Aleksei Oskolkov

We study the cross-section of currencies using a macroeconomic model with heterogeneous countries, segmented asset markets, and risk-averse financial intermediaries. The model admits a two-factor structure of bilateral exchange rates. The factors combine global business cycle shocks and financial shocks that reprice the business cycle risk. Countries’ exposures to these factors are heterogeneous and depend on their reliance on commodity exports and dollar asset holdings. We estimate the resulting factor structure empirically and use it to quantify the model. Our results indicate substantial heterogeneity in the drivers of dollar exchange rates across countries. Economies with high commodity exposure or low dollar holdings are risky, and their exchange rates fluctuate in response to global output and risk shocks. The exchange rates of countries with low commodity incidence and high dollar assets are less volatile, and their fluctuations are mostly determined by idiosyncratic local shocks.

I develop a heterogeneous-country model of the world economy with global financial intermediaries. Imperfect risk-sharing creates a wealth distribution across countries. Risk-off shocks make the intermediaries revise their models of returns, which leads to repricing of country risk and triggers aggregate capital flight. In these events, the risk premia rise steeply in poor countries, while rich countries experience “retrenchment”: local agents use their external assets to replace foreign investors and stabilize asset prices. Risky assets in rich countries endogenously become safe havens. I derive asset price responses to global shocks in closed form and show that accounting for gross capital flows is crucial for explaining the international cross-section of asset returns. The wealth gradient in the cross-section of asset price responses identifies real and financial shocks behind the global financial cycle. A simple two-factor regression implied by the model explains 57% of the variation in the global cross-section of equities.

I build a tractable general equilibrium framework with investors operating under stochastic value-at-risk constraints. Constraints can be regulatory or self-imposed, originating in robustness concerns: intermediaries choose alternative models of asset returns to proof their decisions against model misspecification. Economies with heterogeneous value-at-risk limits admit a form of aggregation. Asset prices and interest rates depend on the representative risk limit, which is the wealth-weighted average of individual risk limits. Wealth distribution dynamics depend on a single stochastic process: the forecast error in the representative misspecified model of the total market return chosen by robust agents. The framework nests both forces driving time-varying risk premia in the literature: shocks to risk-bearing capacity and redistribution between heterogeneous agents. I decompose risk premium dynamics into these two parts and show that risk limit shocks do not redistribute directly, only changing investors’ leverage, which then determines how output shocks affect the wealth distribution.

This paper studies capital misallocation in a tractable model with random fixed costs of adjustment. We identify the distribution of fixed costs and productivity shocks using the entire size distribution and frequency of investments and provide an efficient estimation method. We derive the measure of capital misallocation in the presence of fixed costs and show that it differs from the traditional metric based on the variance of marginal product of capital. The key feature of models with lumpy investments responsible for this is their non-linearity: the distribution of marginal product of capital is not log-normal even with normal shocks and non-degenerate even when shocks are small. We apply our method to 40 years of panel data on Italian firms and find misallocation costs about 0.5-2% of output. Fixed costs contribute about one half to traditional measures of TFPR dispersion, putting an upper bound on potential inefficiencies.

We give a full analytic characterization of a large class of sticky-price models where the firm's price setting behavior is described by a generalized hazard function. Such a function allows for a vast variety of empirical hazards to be fitted. This setup is microfounded by random menu costs as in Caballero and Engel (1993) or, alternatively, by information frictions as in Woodford (2009). We establish two main results. First, we show how to identify all the primitives of the model, including the distribution of the fundamental adjustment costs and the implied generalized hazard function, using the distribution of price changes. Second, we derive a sufficient statistic for the aggregate effect of a monetary shock: given an arbitrary generalized hazard function, the cumulative impulse response of output to a once-and-for-all monetary shock is proportional to the ratio of the kurtosis of the steady-state distribution of price changes over the frequency of price adjustment. We prove that Calvo's model yields the upper bound and Golosov and Lucas's model the lower bound on this measure within the class of random menu cost models.

This paper studies the role of exchange rate regimes in shaping the distributional effects of external monetary shocks. I model a small open economy where agents differ in wealth and in exposure to international trade, producing either tradable or non-tradable goods. The central bank responds to a foreign interest rate hike by a monetary tightening to stabilize the exchange rate or lets the currency depreciate, keeping the interest rate low. I find that exchange rate flexibility distributes consumption gains to the poorer agents. The monetary tightening required to stabilize the currency disproportionately affects their disposable income through interest payments on loans and falling wages. Attempts to fix the exchange rate increase consumption inequality. Flexibility also benefits the non-tradable sector because conditions in this sector are more sensitive to domestic demand and sharply deteriorate when domestic interest rates rise.

This paper studies macroprudential policy for dollarization of domestic financial flows. We model a small open economy with entrepreneurs and workers who can save and borrow in domestic currency and in dollars. Entrepreneurs face a borrowing limit denominated in domestic currency, which makes dollar debt on their balance sheets especially disruptive in times of exchange rate depreciation. Falling output causes additional depreciation, creating a debt-deflation spiral that provides a rationale for de-dollarization. On the other hand, much of this dollar debt constitutes savings of domestic workers and provides them with insurance against depreciation events. We characterize social marginal benefits and costs of de-dollarization under this trade-off. We find that social marginal costs are associated with a deterioration in risk-sharing and can be expressed in terms of the interest rate premium on the dollar assets. These costs are of second order around the allocation without intervention.