The effects of reducing food losses and food waste on global food insecurity, natural resources, and greenhouse gas emissions

Abstract

The Food and Agriculture Organization of the United Nations (FAO) has estimated that in 2010–2012, 868 million people were undernourished worldwide. At the same time, FAO reported that approximately 1.3 billion tons of food were lost or wasted globally in 2007, which was equivalent to approximately one-third of the food produced for human consumption at the time. Food losses and waste deprive the poor living in developing regions of opportunities to access food, cause significant depletion of resources such as land, water, and fossil fuels, and increase the greenhouse gas emissions associated with food production. In the present study, the effects of reducing food losses and waste on global food security, natural resources, and greenhouse gas emissions were evaluated using a food trade model operating under the assumption that in 2007, developed regions, as defined by the FAO, would reduce food losses and waste by up to 50 % during the stages of postharvest handling and storage, processing and packaging, distribution, and consumption. The results obtained show quantitatively that reductions in food losses in developed regions decrease the number of undernourished people in developing regions by up to 63 million, leading to decreases in the harvested area, water utilization, and greenhouse gas emissions associated with food production.

Appendices

Appendix 1: Simple method for estimating per capita income change

The per capita income for region r in the base year 2007, pcINC _r,base, is defined as follows:

$$ {\text{pcINC}}_{\text{r,base}} = {\text{GDE}}_{\text{r,base}} /{\text{POP}}_{\text{r, base}} $$

(1)

where GDE_r,base is the GDE and POP_r,base is the total population in the base year.

After the reduction of food losses and food waste in developed regions, per capita income in developing regions, pcINC_r, is recalculated simply by revising the GDE_r as follows:

$$ {\text{pcINC}}_{\text{r}} = {\text{GDE}}_{\text{r}} /{\text{POP}}_{\text{r,base}} $$

(2)

$$ \begin{aligned} {\text{GDE}}_{\text{r}} & = \sum\limits_{i} {\left( {{\text{AgGDE}}_{{{\text{CN}}_{{i,{\text{base}}}} }}\, { \cdot }\,\frac{{{\text{PC}}_{i}\, { \cdot }\,{\text{QF}}_{i} }}{{{\text{PC}}_{{i,{\text{base}}}}\, { \cdot }\,{\text{QF}}_{{i,{\text{base}}}} }}} \right)} \\ & \quad + \sum\limits_{i} {\left( {{\text{AgGDE}}_{{{\text{EX}}_{{i,{\text{base}}}} }}\, { \cdot }\,\frac{{{\text{PW}}_{i}\, { \cdot }\,{\text{QEX}}_{i} }}{{{\text{PW}}_{{i,{\text{base}}}}\, { \cdot }\,{\text{QEX}}_{{i,{\text{base}}}} }}} \right)} \\ & \quad - \sum\limits_{i} {\left( {{\text{AgGDE}}_{{{\text{IM}}_{{i,{\text{base}}}} }}\, { \cdot }\,\frac{{{\text{PW}}_{i}\, { \cdot }\,{\text{QIM}}_{i} }}{{{\text{PW}}_{{i,{\text{base}}}}\, { \cdot }\,{\text{QIM}}_{{i,{\text{base}}}} }}} \right)} + {\text{RtGDE}}_{\text{r,base}} \\ \end{aligned} $$

(3)

where PC _i and PW _i are consumer and world prices; QF _i , QEX _i and QIM _i are consumer demand, exports and imports, respectively, for food commodity i after the reduction of food losses and food waste in developed regions; and AgGDE_CN _i,base, AgGDE_EX _i,base, and AgGDE_IM _i,base are agricultural consumption, exports, and imports, respectively, in the base year. RtGDE_r,base is determined by subtracting these agricultural terms from gross domestic expenditure, GDE_r,base values taken from the GTAP8 database (Narayanan et al. 2012).

