Technology Diffusion and International Catch-Up

Why Diffusion Matters

A country does not need to invent every production method it uses. A firm can adopt machinery, software, or a production process first developed elsewhere; a country can learn from foreign firms, technical publications, or imported equipment. Such adoption can raise productivity even when the adopter conducts no original research on the technology. This process is technology diffusion.

Diffusion is distinct from invention. Invention creates or substantially improves knowledge or a technology. Diffusion moves existing knowledge into use by firms or countries that did not originate it. The distinction matters for growth analysis: a country’s productivity can increase through domestic innovation, through adoption of foreign innovations, or through both. The earlier material on knowledge spillovers explains why knowledge created in one place can benefit others; here the focus is on how adoption changes productivity and when it can support catch-up.

Definition: The technology frontier is the highest productivity level available to a relevant set of producers at a given time. Technology diffusion is the spread and adoption of technologies developed elsewhere. A follower is a country or firm below the frontier that may adopt available technologies. Catch-up means reducing the follower’s productivity gap relative to the frontier; it does not necessarily mean reaching or surpassing the frontier.

A simple model isolates the mechanics. Let $A_t$ be a follower’s productivity at time $t$, and let $F_t$ be frontier productivity. Assume $A_t \leq F_t$, and suppose the follower adopts a constant fraction $\lambda$ of the gap between its productivity and the frontier each period:

A_{t+1}=A_t+\lambda(F_t-A_t), \qquad 0\leq\lambda\leq 1.

The parameter $\lambda$ is the adoption rate in this model. A larger $\lambda$ means a larger share of the available gap is closed each period. This equation is a simplifying assumption, not a universal law: it treats the adoption rate as constant and summarizes a range of adoption decisions in one parameter.

Closing a Gap When the Frontier Is Fixed

First suppose the frontier is fixed at $F$, so $F_t=F$ in every period. Define the productivity gap as $G_t=F-A_t$. Subtracting the adoption equation from $F$ gives:

G_{t+1}=F-A_{t+1} =F-\left[A_t+\lambda(F-A_t)\right] =(1-\lambda)(F-A_t) =(1-\lambda)G_t.

Thus the gap is multiplied by $1-\lambda$ each period. Repeated application gives $G_t=(1-\lambda)^tG_0$. If $0<\lambda\leq1$, the gap shrinks toward zero. If $\lambda=0$, there is no adoption and the gap remains unchanged. This result says that adoption closes the level gap under the model’s assumptions; it does not establish why a particular country has a particular adoption rate.

Worked Example: Adoption Toward a Fixed Frontier

A hypothetical producer has productivity of 40 units per worker. A foreign technology frontier is fixed at 100 units per worker. Each year, the producer adopts one-quarter of its current productivity gap. Find its productivity and remaining gap after three years.

Method. Apply the adoption equation successively, holding the frontier at $F=100$ and the adoption rate at $\lambda=0.25$.

Initially, $A_0=40$, so $G_0=100-40=60$. In year 1, the producer adopts $0.25(60)=15$ units of productivity: $A_1=40+15=55$, and $G_1=100-55=45$.

In year 2, it adopts $0.25(45)=11.25$: $A_2=55+11.25=66.25$, and $G_2=100-66.25=33.75$.

In year 3, it adopts $0.25(33.75)=8.4375$: $A_3=66.25+8.4375=74.6875$, and $G_3=100-74.6875=25.3125$.

The gap formula checks the result: $G_3=(1-0.25)^3(60)=0.75^3(60)=25.3125$. Therefore productivity after three years is $100-25.3125=74.6875$ units per worker. The common mistake is to add 25% of the original gap in every year; the rule applies to the remaining gap, which declines each period.

Catch-Up While the Frontier Moves

In practice, frontier productivity may keep growing as frontier firms invent and improve technologies. To analyze this case, suppose the frontier grows at a constant rate $g$ per period, so $F_{t+1}=(1+g)F_t$, and the follower still adopts a fraction $\lambda$ of the gap relative to the current frontier:

A_{t+1}=A_t+\lambda(F_t-A_t).

A useful measure is the follower’s productivity relative to the frontier, $q_t=A_t/F_t$. Dividing the adoption equation by $F_{t+1}=(1+g)F_t$ yields:

q_{t+1}=\frac{q_t+\lambda(1-q_t)}{1+g} =\frac{(1-\lambda)q_t+\lambda}{1+g}.

