Intertemporal Substitution and Household Consumption: Evidence from Structural Shocks

 The central thesis of the provided research is that households do not substitute consumption intertemporally in response to changes in real interest rates. While standard macroeconomic models assume that households adjust their spending paths to take advantage of higher interest rates (saving now to consume more later), the source finds no empirical evidence for this behavior across ten different structural shocks. Instead, the data suggests that aggregate consumption responses are almost entirely driven by changes in the expected path of labor income.

The Context of Intertemporal Substitution

In traditional macroeconomic theory, households are expected to "smooth" their consumption over time. This behavior is governed by the elasticity of intertemporal substitution (EIS), which measures how sensitive consumption growth is to changes in the real interest rate. If the EIS is positive, a rise in interest rates should theoretically induce households to reduce current consumption to increase future consumption.

However, the source highlights a long-standing "lack of evidence" for this sensitivity:

• Historical Precedent: Seminal research by Hall (1988) previously found no strong relationship between interest rates and consumption growth, suggesting the EIS may be zero.
• Near Rationality: The source notes that for many households, the utility gain from perfectly optimizing consumption based on interest rate fluctuations is extremely small—potentially as little as $0.08 to $1.45 per quarter. Consequently, it may be "rational" for consumers to ignore interest rates entirely.
• Popular Advice: Interestingly, popular financial advice books rarely, if ever, advise households to change their spending patterns based on interest rate movements, focusing instead on saving fixed fractions of income.

The Author’s Identification Strategy

To isolate the role of intertemporal substitution, the author uses a "consumption block" approach, which allows for estimation without needing to specify the entire structure of the economy. The methodology relies on:

• Structural Shocks: The author examines 10 different types of shocks—including monetary policy, government spending, and technology shocks—some of which produce large, persistent changes in real interest rates.
• Consumption Jacobians: These are "sufficient statistics" used to map how consumption responds to four specific inputs: labor income, the real fed funds rate, stock returns, and real estate returns.
• Sticky Expectations: By incorporating a parameter for "sticky expectations," the author ensures that the observed lack of response isn't simply an artifact of households being slow to process new information.

Findings and Evidence

The core result is that the estimated parameter for intertemporal substitution is statistically indistinguishable from zero for all ten shocks examined.

• Income as the Primary Driver: In every shock series, the "income component" of the consumption response is the dominant factor. For example, in the Romer and Romer (2004) monetary policy shocks, the rise in consumption is explained by the gradual rise in labor income, not the drop in interest rates.
• The Failure of Substitution Models: When the author applies a standard model with "rational expectations" (assuming an EIS of 1), the predicted consumption response moves far outside the confidence intervals of actual empirical data.
• Robustness: These findings remain consistent even when the author tests different values for the elasticity of substitution or uses alternative models like "finite horizon planning".

Implications for Economic Theory

This evidence challenges the "pivotal factor" role that intertemporal substitution plays in many macroeconomic models. If consumption does not respond to interest rates, it suggests that monetary policy may transmit through the economy differently than previously thought—perhaps primarily through its impact on investment or by directly altering household income through labor market channels rather than through the traditional "substitution" incentive.

The methodology employed in the provided research focuses on isolating the "consumption block" of macroeconomic models to determine if households actually shift their spending in response to real interest rate changes. By narrowing the scope to this specific block, the author can remain agnostic about the rest of the model (such as firm-side frictions or price-setting mechanisms), allowing for the use of a wider range of structural shocks than traditional general equilibrium models.

The following sections detail the specific components of this methodological approach:

1. The Input-Output Framework (Consumption Jacobians)

The author treats the consumption decision as an input-output structure where the dynamics are fully described by consumption Jacobians. These Jacobians are "sufficient statistics" that map how aggregate consumption at time $t$ responds to a marginal change in an input at time $s$.

• Inputs: The expected paths for aggregate labor income, the real federal funds rate, and real returns on stocks and real estate.
• Outputs: Aggregate consumption and savings.

2. Use of Structural Shocks for Identification

To overcome the challenge of interest rates co-moving with other economic factors, the author utilizes 10 different structural shocks identified in existing literature (Ramey, 2016). These include monetary policy surprises, defense spending news, tax shocks, and technology shocks.

