Research topics in Computational
Economics,Finance, and Business related
Eojin Kim
November 2024
1 Introduction
In this document, some interesting data/computational science related research
in Econ, Finance, and Business are listed.
2 Gravity equation
Gravity equation in economics presents an interesting system that has a lot of
its equation property similar to dynamical equation governing Physical systems.
Basic model for trade between two countries i and j takes the form of
Fij = G · Mi Mj /Dij
where G is constant like gravity F is trade flow D is distance between the coun-
try and M stands for economic dimensions of the countries that are being
meaasured.
At the end, this looks just like the gravitational potential/force that exists in
physics. Even though it is in economics context, it’d become a system just like N-
body simulation in physics. There are many questions regarding whether
same properties that are held for physics gravity holds the same for
gravity equation in economics For example, what are the conditions in
Economics trade where the system behaves just like Hamiltonian sys-
tem in Physics? Does Sensitivity of various numerical methods that
are observed in Physics N body simulation carry over for Economics
gravity equation modeling?
3 Dynamic Stochastic General Equilibrium model
DSGE model in macroeconomics presents another interesting research topic.
Let’s at first address three basic properties. Firms maximize profits by de-
manding capital and labor and supplying output. Firms take the price
1
of capital (r) and the price of labor (w) as outside their control. Sec-
ond, Households maximize utility subject to their budget constraint
by demanding output (to be split into consumption and savings) and
supplying labor and capital. Households take the price of capital (r)
and the price of labor (w) as given and outside their control. Lastly,
An equilibrium is a set of prices, r and w, such that firms maximize
profits, households maximize utility, and markets clear: Output sup-
plied by firms is equal to output demanded by households, and capital
and labor demanded by firms is equal to capital and labor supplied
by households.
Let’s briefly discuss sources of fluctuations. Policy is one of them. From classical
view, monetary policy is ineffective. Fiscal policy through government spend-
ing and tax changes affects both labor supply and business cycles. There is
also Keynesian view where assumption is prices do not adjust in short-run.
And, rather Business cycles can be caused by monetary shocks. Ito calculus
maybe more releveant for this kind of scenario.
3.1 Classical version growth model
Lt = (1 + n)Lt−1
zt = (1 + g)zt−1
L and z represent population growth and technology growth. Technology is
considered to be stochastic. Problem facing household is
∞
X
maxEo β t u(ct , 1 − lt )Lt
t=0
and this is subject to
Ct + At+1 = Lt ŵt lt + At (1 + rˆt ) − Tt
T represents lump-sum taxes by government. we can express problem facing
firm as
maxYt − wt Nt − rt Kt
where
Yt = Ktα (zt Nt )1−α
where α is between zero and one.
Our market clearing conditions are Labor market: Nt = Lt lt and Capi-
tal/asset market: Kt = At Lastly goods market
Xt = Yt − Ct − Tt + Tt − Gt = Yt − Ct − Gt
Xt = Kt+1 − (1 − δ)Kt
is investment.
Before we move onto next section, applying different condition for some of the
variables above are definitely of interest.
2
3.2 Adding Uncertainty
Lets now add assumption technology and goverment purchases are stochastic.
Such that
lnzt = lnzo + gt + z˜t
z˜t = ρz zt˜− 1 + ϵzt
Government purchases:
æG̃t
ln(Gt ) = lnGo + (g + n)+
˜ + ϵgt
G̃t = ρg Gt−1
First order conditions for household are
ct : [β(1 + n)]t u′ct = λt
lt : −[β(1 + n)]t u′lt = ŵλt
at+1 : λt (1 + n) = Et λt+1 (1 + r̂t+1 )
Firm’s first order condition:
yt
kt : α = rt
kt
yt
lt : (1 − α) = wt
lt
Rest of steps for Dynamica stochastic general equilibrium(DSGE) models are
finding first order condition of model, Find steady state, Linearize arond
steady state, and solve linearized system of equations There is certainly
many additional extension/questions that can be addressed further regarding
DSGE models including incorportating Quasi-linearity into the system of equa-
tion, extending its linearity with Cumulant expansion form and truncating at
certain higher order instead.
Also there is question of how those systems of equations are sensitive to choice
of some of the developed Numerical Methods/Scientific Computing techniques
such as Finite Element Method and Chebyshev Spectral Methods.
4 Cox-Ingersoll-Ross model
This is a model that describes evolution of interest rates.
It describes instantaneous interest rate rt with Feller square-root process and
its equation result in
√
drt = a(b − rt )dt + σ rt dWt
where Wt represents Wiener process.
a is speed of adjustment to mean b . σ is volatility. Overall, a(b − rt ) represents
drift factor of interest rate towrad long run value b Research around simulation
of this system is interesting topic.
3
5 Hull-White model
This is model described to model future interest rates.
5.1 one factor or short-rate model
Its governing equation looks like
dr(t) = [θ(t)α(t)r(t)]dt + σ(t)dW (t)
when θ has time dependence it is called Hull-White model.
When both α and θ has time dependence, it is often called extended vasicek
model
5.2 Two-factor model
There is extension of one factor model with additional disturbance term whose
mean is zero.
df (r(t)) = [θ(t) + u − α(t)f (r(t))]dt + σ1 (t)dW1 (t)
f is deterministic function of r. And u has initial value of zero and varies in time
governed by
du = −budt + σ2 dW2 (t)
There are many interesting research topics to explore around the model frame-
work of Hull-White.
6 Chen(three factor) model
You can vie this model as further extension of models from previous section. It
is viewed as three factor model with equations Equations are written as
√ √
drt = κ(θt − rt )dt + rt σt dW1
p
dθt = ν(γ − θt )dt + α θt dW2
√
dσt = µ(β − σt )dt + η σt dW3
This model can be extended even further. There is many interesting research
topics around this model including model’s feasibility for interest rate deriva-
tives.
