1 of 28

Finance and �Decreasing Business Dynamics:� �Using Macroeconomic Agent-Based Simulation

Tae-Sub Yun

Department of Industrial and Systems Engineering

KAIST

1

2 of 28

Contents outline

2

Introduction

Motivation from empirical evidences
Research questions

Model description

Brief overview of K+S model
Model modification points

Result

Policy experiment results
Summary of policy implication

Limitations and further works

3 of 28

Motivation – Changing perspectives on finance

Traditional perspectives on Finance

An economy needs intermediation to match borrowers and lenders, channeling resources to their most efficient uses.
Past studies provide empirical evidence that finance is good for the economy:

Economic growth (Goldsmith, 1969; McKinnon, 1973; Shaw, 1973)
Economic stability (Acemoglu and Zilibotti, 1997; Aghion et al., 1999; Caballero and Krishnamurty, 2001)
Economic inequality (Clarke et al., 2006; Beck et al., 2007)

3

Perspectives on Finance after the financial crisis: “Too much finance”

Excessive financing has a negative effect on �economic growth. �(Cecchetti and Kharroubi, 2012; Arcand et al., 2015)
Non-linear relationship exists between �financial development and economic volatility�(Easterly et al., 2000; Dalba-Norris and Srivisal, 2013;�Sahay et al., 2015)

Economic growth rate according to private sector credit

4 of 28

Motivation – Trends in business dynamism

4

Definition of business dynamism :

The rates at which firms enter the market, grow, and leave the market.
Related to higher rates of productivity growth, as unproductive firms leave and more productive firms enter. (Bartelsman and Doms, 2000)
Market concentration and labor share are also related to business dynamics.

Source: Karabarbounis and Neiman (2014) [36]

Labor share of GDP:�GDP = Labor share + Capital share + Profit share

Source: Autor et al. (2020) [35]

5 of 28

Motivation – Trends in economic inequality

5

Income inequality in the U.S. (Piketty and Saez, 2003)

Thomas Piketty (1971 ~ )

“These works have led to radically question the optimistic relationship between development and inequality posited by Kuznets, and to emphasize the role of political and fiscal institutions in the historical evolution of income and wealth distribution.”

� - Short bio on Piketty’s CV

Income inequality in developed countries

“Time” no longer resolves inequality

6 of 28

Research questions

The financial system is one of the economic systems that affects economic inequality.

6

Research Questions

RQ1

Does financial quantitative easing induce economic inequality?

RQ2

Does financial quantitative regulation policy help reduce economic inequality?

RQ3

Are there any side effects of financial quantitative regulation?

RQ 1,2,3

Housing finance

Corporate finance

RQ 1,2,3

Wealth inequality

Income inequality

Housing market

Macro-economic

Chapter 1.

Chapter 2.

Today’s presentation

7 of 28

Schumpeter meeting Keynes model

7

Base model: Macro-economic agent-based model with financial sector (K+S model)

Capital-good firm agent: making decisions related to technology innovation.
Consumption-good firm agent: making decisions related to investment and production.
Household agent: make decisions related to consumption.
Bank agent: making decisions related to corporate financing and market liquidity.

Consumption-good�firms

wages

consumption

Households

Capital-good�firms

wages

consumption

Banks

Schumpeter meeting Keynes model

Dosi et al. (2010,2013)

Model limitation (inequality perspective)

Homogeneous household agents:�Labor productivity and wage of all households are the same.
Fixed proportional assumption in production:�A fixed number of capital goods and labor combine to produce consumption goods.

8 of 28

Model assumptions

8

Macroeconomic ABMs	Consumption good production		Consumption good pricing		Endogenous�labor share
Macroeconomic ABMs	Fixed proportional	Cobb-Douglas function	Fixed mark-up	Adjusted mark-up	Endogenous�labor share
AGH (2017, JEBO)	✔		✔		N
CATS (2011, JEDC)	✔			✔	N
Eurace (2018, JEDC)	✔			✔	N
K+S (2013, JEDC)	✔			✔	N
LEN (2013, JEBO)	✔			✔	N
EUGE (2018, JEBO)		✔	✔		Y
Proposed model		✔		✔	Y

Comparison with existing macroeconomic ABMs

9 of 28

Model validation: Output validation

9

(iteration # = 20)

Modeling goal of macroeconomic ABMs:

Dynamics of aggregated economic variables, such as GDP, unemployment rate, etc.�(Combining capital and labor to produce final goods.)

Output validation comparing with historical data of Korean macro economy.