• The strategy relies on finding shocks that affect consumption only through their impact on income and asset returns, rather than through changes in household preferences (discount factors).
• By using ten distinct shocks, the researcher aims to provide "overwhelming" aggregate evidence that is not sensitive to a single identification method.

3. Two-Step Estimation Process

The methodology follows a precise two-step process to calculate the role of intertemporal substitution:

• Step 1: Local Projections. The author uses local projections to calculate empirical impulse response functions (IRFs) for every input (income, interest rates, asset prices) and for consumption itself across all ten shocks.
• Step 2: Minimum-Distance Estimation. The author then calculates a "Jacobian-implied" consumption IRF by summing the four Jacobians multiplied by the empirical inputs. A distance metric is used to find the parameters that minimize the difference between this theoretical response and the actual empirical consumption response observed in the data.

4. Incorporating Sticky Expectations

A critical feature of the methodology is the inclusion of a parameter for sticky expectations. This ensures that if households appear unresponsive to interest rates, it is not simply because they are slow to process new information or sluggish in belief updating. By explicitly modeling this friction, the author can more rigorously test if the near-zero response is a "genuine lack of intertemporal substitution" rather than a delay in reaction.

5. Key Assumptions and Limitations

• Separability of Labor and Consumption: The methodology assumes that the consumption-saving decision can be separated from the labor decision through aggregate labor income. This assumption might not hold if leisure and consumption are not separable in the household's utility function.
• Linearization: The model is limited to analyzing small deviations around a steady state and does not account for time-varying risk premiums or state-dependent responses.
• Asset Income Effects: Because households' primary assets (stocks/real estate) are not well-approximated by short-term bonds, the author discards the model-based interest rate income effect and replaces it with empirical estimates of consumption responses to asset price changes.

In the research provided, the consumption block is modeled as a specific input-output structure within the larger macroeconomy. The central thesis is that the dynamics of aggregate consumption can be fully described by how it responds to four specific inputs, mediated through "consumption Jacobians" which act as sufficient statistics for these responses.

The Four Primary Inputs

The sources identify the following as the essential inputs that determine the household consumption decision:

• Aggregate Labor Income: This is the most significant input for determining actual consumption behavior. The author specifically defines this as the sum of employee compensation, proprietors' income, and transfers, minus specific taxes and social security contributions.
• Real Federal Funds Rate: This input is the theoretical centerpiece of intertemporal substitution. It is measured as the nominal federal funds rate minus one-year-ahead expected inflation. Standard theory suggests that changes in this input should incentivize households to shift consumption across time.
• Real Returns on the Stock Market: This input helps capture the "income effect" of wealth. It uses inflation-adjusted prices from Fama French.
• Real Returns on Real Estate: Similarly used for the income effect, this input is based on the Case-Shiller house price index adjusted for inflation.

Inputs in the Context of Intertemporal Substitution

The methodology uses these inputs to isolate why households change their spending. Theoretically, if the elasticity of intertemporal substitution (EIS) is positive, the "Real Federal Funds Rate" input should be a major driver of consumption fluctuations.

However, the sources reveal a disconnect between theory and empirical data regarding these inputs:

• Dominance of the Income Input: Across ten different types of structural shocks (such as monetary policy or technology shocks), the labor income path explains almost the entire aggregate consumption response. Even when interest rate inputs move significantly and persistently, they fail to induce a corresponding shift in consumption.
• The Residual Role of Interest Rates: The author calculates the effect of interest rates as a residual—the change in consumption that remains after accounting for the responses to income and asset prices. In practice, this residual is found to be close to zero, suggesting that the interest rate input has a negligible effect on household timing.
• Asset Inputs as Refinements: The sources explicitly replace the "income effect" of interest rates found in simple models with the stock and real estate inputs. This is because most household wealth is held in these assets rather than the short-term bonds typically used in Euler equation models. Despite this more realistic calibration, these asset inputs are found to play a negligible role in consumption responses compared to labor income.

