|dc.description.abstract||Chapter 1 evaluates if MFIs can realize cost savings from lending in both rural and urban markets, which would mean that the MFIs would exhibit economies of scale and cost complementarities. The presence of cost complementarities would suggest that the marginal costs of producing urban loans would decline as the rural loan portfolio expanded, and vice versa. A translog cost function with its respective share equations are estimated as a SUR model via maximum likelihood estimation separately for lending-only MFIs and deposit-lending MFIs. This was done to consider potential technological heterogeneity across the MFIs. In addition, two sets of translog cost functions were estimated for each MFI type. The first imposed both homogeneity and symmetry while the second relaxed the homogeneity restrictions. Based on the likelihood ratio test, the less restrictive model was a superior statistical fit to the MFI data. The results indicate that both MFI types exhibited economies scale averaging about 1.55 for lending-only MFIs and 1.52 for deposit-lending MFIs. Moreover, lending-only MFIs derived greater scale savings from rural loans which yielded a cost elasticity of 0.275 compared to urban loans which yielded a cost elasticity of 0.37. On the other hand, deposit-lending MFIs derived greater scale savings from urban loans which yielded a cost elasticity of 0.211 as compared to rural loans which yielded a cost elasticity of 0.45. Positive sample estimates for cost complementarities, and computed point estimates illustrate the potential for cost savings from the joint production of both rural and urban loans. Lending-only MFIs had 19.81% of the sample benefiting from marginal cost reductions while lending-deposit MFIs had 33.44% of the sample benefiting from marginal cost reductions because of the joint production of rural loans and urban loans.
Chapter 2 assesses the impact of the 2010 increase in the up-front mortgage insurance premium from 1.75% - 2.25% on the LTV ratios pursued by homebuyers. The results indicate that first-time home buyers are affected the worst as their LTV ratios decline by 2.4% - 3.7%. The impact of the policy was also analyzed based on income groups. The first income group had incomes ranging between $37,067 - $73,124 and the second income group had incomes ranging $74,000 - $149,987. As expected, the impact of the policy was on the lower income group led to an LTV ratio decline of between 2.1% - 2.9% while the higher income group had decline in the LTV ratio of 1.5%. The credit constraint hypothesis and the portfolio substitution hypothesis were as well analyzed. The results indicate that First-time home buyers faced a credit constraint in their LTV ratio decision while higher income group did not face the constraint. In the case of the portfolio substitution hypothesis, first-time home buyers and borrowers in both income groups seem to consider housing assets and non-housing assets as substitutes. Based on the two hypotheses, there are potential indirect effects of the increase in the up-front MIP which warrant further empirical analysis: (1) a possible rebalancing of homebuyers’ housing and non-housing consumption, (2) possible shift in homebuyers’ in investment towards non-housing assets, (3) possible liquidation of non-housing assets to overcome the liquidity constraint, and (4) possible down-shift for cheaper homes by homebuyers who face a high income constraint and lack assets to liquidate.
Chapter 3 analyzes the lending dynamics of US agricultural banks over the period 2011-2019 primarily focusing on liquidity and the capital-gap, their potential interaction and how this interaction manifests itself in both capital-deficit states and capital-surplus states. The results indicate a symmetric effect of liquidity on lending across both capital-states. Small banks in the sample show no link between liquidity and lending. On the other hand, medium-sized banks, and large banks report -0.013% and -0.049% decline in lending expected from a 1% increase in liquidity. However, this negative effect diminishes as the capital deficit is reduced but increases as the capital surplus is increased. The effect of the capital-gap is asymmetric across capital-states in all bank samples. In the case of small banks, it is only statistically significant in the capital-deficit state and the results indicate that a 1% decline in the capital-gap will lead to a 0.21% increase in lending. In the case of medium-sized banks and large banks the capital-gap is only statistically significant in the capital-surplus state. A 1% increase in the capital gap will lead to a 0.19% and a 0.29% increase in lending in medium-sized banks and large banks, respectively. In addition, both the effect of the capital-deficit gap in small banks and the effect of the capital-surplus gap in the case of medium-sized banks and large banks was lower at higher liquidity positions. The interaction between the capital-gap and liquidity, across all bank sizes, was found to be statistically significant but only in the capital-deficit state. In small banks, the capital-gap and liquidity interacted as substitutes in the lending process while in medium-sized banks and large banks they were found to be complements. In addition, when medium-sized banks and large banks were in a capital-deficit state the marginal effects of their capital-gap on lending were greater at higher liquidity levels. However, in the case of small banks the marginal effects of the capital gap across capital-states decline at higher liquidity levels.||en_US