Multiperiod Loans, Occasionally Binding Constraints, and Monetary Policy: A Quantitative Evaluation

• Author: KRISTINA BLUWSTEIN,MARCIN KOLASA,MICHAł BRZOZA‐BRZEZINA,PAOLO GELAIN
• Date: 01 October 2020
• DOI: http://doi.org/10.1111/jmcb.12689
• Published date: 01 October 2020

DOI: 10.1111/jmcb.12689

KRISTINA BLUWSTEIN

MICHAŁ BRZOZA-BRZEZINA

PAOLO GELAIN

MARCIN KOLASA

Multiperiod Loans, Occasionally Binding

Constraints, and Monetary Policy: A Quantitative

Evaluation

Weintroduce multiperiod mortgage loans, xed interest rate, a lower bound

constraint on newly granted loans, and a possibly slack collateral constraint,

in an otherwise standard Dynamic Stochastic General Equilibrium (DSGE)

model with housing. Our nonlinear estimation shows that all those features

are important to understand the evolutionof mortgage debt during the recent

U.S. housing market boom and bust. The transmission of monetary policy

becomes dependent on the housing cycle, with weaker effects when house

prices are high or start falling sharply. Higher averageloan duration makes

monetary policy less effective, eventuallyleading to asymmetric responses

to positive and negativemonetary shocks.

JEL codes: E44, E51, E52

Keywords: mortgages, xed-rate contracts, monetary

A     nancial crisis, mortgage

markets can play an important role in driving businesscycles. Moreover, they strongly

A previous version of the paper was circulated under the title “Monetary and macroprudential policy

with multiperiod loans.” We would like to thank Francesca Barbiero, Markus Brunnermeier,Carlos Gar-

riga, Richard Harrison, Matteo Iacoviello, Michael Kiley, Caterina Mendicino, Gisle J. Natvik, Johannes

Pfeifer, Giorgio Primiceri, Soren Ravn, Roman Šustek, Lars Svensson, Andrea Tambalotti, and Harald

Uhlig for useful discussions, and two anonymous referees for suggestions to the earlier draft. We also

beneted from comments by participants at the Computing in Economics and Finance conference in Oslo,

Annual Meeting of the Society for Economic Dynamics in Toronto, Dynare conference in Paris, Central

Bank Macroeconomic Modeling Workshop in Rome, Society for Nonlinear Dynamics and Econometrics

Symposium in Oslo, WGEM meeting at the European Central Bank, NBP Summer Workshopin Warsaw,

and seminars at the Bank of Finland and the Halle Institute for Economic Research. This paper does not

necessarily represent the views of the Bank of England, Narodowy Bank Polski, the Federal ReserveBank

of Cleveland or the Board of Governors of the Federal Reserve System.

K B is a Research Economist at Bank of England, Financial Stability Strategy

and Risk (E-mail: kristina.bluwstein@bankofengland.co.uk). M B-B is a Research

Economist at Narodowy Bank Polski and Associate Professor at SGH Warsaw School of Economics

(E-mail: michal.brzoza-brzezina@nbp.pl). P G is a Research Economist at the Federal Re-

serve Bank of Cleveland (E-mail: Paolo.Gelain@clev.frb.org). M K is an Economic Advisor

at the NarodowyBank Polski and an Associate Professor at the SGH WarsawSchool of Economics (E-mail:

marcin.kolasa@nbp.pl).

Received February 22, 2016; and accepted in revised form June 26, 2019.

Journal of Money, Credit and Banking, Vol. 52, No. 7 (October 2020)

© 2020 The Ohio State University

1692 :MONEY,CREDIT AND BANKING

interfere with macro-economic policies, and monetary policy in particular. These ob-

servations led to an unprecedented boom in the creation (and publication) of macro-

economic models featuring nancial intermediation, housing markets, and mortgage

loans, the early examples of which were Iacoviello (2005), Iacovielloand Neri (2010),

and Gerali et al. (2010).

In spite of its apparent importance, structural modeling of mortgage and housing

markets usually assumes—needless to say,counterfactually—that mortgage loans are

granted for a single period, which in most models corresponds to one quarter. This

makes it impossible to incorporate xed-rate contracts, despite their prevalence in

some countries, and most notably in the United States. Second, most models that fol-

low the seminal contribution of Iacoviello (2005) and build nancial imperfections

on the concept of collateral constraints ignore another important fact: that such con-

straints usually bind only occasionally. However, while a potential creditor can be

prevented from taking a loan, she cannot be forced to take one. Last, but not least,

another nonlinearity associated with mortgages seems important: a lower bound on

new loans. Again, there is an asymmetry at play: a borrower can be prevented from

taking a new loan, but usually cannot be forced to accelerate the repayment of old

loans when the collateral value declines.

This brings three features of mortgage markets to our attention: the multiperiodic-

ity of loan contracts, the occasionally binding nature of collateral constraints, and the

existence of a lower bound for new loans. It should be highlighted that these are not

only potentially important stand-alone features of the mortgage market, but they can

also enter into powerful interactions. For instance, both the already mentioned xed-

rate contracts and the lower bound on new loans make sense only in a multiperiod

contract setting. While all three features have been dealt with in the literature before,

we are not aware of any study that offers a thorough, quantitative assessment of the

role they played in shaping macro-economic dynamics or that explains how they af-

fect the transmission of monetary policy. We believe that these are highly relevant

topics and we try to ll the existing gap in the literature.

Our paper offers three contributions to the literature reviewed in the next section.

First, we conrm the key importance of multiperiod contracts and the two nonlinear-

ities in explaining the U.S. housing and nancial cycle in the past 25 years, which

linear models have a hard time matching. In particular, we document, period by pe-

riod, when and how the nonlinearities were relevant. Second, we demonstrate and

explain how the phase of the housing cycle affects the transmission of monetary pol-

icy. Last, butnot least, we are the rst to provide a nonlinear Bayesian estimation of a

Dynamic Stochastic General Equilibrium (DSGE) model with housing and nancial

frictions, where mortgage contracts are multiperiod, the collateral constraint faced by

borrowers can be occasionally binding, and new loans are subject to a lower bound.

More specically, we construct a DSGE model with real and nominal rigidities

as well as housing and nancial intermediaries. In contrast to much of the literature,

we allow loans to be multiperiod and carry a constant interest rate over the contract’s

duration. Weintroduce the two nonlinearities mentioned above, which have also been

recently highlighted by Guerrieri and Iacoviello (2017) and Justiniano, Primiceri, and