📈 economics

Monetary Policy, Exchange Rate Shocks, and Asset Prices in a Small Open Frontier Economy: A Restricted-Sample, Multi-Horizon Distributional Analysis

By employing a rigorous restricted-sample design and advanced multi-horizon distributional methods to eliminate interpolation artifacts and control for false discoveries, this study finds that Ghana's equity market exhibits only a loose and uneven coupling to domestic monetary policy, with global commodity prices and short-term funding costs playing a more significant, state-dependent role in asset price dynamics.

Original authors: Isaac Nyame, Gabriel Osei Forkuo

Published 2026-10-06

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Isaac Nyame, Gabriel Osei Forkuo

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of finance, central banks act as the thermostat for an economy, turning interest rates up to cool down inflation or down to warm up growth. In wealthy nations with deep, liquid stock markets, this thermostat works like a well-oiled machine: when the central bank changes the rate, investors quickly adjust the prices of their stocks, sending a clear signal through the financial system. This is the standard story of monetary policy transmission. But in frontier markets like Ghana, where the stock exchange is small, thinly traded, and dominated by a handful of large companies, the story is far less certain. Here, the connection between the central bank's decisions and the price of shares is often murky, tangled with currency fluctuations, global commodity prices, and the unique quirks of a market that lacks the depth to absorb shocks smoothly. Understanding whether the central bank's policy rate actually moves the stock market in these fragile economies is not just an academic exercise; it determines whether policymakers can rely on stock prices as a barometer of their success and whether investors can trust that rate changes will protect or punish their portfolios.

A team of researchers set out to untangle this specific knot in Ghana, a country that has experienced wild swings in interest rates, currency values, and inflation over the last decade. Their goal was to see if changes in the Bank of Ghana's policy rate actually cause changes in the value of stocks listed on the Ghana Stock Exchange. To do this, they had to be exceptionally careful about the data they used. The Ghana Stock Exchange recently introduced new ways of calculating its main stock indices, and for several years, the data for one of these indices was not directly reported but rather filled in using mathematical guesses, or interpolation. The researchers realized that using these guessed numbers would create a false sense of precision, making the market look smoother and more predictable than it really was. So, they made a bold methodological choice: they threw out four years of data for the financial stock index that relied on these interpolations. They worked only with the months where the closing prices were independently verified and actually observed, even though this meant working with a much shorter timeline. This decision was crucial, as it ensured their conclusions were built on real market movements rather than mathematical artifacts.

Using this cleaned-up data, the researchers applied a suite of sophisticated statistical tools designed to look at the market from every angle. They tested whether past changes in interest rates could predict future stock prices, examined how the market reacted during calm times versus turbulent ones, and measured how much of the stock market's movement could be blamed on the central bank compared to other factors like oil prices or the exchange rate. The results painted a picture that was far more modest than the standard theory suggests. The study found no convincing evidence that the central bank's policy rate has a direct, causal link to the overall stock market index. When the researchers corrected for the fact that they were running many different tests at once—a step necessary to avoid finding patterns that are just random noise—the link between the policy rate and stock returns vanished completely. The policy rate explained less than four percent of the changes in the main stock index over a year, a figure so small it suggests the central bank's thermostat is barely touching the market.

Instead of the policy rate, the study pointed to other forces as the true drivers of the market. For the broad stock index, global oil prices emerged as a powerful influence, but only when the market was doing well. When the Ghanaian market was in a bullish, optimistic phase, rising oil prices were strongly associated with higher stock returns, likely because Ghana is an oil producer and higher prices boost confidence in the national economy. However, this connection disappeared when the market was struggling. For the financial sector index, which is made up mostly of banks, the story was different. Here, the cost of short-term borrowing, measured by the Treasury bill rate, seemed to matter, but only when the market was in a bearish, stressed state. This suggests that when banks are under pressure, the cost of their immediate funding becomes a critical factor, a nuance that would be missed by looking only at average market behavior.

The researchers also looked at whether the relationship between interest rates and stocks changed over time, specifically whether the connection grew stronger the further out one looked. Earlier studies had suggested that the effect of monetary policy might take a long time to fully play out. However, when this team tested that idea against a statistical baseline that accounted for the natural randomness of the data, the pattern disappeared. The apparent strengthening of the link over time was revealed to be an illusion created by the way the data was smoothed out, not a real economic phenomenon. The market's reaction to policy shocks remained weak and inconsistent, regardless of whether the researchers looked at the short term or the long term.

These findings carry significant weight for anyone involved in Ghana's economy. For the central bank, it suggests that using the stock market as a real-time signal of how their policies are working is unreliable; the market is too thin and too influenced by outside factors like global oil prices to reflect the central bank's moves clearly. For investors, the study highlights that in frontier markets, the risks are not uniform. The danger of rising interest rates for financial stocks is concentrated in times of stress, while the benefits of global commodity booms are concentrated in times of optimism. The research underscores that in small, fragile markets, the simple rules that apply in wealthy economies often do not hold. The central bank's policy rate is not the master switch for stock prices; instead, the market is a patchwork of different reactions to global commodities, currency values, and short-term funding costs, all operating in a way that is highly dependent on the specific state of the economy at that moment.

Technical Summary: Monetary Policy, Exchange Rate Shocks, and Asset Prices in a Small Open Frontier Economy

Problem Statement and Context
This study investigates the transmission of monetary policy shocks to equity returns in Ghana, a frontier market characterized by shallow liquidity, a concentrated investor base, and structural discontinuities in index construction. While monetary policy transmission to asset prices is well-established in advanced economies via discount-rate and expected-earnings channels, its operation in frontier markets remains empirically unsettled. The authors argue that prior literature on Ghana has been limited by dated evidence, a focus on growth rather than monetary transmission, and, critically, the use of index data that includes periods of interpolated (constructed) values rather than observed market data. Specifically, the Ghana Stock Exchange Financial Stock Index (GSE-FSI) lacked monthly observations for 2014–2017 in official reports, leading earlier studies to rely on log-linear interpolation, which renders implied monthly returns deterministic and statistically invalid for causal inference.

