax (GST), which has few exemptions and is essentially levied at a single rate (by contrast, SACU countries zero-rate a limited number of unprocessed foodstuffs and kerosene). The design of the threshold level of turnover above which businesses are required to register and charge VAT on their output and entitled to reclaim tax on their inputs invoiced by other registered businesses, is one of the most important issues under VAT [
20]. Too high a threshold may be costly in terms of revenue and unduly distort business organization around the threshold. Too low a threshold may involve high administration costs for the VAT department and disproportionately high compliance costs for small taxpayers. As a result, the VATs in Southern Africa are exceptionally revenue productive. Kenya also moved closer to the New Zealand model.
The VATs in all other African countries are based on the EU VAT taxation, be it the UK, France, or Portugal version, although with some variability. These VAT versions have more exemptions and differentiated rates than the New Zealand GST and, therefore are more complicated. Southern African Development Community (SADC, Gaborone, Botswana) member countries (outside SACU and with the exception of DR Congo, Madagascar, and Mozambique), the East African Community (EAC) countries (except Burundi and Rwanda), the West African Monetary Zone (WAMZ) countries (except Cabo Verde and Guinea), and the members of the Intergovernmental Authority on Development (IGAD) based their VATs mainly on the UK legislation (Malawi is one of these cases). Like the UK, most of these countries apply a zero rate to various essential products, which is not permitted under the Common Directive. Most other countries follow the French version of VAT but, according to [
17], this has a civil law flavour different from the UK VAT. Cabo Verde, Guinea–Bissau, and Mozambique have a Portuguese type of VAT.
According to [
17], the standard VAT rate ranges from 10 percent in Djibouti to 20 percent in Madagascar, Morocco, and Tunisia. The standard rate of most VATs is 15 percent or higher. The lower rates are found predominantly in Southern Africa, where VATs are particularly broad-based. The zero rate, mainly applied to some unprocessed foodstuffs, is generally confined to countries with Anglo-Saxon taxing traditions. Countries with French taxing traditions do not have a zero rate on domestically produced or consumed goods.
The introduction of the VAT in Malawi was triggered by a deterioration in the country’s economic situation. In the early 1980s, the flow of external funds to Malawi decreased significantly, which coincided with the loss of its primary foreign trade route (between 80 to 90 percent of exports and imports) due to the closure of rail lines in Mozambique. These shocks resulted in a sharp increase in the servicing of Malawi’s external debt and defence spending, thereby creating a pressing need for more revenue. At first, the Government raised the rates on those tax bases that were, administratively, the easiest to tax, such as trade. By 1985, the tax to GDP ratio had increased by almost 50 percent. However, it was increasingly apparent that the ad hoc, temporary measures were inconsistent with the creation of a more liberal economic environment in the long run, which led to a reform of the tax system [
21]. A comprehensive tax reform was carried out in the second half of the 1980s with assistance from the World Bank, with the VAT being formally introduced in Malawi in 1989 [
22] under the name of “surtax.” The registration threshold (minimum annual turnover) is 10 million Malawi Kwachas (approximately US
$13,500) and the standard rate is 16.5 percent. The supply of some goods and services is exempt from VAT. However, all imports are generally subject to VAT except for goods used in agricultural trade or exempt from duty.
In economics, tax incidence or tax burden is the effect of a particular tax on the distribution of economic welfare. The tax burden represents the true economic weight of the tax, measured by the difference between real incomes or utilities before and after imposing the tax. As pointed out by [
23], the equity and distributional effects of VAT and its potentially distorting economic effects are always a matter of concern (see also [
24,
25,
26,
27,
28]). Reality, however, indicates that most countries, particularly developing countries, cannot finance the education, health, and infrastructure development they need to sustain growth without recourse to some form of general consumption tax. This is similar to the views expressed by [
29] who argue from a policy viewpoint that net fiscal incidence is the relevant equity measure that government authorities need to use in judging particular policies. For example, an increase in VAT may be rejected on equity grounds as being regressive but may be desirable from an equity standpoint if the resulting revenues are used to finance, for instance, primary-school services in poor neighbourhoods. Taxes can be progressive but, if the transfers to the poor are not sufficiently large, they may exacerbate poverty.
