Research Statement


I am a macroeconomist with interests in theory and methods. My subfields are listed below. 

In Macro theory I try to combine general equilibrium theory with the stochastic process literature in order to derive computable structures for infinite horizon non-optimal economies. In this sense, I am extending GE theory beyond existence results in order to be able to characterize numerically and empirically macro models. I have 3 papers in this field: i) an applications to heterogeneous agents economies, ii) a project on sudden stop models, iii) a paper on asymmetric fiscal policy in closed economies.

In Applied Recursive Macro I develop applications of minimal state space recursive macro models. I evaluate these models empirically using different structural estimation techniques. I have 5 papers in this field: i) a project on financial crises and macro prudential policies estimated using “hitting times” of markov processes. ii) I also study the pass through of monetary policy instruments, calibrating the model to Latin American data. iii) I estimate a state dependent tax policy multiplier using indirect inference with quantile VARs. iv) I develop a theoretical framework to understand the interaction between sovereign default and sudden stops. v) I study the relevance of lenders’ heterogeneity (in preferences and wealth) on financial market crises using a novel data base containing a panel of brokers’ sovereign debt transactions across time. 

In Recursive methods I combine recursive equilibrium theory and computer science to measure the performance of standard algorithms. I have 2 papers in this field: i) a project on the accuracy of minimal state space methods for non-optimal economies. ii) I also estimate the rate of convergence and accuracy of the Value and Policy Function Iteration algorithms for efficient economies with price independent inequality constraints.

In Search models I study recursive economies with risk averse agents and non-segmented labor markets. I use them to undertand the effects of pension reforms and different regulations on several unemployment characteristics. I have 2 papers in this field: i) a project on the effects of social-security reforms in Chile, ii) I study the interaction between income risks and liquidity constraints for the 2008 pension reform in Argentina using a novel data base with a panel of 2 million observations.

Below you can find a list of extended abstracts for each project 

Extended Abstracts, Links to the papers & Presentations

Macro Theory


i) Useful Results for the Simulation of Non Optimal General Equilibrium Economies

Job Market Paper

This paper investigates the empirical evaluation of infinite horizon non-optimal economies by means of numerical simulations. The paper answers the following question: is it possible to derive a general framework which guarantees that numerical simulations truly reflect the behavior of endogenous variables in the model? Under mild assumptions, this paper provides an affirmative answer to this question for endowment economies with incomplete markets, heterogeneous agents and infinitely many exogenous states. For this type of models, the paper also presents an accurate calibration method. For economies with finitely many shocks, even under stronger assumptions, it is only possible to show that a numerically computable, time independent and recursive representation of the sequential equilibrium generates a stationary Markov process, which is a necessary condition to answer the above mentioned question.

Link to the paper :

ii) Memory, Collateral and Emerging Market Crises: A Qualitative Approach with Quantitative Implications

Joint with K. Reffett

This paper presents a flexible and yet computable markovian equilibrium notion, called Generalized Markov Equilibrium (GME), that is suitable for a class of macro models typically used to represent financial crises in emerging countries. This type of equilibrium is able to represent a larger fraction of the (multiple) sequential equilibrium set when compared with a minimal state space recursive equilibrium (MSSRE). Thus it incorporates the role of "memory", which characterize the sequential equilibria, to a recursive representation. For a particular model in the mentioned class, we show that there exist an ergodic GME which at the same time replicates all the observed phases in a macroeconomic crises (boom, collapse and recovery / spiralized recession) and is able to capture the stylized behavior of data due to its ergodic nature. Moreover, we derive a computation and calibration method that captures the short and long run behavior of data accurately. Finally, we show the existence and characterize a MSSRE using a constructive method, which allows us to derive a convergent algorithm and robust comparative statics. This procedure uses a different state space when compared with the standard practice.

Link to the presentation:

iii) Asymmetric fiscal policy: theory and practice

Joint with Juan Pablo Gama and Juan Pablo Rincón Zapatero

We build on the asymmetric business cycle theory (Hansen and Prescott, 2005) to develop a novel equilibrium notion which allows to replicate the global dynamics in a possibly inefficient economy subject to fiscal shocks. The theoretical structure is sufficiently general to compute and simulate tax and Government expenditure shocks of an arbitrary magnitude which may affect the behavior of the economy differently depending on the position of the economy along the cycle. We prove the existence of a sequential and of a recursive equilibrium for this economy. Under an additional set of assumptions, the sequential equilibria can be computed directly from the system of equations characterizing the dynamic behavior of the economy. Under milder assumptions, we can compute the recursive equilibria.