A steady relative position $q^*$ is a value for which $q_{t+1}=q_t=q^*$. Solving the equation gives:

q^*=\frac{(1-\lambda)q^*+\lambda}{1+g} \quad\Longrightarrow\quad (1+g)q^*=(1-\lambda)q^*+\lambda \quad\Longrightarrow\quad q^*=\frac{\lambda}{\lambda+g}.

For positive frontier growth and a positive adoption rate, this steady relative position is below one: the follower remains behind the frontier. Yet its productivity growth rate at that position equals frontier growth. To see why, divide its adoption equation by $A_t$:

\frac{A_{t+1}}{A_t} =1+\lambda\frac{F_t-A_t}{A_t} =1+\lambda\frac{1-q^*}{q^*} =1+g.

The last equality follows from $q^*=\lambda/(\lambda+g)$, which implies $\lambda(1-q^*)/q^*=g$. Thus the follower grows at the frontier’s rate while maintaining a constant proportional gap. This is catch-up in relative position during the transition toward $q^*$, not a guarantee that the follower reaches the frontier. The result depends on the constant-growth frontier and constant-rate adoption assumptions.

Worked Example: Catch-Up With a Growing Frontier

A hypothetical country begins with productivity $A_0=50$, while the frontier has $F_0=100$. The frontier grows by 10% each year, and the country adopts one-half of the gap each year. Calculate its productivity and its relative productivity after two years, then find the model’s steady relative position.

Method. Update the frontier and follower productivity separately each year. The adoption rule uses the gap against the frontier at the start of that year.

At year 0, the gap is $100-50=50$. The country adopts $0.5(50)=25$, so $A_1=50+25=75$. Meanwhile, $F_1=100(1.10)=110$. The relative productivity is $q_1=75/110\approx0.6818$.

At year 1, the gap is $110-75=35$. The country adopts $0.5(35)=17.5$, so $A_2=75+17.5=92.5$. The frontier becomes $F_2=110(1.10)=121$, and $q_2=92.5/121\approx0.7645$.

The steady relative position is $q^*=\lambda/(\lambda+g)=0.5/(0.5+0.10)=0.5/0.60\approx0.8333$. The relative position rises from $0.5000$ initially to approximately $0.6818$, then $0.7645$, moving toward $0.8333$. At that steady relative position, productivity grows at 10% per year, like the frontier, while remaining about $1-0.8333=0.1667$, or 16.67%, below it. The common mistake is to assume that matching the frontier’s growth rate means matching its productivity level; equal growth rates preserve a nonzero proportional gap.

Adoption Costs and Foreign Knowledge Access

The gap model describes adoption after it occurs. Whether an adopter chooses to use a technology depends in part on its expected benefits and costs. Adoption may require installation, worker training, redesign of production, or adapting equipment to local conditions. Access can also be limited: the relevant equipment, technical information, or licensing may not be available at a reasonable cost. These features can delay adoption even when the technology is known to exist.

Adoption decision: In a simple cost-benefit model, a firm adopts a technology when the present value of its expected incremental benefits is at least as large as the adoption cost. This is a decision rule under the specified benefit, cost, and discounting assumptions; it does not imply that every socially beneficial technology will be privately adopted.

For example, suppose a technology creates a constant additional benefit $B$ at the end of each year forever, adoption requires a one-time cost $C$ now, and the effective annual interest rate is $i>0$. If benefits are certain and there are no maintenance costs or other changes, their present value is $B/i$. Adoption is privately worthwhile if $B/i\geq C$. Lower access costs can make this condition hold; better access can also make a technology available sooner. The rule does not account for spillovers to other firms unless those benefits are included in the adopting firm’s own return.

Worked Example: How Access Costs Change an Adoption Decision

A hypothetical manufacturer is considering a foreign production system that would generate an additional \$8,000 in operating benefit at the end of each year forever. The effective annual interest rate is 10%, benefits are certain, and there are no maintenance costs. The system’s installation and access cost is \$90,000 through one channel and \$70,000 through another. Determine whether the manufacturer adopts under each cost.

Method. Value the constant perpetual annual benefit, then compare that present value with each one-time cost. The stated 10% rate is the annual discount rate for this decision.

