Relation with Poverty: Endogenous Incomes and Poverty Line

How does the above analysis relate to the measure of poverty? The standard poverty measure depends critically on the poverty line (PL) and incomes of the poor (Y^Poor). The starting point to determine PL is to select a basket of basic needs reflecting the consumption pattern of households near the presumed poverty line and yielding threshold caloric requirements. Food is typically by far the most important commodity in the basket of basic needs. Denoting the basket of basic needs with π_com, the poverty line is essentially Σ_com π_com. P_com, where P_com is the endogenously derived poverty line prices. To arrive at the poverty measure, one has to determine first the intra-group income distributions corresponding to the characteristics of each group. Given such a distribution, various poverty measures may then be used.⁶

Clearly, PL and INC_h hold the key to the poverty measure. In an expansionary policy, the price increase can be larger or smaller than the increase in PL; and, GDP expansion may be greater or smaller than the increase in incomes of the poor (Y^Poor). A contractionary policy may also generate a fall of PL that is smaller or larger than the decline in GDP. The results vary by country and type of policies being implemented (e.g., fiscal versus monetary). The precise relation will be known only after the transmission mechanisms are evaluated empirically. In the cases of Thailand and Indonesia, it is known only from the earlier analysis that the AS curve is flat, but how it translates into poverty is yet to be analyzed.

If under an expansionary policy the relation between PL and Y^Poor form a convex curve (PL being the x-axis), the likelihood of improved poverty conditions is high because the increase of PL is smaller than the increase of Y^Poor. However, there is no reason that a convex curve cannot result in a larger increase of PL than of Y^Poor. Two scenarios may arise under a convex curve case: worsening poverty condition—represented by the thin curves and thin arrows; and improved poverty condition—represented by the bold curves and bold arrows in quadrant 2, 3, and 4 (Figure 12 [ PDF 49.5KB | 1 page ]). Thus, even if we know the precise shape of the curves, it is still uncertain whether the expansionary policy will generate a favorable or unfavorable poverty outcome. The same is true for the contractionary policy, where all curves are concave: the case of worsening poverty is depicted by the thin curves, while the bold curves represent a scenario of improved poverty condition (Figure 13 [ PDF 49.5KB | 1 page ]).

As far as the relation between GDP and Y^Poor (in quadrant 4) is concerned, the dynamic slope of the curve also predicts what happens with the income distribution: in an expansionary policy, a concave curve suggests that inequality tends to worsen after the policy shock, whereas in a contractionary policy, a concave curve implies an improvement in income distribution after the shock. On the other hand, in an expansionary policy, a convex curve suggests that inequality tends to improve after the policy shock, and in a contractionary policy a convex curve implies a deterioration in the post-shock income distribution.

To the extent that the relations between PL and price level (PINDEX), as well as between output (GDP) and incomes of the poor (Y^Poor) are too complex to be estimated in a partial equilibrium setting, a general equilibrium model with a detailed financial sector is used to estimate those relations.

First, I started with the relation that captures the response of poverty line (PL) to changes in prices. The general price level is endogenously determined through the interactions between supply and demand of both domestic and foreign goods, in which the demand consists of domestic and import demand (PD.D + PM.M):

where Q, D, and M refer to total supply of goods available, goods produced and sold domestically, and imported goods, respectively, and subscript p denotes the economic sector. PQ, PD, and PM are the corresponding prices. A similar notion applies to the prices of domestic output, PX:

where tdom and ttd are indirect tax rates on domestic goods and the trade and transport margin rate on domestic goods, respectively.

The above specification is based on a production structure that is modeled as a set of nested constant elasticity of substitution (CES) function. In the first stage, the production function (expressed as value-added) is determined, in which the primary inputs are the righthand side variables. Since a considerable portion of intermediate inputs are imported, the composite intermediate inputs INTM are modeled as a CES function of domestic and imported inputs (DOMINTM and FORINTM).⁷ In the second stage, the domestic output is specified as a CES function of the value-added VA and the composite intermediate inputs. The resulting price of value-added PV is:

where PINTM is the price of intermediate inputs. The unit price of imported and domestically produced intermediate inputs (PDINTM and PFINTM) are, respectively:

where aad and aam are the share parameters, and subscripts p and pp refer to the production sector. From these two equations, the price of composite intermediate inputs is derived:

The value-added price PV determines the nominal value-added. After taking into account the indirect tax (INDTAX), TARIFF, and subsidy (SUB), the nominal GDP (GDP at current price, or GDPCUR) can be derived:

The general price level (PINDEX) is derived as the GDP deflator:

where the GDP at constant prices is derived from the expenditure side.

The prices of the basic needs presumably consumed by the poor, are classified according to urban and rural, and formal and informal. Rural poverty line prices are distinguished from prices in urban areas, and so are the consumption patterns. Hence, the relation between P and PL is:

where α_p^r,u is the sectoral consumption parameter that captures different consumption patterns between rural and urban.

