- Armantier, Olivier, James McAndrews, and Jeffrey Arnold. 2008. “Changes in the Timing Distribution of Fedwire Funds Transfers.” FRBNY Economic Policy Review 14(2): 83–112. [url].
- Soramäki, Kimmo, Morten L Bech, Jeffrey Arnold, Robert J Glass, and Walter E Beyeler. 2007. “The topology of interbank payment flows.” Physica A: Statistical Mechanics and its Applications 379(1): 317–333. url
Financial Markets and the Onset of the American Civil War
I estimate the market implied ex ante probability of the onset of the American Civil War using U.S. government and state bonds. Surprisingly, financial markets were surprised by the Battle of Fort Sumter and the start of the war. Prior to Abraham Lincoln’s election in November 1860, the market assigned almost no probability to a war. Even after the secession of several states, the week before the Battle of Fort Sumter, the market assigned a negligible probability, approximately 5%, to war onset.
Pricing the Costly Lottery:
Financial Market Reactions to Battlefield Events in the American Civil War
This paper estimates the effects of major battles in the American Civil War on the price of Union bonds. Bond prices are a function of expected future cash flows, and war results can have a large influence on the the expectation of the timely payment of those cash flows. Thus, bond prices provide a method to assess the effects of war events on the expected war result. This allows for understanding how events within war influenced the expected war result. Assessing the importance of war events on war termination provides another method for researchers to “open up the black box of war”. In this application, initial results suggest that for the Union bonds considered, the average Confederate victory decreased the price by 3.3% Union victory increased it by 1%.
Bayesian Shrinkage in Dynamic Linear Models
Unifying Smoothing and Structural Break Models
This develops a method for estimating time varying parameters that unifies change-point and smoothing models.
It unifies these approaches by noting that change point models are simply a case of a sparse parameter model in which the sparsity is over the difference rather than the level of the parameters.
The method presented in the essay consists of combining dynamic linear models with newly developed Bayesian shrinkage prior distributions.
This method has several advantages over existing change point models.
First, it does not require choosing a specific number of change points.
In fact, it does not require the researcher to choose between smoothly varying and change point models for the parameter of interest since the sparsity of parameter changes can itself be estimated.
Second, This method is extremely flexible, and can be extended to parameters in a variety of regression and time-series models.
Third, it can also be used to model change points in multiple parameters that could be either independent or correlated.
Fourth, since it is a special case of dynamic linear models, computationally efficient methods exist to estimate and sample from it.
While the computationally efficient methods to estimate dynamic linear models are well known, they are not included in general purpose Bayesian software.
In this essay, I provide implementations of these methods in
Stan, which can be found on github.
Civil War as a Hidden Markov Model
This paper estimates a model of civil war prevalence with a hidden Markov model (HMM) to account for measurement error in classifying civil wars. Since there exist multiple coding rules for civil war with substantial disagreement, each of these coding rules is treated as a noisy indicator of a latent concept. A HMM is used to classify country-years into latent states of civil war and peace and to model the probability of civil war onset and continuation with these latent states. The estimated classification supports an inclusive concept of civil war with a low threshold for what constitutes a civil war. The only statistically significant covariates in the civil war transition models are population and GDP on civil war onset.
Estimating the Offense-Defense Balance from the Production Function of Battle Casualties
250 years ago wars were fought with muskets. Today they are fought with missiles. Yet most empirical and theoretical literature has treated war as a homogeneous process across time. This paper develops a theoretical and quantitative apparatus to describe a key source of heterogeneity in wars, the technology of combat. The technology of combat is the shape of the function that takes military forces as inputs and produces casualties as its outputs, i.e. the production function of war. Using a Bayesian hierarchical model, I estimate the coefficients of this function at the war level. These estimates provide theoretically derived empirical estimates of the offense-defense balance and the relationship between the quantity and quality of forces over time and across wars
With Hein Goemans
We develop a measure of the historical reputation of leaders‘ using their Wikipedia pages.