Although the expenditure for the other goods (RtGDE_r) is assumed to be constant at the base year level in the formula (3), it could be reduced due to food wastage reduction in developed regions for the following two relations. First, RtGDE_r could be reduced due to the price decrease of food considering substitution between food and the other goods. Generally, the more substitutable with other goods, the more elastic to the price the demand of good tends to be. As is shown in Table 13, food demand is less elastic to the price than the other goods are, which means food has little substitution goods. Therefore, RtGDE_r might be reduced slightly due to the decreases of food prices in developing regions. Secondly, RtGDE_r could be reduced due to agricultural income reduction arising from the decreases of food price and food production in developing regions because the other goods are more elastic to income than food is (Table 13). Consequently, GDEs (GDPs) in developing regions could be reduced further due to the decrease of RtGDE_r, which leads to the increase in the number of undernourished people for developing regions.

Table 13 Income and price elasticities for aggregate consumption categories

Appendix 2: The FAO methodology for assessing food insecurity: a skew-normal distributional model

In 2012, the FAO methodology for estimating the prevalence of undernourishment went through a review to choose the most appropriate model to describe the dietary energy consumption of the population and improve the estimation of its parameters. A skew-normal distribution (Azzalini 1985) was adopted as a result of this review (FAO, WFP, IFAD 2012). In the model, the proportion of undernourished individuals in the total population is defined by the following cumulative distribution function:

$$ F({\text{MDER}}) = \varPhi \left( {\frac{{{\text{MDER}} - \xi }}{\omega }} \right) - 2\,{\text{Owen}}T\left( {\frac{{{\text{MDER}} - \xi }}{\omega },\alpha } \right) $$

(4)

where MDER is the minimum dietary energy requirement (kcal/person/day) and ω, ξ, and α are scale, location, and shape parameters, respectively. Φ is the cumulative distribution function of the standard normal distribution and OwenT is Owen’s T function. These two parameters are defined as follows:

$$ \varPhi (x) = \frac{1}{2}\left[ {1 + {\text{erf}}\left( {\frac{x}{\sqrt 2 }} \right)} \right]( - \infty < x < + \infty ) $$

(5)

$$ {\text{Owen}}T(h,a) = \frac{1}{2\pi }\int\limits_{0}^{a} {\frac{{{\text{e}}^{{ - \frac{1}{2}h^{2} \left( {1 + x^{2} } \right)}} }}{{1 + x^{2} }}} {\text{d}}x( - \infty < h,a < + \infty ) $$

(6)

where erf is the error function.

The values of the scale, location, and shape parameters are derived from the mean, variance, coefficient of variation (CV), and skewness of the distribution and the food loss rate (Loss), as follows:

$$ {\text{Mean}} = {\text{DES}}{ \cdot }\left( {1 - {\text{Loss}}} \right) = \xi + \omega \delta \sqrt {\frac{2}{\pi }\;}\, {\text{where}}\;\delta = \frac{\alpha }{{\sqrt {1 + \alpha^{2} } }} $$

(7)

$$ {\text{Variance}} = \omega^{2} (1 - (2\delta^{2} )/\pi ) $$

(8)

$$ {\text{CV}} = \frac{{\sqrt {\text{Variance}} }}{\text{Mean}} $$

(9)

$$ {\text{Skewness}} = \frac{4 - \pi }{2}{ \cdot }\frac{{\left( {\delta \sqrt {2/\pi } } \right)^{3} }}{{(1 - 2\delta^{2} /\pi )^{3/2} }} $$

(10)

where DES is the average dietary energy supply (kcal/person/day), which is computed from the per capita consumption of food commodities (kg/person/year) as the outcome of the modified PEATSim model. Finally, the number of undernourished people is determined by multiplying the proportion of undernourished people by the total population.

Appendix 3: How to calculate the reduction target for food losses and food waste in developed regions

In the modified PEATSim model, the regional food balance for each commodity in 2007 is expressed as follows:

$$ {\text{QS}}0 + {\text{QIM}}0 - {\text{QEX}}0 + {\text{QSt}}0 = {\text{QL}}0 + {\text{QR}}0 + {\text{QP}}0 + {\text{QE}}0 + {\text{QF}}0 $$

(11)

where QS0, QIM0, QEX0, and QSt0 on the left-hand side are production, import, export, and stock variation, respectively, and QL0, QR0, QE0, QP0, and QF0 on the right-hand side are feed, seed, other utility, processing, and consumer demand, respectively, derived from the food balance sheets in FAOSTAT (FAO 2012).

