Miguel Herculano, Santiago Montoya-Blandon and Jorge Pinheiro
Financial conditions indices are widely used by central banks and policymakers, including in the UK, to summarise and monitor in real-time the state of financial conditions in an economy. But most indices focus on what happens to average conditions, rather than the risks policymakers are often most concerned about – such as sharp downturns in economic activity or surges in inflation. We develop novel targeted financial conditions indices (TFCIs), using US data, that focus directly on measuring and forecasting these risks. Instead of asking which financial variables move together on average, the approach identifies which ones matter for specific outcomes. We show that different risks are linked to different financial factors – and that focusing on these can improve the forecasting performance of these key macroeconomic targets.
Financial systems are typically assessed using measures that focus on what happens most of the time or on an average basis. But crises are defined by the opposite: they are shaped by rare, extreme events that sit in the tails of the distribution. Financial conditions indices (FCIs) are widely used by central banks and policymakers to summarise and monitor the stance of the financial system. Most FCIs are built to track the average co-movement across financial variables and markets. That makes them useful for predicting the central tendency of macroeconomic outcomes – but much less so for understanding the extreme outcomes policymakers tend to worry about most.
In practice, risks are rarely symmetric. When it comes to real activity, policymakers are often concerned with downside tail risks – sharp contractions rather than average growth. For inflation or unemployment, concern may lie in the upper tail – unexpectedly high inflation (inflation-at-risk), downside risks to growth (growth-at-risk), or spikes in joblessness. Yet standard FCIs, often constructed using Principal Components Analysis (PCA, a statistical method that summarises the common movement across many financial variables), are agnostic about which part of the distribution matters. They summarise what moves together, not what drives tail events.
This blog introduces a new approach that starts from the opposite direction. Rather than asking which financial variables comove on average, it asks which ones matter for specific macroeconomic risks of interest. We construct targeted financial conditions indices (TFCIs) that are explicitly designed to predict tail risk of key macroeconomic variables – such as the lower tail of growth or the upper tail of inflation.
The key insight is simple. If the policy question is about tail risk of a key target, the conditioning information should be tailored both to that tail and target from the outset. Using a quantile-based extension of the three-pass regression filter of Kelly and Pruitt (2015), the method extracts financial factors that are specific not only to the macroeconomic variable of interest, but also to the quantile being forecast, rather than the variation that dominates on average.
Once we do this, the picture of financial conditions changes meaningfully. The financial drivers of downside risks to activity look different from those associated with upside risks to inflation or unemployment. In other words, there is no single ‘financial conditions’ factor – there are multiple, tail-specific ones, each with distinct economic content.
Why targeting tails changes the picture
In practice, this is done by adapting an existing factor-based approach so that it focuses on specific parts of the distribution – using quantile methods rather than standard average-based techniques.
Put differently, standard methods ask which financial variables move together most strongly. The targeted approach asks which combination of financial variables is most informative about the specific tail risk policymakers care about. Those are different questions – and, as we show, they lead to different answers.
What the data shows
We apply this approach to a large panel of 105 monthly financial indicators, spanning credit, leverage, and risk measures in the United States, and covering the same data set used in the Chicago Fed’s National Financial Conditions Index. We examine three macroeconomic targets – CPI inflation, industrial production, and unemployment – across multiple forecast horizons and across parts of the distribution. The focal tail is the part of the distribution that the index is designed to capture. For example, on inflation, we focus on the upper tail (the highest inflation outcomes), allowing the index to identify the financial signals that matter most when inflation is unusually high rather than when it is close to average.
A first key result is that targeting materially changes the economic content of the extracted factor.
Chart 1: Full-sample focal-tail TFCIs versus PCA
Notes: The focal tails are 𝜏 = 0.90 for inflation and unemployment and 𝜏 = 0.10 for industrial production.
For downside risks to industrial production, the targeted index loads heavily on variables related to delinquency, volatility, and money-market conditions. By contrast, for upside risks to inflation, the index places greater weight on commodity prices, term spreads, and liquidity-sensitive credit variables. The full list of variables can be found in the same data source above, in the ‘Contributions‘ file.
Chart 2: Top predictor-level contributors for the focal tails
Notes: Bar length is the mean absolute contribution of each financial series to the corresponding TFCI. To describe a few, SPOVX is the CBOE Crude Oil Volatility Index, COMMODLIQ is the COMEX gold/NYMEX WTI futures market depth, and the BONDGR is the New US corporate debt issuance relative to its 12-month moving average.
This distinction matters. It implies that ‘financial conditions’ cannot be summarised by a single metric if the goal is to understand different macroeconomic risks. The financial signals associated with downside risks to activity are not the same as those associated with upside risks to inflation or unemployment.
Does targeting improve forecasting performance?
We next assess whether these targeted indices improve the ability to forecast macroeconomic outcomes. We measure forecast accuracy using a standard metric (more specifically, tick loss) where lower values indicate better performance.
In practical terms, the question is whether targeting specific risks – such as periods of very high inflation – helps us make better predictions than focusing on average outcomes.
The results show that targeted indices can deliver meaningful improvements in forecasting performance, particularly for inflation. For example, when forecasting the upper tail of inflation – periods when inflation is unusually high – the targeted index consistently produces more accurate forecasts than both a simple benchmark model and one based on a standard FCI. In several cases, the improvement is sizeable and statistically significant at 1%.
For industrial production and unemployment, the improvements are more nuanced. The targeted indices frequently outperform the autoregressive benchmark and, in some cases, also improve on PCA based measures – particularly at shorter horizons or for specific parts of the distribution. But the gains are not uniform across all settings.
This pattern is informative. Targeting does not automatically dominate traditional approaches in every context. Instead, its advantages are most pronounced when the forecasting objective is closely aligned with a particular tail risk.
Why this matters for policymakers
For policymakers, FCIs are valuable because they provide a compact summary of a large and complex financial system. But the relevant summary depends on the question being asked.
If the objective is to monitor broad financial conditions, a conventional FCI may suffice. But if the objective is to assess risks – such as the probability of a sharp economic downturn or an inflation spike – then a more targeted measure may be more informative.
The results in this paper suggest that tailoring financial conditions indices to specific macroeconomic risks can change both the interpretation of financial conditions and the inferred drivers of those risks. This can, in turn, support more targeted policy analysis and communication.
Although the empirical application uses US financial and macroeconomic data, the broader lesson is not specific to the United States. Policymakers in the UK and elsewhere often focus on risks that are concentrated in particular parts of the distribution, such as periods of unusually high inflation or sharp economic downturns. While the specific financial indicators associated with those risks may differ across countries, the framework illustrates how measures of financial conditions can be tailored to the policy question at hand, helping to identify the financial signals that are most relevant for assessing particular macroeconomic risks.
Conclusion
The central message is straightforward. If policymakers care about tail risks, the tools used to measure financial conditions should reflect that focus.
Targeted financial conditions indices provide one way to do this, by identifying the financial signals that matter for specific macroeconomic outcomes rather than relying on a single, broad measure.
In doing so, they shift the focus from average conditions to the risks that are most relevant for policy decisions – where financial conditions may matter most.
Miguel Herculano is a Lecturer (Assistant Professor) in Financial Economics at the University of Glasgow, Santiago Montoya-Blandon is a Lecturer (Assistant Professor) in Economics at the University of Glasgow and Jorge Pinheiro works in the Bank’s Banking Capital Policy Division.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk or leave a comment below.
Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.
Share the post “Targeted Financial Conditions Indices (TFCI): why average financial conditions miss what matters most”