10 of 28

Finance and market concentration

10

Type	Name	Description	Range
Dependent variables	HHI index	Herfindahl-Hirschman Index	-
Independent variables	DSR	Debt-to-Sales ratio regulation limit	[0.5, 10]
	Interest rate	Interest rate of corporate finance	[0.01, 0.1]
Experimental settings	Policy case number	The number of policy cases	200
	Replication number	The number of repeat experiments for each policy case	20

The effect of changes in the quantity and cost of finance on market concentration

In the early stage of development, finance encourages market competition, �but after a certain level, it rather increases market polarization.

11 of 28

Finance and labor share

11

Type	Name	Description	Range
Dependent variables	Labor share	Labor share	-
Independent variables	DSR	Debt-to-Sales ratio regulation limit	[0.5, 10]
	Interest rate	Interest rate of corporate finance	[0.01, 0.1]
Experimental settings	Policy case number	The number of policy cases	200
	Replication number	The number of repeat experiments for each policy case	20

The effect of changes in the quantity and cost of finance on labor share

Labor share and market concentration change in the opposite directions.

RA1: Corporate finance exacerbates income inequality.

12 of 28

Firm-level debt distribution analysis

12

Comparison of individual firms’ debt amount between the DSR scenario

In scenarios with a DSR of 2 or 4, all firms receive relatively similar debt.
In scenarios with a DSR of 8, some firms with lower market share receive no debt.
The bank is more favorable towards larger firms.

RA2: DSR regulation mitigates market concentration and further income inequality.� (When the regulation is tight, the benefit of finance is distributed relatively equally.)

13 of 28

Finance and economic growth

13

Type	Name	Description	Range
Dependent variables	Real GDP growth rate	Real GDP growth compared to the previous quarter	-
Independent variables	DSR	Debt-to-Sales ratio regulation limit	[0.5, 10]
	Interest rate	Interest rate of corporate finance	[0.01, 0.1]
Experimental settings	Policy case number	The number of policy cases	200
	Replication number	The number of repeat experiments for each policy case	20

The effect of changes in the quantity and cost of finance on economic growth

Improved access to financial services have a positive impact on the economic growth.��RA3: Financial regulation harms economic growth.

14 of 28

Summary of macroeconomic research

14

Findings for RQ 1 (finance and income inequality)

Finance influences not only the distribution of market share among firms, but also the distribution among production factors.
Corporate finance exacerbates income inequality.

Findings for RQ 2 (financial regulation)

Adequate level of DSR regulation can make finance more favorable towards small firms and new entrants.
DSR regulation lowers market concentration and raises the labor share.

Findings for RQ 3 (side effect)

Financial regulation slows economic growth.

Innate preference of financial institutions

Raising �market entry barrier

Aggravation of �market concentration

Lowering labor share

Worsening

income inequality

Too relaxed�DSR regulation

Empirical Correlation

15 of 28

Model limitation – R&D modeling

15

R&D investment

16 of 28

Further works

16

Where are we now?

Model should be more close to reality!

Data-driven simulation: initial agent state distribution and parameters
Model fitting: focusing on the latest

17 of 28

APPENDIX

17

18 of 28

Indicator of finance at the policy level

18

	Domain	Key agent	Financial policy	Micro-interaction	Macro-pattern
Chapter 2.	Housing market	Household agent	LTV, DTI, �interest rate	Housing transaction	Market index, wealth distribution
Chapter 3.	Factor market, product market, and labor market	Firm agent	DSR, interest rate	Investment & production	Macro-economic index, labor share

Proposed ABM models

19 of 28

The capital-good industry

19

Capital-good�firms

* Survey data evidence summarized in Fabiani et al. (2006) show that European firms mostly set price according to mark-up rules.

related to machine quality

related to machine�price competitiveness

20 of 28

Capital-good firms investment (1)

20

21 of 28

Capital-good firms investment (2)

21

22 of 28

The consumption-good industry

22

Consumption-good�firms

* In Dosi et al. (2006), the authors check the robustness of the simulation results employing more sophisticated expectation formation rules.

23 of 28

Consumption-good firms investment

23

* In Dosi et al. (2006), the authors check the robustness of the simulation results employing more sophisticated expectation formation rules.

24 of 28

Consumption-good market (1)

24

* This is close to the spirit of “customer market” models originated by the seminal work of Phelps and Winter (1970)

Labor share:�proxy of the income inequality

25 of 28

Consumption-good market (2)

25

26 of 28

Labor market and consumption

26

27 of 28

Commercial bank and credit multiplier

27

Debt-to-Sales ratio�(Macro-prudential)

Monetary policy

28 of 28

Macroeconomic model input parameters

28