arXiv subjects

Marko Mlikota

Publications and source records attributed to Marko Mlikota.

4 recordsLinked to original sources

arXiv2026

Dynamic Innovation Transmission Through Networks: Theory, Large T-Inference, and Business Cycles by Lagged Input-Output Conversion

I develop an econometric framework that rationalizes the dynamics of a cross-sectional variable by lagged transmissions of innovations along bilateral links between units. While nesting the Spatial Autoregression and Spatial Error Model as limits and producing equivalent impulse-responses in the long run, the Network-Vector-Autoregression (NVAR) I propose can accommodate general transmission patterns over time and yields "networked" transition dynamics distinct from those implied by autocorrelated innovations. I discuss large $T$-inference conditional on a network. I then estimate an NVAR for US sectoral output, as derived under a Real Business Cycle economy with lagged input-output conversion. Under the preferred specification, lagged transmissions of productivity shocks along supply chains account for 85% of the persistence in aggregate output growth and reduce shock-variances relative to an economy with contemporaneous input-output conversion by 73% on average across sectors.

econ.EM↗

arXiv2026

Parameter Identification and Inference in Discretely Sampled or Temporally Aggregated Autoregressions

I consider an AR($p$) process that is observed every $q$ periods, either as a snapshot (stock variable) or as a sum over the sampling interval (flow variable). I first characterize the resulting ARMA process followed by observables. Under fairly mild assumptions, I then derive the identified set for general lag lengths $p \in \mathbb{N}$ and sampling frequencies $q \in \mathbb{N}$, I bound its cardinality, and I provide an algorithm to compute all candidate points and determine their membership in the identified set. My exact but implicit characterization supports the following conjecture that I prove in some settings and verify numerically more broadly: (i) the error term-variance is point-identified, (ii) under temporal aggregation, the autoregressive parameters are point-identified, and (iii) under discrete sampling they are point-identified for odd $q$ and identified up to alternating sign for even $q$. My analysis supplements existing inference results that show consistency and asymptotic Normality of the Gaussian Maximum Likelihood estimator conditional on point-identification. Holding the number of observations fixed, I show that its precision does not necessarily decrease with $q$.

econ.EM↗

arXiv2025

Origins and Nature of Macroeconomic Instability in Vector Autoregressions

For a general class of dynamic and stochastic structural models, we show that (i) non-linearity in economic dynamics is a necessary and sufficient condition for time-varying parameters (TVPs) in the reduced-form VARMA process followed by observables, and (ii) all parameters' time-variation is driven by the same, typically few sources of stochasticity: the structural shocks. Our results call into question the common interpretation that TVPs are due to "structural instabilities". Motivated by our theoretical analysis, we model a set of macroeconomic and financial variables as a TVP-VAR with a factor-structure in TVPs. This reveals that most instabilities are driven by a few factors, which comove strongly with measures of macroeconomic uncertainty and the contribution of finance to real economic activity, commonly emphasized as important sources of non-linearities in macroeconomics. Furthermore, our model yields improved forecasts relative to the standard TVP-VAR where TVPs evolve as independent random walks.

econ.EM↗

arXiv2022

Sequential Monte Carlo With Model Tempering

Modern macroeconometrics often relies on time series models for which it is time-consuming to evaluate the likelihood function. We demonstrate how Bayesian computations for such models can be drastically accelerated by reweighting and mutating posterior draws from an approximating model that allows for fast likelihood evaluations, into posterior draws from the model of interest, using a sequential Monte Carlo (SMC) algorithm. We apply the technique to the estimation of a vector autoregression with stochastic volatility and a nonlinear dynamic stochastic general equilibrium model. The runtime reductions we obtain range from 27% to 88%.

econ.EM↗