Work in progress


Applied Recursive macroeconomics


i) The empirical dimension of overborrowing

Joint with G. Montes Rojas and P. Mira

Persistent current account deficits are common among low and middle income countries. When is this situation dangerous? Is there a critical value for the yearly current account deficit just before the crisis sets off
? We provide a positive answer to the last question; a finding that gives rise to an empirical measure of overborrowing. We observe that countries that have increased their external indebtedness by at least 26%-31% of the GDP in a time span of 3 to 5 years are more prone to be hit by a sudden stop. The typical crisis produces a consumption drop of 4% of GDP and a current account reversal of 2.5-4.5% of GDP. We also contribute to the structural characterization of sudden stops. Using a canonical model we are able to replicate these stylized facts matching the “time between crises”. Moreover, we compute the ratio of net debt to GDP. This parameter is two or three times bigger than the benchmark value in the literature, a fact that improves the empirical performance of the model. From a policy perspective, our findings help to elaborate leading indicators to anticipate a sudden stop.

Link to the paper:

ii) Incomplete interest rate pass through: a recursive partial equilibrium approach

Joint with M. Cherkasky and L. Trajtenberg

The empirical literature provides robust evidence on the incomplete pass through (IPT) of monetary policy instruments on market rates. However, the coefficients found are not similar across regions, ranging between 0.25 and 0.75. Also, the theoretical literature has not provided a comprehensive framework which can explain this fact satisfactorily. This paper proposes a recursive partial equilibrium model that generates a PT coefficient between 0.5 and 0.6. The model is calibrated for Latin American economies where the active (passive) rate is above (below) the policy rate and the spread is stable. These facts allow explaining the IPT: after a tightening in the monetary policy, banks switch the composition of their assets from loans to central bank notes. This change allows them to reduce the effects on earnings caused by the tightening provided that there is an IPT in the passive rate and that this rate is below the policy rate. The stability of the spread explains the PT in the active rate. In order to obtain the IPT coefficients, this paper derives a structural estimation of the supply elasticity of deposits and demand elasticity of loans. The former ranges between 3.0 and 4.0 and the latter between -1.5 and -2.0. These results suggest that the efficiency of the monetary policy might be affected more by an incomplete PT than by the lack of sensibility of loans and deposits.

Link to the paper:

iii) Asymmetric business cycles and fiscal policy: the effect of idle capacity.

Joint with G. Montes Rojas

We derive a simple extension of the model in Hansen and Prescott (2005) to allow for income taxes in a decentralized framework. We then structurally estimate the differentiated effects of sing and state dependent fiscal shocks using indirect inference with quantile VARs. We impose a set of assumptions on preferences which simultaneously allow to compute the model using a convergent global algorithm and replicate the observed effects of income tax shocks on consumption and investment.

Work in progress  

iv) Default Risk and Fiscally induced sudden stop

Joint with H. Seoane and A. Vicondoa

To our knowledge the literature has taken care separately of sudden stops on private external debt and of defaults on public debt. This paper is the first attempt to model the effects of sovereign default on private expenditure when external debt is mainly issued by the public sector and there is a positive probability of suffering a sudden stop. The model takes the standard structure in models with financial frictions in small open economies as in Bianchi (2011) and introduce default sobering risk using a framework adapted from Lorenzoni and Werning (2019). In order to structurally estimate the model, we compute directly the sequential equilibria exploiting the finite time assumption inherited from Lorenzoni and Werning (2019). 

Work in progress.

v) The importance of the buy side in emerging market crises

Joint with S. Barraza and F. Roldan

The macro literature since Mendoza (91) have modeled emerging market crises as a shortage in international lending. The focus is always on the behavior of the borrower. But, what about the lender? As a consequence of assuming that lenders have great tolerance to risk and that are price takers, financial market equilibrium is dichotomous: the borrowers fix the amount of the debt issuance and the lender decides the price. More to the point, the wealth distribution of lenders is irrelevant.  Based on a novel data set, this paper challenges these assumptions. We document the position in Argentinean sovereign debt bond by bond and broker by broker. We found not only that the availability of funds, represented by the net worth of each “brand” of mutual funds, is crucial but also that investment decisions are irreversible. We also observe that in each debt issuance (primary auctions) the fraction purchased by a particular hedge fund is no larger than 5%, suggesting a competitive behavior.