As it is well known in public finance [
30], key for the incidence of a tax such as the VAT are both the price elasticity of demand and price elasticity of supply. General rule claims that the steeper the demand curve and the flatter the supply curve, the more of the tax will be borne by consumers; and the flatter the demand curve and the steeper the supply curve, the more of the tax will be borne by producers [
30]. Ref. [
10] analysed the effects of a decrease in the VAT on food in Norway. Her findings suggest that taxes levied on food are completely shifted to consumer prices, with only a few spill-over effects to most other goods. Moreover, she found that lowering the VAT on food attenuates inequality in consumer welfare, in part because households adjust their spending patterns in response to the price change. It should be noted that [
31], indicated that in contrast to the standard incidence model—where the direction of a tax change does not matter for incidence, as supply and demand elasticities are sufficient to determine the proportion of the tax borne by each agent—it is possible for the response to prices to be asymmetric to variation in VATs showing that there is a consistently higher pass-through to prices for tax increases than for tax decreases. Ref. [
32], using monthly observations on consumer prices and VAT rates for around 70 commodity groups in 17 Eurozone countries over 1999–2013 studied the extent of pass through. They found the impact on consumer prices to vary systematically and sharply across the different types of VAT reform, with full pass confirmed for changes in the standard rate, but pass-through for reduced rates to be lower. Ref. [
33] analysed the incidence of VAT in Mexico following two tax reforms that increased the VAT rate for a group of cities and left the rest unaffected. They compared the inflation rate in the affected cities with that in the exempted cities before and after the law changed, found that the effect on prices was limited and concluded that the burden of the tax was shared between producers and consumers. Ref. [
34] studied the introduction of a 3 percent VAT for retailers against the existing normal rate of 17.5 percent in Ghana and concluded that the incidence of VAT is on the consumers. Ref. [
35] studied the introduction of VAT in South Africa using a computable general equilibrium (CGE) model and found that it negatively impacted the welfare of low-income households. Ref. [
36], used a CGE model to study the increase by 1 percent of the VAT in South Africa, and found that it affected not only the cost of living (i.e., consumers) but also the short-run cost to the employer.
The literature has focused on significant tax reforms in developed or developing countries applying to large groups of goods (e.g., food) or all goods and services, in contrast to changes in the VAT rate for a final good, which is the focus of this paper. This allows concentrating the analysis in the context of a particular market and highlighting some of the particularities of the supply chains (e.g., [
37]), such as the presence of an informal market. In this sense, the present paper aims to contribute to the VAT literature not only by measuring the incidence of the tax but the effects that it brings to the supply chain.
3. The Dairy Sector in Malawi and the VAT
Figure 1 is useful not only for presenting the major characteristics of the dairy sector in Malawi but also for providing an overview of its evolution [
7]. The dairy market is characterised by the coexistence of formal and informal trading of milk. The formal sector markets pasteurised milk (amongst other products), whilst the informal sector trades raw milk.
Figure 1 provides an indication of the size of both markets, which, as shown by the evolution, is not constant.
The red line in
Figure 1 represents the total milk produced in Malawi according to the Agriculture Production Estimates Survey (this includes milk from the dairy and beef herds). The blue line in
Figure 1 indicates the collection of milk by milk bulking groups (MBGs), i.e., the milk that is pasteurised, packaged, and sold by retailers. The level difference between the red and blue lines represents milk that remains outside the formal market and is either consumed on farms, wasted, or sold by informal vendors.
The dotted line in
Figure 1 (measured on the right axis) represents the proportion of the total milk production outside the formal sector.
Figure 1 indicates that a sizable part of the milk supply is consumed unpasteurised and without any quality control (between 70–80 percent at the end of period recorded in
Figure 1).
In terms of international trade, Malawi imported US $735,522 worth of products under the United Nations (UN, New York, NY, USA) Comtrade’s 04.01 “Milk and cream, not concentrated nor containing added sugar or other sweetening matter” (714,731 kg) in 2015, of which 89.9 percent were exempted from tariffs due to the fact that they originated from countries members of the Common Market for Eastern and Southern Africa (COMESA) or SADC (Gaborone, Botswana). Among these, the most important trading partner is South Africa, which represents the source of 78.9 percent of the Malawian imports.
