Heterogeneous Agent Model Research Articles

DeepHAM: A Global Solution Method for Heterogeneous Agent Models with Aggregate Shocks

DeepHAM: A Global Solution Method for Heterogeneous Agent Models with Aggregate Shocks

Branch Expansion versus Digital Banking: The Dynamics of Growth and Inequality in a Spatial Equilibrium Model

We develop a heterogeneous-agent model with local spatial markets to study the relationships among bank expansion, growth, and inequality. In the model, households choose their occupations, consumption, and holdings of loans and portfolio assets that vary by liquidity. Banks choose the locations of new branches, which affect the financial frictions facing households across regions. We calibrate the model using a geographic information system to evaluate the rapid bank expansion in Thailand between 1986-1996. The model quantifies the sources of growth and inequality over time and across space and the potential role of digital banking in substantially reducing regional heterogeneity.

Universal Basic Income in Developing Countries: Pitfalls and Alternatives

We study the macroeconomic and social effects of Universal Basic Income (UBI) programs in a developing economy, comparing them with policies that condition cash transfers on household characteristics (CCT). We construct a dynastic heterogeneous-agent model with human capital investment and choice of labor effort and calibrate it to Brazilian data. In the short run, UBI alleviates poverty and increases the welfare of the poor. Over time, however, income falls and poverty and inequality increase since investments in physical and human capital decrease along with labor supply. In most dimensions, CCT programs outperform UBI policies, largely due to school enrollment requirements.

Status Externalities and Low Birth Rates in Korea

East Asians, especially South Koreans, appear to be preoccupied with their offspring's education—most children spend time in expensive private institutes and in cram schools in the evenings and on weekends. At the same time, South Korea currently has the lowest total fertility rate in the world. In this paper, we propose a theory with status externalities and endogenous fertility that connects these two facts. Using a quantitative heterogeneous-agent model calibrated to Korea, we find that fertility would be 16% higher in the absence of the status externality. Furthermore, childlessness in the poorest quintile would fall from five to less than one percent. We then explore the effects of various government policies. A pro-natal transfer increases fertility and reduces education while an education tax reduces both education and fertility, with heterogeneous effects across the income distribution. The policy mix that maximizes the current generation's welfare consists of an education tax of 12% and moderate pro-natal transfers—a monthly child allowance of 3% of average income for 18 years. This would raise average fertility by about 5% and decrease education spending by 16%. Although this policy increases the welfare of the current generation, it may not do the same for future generations as it lowers their human capital.

Common Idiosyncratic Volatility and Carry Trade Returns

I provide new evidence that incomplete consumption risk sharing across countries is an important determinant of carry trade returns. I show that there is a strong co-movement in idiosyncratic volatilities over time, and that shocks to the common idiosyncratic volatility (CIV) factor, defined as the equally weighted average of the idiosyncratic volatilities in the cross-section, are priced. I find that high-interest rate currencies deliver low returns when the CIV increases, which are bad times for investors. Low-interest rate currencies provide a hedge by yielding positive returns. CIV shocks remain an empirically powerful risk factor in explaining the cross-section of carry trade returns after controlling for global foreign exchange (FX) volatility risk. Furthermore, CIV risk is correlated with cross-country income risk faced by households. My findings are consistent with a heterogeneous-agent model with persistent, uninsurable idiosyncratic shocks in consumption growth. The calibrated model quantitatively accounts for the cross-sectional differences in average returns across CIV-beta sorted portfolios for plausible market prices of CIV risk.

Labor Market Policies in a Dual Economy

A structural model of heterogeneous agents is built to account for the labor market dynamics of an economy with a large informal sector and to quantify effects of labor market policies on employment, worker flows, savings and welfare in a dual economy. The introduction of unemployment insurance has only a small impact on unemployment but induces a sectoral reallocation of formal labor into informality. Generous severance payments lower the wage of formal jobs and reduce flows from unemployment to formality. In financing expenditures, shifting a tax burden from labor income to consumption increases the size of a formal sector, enhancing productivity and welfare. Economic and welfare consequences of policy reform can significantly differ from those in a single-sector economy due to workers’ incentives to avoid taxes and maximize transfers by moving across sectors.

Who suffers from the COVID-19 shocks? Labor market heterogeneity and welfare consequences in Japan

Effects of the COVID-19 shocks in the Japanese labor market vary across workers of different age groups, genders, employment types, education levels, occupations, and industries. We document heterogeneous changes in employment and earnings in response to the COVID-19 shocks, observed in various data sources during the initial months after the onset of the pandemic in Japan. We then feed these shocks into a life-cycle model of heterogeneous agents to quantify welfare consequences of the COVID-19 shocks. In each dimension of the heterogeneity, the shocks are amplified for those who earned less prior to the crisis. Contingent workers are hit harder than regular workers, younger workers than older workers, females than males, and workers engaged in social and non-flexible jobs than those in ordinary and flexible jobs. The most severely hurt by the COVID-19 shocks has been a group of female, contingent, low-skilled workers, engaged in social and non-flexible jobs and without a spouse of a different group.

