Climate and Monetary Policy in BRICS Economies: Do Temperature Shocks Trigger Financial Instability?

2.1. Climate shocks and economic activity

Research findings show that rising temperatures produce adverse effects on economic performance, according to key empirical studies. The research by Dell et al. (2012)⁶ demonstrates that temperature increases lead to decreased output levels mainly in tropical and developing areas through their impact on labour productivity and agricultural production. The study by Burke et al. (2015)⁵ shows that the relationship between temperature and productivity is non-linear and that output losses become more severe after certain temperature limits are exceeded. This study demonstrates that temperature fluctuations cause disruptions to energy consumption and agricultural yields and industrial operations, resulting in economic instability and price increases. The research demonstrates that temperature anomalies function as external supply disturbances that modify the path of economic development.

2.2. Climate risk and financial stability

Research shows that climate risk creates financial instability through four main transmission channels, which include credit risk and insurance losses, market repricing, and macroeconomic uncertainty. The (BIS, 2020) and the (NGFS, 2020)^10,3 identify climate-related risks as fundamental threats that can cause financial institution failures and make standard monetary policy instruments useless. Battiston et al. (2017) and Dafermos et al. (2020)^8,9 show that climate shocks spread through connected financial balance sheets, resulting in non-linear growth patterns in the financial system. The research shows environmental disturbances create systemic risk because they activate risk transmission through market-based expectations and financial exposures that affect financial markets.

2.3. Monetary policy and climate transmission

The traditional monetary systems that use exogenous output and inflation shocks do not account for how climate variability impacts economic operations. Research findings from recent times have started to question this established belief system. Campiglio et al. (2018)¹³ demonstrates how climate shocks affect the Phillips curve relationship and monetary policy transmission, which requires central banks to balance price stability with financial stability. Supply chain disruptions caused by climate change led to higher prices, which forced central banks to increase interest rates and create more severe credit problems, and reduced financial flexibility. The implementation of growth-supporting policies through policy easing will generate additional inflationary challenges for countries that depend on energy and food supplies. The trade-offs become more pronounced in emerging markets because of their limited capital markets and their tendency for capital flow cycles and unstable exchange rates ^11,12.

2.4. Empirical gaps in the BRICS context

The BRICS economies face limited research about their exposure to physical and transition risks, despite their rising vulnerability to these threats Reinders et al. (2020)¹⁴ developed a taxonomy of sustainable finance from a policy perspective; Kahn et al. (2016)¹¹ examines how climate change affects developing regions economically. Most research studies have concentrated on long-term sustainability and carbon emissions and transition risks, but they fail to examine the immediate macro-financial impacts of temperature changes. The research identifies a missing link because no existing study has combined temperature anomaly effects on monetary policy changes with their impact on financial stability indicators within the BRICS bloc through quarterly data analysis. The existing research gap requires immediate attention because BRICS central banks must handle unstable capital movements and currency fluctuations and commodity price relationships, which produce complex climate shock effects that standard annual data analysis methods fail to detect. The current research lacks sufficient studies that use system-based approaches, which can detect feedback relationships between climate variables and monetary and financial elements. The PVAR model shows its ability to study macroeconomic relationships between countries according to Love and Zicchino (2006) and Towbin and Weber (2013)^15,16, but its use for climate research in BRICS countries remains in its initial stages. The research constructs a single PVAR model spanning 2000 to 2024 to analyse how monetary policy affects financial stability when dealing with external temperature shock responses. The model enables researchers to study both short-term market responses and long-term equilibrium changes, which provides a complete analysis of environmental volatility effects on macro-financial instability in emerging markets.

3.1. Model specification

Following Love and Zicchino¹⁵, a standard PVAR(p) with country-specific fixed effects can be written as:

$Y_{i,t}=A_1{Y}_{i,t-1}+A_2{Y}_{i,t-2}+\cdots +A_p{Y}_{i,t-p}+{\mu }_i+{\epsilon }_{i,t}$

where:

• $Y_{i,t}$ is a vector of endogenous variables for country i at time t,

• $A_j$ are coefficient matrices of lag order j,

• ${\mu }_i$ captures unobserved country-specific fixed effects,

• ${\epsilon }_{i,t}$ denotes a vector of innovations assumed to be serially uncorrelated with the covariance matrix $\text{Σ}$.