Key Methodological Assumption

The validity of using these inputs relies on a "Key Assumption": that the aggregate consumption response is a function only of these four inputs (income, interest rates, and asset returns) and other factors independent of the shocks being analyzed. This allows the researcher to be agnostic about the rest of the economy—such as how firms set prices or how the government balances its budget—so long as those factors only affect households through these four specific channels.

In the source material, the author utilizes 10 distinct structural shocks to investigate whether households adjust their consumption timing in response to changes in real interest rates. These shocks, selected from the comprehensive review by Ramey (2016), serve as the primary tool for identifying the role of intertemporal substitution because they induce varied and often persistent fluctuations in real interest rates.

The analyzed structural shocks are categorized into three main areas:

1. Monetary Policy Shocks

The author examines two prominent methods for identifying monetary policy surprises:

• Romer and Romer (2004): Shocks are identified by the residual of the federal funds target rate after accounting for the Federal Reserve’s internal "Greenbook" forecasts.
• Gertler and Karadi (2015): This method uses high-frequency identification, looking at surprise changes in fed funds futures around Federal Open Market Committee (FOMC) announcements.

While these shocks lead to immediate drops in the real federal funds rate, the source finds that the resulting rise in consumption is almost entirely explained by a gradual rise in labor income rather than the interest rate change itself.

2. Fiscal Shocks

These shocks involve changes in government spending and anticipated tax policies:

• Military and Defense News: This includes military news shocks from Ramey and Zubairy (2018) and defense spending news from Ben Zeev and Pappa (2017). The latter is particularly informative because it produces a persistent negative real interest rate for four years.
• Government Spending: Blanchard and Perotti (2002) identify shocks through a recursive structure in a Vector Autoregression (VAR).
• Tax Shocks: These include "tax news" shocks from Mertens and Ravn (2011) and "expected taxes" from Leeper et al. (2012), the latter of which is derived from the spread between federal and municipal bonds.

3. Technology Shocks

The final category focuses on shifts in productivity and investment:

• Investment-Specific Technology (IST) News: Ben Zeev and Khan (2015) identify these through changes in the real price of investment.
• Total Factor Productivity (TFP): This includes utilization-adjusted TFP shocks from Fernald (2014) and unanticipated TFP shocks from Francis et al. (2014).

Findings in the Context of Intertemporal Substitution

The central finding across all ten shocks is that the estimated parameter for intertemporal substitution is statistically indistinguishable from zero. The source provides several insights into why these shocks support this conclusion:

• Failure of Traditional Models: If households behaved according to standard rational expectations models (with an elasticity of 1), the predicted consumption response to these shocks would fall far outside the empirical confidence intervals.
• Income as the Dominant Factor: In every shock series, the "income component" forms the bulk of the consumption response. For instance, despite the long period of lowered interest rates following a defense spending shock, households do not show the increased spending that intertemporal substitution theory would predict.
• Robustness of Evidence: By using a broad range of shocks—covering different identification methods and economic sectors—the author argues that the aggregate evidence against intertemporal substitution is "overwhelming" and not dependent on any single, potentially flawed identification strategy.

The theoretical framework of the source centers on isolating the "consumption block" of the economy to test the standard macroeconomic assumption that households smooth consumption by adjusting their spending in response to real interest rates. The author moves away from fully-specified general equilibrium models, instead using a framework that treats aggregate consumption as an input-output structure defined in sequence space.

1. The Consumption Block and Input-Output Structure

The framework assumes that the household decision-making process can be viewed as a sub-block of a larger model. The key theoretical assumption is that the impulse response of aggregate consumption is a function of specific expected paths: aggregate labor income, asset returns (stocks and real estate), and the real federal funds rate.

• Separability: This requires that the labor supply decision be separable from the consumption-saving decision, meaning leisure and consumption are not interconnected in the household's utility function.
• Agnosticism: Because the focus is strictly on this "block," the framework remains valid regardless of how other parts of the economy are modeled—such as price stickiness, international trade, or investment frictions—so long as they do not violate the block's internal logic.

2. Sequence Space and Consumption Jacobians

To describe the dynamics of this block, the framework utilizes "consumption Jacobians" as sufficient statistics.

• A Jacobian is a matrix that maps how aggregate consumption at time $t$ responds to a marginal change in an input (like income or interest rates) at time $s$.
• By linearizing the model around a steady state, the entire behavior of the consumption block is captured by four Jacobians—one for each primary input.