Next is to identify the relation between GDP and incomes of the poor in the lower right quadrant of Figures 12 and 13. Income of different households consists of factor income (wages), transfers, and income from financial assets. Given labor market segmentation (wages being strongly sector-specific), labor income is specified as follows:

where PDL0 and FACDEM are, respectively, labor productivity before the shock and factor (labor) demand. Note that ρ, the price electivity of wages, can play a critical role in determining the effect of a policy shock that causes changes in prices on wage income. In particular, an expansionary macroeconomic policy (e.g., a positive AD shock) could affect the wages of the low income group differently when the value of ρ is altered. This implies that the value of ρ can determine the resulting poverty, depending on a given poverty line.

The average wage rates for each labor category are arrived at on the basis of the above sectoral wage rates and the wage shares of each type of labor in each sector (wsharep,fl):

The unemployment and rural-urban migration reflect the slack in the labor market as such that the total supply of labor equals the demand for labor plus the unemployed labor force.

Household income from sources other than factor income is denoted by ITRAN. It consists of transfers among households, firms, and rest of the world (OTRAN), government subsidies (GTRAN), and returns on financial assets (RTRAN):

where i, j reflect different institutions.

The inclusion of financial assets, which is the core feature of the CFGE model, is particularly important amid what has been happening during the last few years in most emerging markets, including Thailand and Indonesia. In these two countries, there exists an excess liquidity (saving) characterized by a faster growth of investment in financial assets than in the real sector. This phenomenon is supported by existing data which shows total saving as greater than total investment in the real sector (most excess saving goes to financial investments).

Unlike in a standard non-financial computable general equilibrium (CGE) model, investment is endogenously determined by the investment function and institutional portfolio allocations (“fixed” asset investments). Institutional savings will also be a part of the institutional balance sheet as they represent changes in wealth.⁸ While, in general, the rate of return for each asset is determined based upon the supply and demand of financial assets, some returns determine the supply (e.g., the supply of time deposit follows the demand and given deposit rates)—and others determine the demand (e.g., the demand for government bonds is determined by how much is offered and at what rate).

The saving-investment closure in the model departs drastically from the neo-classical specifications. Based on a number of empirical studies (Azis, 2002), investment in sector p can be specified as a function of value added VA (output accelerator), loan interest rates rn_c, and exchange rate EXR:

where rn_c is the loan interest rate and λs are constant; the size of λ3 depends on the sensitivity of investment on exchange rate fluctuations. This specification reflects the financing behavior of agents (i.e., bank-dependent) and balance sheet constraints (Bernanke & Gertler, 1989, Krugman, 2001, Aghion, Bacchetta and Banerjee, 2001). When the exchange rate is stable, few firms are constrained by their balance sheets: the direct effect of EXR on aggregate demand is minor. On the other hand, if the exchange rate depreciates sharply, agents’ ability to expand is adversely affected. Since the balance sheet effect in Thailand during the post-crisis period has declined substantially, EXR is not included in Thailand’s investment function.

The portfolio allocation of institutions is specified based upon the assumption that there is no perfect substitutability, as suggested by Tobin (1970); Brunner and Meltzer (1972); Bernanke and Blinder (1988), and used in Bouguignon, Branson, and de Melo (1989), and Thorbecke et al. (1992). In the Thai model, after specifying the money demand of household hh (MD_hh), the following equations determine how much households want to hold in cash and demand deposits:

where (cush_hh) is a fixed share. The weighted average rate of return on other assets (i.e., time deposit, equity, and government bond), rnh1_hh, is defined as:

where subscripts td, eq, and gb refer to, respectively, time deposit, equity, and government bond. The ratio of time deposit to equity (gh1/(1-gh1)) depends on the ratio of their returns, i.e., interest rate on time deposit (rn_td) and return on equity (rn_eq):

where thetah1hhf and sigmah1_hh are constant. Thus, the values of time deposit and equity are, respectively:

In the Indonesian model, incomes received by household h from institution j are determined as follows:

where rns is the return on asset type s, LiablGj,s is the stock of asset type s transferred from institution j at the beginning of the period and AssetsH_hh is the share of total assets held by household j. For institution i, the latter is defined as:

where AssetlG_s,i and AssetlGs,j are, respectively, the stock of asset type s in the beginning period held by institution i, and the stock of asset type s transferred from institution j. For example, if i is urban-rich household and s is the time deposit, the above equation indicates the ratio of time deposit held by urban-rich household over the time deposit of all households. Thus, the ratio shows how much of the total time deposit is held by the urban rich.

From the above specifications, it is clear that if the interest rate rns is raised to combat inflation (s is the time deposit), the RTRAN of households hh which hold savings will also increase. Hence, those who own more time deposits will receive higher incomes. From this mechanism alone, the relative income distribution can be altered, implying that the link between macroeconomic policy and poverty, as well as the growth-poverty nexus, can also change.

Download this Discussion Paper [ PDF 978KB| 34 pages ].

Comment(s)

There are [1] comment(s) for this entry. Post a comment.

1. Ruly
   (posted 11 July 2008 / 02:08:15 PM)
   
   This paper gives me a new insight about the real condition of my country economy, and how the authority (government) had applied many conventional economic policies, which is not suitable for Indonesian economy that has different characters and conditions from the west, the origin of these conventional policies..
   
   Furthermore, I was very lucky because the author presented this paper in my class as a guess lecturer about 2 months ago. For the first time after 1.5 years studying economics, a lot of my "unanswered" questions being answered, and I even didn't have to ask..