In this study, the vector of endogenous variables is defined as: $Y_{i,t}=\left(\begin{array}{c}TEM{P}_{i,t}\\ M{P}_{i,t}\\ FS{I}_{i,t}\\ IN{F}_{i,t}\\ GD{P}_{i,t}\end{array}\right)$

where:

• TEMP represents temperature anomalies (physical climate shocks), which are defined as deviations from a long-run climatological baseline. We apply seasonal adjustment to quarterly temperature data because they need to eliminate seasonal patterns to detect climate shocks instead of seasonal fluctuations.

• Monetary policy (MP) proxies monetary policy stance (policy rate or synthetic indicator).

• FSI is the financial stress index; It is a combined metric that tracks financial system-wide stress through its assessment of banking sector problems, stock market instability, government bond yield differences, and currency market tensions. All components are standardised and aggregated using an equal-weighting approach consistent with the International Monetary Fund (IMF) and BIS methodologies.

• INF denotes the inflation rate.

• GDP represents quarterly real GDP growth.

This ordering reflects the assumption that climate shocks are exogenous within the quarter, monetary policy responds with some lag, and financial stability conditions react to both climate and policy disturbances. Note that Table 1 below describes the variables used and their sources.

3.2. Treatment of fixed effects and stationarity

The coefficients of PVAR models become biased when heterogeneity exists in the data because unobserved heterogeneity leads to biased estimates. To address this, country-specific fixed effects µ_i are removed using the forward-mean-differencing (Helmert transformation), which preserves orthogonality between transformed regressors and lagged variables. This ensures consistent estimation of the coefficient matrices Aj.

The studies based on PVAR models usually need strict stationarity, but the quarterly macro-financial variables in this analysis show persistent behaviour that returns to their mean. Rather than imposing differencing (which may remove valuable long-run information), we rely on: demeaning and standardising, including multiple lags (p=2), and the process of validating stationarity after the fact involves using a unit-root test.

3.3. Lag length selection

The lag order is determined using standard information criteria Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Hannan–Quinn Information Criterion (HQIC), and final prediction error (FPE). For each BRICS country, we estimated VAR models with up to four lags. Although some information criteria suggested different optimal lag lengths, a PVAR(2) specification (two lags) was selected as the baseline because the model follows the quarterly frequency literature, which requires at least two lags to detect monetary and financial cycles. This could be reinforced in the following section through the results of the aforementioned criteria.

3.4. Estimation procedure

The PVAR is estimated using the Generalised Method of Moments (GMM) estimator designed for dynamic panels, following the procedure of Holtz-Eakin, Newey, and Rosen (1988)¹⁷. The estimation follows this sequence.

• 1.

  Helmert transformation functions as a technique that removes fixed effects from data.

• 2.

  Using deeper lags of endogenous variables as instruments.

• 3.

  The process of computing GMM coefficient matrices with robust standard errors.

• 4.

  Obtaining orthogonalised and generalised impulse responses.

The method ensures that the estimates remain stable when the model includes lagged dependent variables and when there are endogeneity issues between the variables.

3.5. Impulse response functions (IRFs)

The analysis of shock propagation through time requires us to generate IRFs which extend across twelve quarters, amounting to three years. The analysis includes two different methods to generate IRFs.

4.1. Panel unit-root test with Im–Pesaran–Shin (IPS)

The PVAR framework requires validation through stationarity tests, which examine the stability properties of all variables included in the system. The PVAR model uses level estimation for structural analysis of macroeconomic quarterly data, so we checked all series for explosive patterns. In addition, the IPS panel unit-root test evaluates stationarity of the BRICS quarterly macro-financial variables because they show persistence but no explosive patterns. The IPS t-bar statistic and Fisher–ADF p-values operate as a diagnostic system that enables effective evaluation.

The results of the IPS unit root test [Table 2] show that all variables are stationary at the level, satisfying the IPS condition. This validates the reliability of IRFs computed from the PVAR.

4.2. GMM diagnostic tests and lag-length justification

Table 3 reports the lag length selection criteria for the PVAR model based on AIC, BIC, HQIC, and the FPE. The lag length selection results indicate that a two-lag specification is the most appropriate choice for the PVAR model.