The present value of the benefits is $B/i=\$8{,}000/0.10=\$80{,}000$. With the \$90,000 cost, net present value is $\$80{,}000-\$90{,}000=-\$10{,}000$, so the manufacturer does not adopt under that channel. With the \$70,000 cost, net present value is $\$80{,}000-\$70{,}000=\$10{,}000$, so it adopts under the lower-cost channel.

The arithmetic checks: the same benefit stream is worth \$80,000 in both cases; only the access and installation cost changes. The common mistake is to compare the annual benefit of \$8,000 directly with the one-time cost. They are different time measures: the benefit must first be converted to present value. The result also depends on the perpetual, certain-benefit assumption; a finite or risky benefit stream would require a different valuation.

What Catch-Up Does—and Does Not—Mean

Diffusion provides one route for productivity growth without requiring the follower to originate every technology it uses. But the simple equations clarify why “technology is available abroad” is not sufficient to conclude that a country will catch up. The country must have a positive effective adoption rate, and that rate is influenced by the costs and access conditions represented in the model. If adoption is absent, the follower can fall further behind a growing frontier even if it began with a substantial level of productivity.

The term conditional convergence is useful here: catch-up is expected conditional on circumstances that allow adoption and on the model’s other assumptions. It is not an unconditional prediction that every low-productivity country will grow faster than every high-productivity country. A follower with a low adoption rate can remain far behind; one with a higher rate can approach a larger share of the frontier. Differences in adoption barriers can therefore contribute to persistent productivity differences.

This tutorial represents adoption capacity through a reduced-form adoption rate and treats the available technology as usable once adopted. The next topic examines more closely the capabilities required to recognize, adapt, and use external knowledge. For the present analysis, keep the distinction clear: a technology can exist at the frontier, yet adoption still depends on access and costs.

Common Exam Traps

Equating invention with diffusion. A country that adopts a foreign process can gain productivity without having invented that process.
Applying the adoption fraction to the original gap every period. The model applies $\lambda$ to the current gap, so the amount adopted usually declines as the gap closes.
Confusing growth-rate convergence with level convergence. With a growing frontier, a follower can grow at the same rate as the frontier while remaining below it.
Assuming foreign availability guarantees adoption. A technology’s presence elsewhere does not remove local access or implementation costs.
Calling catch-up automatic. The convergence result depends on the model assumptions, including a positive adoption rate; it is not a universal prediction for all countries.

Key takeaway: Diffusion raises productivity by bringing existing technology into use. With a fixed frontier, a constant fraction of the remaining gap closes each period. With a growing frontier, the follower’s long-run relative position depends on both frontier growth and adoption; matching frontier growth does not require reaching the frontier.

The framework is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus: identify the channel of productivity growth, apply the adoption-gap equation carefully, and interpret level and growth comparisons separately.

Check Your Understanding

Work each problem by stating the relevant model assumption before calculating. Each problem is original and uses the definitions developed above.

A hypothetical firm has productivity of 30 units per worker, with a fixed available frontier of 90. It adopts one-third of its current gap per year. Find its productivity after two years. (Answer: $190/3$, or approximately 63.33 units per worker.)
A follower has relative productivity $q_0=0.5$. Its frontier grows at 5% per period, and it adopts 40% of the current gap each period. Calculate its relative productivity after one period using the moving-frontier equation. (Answer: $2/3$, or approximately 0.6667.)
A hypothetical technology produces certain additional benefits of \$12,000 at each year-end forever. The effective annual interest rate is 8%, there are no maintenance costs, and adoption costs \$140,000. Should the firm adopt under the perpetual-benefit model, and what is net present value? (Answer: Yes; present value is \$150,000 and net present value is \$10,000.)
A country starts at productivity 20 while a fixed frontier is 100. It uses the constant-gap adoption model and wants productivity to be at least 80 after three periods. What is the minimum adoption rate, to four decimal places? (Answer: $\lambda\geq1-\sqrt[3]{0.25}\approx0.3700$.)
A follower has adoption rate $\lambda=0.12$, while the frontier grows at 4% per period. Under the moving-frontier model, calculate the steady relative productivity and state the proportional gap that remains. (Answer: $q^*=0.12/(0.12+0.04)=0.75$; the follower remains 25% below the frontier.)

What You'll Learn