In the following formulas for calculating food wastage, QW0 represents food losses in the postharvest handling and storage stages in 2007, derived from “waste” in the commodity balances in FAOSTAT, which is included in the other utility (QE0) in the food balance sheets. QE, QP, and QF are the other utility, processing, and consumer demand, respectively, in developed regions after reduction of food losses and food waste, which are fixed at the reduced levels during simulation processes. R _m, R _p, R _d, and R _c are the ratios of waste generation in the milling, processing and packaging, distribution, and consumption stages, respectively, in the SIK report (Gustavsson et al. 2013). R _r is the target reduction ratio for food losses and food waste for all food commodities, which is taken as the maximum of 0.5 in the present study. S _QE, S _QP, and S _QF are the maximum reduction quantities for other utility, processing, and consumer demand, respectively, in the model.

3.1 Cereals

Food losses and food waste are generated as follows at each stage of the food supply chain:

Postharvest handling and storage: QW0

Milling: R _m · QF0

Processing and packaging: R _p · QP0 + R _p · (1 − R _m) · QF0

Distribution: R _d · (1 − R _p) · QP0 + R _d · (1 − R _p) · (1 − R _m) · QF0

Consumption: R _c · (1 − R _d) · (1 − R _p) · QP0 + R _c · (1 − R _d) · (1 − R _p) · (1 − R _m) · QF0

Therefore, the maximum reduction quantity for each demand is calculated as follows:

Other utility (S _QE): QW0

Processing (S _QP): [R _p + R _d · (1 − R _p) + R _c · (1 − R _d) · (1 − R _p)] · QP0

Consumer (S _QF): [R _m + R _p · (1 − R _m) + R _d · (1 − R _p) · (1 − R _m) + R _c · (1 − R _d) · (1 − R _p) · (1 − R _m)] · QF0

3.2 Oilseed crops and vegetable oils

Food losses and food waste are generated as follows at each stage of the food supply chain:

Postharvest handling and storage: QW0_crop

Processing and packaging: R _p · QP0_oil + R _p/(1 − R _p) · QF0_oil

Distribution: R _d · (1 − R _p) · QP0_oil + R _d · [1 − R _p/(1 − R _p)] · QF0_oil + R _d · QF0_crop

Consumption: R _c · (1 − R _d) · (1 − R _p) · QP0_oil + R _c · (1 − R _d) · [1 − R _p/(1 − R _p)] · QF0_oil + R _c · (1 − R _d) · QF0_crop

Therefore, the maximum reduction quantity for each demand is calculated as follows:

Other utility (S _QE,crop): QW0_crop

Processing (S _QP,oil): [R _p + R _d · (1 − R _p) + R _c · (1 − R _d) · (1 − R _p)] · QP0_oil

Consumer (S _QF,oil): {R _p/(1 − R _p) + R _d · [1 − R _p/(1 − R _p)] + R _c · (1 − R _d) · [1 − R _p/(1 − R _p)]} · QF0_oil

Consumer (S _QF,crop): [R _d + R _c · (1 − R _d)] · QF0_crop

3.3 Livestock products

Food losses and food waste are generated as follows at each stage of the food supply chain:

Postharvest handling and storage: QW0

Processing and packaging: R _p · QP0 + R _p · QF0

Distribution: R _d · (1 − R _p) · QP0 + R _d · (1 − R _p) · QF0

Consumption: R _c · (1 − R _d) · (1 − R _p) · QP0 + R _c · (1 − R _d) · (1 − R _p) · QF0

Therefore, the maximum reduction quantity for each demand is calculated as follows:

Other utility (S _QE): QW0

Processing (S _QP): [R _p + R _d · (1 − R _p) + R _c · (1 − R _d) · (1 − R _p)] · QP0

Consumer (S _QF): [R _p + R _d · (1 − R _p) + R _c · (1 − R _d) · (1 − R _p)] · QF0

Consequently, other utility, processing, and consumer demands in developed regions after reducing food losses and food waste, namely, QE, QP, and QF, respectively, amount to (QE0 − R _r · S _QE), (QP0 − R _r · S _QP), and (QF0 − R _r · S _QF), respectively, which are fixed at these levels during simulation processes.