Work in progress.

Recursive Methods


i) Recursive Equilibria, are you there? Measuring the accuracy of minimal state space methods

Joint with J. Martinez 

Duffie, et. al. (1994) shows that in non-optimal economies with a finite number of exogenous shocks there is a tradeoff between the generality of a recursive representation and a well behaved steady state, which is defined by an ergodic invariant measure of a stationary Markov Process. The authors "convexified" the state space using "sunspots" in order to prove the ergodicity of the measure but this procedure is not compatible with the computation and simulation of the model as it is not clear how sunspots affect the economy. The purpose of this paper is to show that, in certain environments, it is possible to obtain a recursive representation of a non-optimal general equilibrium model with a finite number of exogenous shocks that has an ergodic invariant measure, a compact and stationary state space and no "sunspots". By enlarging the number of variables in the state space, this paper proves the existence of multiple continuous markovian representations; which allows deriving an ergodic invariant measure for each of them using standard results. These facts show, contrarily to what is claimed in Blume (1982), that it is possible to obtain an economy with multiple equilibria and a continuous markovian representation. Moreover, for a stochastic RBC model with taxes, this paper derives a closed form recursive representation. These results are then used to test the performance of minimal state space recursive equilibrium methods. Even if the algorithm converges, the numerically simulated distribution does not match any of the possible ergodic measures (it may over/sub-estimate concentration and dispersion measures of the true ergodic distribution)

Link to the paper:

ii) Theoretical Error Bounds for the Value and Policy Function Iteration Algorithms: An Application for Recursive Dynamic Models with Inequality Constraints.

Winner of the Society for Computational Economics  Contest 2011

This paper derives theoretical error bounds for the Policy and Value Function Iteration algorithms applied to Recursive Dynamic Models with continuous decision variables and inequality constraints. This paper proves two main theorems. The first one uses a recent result due to Santos and Rust (2004). The theorem extends the result by combining a feasible version of the Policy Function Iteration algorithm with the barrier method for a model with an arbitrary number of state and decision variables. This constitutes a significant difference with the original theorem, since it is no longer necessary to assume the interiority of the solutions. The Algorithm converges at a rate of 1.5 for a given grid size. The second theorem, for problems with only one continuous endogenous state variable, uses a feasible version of the Value Function Iteration Algorithm, the barrier method and a cubic Variation Diminishing Spline Approximation. The algorithm converges at a linear rate, given the grid size. Finally, under a certain configuration of the parameters, the maximization problem in this last theorem is in the convex class, which can be solved in polynomial type complexity, and the policy function is first order differentiable. These last results enables the algorithm to avoid the Course of Dimensionality for the maximization problem in the Bellman Equation and the use of first order perturbation methods, thus, constitutes an extension of existing theorems that deals only with equality constraints (see Judd, 1994).

Link to  the paper & Presentation (a new version is  coming joint with L. Bali):

Partial equilibrium search models


i) Evaluating Pension Reform and Labor-Market Policy Changes in Chile through a Structural Search Model

Joint with E. Kawamura

This paper estimates and calibrates a structural search model with formal and informal sectors and life-cycle elements to analyze counterfactual social-security reforms relevant for Chile using data from the harmonized Longitudinal Social Protection Survey (LSPS) and other official sources. The paper considers a drop in contribution rates and an increase in the non-contributory pension payments after retirement. A drop in the contribution rates implies increase in the average duration of unemployment, rejecting both types of job offers more often. Also, an increase in the non-contributory pension reduces the average duration of unemployment by increasing the acceptance of informal job offers.

Link to the paper:

ii) A measure of the wealth distribution for Argentina: implications for the labor market and the pension system.

Joint with D. Trupkin

Using a panel of 2 million observations we can keep track of the labor market history of a subset of the formal labor market in Argentina. The data allows me to estimate the wealth distribution of non-retired agents using a search model with risk averse agents. We then can compute the effect of different fiscal policies on the labor market conditional on the position on the income and wealth distribution of each worker. 

Work in progress.