There is limited published information on the food consumption in Malawi, and dairy is not an exception. Available data are either estimates based on the apparent consumption (i.e., production plus imports plus net decrease in stocks minus exports) or sourced from the Living Standards Measurement Surveys [
38]. Malawi has the lowest consumption of milk per capita among African countries, which was estimated at 4.7 kg/capita/year compared to an African average of 15 kg/capita/year [
6,
38]. The demand for dairy products in Malawi includes mostly fresh milk, followed by powdered milk (which is imported for the more affluent population), margarine, butter, chambiko (a type of liquid yoghurt), yoghurt, and cheese.
The introduction in 2016 of a standard VAT rate of 16.5 percent for domestic pasteurised milk (excluding infant milk) previously exempted not only the milk produced domestically but also remaining imports of milk under the customs code 04.01 (UN Comtrade database). The heading 4.1 includes the subheadings 4.1.1 (fat content not exceeding 1 percent), 4.1.2 (fat content exceeding 1 percent but not exceeding 6 percent), 4.1.4 (fat content exceeding 6 percent but not exceeding 10 percent) and 4.1.5 (fat content exceeding 10 percent). All subheadings are taxed with a custom tariff of 30 percent (e.g., from the European Union) with some exemptions. Namely, for imports from COMESA, the rate is 3 percent, while imports from SADC, either from South Africa or other countries, are tax free. In addition, there is a withholding tax of 3 percent.
Hence, with the implementation of VAT in November 2016 for both imported and domestic products, milk products sold at the retail level were all subject to the VAT rate of 16.5 percent, which implicitly led to an increase in the retail price and a further decrease in demand during what is typically the lowest sales period in the year (January). Moreover, in February 2017, processors reduced milk price by 9 percent, i.e., from 170 Kwachas (US$ 0.23) to 155 Kwachas (US$ 0.21) in order to reduce the increase in the retail price. After discussions between the government, milk processors and other stakeholders, the tax was eliminated in May 2017.
Figure 2 shows the milk collection of the MBGs belonging to Shire Highlands Milk Producers Association (SHMPA). Despite the seasonality in the milk collection, the contraction in the collection of milk following implementation of VAT is clear. When comparing the months January and February 2017 against the same period the previous year, the decrease in milk collection was 7.3 percent.
Figure 3 shows the evolution in nominal terms of the prices paid by processors and the prices received by farmers. Note that processors pay for their milk collection to the MBGs and the MBGs pay the farmers. The difference between prices represents the administrative costs of the MBGs. The figure shows the decrease in the price paid by processors in February 2017, which was not yet passed to farmers by most MBGs. This situation, however, is only sustainable in the short term and only as far as the MBGs have the capacity to incur temporary losses.
The following section presents the data and model used to analyse the effects of the introduction of the VAT to dairy products.
6. Discussion
This section discusses the implications of the VAT introduction in terms of its incidence and effects on the agri-food value chains.
Table 9 presents the results of the VAT incidence analysis. These indicate that consumers are expected to bear about 24.3 percent of the tax collection, whilst the industry bears the remaining 75.7 percent. These results contrast with evidence that indicates that the VAT is mostly borne by consumers (e.g., [
34]). However, this is not surprising as the main consumers of domestic milk are poor urban consumers who are very sensitive to changes in prices. An increase in the price of a good of the magnitude of the VAT rate analysed here reduces substantively its consumption, subsequently forcing the price increase to be shared among actors upstream the supply chain. This is a negative outcome for both consumers and the supply chain as the former reduce their consumption of a staple food product and the latter faces a reduction in production with consequences in terms of its economic and social sustainability.
Table 9 also presents a calculation of the inefficiency created by the introduction of the VAT (i.e., the economic benefit foregone due to the tax and thus not reaching either consumers, producers, or the government). This was estimated to be 0.2 percent of the total milk sales before VAT.
