Publications

In this paper, we analyze how borrower characteristics influence the choice between the short-term fixed-rate mortgage (STFRM) and the long-term fixed-rate mortgage (LTFRM) types in an emerging market. We use the national Survey of Households with Housing Loans conducted by the Bank of Lithuania between 2009 and 2012. This paper is the first to empirically test the findings of Campbell and Cocco in an emerging market. Following the model of Campbell and Cocco (2003), our analysis focuses on the interaction of demand and supply. We focus on the contract outcome; we do not specify who initiated such outcome – the household or the mortgage provider. We supplement the findings of previous literature by testing the model among financially constrained households in a different economic and institutional setting; that is, in Lithuania.
We define financial constraints of a household in multiple ways: high mortgage payment-toincome ratio, low residual income, high loan-to-value ratio, absence of savings, existence of other obligations, a single breadwinner in the household, and the existence of dependants in the household. Estimates based on these measures indicate that constrained households are more likely to choose a safer, but more expensive, long-term interest rate mortgage. Our results are in line with Campbell and Cocco’s (2003) suggestion that, when borrowing constraints are binding, financially constrained households should choose a long-term interest rate mortgage. Our results contradict the empirical evidence of Coulibaly and Li (2009), of Damen and Buyst (2013), of Ehrmann and Ziegelmeyer (2013), and of Hullgren and Söderberg (2013) that financially constrained households prefer short-term interest rate mortgages. We argue that the difference arises because of institutional features.
Our study adds to existing literature by showing that, in the world where borrowing constraints are binding, financially constrained households have the safer mortgage type that reduces consumption shock and liquidity risks.