3. Decomposing the Interest Rate Effect

A critical component of the theoretical framework is the decomposition of the real interest rate Jacobian into two distinct effects:

• Intertemporal Substitution Effect: This measures how households shift the timing of consumption solely due to changes in the incentive to save (the Euler equation effect) while keeping the budget constraint fixed.
• Income Effect: This measures how changes in interest rates affect the household's feasible set of consumption plans (their wealth). The author discards the model-based income effect for interest rates—which often assumes households only hold short-term bonds—and instead uses empirical data for stock and real estate returns to capture the "true" wealth effects.

4. Micro-foundations and Behavioral Parameters

The framework is disciplined by a one-asset heterogeneous agent model (often called a buffer-stock model) where infinitely lived households face idiosyncratic income uncertainty and borrowing constraints.

• Sticky Expectations: The author incorporates a parameter for "sticky expectations" ($\theta$) to allow the model to span a range of behaviors, from full rational expectations to complete inattention. This ensures that a zero response to interest rates is interpreted as a genuine lack of substitution rather than a mere delay in processing information.
• Near Rationality: The framework draws on the theory of "near rationality," which suggests that the utility cost of failing to perfectly optimize consumption based on interest rate fluctuations is extremely low (cents per quarter). Theoretically, this makes it rational for households to ignore interest rate signals entirely and follow simpler rules of thumb, such as saving a fixed fraction of income.

5. Identification Strategy

The theoretical framework uses structural shocks as the identification tool. By selecting shocks orthogonal to household preferences (like technology or defense spending shocks), the author ensures that any observed change in consumption is a reaction to the inputs (income, rates, assets) rather than a shift in the households' internal discount factors. This allows for the estimation of the intertemporal substitution parameter as a residual after accounting for the well-documented effects of income and asset price changes.

The primary finding of the research is that households do not substitute consumption intertemporally in response to changes in real interest rates,,. Despite standard macroeconomic theories suggesting that households should shift their spending paths to take advantage of higher interest rates, the source find no evidence of this behavior across ten different types of structural shocks,,.

The key findings and their implications for the study of intertemporal substitution include:

1. Dominance of the Income Component

Aggregate consumption responses to economic shocks are driven almost entirely by changes in the expected path of labor income,,.

• In every shock examined—including monetary policy, defense spending, and technology shocks—the "income component" accounts for the bulk of the consumption response,.
• Movement in asset prices, such as stocks and real estate, plays a negligible role in determining aggregate consumption compared to labor income,.

2. Failure of Standard Macroeconomic Models

The source demonstrates a significant disconnect between empirical data and traditional "rational expectations" models:

• When a standard model with an elasticity of intertemporal substitution (EIS) of 1 is applied, it predicts consumption responses that fall far outside the 90 percent confidence intervals of actual empirical data,,.
• Even for shocks that produce large and persistent changes in real interest rates (such as defense spending news), households do not adjust their spending timing as the theory of intertemporal substitution predicts,,.

3. Robustness of the Near-Zero Estimate

The estimate that households do not intertemporally substitute remains consistent across various testing environments:

• Statistical Significance: For all ten structural shocks analyzed, the parameter for intertemporal substitution is statistically indistinguishable from zero,,.
• Model Variations: The findings are robust to different initial assumptions for the EIS, alternative behavioral frameworks like "finite horizon planning," and different calibrations of household wealth,,,.

4. Theoretical Justification: Near Rationality

The source provides a "near rationality" explanation for why households might ignore interest rate signals:

• The utility gain for a household that perfectly optimizes its consumption based on interest rate fluctuations is extremely small—estimated to be as little as $0.08 to $1.45 per quarter,.
• Because the cost of making a "mistake" by ignoring interest rates is so low, it is rational for consumers to focus on simpler rules, such as saving a fixed fraction of their income,.

5. Implications for Policy Transmission

The findings suggest that the traditional view of monetary policy transmission—where interest rates directly incentivize households to change their spending timing—may be incorrect. Instead, the source suggests that the effects of policy are likely transmitted through its impact on investment behavior or by indirectly altering household labor income.