Consumption dynamics under time-varying unemployment risk

In response to an adverse labor-market shock, a calibrated heterogeneous-agent model predicts that aggregate spending on durable goods falls mainly due to the ex-ante increase in income uncertainty caused by higher unemployment risk. In contrast, aggregate spending on nondurable goods falls mainly due to the ex-post income losses associated with realized unemployment spells. When households hold little liquid assets, the nondurable spending response is amplified, whereas the durable spending response is dampened. These differences stem from micro-level adjustment frictions involved in purchases of durable goods. The model is corroborated with evidence from micro survey data.

ENTREPRENEURSHIP AND INCOME DISTRIBUTION DYNAMICS: WHY IS THE INCOME SHARE OF TOP INCOME EARNERS ACYCLICAL OVER THE BUSINESS CYCLE?

AbstractIn the United States, the income share of the top 5% income group is acyclical over the business cycle. This study attempts to explain the cyclical behavior of the income distribution over the business cycle, particularly focusing on the top 5% income earners' share, using a heterogeneous agent model featuring a choice to become an entrepreneur. The model economy successfully reproduces the acyclical behavior of the income share of the top 5%. During expansions, relatively more people become entrepreneurs at the top, which offsets the decline in the income share of the high‐income earners from workers' side.

Nonlinear Dynamic Analysis of New Product Diffusion considering Consumer Heterogeneity

When a new product enters the market, individual consumers’ decision-making behavior and purchase time are uncertain. Based on the dynamics of epidemic transmission theory and agent modeling technology, this study proposes a new coupling model through the combination of the improved SEIR epidemic model and the heterogeneous agent model. This model considers consumer heterogeneity resulting from three aspects in consumers’ sensitivity, network topology, and considerations of information flow received. It aims to analyze how consumer heterogeneity affects the scale and speed of new product diffusion. The proposed model showed that consumers’ characteristics and behavior combination at the microlevel lead to the diversity of nonlinear diffusion curves at the macrolevel for new products. Moreover, a pilot study is conducted to simulate this model and examine how to estimate the model’s parameters using aggregated data about film products. The pilot study results suggested that different consumer characteristics and behavior combinations affect the scale and speed of new product diffusion to varying degrees. In different scenarios, there were significant differences in the influence of the degree of consumer heterogeneity on diffusion, accompanied by the occurrence of threshold. The results of the empirical analysis in this study are in line with reality.

Behavioral heterogeneity in return expectations across equity style portfolios

AbstractWe estimate a heterogeneous agent model on five prominent equity investment styles—value, size, profitability, investment, and momentum—and find evidence for behavioral heterogeneity in expected return formation. Our model features two groups of boundedly rational investors, fundamentalists and chartists, whose demand functions for the investment styles depend on their respective expected style return forecasts. The fundamentalists form return expectations using a model based on time‐varying stock‐level characteristics and dynamic factor premia, and the chartists do so based on heuristics commonly employed by technical analysts, such as moving average rules. Our results cast doubt on the theories that assume perfect rationality of the representative agent in financial markets, and give support to the behavioral theories with heterogeneous agents.

A fast and low computational memory algorithm for non-stochastic simulations in heterogeneous agent models

Heterogeneous agent models in macroeconomics generally require numerical computation of the cross-sectional distribution of agents. The standard textbook approach is to fully approximate the Markov kernel that iterates the distribution forward in time as a Markov transition matrix, which can be costly in terms of computational time and memory when the state space is large. This note provides an alternative algorithm that is simple, requires much less computational memory, and is substantially faster than the standard algorithm.

Education in a Heterogeneous-Agent Economy: Revisiting Transatlantic Differences

I propose a novel heterogeneous-agent model featuring public and private education in the choice set of households, positive human capital externalities, and distortionary taxes for public education financing to study transatlantic differences in public versus private education investment, tax rates, and economic distributions. I show that exogenous tax differences alone can generate most of the observed transatlantic disparities in this model. I also demonstrate that this model can explain how either of the calibrated US and European tax regimes can gain public support due to the U shape of aggregate variables over taxes, as also seen in the data.

Do we really need heterogeneous agent models?

