What if economists could zoom in on the exact impact of a policy—say, a new tax cut or central bank rate hike—as it ripples through the economy, not just in broad averages but with a precise sense of timing and local context? That’s the promise of targeted local projections, a modern econometric tool that has become increasingly influential in research on macroeconomic policy effects. If you’re curious about how top economists estimate the direct, nuanced consequences of major policy moves, understanding targeted local projections is key.
Short answer: Targeted local projections are an econometric method designed to estimate how specific macroeconomic policies affect economic outcomes over time and across different groups or locations. Unlike traditional models, they allow researchers to pinpoint the timing and distribution of policy impacts—making it easier to see not just whether a policy works, but how and when its effects unfold in detail.
The Need for Precise Policy Measurement
Macroeconomic policies—like changes in interest rates, government spending, or the introduction of new technologies—can have widespread consequences. However, these effects aren’t always uniform or immediate. For example, as highlighted in the research of Barrios, Hochberg, and Yi (2020) from nber.org, the arrival of ridehailing services led to a measurable increase of about 3% in traffic fatalities in U.S. cities. This change wasn’t the same everywhere or for everyone: it varied with local conditions, such as previous reliance on public transportation or carpooling, and persisted over time. The authors controlled for factors like “proxies for smartphone adoption patterns” to ensure that their results isolated the effect of ridehailing itself, not just broader technology trends (nber.org).
Traditional models often struggle to capture these nuances. They might average effects across the whole country or assume impacts are felt instantly. But economic reality is rarely so neat. That’s where targeted local projections come in—they’re built to answer the kind of questions policymakers and researchers care about: Who is affected, when, and by how much?
What Are Targeted Local Projections?
At their core, targeted local projections are a statistical technique used to map out the dynamic response of an economic variable (like unemployment, GDP, or accident rates) to a specific policy change or shock, over a series of future time periods. The “local” part refers to the method’s ability to focus on particular points in time (for example, one year out, two years out, etc.), or on particular subgroups or regions (“targets”) within the data.
The method works by estimating a sequence of separate regressions for each future period of interest. For example, if researchers want to know how a new policy affects employment one year, two years, and five years down the line, they would run separate regression models for each of those periods, rather than relying on a single, all-encompassing equation. This approach is especially powerful for identifying the timing and persistence of effects: does the impact of a policy fade quickly, or does it build over time?
In the context of the ridehailing study, for instance, the researchers could use targeted local projections to track how traffic fatalities changed in the months and years following the introduction of ridehailing services, controlling for other local trends. This allows them to say with confidence that the “increase of approximately 3% in the number of fatalities and fatal accidents” was not just a statistical blip, but a persistent change that could be observed over multiple periods (nber.org).
Advantages Over Traditional Models
One major advantage of targeted local projections is their flexibility. Traditional vector autoregression (VAR) models, for example, require strong assumptions about the structure of the economy, and often impose the same relationship between variables at every point in time. Local projections, by contrast, require fewer assumptions and can accommodate changing relationships across different periods.
This flexibility also extends to the kinds of heterogeneity researchers can study. If a policy is expected to have bigger effects in some places than others (say, in cities with more public transport use, as was seen with ridehailing), targeted local projections can directly estimate these differences. This was clear in Barrios, Hochberg, and Yi’s work, where they found the effects of ridehailing “higher in cities with prior higher use of public transportation and carpools, consistent with a substitution effect, and in larger cities” (nber.org). Such localized detail is difficult to extract from more rigid models.
Another benefit is transparency. Since each projection is estimated separately, it’s easier to see how the estimated policy effect changes over time—say, whether it spikes immediately and then dissipates, or builds slowly. This is crucial for policymakers who need to anticipate not just the size but also the timing of a policy’s consequences.
Real-World Relevance: Examples and Concrete Impacts
To put this in practical terms, consider the ridehailing case again. Using targeted local projections, the researchers established that the introduction of ridehailing correlated with “an increase in arterial vehicle miles traveled, excess gas consumption, and annual hours of delay in traffic,” as well as a persistent uptick in traffic fatalities (nber.org). What’s more, their “back-of-the-envelope estimates of the annual cost in human lives range from $5.33B to $13.24B.” These are concrete, policy-relevant numbers that illustrate not just a statistical relationship, but a meaningful societal impact.
The method also allows for robust tests of causality. By controlling for other possible influences—like changes in smartphone adoption or broader trends in urban mobility—the researchers could home in on the contribution of ridehailing specifically. This is crucial for informing policy debates: if a city is considering new regulations on ridehailing services, it needs to know not just whether these services are associated with more accidents, but whether the services themselves are the cause, and how that effect plays out over time.
Targeted local projections have been applied in a variety of other macroeconomic contexts as well. For instance, economists use them to study the effects of monetary policy shocks on inflation, employment, and output, or to trace the impact of government spending programs across different regions or demographic groups. Their ability to reveal “local” dynamics—whether in time, space, or population—makes them indispensable for modern empirical macroeconomics.
Contrasts and Limitations
While targeted local projections are powerful, they aren’t without limitations. One challenge is that, by estimating separate models for each time period or subgroup, the method can become statistically “noisy,” especially with limited data. It also requires careful attention to the choice of control variables and the possibility of confounding factors, as any omitted variable could bias the results.
Researchers must also grapple with the interpretation of local effects: is a spike in fatalities, for example, due to the policy itself, or to other local changes that happened to coincide with its rollout? In the ridehailing example, the authors went to great lengths to control for “proxies for smartphone adoption patterns” and other local trends, but full certainty is always elusive in social science (nber.org).
Nonetheless, the growing use of targeted local projections reflects a broad consensus in applied macroeconomics: nuanced, context-aware measurement is essential for understanding how policies play out in the real world.
Broader Implications for Policy and Research
The rise of targeted local projections has significant implications for both policymakers and researchers. For policymakers, the technique offers a way to forecast not just the average effect of a policy, but its specific timing and distribution—crucial information for designing interventions and managing unintended consequences. For researchers, it opens up new avenues for understanding the complexity of macroeconomic dynamics.
Moreover, as demonstrated by studies like the one by Barrios, Hochberg, and Yi, this approach can uncover important, sometimes counterintuitive effects. While some headlines suggested ridehailing services would reduce traffic fatalities by curbing drunk driving, the detailed, locally targeted analysis revealed a more complex reality: both vehicle occupants and pedestrians saw increased risk, especially in larger urban areas and places with more public transport substitution (nber.org).
Summary
In summary, targeted local projections are a flexible, transparent, and powerful econometric technique for estimating the dynamic effects of macroeconomic policies. They allow researchers to track not just whether a policy works, but how, when, and for whom—providing a level of detail and context often missing from traditional models. Their use in studies like the analysis of ridehailing and traffic fatalities underscores their value in revealing the true, often nuanced, impact of major policy shifts. As policymakers and researchers continue to grapple with complex economic challenges, targeted local projections will remain an essential part of the empirical toolkit, helping to ensure that policy debates are grounded in robust, context-sensitive evidence.