The diagnostic results presented in Table 4 validate the econometric accuracy of the panel vector autoregression–generalised method of moments (PVAR–GMM) estimation method. The Hansen test of over-identifying restrictions fails to reject the null hypothesis, which confirms that the instrument set is valid and appropriately exogenous. The Arellano–Bond AR(2) test shows no evidence of second-order serial correlation in the differenced residuals, which proves that the model lacks dynamic misspecification. Taken together, these results provide strong evidence that the PVAR–GMM estimates are reliable and that the inferred impulse response functions are based on a well-specified dynamic panel model.

4.3. PVAR estimation

Table 5 reports the estimated lag(1) effects of each variable on the others, averaged across the five BRICS economies. It shows that Temperature shocks reduce financial stability significantly. The monetary policy tightening response to temperature-driven inflation does not effectively control FSI stability because it creates only partial policy stabilisation. The real economy experiences negative effects from temperature shocks because GDP shows a small decrease after these events. The coefficient structure shows how climate change causes inflation, which monetary policy controls, but also impacts financial stability and economic growth. Indeed, a one-standard-deviation increase in temperature anomaly reduces the financial stress index and real GDP growth, respectively, by 0.187 and 0.073 standard deviations in the following quarter. Yet it raises the monetary policy rate by 0.042 percentage points in the following quarter. Furthermore, a one-percentage-point increase in the policy rate and in real GDP growth reduces the financial stress index, respectively, by 0.116 and 0.065 standard deviations in the following quarter, but in inflation, it raises the policy rate by 0.213.

4.4. Impulse response functions (IRFs) analysis

The BRICS-averaged impulse responses appear in Figures 1-5, which use generalised IRFs (GIRFs) to generate the responses across a 12-quarter time period. In fact, the impulse response traces the dynamic reaction of a variable to a one-standard-deviation positive temperature shock essentially across BRICS countries, averaged from individual country VAR(2) models.

Close

Figure 1: Response of the financial stress index (FSI) to a temperature shock. IRF: Impulse response function.

Export to PPT

Close

Figure 2: Response of monetary policy (MP) rate to a temperature shock. IRF: Impulse response function.

Export to PPT

Close

Figure 3: Response of financial stress index (FSI) to a monetary policy (MP) shock. IRF: Impulse response function.

Export to PPT

Close

Figure 4: Response of GDP growth to a temperature shock.

Export to PPT

Close

Figure 5: Response of GDP growth to a financial stress shock (FSI).

Export to PPT

The graph in Figure 1 demonstrates that a temperature shock equal to one standard deviation leads to immediate financial stress growth, which reaches its highest point between quarters two and three before returning to normal levels after eight to ten quarters. This shows that climate anomalies lead to immediate financial instability because they impact businesses that depend on weather conditions, including farming operations, energy consumption, and insurance claims. The policy rate shows a small increase during the first three quarters after a temperature shock before it returns to its original level, according to Figure 2. This indicates that monetary policy narrows following a temperature shock. The financial stress levels increase when monetary policy becomes more stringent, according to Figure 3, and these effects last between four and six quarters. The short-term economic growth emerges from two factors in Figure 4: climate-induced demand increases and seasonal productivity improvements. Climate change will generate permanent economic damage because it decreases productivity levels and makes financial markets more unstable. The BRICS economies face short-term economic cycles that lead to lasting economic decline because of temperature changes, which need climate risk evaluation for macroeconomic modelling and fiscal management. The graph in Figure 5 demonstrates how real GDP growth responds to a one-standard-deviation Financial Stress Index shock across all BRICS countries. The GDP level reaches -0.06 during Quarter 1. The GDP index shows a strong upward movement in Quarter 2 before reaching its highest point at +0.12. The GDP index demonstrates a small positive expansion between Quarter 3 and Quarter 4 at 0.025-0.05. The GDP index shows a gradual decline toward zero, which starts during Quarter 5. The GDP responses during Quarters 8-12 stay at minimal levels between 0.005 and 0.01. That being said, the economic output reduction from financial stress continues to affect growth rates, resulting in a 0.3% total decline during the first two years.