Methodological Framework
The paper employs a rigorous, multi-method econometric approach designed to address data integrity and model uncertainty:

Restricted-Sample Design: A primary methodological contribution is the exclusion of 48 months (January 2014–December 2017) from the GSE-FSI analysis where data were interpolated. The GSE-CI analysis uses the full 2014–2024 window ( $N=131$ returns), while the GSE-FSI analysis is restricted to February 2018–December 2024 ( $N=83$ returns), ensuring all levels are independently verified.
Unit Root and Structural Break Testing: Series are tested for integration using the Augmented Dickey-Fuller (ADF) test with a constant-plus-trend specification and the Zivot-Andrews (ZA) test. A single, pre-specified classification rule is applied: a series is $I(1)$ if the ADF test fails to reject a unit root in levels but rejects it in first differences.
Causality Testing: Granger causality is assessed using the Toda-Yamamoto augmented-lag procedure. Crucially, the authors apply the Benjamini-Hochberg procedure to control the false discovery rate (FDR) across the 24 bivariate tests conducted, addressing the multiple-testing problem often overlooked in this literature.
Bayesian Quantile Regression (BQR): To capture distributional effects, the authors estimate BQR with an embedded spike-and-slab stochastic search variable selection (SSVS) prior. This allows for variable selection and coefficient estimation within the same Gibbs sampler, providing posterior inclusion probabilities (PIP) for predictors at quantiles $\tau = 0.10, 0.25, 0.50, 0.75, 0.90$ . The Treasury bill rate (TBR) is winsorized to mitigate the influence of extreme observations during the 2022–2023 debt restructuring.
Structural VAR Analysis: A six-variable Vector Autoregression (VAR) is estimated with a corrected recursive ordering (Oil $\to$ Inflation $\to$ Policy Rate $\to$ TBR $\to$ Exchange Rate $\to$ Equity). Impulse response functions (IRFs) and forecast error variance decompositions (FEVD) are computed, alongside Generalized IRFs (GIRF) to ensure order invariance.
Wavelet Coherence: Continuous Morlet wavelet coherence is used to analyze time-frequency co-movement. Raw coherence is benchmarked against a Monte Carlo red-noise null (150 AR(1) surrogate pairs) to distinguish genuine horizon-dependence from artifacts of the smoothing operator.

Key Results

Monetary Policy Rate (MPR): The domestic policy rate shows no causal relationship with either the GSE-CI or GSE-FSI that survives multiple-testing correction. In Bayesian quantile regressions, the posterior inclusion probability (PIP) for the MPR never exceeds 0.23 for either index, indicating weak evidence for its inclusion as a predictor across the entire return distribution. In the VAR framework, MPR shocks account for only 0.5% to 3.1% of GSE-CI forecast error variance and 0.7% to 7.4% of GSE-FSI variance at a 12-month horizon.
Money Supply and Exchange Rates: Money supply growth (M2) is the only macro-financial variable showing a causal link to GSE-CI that survives FDR correction, though the authors note this may reflect shared trend dynamics rather than pure causality. For GSE-FSI, exchange rate shocks show a significant causal link to equity returns in the restricted sample, a finding not detectable in earlier interpolated-data drafts.
Oil Prices: Oil price shocks exhibit strong posterior support for GSE-CI at upper (bullish) quantiles ( $\tau=0.75$ and $0.90$), with PIPs of 0.98 and 0.94 respectively. This suggests global commodity prices drive bullish market states more than the policy rate, consistent with Ghana's status as an oil producer.
Treasury Bill Rates: For GSE-FSI, TBR changes show "weak-to-moderate" posterior support (PIP = 0.64) at the 10th percentile (bearish states), suggesting a funding-cost channel concentrated in financial-sector stress periods. However, this estimate relies on a small effective sample size (~8 observations) and is presented as suggestive rather than conclusive.
Wavelet Coherence: While raw wavelet coherence between policy shocks and equity returns appears to rise with horizon, benchmarking against the red-noise null reveals that the statistically significant share of coherence is low (3.4% to 7.4%) and non-monotonic. This indicates that the apparent horizon-dependence is largely an artifact of the estimator's smoothing properties rather than a substantive economic finding.

Significance and Claims
The paper claims a "modest" but methodologically significant contribution. It demonstrates that in thin, structurally discontinuous frontier markets like Ghana, the domestic policy rate is loosely and unevenly coupled to equity returns. The study argues that claims about monetary transmission in such markets require index-construction safeguards (excluding interpolated data) and multiple-testing corrections that are not standard in the existing literature.

The authors conclude that:

Policy Rate Limitations: The policy rate is not a reliable lever for managing equity market conditions in Ghana, nor should equity movements be viewed as a real-time signal of policy transmission.
Alternative Drivers: Global commodity prices (specifically oil) and short-term funding costs (Treasury bill rates) play larger, more state-dependent roles in driving equity returns than the headline policy rate.
Methodological Necessity: The use of interpolated data and the failure to correct for multiple testing can lead to spurious findings of causality and horizon-dependence. The corrected results suggest that monetary transmission in frontier markets is weaker and less systematic than often assumed in advanced-economy models.

The study does not claim to have identified a dominant transmission channel but rather to have rigorously ruled out the standard discount-rate channel as a primary driver, pointing instead to liquidity and currency conditions as more proximate factors.

More like this