📄 other

A Method for Estimating Technological Innovation in Macroeconomy: An Extension of Econometric Structural Estimation by Simulation

This paper proposes a simulation-based extension of structural econometric estimation to measure technological innovation's impact on macroeconomic growth, finding that while the US has significantly outperformed its optimal growth path since the 2000s due to IT-driven innovation, Japan has lagged behind due to offsetting macroeconomic inefficiencies, though the method's accuracy remains limited by current estimates of intertemporal elasticity of substitution.

Original authors: Shungo Sakaki

Published 2026-07-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Shungo Sakaki

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the economy as a giant, complex video game where countries are players trying to level up their "GDP" score. In this game, there are two main ways to get stronger: you can build more stuff (like factories and roads), or you can get smarter at using the stuff you already have. Getting smarter is called "technological innovation," and in the world of economics, it's often measured by something called Total Factor Productivity (TFP). Think of TFP as the "efficiency stat" of a country. If two countries have the same amount of robots and workers, but one produces twice as much pizza, that extra pizza is thanks to a higher efficiency stat.

But how do we know if a country is actually getting smarter, or if it's just working harder? That's the tricky part. This paper tries to solve that mystery by building a "perfect world" simulation. Imagine a GPS that calculates the absolute fastest, most efficient route a car could take to get from Point A to Point B, assuming the engine never breaks and the driver never gets tired. That GPS route is the "optimal growth path." If the car actually drives faster than the GPS says is possible, we know the car must have a secret turbo boost (technological innovation). If the car drives slower than the GPS, something is wrong with the driver or the road, even if the engine is fine. This paper uses that GPS idea to see how the US and Japan have performed since the "IT Revolution" (the rise of computers and the internet) changed the game.

The Great Economic Race: US vs. Japan

This paper, written by independent scholar Shungo Sakaki, is a deep dive into how the United States and Japan have performed since the mid-1990s, when the "IT economy" (the era dominated by computers and the internet) really kicked off. The author wanted to know: Did these countries actually get a "turbo boost" from new technology, or did they just stumble along?

To find out, the author didn't just look at how much money the countries made. Instead, he built a mathematical simulation of what the "perfect" economy would look like if it were running on the old technology (pre-1990s) but managed by a super-smart, perfectly rational planner. This planner knows exactly how to balance saving money for the future and spending it today to get the best possible result. This perfect plan is the "optimal growth path."

The author then took the real-world data from the US and Japan from 1996 to 2022 and compared it to this perfect plan. The logic is simple: if the real economy is doing better than the perfect plan (which was based on old tech), the extra success must be because of new, shiny technological innovations. If the real economy is doing worse, then something is holding it back.

The Results: A Tale of Two Countries

The simulation painted a very different picture for the two nations.

For the United States, the results were exciting. The simulation showed that the US economy was actually zooming past the "perfect plan" based on old technology. The paper estimates that technological innovation has boosted US GDP by about 20% since the 2000s, and that this cumulative boost has reached 30–40% by the 2020s. In video game terms, the US player didn't just play well; they unlocked a secret "tech tree" that made them significantly stronger than the rules of the old game allowed.

For Japan, however, the story was much less cheerful. The simulation showed that Japan's actual GDP has been falling significantly short of the "optimal growth path" since 1996. This doesn't necessarily mean Japan failed to invent or use new IT technology. Instead, the paper suggests that "macroeconomic inefficiencies"—like policy mistakes, slow corporate adaptation, or other economic glitches—were so strong that they completely canceled out any benefits Japan might have gotten from new technology. It's like having a Ferrari engine (new tech) but driving it through a traffic jam of construction and bad rules; you never reach the speed you're capable of. In fact, the paper notes that Japan's actual performance was so far below the optimal path that the "improvement rate" was negative, meaning the economy was underperforming even compared to what was theoretically possible with old tech.

The Catch: The "Elasticity" Problem

There is a big "but" to this story, and it's a technical one that the author is very honest about. To run this perfect GPS simulation, the author needed a specific number called the "Intertemporal Elasticity of Substitution" (IES). You can think of IES as a measure of how willing people are to trade spending money today for spending money tomorrow.

The paper found that depending on which number you use for IES (based on previous research), the simulation sometimes broke down. For the US, when using one specific estimate for IES, the computer couldn't calculate a reliable "perfect path" at all. This means the method the author invented has limits. If we don't know the exact "willingness to wait" number, we can't be 100% sure how accurate our "turbo boost" calculations are. The author concludes that while the method is a cool new way to look at innovation, it relies heavily on getting that one specific number right, and right now, we aren't quite sure enough to make the method perfect for every situation.

The Bottom Line

In short, this paper suggests that the US has successfully harnessed the IT revolution to supercharge its economy, pushing it far beyond what was possible in the past. Japan, on the other hand, seems to have been weighed down by other economic problems that masked any potential gains from new technology. The study offers a fresh, simulation-based way to measure these invisible "efficiency boosts," but it also warns us that our tools for measuring them are still a bit shaky until we understand the "willingness to wait" factor a little better.

