Designing a Behavioral Mechanism for Managing Inflation Expectations under Bounded Rationality: Evidence from Iran (2016–2024)

Abstract
In economies with chronic instability, the formation of inflation expectations does not rely solely on macroeconomic variables; it is also strongly influenced by behavioral factors, media signals, and collective psychological reactions. This article aims to design a behavioral policy mechanism for managing inflation expectations in Iran, using monthly data from March 2016 to December 2024. In the first step, employing an Autoregressive Distributed Lag (ARDL) model and a Random Forest algorithm, the existence of a long-run relationship between inflation expectations and a set of macroeconomic and behavioral variables, as well as their relative importance, is examined. Then, a Time-Varying Parameter Vector Autoregressive (TVP-VAR) model is used to estimate these relationships dynamically. The TVP-VAR estimates show that the average coefficient of the free-market exchange rate over the entire period is about 0.25 and is significantly larger than the coefficient of the monetary base, which is found to be about 0.05. Moreover, the inflation search index and public trust have coefficients of approximately 0.15 and 0.10, respectively. These results indicate that inflation expectations are influenced more by exchange-rate shocks and informational and perceptual stimuli than by traditional monetary variables, and that the strength of this effect increases during episodes of exchange-rate surges. Based on this evidence, a behavioral mechanism is proposed for the monetary policymaker, including targeted publication of inflation forecasts, structured intervention in the information environment, use of financial instruments to align expectations, and continuous monitoring of behavioral indicators.
 
 
 
Extended Abstract
 
Purpose: The purpose of this study is to design a behavioral mechanism for managing inflation expectations in Iran where bounded rationality, information frictions, and credibility deficits shape how the public forms beliefs about future inflation. Specifically, the research aims to (i) quantify the relative roles of nominal fundamentals (inflation, monetary base, liquidity, interest rate) versus salient behavioral-information signals (search intensity, negative news, consumer confidence) to explain inflation expectations, (ii) assess whether these effects are stable or time-varying across periods of relative stability and crisis, and (iii) translate the empirical findings into an operational policy framework that complements conventional monetary policy with communication and behavioral monitoring tools. The study is motivated by the observation that, in Iran’s recent macroeconomic history, sharp shifts in expectations often occur even without proportional changes in traditional monetary indicators, suggesting that expectation formation is partly driven by salience, heuristics, and social-information channels.
 
Methodology: The study uses monthly Iranian data from March 2016 to December 2024. The dependent variable is an inflation expectations index constructed from domestic survey-based and institutional measures reported by national research and monetary-policy institutions. The explanatory variables are organized into two sets.
The first set includes standard macroeconomic and the nominal variables of monthly CPI inflation (and/or point-to-point inflation), the free-market exchange rate, the monetary base, broad money and liquidity aggregates, and the interbank interest rate. Inflation is obtained from official statistical sources, monetary aggregates and the interest rate are taken from central bank data, and the free-market exchange rate is collected from widely used domestic market sources as a proxy for the most salient nominal anchor in public perception.
The second set includes behavioral and information variables that proxy bounded rationality and attention allocation an inflation-related Google Trends search index, a consumer confidence (or public trust) index, and an economic negative-news intensity index constructed from Persian-language media coverage. These indicators are compiled at monthly frequencies to match the macro data and to capture the shifts in salience and sentiment.
Data preprocessing involves harmonization to a common monthly frequency and transformations to improve comparability and statistical properties. The variables with strong trends are log-transformed and converted into growth rates when appropriate. Stationarity is assessed using standard unit-root procedures, and model specifications avoid retaining non-stationary levels where this could generate spurious relationships.
Empirically, the paper follows a three-step approach. First, ARDL is used to test for long-run relationships between expectations and the combined macro-behavioral set, providing evidence on long-run associations and error-correction dynamics. Second, Random Forest regression is used as a complementary nonlinear approach to evaluate variable importance rankings and detect potential nonlinearities without imposing strong parametric restrictions. Third, to capture the core contribution of this study, a TVP-VAR framework is employed to estimate the time-varying coefficients and dynamic responses. This approach allows the sensitivity of expectations to exchange-rate shocks, monetary disturbances, and behavioral-information signals to evolve over time, which is essential for crisis-prone economies with structural change and credibility fluctuations.
 
Findings and Discussion: The results consistently indicate that inflation expectations in Iran are shaped primarily by the free-market exchange rate and behavioral-information variables rather than by conventional monetary aggregates alone. In the TVP-VAR estimates, the average coefficient on the free-market exchange rate in the expectations equation is about 0.25, and it rises over time. This implies that exchange-rate movements increasingly act as the dominant nominal anchor. By contrast, the monetary base coefficient is roughly 0.05 on average and varies little, indicating weaker direct salience of monetary fundamentals in expectation formation as compared to exchange-rate signals.
Behavioral indicators also display economically meaningful and strengthening effects. The inflation-search index has an average coefficient of about 0.15, and the public trust/consumer confidence index is about 0.10 (with the sign depending on index definition), both increasing in later years. This pattern suggests that attention, salience, and credibility perceptions have become more central to expectation dynamics. ARDL estimates corroborate a long-run relationship between expectations and the combined macro-behavioral set; the error-correction term implies rapid adjustment after short-run deviations, which is consistent with fast reactions to salient disturbances.
Random Forest results reinforce this interpretation by ranking the free-market exchange rate and negative news intensity as the most important predictors of expectations, while monetary aggregates and the interest rate receive substantially lower importance. Overall, the evidence supports bounded rationality. That is, under high uncertainty and information costs, agents rely on salient, easily observable signals such as the exchange rate and negative news rather than complex monetary indicators. Time variation further suggests that the expectation channel intensifies during exchange-rate stress, when credibility is questioned and heuristic-based updating becomes stronger.
 
Conclusions and Policy Implications: The findings imply that managing inflation expectations in Iran requires moving beyond a purely fundamentals-based framework. Conventional tools (e.g., interest-rate adjustments and balance-sheet control) remain relevant, but they are unlikely to suffice when the public anchors expectations on the exchange rate and information sentiment. The study, therefore, proposes a behavioral mechanism on three bases.
First, exchange-rate anchoring and stabilization should be treated as a core expectation-management instrument. This does not require a rigid peg, but it does call for a transparent, rule-like approach to limiting excessive volatility and avoiding contradictory signals.
Second, cognitive-aligned communication should be institutionalized. Inflation reports and forward guidance should be more frequent and predictable. Also, they should be designed for interpretability through clear scenarios, consistent messaging, and salient framing to reduce uncertainty-driven overreactions.
Third, behavioral monitoring should become operational. A “behavioral dashboard” based on search trends, confidence measures, and negative-news indicators can provide early warnings of de-anchoring. When thresholds are crossed, the central bank should activate complementary interventions such as targeted communication, clarification of policy stance, and timely disclosure of exchange-rate and inflation outlook assumptions.
Overall, the study supports a dual-track approach in which managing perceptions is assumed to be as important as managing nominal fundamentals.