As pointed out by [
1], uncertainty is present in smallholder value chains and it relates to a number of interrelated areas such as the insecure nature of agricultural production; market dynamics, with shifts in supply and demand patterns linked to factors such as demographic change; and technological innovation. One of the most significant is the economic environment in which these value chains operate.
The introduction of the VAT provides not only an opportunity to better understand the functioning of the dairy supply chain in Malawi and to observe its vulnerability to changes in the policy framework, but also to extract lessons for similar value chains. Changes in taxation and other policies are a common occurrence, particularly in developing countries where macroeconomic pressures are an important source of uncertainty for economic agents.
An important debate in development economics is how to assist smallholder farmers in moving forward and to facilitate their connection to the market in order to reach sustainable growth (e.g., [
11]). Value chains are a mechanism to achieve this. This type of development approach has been named as a ‘‘demand-side” approach where private agents (entrepreneurs, producer organisations) create incentives for smallholder farmers to modernise through contracting and vertical coordination in value chains [
48].
According to [
48] “a demand-side approach consists in creating incentives for smallholder farmers to modernise through their participation in vertically coordinated value chains that provide links to markets for products with a profitable effective demand, while at the same time potentially offering solutions to market and institutional failures” (p. 9). The advantage of a demand-side approach, in contrast to a supply side one (e.g., aiming to remove barriers in production), is that it seeks broad complementarities to achieve modernisation, specific to the agent in question.
This study clearly points out that changes in taxation policy can be damaging to smallholder supply chains, common in developing countries, and can have significant economic and social sustainability effects as these chains are embedded into rural communities. The analysis raises two related questions, whether these chains can be taxed and when. The answer to both questions is not straightforward, as a solution to how and for how long to support an “infant” industry in developing countries is still under debate [
49]. This needs to be evaluated by carefully taking into account the reaction of a specific chain. As pointed out by [
2], in many cases tax decisions are taken as a leap in the dark by authorities, which creates additional uncertainties to the value chain.
7. Conclusions
The dairy sector in Malawi can be considered an example of an inclusive supply chain where smallholder farmers are integrated into the business and the final product is sold domestically. This paper investigated the effects that the short-lived VAT would have had on the different stages of the domestic dairy supply chain, as well as the incidence of the tax. Some conclusions emerged:
The government tax revenues would be borne by both consumers (24.3 percent of VAT revenues) and the domestic dairy industry (75.7 percent of VAT revenues). This is mainly due to the fact that consumers’ demand for pasteurised milk is very sensitive to changes in prices and cannot absorb the full increase in price due to VAT without contracting significantly. This forced the industry to reduce prices along the supply chain up to retail and shrink production to avoid an unsustainable accumulation of stocks.
The fact that most of the VAT revenues were paid by the industry means that the measure affected not only processors, but also the farmers. This is particularly important because milk is an important source of income for farmers.
The consumption results indicate that the introduction of VAT will not significantly affect the per capita consumption of milk (for all the population), which would only suffer a minor contraction of 0.2 percent due to the reduction in the demand for imported milk. The domestic demand for pasteurised milk (most of which linked to poor urban households) would be expected to decrease by 3 percent. The quantity of milk farmers sold on the informal market would be expected to increase by about 1.6 percent due to the reduction in milk prices paid to farmers on the formal market.
The tax measure hit a sector already greatly affected by infrastructure issues such as electric power cuts, factory breakdowns, and impassable access roads, which, despite the aforementioned challenges, has been a good source of investment in Malawi. Moreover, it has been supported by donors which recognised its potential to increase the income of poor farmers, particularly those in the South of Malawi lacking economic choices.
The implications of the results for the dairy industry provide lessons for international development. The international development community are increasingly interested in stimulating agricultural transformation through improved productivity, increased market access for smallholder farmers and investment in agribusiness [
50]. Moreover, they seek to increase economic opportunities for smallholders by demonstrating the commercial viability of businesses with significant smallholder supply chains and attracting more investment into the sector.
Under the guiding idea that economic growth is most effective in reducing poverty when it occurs in sectors that are important to the poor as a source of income or employment [
51], there has been interest in encouraging private sector investment in inclusive innovative business models, particularly in the manufacturing and agriculture sectors.