Research
Works in Progress
“Predicting forecaster inattention”
Abstract: Economic forecasters have been documented as deviating from having both full information and rational expectations. The sticky information model attempts to explain economic forecasters deviating from having both full information and rational expectations by assuming information rigidity: each period a fraction $\lambda$ of forecasters do not get their information sets updated. To explain the data, empirical findings use calibration or follow Coibion and Gorodnichenko (2015) and estimate $\lambda$ using aggregated regressions. This leads to large estimates, with up to 50% of forecasters not receiving information each quarter. Adapting Andrade and Bihan (2013) and Giacomini et al. (2020), I employ a different estimation strategy using data from the US survey of professional forecasters and provide evidence that $\lambda$ is 10 times smaller. My results imply that some other, stronger source of information rigidity must be present, and thus a model accounting for information rigidity should not use sticky information as its only mechanism. I also document novel state-dependence and show how individual and group forecast errors affect the likelihood of forecasts being sticky, with the effects being heterogeneous and nonlinear across forecast horizons.
“Optimal income tax progressivity over the business cycle”
Abstract: I study optimal progressive income taxation in an incomplete-markets model with aggregate state-dependent skewed and fat tailed idiosyncratic income risk and a two-state regime switching process for aggregate productivity. A Ramsey planner is constrained to two log-linear tax and transfer functions, one for each aggregate state, to allow state-dependent progressivity. I take recent computational techniques used for aggregate MIT shocks and adapt them to work for regime switching processes. I find that the overall negative skewness of income risk leads to higher optimal progressivity overall, but the difference between recessions and expansions is negligible. I also find that the welfare difference from optimal policy between an income process that incorporates skewness and fat tails and a misspecified process that does not can be upward of 8%.
Published Papers
“A study on discrete Ponzi Scheme model through Sturm-Liouville theory”
Recommended citation: Atici, F.M. and Bennett, W.R. (2021) “A study on discrete Ponzi Scheme model through Sturm-Liouville theory”, Int. J. Dynamical Systems and Differential Equations, Vol. 11, Nos. 3/4, pp. 227-240.
Abstract: In this paper, we introduce a second order self-adjoint difference equation which describes the dynamics of Ponzi schemes: a type of investment fraud that promises more than it can deliver. We use the Sturm-Liouville theory to study the discrete equation with boundary conditions. The model is based on a promised, unrealistic interest rate $r_p$, a realised nominal interest rate $r_n$, a growth rate of the deposits $r_i$, and a withdrawal rate $r_w$. Giving some restrictions on the rates $r_p$, $r_i$, and $r_w$, we prove some theorems to when the fund will collapse or be solvent. Two examples are given to illustrate the applicability of the main results.