Makola - Tools and Workflows for Market Integration Analysis

Francis Tsiboe

Acknowledgment: The views expressed in this repository are solely those of the author and do not necessarily reflect the official positions or policies of affiliated institutions.

Overview

makola provides tools, workflows, and supporting materials for market integration and related analysis. The package is designed to support empirical work on price transmission, market comovement, cointegration, threshold adjustment, and related nonlinear time-series dynamics in paired market data.

The current workflow emphasizes: - simulation of toy market price series for tool development and demonstrations, - lag-order selection and ordering of market pairs, - unit root and cointegration screening, - model selection across linear and threshold specifications, and - estimation of selected pairwise integration models.

Current scope

makola is being developed as a research-oriented toolkit for pairwise market integration analysis. At this stage, the package includes functions for:

• simulating synthetic market price series with configurable temporal frequency and numbers of integrated markets,
• selecting lag order for a pair of series,
• ordering paired series using Granger-causality evidence,
• identifying I(1)-consistent deterministic specifications,
• testing pairwise cointegration using Johansen trace procedures,
• testing threshold cointegration using the Hansen-Seo framework,
• testing linearity against threshold alternatives using Hansen-style threshold tests,
• selecting an integration-based model for a pair of series, and
• estimating linear and threshold VECM/VARDL-style models.

Main workflow

A typical workflow is:

1. Prepare or simulate a pair of market price series.
2. Select the lag order for the pair.
3. Screen for integration and cointegration properties.
4. Test for linear versus threshold adjustment.
5. Select an appropriate integration-based model.
6. Estimate the selected model, allowing fallback to simpler nested specifications when needed.

Example

toy_data <- simulate_market_prices(
  start_date = "2000-01-01",
  end_date = Sys.Date(),
  frequency = "month",
  n_markets = 2,
  n_integrated_markets = 2,
  seed = 123,
  output = "wide"
)

res <- estimate_pair_integration_model(
  pair = c("market_1", "market_2"),
  data = toy_data,
  nboot = 100,
  lag.max = 4,
  lag.opt = 2,
  trim = 0.10
)

res$model
res$est

Installation

You can install the development version from GitHub with:

# install.packages("remotes")
remotes::install_github("ftsiboe/makola")

Development notes

This repository is under active development. Function names, arguments, output structures, and estimation workflows may continue to evolve as the package is expanded and standardized.

Planned improvements include: - additional diagnostics and summary tools, - broader support for pairwise and multi-market workflows, - improved estimation wrappers and reporting utilities, and - worked examples and vignettes for empirical applications.

Contributing

Constructive feedback, bug reports, and research collaborations are highly appreciated. Contributions that improve documentation, testing, estimation workflows, and reproducible examples are especially welcome.

Please use the GitHub issue tracker to report bugs or suggest enhancements.

Contact

Francis Tsiboe
Email: ftsiboe@hotmail.com