We investigate the existence of representative agent under various heterogeneities in a recursive utility framework. We provide the analytical solution of household allocations. We numerically explore whether we can find a representative agent whose income is the aggregate income of the society and whose allocations are the aggregate allocations of the society under heterogeneity in the parameter of risk aversion and/or parameter of intertemporal substitution, or discount rate or survival probability. We find that there is no representative agent when the heterogeneity is in survival probability. From the data, we know that there is heterogeneity among people in terms of survival probability. For instance, women tend to live longer than men, rich people tend to live longer than poor people. We conclude that it may be better to use heterogeneous agents models than the representative agent models under the recursive utility framework. Our results support the idea that we do need the heterogeneous agent models indeed.

Bitcoin: Speculative asset or innovative technology?

We unite investment by speculators and tech-savvy investors with a heterogeneous agent model. While speculators seek to profit from extrapolating the price trends, tech-savvy investors trade based on the prospective value of Bitcoin, which is a function of factors that capture the market demand and technical supply of Bitcoin. Estimating the structural model, we find the coexistence of speculators and tech-savvy investors in the Bitcoin market. We further show that, regardless of the market states, tech-savvy investors consistently buy (sell) Bitcoin when its price goes below (above) the prospective value. However, speculators follow a momentum trading strategy in the high-market-volatility regime and switch to a contrarian strategy in the low-market-volatility regime. Our finding suggests that a significant fraction of tech-savvy investors value the potential of Bitcoin as an innovative technology. Incorporating heterogeneous investors yields better in-sample estimation efficiency and out-of-sample forecasting precision than models that consider only speculators or tech-savvy investors.

BANK COMPETITION, FINANCIAL DEVELOPMENT, AND INCOME INEQUALITY

Income inequality rises with financial development initially and then drops. We reach this conclusion by numerically solving a heterogeneous agent model parameterized to the Chinese economy. The model features a banking sector with Cournot competition, and the process of financial development in the model economy begins with the deregulation of the banking sector. Based on regressions with the fixed effects and the system generalized method of moments, the empirical analysis also suggests an inverted‐U relationship between income inequality and financial development using provincial data from China. (JEL E25, G21, G28)

Entrepreneurship and Income Distribution Dynamics: Why is the Income Share of Top Income Earners Acyclical Over the Business Cycle?

The U.S. economy shows a contrasting difference in cyclical properties between low- and high-income groups. Income shares of the bottom three income quintiles are procyclical; while those of the next 35 percent are countercyclical. However, the income share of the very top five percent income group is acyclical over the business cycle. This study attempts to explain the cyclical behavior of the income distribution over the business cycle, particularly focusing on the top five percent income earners' share, using a heterogeneous agent model featuring an extensive labor supply margin and a choice to become an entrepreneur. The model economy successfully reproduces the acyclical behavior of the income share of the top five percent income earners as well as the overall cyclicality of income shares of the income groups. In an economic expansion, relatively more people in the low income groups newly begin employment as workers, which raises the income shares of the bottom three quintiles and lowers the income shares of the top two quintiles, while at the top, relatively more people become entrepreneurs, which offsets the decline in the income share of the high-income earners from the workers' side

Estimating linearized heterogeneous agent models using panel data

We develop a method to estimate heterogeneous agent models that uses not only time series of macroeconomic aggregates, but can also incorporate micro level data (repeated cross-section or panel). The micro data may be collected at lower frequency and time-aggregated. The method is based on the linearization approach of Reiter (2009), combined with optimal state aggregation as in Reiter (2010). The model may contain decision problems with both continuous and discrete choice. Linearity of the model solution allows fast computation of second moments and likelihood. We discuss various computational devices to maximize the speed of the estimation.

Discussion of “Estimating linearized heterogeneous agent models using panel data”

The techniques proposed in Papp and Reiter (2020) allow the use of cross-sectional and aggregate data observed at different frequencies in the estimation of dynamic stochastic macroeconomic models. However, the question is whether technique is getting ahead of what is sensible in terms of currently available empirical strategies to estimate macroeconomic models which are – without exception – misspecified.

How Unsecured Credit Policies Influence Mortgage and Unsecured Loan Defaults

AbstractBefore the global financial crisis, the proportion of households defaulting on the mortgage while remaining current on the unsecured loan was almost the same as the proportion of households current on the mortgage but defaulting on the unsecured loan. After the crisis, the former ratio became higher than the latter. By using a heterogeneous agent model with the mortgage and the unsecured loan, I examine how the order of defaults changed before and after the crisis. I then analyze the impacts of unsecured credit policies on households' mortgage and unsecured loan defaults. My quantitative exercise shows that both default rates can decrease as the cost for unsecured loan defaults increases.