Literature Review

The linkages between economic reform and poverty, and the development of ways to quantitatively assess those linkages have been the subject of intense recent research.² Hertel and Reimer (2005) and Hertel and Winters (2005) reviewed ex ante studies and provided a method of classification by simulation type: partial equilibrium models; general equilibrium models; and micro/macro simulation models, which combine (not always with feedback) macro-level simulation with micro-level household models. They concluded that CGE techniques and micro/macro methods have the best potential to fully evaluate the complex web of determinants of changes in poverty ex ante.

Within the recent CGE literature there is a range of ways of addressing poverty impacts of policy reform. Aggregate results at the regional level, such as impacts on prices of staples or returns to unskilled labor, may be combined with ad hoc observations on potential effects on poverty. Other studies take results from a global model with a single representative household, and pass them through a sub-model to determine the poverty impact. The sub-model may be small, producing rough assessments of poverty impacts in the form of headcount indices that are calculated using established elasticities, as in Anderson and Martin (2005). Another approach at the country level is to build more sophisticated sub-models of household behavior, or to directly incorporate income distribution at the household level within a regional CGE model. Several studies of this type have been conducted recently for selected economies (Hertel and Winters 2006; OECD 2006).

The most straightforward method of dealing with income distribution and poverty impacts in a CGE framework is to abandon the single-representative consumer approach and to incorporate multiple representative household groups in the model. This allows the model to track inter-group income changes directly, but still leaves the issue of intra-group income changes external to the model. The greater the number of household groups, the better the model will be able to capture the pattern of income changes. Recent efforts typically feature between five and 10 household categories, although in some cases there are many more. The approach can be extended by incorporating any available information on the distribution of income within household categories.

Several recent CGE studies have used the multiple household approach to analyze the consequences of reform (mostly trade policy) in South Asia. Since the reduction in the cost of transportation has very similar theoretical implications to reductions in tax-based distortions, a review of the literature is worthwhile. These studies have tended to use single-economy models for the distribution analysis, often in combination with a global model to estimate the impacts of the proposed scenarios on world prices.

There are a number of studies of India. Gilbert (2007) considered the impact of the current proposed modalities for reform in agriculture under Doha at the household level in India, in addition to the effects of more comprehensive agricultural reform. The study used the Global Trade Analysis Project (GTAP) model to estimate the world market effects, after first modifying the underlying GTAP6 data to reflect the latest available applied protection levels, using the trade analysis and information system (TRAINS) database. The global results were then input into a single-economy CGE model of India, which is a competitive CGE of the Armington variety. Forty-three productive sectors and five factors of production were identified in that study, along with nine households (four rural and five urban with Stone-Geary preferences). Household data were obtained from Pradhan and Sahoo (2006) and matched to the GTAP data on aggregate consumption, production and trade using RAS methods. The base year was 2001, although protection data was more recent, as noted above. The simulations were run as comparative statics, with two different adjustment time horizons (short and long run) represented by mobility/immobility of capital across productive sectors. Tax replacement was (implicitly) through lump sum transfers from the households.

In the Doha scenarios, the welfare of the poorest households (agricultural labor and other rural labor) fell in both the short and long run, whereas the welfare of the richest group (urban self-employed) rose. The income of the rural self-employed (landowners) also rose in the Doha scenarios, suggesting that ownership of land and capital helps to insulate this group from the effects of trade shifts. The result was similar to that of Annabi et al. (2006) for Bangladesh, but the change was not statistically robust. Under comprehensive reform the results were quite different. The aggregate welfare gains were several orders of magnitude larger, and the welfare of all households except the rural self-employed rose. The results were statistically significant, and suggest that India's landowning class is able to benefit from rising world prices under Doha reform when India does not engage in significant reforms of its own, but faces considerable drops in income if domestic prices are allowed to fall.

The results also indicate that, in the long run, income inequality improves in all scenarios except for comprehensive reform of agriculture. India would gain overall from agricultural reform, but a small increase in rural poverty is possible under the Doha agreement as it stands. On the other hand, comprehensive reform is likely to increase the incomes of the poorest groups, but at the expense of a slight increase in income inequality, and a substantial reduction in the incomes of landowners.

Pradhan and Sahoo (2006) used a similar CGE structure in their analysis of potential trade reform scenarios for India, although it was not connected to a global CGE framework, and obtained similar results.

Panda and Ganesh-Kumar (2008) specifically considered the issue of food security with changes in trade policy. Their modeling approach was very similar to that used in Gilbert (2007), with the exception that they used the modeling international relationships in applied general equilibrium (MIRAGE) model developed by the International Food Policy Research Institute (IFPRI) as the source of their global price changes rather than GTAP. They considered a Doha scenario and found that all households experienced a rise in welfare and a decline in poverty. However, they argued that this does not necessarily translate into increased food security as the poorest households decreased their consumption of protein and calories, while increasing their consumption of fats. These conclusions are based on an ex post assessment of the household consumption patterns which drive the CGE model.

Finally, Polaski et al. (2008) used a single-economy CGE of India, based on a very detailed social accounting matrix (SAM) for 1998–1999. The SAM included 115 commodities, 48 labor types, and 352 households (classified by social group, income class, region, and urban or rural). The model was a competitive Armington type, and varied from the standard model only in that the labor market for unskilled workers was closed by fixing the wage and allowing the supply to be endogenous (to capture potential underemployment of unskilled labor). They considered the impact of price changes in agricultural commodities, and found that a decrease in the price of rice could have a significant negative impact on Indian poverty levels.