Technical Summary: A Method for Estimating Technological Innovation in Macroeconomy

Problem Statement
The paper addresses the challenge of estimating the magnitude of technological innovation within a macroeconomic framework, specifically during the transition to an information technology-based economic structure (the "IT economy"). While existing methods exist—such as calculating Total Factor Productivity (TFP) as a Solow residual, using endogenous growth models to internalize R&D, or employing Dynamic Stochastic General Equilibrium (DSGE) models to estimate policy spillovers—this paper identifies a gap in directly benchmarking actual economic performance against a theoretical optimal path derived from pre-IT technological constraints. The core problem is distinguishing between GDP growth driven by optimal resource allocation under existing technology versus growth driven by genuine technological advancement.

Methodology
The proposed method extends the structural estimation of production functions by integrating simulations of an optimal growth model. The approach proceeds in three main stages:

Pre-IT Structural Estimation:
Using OECD time-series data (1970–1990 for the U.S., 1980–1995 for Japan), the authors estimate production functions and TFP for the pre-IT era using the Generalized Least Squares (GLS) method. The production function is specified as a Cobb-Douglas form ( $y_t = a k_t^b$ ), where $a$ represents TFP and $b$ is the capital elasticity.
Optimal Growth Path Simulation:
The authors construct a benchmark "optimal growth path" assuming the technology level remains fixed at the pre-IT era estimates. This utilizes the Ramsey-Cass-Koopmans optimal growth model, defined by:
- State Equation: $\dot{k}_t = f(k_t) - c_t - n k_t$ (capital accumulation).
- Objective: Maximizing the discounted sum of household utility $u(c_t)$ over an infinite horizon.
- Keynes-Ramsey Rule: $\dot{c}_t/c_t = \frac{1}{\sigma}[f'(k_t) - n - \rho]$ , where $\sigma$ is the Intertemporal Elasticity of Substitution (IES), $n$ is population growth, and $\rho$ is the subjective discount rate.
The simulation starts from the actual fixed asset balances of 1996 (marking the onset of the IT economy in Japan and the mid-point for the U.S. in this study) and numerically solves the system of differential equations to approach the "modified golden rule" steady state. The authors use Mathematica 13.2 for numerical solutions, setting a time interval $\Delta t = 0.1$ .
Innovation Estimation:
The "effect of technological innovation" is defined as the percentage by which actual GDP exceeds the GDP calculated on the simulated optimal growth path (which assumes no technological improvement). This difference is attributed to TFP growth driven by the IT revolution. The authors acknowledge that this metric also captures year-specific economic conditions and potential inefficiencies.

Key Contributions

Novel Benchmarking Approach: The paper introduces a method that does not rely solely on residual calculation but uses the deviation from a theoretically optimal intertemporal allocation of resources (under constant pre-IT technology) as a proxy for technological progress.
Comparative Analysis of US and Japan: The study applies this framework to two major economies, utilizing specific IES estimates derived from prior research (Yagihashi and Katano, 2020) to calibrate the optimal paths.
Identification of Macroeconomic Inefficiencies: The methodology allows for the differentiation between a lack of technological innovation and a lack of efficient resource allocation. In the case of Japan, the model reveals that actual GDP fell below the optimal path, suggesting that macroeconomic inefficiencies may have offset any potential IT-driven gains.

Results

United States: The simulation indicates significant positive deviations. Since the 2000s, actual GDP has exceeded the optimal growth path by approximately 20%. This gap widened in the 2020s, with actual levels exceeding the benchmark by 30–40%. These figures are interpreted as the boost in GDP attributable to technological innovation.
Japan: In contrast, actual Japanese GDP levels since 1996 have consistently fallen short of the calculated optimal growth path. The study reports negative improvement rates (ranging from roughly -1.7% to -48% depending on the year and IES parameter). The authors conclude that while IT-driven innovation may have occurred, it was more than offset by macroeconomic inefficiencies, including policy failures (e.g., non-performing loan resolution, monetary easing, consumption tax hikes) and lagging corporate adaptation.
Technical Limitations: The study highlights a critical dependency on the accuracy of IES estimates. For the U.S. with an IES of 0.59, the authors could not compute an approximate optimal path with sufficient accuracy, preventing a reliable estimation of adjusted TFP for that specific parameter set. This suggests the method's applicability is constrained by the precision of available IES data.

Significance and Claims
The paper claims to offer a new perspective on estimating TFP by shifting the focus from direct statistical residuals to a comparison against a counterfactual optimal growth trajectory. The authors modestly frame the results, noting that the "improvement rate" is a composite measure that includes both technological innovation and specific economic conditions.

The primary significance lies in the finding that for Japan, the failure to meet the optimal growth path implies that structural and policy inefficiencies have been a more dominant factor than technological stagnation in explaining economic performance relative to the theoretical optimum. However, the authors explicitly caution that the method's reliability is currently limited by the availability of accurate IES estimates; without these, the computational accuracy required to define the saddle-point path is not guaranteed. Thus, the paper presents the method as a promising extension of structural estimation that requires further refinement in parameter estimation to be broadly applicable.

Technical Summary: A Method for Estimating Technological Innovation in Macroeconomy

More like this