Results for Bangladesh are available from Annabi et al. (2006) and Raihan (2008). Annabi et al. (2006) used the GTAP model to estimate the overall effect of trade reform under the Doha proposals (both agriculture and non-agriculture) at the world level, and then input the world market effects into a single-economy CGE model for Bangladesh. The single-country model was used to generate detailed results at the household level. In addition to the Doha agenda, the study also considered the potential impact of more comprehensive global reform, and of unilateral reform by Bangladesh.

The Bangladesh model was a standard, competitive CGE model of the Armington type. It identified 15 productive sectors, four factors of production, and nine households (five rural and four urban with Stone-Geary preferences), with 2000 as the base year. The simulation procedure is recursive dynamic, setting the growth of the labor stock and productivity at fixed levels, and making the capital stock growth path endogenous by applying a simple, sector-specific investment rule. Tax replacement is (implicitly) through lump sum transfers from the households. The simulations extend for a 20-year period, and are compared with a baseline growth path.

The results indicated aggregate welfare losses for Bangladesh in the Doha scenarios, along with small increases in the headcount ratio (diminishing somewhat but remaining negative in the long run). The negative aggregate welfare effect was driven by adverse terms of trade movements. These remain even in a scenario with complete liberalization in the rest of the world. The poverty effect was driven by increased prices, even though nominal unskilled wages rose slightly. When broken down to the household level, Annabi et al. (2006) found poverty increases for all household categories except large farms.

The study by Raihan (2008) used a single-economy model for Bangladesh. The SAM was based on 2005, with 26 sectors, nine production factors, and six households. Again, the model used the competitive Armington structure to analyze various unilateral liberalization scenarios under both a neoclassical labor market closure and an unemployment closure, where it is assumed that only a proportion of the workforce is able to change production activities. Raihan (2008) argued that the effects of unilateral reform in the aggregate are positive but small, suggesting that the export bias of the current regime is minimal. Unfortunately, the paper did not directly discuss poverty or income distribution impacts, although the model was clearly able to generate information on this aspect of adjustment to trade reforms.

Two recent studies are available for Nepal, Cockburn (2002) and Acharya and Cohen (2008). Cockburn (2002) constructed a CGE model of Nepal that explicitly models all households from a nationally representative household survey. The model was an archetype competitive CGE, but with a high degree of household disaggregation. The base year of the model was 1986, and the model incorporated 15 production activities. They considered a trade liberalization scenario and found that urban poverty decreased and rural poverty increased, as initial tariffs were highest for agriculture. Impacts increased with income level, resulting in rising income inequality.

Acharya and Cohen (2008) based their work on a 1996 SAM of Nepal, with four household groups. The model was very small scale, identifying only four production activities and two factors. The scenario considered is a 10% reduction in tariffs in the presence of both a fixed and a flexible exchange rate regime. They concluded that the results were not conclusively pro-poor, as one of the richer household groups benefited most, whereas the benefit to the poorest household group was only modest. Therefore, they suggested that complementary policies are required to make trade liberalization pro-poor in Nepal.

Although Sri Lanka is not part of the SASEC region, it is an important economy in South Asia and has been the focus of several CGE studies, many surveyed by Dasanayake (2000).³ The most recent study was by Naranpanawa (2005), and used a 1995 SAM. The model identified 38 production activities, three factors, and five households. The poverty analysis was extended to intra-household groups (post simulation) by estimating the distributions of income within household groups and using these to estimate disaggregate poverty metrics. Naranpanawa considered a manufactured goods trade liberalization scenario, and found that the potential benefits accruing to low income rural groups were low relative to other groups in the model, a fact attributed to a reduction in transfer payments from the government to households following falls in government revenue. Long-term liberalization reduced absolute poverty in all groups, with manufacturing reform being more pro-poor than agricultural reform, but might increase relative poverty, i.e., would not support a more even distribution of income.

Pakistan, while also not a member of SASEC, has been the subject of several recent CGE studies. Ahmed and O'Donoghue (2008) used a single-economy CGE to examine various macro-economic shocks. The SAM identified 44 sectors, 11 factors and 17 household categories, with 2001 as a base year. Although based on the standard competitive Armington framework, several modifications of the model were made to better represent the developing economy characteristics of Pakistan. In particular, the model incorporated household consumption of non-market commodities as well as market commodities and an explicit treatment of transaction costs for marketed commodities. Unlike many other CGE models (such as GTAP), the model also separated production activities and commodities (i.e., joint production was allowed). The CGE model was combined with a microsimulation model to generate poverty information. Prices, wages, and aggregate employment variables from the CGE model were used as input to a microsimulation model that generated changes in individual wages, self-employment incomes, and employment status. The microsimulation model was based on household and individual level data from the survey data for the year 1996 and simulated income generation mechanisms for 1150 households.

The model was used to simulate several international price shocks, both aggregate and sector-specific. Among the most important results, Ahmed and O'Donoghue (2008) found that external oil price shocks have the highest potential for socioeconomic impact. Poverty increased when the overall import price rose, but increases in the import price of petroleum had the greatest effect on inequality.

All of the preceding studies used single-economy models, sometimes in combination with a global model such as GTAP or LINKAGE, to analyze the socioeconomic impacts of policy changes on a single economy in the region. As far as we are aware, this is the first CGE study that attempts to deal with household income distribution issues in the context of the whole region simultaneously, using a disaggregated CGE model. Khan (2008) presented very preliminary results for a prototype model for South Asia. The model was an interesting approach, incorporating several non-standard features, including technological dualism and rural-urban migration of the Harris-Todaro type. Although the model included eight household groups, it had only three production activities, and was calibrated to a single country (India). Hence the results from Khan (2008) are relevant for other countries in the region only by extension in the model's current form, although the approach might be usefully adapted to other countries in the future.

Download this Paper [ PDF 210.1KB| 26 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.